view src/ChangeLog @ 12029:ee264f77eaed release-3-2-x

Fix 'format rat' for values like '1 - eps'
author David Bateman <dbateman@free.fr>
date Thu, 02 Jul 2009 06:31:29 +0200
parents 2de1ae881cea
children 3335e82622ba
line wrap: on
line source

2008-07-01  David Bateman  <dbateman@free.fr>
	
	* pr-output.cc (static inline std::string rational_approx (double, 
	int)): Test for underflow of fractional part of rational approximation
	earlier in the loop.

2009-07-01  Joe Rothweiler <octaveuser@sensicomm.com>

	* input.cc (raw_mode): Use TCSADRAIN if no wait.

2009-06-30  Jaroslav Hajek  <highegg@gmail.com>

	* symtab.h (force_variable): Remove assertion.

2009-06-26  John W. Eaton  <jwe@octave.org>

	* load-path.cc (Faddpath): Preserve order of prepended elements.

2009-06-24  Thorsten Meyer  <thorsten.meyier@gmx.de>

	* oct-map.cc (Octave_map::squeeze, Octave_map::permute,
	Octave_map::transpose, Octave_map::reshape, Octave_map::concat,
	Octave_map::index): Add tests for preservation of key order in
	struct arrays.

2009-06-24  Jaroslav Hajek  <highegg@gmail.com>

	* pt-mat.cc (get_concat_class): Use empty string as zero value.
	(tm_row_const_rep::tm_row_const_rep): Initialize class_nm to empty
	string.
	(tm_const::tm_const): Ditto.

2009-06-24  John W. Eaton  <jwe@octave.org>

	* pt-assign.cc (former_built_in_variables): Remove "ans" from the list.

2009-06-24  Jaroslav Hajek  <highegg@gmail.com>

	* pt-eval.cc (do_unwind_protect_cleanup_code):
	Workaround sig_atomic_t != int.

2009-06-23  Jaroslav Hajek  <highegg@gmail.com>

	* quit.h (octave_quit_exception): Delete.
	(exit_status, quitting_gracefully): New globals.
	* quit.cc: Initialize them.
	(Fquit): Set the globals, simulate interrupt.
	(main_loop): Handle exit properly.
	* octave.cc (execute_eval_option_code): Ditto.
	(execute_command_line_file): Ditto.
	* pt-eval.cc (do_unwind_protect_cleanup_code):
	Fix order of unwind_protect calls.

2009-06-23  John W. Eaton  <jwe@octave.org>

	* oct-map.cc (Octave_map::squeeze, Octave_map::permute,
	Octave_map::transpose, Octave_map::reshape, Octave_map::concat):
	Preserve key order.

2009-06-22  Jaroslav Hajek  <highegg@gmail.com>

	* ov-cell.cc (octave_cell::all_strings): Avoid duplicate conversions.

2009-06-20  Jaroslav Hajek  <highegg@gmail.com>

	* ov.cc (Fsubsasgn): Uniquify shared value before assigning to it.

2009-06-17  John W. Eaton  <jwe@octave.org>

	* mex.cc (mxArray_octave_value::get_data):
	If octave_value::mex_get_data returns a valid pointer, mark it as
	foreign and return it instead of enumerating the types that can be
	handled directly.

	* ov-re-mat.h (octave_matrix::mex_get_data): New function.
	* ov-bool-mat.h (octave_bool_matrix::mex_get_data): Ditto.
	* ov-flt-re-mat.h (octave_float_matrix::mex_get_data): Ditto.
	* ov-intx.h (OCTAVE_VALUE_INT_MATRIX_T::mex_get_data): Ditto.
	* ov-base-mat.h (octave_base_matrix::mex_get_data): Delete.

2009-06-16  Jaroslav Hajek  <highegg@gmail.com>

	* symtab.h (symbol_table::set_scope_and_context): Avoid checking
	error_state for setting context.

2009-06-12  Kai NODA  <nodakai@gmail.com>

	* ls-mat4.h: Fix include guard
	* ov-type-conv.h: Add missing include guard.

2009-06-12  Jaroslav Hajek  <highegg@gmail.com>

	* ov-cell.cc (octave_cell::sort (Array<octave_idx_type>&)): Simplify.

2009-06-10  John W. Eaton  <jwe@octave.org>

	* pt-fcn-handle.cc (tree_anon_fcn_handle::dup): Don't convert to
	tree_constant object here.  Do inherit from current symbol table
	scope and context.  New test.

	* lex.l	(handle_identifier): Set lexer_flags.looking_for_object_index
	false if identifier is a keyword.
	({IDENT}{S}*): Don't set lexer_flags.looking_for_object_index here.

2009-06-09  Jaroslav Hajek  <highegg@gmail.com>

	* octave.cc (octave_main): Call initialize_command_input
	conditionally. Move line_editing update in front of it.
	Call command_editor::force_default_editor if not line_editing.

2009-06-09  Jaroslav Hajek  <highegg@gmail.com>

	* data.cc (Flog2): Fix tests.

2009-06-08  John W. Eaton  <jwe@octave.org>

	* variables.cc (symbol_exist): Returnn 1 for function handles and
	inline function objects.

2009-06-08  Jaroslav Hajek  <highegg@gmail.com>

	* symtab.h (symbol_table::do_clear_global,
	symbol_table::do_clear_global_pattern): Properly erase from both local
	and global table.

2009-06-08  Jaroslav Hajek  <highegg@gmail.com>

	* variables.cc (Fclear): Clear also globals when called without
	arguments.

2009-06-07  Jaroslav Hajek  <highegg@gmail.com>

	* graphics.cc (color_property::do_set): Allow a wider range of types.

2009-06-07  Jaroslav Hajek  <highegg@gmail.com>

	* DLD-FUNCTIONS/find.cc (Ffind): Fix docs. Improve second argument
	handling. Add regression tests.

2009-06-06  Rik  <rdrider0-list@yahoo.com>

	* data.cc: Update documentation for 'complex' function

2009-06-06  Rik  <rdrider0-list@yahoo.com>

	* load-save.cc: Update documentation for NA and isna functions

2009-06-06  Rik  <rdrider0-list@yahoo.com>

	* load-save.cc: Update documentation for load and save

2009-06-06  Rik  <rdrider0-list@yahoo.com>

	* pr-output.cc: Update documentation for 'format'

2009-06-05  Rik  <rdrider0-list@yahoo.com>

	* variables.cc: Update documentation for 'who' family of functions

2009-06-03  Rik  <rdrider0-list@yahoo.com>

	* input.cc: Correct documentation for keyboard function

	2009-05-25  Jaroslav Hajek  <highegg@gmail.com>

	Version 3.2.0 released.

2009-06-02  Rob Mahurin  <rob@utk.edu>

	* Makefile.in: Add CARBON_LIBS to OCTINTERP_LINK_DEPS.
	From Bernard Desgraups <bdesgraupes@orange.fr>.
	
2009-06-02  Jaroslav Hajek  <highegg@gmail.com>

	* ov-base.cc (octave_base_value::numeric_assign): Gripe when no index
	given.
	* ov-cell.cc (octave_cell::subsasgn): Ditto.
	* ov-struct.cc (octave_struct::subsasgn): Ditto.
	* ov-list.cc (octave_list::subsasgn): Ditto.

2009-05-28  John W. Eaton  <jwe@octave.org>

	* load-path.cc (load_path::do_files): Avoid shadow warning from GCC.

2009-05-27  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/__magick_read__.cc (write_image):
	Bail out if given indexed image.

2009-05-27  Jaroslav Hajek  <highegg@gmail.com>

	* DLD-FUNCTIONS/__magick_read__.cc (encode_map): Fix RGB color
	construction.

2009-05-26  John W. Eaton  <jwe@octave.org>

	* load-path.h, load-path.cc (load_path::files, load_path::do_files): 
	New arg, OMIT_EXTS.
	* help.cc (F__list_functions__): Call load_path::files with
	omit_exts set to true.

	* symtab.h
	(symbol_table::symbol_record::symbol_record_rep::is_variable):
	Use "! is_local ()" instead of storage_class != local.
	(symbol_table::do_variable_names): Only add variables to the list.
	(symbol_table::unmark_forced_variables): New static function
	* variables.cc (do_who): Use is_variable instead of is_defined.
	Also limit output to variables when using regexp pattern.
	* octave.cc (unmark_forced_vars): Delete.
	(execute_eval_option_code): Don't add unmark_forced_vars to 
	unwind_protect stack here.
	* toplev.cc (main_loop): Add symbol_table::unmark_forced_variables
	to the unwind_protect stack here.
	* input.cc (get_debug_input): Likewise.
	* parse.y (parse_fcn_file, eval_string): Likewise.

2009-05-25  Jaroslav Hajek  <highegg@gmail.com>

	* toplev.h (quit_allowed): New global variable.
	* toplev.cc (quit_allowed): Declare it.
	(Fquit): Raise error if quitting is not allowed.
	* octave.cc (octave_main): if running as embedded, disable quit by
	default.

2009-05-25  Jaroslav Hajek  <highegg@gmail.com>

	* variables.cc (do_who): Only output symbols with a defined value.

2009-05-22  Michael Goffioul  <michael.goffioul@gmail.com>

	* toplev.h (main_loop): Tag with OCTINTERP_API.
	* input.h (octave_read, get_input_from_file, get_input_from_stdin):
	Ditto.
	* lex.h (create_buffer, current_buffer, switch_to_buffer,
	delete_buffer, restore_input_buffer, delete_input_buffer): Ditto.

2009-05-22  Robert T. Short  <octave@phaselockedsystems.com>

	* ov-class.h, ov-class.cc (octave_class::clear_exemplar_map):
	New function.
	* symtab.h (symbol_record::clear_objects,
	symbol_record::do_clear_objects): New functions.
	* variables.cc (do_matlab_compatible_clear, clear):
	Handle -classes option.

2009-05-22  Jaroslav Hajek  <highegg@gmail.com>

	* src/ov-base-mat.cc (octave_base_matrix<MT>::assign (const
	octave_value_list, typename MT::element_type)): Avoid out of bounds
	assignments.

2009-05-21  Michael Goffioul  <michael.goffioul@gmail.com>

	* graphics.cc (figure::properties::get_boundingbox,
	figure::properties::set_boundingbox): Get screen size from root
	object.
	(convert_position): Get screen resolution from root object and remove
	unneeded "backend" argument.
	(axes::properties::get_boundingbox,
	figure::properties::get_boundingbox,
	figure::properties::set_boundingbox): Remove unneeded backend argument
	to convert_position call.

	* debug.h (class bp_table): Tag with OCTINTERP_API.
	* input.h (Vdebugging): Ditto.
	* pt-eval.h (class tree_evaluator): Ditto.
	* toplev.h (class octave_call_stack): Ditto.

	* file-io.cc (mkstemp): Add mktemp-based implementation of mkstemp on
	platforms that do not have it (mkstemp is required by the new help
	system).

	* TEMPLATE-INST/Array-os.cc: Add "extern template" declaration for
	Array<octave_idx_type> to avoid implicit instantiation (and duplicate
	symbols at link-time) [Win32]

2009-05-20  John W. Eaton  <jwe@octave.org>

	* pt-assign.cc (maybe_warn_former_built_in_variable):
	Set initialized to true when done initializing vars set.
	From Michael Goffioul  <michael.goffioul@gmail.com>.

2009-05-20  Jaroslav Hajek  <highegg@gmail.com>

	* ov-typeinfo.h
	(octave_value_typeinfo::unary_class_ops,
	octave_value_typeinfo::unary_ops,
	octave_value_typeinfo::non_const_unary_ops,
	octave_value_typeinfo::binary_class_ops,
	octave_value_typeinfo::binary_ops,
	octave_value_typeinfo::compound_binary_class_ops,
	octave_value_typeinfo::compound_binary_ops,
	octave_value_typeinfo::cat_ops,
	octave_value_typeinfo::assign_ops,
	octave_value_typeinfo::assignany_ops,
	octave_value_typeinfo::pref_assign_conv,
	octave_value_typeinfo::type_conv_ops,
	octave_value_typeinfo::widening_ops): Declare as Array<void *>.
	* ov-typeinfo.cc: Reflect changes.

2009-05-20  Jaroslav Hajek  <highegg@gmail.com>

	* toplev.h (octave_exit_func): New typedef.
	(octave_exit): Change to octave_exit_func.
	
2009-05-20  Jaroslav Hajek  <highegg@gmail.com>
	
	* ov-typeinfo.cc: Don't include oct-obj.h.

2009-05-20  Jaroslav Hajek  <highegg@gmail.com>

	* toplev.h (octave_quit_exception): New class.
	(octave_exit): New global variable.
	* toplev.cc (octave_exit): Initialize to ::exit.
	(clean_up_and_exit): Call octave_exit if set.
	(Fquit): Raise octave_quit_exception to quit.
	(main_loop): Catch octave_quit_exception.
	* octave.cc (execute_command_line_file): Ditto.
	(execute_eval_option_code): Ditto.

2009-05-19  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/fltk_backend.cc (F__fltk_redraw__): New function.
	(F__init_fltk__): Call add_input_event_hook with feval to add
	__fltk_redraw__ to the list of even hook functions to call.
	Call mlock when initializing.
	(F__remove_fltk__): Call remove_input_event_hook with feval to
	remove __fltk_redraw__ from the list of event hook functions.
	Unlock __init_fltk__ when shutting down.

	* input.cc (Finput_event_hook): Delete.
	(input_event_hook): Handle set of functions instead of just one.
	(Fadd_input_event_hook, Fremove_input_event_hook): New functions.

2009-05-19  Jaroslav Hajek  <highegg@gmail.com>

	* Makefile.in: Add X11_LIBS to OCTINTERP_LINK_DEPS.

2009-05-17  Jaroslav Hajek  <highegg@gmail.com>

	* TEMPLATE-INST/Array-tc.cc: Change #include.

2009-05-15  Robert T. Short  <octave@phaselockedsystems.com>

	* ov-class.cc (octave_class::exemplar_info): Fix typo in error message.
	* ov-class.cc (Fmethods): Methods returns cell array.
	* ls-mat5.cc (read_mat5_binary_element):
	Construct exemplar array and ensure inheritance is correct.

2009-05-15  John W. Eaton  <jwe@octave.org>

	* error.cc (Ferror): Handle error struct argument.

	* ls-mat5.cc (save_mat5_binary_element): Avoid multiple calls to
	contents method.  From David Bateman <dbateman@free.fr>.
	(save_mat5_element_length): Use const Cell to avoid making copies
	when indexing.

2009-05-15  Jaroslav Hajek  <highegg@gmail.com>

	* oct-stream.cc: Don't instantiate Array2<read_fptr>.
	(octave_stream::read): Make read_fptr_table a static 2d array.
	(FILL_TABLE_ROW): Update.

2009-05-12  Jaroslav Hajek  <highegg@gmail.com>

	* data.cc (F__accumarray_sum__): Optimize the constant range case.

2009-05-11  John W. Eaton  <jwe@octave.org>

	* sparse-xdiv.cc (do_rightdiv_sm_dm, do_leftdiv_dm_sm):
	Avoid apparent MSVC bug with typedef.

	* Makefile.in (install-lib): Remove
	$(DESTDIR)$(octlibdir)/$(SHLLIBPRE)octinterp.$(SHLLIB_VER), not
	$(DESTDIR)$(octlibdir)/$(SHLPRE)octinterp.$(SHLEXT_VER).

2009-05-08  Carsten Clark <tantumquantum+gnuoctave@gmail.com>

	* data.cc (Fissorted): Fix typo in documentation entry.

2009-05-07  Robert T. Short  <octave@phaselockedsystems.com>

	* ov-class.h, ov-class.cc (octave_class::reconstruct_exemplar):
	New function.
	* ov-class.cc (octave_class::load_binary, octave_class::load_hdf5,
	octave_class::load_ascii): Construct exemplar table and ensure
	inheritance is correct.
        * ov-struct.cc (struct): Return struct from object.

2009-05-07  John W. Eaton  <jwe@octave.org>

	* graphics.h.in (base_graphics_object::set): Undo previous change.
	(base_properties::set (const caseless_str&, const octave_value&):
	New virtual function.
	(base_properties::set (const caseless_str&, const std::string&,
	const octave_value&)): No longer virtual.

	* toplev.cc (octave_config_info): Remove F2C and F2CFLAGS from the
	config infor struct.
	* oct-conf.h.in (OCTAVE_CONF_F2CFLAGS, OCTAVE_CONF_F2C):
	Delete definitions.

2009-05-07  Marco Atzeri  <marco_atzeri@yahoo.it>

        * Makefile.in: (SHLPRE): Rename from SHLLIBPRE.

2009-05-07  John W. Eaton  <jwe@octave.org>

	* genprops.awk (emit_declarations): Emit decls for static
	has_property functions.
	(emit_source): Add class name argument to base_properties::set
	function.  Pass class name to set_dynamic and
	base_properties::set.  Emit definitions for has_property
	functions.
	* graphics.h.in, graphics.cc
	(base_properties::set, base_properties::set_dynamic):
	New argument CNAME.
	(base_properties::all_dynamic_properties): New static data member.
	(base_properties::has_dynamic_property): New static function.
	(base_graphics_object::set): Pass class name to
	base_properties::set function.
	(property_list::set): Check for invalid property names.

2009-05-06  Jaroslav Hajek  <highegg@gmail.com>

	* ov-re-mat.cc (Fdouble): Fix order of branches.

2009-05-05  John Swensen  <jpswensen@comcast.net>

	* debug.h, debug.cc (breakpoints): Rename from bp_map, use a
	std::set instead of a std::map object.  Change all uses.
	(bp_table::do_get_breakpoint_list): Simplify.

2009-05-05  Robert T. Short  <octave@phaselockedsystems.com>

	* ov-class.h, ov-class.cc (octave_class::reconstruct_parents):
	New function.
	* ov-class.cc (octave_class::load_binary, octave_class::load_hdf5):
	Contstruct parent list.
	(get_current_method_class): Clean up method class extraction.

2009-05-05  John W. Eaton  <jwe@octave.org>

	* graphics.cc (array_property::validate): Require object to be any
	numeric type, not necessarily a double precision value.

	* variables.cc (set_internal_variable): Pass NM in call to error.

2009-05-04  Peter O'Gorman  <pogma@thewrittenword.com>

	* utils.cc: Don't define HAVE_C99_VSNPRINTF here.

2009-05-01  John W. Eaton  <jwe@octave.org>

	* error.cc (Vlast_error_file, Vlast_error_name, Vlast_error_line,
	Vlast_error_column): Delete.
	(Vlast_error_stack): New static variable.
	(initialize_last_error_stack): New static function.
	(verror, Frethrow, Flasterror): Set or use Vlast_error_stack, not
	Vlast_error_file, etc.

2009-04-27  John W. Eaton  <jwe@octave.org>

	* ov-class.cc (octave_class::dotref): Handle empty parent class.
	(octave_class::get_current_method_class): Allow result to be empty.
	(called_from_builtin): New static function.
	(octave_class::subsref, octave_class::subsasgn): Use it.

2009-04-23  John W. Eaton  <jwe@octave.org>

	* ov-class.cc (Fclass): Check newly constructed classes against
	the first constructed object of the class.

	* ov-class.h, ov-class.cc (octave_class::exmplar_info): New class.
	(exemplar_map): New static data member.
	(exemplar_iterator, exemplar_const_iterator): New typedefs.

	* ov-class.h (octave_class::nparents,
	octave_class::parent_class_name_list): New functions.
	* ov.h (octave_base_value::nparents,
	octave_base_value::parent_class_name_list): New functions.
	* ov-base.h, ov-base.cc (octave_base_value::nparents,
	octave_base_value::parent_class_name_list): New functions.
	(parent_class_names): Error if called for wrong type argument.

	* symtab.cc (load_out_of_date_fcn): New arg, dispatch_type.
	(out_of_date_check_internal): Pass dispatch type to
	load_out_of_date_fcn.

2009-04-22  John W. Eaton  <jwe@octave.org>

	* ov-base-int.cc (octave_base_int_helper<T, false,
	false>::char_value_out_of_range): Correct result for specialization.

	* ov-class.cc (octave_class::dotref, octave_class::subsasgn):
	Protect against possibly invalid octave_value -> string conversions.

2009-04-22  Robert T. Short  <octave@phaselockedsystems.com>

	* variables.cc (symbol_exist): Also return 1 for objects.
	* ov-base.h (octave_base_value::assign): New virtual function.
	* ov-class.h (octave_class::assign): New function.
	* ov-class.cc (octave_class::find_parent_class): Simplify.
	(octave_class::octave_class): Don't allow duplicate parent classes.
	(octave_class::subsasgn): Allow cls = method (cls, value) to work
	properly when method is a parent-class method.
	(get_method_class): New static function.
	(octave_class:subsasgn, octave_class::dotref): Use it.
	(F__isa_parent__): New function.

2009-04-22  Jaroslav Hajek  <highegg@gmail.com>

	* ov-range.cc (octave_range::char_array_value): New virtual function
	override.
	* ov-range.h: Declare it.

2009-04-22  Jaroslav Hajek  <highegg@gmail.com>

	* pt-mat.cc (tm_row_const::tm_row_const_rep::do_init_element):
	Update class name even for all-zeros elements.
	(get_concat_class): Update the default value if possible.

2009-04-21  John W. Eaton  <jwe@octave.org>

	* parse.y (Fassignin): Add missing unwind_protect frame.

	* toplev.h (push (symbol_table::scope_id, symbol_table::context_id)):
	New function.

	* toplev.cc (main_loop): Don't call symbol_table::reset_scope.

	* mex.cc (mexGetVariable, mexPutVariable): Use unwind_protect to
	restore call stack and scope.

	* ov-usr-fcn.cc (octave_user_function::do_multi_index_op):
	Don't use symbol_table::push_scope.
	* variablees.cc (do_who): Likewise.  Use octave_call_stack and
	unwind_protect to manage change in scope.
	* ov-fcn-handle.cc (octave_fcn_handle::load_ascii,
	octave_fcn_handle::load_binary, octave_fcn_handle::load_hdf5):
	Likewise.
	* ls-mat5.cc (read_mat5_binary_element): Likewise.

	* symtab.h (erase_scope (void*)): New function, for unwind_protect.
	(symbol_table::push_scope, symbol_table::pop_scope,
	symbol_table::reset_scope): Delete.
	(symbol_table::scope_stack): Delete static member.
	* symtab.cc (symbol_table::scope_stack): Delete definition.

2009-04-17  Jaroslav Hajek  <highegg@gmail.com>

	* oct-map.h (Octave_map::contents (const_iterator) const,
	Octave_map::contents (iterator)): Simplify.

2009-04-17  Jaroslav Hajek  <highegg@gmail.com>

	* oct-map.cc (Octave_map::assign (const octave_value_list&, const
	std::string&, const Cell&)): Fix & simplify.
	(common_size): Remove.

2009-04-16  Jaroslav Hajek  <highegg@gmail.com>

	* pt-idx.cc (make_value_list): Gripe on magic end query for undefined
	variables.

2009-04-16  Jaroslav Hajek  <highegg@gmail.com>

	* xpow.cc (same_sign): New helper function.
	(elem_xpow (double, const Range&), elem_xpow (const Complex&, const Range&)): 
	Only optimize monotonic-magnitude integer ranges, start always from
	the smaller end.

2008-04-11  David Bateman  <dbateman@free.fr>

	* ov-cell.cc (Fstruct2cell): Treat possible trailing singleton for
	creation of cell array from column structure vector.

2009-04-09  Jaroslav Hajek  <highegg@gmail.com>

	* ov-cell.cc (octave_cell::subsasgn): Fix reference counting
	optimization.

2009-04-08  John W. Eaton  <jwe@octave.org>

	* load-path.cc (rehash_internal): New function.
	(Frehash): Use it.
	(Fpath, Faddpath, Frmpath): Call rehash_internal if path is modified.

2009-04-08  Jaroslav Hajek  <highegg@gmail.com>

	* xpow.cc (elem_xpow (double, const Range&),
	elem_xpow (const Complex&, const Range&)): New functions.
	* xpow.h: Declare them.
	* OPERATORS/op-range.cc: Define scalar .^ range and complex .^ range &
	install them.

2009-04-04  Jaroslav Hajek  <highegg@gmail.com>

	* ov-struct.cc (octave_struct::subsasgn): Fix reference counting
	optimization.

2008-04-03  David Bateman  <dbateman@free.fr>

	* DLD-FUNCTIONS/max.cc (MINMAX_SPARSE_BODY): Allow sparse boolean 
	values.

2009-04-01  Jaroslav Hajek  <highegg@gmail.com>

	* ov-str-mat.cc (default_numeric_conversion_function):
	Create an octave_scalar if possible.

2009-03-29  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/fltk_backend.cc (plot_window::button_press):
	Don't pass arbitrary input to fl_message as a format string.

2009-03-29  Jaroslav Hajek  <highegg@gmail.com>

	* pt-eval.cc (do_unwind_protect_cleanup_code): Protect also
	octave_interrupt_state.

2009-03-29  Jaroslav Hajek  <highegg@gmail.com>

	* DLD-FUNCTIONS/cellfun.cc (Fcellslices): Index n-d arrays along the
	last dimension.

2009-03-27  Jaroslav Hajek  <highegg@gmail.com>

	* DLD-FUNCTIONS/balance.cc (Fbalance): Fix order of output args.

2009-03-26  Jaroslav Hajek  <highegg@gmail.com>

	* DLD-FUNCTIONS/find.cc 
	(find_nonzero_elem_idx (const Array<T>&, ...)): Move dimensions
	fixup to liboctave.

2009-03-26  Jaroslav Hajek  <highegg@gmail.com>

	* DLD-FUNCTIONS/find.cc 
	(find_nonzero_elem_idx (const Array<T>&, ...)): Simplify.
	Instantiate for bool and octave_int types.
	(find_nonzero_elem_idx (const Sparse<T>&, ...)): 
	Instantiate for bool.
	(Ffind): Handle bool and octave_int cases.

2009-03-25  John W. Eaton  <jwe@octave.org>

	* version.h (OCTAVE_VERSION): Now 3.1.55+.
	(OCTAVE_API_VERSION): Now api-v37+.

	* version.h (OCTAVE_VERSION): Now 3.1.55.
	(OCTAVE_API_VERSION): Now api-v37.
	(OCTAVE_RELEASE_DATE): Now 2009-03-25.

	* DLD-FUNCTIONS/find.cc (find_nonzero_elem_idx): Also return
	[](0x0) if the array has 0 rows and it is not a column vector.

	* oct-stream.cc (octave_stream::write (const Array<T>&,
	octave_idx_type, oct_data_conv::data_type, octave_idx_type,
	oct_mach_info::float_format)): Seek to skip if still inside bounds
	of existing file.  Otherwise, write NUL to skip.

	* Makefile.in (%.df : %.cc): Write source file name to output,
	wrapped in XDEFUN_FILE_NAME macro.
	* mkbuiltins: Provide definition for XDEFUN_FILE_NAME.
	* mkgendoc: Likeiwse.
	(XDEFUN_DLD_INTERNAL, XDEFUNX_DLD_INTERNAL, XDEFUN_INTERNAL,
	XDEFCONSTFUN_INTERNAL, XDEFUNX_INTERNAL, XDEFVAR_INTERNAL,
	XDEFCONST_INTERNAL): Pass file_name to print_doc_string.
	(print_doc_string): New arg, FILE_NAME.  Print file name as
	comment.

2009-03-24  John W. Eaton  <jwe@octave.org>

	* ov-class.cc (F__parent_classes__): New function.

2009-03-24  Robert T. Short  <octave@phaselockedsystems.com>

	* ov-class.h, ov-class.cc (octave_class::octave_class (const
	Octave_map&, const std::string&, const octave_value_list&)):
	New constructor.
	(octave_class::find_parent_class, octave_class::parent_classes):
	New functions.
	(octave_class::dotref): Also look in parent class.
	(Fclass): Handle parent class arguments.

	* ov-base.h (octave_base_value::get_parent_list,
	octave_base_value::parent_classes): New virtual functions.

	* load-path.h, load-path.cc (load_path::parent_map): New data member. 
	(load_path::add_to_parent_map): New static function.
	(load_path::do_add_to_parent_map): New member function.
	(load_path::do_find_method): Also look in parent classes for methods.
	(load_path::parent_map_type, load_path::const_parent_map_iterator,
	load_path::parent_map_iterator): New typedefs.

2009-03-23  John W. Eaton  <jwe@octave.org>

	* symtab.h
	(symbol_table::symbol_record::symobl_recoord_rep::is_variable):
	Also return true if symbol is tagged as a variable.
	(symbol_table::symbol_record::symobl_recoord_rep::force_variable):
	Don't set variable value.
	(symbol_table::symbol_record::symobl_recoord_rep::clear_forced,
	symbol_table::symbol_record::clear_forced): Delete.
	(symbol_table::unmark_forced_variables): Rename from
	symbol_table::clear_forced_variables.
	(symbol_table::do_unmark_forced_variables): Rename from
	symbol_table::do_clear_forced_variables.
	* parse.y (make_script, finish_function): Call
	symbol_table::unmark_forced_variables instead of
	symbol_table::clear_forced_variables.
	* octave.cc (unmark_forced_vars): New function.
	(execute_eval_option_code): Add it to the unwind-protect stack.

2009-03-22  Jaroslav Hajek  <highegg@gmail.com>

	* pt-eval.cc (tree_evaluator::visit_simple_for_command):
	Remove struct branch, handle structs by the generic code.
	(tree_evaluator::visit_complex_for_command):
	Add missing const qualifiers.

2009-03-21  Jaroslav Hajek  <highegg@gmail.com>

	* ov-base-diag.cc: Add missing include.
	* sparse-xdiv.cc: Ditto.
	* DLD-FUNCTIONS/__glpk__.cc: Ditto.
	* DLD-FUNCTIONS/__lin_interpn__.cc: Ditto.
	* DLD-FUNCTIONS/__voronoi__.cc: Ditto.

2009-03-20  Jaroslav Hajek  <highegg@gmail.com>

	* ov-re-mat.cc (octave_matrix::load_ascii): Simplify.
	* ov-flt-re-mat.cc (octave_float_matrix::load_ascii): Simplify.
	* ov-cx-mat.cc (octave_complex_matrix::load_ascii): Simplify.
	* ov-flt-cx-mat.cc (octave_float_complex_matrix::load_ascii): Simplify.

2009-03-20  Jaroslav Hajek  <highegg@gmail.com>

	* ov-re-mat.cc (octave_matrix::isnan, octave_matrix::isinf,
	octave_matrix::finite): Simplify.
	* ov-flt-re-mat.cc (octave_float_matrix::isnan,
	octave_float_matrix::isinf, octave_float_matrix::finite): Simplify.
	* ov-cx-mat.cc (octave_complex_matrix::isnan,
	octave_complex_matrix::isinf, octave_complex_matrix::finite):
	Simplify.
	* ov-flt-cx-mat.cc (octave_float_complex_matrix::isnan,
	octave_float_complex_matrix::isinf,
	octave_float_complex_matrix::finite): Simplify.

2009-03-19  Benjamin Lindner  <lindnerb@users.sourceforge.net>

	* ls-oct-ascii.cc (extract_keyword): Replace loop with call to
	read_until_newline to avoid leaving stray '\r' in stream when
	reading files with CRLF line endings.

2009-03-18  Jaroslav Hajek <highegg@gmail.com>

	* DLD-FUNCTIONS/cellfun.cc
	(scalar_col_helper, scalar_col_helper_def, scalar_col_helper_nda,
	scalar_query_helper): New classes.
	(make_col_helper): New function.
	(Fcellfun): Use col_helper classes for UniformOutput. Avoid copying by
	using const variables.

2009-03-17  Jaroslav Hajek  <highegg@gmail.com>

	* OPERATORS/op-b-bm.cc: Add missing INSTALL_BINOPs.
	* OPERATORS/op-bm-b.cc: Ditto.
	* OPERATORS/op-fm-fm.cc: Ditto.
	* OPERATORS/op-m-m.cc: Ditto.

2009-03-17  Jaroslav Hajek  <highegg@gmail.com>

	* ov.cc (octave_value::octave_value): Move to ov.h
	* ov.h (octave_value::octave_value): Implement fast inline constructor
	using a static rep.

2009-03-15  Jaroslav Hajek  <highegg@gmail.com>

	* DLD-FUNCTIONS/cellfun.cc (Fcellslices): New DLD function.

2009-03-13  Jaroslav Hajek  <highegg@gmail.com>

	* ov.h (octave_value::compound_binary_op): Support bool compound ops.
	* ov.cc (do_binary_op, decompose_binary_op, binary_op_fcn_name):
	Ditto.
	* pt-cbinop.cc (strip_not, simplify_and_or_op): New funcs.
	(maybe_compound_binary_expression): Support bool compound ops.
	* OPERATORS/op-int.h: Support bool compound ops.
	* OPERATORS/op-b-bm.cc: Ditto.
	* OPERATORS/op-bm-b.cc: Ditto.
	* OPERATORS/op-bm-bm.cc: Ditto.
	* OPERATORS/op-fm-fm.cc: Ditto.
	* OPERATORS/op-m-m.cc: Ditto.

2009-03-13  Jaroslav Hajek  <highegg@gmail.com>

	* xpow.cc (xpow (const NDArray&, double), xpow (const ComplexNDArray&, double),
	xpow (const FloatNDArray&, float), xpow (const FloatComplexNDArray&,
	float)): Use xisint for testing ints. Optimize w.r.t int exponents.

2009-03-13  John W. Eaton  <jwe@octave.org>

	* parse.y (maybe_warn_assign_as_truth_value, make_binary_op,
	maybe_warn_variable_switch_label, maybe_warn_associativity_change): 
	Print file and line number info if available.
	* lex.l (gripe_matlab_incompatible): Likewise.

	* error.cc (pr_where): Use octave_call_stack::backtrace to print
	complete stack trace at once.  Don't attempt to print code.
	(error_2): Set error_state to 0 before calling pr_where.
	(warning_1): Switch sense of test on symbol_table::at_top_level.
	Call pr_where after printing primary warning message.

2009-03-13  Jaroslav Hajek  <highegg@gmail.com>

	* ov-range.h (octave_range::octave_range (const Range&)): Allow
	constructing from invalid range op result.
	* ov-range.cc (octave_range::try_narrowing_conversion): Validate
	invalid range op results.
	* data.cc (fill_matrix): Return packed form (zero-step range) if
	possible.

2009-03-10  Jason Riedy  <jason@acm.org>

	* DLD-FUNCTIONS/lu.cc (lu): Call fact.Pr_mat () and fact.Pc_mat ()
	to return permutation matrices in the sparse case.

2009-03-12  John W. Eaton  <jwe@octave.org>

	* ls-mat-ascii.cc (get_mat_data_input_line): If we are looking at
	'\r' or '\n', skip current line ending instead of skipping until
	the next.

2009-03-12  Ben Abbott <bpabbott@mac.com>

	* graphics.cc: Fix default "papersize" property value.

2009-03-12  Jaroslav Hajek  <highegg@gmail.com>

	* ov-flt-perm.h, ov-flt-perm.cc: Remove sources.
	* ov.h, ov.cc (octave_value::octave_value (const PermMatrix&)): Remove
	use "single" argument.
	* ov.cc (install_types): Update.
	* xpow.cc, xpow.h (xpow (PermMatrix, float)): Remove.
	* ov-base-diag.cc (octave_base_diag::do_index_op): Remove permutation
	creating block.
	* ov-re-diag.cc (octave_diag_matrix::do_index_op): Move it here.
	* ov-re-diag.h (octave_diag_matrix::do_index_op): New decl.

	* DLD-FUNCTIONS/det.cc (Fdet): Update.
	* DLD-FUNCTIONS/inv.cc (Finv): Update.
	* DLD-FUNCTIONS/pinv.cc (Fpinv): Update.

	* OPERATORS/op-fcm-pm.cc <-- OPERATORS/op-fcm-fpm.cc
	* OPERATORS/op-fcm-pm.cc: Update.
	* OPERATORS/op-fm-pm.cc <-- OPERATORS/op-fm-fpm.cc
	* OPERATORS/op-fm-pm.cc: Update.
	* OPERATORS/op-pm-fcm.cc <-- OPERATORS/op-fpm-fcm.cc
	* OPERATORS/op-pm-fcm.cc: Update.
	* OPERATORS/op-pm-fm.cc <-- OPERATORS/op-fpm-fm.cc
	* OPERATORS/op-pm-fm.cc: Update.
	* OPERATORS/op-pm-pm.cc: Update.
	* OPERATORS/op-pm-template.cc: Update.

2009-03-11  Jaroslav Hajek  <highegg@gmail.com>
	
	* xpow.cc (xpow (const PermMatrix&, double), xpow (const PermMatrix&,
	float)): New functions.
	* xpow.h: Declare them.
	* DLD-FUNCTIONS/op-pm-pm.cc: Support permutation matrix ^ scalar.
	* DLD-FUNCTIONS/op-fpm-fpm.cc: Ditto.

2009-03-11  Jaroslav Hajek  <highegg@gmail.com>

	* ov-ch-mat.cc (octave_char_matrix::double_value, 
	octave_char_matrix::float_value, octave_char_matrix::complex_value,
	octave_char_matrix::float_complex_value): Cast to unsigned char.

2009-03-10  Jason Riedy  <jason@acm.org>

	* OPERATORS/op-pm-sm.cc (mul_pm_sm): New Octave binding for
	perm * sparse.
	(ldiv_pm_sm): New Octave binding for perm \ sparse.
	(mul_sm_pm): New Octave binding for sparse * perm.
	(div_sm_pm): New Octave binding for sparse / perm.
	(install_pm_sm_ops): Install the above bindings.

	* OPERATORS/op-pm-scm.cc (mul_pm_scm): New Octave binding for
	perm * sparse complex.
	(ldiv_pm_scm): New Octave binding for perm \ sparse complex.
	(mul_scm_pm): New Octave binding for sparse complex * perm.
	(div_scm_pm): New Octave binding for sparse complex / perm.
	(install_pm_scm_ops): Install the above bindings.

	* Makefile.in (PERM_OP_XSRC): Add op-pm-sm.cc and op-pm-scm.cc for
	operations between permutations and sparse matrices.

2009-03-10  Jason Riedy  <jason@acm.org>

	* DLD-FUNCTIONS/find.cc (find_nonzero_elem_idx): New override
	for find on PermMatrix.
	(find): Add a branch testing arg.is_perm_matrix () and calling the
	above override.

2009-03-10  John W. Eaton  <jwe@octave.org>

	* c-file-ptr-stream.cc, dynamic-ld.cc, error.cc, lex.l, pager.cc,
	zfstream.cc: Include <iostream>.
	* zfstream.h: Include <iosfwd> instead of <istream> and <ostream>.
	* c-file-ptr-stream.h, help.h, load-path.h, load-save.h,
	ls-ascii-helper.h, oct-iostrm.h, oct-stream.h, ov-base-diag.h,
	ov-base-int.h, ov-base-mat.h, ov-base-scalar.h, ov-base-sparse.h,
	ov-base.h, ov-bool-mat.h, ov-bool-sparse.h, ov-bool.h, ov-cell.h,
	ov-ch-mat.h, ov-class.h, ov-colon.h, ov-complex.h, ov-cs-list.h,
	ov-cx-mat.h, ov-cx-sparse.h, ov-fcn-handle.h, ov-fcn-inline.h,
	ov-float.h, ov-flt-complex.h, ov-flt-cx-mat.h, ov-flt-re-mat.h,
	ov-intx.h, ov-list.h, ov-range.h, ov-re-mat.h, ov-re-sparse.h,
	ov-scalar.h, ov-str-mat.h, ov-struct.h, ov.h, pager.h,
	pr-output.h, procstream.h, pt-assign.h, pt-cell.h, pt-const.h,
	pt-fcn-handle.h, pt-id.h, pt-mat.h, pt.h, utils.h:
	Include <iosfwd> instead of <iosstream>.

2009-03-10  Jaroslav Hajek  <highegg@gmail.com>

	* data.cc (Fsize_equal): Allow single argument call.
	* DLD-FUNCTIONS/max.cc (Fcummin, Fcummax): Update docs.

2009-03-09  Benjamin Lindner  <lindnerb@users.sourceforge.net>

	* ls-ascii-helper.h, ls-ascii-helper.cc: New files.
	* Makefile.in: Add them to the appropriate lists.
	* load-save.cc (Fload): Open all files in binary mode.
	* ov-range.cc (load_ascii): Explicitly handle CR and CRLF line endings.
	* ov-fcn-handle.cc (load_ascii): Likewise.
	* ov-fcn-inline.cc (load_ascii): Likewise.
	* ov-str-mat.cc (load_ascii): Likewise.
	* ls-mat-ascii.cc (get_mat_data_input_line): Likewise.
	* ls-oct-ascii.cc (extract_keyword, read_ascii_data): Likewise.
	* ls-oct-ascii.h (extract_keyword): Likewise.
	
2009-03-09  John W. Eaton  <jwe@octave.org>

	* graphics.h.in (OCTAVE_DEFAULT_FONTNAME): New macro, defaults to "*".
	(axes::properties, text::properties): Use it to set default fontname.
	* graphics.cc (axes::properties::set_defaults): Likewise.

2009-03-09  Rafael Laboissiere  <rafael@debian.org>

	* Makefile.in (maintainer-clean): Remove y.tab.h here.
	(distclean): Not here.

2009-03-09  Jason Riedy  <jason@acm.org>

	* OPERATORS/op-dm-sm.cc (add_dm_sm): Octave binding for diag + sparse.
	(sub_dm_sm): Octave binding for diag - sparse.
	(add_sm_dm): Octave binding for diag + sparse.
	(sub_sm_dm): Octave binding for diag - sparse.
	(install_dm_sm_ops): Install above bindings.

	* OPERATORS/op-dm-scm.cc (add_cdm_sm): Octave binding for diag + sparse.
	(add_dm_scm): Octave binding for diag + sparse.
	(add_cdm_scm): Octave binding for diag + sparse.
	(sub_cdm_sm): Octave binding for diag - sparse.
	(sub_dm_scm): Octave binding for diag - sparse.
	(sub_cdm_csm): Octave binding for diag - sparse.
	(add_sm_cdm): Octave binding for diag + sparse.
	(add_scm_dm): Octave binding for diag + sparse.
	(add_scm_cdm): Octave binding for diag + sparse.
	(sub_sm_cdm): Octave binding for diag - sparse.
	(sub_scm_dm): Octave binding for diag - sparse.
	(sub_scm_cdm): Octave binding for diag - sparse.
	(install_dm_scm_ops): Install above bindings.

2009-03-08  Jason Riedy  <jason@acm.org>

	* sparse-xdiv.h (xleftdiv): Declare overrides for
	xleftdiv (diagonal, sparse), both real and complex.
	(xdiv): Declare overrides for xdiv (sparse, diagonal), both real
	and complex.

	* sparse-xdiv.cc (do_rightdiv_sm_dm): New template function.
	Implementation for xdiv (sparse, diagonal).
	(xdiv): Call do_rightdiv_sm_dm to implement overrides, real and
	complex.
	(do_leftdiv_dm_sm): New template function.  Implementation for
	xleftdiv (diagonal, sparse).
	(xleftdiv): Call do_leftdiv_dm_sm to implement overrides, real and
	complex.

	* OPERATORS/op-dm-sm.cc (ldiv_dm_sm): Octave binding for real left
	division, diagonal into sparse.
	(div_sm_dm): Octave binding for real right division, sparse by
	diagonal.
	(install_dm_sm_ops): Install above bindings.

	* OPERATORS/op-dm-scm.cc (ldiv_dm_scm): Octave binding for real
	diagonal \ complex sparse.
	(ldiv_cdm_sm): Octave binding for complex diagonal \ real sparse.
	(ldiv_cdm_scm): Octave binding for complex diagonal \ complex sparse.
	(div_scm_dm): Octave binding for complex sparse / real diagonal.
	(div_sm_cdm): Octave binding for real sparse / complex diagonal.
	(div_scm_cdm): Octave binding for complex sparse / complex diagonal.
	(install_dm_scm_ops): Install above bindings.

2009-03-08  Jason Riedy  <jason@acm.org>

	* OPERATORS/op-dm-scm.cc: New file.  Implement multiplication
	between diagonal matrices (both real and complex) and complex
	sparse matrices.
	* OPERATORS/op-dm-sm.cc: New file.  Implement multiplication
	between diagonal matrices and sparse matrices, all real.
	* Makefile.in (DIAG_OP_XSRC): Add op-dm-sm.cc and op-dm-scm.cc.

2009-03-09  Jaroslav Hajek  <highegg@gmail.com>

	* data.cc (F__accumarray_sum__): New function.
	(do_accumarray_sum): New helper template function.

2009-03-07  Jaroslav Hajek  <highegg@gmail.com>

	* xdiv.cc (mdm_div_impl, dmm_lelftdiv_impl, dmdm_div_impl,
	dmdm_leftdiv_impl): Optimize.

2009-03-07  John W. Eaton  <jwe@octave.org>

	* pr-output.cc (octave_print_internal (std::ostream&,
	const PermMatrix&, bool, int)): Delete unused variable SCALE.

	* utils.cc (octave_vsnprintf): Avoid uninitialized variable
	warning from GCC.

	* DLD-FUNCTIONS/qz.cc (Fqz): Avoid "maybe clobbered by vfork"
	warning from GCC.

	* version.h (OCTAVE_VERSION): Now 3.1.54.
	(OCTAVE_API_VERSION): Now api-v36.
	(OCTAVE_RELEASE_DATE): Now 2009-03-07.

	* octave.cc (verbose_usage): Include --doc-cache-file in option list.

2009-03-06  Jaroslav Hajek  <highegg@gmail.com>

	* ov.h (octave_value::diag_matrix_value,
	octave_value::complex_diag_matrix_value,
	octave_value::float_diag_matrix_value,
	octave_value::float_complex_diag_matrix_value,
	octave_value::perm_matrix_value): New methods.

	* ov-base.cc (octave_base_value::diag_matrix_value,
	octave_base_value::complex_diag_matrix_value,
	octave_base_value::float_diag_matrix_value,
	octave_base_value::float_complex_diag_matrix_value,
	octave_base_value::perm_matrix_value): New virtual methods.
	
	* ov-base.h: Declare them.

	* ov-base-diag.h (octave_base_diag::diag_matrix_value,
	octave_base_diag::complex_diag_matrix_value,
	octave_base_diag::float_diag_matrix_value,
	octave_base_diag::float_complex_diag_matrix_value,
	octave_base_diag::perm_matrix_value): Remove declarations.

	* DLD-FUNCTIONS/inv.cc (Finv): Simplify handling diag & perm matrices.
	* DLD-FUNCTIONS/pinv.cc (Fpinv): Ditto.
	* DLD-FUNCTIONS/det.cc (Fdet): Ditto.

2009-03-05  Jaroslav Hajek  <highegg@gmail.com>
	
	* ls-hdf5.cc (add_hdf5_data): Handle diag & perm matrices.

2009-03-05  John W. Eaton  <jwe@octave.org>

	* pt-cell.cc, pt-cell.h (tree_cell::dup): Now const.
	* comment-list.cc, comment-list.h (octave_comment_list::dup): Ditto.
	* pt-arg-list.cc, pt-arg-list.h (tree_argument_list::dup): Ditto.
	* pt-assign.cc, pt-assign.h (tree_simple_assignment::dup,
	tree_multi_assignment::dup): Ditto.
	* pt-binop.cc, pt-binop.h (tree_binary_expression::dup,
	tree_boolean_expression::dup): Ditto.
	* pt-cmd.cc, pt-cmd.h (tree_no_op_command::dup,
	tree_function_def::dup): Ditto.
	* pt-colon.cc, pt-colon.h (tree_colon_expression::dup): Ditto.
	* pt-const.cc, pt-const.h (tree_constant::dup): Ditto.
	* pt-decl.cc, pt-decl.h (tree_decl_elt::dup, tree_decl_init_list::dup,
	tree_global_command::dup, tree_static_command::dup): Ditto.
	* pt-except.cc, pt-except.h (tree_try_catch_command::dup,
	tree_unwind_protect_command::dup): Ditto.
	* pt-fcn-handle.cc, pt-fcn-handle.h (tree_fcn_handle::dup,
	tree_anon_fcn_handle::dup): Ditto.
	* pt-id.cc, pt-id.h (tree_identifier::dup): Ditto.
	* pt-idx.cc, pt-idx.h (tree_index_expression::dup): Ditto.
	* pt-jump.cc, pt-jump.h (tree_break_command::dup,
	tree_continue_command::dup, tree_return_command::dup): Ditto.
	* pt-loop.cc, pt-loop.h (tree_while_command::dup,
	tree_do_until_command::dup, tree_simple_for_command::dup,
	tree_complex_for_command::dup): Ditto.
	* pt-mat.cc, pt-mat.h (tree_matrix::dup): Ditto.
	* pt-misc.cc, pt-misc.h (tree_parameter_list::dup,
	tree_return_list::dup): Ditto.
	* pt-select.cc, pt-select.h (tree_if_clause::dup,
	tree_if_command_list::dup, tree_if_command::dup,
	tree_switch_case::dup, tree_switch_case_list::dup,
	tree_switch_command::dup): Ditto.
	* pt-stmt.cc, pt-stmt.h (tree_statement::dup,
	tree_statement_list::dup): Ditto.
	* pt-unop.cc, pt-unop.h (tree_prefix_expression::dup,
	tree_postfix_expression::dup): Ditto.
	* pt-fcn-handle.h (tree_anon_fcn_handle::parameter_list,
	tree_anon_fcn_handle::return_list, tree_anon_fcn_handle::body,
	tree_anon_fcn_handle::scope): Ditto.

2009-03-05  Jason Riedy  <jason@acm.org>

	* ov-base-int.cc (convert_to_str_internal): Replace elt_type with
	element_type throughout.

2009-03-05  Jaroslav Hajek  <highegg@gmail.com>

	* DLD-FUNCTIONS/sparse.cc (Fsparse): Handle diagonal and permutation
	matrices.

2009-03-03  John W. Eaton  <jwe@octave.org>

	* ov-struct.cc (octave_struct::save_ascii,
	octave_struct::save_binary, octave_struct::save_hdf5):
	Preserve order of structure fields.
	* ls-mat5.cc (save_mat5_binary_element): Likewise.

	* symtab.h (symbol_table::do_inherit): Only inherit values for
	symbols from the donor_scope that already exist in the table.
	(symbol_table::symbol_record::symbol_record_rep::dup): Now const.
	(symbol_table::symbol_record::operator=): Decrement rep->count and
	maybe delete rep.
	(symbol_table::fcn_info::operator=): Likewise.

	* pt-fcn-handle.cc: (tree_anon_fcn_handle::dup): Transform
	tree_anon_fcn_handle objects to tree_constant objects containing
	octave_fcn_handle objects.  New tests.

	* pt-assign.cc (tree_simple_assignment::rvalue1): Assign result of
	call to rhs->rvalue1() to an octave_value object, not an
	octave_value_list object.

2009-03-03  Jaroslav Hajek  <highegg@gmail.com>

	* oct-stream.h (octave_stream_list::lookup_cache): New member field.
	(octave_stream_list::octave_stream_list): Initialize it.
	(octave_stream_list::do_lookup): Use it.
	(octave_stream_list::clear): Make flush optional. Do physically erase
	entries from the map. Close files.
	(octave_stream_list::do_remove): Call clear on "all". Do erase deleted
	entry from the map.

2009-03-02  John W. Eaton  <jwe@octave.org>

	* graphics.cc (Fget, F__get__): Return a column vector of property
	values, not a row vector.

2009-03-01  John W. Eaton  <jwe@octave.org>

	* OPERATORS/op-fcm-fcm.cc (DEFNDASSIGNOP_FN (dbl_assign)):
	LHS type is float_complex_matrix, not complex_matrix.  RHS value
	function is float_complex_array, not complex_array.

2009-03-01  Jaroslav Hajek  <highegg@gmail.com>

	* ov-perm.cc (octave_perm_matrix::print_raw): Call
	octave_print_internal.
	(octave_perm_matrix::print_raw): Call print_raw.
	* pr-output.cc (octave_print_internal (...,const DiagMatrix&,...)):
	Indicate diagonal matrix.
	(octave_print_internal (...,const ComplexDiagMatrix&,...)): Ditto.
	(octave_print_internal (...,const PermMatrix&,...)): New function.
	* pr-output.h: Declare it.

2009-02-27  Jaroslav Hajek  <highegg@gmail.com>

	* OPERATORS/op-dms-template.cc (gripe_if_zero): New template static
	function.
	(dmsdiv, sdmldiv): Call it.

2009-02-26  John W. Eaton  <jwe@octave.org>

	* symtab.h (symbol_table::symbol_record::symbol_record_rep::forced):
	New static constant.
	(symbol_table::symbol_record::symbol_record_rep::force_variable,
	symbol_table::symbol_record::force_variable): New functions.
	(symbol_table::symbol_record::symbol_record_rep::is_forced,
	symbol_table::symbol_record::symbol_record_rep::mark_forced,
	symbol_table::symbol_record::symbol_record_rep::unmark_forced,
	symbol_table::symbol_record::symbol_record_rep::clear_forced,
	symbol_table::symbol_record::is_forced,
	symbol_table::symbol_record::mark_forced,
	symbol_table::symbol_record::clear_forced,
	symbol_table::symbol_record::unmark_forced): New fucntions.
	* lex.h, lex.l (force_local_variable): Delete.
	(handle_identifier): Call symbol_table::force_variable instead of
	force_local_variable.
	* parse.y (make_script): Call symbol_table:clear_forced_variables
	after defining script.
	(finish_function): Call symbol_table::clear_forced_variables
	instead of symbol_table::clear_variables.
	
	* DLD-FUNCTIONS/chol.cc: Correct spelling of CHOLMOD in tests.

	* version.h (OCTAVE_VERSION): Now 3.1.53+.
	(OCTAVE_API_VERSION): Now api-v35+.

2009-02-25  John W. Eaton  <jwe@octave.org>

	* version.h (OCTAVE_VERSION): Now 3.1.53.
	(OCTAVE_API_VERSION): Now api-v35.
	(OCTAVE_RELEASE_DATE): Now 2009-02-25.
	(OCTAVE_COPYRIGHT): Update year.

	* Makefile.in (distclean): Remove tags TAGS y.tab.h y.output
	yy.lex.c here, not in maintainer-clean target.  Also remove
	parse.output.

	* DLD-FUNCTIONS/rand.cc: Note that the tests for the old random
	number generator will fail if using 64-bit indexing.

2009-02-25  Jaroslav Hajek  <highegg@gmail.com>

	* DLD-FUNCTIONS/lu.cc (maybe_set_triangular): New function.
	(Flu): Use it.
	* DLD-FUNCTIONS/qr.cc (maybe_set_triangular): New function.
	(Fqr): Use it.

2009-02-25  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/chol.cc: Fix tests for missing cholmod.

	* input.cc (get_debug_input): Write debugging location info
	directly to std::cerr instead of calling message.
	* pt-eval.cc (tree_evaluator::do_breakpoint): Write debugging
	location info directly to std::cerr, not octave_stdout.

	* defaults.h.in (OCTAVE_DOC_CACHE_FILE): New macro.
	* defaults.cc (set_default_doc_cache_file): New function.
	(install_defaults): Call it.

	* help.cc (__list_functions__): Simplify

	* input.cc (get_debug_input): Don't pass arbitrary input to
	message as a format string.

2009-02-24  John W. Eaton  <jwe@octave.org>

	* help.cc, help.h (Vdoc_cache_file): New global variable.
	(Fdoc_cache_file): New function.
	* octave.cc (DOC_CACHE_FILE_OPTION): New option tag.
	(long_opts): Include doc-cache-file in the list.
	(octave_main): Handle DOC_CACHE_FILE_OPTION.

2009-02-24  Jaroslav Hajek  <highegg@gmail.com>

	* ov-scalar.h 
	(octave_scalar::int8_scalar_value): New method.
	(octave_scalar::int16_scalar_value): New method.
	(octave_scalar::int32_scalar_value): New method.
	(octave_scalar::int64_scalar_value): New method.
	(octave_scalar::uint8_scalar_value): New method.
	(octave_scalar::uint16_scalar_value): New method.
	(octave_scalar::uint32_scalar_value): New method.
	(octave_scalar::uint64_scalar_value): New method.

	* ov-float.h 
	(octave_float_scalar::int8_scalar_value): New method.
	(octave_float_scalar::int16_scalar_value): New method.
	(octave_float_scalar::int32_scalar_value): New method.
	(octave_float_scalar::int64_scalar_value): New method.
	(octave_float_scalar::uint8_scalar_value): New method.
	(octave_float_scalar::uint16_scalar_value): New method.
	(octave_float_scalar::uint32_scalar_value): New method.
	(octave_float_scalar::uint64_scalar_value): New method.

	* OPERATORS/op-s-s.cc (scalar_to_float): Remove duplicate conversion.
	* OPERATORS/op-range.cc (range_to_float_matrix): Ditto.

2009-02-24  John W. Eaton  <jwe@octave.org>

	* OPERATORS/op-float-conv.cc: New file.
	* Makefile.in (FLOAT_OP_XSRC): Add it to the list.
	(DOUBLE_OP_XSRC): Move op-double-conv.cc here from INTTYPE_OP_XSRC.
	* ops.h (DEFFLTCONVFN, DEFSTRFLTCONVFN): New macros.

2009-02-23  John W. Eaton  <jwe@octave.org>

	* ov-base-sparse.cc (octave_base_sparse<T>::print_raw):
	Remove extra ")" from output.

	* DLD-FUNCTIONS/eigs.cc (Feigs): If sigma argument is not a
	string, try extraction as complex value and check for error
	instead of inquiring about type first.

	* pt-eval.cc (tree_evaluator::visit_break_command,
	tree_evaluator::visit_return_command,
	tree_evaluator::visit_global_command,
	tree_evaluator::visit_static_command): Handle breakpoint.
	(tree_evaluator::visit_simple_for_command,
	tree_evaluator::visit_complex_for_command): Handle breakpoint once
	at beginning of loop.
	(tree_evaluator::visit_if_command_list): Don't call do_breakpoint
	for else clauses.
	(tree_evaluator::visit_no_op_command): Handle breakpoint if no-op
	command is end of function or script.
	(tree_evaluator::visit_statement): Handle breakpoint for
	expressions but not commands.
	(tree_evaluator::visit_while_command,
	tree_evaluator::visit_do_until_command):
	Check for breakpoint set on cmd, not expr.
	(tree_evaluator::visit_do_until_command): Handle breakpoint
	between check for breaking out of loop and loop control
	expression.

	* pt-cmd.h (tree_no_op_command::eof): New data member
	(tree_no_op_command::tree_no_op_command): Initialize it.
	(tree_no_op_command::is_end_of_fcn_or_script): New function.
	* pt-stmt.cc (tree_statement::is_end_of_fcn_or_script):
	Call is_end_of_fcn_or_script instead of looking at original text
	of no-op command.

	* pt-select.h, pt-select.cc (tree_if_command::set_breakpoint,
	tree_if_command::delete_breakpoint): New functions.
	(tree_switch_command::set_breakpoint,
	tree_switch_command::delete_breakpoint): New functions.
	* pt-stmt.h, pt-stmt.cc (tree_statement::set_breakpoint,
	tree_statement::delete_breakpoint, tree_statement::is_breakpoint):
	Delegate real work to cmd or expr.
	(tree_statement::bp): Delete data member.
	(tree_statement::tree_statement): Don't initialize bp.
	(tree_statement::dup): Don't copy bp.

	* pt.h (tree::line (int), tree:column (int)): New functions.
	(tree:set_breakpoint, tree::delete_breakpoint): Now virtual.

	* parse.y (finish_if_command): Also store line and column info for
	if statement in first element of list.
	(finish_switch_command): Likewise, for switch.

	* input.cc (last_debugging_command): New static variable.
	(octave_gets): Set it here.  Don't insert repeated debugging
	commands in the history list.

2009-02-23  Jaroslav Hajek  <highegg@gmail.com>

	* DLD-FUNCTIONS/pinv.cc: Support diagonal and permutation matrices.

2008-02-21  David Bateman  <dbateman@free.fr>

	* OPERATORS/op-cs-scm.cc, OPERATORS/op-cs-sm.cc, OPERATORS/op-s-scm.cc,
	OPERATORS/op-s-sm.cc, OPERATORS/op-scm-cs.cc, OPERATORS/op-scm-s.cc,
	OPERATORS/op-sm-cs.cc, OPERATORS/op-sm-s.cc: Don't perform any 
	narrowing to full matrices.

2009-02-20  Jaroslav Hajek  <highegg@gmail.com>

	* ov-base-diag.h (octave_base_diag::sqrt): Remove.
	* ov-re-diag.cc (octave_diag_matrix::sqrt): New method.
	* ov-re-diag.h: Declare it.
	* ov-flt-re-diag.cc (octave_float_diag_matrix::sqrt): New method.
	* ov-flt-re-diag.h: Declare it.
	* ov-cx-diag.cc (octave_complex_diag_matrix::sqrt): New method.
	* ov-cx-diag.h: Declare it.
	* ov-flt-cx-diag.cc (octave_float_complex_diag_matrix::sqrt): New method.
	* ov-flt-cx-diag.h: Declare it.

2009-02-20  John W. Eaton  <jwe@octave.org>

	* dynamic-ld.cc (octave_dynamic_loader::do_load_mex): Clear and
	reload mex file if it is out of date.  Don't check path for file.

	* pt-bp.cc (tree_breakpoint::visit_octave_user_script,
	tree_breakpoint::visit_octave_user_function):
	Don't panic.  Call accept on command list if it exists.

2009-02-20  Jaroslav Hajek  <highegg@gmail.com>

	* ov-base-mat.cc (octave_base_matrix<MT>::do_index_op):
	Correctly compute all-scalar index.
	* ov-cell.cc (octave_cell::is_sorted, octave_cell::is_sorted_rows):
	New methods.
	(octave_cell::octave_cell (const Array<std::string>&)): New
	constructor.
	* ov-cell.h: Declare them.
	* ov-cell.cc (octave_cell::sort): Create result already with
	cellstr_cache.
	* strfns.cc (Fstrcmp): Use special code when dealing with
	cellstr arrays.

2009-02-20  John W. Eaton  <jwe@octave.org>

	* ov-base-mat.cc (octave_base_matrix<MT>::assign):
	Correctly compute all-scalar index.

	* symbtab.cc (symbol_table::stash_dir_name_for_subfunctions):
	New function.
	* symtab.h: Provide decl.
	* parse.y (load_fcn_from_file): Call it after parsing a function.

2009-02-19  Jaroslav Hajek  <highegg@gmail.com>

	* ov-cell.h (octave_cell::cellstr_cache): New field.
	(octave_cell::clear_cellstr_cache, octave_cell::make_cellstr_cache,
	octave_cell::assign, octave_cell::delete_elements,
	octave_cell::mex_get_data): New methods.
	(octave_cell::is_cellstr): Reuse cellstr cache if possible, create if
	successful.
	(octave_cell::cellstr_value): Reuse cellstr cache if possible.
	(octave_cell::subsasgn): Clear cellstr cache.

2009-02-19  Jaroslav Hajek  <highegg@gmail.com>

	* DLD-FUNCTIONS/lookup.cc (Flookup): Use Array<T>::lookup if possible.
	Do not compare octave_values directly. Properly check for iscellstr.

2009-02-19  John W. Eaton  <jwe@octave.org>

	* data.cc, graphics.cc, help.cc, lex.l, load-path.cc, parse.y:
	Consistent doc strings for internal functions.

2009-02-18  John W. Eaton  <jwe@octave.org>

	* ov.cc (convert_to_octave_idx_type_array): New static function.
	(octave_value::octave_idx_type_vector_value): New function.
	* ov.h (octave_value::octave_idx_type_vector_value): Provide decl.
	* DLD-FUNCTIONS/qr.cc (Fqrdelete, Fqrinsert): Use it to convert
	octave_value object to array of octave_idx_type values.

	* DLD-FUNCTIONS/find.cc (find_nonzero_elem_idx): Likewise.
	* DLD-FUNCTIONS/time.cc (strptime): Likewise.
	* DLD-FUNCTIONS/quad.cc (Fquad): Eliminate unnecessary cast.
	* toplev.cc (run_command_and_return_output): Likewise.

	* strfns.cc (Fstrvcat): Use octave_idx_type and size_t instead of
	int as needed.
	(Fstrcmp, Fstrncmp): Use octave_idx_type instead of int as needed.

	* DLD-FUNCTIONS/hex2num.cc (Fhex2num, Fnum2hex):
	Use union to avoid reinterpret_cast and GCC warning.

	* mex.cc (call_mex): Declare local nargout variable volatile to
	avoid "maybe clobbered by vfork" warning from GCC.

	* ov-cx-mat.cc (xabs): Comment out unused static function.
	* ov-flt-cx-mat.cc (xabs): Ditto.

	* c-file-ptr-stream.cc (c_file_ptr_buf::seekoff,
	c_file_ptr_buf::seepos, c_zfile_ptr_buf::seekoff,
	c_zfile_ptr_buf::seepos): Avoid unused paramter warnings.

2009-02-18  Jaroslav Hajek  <highegg@gmail.com>

	* data.cc (Fresize): Allow arbitrary number of parameters. Improve
	documentation.

2009-02-17  Benjamin Lindner  <lindnerb@users.sourceforge.net>

	* file-io.cc: (Fmkstemp): Use mkstemps if it is available and
	mkstemp is missing.

2009-02-17  Olaf Till  <olaf.till@uni-jena.de>

	* DLD-FUNCTIONS/lsode.cc, DLD-FUNCTIONS/daspk.cc, 
	DLD-FUNCTIONS/dassl.cc, DLD-FUNCTIONS/dasrt.cc: Documentation fixes.

2009-02-17  Kai Habel  <kai.habel@gmx.de>

	* DLD-FUNCTIONS/convhulln.cc (Fconvhulln): Compute convex hull
	volume and return it as second output.  New tests.

2009-02-17  John W. Eaton  <jwe@octave.org>

	* ov-fcn.h (octave_function::octave_function): Initialize data members.
	(octave_function::dispatch_class): No longer virtual.
	Replace with version from ov-usr-fcn.h.
	(octave_function::xdispatch_class,
	octave_function::stash_dispatch_class): Move here from ov-usr-fcn.h.
	(octave_function::mark_as_private_function,
	octave_function::is_private_function,
	octave_function::is_private_function_of_class): New functions.
	(octave_function::private_function): New data member.
	* ov-usr-fcn.h (octave_user_function::xdispatch_class,
	octave_user_function::stash_dispatch_class,
	octave_user_function::dispatch_class): Delete.
	* ov-usr-fcn.cc (octave_user_function::octave_user_function):
	Don't initialize xdispatch_type.

	* symtab.cc
	(symbol_table::fcn_info::fcn_info_rep::load_private_function):
	If private directory is subdirectory of class directory, qmark
	function as a private directory of the class.

	* ov-class.cc (octave_class::in_class_method): Also return true
	for functions that are private to a class.
	(octave_class::subsasgn): If not in class method, look for
	subsasgn method.  If not found, or if not in a class method,
	perform default struct-style subscripted indexing.

	* ov-class.h (octave_class::empty_clone):
	Preserve fields and class name.
	* oct-map.h (Octave_map::Octave_map (const string_vector&):
	New constructor.

	* ov-class.cc (octave_class::subsref): If indexing directly and
	result is map, return class object, not simple struct.

	* symtab.cc
	(symbol_table::fcn_info::fcn_info_rep::load_private_function):
	Pass dir_name to load_fcn_from_file.
	
2009-02-17  Jaroslav Hajek  <highegg@gmail.com>

	* DLD-FUNCTIONS/data.cc (NATIVE_REDUCTION): Add BOOL_FCN argument.
	(NATIVE_REDUCTION_1): Check integer overflow flags and possibly gripe.
	(Fsum): Reflect change.
	(Fcumsum): USE NATIVE_REDUCTION.
	* gripes.cc (gripe_native_integer_math_truncated): New function.

2009-02-17  Jaroslav Hajek  <highegg@gmail.com>

	* DLD-FUNCTIONS/max.cc (Fcummin, Fcummax): Improve inline docs.

2009-02-17  Jaroslav Hajek  <highegg@gmail.com>

	* DLD-FUNCTIONS/max.cc (MINMAX_DOUBLE_SBODY): New macro.
	(MINMAX_DOUBLE_BODY): Move part of code to MINMAX_DOUBLE_SBODY.
	(MINMAX_SINGLE_SBODY): New macro.
	(MINMAX_SINGLE_BODY): Move part of code to MINMAX_DOUBLE_SBODY.
	(MINMAX_INT_SBODY): New macro.
	(MINMAX_INT_BODY): Move part of code to MINMAX_DOUBLE_SBODY.
	(CUMMINMAX_BODY): New macro.
	(Fcummin, Fcummax): New DLD functions.

2009-02-17  John W. Eaton  <jwe@octave.org>

	* octave.gperf: Eliminate whitespace to allow gperf 2.7.2 to work.

	* file-io.cc (Ffscanf, Fsscanf): Check error_state after call to
	octave_stream::scanf.
	(Ffgetl): Check error state after call to octave_stream::getl.
	(Ffgets): Check error state after call to octave_stream::gets.

	* oct-stream.cc (octave_base_stream::do_gets,
	octave_base_stream::do_scanf):
	Disallow reading from stdin when running interactively.

	* toplev.cc (octave_config_info): Add CARBON_LIBS, X11_INCFLAGS,
	and X11_LIBS to the struct.
	* oct-conf.h.in (OCTAVE_CONF_CARBON_LIBS,
	OCTAVE_CONF_X11_INCFLAGS, OCTAVE_CONF_X11_LIBS): New macros.

2009-02-17  Thomas Treichl  <Thomas.Treichl@gmx.net>

	* Makefile.in (octave$(EXEEXT)): Link with CARBON_LIBS.

2009-02-16  John W. Eaton  <jwe@octave.org>

	* lex.l (next_token_can_follow_bin_op):
	Push all characters read on to buffer stack.

	* genprops.awk (emit_source): Don't use + to concatenate strings.

2009-02-16  Jaroslav Hajek  <highegg@gmail.com>

	* data.cc (NATIVE_REDUCTION): Use boolNDArray::any for native bool
	summation, boolNDArray::sum for double-valued.

2009-02-16  Jaroslav Hajek  <highegg@gmail.com>

	* ov-base-mat.cc (octave_base_matrix<MT>::assign (const octave_value_list&,
	typename MT::element_type): Fix invalid index.
	* ov-re-mat.cc (default_numeric_demotion_function): Use
	float_array_value instead of float_matrix_value.

2009-02-16  John W. Eaton  <jwe@octave.org>

	* input.cc (gnu_readline): Use fputs instead of fprintf.

2009-02-15  John W. Eaton  <jwe@octave.org>

	* defun.cc, defun-int.h, defun-dld.h (DEFUN_INTERNAL,
	DEFCONSTFUN_INTERNAL, DEFUNX_INTERNAL, DEFUN_DLD_INTERNAL,
	DEFUNX_DLD_INTERNAL, install_builtin_function,
	install_dld_function, install_mex_function):
	Delete IS_TEXT_FCN arg.  Fix all uses.
	(DEFMD, DEFUN_TEXT): Delete.
	* mkbuiltins, mkgendoc: Adapt to DEFUN macro changes.

	* lex.h, lex.l (lexer_flags::at_beginning_of_statement):
	New data member.  Set it as needed in rules.
	(is_keyword_token):
	Handle lexer_flags::at_beginning_of_statement.
	(next_token_can_follow_bin_op, looks_like_command_arg): New functions.
	(handle_identifier): Use them to determine 
	Don't check is_command_name or is_rawcommand_name.
	* parse.y (statement): Recognize word_list_cmd here.
	(expression): Not here.

	* lex.l (handle_string): Delete arg TEXT_STYLE.
	(BIN_OP_RETURN, XBIN_OP_RETURN): New arg, BOS. Change all uses.

	* lex.h, lex.l (lexer_flags::doing_raw_command): Delete data
	member and all uses.

	* debug.cc, dirfns.cc, error.cc, input.cc, lex.l, load-path.cc,
	load-save.cc, oct-hist.cc, ov-class.cc, pager.cc, parse.y,
	pr-output.cc, sysdep.cc, utils.cc, variables.cc:
	Replace all uses of DEFCMD with DEFUN.

	* variables.cc (command_set, rawcommand_set): Delete static variables.
	(mark_as_command, unmark_command, is_command_name,
	mark_as_rawcommand, unmark_raw_command, is_rawcommand_name): Delete.
	(Fmark_as_command, Funmark_command, Fiscommand,
	Fmark_as_rawcommand, Funmark_rawcommand, Fisrawcommand):
	Convert to .m files and move to scripts/deprecated.
	* variables.h (mark_as_command, is_command_name,
	is_marked_as_rawcommand, mark_as_rawcommand, unmark_rawcommand,
	is_rawcommand_name): Delete decls.

	* lex.h, lex.l (lexer_flags.looking_at_object_index): Now a
	std::list<bool> object instead of an int.
	Push TRUE to list at start of object index.  Push FALSE at
	beginning of matrix list.  Pop value at end of object index or
	matrix list.
	(lexer_flags.looking_for_object_index): New data member.
	Set it as needed in rules.
	(inside_any_object_index): New function.
	* parse.y (begin_obj_idx, cancel_obj_idx): Delete non-terminals
	and all uses.

2009-02-13 Ben Abbott <bpabott@mac.com>

	* graphics.h.in (class axes::properties): Initialize ticklength.
	* graphics.cc (default_axes_ticklength): New function.

2009-02-13  John W. Eaton  <jwe@octave.org>

	* oct-stream.cc (octave_base_stream::do_gets): Handle CRLF and CR.

	* toplev.cc (do_octave_atexit): Only save history if
	Vsaving_history is true.

2009-02-12  John W. Eaton  <jwe@octave.org>

	* data.cc, ov-base-diag.h, ov-base-mat.h, ov-base-scalar.h,
	ov-base-sparse.h, ov-base.cc, ov-base.h, ov-perm.h, ov-range.h,
	ov.h: Rename internal issorted and issorted_rows functions to
	is_sorted and is_sorted_rows.

	* ov-base-diag.h, ov-base-mat.h, ov-base-scalar.h, ov-base.h,
	ov-cell.h, ov-perm.h, ov-range.h, ov.h, data.cc, ov-base.cc,
	ov-cell.cc: Rename all uses of sortrows_idx to sort_rows_idx.

	* TEMPLATE-INST/Array-tc.cc: Don't instantiate sort functions for
	Arrays of octave_value objects.
	(octave_sort<octave_value>::ascending_compare,
	octave_sort<octave_value>::descending_compare): Delete.

	* ov.h (octave_value::cellstr_value): New function.
	* ov-base.cc, ov-base.h (octave_base_value::cellstr_value):
	New function.
	* ov-cell.h (octave_cell::cellstr_value, octave_cell::sort,
	octave_cell::sortrows_idx): New functions
	* Cell.h, Cell.cc (Cell::Cell (Array<std::string>)): New constructor.

	* TEMPLATE-INST/Array-tc.cc: Undo previous change.
	(octave_sort<octave_value>::ascending_compare,
	octave_sort<octave_value>::descending_compare):
	Pass args by const reference instead of value.

	* version.h (OCTAVE_VERSION): Now 3.1.52+.
	(OCTAVE_API_VERSION): Now api-v34+.

	* TEMPLATE-INST/Array-tc.cc
	(octave_sort<octave_value>::ascending_compare,
	octave_sort<octave_value>::descending_compare):
	Delete unused template specializations.
	Use NO_INSTANTIATE_ARRAY_SORT instead of INSTANTIATE_ARRAY_SORT
	for octave_values.

2009-02-11  Thomas Treichl  <Thomas.Treichl@gmx.net>

	* gl-render.h: Use HAVE_FRAMEWORK_OPENGL.
	* gl-render.cc (opengl_tesselator, opengl_renderer::draw, 
	opengl_renderer::init_marker): Use HAVE_FRAMEWORK_OPENGL.
	
2009-02-11  Jaroslav Hajek  <highegg@gmail.com>

	* ov.h (octave_value::issorted, octave_value::sortrows_idx,
	octave_value::issorted_rows): New methods.
	* ov.cc (octave_base_value::issorted, octave_base_value::sortrows_idx,
	octave_base_value::issorted_rows): New methods.
	* ov.cc: Declare them.

	* ov-base-mat.h (octave_base_matrix::issorted, octave_base_matrix::sortrows_idx,
	octave_base_matrix::issorted_rows): New methods.

	* ov-base-diag.h (octave_base_diag::issorted, octave_base_diag::sortrows_idx,
	octave_base_diag::issorted_rows): New methods.

	* ov-perm.h (octave_perm_matrix::issorted, octave_perm_matrix::sortrows_idx,
	octave_perm_matrix::issorted_rows): New methods.

	* ov-range.h (octave_range::issorted, octave_range::sortrows_idx,
	octave_range::issorted_rows): New methods.

	* data.cc (F__sortrows_idx__, Fissorted): New defuns.

2009-02-11  John W. Eaton  <jwe@octave.org>

	* toplev.cc (octave_config_info): Add octetcdir to the struct.
	* defaults.h.in (OCTAVE_OCTETCDIR): New macro.

2009-02-10  Jaroslav Hajek  <highegg@gmail.com>

	* DLD-FUNCTIONS/__pchip_deriv__.cc (F__pchip_deriv__):
	Add support for computing pchip derivatives along rows of matrix.
	Eliminate redundant copying by using const args where appropriate.

2009-02-09  John W. Eaton  <jwe@octave.org>

	* version.h (OCTAVE_VERSION): Now 3.1.52.
	(OCTAVE_RELEASE_DATE): Now 2009-02-09.
	(OCTAVE_COPYRIGHT): Update year.

	* load-path.cc (dir_info::update, dir_info::initialize):
	Likewise, to allow some functionality if getcwd fails.

	* toplev.cc (main_loop): Also catch octave_execution_exception.

	* DLD-FUNCTIONS/dispatch.cc: Comment out troublesome tests.

	* DLD-FUNCTIONS/eigs.cc: Increase tolerance to 1e-11 on all tests.

	* lex.l (lexical_feedback::looking_at_decl_list): New data member.
	* lex.l (lexical_feedback::init): Initialize it.
	(handle_identifier): Also force local variable if looking_at_decl_list.
	* parse.y (parsing_decl_list): New non-terminal.
	(declaration): Use it.  Set lexer_flags.looking_at_decl_list to
	false after parsing the declaration.

2009-02-09  Jaroslav Hajek  <highegg@gmail.com>

	* TEMPLATE-INST/Array-tc.cc: Reflect changes in octave_sort.

2009-02-08  John W. Eaton  <jwe@octave.org>

	* octave.cc (initialize_pathsearch): Delete.
	(octave_main): Don't call initialize_pathsearch.

2009-02-07  John W. Eaton  <jwe@octave.org>

	* parse.y (eval_string): Use the one true evaluator.

2009-02-06  John W. Eaton  <jwe@octave.org>

	* Makefile.in (INCLUDES): Include oct-hdf5.h in the list.

	* Makefile.in (oct-gperf.h): Don't use pipeline to process gperf
	output.

2009-02-05  John W. Eaton  <jwe@octave.org>

	* symtab.cc (symbol_table::fcn_info::cn_info_rep::xfind):
	New function.
	(symbol_table::fcn_info::cn_info_rep::find):
	Use it to avoid recursive call.

	* graphics.cc (Fdrawnow): Return after errors.  Don't strip
	trailing directory separator from name used in call to file_stat.

	* DLD-FUNCTIONS/eigs.cc (Feigs): Avoid warnings about ambiguous
	else.  Style fixes.  Return after errors.

2009-02-05  Jaroslav Hajek  <highegg@gmail.com>

	* oct-map.cc (Octave_map::index): Optimize.

2009-02-05  Jaroslav Hajek  <highegg@gmail.com>

	* OPERATORS/op-cm-cs.cc: Use scalar_value in scalar-to-matrix
	indexed assignments.
	* OPERATORS/op-fcm-fcs.cc: Ditto.
	* OPERATORS/op-fm-fs.cc: Ditto.
	* OPERATORS/op-m-s.cc: Ditto.
	* OPERATORS/op-int.h (OCTAVE_MS_INT_ASSIGN_OPS): Use homogeneous
	integer scalar-to-matrix assignment.
	(OCTAVE_MM_INT_ASSIGN_OPS): Use homogeneous integer matrix-to-matrix
	assignment.

	* Cell.h (Cell::index): Remove direct idx_vector overloads.
	(Cell::assign): Remove direct idx_vector overloads.
	(Cell::delete_elements): Remove direct idx_vector overloads.
	* Cell.cc (Cell::index (const octave_value_list&)): Call
	ArrayN<octave_value>::index.

	* oct-map.h (Octave_map::index): Remove direct idx_vector overloads.
	* oct-map.cc (Octave_map::index): Ditto.
	* ov-base-mat.cc (octave_base_mat::do_index_op): Optimize scalar
	indexing case.
	(octave_base_mat::assign (const octave_value_list&, const MT&)):
	Specialize variable argument number.
	(octave_base_mat::assign (const octave_value_list&, const 
	typename MT::element_type&)):
	New method overload with optimized scalar indexing case.

	* ov-cell.cc (octave_base_matrix<Cell>::do_index_op,
	octave_base_matrix<Cell>::assign,
	octave_base_matrix<Cell>::delete_elements): Specialize.
	(octave_cell::assign (const_octave_value_list, const octave_value&):
	Delete method.
	* ov-cell.h: Reflect change.

	* ov-cx-mat.cc (octave_complex_matrix::assign): Delete overloads.
	* ov-cx-mat.h: Reflect changes.
	* ov-flt-cx-mat.cc (octave_float_complex_matrix::assign): Delete overloads.
	* ov-flt-cx-mat.h: Reflect changes.
	* ov-list.cc (octave_list::do_index_op): Simplify.
	* ov-struct.cc (octave_struct::do_index_op): Simplify.
	* pt-eval.cc (tree_evaluator::visit_simple_for_command): Optimize
	traversing a row vector.

2009-02-05  John W. Eaton  <jwe@octave.org>

	* variables.cc (Vignore_function_time_stamp):
	Eliminate unused static variable.

	* debug.cc (get_file_line): Override default precedence in logical
	expression.

	* ov-null-mat.cc (default_null_str_numeric_conversion_function,
	default_null_sq_str_numeric_conversion_function,
	default_null_matrix_numeric_conversion_function):
	Avoid unused variable warning.

	* ov-flt-re-diag.cc (octave_float_diag_matrix::save_binary):
	Avoid unused parameter warning.
	* pt-eval.cc (tree_evaluator::visit_if_clause): Likewise.
	* ov-typeinfo.cc (octave_value_typeinfo::do_register_type): Likewise.
	* ov-base.cc (octave_base_value::subsref (const std::string&,
	const std::list<octave_value_list>&, bool): Likewise.
	* ov-flt-cx-diag.cc (octave_float_complex_diag_matrix::save_binary): 
	Likewise.

	* oct-hdf5.h: New file.
	* ls-hdf5.cc, ov-base.h, ov.h: Include oct-hdf5.h instead of hdf5.h.
	* ls-hdf5.h: Include oct-hdf5.h.
	* load-save.cc: Include ls-hdf5.h instead of hdf5.h.
	* ov-base-sparse.cc, load-save.cc:
	Don't protect #include "ls-hdf5.h" with #ifdef.

2009-02-04  Kai Habel  <kai.habel@gmx.de>

	* gl-render.cc (opengl_renderer::draw (surface::properties)):
	Normalize surface normals. Correct calculation of diffuse lighting.
	* gl-render.cc (opengl_renderer::draw (patch::properties)):
	Correct calculation of diffuse lighting.

2009-02-04  John W. Eaton  <jwe@octave.org>

	* help.cc (do_which): If NAME is not in the symbol table, look for
	a function file in the load path.
	(raw_help_from_file): Unwind-protect and set reading_script_file.

	* pt-loop.h, pt-loop.cc (evaluating_looping_command):
	Delete global variable and all uses.
	* parse.y, parse.h (evaluating_function_body): Delete global
	variable and all uses.
	(make_break_command, make_continue_command, make_return_command):
	Use tree_evaluator::in_fcn_or_script_body and
	tree_evaluator::in_loop_command instead of
	evaluating_function_body and evaluating_looping_command.

	* pt-eval.h (tree_evaluator::in_function_or_script_body):
	Delete 	member variable and all uses.
	(tree_evaluator::reset): Delete function and all uses.
	
	* pt-eval.cc, pt-eval.h (tree_evaluator::in_fcn_or_script_body,
	tree_evaluator::in_loop_command): New static variables.
	(tree_evaluator::visit_simple_for_command,
	tree_evaluator::visit_complex_for_command,
	tree_evaluator::visit_while_command,
	tree_evaluator::visit_do_until_command): Unwind-protect and set
	tree_evaluator::in_loop_command instead of
	evaluating_looping_command.
	(tree_evaluator::visit_statement): Only call echo_code if
	evaluating function or script and (Vecho_executing_commands &
	ECHO_FUNCTIONS).  Use tree_evaluator::in_fcn_or_script_body
	instead of evaluating_function_body.  

	* ov-usr-fcn.cc (octave_user_script::do_multi_index_op):
	octave_user_function::do_multi_index_op): Unwind-protect and set
	tree_evaluator::in_fcn_or_script_body.

	* pt-stmt.cc, pt-stmt.h (tree_statement::echo_code):
	Rename from maybe_echo_code.  Simplify.

2009-02-04  Jaroslav Hajek  <highegg@gmail.com>

	* pt-loop.h (tree_simple_for_command::do_for_loop_once,
	tree_complex_for_command::do_for_loop_once): Remove obsolete decls.
	* pt-loop.cc (quit_loop_now): Remove obsolete method.
	* pt-eval.cc (DO_ND_LOOP): Delete macro.
	(tree_evaluator::visit_simple_for_command): Use the generic
	do_index_op for iterating matrices.

2009-02-04  John W. Eaton  <jwe@octave.org>

	New evaluator and debugger derived from tree-walker class.

	* pt-eval.h, pt-eval.cc: New files.  Parse tree evaluator code
	adapted from eval member functions in classes derived from
	tree_command.
	* Makefile.in (PT_INCLUDES, PT_SRC): Add them to the lists

	* pt-cmd.cc (tree_function_def::eval): Delete.
	* pt-cmd.h: Delete decl.
	(tree_command::eval): Delete pure virtual function.
	(tree_no_op_command::eval): Delete.
	(tree_function_def::function): Return octave_value, instead of
	pointer to octave_function.  Change all uses.

	* pt-except.cc (do_catch_code, tree_try_catch_command::eval,
	do_unwind_protect_cleanup_code,
	tree_unwind_protect_command::eval): Delete.
	* pt-except.h: Delete decls.

	* pt-jump.cc (tree_break_command::eval, tree_return_command::eval,
	tree_continue_command::eval): Delete.
	* pt-jump.h: Delete decls.

	* pt-loop.cc (DO_ND_LOOP): Delete macro.
	(tree_while_command::eval, tree_do_until_command::eval,
	tree_simple_for_command::do_for_loop_once,
	tree_simple_for_command::eval,
	tree_complex_for_command::do_for_loop_once,
	tree_complex_for_command::eval):
	* pt-loop.h: Delete decls.

	* pt-select.cc (tree_if_clause::eval, tree_if_command_list::eval,
	tree_if_command::eval, tree_switch_case::eval,
	tree_switch_case_list::eval, tree_switch_command::eval): Delete.
	* pt-select.h: Delete decls.
	(class tree_if_clause, class tree_switch_case): Derive from tree.
	Handle line and column in constructors.

	* pt-stmt.cc (tree_statement::eval): Delete
	* pt-stmt.h: Delete decl.

	* pt-stmt.cc, pt-stmt.h (tree_statement::is_command,
	tree_statement::is_expression, tree_statement::line,
	tree_statement::column): Now const.
	
	* pt-stmt.cc (tree_statement::set_print_flag,
	tree_statement::is_end_of_fcn_or_script): New functions.
	* pt-stmt.h: Provide decl.
	(set_breakpoint, delete_breakpoint, is_breakpoint): New function.
	(bp): New member variable.
	(print_result): Delete member variable.
	(tree_statement_list::anon_function_body): New member variable.
	(tree_statement_list::mark_as_anon_function_body,
	tree_statement_list::is_anon_function_body,
	tree_statement_list::is_script_body): New functions.

	* pt-decl.cc (tree_decl_init_list::eval,
	tree_decl_command::accept, tree_global_command::do_init,
	tree_global_command::eval, tree_static_command::do_init,
	tree_static_command::eval): Delete.
	* pt-decl.h: Delete decls.
	(tree_decl_elt::eval_fcn): Delete typedef.

	* pt-decl.cc (tree_global_command::accept,
	tree_static_command::accept): New functions.
	
	* pt-stmt.cc (tree_statement::print_result): Move here from
	pt-stmt.h.  Return true if expr is printable.

	* ov-base.cc (Vsilent_functions): Delete.
	(octave_base_value::print_with_name): Don't check
	evaluating_function_body && Vsilent_functions here.
	* ov-class.cc (octave_class::print_with_name): Likewise.
	* ov-base.h (Vsilent_functions): Delete decl.
	* pt-eval.cc (Fsilent_functions): Move here from ov-base.cc.
	(Vsilent_functions): New static variable.

	* sighandlers.cc (user_abort): Set tree_evaluator::debug_mode here.

	* pt.h (tree::bp): Rename from tree::break_point.
	(tree::set_breakpoint, tree::delete_breakpoint, tree::is_breakpoint):
	No longer virtual.

	* input.cc (get_debug_input): Use current_evaluator and
	tree-walker to evaluate command.
	* ov-usr-fcn.cc (octave_user_script::do_multi_index_op): Likewise.
	* toplev.cc (main_loop): Likewise.
	* parse.y (eval_string):

	* input.h, input.cc (Vdebugging_current_line): Delete.
	(get_debug_input): Use tree_evaluator::debug_line instead of
	Vdebugging_current_line.

	* ov-usr-fcn.cc (octave_user_function::do_multi_index_op):
	Handle inline functions and anonymous functions as single
	expressions.
	* parse.y (eval_string): Likewise.

	* parse.y (make_do_until_command): Rename first argument from
	do_tok to until_tok.
	(loop_command): Pass UNTIL token instead of DO token to
	make_do_until_command.
	(make_elseif_clause): New arg, elseif_tok.  Pass line and column
	info to tree_if_clause constructor.
	(elseif_clause): Pass ELSEIF token to make_elseif_clause.
	(make_switch_case): New arg, case_tok.  Pass line and column info
	to tree_switch_case constructor.
	(switch_case): Pass CASE token to make_switch_case.
	(make_script): New arg end_script.  Append it to cmds.
	(script): Create no-op command for end of script and pass it to
	make_script.
	(start_function): New arg, end_function. Append it to body.
	(function2): Pass end_function to start_function.
	(make_end): New function.
	(function_end): Declare as tree_statement_type.
	Create no-op command for end of script.
	(make_anon_fcn_handle): Mark body as anonymous function.
	(set_stmt_print_flag): Set print flag for all separator types

	* parse.y (fold (tree_binary_expression*),
	fold (tree_unary_expression *),	finish_colon_expression,
	finish_matrix): Call rvalue1 instead of rvalue.
	Stash line number in new tree_constant object.

	* debug.h (bp_table::have_breakpoints): New static function.
	(bp_table::do_have_breakpoints): New member function.
	* debug.cc (bp_table::do_add_breakpoint,
	bp_table::do_remove_breakpoint,
	bp_table::do_remove_all_breakpoints_in_file,
	bp_table::do_remove_all_breakpoints): Call
	bp_table::have_breakpoints to set tree_evaluator::debug_mode.
	(Fdbnext): Delete function.  Alias to dbstep.
	(Fdbquit, Fdbcont): Set tree_evaluator::dbstep_flag to zero.
	(Fdbstep): Rewrite to use tree_evaluator::dbstep_flag instead of
	tree::break_next, tree::last_line, tree::break_function, and
	tree::last_break_function.
	(Fdbwhere): Use tree_evaluator::debug_line and
	tree_evaluator::debug_column to get current line and column info.
	Don't print column if it is less than zero.
	* pt.cc, pt.h (tree::break_next, tree::last_line,
	tree::last_break_function, tree::break_function,
	tree::break_statement): Delete.

	* pt-bp.cc (tree_breakpoint::visit_global_command,
	tree_breakpoint::visit_static_command,
	tree_breakpoint::take_action (tree_statement&)): New functions.
	* pt-bp.h: Provide decls.

	* pt-bp.h (tree_walker::visit_global_command,
	tree_walker::visit_static_command): New pure virtual functions.
	(tree_walker::visit_decl_command): Delete.

	* pt-bp.cc (tree_breakpoint::visit_decl_command): Delete.
	* pt-bp.h: Delete decl.
	(MAYBE_DO_BREAKPOINT): Delete macro and all uses.

	* pt-bp.cc (tree_breakpoint::visit_no_op_command): Do nothing.
	(tree_breakpoint::visit_argument_list,
	tree_breakpoint::visit_binary_expression,
	tree_breakpoint::visit_colon_expression,
	tree_breakpoint::visit_decl_elt,
	tree_breakpoint::visit_decl_init_list,
	tree_breakpoint::visit_octave_user_script,
	tree_breakpoint::visit_octave_user_function,
	tree_breakpoint::visit_octave_user_function_header,
	tree_breakpoint::visit_octave_user_function_trailer,
	tree_breakpoint::visit_identifier,
	tree_breakpoint::visit_index_expression,
	tree_breakpoint::visit_matrix, tree_breakpoint::visit_cell,
	tree_breakpoint::visit_multi_assignment,
	tree_breakpoint::visit_anon_fcn_handle,
	tree_breakpoint::visit_constant,
	tree_breakpoint::visit_fcn_handle,
	tree_breakpoint::visit_parameter_list,
	tree_breakpoint::visit_postfix_expression,
	tree_breakpoint::visit_prefix_expression,
	tree_breakpoint::visit_return_list,
	tree_breakpoint::visit_simple_assignment): Call panic_impossible
	since breakpoints aren't set on expressions.
	(tree_breakpoint::visit_if_clause): Fold into visit_if_command_list.
	(tree_breakpoint::visit_switch_case): Fold into visit_switch_case_list.
	(tree_breakpoint::visit_try_catch_command,
	tree_breakpoint::visit_unwind_protect_command):
	Don't set breakpoint on command itself, the the statements it
	contains.
	(tree_breakpoint::visit_global_command,
	tree_breakpoint::visit_static_command): Call do_decl_command to do
	actual work.
	(tree_breakpoint::visit_while_command,
	tree_breakpoint::visit_do_until_command,
	tree_breakpoint::visit_simple_for_command,
	tree_breakpoint::visit_complex_for_command,
	tree_breakpoint::visit_statement,
	tree_breakpoint::visit_statement_list,
	tree_breakpoint::visit_switch_case_list): Set breakpoints at
	appropriate places.

	* pt-pr-code.h, pt-pr-code.cc (tree_print_code::visit_global_command,
	tree_print_code::visit_static_command,
	tree_print_code::do_decl_command): New functions.
	(tree_print_code::visit_decl_command): Delete.

	* pt-check.h, pt-check.cc (tree_checker::visit_global_command,
	tree_checker::visit_static_command, tree_checker::do_decl_command):
	New functions. 
	(tree_checker::visit_decl_command): Delete.

	* pt-select.cc (tree_switch_case::label_matches): Call rvalue1
	instead of rvalue.

	* pt-exp.h, pt-exp.cc (tree_expression::rvalue1): New function.
	(tree_expression::rvalue (void)): Delete.
	(tree_expression::is_logically_true): Call rvalue1 instead of rvalue.

	* pt-fcn-handle.h, pt-fcn-handle.cc (tree_fcn_handle::rvalue1,
	tree_anon_fcn_handle::rvalue1): New functions.
	(tree_fcn_handle:rvalue (void),
	tree_anon_fcn_handle::rvalue1 (void)): Delete.
	(tree_fcn_handle::rvalue (int), tree_anon_fcn_handle_rvalue (int)):
	Call rvalue1 instead of rvalue.

	* pt-idx.h, pt-idx.cc (tree_index_expression::rvalue1): New function.
	(tree_index_expression::rvalue (void)): Delete.
	(tree_index_expression::get_struct_index,
	tree_index_expression::rvalue (int)): Call rvalue1 instead of rvalue.

	* pt-mat.h, pt-mat.cc (tree_matrix::rvalue1): New function.
	(tree_matrix::rvalue (void)): Delete.
	(tm_row_const::tm_row_const_rep::init,
	tree_matrix::rvalue (int)): Call rvalue1 instead of rvalue.

	* pt-misc.cc (tree_parameter_list::convert_to_const_vector):
	Call rvalue1 instead of rvalue.
	* pt-arg-list.cc (tree_argument_list::convert_to_const_vector):
	Likewise.

	* pt-unop.h, pt-unop.cc (tree_prefix_expression::rvalue1,
	tree_postfix_expression::rvalue1): New functions.
	(tree_prefix_expression::rvalue (void),
	tree_postfix_expression::rvalue (void)): Delete.
	(tree_prefix_expression::rvalue (int),
	tree_postfix_expression::rvalue (int)):
	Call rvalue1 instead of rvalue.

	* pt-id.h, pt-id.cc (tree_identifier::rvalue1): New function.
	(tree_identifier::rvalue (void)): Delete.

	* pt-assign.h, pt-assign.cc (tree_simple_assignment::rvalue1,
	tree_multi_assignment::rvalue1): New functions.
	(tree_simple_assignment::rvalue (void),
	tree_multi_assignment::rvalue (void): Delete.
	(tree_simple_assignment::rvalue (int)):
	Call rvalue1 instead of rvalue.

	* pt-binop.h, pt-binop.cc (tree_binary_expression::rvalue1,
	tree_boolean_expression::rvalue1): New functions.
	(tree_binary_expression::rvalue (void),
	tree_boolean_expression::rvalue (void)): Delete.
	(tree_binary_expression::rvalue (int),
	tree_boolean_expression::rvalue (int)):
	Call rvalue1 instead of rvalue.

	* pt-cbinop.h, pt-cbinop.cc
	(tree_compound_binary_expression::rvalue1): New function.
	(tree_compound_binary_expression::rvalue (void)): Delete.

	* pt-cell.h, pt-cell.cc (tree_cell::rvalue1): New function.
	(tree_cell::rvalue (void)): Delete.
	(tree_cell::rvalue (int)): Call rvalue1 instead of rvalue.

	* pt-colon.h, pt-colon.cc (tree_colon_expression::rvalue1):
	New function.
	(tree_colon_expression::rvalue (void)): Delete.
	(tree_colon_expression::rvalue (int)):
	Call rvalue1 instead of rvalue.

	* pt-const.h, pt-const.cc (tree_constant::rvalue1): New function.
	(tree_constant::rvalue (void)): Delete.
	(tree_constant::rvalue (int)): Call rvalue1 instead of rvalue.

	* pt-decl.h, pt-decl.cc (tree_decl_elt::ravlue1): New function.
	(tree_decl_elt::ravlue (void)): Delete.
	(tree_decl_elt::eval): Call rvalue1 instead of rvalue.

2009-02-03  Jaroslav Hajek  <highegg@gmail.com>

	* TEMPLATE-INST/Array-tc.cc: Replace vec_index by pointers.

2009-02-02  Jaroslav Hajek  <highegg@gmail.com>

	* ov-re-mat.cc (octave_matrix::abs, octave_matrix::real,
	octave_matrix::imag, octave_matrix::conj): Use special code.
	* ov-flt-re-mat.cc (octave_float_matrix::abs, octave_float_matrix::real,
	octave_float_matrix::imag, octave_float_matrix::conj): Use special code.
	* ov-cx-mat.cc (octave_complex_matrix::abs, octave_complex_matrix::real,
	octave_complex_matrix::imag, octave_complex_matrix::conj): Use special code.
	* ov-flt-cx-mat.cc (octave_float_complex_matrix::abs, octave_float_complex_matrix::real,
	octave_float_complex_matrix::imag, octave_float_complex_matrix::conj): Use special code.

2009-01-30  Jaroslav Hajek  <highegg@gmail.com>

	* DLD-FUNCTIONS/lookup.cc (Flookup): Add missing check.

2009-01-29  John W. Eaton  <jwe@octave.org>

	* pt-stmt.cc (tree_statement::eval): Check
	in_function_or_script_body instead of symbol_table::at_top_level ()
	to decide whether to call octave_call_stack::set_statement.

	* graphics.cc (properties::set_defaults): Also set
	horizontalalignment, verticalalignmnt, and rotation properties for
	new label objects here.

	* ov-base-scalar.cc (octave_base_scalar<ST>::is_true (void) const):
	Error if scalar is NaN.
	* ov-base-mat.cc (octave_base_matrix<MT>::is_true (void) const):
	Likewise, if any element of matrix is NaN.
	* ov-cell.cc (octave_cell::is_true): New function.
	* Cell.h (Cell::any_element_is_nan): New function.

2009-01-29  Jaroslav Hajek  <highegg@gmail.com>

	* pr-output.cc:
	(void octave_print_internal (..., const DiagMatrix& m,...),
	(void octave_print_internal (..., const ComplexDiagMatrix& m,...),
	(void octave_print_internal (..., const FloatDiagMatrix& m,...),
	(void octave_print_internal (..., const FloatComplexDiagMatrix& m,...)):
	New functions.
	* pr-output.h: Declare them.
	* ov-base-diag.cc (octave_base_diag::print_raw): Call
	octave_print_internal.

2009-01-28  John W. Eaton  <jwe@octave.org>

	* Makefile.in (install, uninstall): Handle SHLLIBPRE and SHLBINPRE
	library prefixes.  From Marco Atzeri <marco_atzeri@yahoo.it>.

2008-01-28  David Bateman  <dbateman@free.fr>

	* DLD-FUNCTIONS/regexp.cc (octregexp_list): Don't break for zero
	length match, but rather advance the index by one character and
	try again.

2009-01-28  Jaroslav Hajek  <highegg@gmail.com>

	* DLD-FUNCTIONS/lookup.cc (Flookup): Fix doc string.

2009-01-27  John W. Eaton  <jwe@octave.org>

	* load-path.cc (Fcommand_line_path): Rename from Fcommandlinepath.

2009-01-27 Ben Abbott <bpabbott@mac.com>

	* graphics.cc (convert_position): Handle character units.
	(default_axes_tick, default_figure_papersize,
	default_figure_paperposition): New functions.
	* graphics.h.in (class figure::properties): Use them to
	initialize papersize, paperposition, xtick, ytick, and ztick
	properties.

2009-01-27  Jaroslav Hajek  <highegg@gmail.com>

	* DLD-FUNCTIONS/qr.cc (Fqr): Treat empty matrices correctly.

2009-01-26  Thomas Treichl  <Thomas.Treichl@gmx.net>

	* display.cc (display_info::init): Use double instead of CGFloat.
	Use HAVE_FRAMEWORK_CARBON instead of OCTAVE_USE_OS_X_API.

2009-01-26  John W. Eaton  <jwe@octave.org>

	* load-path.cc (load_path::do_find_fcn): Handle @foo/bar.

2009-01-24  Jaroslav Hajek  <highegg@gmail.com>

	* pt-cell.cc (tree_cell::rvalue): Optimize the single row case.

2009-01-24  Jaroslav Hajek  <highegg@gmail.com>

	* ov-struct.cc (octave_struct::subsasgn): Reshape cs-list on rhs 
	according to the index expression.
	* ov-cell.cc (octave_cell::subsasgn): Ditto.

2008-12-25 Ben Abbott <bpabbott@mac.com>

	* load-path.cc (Fcommandlinepath): New function.
	* load-path.h (load_path::get_command_line_path,
	load_path::do_get_command_line_path): New functions.

2009-01-24 Ben Abbott <bpabbott@mac.com>

	* DLD-FUNCTIONS/eigs.cc: eigs.cc: skip tests if ARPACK is missing.

2009-01-25  Søren Hauberg  <hauberg@gmail.com>

	* help.cc (do_get_help_text, raw_help_from_symbol_table): new output to
	flag the a function is found but not documented.
	
2009-01-25  Søren Hauberg  <hauberg@gmail.com>

	* help.cc (raw_help_from_file): No longer search for files called
	'Contents.m', as this is moved to 'script/help.m'.
	

2009-01-23  John W. Eaton  <jwe@octave.org>

	* toplev.cc (octave_call_stack::do_goto_caller_frame):
	Also allow caller frame to be base frame, not just user code.

2009-01-23  Jaroslav Hajek  <highegg@gmail.com>

	* gripes.cc (gripe_indexed_cs_list, gripe_invalid_inquiry_subscript):
	New functions.
	* gripes.h: Declare them.
	* pt-idx.cc: Remove definitions of the above funcs.
	* ov-cell.cc (octave_cell::subsref): Declare constants as const.
	(octave_cell::subsasgn): Remove dead branch, declare constants as const.
	(octave_cell::list_value): Optimize.
	* ov-struct.cc 
	(octave_struct::subsref): Declare constants as const.
	(octave_struct::subsasgn): Remove dead branch, declare constants as const.
	* ov-cs-list.cc (octave_cs_list::octave_cs_list (const Cell&)):
	Optimize.
	* oct-obj.cc (octave_value_list::octave_value_list (const
	std::list<octave_value_list>&)): New constructor.
	* oct-obj.h: Declare it.
	* pt-arg-list.cc (convert_to_const_vector): Optimize.
	* symtab.cc (symbol_table::fcn_info::fcn_info_rep::find): Use const
	reference to avoid redundant copy.
	* ov-usr-fcn.cc (octave_user_function::bind_automatic_vars): Optimize.
	(octave_user_function::octave_all_va_args): Optimize.

2009-01-22  Jaroslav Hajek  <highegg@gmail.com>

	* Cell.h (Cell::Cell (octave_value_list)): Only declare.
	(Cell::index (*)): Change resize_ok type to bool.
	* Cell.cc (Cell::Cell (octave_value_list)): Redefine.
	* oct-obj.h, oct-obj.cc: Change octave_value_list::data to
	Cell variable, reflect changes.

2009-01-22  John W. Eaton  <jwe@octave.org>

	* help.cc (do_which (std::ostream&, const std::string&), Fwhich):
	Delete.
	(do_which (const std::string&, std::string&), F__which__):
	New functions.
	* do_which (const std::string&):
	Call do_which (const std::string&, std::string&) to do the work.

2009-01-22  Søren Hauberg  <hauberg@gmail.com>

	* defun-int.h (print_usage): No longer mark as deprecated.
	* defun.cc (print_usage): Simply call feval to execute print_usage.m.
	* help.cc (additional_help_message, display_names_from_help_list,
	display_symtab_names, simple_help, try_info, help_from_info,
	display_help_text, display_usage_text, raw_help_from_list,
	help_from_list, help_from_symbol_table, help_from_file,
	builtin_help, Fhelp, display_file, do_type, Ftype,
	first_help_sentence, print_lookfor, Flookfor): Delete.
	(looks_like_html, raw_help_from_map, raw_help, do_get_help_text,
	F__operators__, F__keywords__, F__builtins__,
	file_is_in_dir, F__list_functions__): New functions.
	(pair_type, map_iter): New typedefs.
	(operators, keywords): Use pair_type for elements of list.
	(names): Use map for (keyword, doc) lists.
	(make_name_list): Sprinkle with const.
	* help.h (display_help_text, display_usage_text,
	additional_help_message): Delete decls.

2009-01-22  John W. Eaton  <jwe@octave.org>

	* toplev.cc (octave_config_info): Check OCTAVEUSE_OS_X_API instead
	of __APPLE__ && __MACH__.

	* display.cc (display_info::init): Get info for Windows and OS X
	systems.

2009-01-22  Jaroslav Hajek  <highegg@gmail.com>

	* pt-idx.cc (tree_index_expression::lvalue): Correct tmpi when seeing
	a plain struct component.
	* ov-struct.cc (octave_struct::subsasgn): Use proper type substr in
	empty_conv.

2009-01-22  Jaroslav Hajek  <highegg@gmail.com>

	* DLD-FUNCTIONS/qr.cc: Remove HAVE_QRUPDATE check.
	* DLD-FUNCTIONS/chol.cc: Ditto.

2009-01-21  John W. Eaton  <jwe@octave.org>

	* Makefile.in (display.o): Add X11_INCFLAGS to CPPFLAGS.
	(octave$(EXEEXT)): Link with X11_LIBS.

	* display.h, display.cc: New files.
	* graphics.cc (default_screendepth, default_screensize,
	default_screenpixelsperinch): New functions.
	* graphics.h.in (class root_figure::properties): New property,
	screendepth.  Use default_screensize to initialize screensize
	property.  Use default_screenpixelsperinch to initialize
	screenpixelsperinch property.

	* graphics.cc (properties::init): Set default xlabel, ylabel,
	zlabel, and title properties that differ from default text
	properties.

2009-01-21  Jaroslav Hajek  <highegg@gmail.com>
	
	* OPERATORS/op-range.cc: Define & register range-scalar ops.

2009-01-21  Jaroslav Hajek  <highegg@gmail.com>

	* ov.h (octave_value::subsref (..., bool auto_add)): New method.
	(octave_value::next_subsref (bool auto_add, ...)): New method.
	* ov.cc (octave_value::next_subsref (bool auto_add, ...)): New method.
	* ov-base.h (octave_base_value::subsref (..., bool auto_add)): New
	virtual method.
	* ov-base.cc (octave_base_value::subsref (..., bool auto_add)): New
	virtual method.
	* ov-cell.cc (octave_cell::subsref (..., bool auto_add)): New virtual
	method.
	* ov-cell.h (octave_cell::subsref (..., bool auto_add)): Declare it.
	* ov-struct.cc (octave_struct::subsref (..., bool auto_add)): New
	virtual method.
	(octave_struct::subsref (const std::string& type, const
	std::list<octave_value_list>& idx)): Do not allow resizing.
	* ov-struct.h (octave_struct::subsref (..., bool auto_add)): Declare
	it.
	* ov-struct.cc (octave_struct::dotref (..., bool auto_add)): New
	virtual method.
	* ov-struct.h (octave_struct::dotref (..., bool auto_add)): Declare it.
	* pt-idx.cc (tree_index_expression::rvalue): Do not reevaluate already
	evaluated part of the index chain.
	(tree_index_expression::rvalue): Do not reevaluate already evaluated
	part of the index chain. Do not actually perform trailing indexing. 
	Do not allow indexing cs-lists.

2009-01-20  John W. Eaton  <jwe@octave.org>

	* file-io.cc (Ffstat): New function.

2009-01-17  Jaroslav Hajek  <highegg@gmail.com>

	* DLD-FUNCTIONS/qr.cc (Fqrupdate, Fqrinsert, Fqrdelete, Fqrshift):
	Reflect changes in liboctave.
	* DLD-FUNCTIONS/chol.cc (Fcholupdate, Fcholinsert):
	Reflect changes in liboctave.

2009-01-19  Jaroslav Hajek  <highegg@gmail.com>

	* ov.h (octave_value::make_unique (int)): New method.
	* oct-obj.h (octave_value_list::octave_value_list (const Cell&)): New
	constructor.
	* ov-cell.cc (octave_cell::subsasgn): Allow composed {} indexing
	in multi-assignments. Optimize.
	* ov-struct.cc (octave_struct::subsasgn): Correct composed {} indexing
	in multi-assignments. Optimize & fix bugs.
	* pt-idx.cc (tree_index_expression::lvalue): Rewrite to allow
	specifying cs-list anywhere in index chain, be more robust.

2009-01-19  John W. Eaton  <jwe@octave.org>

	* lex.l (grab_comment_block): If not reading input from a file,
	bail out at first newline inside a comment.

	* lex.l (lexer_debug_flag): New static variable.
	(F__lexer_debug_flag__): New function.
	(LEXER_DEBUG): New macro.  Use it in all patterns.
	(DISPLAY_TOK_AND_RETURN): Also display token if lexer_debug_flag
	is set.
	(process_comment): Display comment if lexer_debug_flag is set.
	(display_character, display_state, lexer_debug): New static functions.
	(xunput): Display character if lexer_debug_flag is set.

2009-01-17  Jaroslav Hajek  <highegg@gmail.com>

	* ov.h: Describe usage of storable_value and make_storable_value.
	* ov.cc: Remove FIXME comment.

2009-01-15  John W. Eaton  <jwe@octave.org>

	* data.cc (Freshape): Include mismatched dimensions in error message.
	From Robert Millan <rmh@aybabtu.com>.

2009-01-14  Jaroslav Hajek  <highegg@gmail.com>

	* ov.cc (octave_value::maybe_economize): New method.
	(octave_value::non_null_value): rename to storable_value.
	(octave_value::make_non_null_value): rename to make_storable_value.
	* ov-base.h (octave_base_value::maybe_economize): New method.
	* ov-base-mat.h (octave_base_mat::maybe_economize): New override.
	* oct-obj.cc (octave_value_list::normalize_null_values):
	Rename to make_storable_values, use make_storable_value.
	* oct-obj.h: Ditto.
	* ov-builtin.cc: non_null_value -> storable_value.
	* ov-cell.cc: Ditto.
	* ov-struct.cc: Ditto.
	* pt-decl.h: Ditto.

2009-01-15  Søren Hauberg  <hauberg@gmail.com>

	* DLD-FUNCTIONS/__magick_read__.cc (encode_uint_image):
	Initialize bitdepth.

2009-01-14  Søren Hauberg  <hauberg@gmail.com>

	* DLD-FUNCTIONS/betainc.cc, DLD-FUNCTIONS/chol.cc,
	DLD-FUNCTIONS/daspk.cc, DLD-FUNCTIONS/dasrt.cc,
	DLD-FUNCTIONS/dassl.cc, DLD-FUNCTIONS/filter.cc,
	DLD-FUNCTIONS/gammainc.cc, DLD-FUNCTIONS/gcd.cc,
	DLD-FUNCTIONS/givens.cc, DLD-FUNCTIONS/hess.cc,
	DLD-FUNCTIONS/lsode.cc, DLD-FUNCTIONS/qr.cc, DLD-FUNCTIONS/qz.cc,
	DLD-FUNCTIONS/schur.cc, DLD-FUNCTIONS/svd.cc,
	DLD-FUNCTIONS/syl.cc, data.cc, mappers.cc:
	Use ifnottex instead of ifinfo.

2009-01-14  John W. Eaton  <jwe@octave.org>

	* load-path.cc (load_path::do_set): Call do_clear after disabling
	add_hook.

2009-01-13  Jaroslav Hajek  <highegg@gmail.com>

	* dirfns.cc (Freadddir): qsort -> sort.
	* graphics.cc (maybe_set_children): Ditto.
	* help.cc (simple_help): Ditto.
	* input.cc (generate_possible_completions): Ditto.
	* oct_map.cc (keys_ok): Ditto.
	* ov-fcn-inline.cc (Finline): Ditto.
	* variables.cc (Fisrawcommand): Ditto.

2009-01-12  John W. Eaton  <jwe@octave.org>

	* octave.cc (maximum_braindamage): Set confirm_recursive_rmdir to
	false.

	* strfns.cc (Fstrcmp, Fstrncmp): Simplify by initializing all
	elements of output to false with constructor argument.

2009-01-12  Thorsten Meyer  <thorsten.meyier@gmx.de>

	* strfns.cc (Fstrncmp): New test.
	
2009-01-12  John W. Eaton  <jwe@octave.org>

	* pt-stmt.h (tree_statement::is_null_statement): New function.
	* parse.y (set_stmt_print_flag): Return list.  Remove trailing
	null statements from list.
	(list, simple_list): Simplify action.
	(make_statement, make_statement_list, append_statement_list):
	Always create statements and them to the list.

	* base-list.h (octave_base_list::push_front,
	octave_base_list::push_back, octave_base_list::pop_front,
	octave_base_list::pop_back): New functions.

	* DLD-FUNCTIONS/qr.cc (Fqrupdate, Fqrinsert, Fqrdelete, Fqrshift):
	Require args to be numeric, not necessarily matrix objects.

2009-01-12  Ben Abbott  <bpabbott@mac.com>

	* graphics.h.in (text::properties::fontunits): Fix typo in list of
	possible values.

2009-01-12  John W. Eaton  <jwe@octave.org>

	* strfns.cc (Fstrvcat, Fchar): Use queue to avoid duplicate calls
	to all_strings.

2009-01-12  Thorsten Meyer  <thorsten.meyier@gmx.de>

	* strfns.cc (Fstrvcap): New function.
	
2008-11-29  Thorsten Meyer  <thorsten.meyier@gmx.de>

        * strfns.cc (Fchar, Fischar, Fstrncmp, Flist_in_columns):
	Improve documentation strings.  Add examples and tests.
        
2009-01-12  Jaroslav Hajek  <highegg@gmail.com>

	* ov-perm.cc (octave_perm_matrix::do_index_op): Return shallow copy
	when indexed by (:,:).
	* ov-base-mat.cc (octave_base_diag::do_index_op): Return shallow copy
	when indexed by (:,:). Optimize indexing equivalent to resizing.
	* data.cc: Fix tests.
	* mappers.cc: Fix tests.

2009-01-12  Jaroslav Hajek  <highegg@gmail.com>

	* ov-base-diag.h, ov-base-mat.h, ov-base-scalar.h, ov-base-sparse.h,
	ov-base.cc, ov-base.h, ov-perm.h, ov-range.h, ov.h: Add full_value
	methods to octave_value, octave_base_value and its relevant
	subclasses.
	* data.cc (Ffull): New DEFUN.
	* DLD-FUNCTIONS/sparse.cc (Ffull): move to data.cc.
2009-01-11  Jaroslav Hajek  <highegg@gmail.com>
	
	* ov.h (octave_value::is_copy_of): New method.
	* ov-struct.cc (numeric_conv): Add output parameter orig.
	(octave_struct::subsasgn): Temporarily erase duplicate lhs value prior
	to assignment.

2009-01-11  Jaroslav Hajek  <highegg@gmail.com>

	* oct-obj.cc (octave_value_list::all_scalars): New method.
	* oct-obj.h: Declare it.
	* ov-cell.cc (octave_cell::subsasgn): Disable assigning
	to non-scalar {}-indexed cell, unless in a multi-assignment.

2009-01-10  Jaroslav Hajek  <highegg@gmail.com>

	* DLD-FUNCTIONS/sparse.cc (Ffull): Convert diagonal & permutation
	matrices, leave other classes untouched.

2009-01-09  Jaroslav Hajek  <highegg@gmail.com>

	* OPERATORS/op-cdm-cdm.cc: Install missing widening op.
	* OPERATORS/op-dm-dm.cc: Ditto.
	* OPERATORS/op-fcdm-fcdm.cc: Ditto.
	* OPERATORS/op-fdm-fdm.cc: Ditto.
	* OPERATORS/op-pm-pm.cc: Ditto.
	* OPERATORS/op-fpm-fpm.cc: Ditto.

2009-01-09  Jaroslav Hajek  <highegg@gmail.com>

	* ov-cell.cc (octave_cell::subsasgn): Put erased copy back after
	make_unique has been called.

2009-01-09  Kai Habel <kai.habel@gmx.de>

        * graphics.cc: Calculate normals for surface boundaries, use
        more neighboring points to calculate interior normals

2009-01-09  John W. Eaton  <jwe@octave.org>

	* input.cc (get_user_input (void)): Don't increment input_line_number.
	* lex.l (xunput): New function.  Use it in place of yyunput
	anywhere a newline character may be put back on the input stream.
	Increment input_line_number in all rules that consume newlines
	characters.
	(text_yyinput): Increment input_line_number
	(fixup_column_count): Increment input_line_number.
	(prep_for_function): Set input_line_number to 1, not 0.
	(reset_parser): Set input_line_number to current_command_number,
	not current_command_number-1.
	(flex_stream_reader::ungetc): Call xunput, not yyunput.
	* parse.y (input_line_number): Initialize to 1, not 0.
	(text_getc): Increment input_line_number correctly.
	(stdio_stream_reader::ungetc): Decrement input_line_number if
	putting back a newline character.
	(parse_fcn_file): Set input_line_number to 1, not 0.
	(eval_string): Unwind-protect input_line_number and
	current_input_column.

	* parse.y (make_statement): New function.
	(statement): Use it.
	(make_statement_list, append_statement_list): New functions.
	(simple_list1, list1): Use them.

2009-01-08  Jaroslav Hajek  <highegg@gmail.com>

	* ov-cell.cc (octave_cell::subsasgn): Erase duplicate lhs value 
	prior to assignment to avoid a redundant copy.

2008-12-26  Thorsten Meyer  <thorsten.meyier@gmx.de>

	* mappers.cc (Ftoascii), mappers.cc (Ftolower), mappers.cc
	(Ftoupper), DLD-FUNCTIONS/regexp.cc (Fregexp),
	DLD-FUNCTIONS/regexp.cc	(Fregexpi), DLD-FUNCTIONS/regexp.cc
	(Fregexprep): Add references.
	
2008-12-27  Jaroslav Hajek <highegg@gmail.com>

	* oct-obj.h, oct-obj.cc (octave_value_list::valid_scalar_indices): Remove.
	* ov-base-diag.cc, ov-base-diag.h, ov-base-scalar.cc, ov-base.h,
	ov-bool-mat.cc, ov-bool-mat.h, ov-bool-sparse.cc, ov-bool-sparse.h,
	ov-bool.cc, ov-bool.h, ov-ch-mat.cc, ov-ch-mat.h, ov-colon.h,
	ov-complex.cc, ov-complex.h, ov-cx-mat.cc, ov-cx-mat.h,
	ov-cx-sparse.cc, ov-cx-sparse.h, ov-float.cc, ov-float.h,
	ov-flt-complex.cc, ov-flt-complex.h, ov-flt-cx-mat.cc,
	ov-flt-cx-mat.h, ov-flt-re-mat.cc, ov-flt-re-mat.h, ov-intx.h,
	ov-perm.cc, ov-perm.h, ov-range.h, ov-re-mat.cc, ov-re-mat.h,
	ov-re-sparse.cc, ov-re-sparse.h, ov-scalar.cc, ov-scalar.h,
	ov-str-mat.cc, ov-str-mat.h, ov.h:
	Remove valid_as_scalar_index and valid_as_zero_index methods
	from octave_value, octave_base_value and all of its subclasses.

2008-12-26  Francesco Potortì  <pot@gnu.org>

	* data.cc (resize): Add cross reference to postpad.

2008-12-25  Jaroslav Hajek <highegg@gmail.com>

	* load-save.h (load_save_format_type, load_save_format_options): New
	enums.
	(load_save_format): Convert to a class.
	* load-save.cc (do_save(..., const octave_value&,...)): Use fmt.type
	in switch. Use options of LS_MAT_ASCII.
	(do_load, write_header): Use format.type in switch.
	* ls-mat-ascii.h (save_mat_ascii_data): Add tabs option.
	* ls-mat-ascii.cc (save_mat_ascii_data): Implement it.

2008-12-24  Ben Abbott  <bpabbott@mac.com>

	* DLD-FUNCTIONS/chol.cc: Compare results to 0 in tests.
	Delete spurious function calls in test block.

2008-12-24  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/eigs.cc: Use "testif HAVE_ARPACK" to skip tests if
	ARPACK is not available.

	* Makefile.in (DLD_XSRC): Add eigs.cc to the list.

	* load-save.cc (parse_save_options): Handle -double.

2008-12-23  David Bateman  <dbateman@free.fr>

	* DLD-FUNCTIONS/eigs.cc: New file.
	* Makefile.in (DLD_XSRC): Add it here.

2008-12-22  David Bateman  <dbateman@free.fr>

	* DLD-FUNCTIONS/__voronoi__.cc (F__voronoi__): Resize AtInf array
	before returning it.

2008-12-12  David Bateman  <dbateman@free.fr>

	* DLD-FUNCTIONS/chol.cc (Fcholinv): Add test.

2008-12-12  Jaroslav Hajek  <highegg@gmail.com>

	* ls-hdf5.cc (add_hdf5_data): Check for diagonal & permutation
	matrices and convert them to full prior to saving.

2008-12-11  Jaroslav Hajek  <highegg@gmail.com>
	
	* ov-re-mat.cc (Fdouble): Handle diagonal & perm matrices. 
	* ov-flt-re-mat.cc (Fsingle): Handle diagonal & perm matrices. 

2008-12-11  Jaroslav Hajek  <highegg@gmail.com>

	* ov-base-diag.cc (octave_base_diag<DMT,MT>::subsasgn): New method.
	* ov-base-diag.h (octave_base_diag<DMT,MT>::subsasgn): Declare it.
	(octave_base_diag<DMT,MT>::chk_valid_scalar): New method decl.

	* ov-re-diag.cc (octave_diag_matrix::chk_valid_scalar): New method
	override.
	* ov-re-diag.h: Declare it.
	* ov-flt-re-diag.cc (octave_float_diag_matrix::chk_valid_scalar): New
	method override.
	* ov-flt-re-diag.h: Declare it.
	* ov-cx-diag.cc (octave_complex_diag_matrix::chk_valid_scalar): New 
	method override.
	* ov-cx-diag.h: Declare it.
	* ov-flt-cx-diag.cc (octave_float_complex_diag_matrix::chk_valid_scalar): 
	New method override.
	* ov-flt-cx-diag.h: Declare it.

2008-12-10  Jaroslav Hajek  <highegg@gmail.com>

	* DLD-FUNCTIONS/expm.cc: Remove.
	* Makefile.in: Update.

2008-12-10  Jaroslav Hajek  <highegg@gmail.com>

	* ov-intx.h (OCTAVE_VALUE_INT_SCALAR_T::empty_clone): Construct an
	empty matrix instead of zero.

2008-12-07  Thorsten Meyer  <thorsten.meyier@gmx.de>

        * mappers.cc (Ftolower): Make lower alias of tolower, add tests
        * mappers.cc (Ftoupper): Make upper alias of toupper, add tests

2008-12-09  Jaroslav Hajek  <highegg@gmail.com>

	* DLD-FUNCTIONS/balance.cc (Fbalance): Exploit the new AEPBAL functionality.

2008-12-08  Jaroslav Hajek  <highegg@gmail.com>
	
	* xpow.cc ( xpow (const DiagMatrix& a, double b), 
	xpow (const DiagMatrix& a, const Complex& b), 
	xpow (const ComplexDiagMatrix& a, double b), 
	xpow (const ComplexDiagMatrix& a, const Complex& b), 
	xpow (const FloatDiagMatrix& a, float b), 
	xpow (const FloatDiagMatrix& a, const FloatComplex& b), 
	xpow (const FloatComplexDiagMatrix& a, float b), 
	xpow (const FloatComplexDiagMatrix& a, const FloatComplex& b)):
	New methods.
	* xpow.h: Declare them.
	* OPERATORS/op-dms-template.cc: Support diagonal matrix ^ scalar.

2008-12-08  Jaroslav Hajek  <highegg@gmail.com>

	* ov-re-diag.cc (octave_diag_matrix::save_binary,
	octave_diag_matrix::load_binary): New methods.
	* ov-re-diag.h: Declare them.
	* ov-flt-re-diag.cc (octave_float_diag_matrix::save_binary,
	octave_float_diag_matrix::load_binary): New methods.
	* ov-flt-re-diag.h: Declare them.
	* ov-cx-diag.cc (octave_complex_diag_matrix::save_binary,
	octave_complex_diag_matrix::load_binary): New methods.
	* ov-cx-diag.h: Declare them.
	* ov-flt-cx-diag.cc (octave_float_complex_diag_matrix::save_binary,
	octave_float_complex_diag_matrix::load_binary): New methods.
	* ov-flt-cx-diag.h: Declare them.
	* ov-perm.cc (octave_perm_matrix::save_binary,
	octave_perm_matrix::load_binary): New methods.
	* ov-perm.h: Declare them.

2008-12-06  Jaroslav Hajek  <highegg@gmail.com>

	* ov-fcn-handle.cc (octave_fcn_handle::load_binary): Call istream::get
	rather than istream::read to supply null terminating character.

2008-10-29  Jaroslav Hajek  <highegg@gmail.com>
	* file-io.cc, gl-render.cc, load-save.cc, ls-hdf5.cc, ls-mat4.cc,
	ls-mat5.cc, ls-oct-binary.cc, mex.cc, oct.h, ov-base-int.cc,
	ov-bool-mat.cc, ov-bool-sparse.cc, ov-cell.cc, ov-class.cc,
	ov-cx-mat.cc, ov-cx-sparse.cc, ov-fcn-handle.cc, ov-fcn-inline.cc,
	ov-flt-cx-mat.cc, ov-flt-re-mat.cc, ov-re-mat.cc, ov-re-sparse.cc,
	ov-str-mat.cc, ov-struct.cc, toplev.cc,
	DLD-FUNCTIONS/__lin_interpn__.cc, DLD-FUNCTIONS/amd.cc,
	DLD-FUNCTIONS/ccolamd.cc, DLD-FUNCTIONS/cellfun.cc,
	DLD-FUNCTIONS/colamd.cc, DLD-FUNCTIONS/gcd.cc,
	DLD-FUNCTIONS/matrix_type.cc, DLD-FUNCTIONS/regexp.cc,
	DLD-FUNCTIONS/symbfact.cc, DLD-FUNCTIONS/symrcm.cc: 
	Include oct-locbuf.h.

2008-12-05  Jaroslav Hajek  <highegg@gmail.com>

	* ov-base-diag.cc (save_ascii, load_ascii): Save natively.
	* ov-base-diag.h (load_ascii): Declare.
	* ov-perm.cc (save_ascii, load_ascii): Save natively.
	* ov-perm.h (load_ascii): Declare.


2008-12-04  Thorsten Meyer  <thorsten.meyier@gmx.de>

        * strfns.cc (Fchar): Add test from str2mat.m
        
2008-12-04  Jaroslav Hajek  <highegg@gmail.com>

	* ov.h (octave_value::is_perm_matrix): New method.
	* ov-base.h (octave_base_value::is_perm_matrix): New method.
	* ov-perm.h (octave_perm_matrix::is_perm_matrix): New method.
	* DLD-FUNCTIONS/inv.cc (Finv): Handle permutation matrices specially,
	compute rcond for diagonal matrices.
	* DLD-FUNCTIONS/det.cc (Fdet): Handle permutation & diagonal matrices
	specially.

2008-12-03  Jaroslav Hajek  <highegg@gmail.com>

	* ov-perm.h: New source.
	* ov-perm.cc: New source.
	* ov-flt-perm.h: New source.
	* ov-flt-perm.cc: New source.
	* ov-base-diag.cc (octave_base_diag<DMT, MT>::do_index_op):
	If subscripting an identity matrix by permutation(s), return a
	permutation matrix object.
	* ov.cc (octave_value::octave_value (const PermMatrix&)): New
	constructor.
	* ov.cc (octave_value::octave_value (const PermMatrix&)): Declare it.
	* op-cm-pm.cc: New source.
	* op-fcm-fpm.cc: New source.
	* op-fm-fpm.cc: New source.
	* op-fpm-fcm.cc: New source.
	* op-fpm-fm.cc: New source.
	* op-fpm-fpm.cc: New source.
	* op-m-pm.cc: New source.
	* op-pm-cm.cc: New source.
	* op-pm-m.cc: New source.
	* op-pm-pm.cc: New source.
	* op-pm-template.cc: New source.
	* Makefile.in: Include new sources.
	* DLD-FUNCTIONS/qr.cc (Fqr): Reflect interface changes of QR classes.

2008-12-03  John W. Eaton  <jwe@octave.org>

	* debug.cc (bp_table::do_get_breakpoint_list): Style fixes.

2008-12-01  Jaroslav Hajek  <highegg@gmail.com>

	* ov-base.h (octave_base_value::is_diag_matrix): New virtual method.
	* ops.h (CONCAT2, CONCAT3): New macros. Use CONCAT macros instead of 
	direct token pasting to avoid disabling argument prescan.
	* xdiv.cc, xdiv.h: Implement xdiv and xleftdiv overloads for diagonal
	and mixed dense-diagonal operands.

	* ov-re-diag.h: New source.
	* ov-re-diag.cc: New source.
	* ov-flt-re-diag.h: New source.
	* ov-flt-re-diag.cc: New source.
	* ov-base-diag.h: New source.
	* ov-base-diag.cc: New source.
	* OPERATORS/op-m-dm.cc: New source.
	* OPERATORS/op-m-cdm.cc: New source.
	* OPERATORS/op-fm-fdm.cc: New source.
	* OPERATORS/op-fm-fcdm.cc: New source.
	* OPERATORS/op-fdm-fs.cc: New source.
	* OPERATORS/op-fdm-fm.cc: New source.
	* OPERATORS/op-fdm-fdm.cc: New source.
	* OPERATORS/op-fdm-fcs.cc: New source.
	* OPERATORS/op-fdm-fcm.cc: New source.
	* OPERATORS/op-fdm-fcdm.cc: New source.
	* OPERATORS/op-fcm-fdm.cc: New source.
	* OPERATORS/op-fcm-fcdm.cc: New source.
	* OPERATORS/op-fcdm-fs.cc: New source.
	* OPERATORS/op-fcdm-fm.cc: New source.
	* OPERATORS/op-fcdm-fdm.cc: New source.
	* OPERATORS/op-fcdm-fcs.cc: New source.
	* OPERATORS/op-fcdm-fcm.cc: New source.
	* OPERATORS/op-fcdm-fcdm.cc: New source.
	* OPERATORS/op-dms-template.cc: New source.
	* OPERATORS/op-dm-template.cc: New source.
	* OPERATORS/op-dm-s.cc: New source.
	* OPERATORS/op-dm-m.cc: New source.
	* OPERATORS/op-dm-dm.cc: New source.
	* OPERATORS/op-dm-cs.cc: New source.
	* OPERATORS/op-dm-cm.cc: New source.
	* OPERATORS/op-dm-cdm.cc: New source.
	* OPERATORS/op-cm-dm.cc: New source.
	* OPERATORS/op-cm-cdm.cc: New source.
	* OPERATORS/op-cdm-s.cc: New source.
	* OPERATORS/op-cdm-m.cc: New source.
	* OPERATORS/op-cdm-dm.cc: New source.
	* OPERATORS/op-cdm-cs.cc: New source.
	* OPERATORS/op-cdm-cm.cc: New source.
	* OPERATORS/op-cdm-cdm.cc: New source.
	* Makefile.in: Include them.

	* ov-re-mat.cc (octave_matrix::diag): New method override.
	* ov-re-mat.h: Declare it.
	* ov-cx-mat.cc: Likewise with octave_complex_matrix.
	* ov-cx-mat.h: Likewise with octave_complex_matrix.
	* ov-flt-re-mat.cc: Likewise with octave_float_matrix.
	* ov-flt-re-mat.h: Likewise with octave_float_matrix.
	* ov-flt-cx-mat.cc: Likewise with octave_float_complex_matrix.
	* ov-flt-cx-mat.h: Likewise with octave_float_complex_matrix.
	* ov.cc (octave_value::octave_value (const DiagMatrix&))
	(octave_value::octave_value (const FloatDiagMatrix&))
	(octave_value::octave_value (const ComplexDiagMatrix&))
	(octave_value::octave_value (const FloatComplexDiagMatrix&)):
	Construct a diagonal matrix object.
	* data.cc (Fdiag): Support explicit dimensions. Fix tests.
	(Feye): Return diagonal matrix objects if possible. Fix tests.
	* mappers.cc (Freal, Fimag): Fix tests.
	* DLD-FUNCTIONS/inv.cc (Finv): Handle diagonal matrix objects.
	* ov-range.h (octave_range::diag): Declare only.
	* ov-range.cc (octave_range::diag): Return DiagMatrix if possible.

2008-11-25  Jaroslav Hajek  <highegg@gmail.com>
	* ov.cc (octave_value::is_equal): New member function.
	* ov.h: Declare it.
	* pt-select.cc (tree_switch_case::label_matches): Call
	octave_value::is_equal.

2008-12-04  Thomas Treichl  <Thomas.Treichl@gmx.net>

	* DLD-FUNCTIONS/cellfun.cc: Added more tests for cellfun.

2008-11-25  Jaroslav Hajek  <highegg@gmail.com>

	* ov-base.h (octave_base_value::type_conv_info): New class.
	* ov-base.h, ov-bool-mat.cc, ov-bool-mat.h, ov-bool-sparse.cc,
	ov-bool-sparse.h, ov-bool.cc, ov-bool.h, ov-complex.cc, ov-complex.h,
	ov-cx-mat.cc, ov-cx-mat.h, ov-null-mat.cc, ov-null-mat.h,
	ov-range.cc, ov-range.h, ov-re-mat.cc, ov-re-mat.h, ov-scalar.cc,
	ov-scalar.h, ov-str-mat.cc, ov-str-mat.h, ov.h:
	Return type_conv_info instead of type_conv_fcn from
	numeric_conversion_function and numeric_demotion_function
	methods.
	* ov-base.cc (octave_base_value::numeric_assign): Try biased conversion
	first if possible.
	* ov.cc (do_binary_op): Likewise. Also, search recursively.
	* ov.cc (do_cat_op): Likewise.

2008-11-24  David Bateman  <dbateman@free.fr>

	* graphics.cc (F__go_delete__): Check validity of handles before
	deleting them to avoid issues with callback function also deleting
	the handles.
	
2008-11-21  John Swensen  <jpswensen@comcast.net> 

	* debug.cc (bp_table::do_get_breakpoint_list): Avoid modifying
	bp_map while iterating over it.

2008-11-21  Jarkko Kaleva  <d3roga@gmail.com>

	* DLD-FUNCTIONS/eig.cc (Feig): Handle generalized eigenvalues and 
	eigenvectors.

2008-11-19  Jaroslav Hajek  <highegg@gmail.com>

	* DLD_FUNCTIONS/det.cc: Include only DET.h. Retrieve & matrix type &
	store it after calculation if possible.

2008-11-17  John W. Eaton  <jwe@octave.org>

	* load-path.cc (load_path::dir_info::update): Simplify previous
	change.

	* load-path.h (load_path::dir_info::abs_dir_name): New data member.
	(load_path::dir_info::dir_info, load_path::dir_info::operator =):
	Copy abs_dir_name.
	(load_path::abs_dir_cache_type, load_path::dir_cache_iterator,
	load_path::const_dir_cache_iterator): New typedefs.
	(load_path::abs_dir_cache): New static data member.
	(load_path::dir_info:dir_info): New constructor.
	* load-path.cc (load_path::abs_dir_cache): Define new static data
	member.
	(load_path::dir_info::update): Look in abs_dir_cache for relative
	directory names.
	(load_path::dir_info::initialize): Update abs_dir_cache here.

2008-11-13  John W. Eaton  <jwe@octave.org>

	* ov-int8.h, ov-int16.h, ov-int32.h, ov-int64.h, ov-uint8.h,
	ov-uint16.h, ov-uint32.h, ov-uint64.h: Don't define
	OCTAVE_INT_NDARRAY_T.
	* ov-intx.h (class OCTAVE_VALUE_INT_MATRIX_T,
	class OCTAVE_VALUE_INT_SCALAR_T): Use intNDArray<OCTAVE_INT_T>
	instead of OCTAVE_INT_NDARRAY_T.
	(OCTAVE_VALUE_INT_MATRIX_T::increment,
	OCTAVE_VALUE_INT_MATRIX_T::decrement,
	OCTAVE_VALUE_INT_SCALAR_T::increment,
	OCTAVE_VALUE_INT_SCALAR_T::decrement): Convert 1 to OCTAVE_INT_T.

2008-11-15  Thorsten Meyer  <thorsten.meyier@gmx.de>

        * strfns.cc (Fchar): adapt to conserve empty strings, add tests
        
2008-11-12  Michael Goffioul  <michael.goffioul@gmail.com>

	* dirfcn.cc (Ffilesep): Make it return all file separators when 'all'
	is given as argument.

	* DLD-FUNCTIONS/getrusage.cc: Undefine min/max.
	* TEMPLATE-INST/Array-os.cc: Declare Array<bool> as extern template.

2008-11-12  David Bateman  <dbateman@free.fr>

	* OPERATORS/op-cell.cc (op_catop_matrix_cell): Cast args in the correct
	order.
	* OPERATORS/op-struct.cc (op_catop_matrix_struct): Ditto.

2008-11-11  Jaroslav Hajek <highegg@gmail.com>

	* pt-assign.cc: Fix handling of empty cs-lists in assignment LHS.

2008-11-11  John W. Eaton  <jwe@octave.org>

	* lex.l (<MATRIX_START>{S}+): Don't insert separator if next token
	is a separator.

	* parse.y (F__parser_debug_flag__): New function.

2008-10-31  Jaroslav Hajek <highegg@gmail.com>

	* xnorm.cc: New source.
	* xnorm.h: New header file.
	* Makefile.in: Include xnorm.cc in the build process.
	* data.cc (Fnorm): Call xnorm, xcolnorms, xrownorms or xfrobnorm
	to do the actual work.


2008-10-31  David Bateman  <dbateman@free.fr>

	* ov-base-scalar.h (octave_value octave_base_scalar::reshape 
	(const dim_vector&) const): New method.

2008-10-30  John W. Eaton  <jwe@octave.org>

	* oct-map.cc (Octave_map::index): Copy key_list.

2008-10-31  David Bateman  <dbateman@free.fr>

	* graphics.h.in (void image::init (void)): Add a constraint for
	logical cdata.
	
	* graphics.h.in (base_property::delete_listener): New method.
	(property::delete_listener): New method.
	(base_graphics_object::delete_listener): New method.
	(base_graphics_object::delete_property_listener): New method.
	(base_graphics_object::remove_all_listeners): New method.
	(graphics_object::delete_property_listener): New method.
	(axes::set_defaults): Call remove_all_listeners.
	* graphics.cc (void base_properties::delete_listener): New method
	(void base_graphics_object::remove_all_listeners (void)): New method
	(Fdellistener): New command to remove listener functions associated
	with a property.
	
2008-10-30  David Bateman  <dbateman@free.fr>

	* graphics.h.in (axes::properties): Add keyreverse property.
	* graphics.cc (axes::properties::set_defaults): Initialize
	keyreverse property.

2008-10-28  Jaroslav Hajek <highegg@gmail.com>

	* Cell.h (Cell::assign (const Array<idx_vector>&, ...),
	Cell::delete_elements (const Array<idx_vector>&, ...)):
	New member functions.
	* Cell.cc (Cell::assign (const octave_value_list&, ...),
	Cell::delete_elements (const octave_value_list&, ...)):
	Call Array<T>::assign.
	* DLD-FUNCTIONS/dlmread.cc: Call Array<T>::resize_fill.
	* ov-base-mat.cc (octave_base_matrix::assign): Call Array<T>::assign.
	(octave_base_matrix::delete_elements):: Call Array<T>::delete_elements.
	* ov-cell.cc (Fcell): Call Array<T>::chop_trailing_singletons,
	simplify.
	* ov-cx-mat.cc (octave_complex_matrix::assign): Call Array<T>::assign.
	* ov-flt-cx-mat.cc (octave_float_complex_matrix::assign): Call
	Array<T>::assign.
	* ov-list.cc (octave_list::subsasgn): Call Array<T>::assign.
	* pr-output.cc (PRINT_ND_ARRAY): Use zero-based indices.


2008-10-29  Thorsten Meyer  <thorsten.meyier@gmx.de>

	* data.cc (Fcolumns): Remove "and" from @seealso string.
	* mappers.cc (Fgamma, Flgamma): Replace @seealso reference to
	deprecated gammai by gammainc.
	* DLD-FUNCTIONS/sqrtm.cc (Fsqrtm): Remove reference to Octave
	Forge function funm.
	* DLD-FUNCTIONS/qz.cc (Fqz): removed @seealso reference to Octave
	Forge function dare.
	* DLD-FUNCTIONS/rcond.cc (Frcond): removed @seealso reference to
	mldivide.
	* debug.cc (Fdbstop): Fix @seealso references and function name in
	docstring of dbnext.
	* DLD-FUNCTIONS/cellfun.cc (Fcellfun): Remove @seealso reference
	to isclass.
	* syscalls.cc (Fwaitpid): Replace WCONTINUED by WCONTINUE.
	* oct-hist.cc (Fhistory_size, Fsaving_history): Fix @seealso
	references to history_timestamp_format_string.
	* DLD-FUNCTIONS/luinc.cc (Fluinc): Remove reference to missing
	function cholinc.
        
2008-10-29  Jaroslav Hajek  <highegg@gmail.com>

	* DLD-FUNCTIONS/qr.cc (Fcholinsert, Fcholdelete, Fcholshift): Fix
	inline docs.

2008-10-28  John W. Eaton  <jwe@octave.org>

	* parse.y (finish_function): Clear local variables in function scope.
	* symtab.h (symbol_table::clear_variables): New argument, scope.

2008-10-28  Brian Gough  <bjg@gnu.org>

	* DLD-FUNCTIONS/besselj.cc: Added tests.

2008-10-23  John W. Eaton  <jwe@octave.org>

	* oct-hist.c (initialize_history): New arg, read_history_file)
	* oct-hist.h: Fix decl.
	* octave.cc (octave_main): Set read_history_file to false if
	--no-history option is specified.  Pass read_history_file to
	initialize_history.

	* DLD-FUNCTIONS/fltk_backend.cc: Update initialization comment.

2008-10-22  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/fltk_backend.cc (figure_manager::do_close_all):
	Rename from close_all.  Now private.
	(figure_manager::do_new_window): Rename from new_window.
	Now private.
	(figure_manager::do_delete_window): Rename from delete_window.
	Now private.
	(figure_manager::do_mark_modified): Rename from mark_modified.
	Now private.
	(figure_manager::do_get_size): Rename from get_size.  Now private.
	(figure_manager::close_all, figure_manager::do_new_window, 
	figure_manager::do_delete_window,
	figure_manager::do_mark_modified, figure_manager::do_get_size):
	New static functions.
	(figure_manager::instance_ok): New function.
	(figure_manager::Instance): Delete.  Eliminate all uses in favor
	of static methods that call instance_ok and forward to private
	member functions.
	(figure_manager::instance): New static data member.
	(figure_manager::hnd2idx, figure_manager::figprops2idx,
	figure_manager::figprops2idx, figure_manager::str2idx): Now static.

	* DLD-FUNCTIONS/fltk_backend.cc (figure_manager::default_size):
	Increase default window size.

	* graphics.cc (figure::properties::remove_child): New function.
	* graphics.h.in: Provide decl.

	* gl-render.cc (opengl_renderer::draw): Get all children.

	* DLD-FUNCTIONS/fltk_backend.cc (__fltk_redraw__):
	Check for "root" object, not "root_figure".

	* graphics.h.in (base_properties::get_all_children): New function.

2008-10-22  David Bateman  <dbateman@free.fr>

	* graphics.h.in (line::properties, surface::properties): Add data
	source properties.

2008-10-21  John W. Eaton  <jwe@octave.org>

	* graphics.h.in (axes::properties::delete_children): Delete decl.

	* graphics.h.in (gh_manager::is_handle_visible): New function.
	(axes::properites): Move title property after label properties.
	(graphics_object::is_handle_visible): New function.
	(base_properties::is_handle_visible): New function.
	(base_properties): Mark children property with G.
	* graphics.cc (axes::properties::set_text_child): New function.
	(axes::properties::set_title, axes::properties::set_xlabel,
	axes::properties::set_ylabel, axes::properties::set_zlabel): Use it.
	(base_property::get_children): New function.  Only
	return handles that are visible.
	(axes::properties::delete_text_child): Set handlevisibility to
	"off" for newly created object and add it to the list of children.
	Call base_properties::remove_child on previous handle value.
	(axes::properties::set_defaults): Don't explicitly delete xlabel,
	ylabel, zlabel, and title handles.  Reinitialize them after
	deleting children.
	(axes::properties::remove_child): Don't explicitly delete xlabel,
	ylabel, zlabel, and title handles.
	(axes::properties::init): Move here from graphics.h.in.  Set
	handlevisibility to "off" for xlabel, ylabel, zlabel, and title
	and add them to the list of children.
	(axes::properties::delete_children): Delete.

	* genprops.awk: Allow whitespace between BEGIN_PROPERTIES and
	opening paren.  Accept optional second argument for graphics
	object name.
	* graphics.h.in (root_figure properties): Pass "root" as second
	arg to BEGIN_PROPERTIES.

2008-10-16  John W. Eaton  <jwe@octave.org>

	* graphics.cc (make_handle_fraction): New static function.
	(gh_manager::get_handle): Use it.
	(gh_manager::do_free): Call make_handle_fraction to replace
	fractional part of non-figure handles.

	* graphics.cc (base_properties::remove_child): Handle children as
	a column vector instead of a row vector.

	* utils.cc (Fis_absolute_filename, Fis_rooted_relative_filename,
	Fmake_absolute_filename, Ffind_dir_in_path): New functions.

2008-10-16  David Bateman  <dbateman@free.fr>

	* graphics.cc (void axes::properties::set_defaults 
	(base_graphics_object&,	const std::string&)): Preserve font
	and position properties if the axis is "replaced".

2008-10-16  John W. Eaton  <jwe@octave.org>

	* graphics.h.in (class axes::properties): New property: interpreter.

2008-10-15  David Bateman  <dbateman@free.fr>

	* ov-class.c (Fsuperiorto, Finferiorto): Allow more than one class
	to be defined at a time.

2008-10-12  David Bateman  <dbateman@free.fr>

	* pt-colon.cc (octave_value tree_colon_expression::make_range 
	(const octave_value&, const octave_value&, const octave_value&)):
	Treating class overloading of colon operator.
	(octave_value tree_colon_expression::rvalue (void)): Ditto.

2008-10-10  John W. Eaton  <jwe@octave.org>

	* graphics.h.in (base_properties::adopt): Place new child at front
	of list, not end.  Make list of children a column vector instead
	of a row vector.

	* graphics.cc (gh_manager::do_free): Set the beingdeleted
	property, then delete the children, then execute the deletefcn.
	(axes::properties::set_defaults, axes::properties::remove_child,
	axes::properites::delete_children): Call delete_text_child to
	manage the title, xlabel, ylabel, and zlabel properties.
	(axes::properties::get_title, axes::properties::get_xlabel,
	axes::properties::get_ylabel, axes::properties::get_zlabel): Delete.

	* graphics.h.in (axes::properites::title,
	axes::properites::xlabel, axes::properites::ylabel,
	axes::properites::zlabel): No longer mutable.  Don't generate
	factory default values or custom get functions for these
	properties.

	* graphics.cc (axes::properties::delete_text_child): New function.
	* graphics.h.in: Provide decl.

	* graphics.h.in (graphics_object::type): New function.

2008-10-10  David Bateman  <dbateman@free.fr>

	* graphics.h.in (radio_property axes::properties::__colorbar__):
	Delete.
	* graphics.cc (void axes::properties::sync_positions (void)):
	Disable code synchronizing outerposiiton and position.
	(void axes::properties::set_defaults (base_graphics_object&,
	const std::string&): Remove __colorbar__.
	(F__go_delete__): Don't delete if already being deleting to avoid
	recursion in callback functions.
	(F__go_axes_init__): Flag error if handle is deleted during
	initialization due.
	
2008-10-09  David Bateman  <dbateman@free.fr>

	* ov-class.cc (Octave_map octave_class::map_value (void) const): 
	Remove.
	(bool octave_class::save_ascii (std::ostream&), 
	bool octave_class::load_ascii (std::istream&),
	bool octave_class::save_binary (std::ostream&, bool&),
	bool octave_class::save_binary (std::ostream&, bool&),
	bool octave_class::save_hdf5 (hid_t, const char *, bool),
	bool octave_class::load_hdf5 (hid_t, const char *, bool)):
	Allow saving and reloading of classes.
	* ov-class. (Octave_map octave_class::map_value (void) const): 
	Remove.
	* ls-mat.cc (std::string read_mat5_binary_element (std::istream&, 
	const std::string&, bool, bool&, octave_value&), 
	int save_mat5_element_length (const octave_value&, const std::string&,
	bool, bool)): Allow saving and loading of classes. 
	* ls-oct-ascii.h (template <class T> bool extract_keyword 
	(std::istream&, const char *, T&, const bool)): Initialize value with
	null constructor rather than 0.
	
2008-10-07  David Bateman  <dbateman@free.fr>

	* graphics.cc (F__go_delete__): Allow arrays of graphic handles.

2008-10-01  Jaroslav Hajek <highegg@gmail.com>

	* OPERATORS/op-i64-i64.cc: Instantiate all 64-bit operations.
	* OPERATORS/op-ui64-ui64.cc: Instantiate all unsigned 64-bit operations.

2008-10-06  John W. Eaton  <jwe@octave.org>

	* ov-type-conv.h (octave_type_conv_body): Avoid shadow warning
	from GCC.

2008-10-06  David Bateman  <dbateman@free.fr>

	* graphics.cc (static bool is_handle (const octave_value&)): Delete.
	(static octave_value is_handle (const octave_value&)): New function.

2008-10-03  Jaroslav Hajek  <highegg@gmail.com>

	* Cell.cc (Cell::delete_elements): Don't pass second arg to
	Array<T>::maybe_delete_elements.
	* ov-base-mat.cc (octave_base_matrix<MT>::delete_elements): Ditto.

2008-10-02  Jaroslav Hajek  <highegg@gmail.com>

	* Cell.h (Cell::delete_elements): Declare new member function.
	* Cell.h (Cell::delete_elements): Define it.
	* oct-map.cc (Octave_map::maybe_delete_elements): Call delete_elements
	instead of assign with empty Cell.

2008-10-02  John W. Eaton  <jwe@octave.org>

	* pt-arg-list.cc (tree_argument_list::convert_to_const_vector):
	Don't insert undefined elements in return list.

2008-09-30  Jaroslav Hajek  <highegg@gmail.com>

	* ov-str-mat.h (octave-char-matrix-str::assign): Remove declaration.
	* ov-str-mat.cc (octave-char-matrix-str::assign): Remove definition.

2008-09-30  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/__magick_read__.cc: Only define
	magick_to_octave_value if HAVE_MAGICK is defined.

2008-09-29  John W. Eaton  <jwe@octave.org>

	* symtab.cc (symbol_table::fcn_info::fcn_info_rep::load_class_method):
	Call load_class_constructor if name and dispatch_type are the same.

2008-09-29  David Bateman  <dbateman@free.fr>

	* symtab.cc (octave_value symbol_table::find_function 
	(const std::string&, tree_argument_list *, const string_vector&,
	octave_value_list&, bool&)): If first character of function name
	is "@" then look for class specific method.

2008-09-26  John W. Eaton  <jwe@octave.org>

	* symtab.cc (out_of_date_check_internal):
	Fix order of arguments in call to load_path::find_method.

2008-09-26  David Bateman  <dbateman@free.fr>

	* ov-class.h (idx_vector index_vector (void) const): Declare new
	maethod.
	* ov-class.cc (idx_vector index_vector (void) const): Define new
	method.
	* (Fismethod): New function.

2008-09-26  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/urlwrite.cc (urlwrite_cleanup_file) New function.
	(urlwrite_delete_file, urlwrite_filename): New static variables.
	(Furlwrite): Only return filename if urlget succeeds.  Use
	unwind_protect to delete files we create if download fails.

2008-09-26  Jaroslav Hajek  <highegg@gmail.com>

	* ov-null-mat.h: New header file.
	* ov-null-mat.cc: New source.
	* ov.h (octave_value::is_null_value, octave_value::non_null_value,
	octave_value::make_non_null_value):
	Declare new member functions.
	* ov.cc (octave_value:non_null_value, octave_value::make_non_null_value): 
	Define them.
	(octave_value::assign (assign_op op, const octave_value& rhs)): 
	(register_types): Register null types.
	Call non_null_value ().
	* oct-obj.cc (octave_value_list::normalize_null_values): New member
	function. 
	* oct-obj.h (octave_value_list::normalize_null_values): Declare it.
	* ov-base.h (is_null_value): New virtual member function.
	* ops.h (NULLASSIGNOPDECL, DEFNULLASSIGNOP_FN): New macros.
	* ov-base-mat.cc (octave_base_mat<MT>::delete_elements): New member func.
	* ov-base-mat.h: Declare it.
	* ov-base-sparse.cc (octave_base_sparse<MT>::delete_elements): New member func.
	* ov-base-sparse.h: Declare it.
	* ov-cell.cc (octave_cell:subsasgn): Handle null values.	
	* ov-struct.cc (octave_cell:subsasgn): Handle null values.
	* ov-builtin.cc (octave_builtin::do_multi_index_op): Normalize return
	values.
	* pt-misc.cc (tree_parameter_list::define_from_arg_vector): Call
	octave_lvalue::define instead of octave_lvalue::assign.
	* pt-decl.h (tree_decl_elt::rvalue): Call non_null_value ().
	* OPERATORS/op-int.h (OCTAVE_INT_NULL_ASSIGN_OPS,
	OCTAVE_INSTALL_INT_NULL_ASSIGN_OPS): New macros.
	* OPERATORS/op-m-m.cc: Install & define assignment & conversion operators.
	* OPERATORS/op-bm-bm.cc: Ditto.
	* OPERATORS/op-cell.cc: Ditto.
	* OPERATORS/op-cm-cm.cc: Ditto.
	* OPERATORS/op-cs-cs.cc: Ditto.
	* OPERATORS/op-fcm-fcm.cc: Ditto.
	* OPERATORS/op-fcs-fcs.cc: Ditto.
	* OPERATORS/op-fm-fm.cc: Ditto.
	* OPERATORS/op-fs-fs.cc: Ditto.
	* OPERATORS/op-int.h: Ditto.
	* OPERATORS/op-m-m.cc: Ditto.
	* OPERATORS/op-range.cc: Ditto.
	* OPERATORS/op-s-s.cc: Ditto.
	* OPERATORS/op-scm-scm.cc: Ditto.
	* OPERATORS/op-sm-sm.cc: Ditto.
	* OPERATORS/op-str-str.cc: Ditto.
	* OPERATORS/op-i16-i16.cc: Ditto.
	* OPERATORS/op-i32-i32.cc: Ditto.
	* OPERATORS/op-i64-i64.cc: Ditto.
	* OPERATORS/op-i8-i8.cc: Ditto.
	* OPERATORS/op-ui16-ui16.cc: Ditto.
	* OPERATORS/op-ui32-ui32.cc: Ditto.
	* OPERATORS/op-ui64-ui64.cc: Ditto.
	* OPERATORS/op-ui8-ui8.cc: Ditto.

2008-09-25  David Bateman  <dbateman@free.fr>

	* pt-mat.cc (class tm_row_const): Add any_class test
	(class tm_const): Ditto.
	(octave_value tree_matrix::rvalue (void)): If any object to
	concatenate is a class object, dispatch to the appropriate
	vertcat/horzcat function.

2008-09-25  John W. Eaton  <jwe@octave.org>

	* symtab.cc (symbol_table::do_find): Don't set evaluated_args and
	args_evaluated here, prior to call to symbol_table::fcn_info::find.

2008-09-24  Søren Hauberg  <hauberg@gmail.com>

	* DLD-FUNCTIONS/__magick_read__.cc (magick_to_octave_value): New
	template function with specializations for various
	GraphicsMagick++ types.
	(F__magick_finfo__): New function.

2008-09-24  John W. Eaton  <jwe@octave.org>

	* ov-usr-fcn.cc (octave_user_function::do_multi_index_op):
	Add symbol_table::clear_variables cleanup function to the
	unwind_protect stack after the parameter list cleanup functions.

	* load-path.cc (load_path::do_initialize): Check for OCTAVE_PATH
	in the environment, not OCTAVE_LOADPATH.

2008-09-24  Thomas Weber  <thomas.weber.mail@gmail.com>

	* DLD-FUNCTIONS/regexp.cc (octregexp_list): Retry if match limit
	is exceeded.

2008-09-24  John W. Eaton  <jwe@octave.org>

	* genprops.awk (emit_declarations): Emit call to set_mode before
	updaters.

2008-09-23  John W. Eaton  <jwe@octave.org>

	* pt-arg-list.cc (F__end__): Dispatch to user-defined end
	function for classes if one is defined.
	* lex.l (is_keyword_token): Allow "end" as a functon name.

2008-09-23  David Bateman  <dbateman@free.fr>

	* variables.cc (static octave_value do_who (int, const string_vector&,
	bool, bool, std::string): Add final message argument and simple
	treatment of the "-file" option.

2008-09-22  David Bateman  <dbateman@free.fr>

	* debug.cc (static octave_user_code * get_user_code 
	(const std::string&)): Only check user code as break points can't
	be set in builtins or oct-files.
	(bp_table::intmap bp_table::do_remove_all_breakpoints_in_file 
	(const std::string&, bool)): Add flag to silence the error message 
	from this function if a user code with breakpoints is not found.
	(bp_table::fname_line_map bp_table::do_get_breakpoint_list (const 
	octave_value_list&)): Do an ourt of date check on the function
	before checking the breakpoints.
	* debug.h (do_remove_all_breakpoints_in_file, 
	remove_all_breakpoints_in_file): Add flag to silence error
	message.
	* symtab.cc (out_of_date_check_internal): Clear breakpoints in
	function if out_of_date. split into two versions taking the 
	octave_function pointer seperately or not.
	* symtab.h (bool out_of_date_check (octave_function*)): New function.
	
2008-09-18  David Bateman  <dbateman@free.fr>

	* DLD-FUNCTIONS/fftw.cc (Ffftw): Clarify the documentation.

2008-09-17  Krzesimir Nowak  <qdlacz@gmail.com>

	* debug.cc (Fisdebugmode): New function.

2008-09-17  David Bateman  <dbateman@free.fr>

	* data.cc (do_cat): Chopping trailing singleton dimensions.

	* data.cc (SINGLE_TYPE_CONCAT, DO_SINGLE_TYPE_CONCAT): New macros
	(do_cat): Special case single type concatenations for speed.
	* pt.mat.cc (std::string get_concat_class (const std::string&,
	const std::string&), void maybe_warn_string_concat (bool, bool)):
	Remove static declaration.
	* pt-mat.h (std::string get_concat_class (const std::string&,
	const std::string&), void maybe_warn_string_concat (bool, bool)):
	Define extern here.
	
	* DLD-FUNCTIONS/sparse.cc (Fsparse): Clarify the help string.

2008-09-10  John W. Eaton  <jwe@octave.org>

	* octave.cc (octave_main): Make all command-line arguments
	available to startup scripts and PKG_ADD files.

2008-09-09  David Bateman  <dbateman@free.fr>

	* DLD-FUNCTIONS/regexp.cc (octregexp_list): Distinguish between
	matlab named tokens and perl lookbehind expressions. For
	lookbehind expression replace "*" and "+" with a limited number of
	fixed length expressions to simulate arbitrary length look behind.

2008-09-08  Michael Goffioul  <michael.goffioul@gmail.com>

	* DLD-FUNCTIONS/__magick_read__.cc (encode_uint_image): Resolve "pow"
	ambiguity.

	* graphics.cc (base_properties::get_dynamic,
	base_properties::set_dynamic, base_propertyes::get_property_dynamic):
	Add the 3rd template parameter to std::map, corresponding to all_props
	field.

2008-09-08  John W. Eaton  <jwe@octave.org>

	* ls-oct-ascii.cc (std::string extract_keyword (std::istream&,
	const char *, const bool)): Also accept \r as line ending character.
	* ls-oct-ascii.h (template <class T> bool extract_keyword
	(std::istream&, const char *, T&, const bool)): Likewise.
	(template <class T> bool extract_keyword (std::istream&, const
	string_vector&, std::string&, T&, const bool)): Likewise.

2008-09-03  Ben Abbott  <bpabott@mac.com>

	* mappers.cc: Use eps(single) as tolerance for single precesion test.

2008-09-03  John W. Eaton  <jwe@octave.org>

	* variables.cc (mlock): Lock current function on the call stack.
	(Fmlock): Lock caller.

	* symtab.cc (out_of_date_check_internal): If dispatch type is set,
	check for method.  If that fails, check for regular function. 

2008-09-02  Michael Goffioul  <michael.goffioul@gmail.com>

	* graphics.cc (hggroup::update_axis_limits): Also reacts on
	[xyzac]liminclude changes.

2008-08-31  Michael Goffioul  <michael.goffioul@gmail.com>

	* DLD-FUNCTIONS/fltk_backend.cc (fltk_backend::redraw_figure,
	fltk_backend::print_figure): Change graphics_handle argument to
	graphics_object.

2008-08-29  David Bateman  <dbateman@free.fr>

	* src/graphics.cc: Replace int by octave_idx_type in multiple
	places.
	(array_property::is_equal): In macro CHECK_ARRAY_EQUAL, special
	case scalar case and keep copy of array value to avoid early
	destruction of converted array values.
	
2008-08-28  David Bateman  <dbateman@free.fr>

	* graphics.h.in (root_figure::properties::showhiddenhandles):
	New property.

2008-08-27  Jaroslav Hajek  <highegg@gmail.com>

	* DLD-FUNCTIONS/__glpk__.cc (F__glpk__): Initialize output vectors 
	by NA value.

2008-08-26 Ben Abbott  <bpabott@mac.com>

	* mappers.cc: Adjust tolerance for single precesion tests.

	* DLD-FUNCTIONS/chol.cc: Double tolerance.

2008-08-26  Jaroslav Hajek  <highegg@gmail.com>

	* parse.y (Fassignin): Use default  scope when calling varref.

2008-08-26  Michael Goffioul  <michael.goffioul@gmail.com>

	* genprops.awk (emit_declarations): Call updaters/listeners only when
	'set' method returned true.
	* graphics.h.in (base_property::set, base_property::do_set):
	Return bool value.
	(property::set): Likewise.
	(string_property::do_set): Check value change and return bool value.
	(radio_property::do_set, double_property::do_set,
	array_property::do_set): Likewise.
	(color_property::do_set, double_radio_property::do_set,
	row_vector_property::do_set, bool_property::do_set,
	handle_property::do_set): Return bool value.
	(any_property::do_set, callback_property::do_set): Return always true.
	(color_values::operator==, color_values::operator!=): Add.
	(array_property::is_equal): Add.
	* graphics.cc (base_property::set): Executes listeners/notifiers only
	when do_set returns true. Return bool value.
	(color_property::do_set, double_radio_property::do_set): Check value
	change and return bool value.
	(array_property::is_equal): Add.

	* genprops.awk (emit_declarations, emit_source): Change code emission
	when emitting base_properties code (base is 1).
	(BEGIN): Initialize pcount to 0.
	(BEGIN_PROPERTIES): Initialize base to 0.
	(BEGIN_BASE_PROPERTIES): New statement.
	(main): Add support for 'f' modifier. Add support for non xxx_property
	typed properties.
	* graphics.h.in (base_properties::base_properties): Remove
	implementation.
	(base_properties::get_tag, base_properties::get_parent,
	base_properties::get_type, base_properties::get___myhandle__,
	base_properties::get_busyaction, base_properties::get_buttondownfcn,
	base_properties::is_clipping, base_properties::get_clipping,
	base_properties::execute_createfcn, base_properties::get_createfcn,
	base_properties::execute_deletefcn, base_properties::get_deletefcn,
	base_properties::get_handlevisibility, base_properties::is_hittest,
	base_properties::get_hittest, base_properties::is_interruptible,
	base_properties::get_interruptible, base_properties::is_selected,
	base_properties::get_selected, base_properties::is_selectionhighlight,
	base_properties::get_selectionhighlight,
	base_properties::get_uicontextmenu, base_properties::get_userdata,
	base_properties::is_visible, base_properties::get_visible,
	base_properties::is_beingdeleted, base_properties::get_beingdeleted,
	base_properties::set_busyaction, base_properties::set_buttondownfcn,
	base_properties::set_clipping, base_properties::set_createfcn,
	base_properties::set_deletefcn, base_properties::set_handlevisibility,
	base_properties::set_hittest, base_properties::set_interruptible,
	base_properties::set_selected,
	base_properties::set_selectionhighlight,
	base_properties::set_uicontextmenu, base_properties::set_userdata,
	base_properties::set_visible, base_properties::set_beingdeleted,
	base_properties::get_children): Remove (now auto-generated).
	(base_properties::enum): Likewise.
	(base_properties::is_modified): Call is___modified__.
	(base_properties::set_modified): Call set___modified__.
	(base_properties::set_children): Add.
	(base_properties::set_dynamic, base_properties::get_dynamic,
	base_properties::get_property_dynamic): New methods.
	(class base_properties): Surround properties declaration with
	BEGIN_BASE_PROPERTIES/END_PROPERTIES pair.
	* graphics.cc (base_properties::get, base_properties::set,
	base_properties::get_property): Remove (now auto-generated).
	(base_properties::set_dynamic, base_properties::get_dynamic,
	base_properties::get_property_dynamic): New methods.
	(base_properties::set_children): Likewise.

	* DLD_FUNCTIONS/fltk_backend.cc (F__fltk_redraw__): Make static.
	(F__init_fltk__): Protect from mutliple invocations.
	(F__remove_fltk__): Likewise.
	(F__init_fltk__): Register input event hook.
	(F__remove_fltk__): Unregister input event hook.

	* DLD_FUNCTIONS/fltk_backend.cc (fltk_backend::close_figure): Remove.
	(fltk_backend::object_destroyed, fltk_backend::property_changed): New
	methods.
	* genprops.awk (emit_declarations): Generate "enum" property fields.
	(emit_sources): Emit set_id calls.
	* graphics.h.in (base_property::id): New field.
	(base_property::base_property): Initialize it.
	(base_property::get_id, base_property::set_id): Add accessors.
	(property::get_id, property::set_id): Likewise.
	(base_graphics_backend::close_figure, graphics_backend::close_figure):
	Remove methods.
	(base_graphics_backend::redraw_figure,
	base_graphics_backend::print_figure, graphics_backend::redraw_figure,
	graphics_backend::print_figure): Change graphics_handle argument into
	graphics_object.
	(base_graphics_backend::property_changed,
	base_graphics_backend::object_created,
	base_graphics_backend::object_destroyed,
	graphics_backend::property_changed, graphics_backend::object_created,
	graphics_backend::object_destroyed): New signature with
	graphics_object argument.
	(base_properties::base_properties): Add set_id calls.
	(class base_properties): Add "enum" property fields.
	(root_figure::properties::remove_child): Add overloaded method.
	(figure::properties::close): Remove.
	(figure::properties::set_backend): Call object_destroyed instead of
	close.
	(figure::~figure): Remove close call.
	(figure::properties::get_title): New method.
	* graphics.cc (base_property::set): Call property_changed only for
	valid id (>=0); Use graphics_object argument.
	(gh_manager::do_free): Call object_destroyed with graphics_object
	argument.
	(base_graphics_backend::property_changed,
	base_graphics_backend::object_created,
	base_graphics_object::object_destroyed): Implement wrappers for
	graphics_handle argument case.
	(gnuplot_backend::close_figure): Remove.
	(gnuplot_backend::object_destroyed,
	gnuplot_backend::property_changed): New methods.
	(gnuplot_backend::redraw_figure, gnuplot_backend::print_figure):
	Change graphics_handle argument to graphics_object.
	(root_figure::properties::remove_child): Add.
	(figure::properties::close): Remove.
	(figure::properties::set_position): Do not call set_figure_position.
	(figure::properties::get_title): New method.
	(gh_manager::do_make_graphics_handle): Call object_created with
	a graphics_object argument.
	(gh_manager::do_make_figure_handle): Likewise.
	(Fdrawnow): Call redraw_figure and print_figure with a graphics_object
	argument.

2008-08-26  Maciek Gajewski <maciej.gajewski0@gmail.com>

	* graphics.h.in (base_property::set): Remove inline implementation.
	(base_graphics_backend::property_changed,
	base_graphics_backend::object_created,
	base_graphics_backend::object_destroyed): New method.
	(graphics_backend::property_changed, graphics_backend::object_created,
	graphics_backend::object_destroyed): Likewise.
	* graphics.cc (base_property::set): Moved from header file.
	(gh_manager::do_free): Add backend notification.
	(gh_manager::do_make_graphics_handle): Likewise.
	(gh_manager::do_make_figure_handle): Likewise.

2008-08-25  Thomas L. Scofield  <scofield@calvin.edu>

	* DLD-FUNCTIONS/__magick_read__.cc: Untabify.
	(jpg_settings, encode_bool_image, encode_uint_image, encode_map):
	New functions.
	(write_image): Use them to handle various image types.

2008-08-25  David Bateman  <dbateman@free.fr>

	* graphics.cc (void gh_manager::do_execute_callback 
	(const graphics_handle&, const octave_value&, const
	octave_value&)): Don't pass the function handle as first arg of
	property listener functions

	* graphics.h.in (class base_properties): Add functor for caseless
	string comparison. Use it in the property map, so that user added
	properties are found in a case insensitive fashion.

2008-08-22  John W. Eaton  <jwe@octave.org>

	* symtab.h (symbol_table::inherit): Pass reference to symbol table
	to do_inherit instead of scope.
	(symbol_table::do_inherit): First arg is now reference to
	symbol_table, not scope id.  Insert all variables from donor scope.

2008-08-21  Thomas Treichl  <Thomas.Treichl@gmx.net>
	
	* mappers.cc: Increase test script tolerance.

2008-08-20  John W. Eaton  <jwe@octave.org>

	* help.cc (builtin_help): Go to next symbol name on error.
	(do_type, do_which): Omit separate path search for files.
	(help_from_symbol_table): Only insert Texinfo markup if it looks
	like the help message is already in Texinfo.

2008-08-19  David Bateman  <dbateman@free.fr>

	* load-path.cc (load-path::do_find_dir (const std:string&) const)):
	Method to find a directory on the load-path corresponding to the
	argument.
	* load-path.h (load-path::do_find_dir (const std:string&) const),
	load-path::find_dir (const std::string&) const): New methods.
	* utils.cc (std::string contents_file_in_path (const std::string&)):
	New function.
	* utils.h  (std::string contents_file_in_path (const std::string&)): 
	Declare it.
	* help.cc (static bool raw_help_from_file (const std::string&,
	std::string&, std::string&, bool&)): Also check is requested
	argument is a directory and contains the file Contents.m.

	* OPERATORS/op-int-conv.cc (DEFINTCONVFN): New macro that warn
	for integer conversion issues. Use it to replace DEFCONVFN.
	* OPERATORS/op-int.h (DEFINTBINOP_OP, DEFINTNDBINOP_OP,
	DEFINTBINOP_FN, DEFINTNDBINOP_FN): New macros that warn for
	integer truncation issues. Use them to replace the corresponding
	macros in the integer arithmetic macros. Update other integer
	arithmetic functions that don't use these macros individually.
	* error.cc (initialize_default_warning_state (void)): Initialize
	the default warning state for the integer warnings to be off.
	* gripes.cc (void gripe_binop_integer_math_truncated (const char *,
	const char *, const char *), void gripe_unop_integer_math_truncated
	(const char *, const char *), void gripe_non_integer_conversion
	(const char *, const char *), void gripe_nan_conversion (const char *,
	const char *)): Warning functions for integer conversion and math
	issues.
	* gripes.h (void gripe_binop_integer_math_truncated (const char *,
	const char *, const char *), void gripe_unop_integer_math_truncated
	(const char *, const char *), void gripe_non_integer_conversion
	(const char *, const char *), void gripe_nan_conversion (const char *,
	const char *)): Declare them.
	* ov-intx.h (OCTAVE_VALUE_INT_MATRIX_T::convert_gripe,
	OCTAVE_VALUE_INT_SCALAR_T::convert_gripe): Adapt for new means of
	flagging integer truncation.
	(OCTAVE_VALUE_INT_MATRIX_T::decrement,
	OCTAVE_VALUE_INT_MATRIX_T::increment,
	OCTAVE_VALUE_INT_SCALAR_T::decrement,
	OCTAVE_VALUE_INT_SCALAR_T::increment): Check for integer
	truncation.
	* ov.cc (convert_to_int_array): Adapt for new means of
	flagging integer truncation.

2008-08-19  Jaroslav Hajek  <highegg@gmail.com>

	* pt-assign.h (tree_simple_assignment::first_execution): New
	member field.
	(tree_simple_assignment::first_execution): Ditto.
	* pt-assign.cc (tree_simple_assignment::tree_simple_assignment):
	Initialize first_execution.
	(tree_multi_assignment::tree_multi_assignment): Ditto.
	(tree_simple_assignment::rvalue): Check for obsolete built-in
	variables only at first execution.
	(tree_multi_assignment::rvalue): Ditto.

	* DLD-FUNCTIONS/__glpk__.cc (F__glpk__): Checks whether LB and UB are
	of proper size.

	* oct-obj.cc, oct-obj.h (octave_value_list::make_argv): 
	Allow calling without fcn_name.
	* load-save.cc (parse_save_options (const string_vector&, ...)): 
	Return string_vector with parsed arguments removed instead.
	(parse_save_options (const std::string&, ...)):	Ditto.
	(Fsave): Fix calls to parse_save_options.

2008-08-19  John W. Eaton  <jwe@octave.org>

	* parse.y (Feval): Return value produced by evaluating CATCH string.

2008-08-12  John W. Eaton  <jwe@octave.org>

	* pt-fcn-handle.cc (tree_anon_fcn_handle::rvalue):
	Call stash_parent_function_scope for user function.
	* ov-fcn-inline.cc (octave_fcn_inline::octave_fcn_inline): 
	Likewise.

	* ov-struct.cc (octave_struct::subsref): Don't allow resizing for
	simple x(idx) case.

2008-08-07  John W. Eaton  <jwe@octave.org>

	* ov.cc (octave_value::idx_type_value): Don't include default
	argument values in definition.

	* src/glrender.h, src/zfstream.h: Don't include config.h.

	* oct-errno.h: Include <cerrno>.

	* Cell.cc, DLD-FUNCTIONS/regexp.cc, DLD-FUNCTIONS/dlmread.cc,
	debug.cc, error.cc, file-io.cc, graphics.cc, help.cc, input.cc,
	lex.l, load-path.cc, load-save.cc, ls-mat-ascii.cc,
	ls-oct-ascii.cc, oct-stream.cc, octave.cc, variables.cc,
	ov-fcn-handle.cc, parse.y, pr-output.cc, symtab.cc, sysdep.cc,
	utils.cc, graphics.h.in: Replace all uses of NPOS with
	std::string::npos.

	* ov.cc (octave_value::idx_type_value): Move definition here.
	* ov.h: From here.

	* DLD-FUNCTIONS/fsolve.cc (override_options): Don't fail if
	options_map does not contain an expected keyword.
	Fix typo in warning identifier.
	(make_unimplemented_options): Use CamelCase names here.

2008-08-06  Søren Hauberg  <hauberg@gmail.com>

	* error.cc (Ferror): Update format of error messages in exmple.
	* parse.y: (Feval): Likewise.

2008-08-06  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/fsolve.cc (hybrd_info_to_fsolve_info):
	Update INFO values to be compatible with Matlab's current fsolve.
	(make_unimplemented_options, override_options): New functions.
	(Ffsolve): Handle optimset options.  Update doc string.

	* ov-usr-fcn.cc (octave_user_function::do_multi_index_op,
	octave_user_script::do_multi_index_op):
	Call octave_call_stack::backtrace_error_message.
	* toplev.h (octave_call_stack::do_pop): Don't call
	backtrace_error_message here.
	* toplev.cc, toplev.h (octave_call_stack::do_backtrace_error_message):
	Rename from octave_call_stack::backtrace_error_message.
	* toplev.h (octave_call_stack::backtrace_error_message):
	New public static function. 

2008-08-05  John W. Eaton  <jwe@octave.org>

	* gripes.cc (gripe_truncated_conversion,
	gripe_logical_conversion): Don't declare as extern here.
	* utils.cc (octave_format, octave_vformat): Likewise.
	* pr-output.cc (octave_print_internal (std::ostream&, const Cell&,
	bool, int, bool)): Likewise.

	* toplev.cc (octave_call_stack::backtrace_error_message): New
	function.
	* toplev.h: Provide decl.
	(octave_call_stack::do_pop): Call backtrace_error_message if
	error_state is nonzero.

	* pt-unop.cc (tree_prefix_expression::eval_error, 
	tree_postfix_expression::eval_error): Delete.  Eliminate all uses.
	* pt-unop.h: Delete decls.

	* pt-select.cc (tree_switch_case::eval_error,
	tree_switch_command::eval_error): Delete.  Eliminate all uses.
	(tree_if_command::eval): Omit "evaluating if command" message.
	* pt-select.h: Delete decls.

	* pt-loop.cc (tree_while_command::eval_error,
	tree_complex_for_command::eval_error,
	tree_simple_for_command::eval_error,
	tree_do_until_command::eval_error): Delete.  Eliminate all uses.
	* pt-loop.h: Delete decls.

	* pt-exp.cc (tree_expression::is_logically_true): Omit "evaluating
	conditional expression" error message.

	* pt-decl.cc (tree_global_command::eval): Omit "evaluating global
	command" error message.
	(tree_static_command::eval): Omit "evaluating static command"
	error message.

	* pt-colon.cc (tree_colon_expression::eval_error): Omit
	"evaluating colon expression" error message.
	* pt-colon.h (tree_colon_expression::eval_error): Eliminate
	default argument value.

	* pt-idx.cc (tree_index_expression::eval_error): Delete.
	Eliminate all uses.
	* pt-idx.h: Delete decl.

	* pt-binop.cc (tree_binary_expression::eval_error): Delete.
	* pt-binop.cc, pt-cbinop.cc: Eliminate all uses of eval_error.
	* pt-binop.h: Delete decl.

	* pt-assign.cc (tree_simple_assignment::eval_error,
	tree_multi_assignment::eval_error): Delete.  Eliminate all uses.
	* pt-assign.h: Delete decls.

	* ov-usr-fcn.cc (octave_user_script::traceback_error,
	octave_user_function::traceback_error): Delete.  Eliminate all uses.
	* ov-usr-fcn.h: Delete decls.

	* ov-builtin.cc (octave_builtin::do_multi_index_op): Call
	gripe_library_exectuion_error on octave_execution_exception.
	* ov-mex-fcn.cc (octave_mex_function::do_multi_index_op): Likewise.
	* ov.cc (do_binary_op, do_cat_op, do_unary_op,
	do_non_const_unary_op): Likewise.
	* pt-stmt.cc (tree_statement::eval): Likewise.

	* gripes.cc (gripe_library_execution_error): New function.
	* gripes.h: Provide decl.

	* dirfns.cc (Fpathsep): Fix usage of dir_path::path_sep_str.
	* defaults.cc (set_exec_path, set_image_path): Likewise.
	* load-path.h (load_path::set_command_line_path): Likewise.
	* load-path.cc (maybe_add_path_elts, load_path::do_initialize,
	load_path::do_path, genpath, Fpath): Likewise.
	(split_path): Fix usage of dir_path::path_sep_char.

2008-08-04  John W. Eaton  <jwe@octave.org>

	* symtab.cc (symbol_table::fcn_info::fcn_info_rep::find_autoload):
	Fix usage of file_ops::dir_sep_chars.
	* variables.cc (looks_like_struct):
	Likewise.
	* ov-fcn-handle.cc (octave_fcn_handle::set_fcn): Likewise.
	* octave.cc (execute_command_line_file): Likewise.
	* ls-mat5.cc (read_mat5_binary_element): Likewise.
	* load-save.cc (find_file_to_load): Likewise.
	* load-path.cc (load_path::do_find_file): Likewise.
	* graphics.cc (drawnow): Likewise.
	* parse.y (frob_function): Likewise.

	* octave.cc (initialize_pathsearch): Fix usage of
	file_ops::dir_sep_str.
	* help.cc (Flookfor): Likewise.
	* dirfns.cc (Ffilesep): Likewise.
	(Fautoload): Likewise.

	* defaults.cc (subst_octave_home): Fix usage of
	file_ops::dir_sep_char.
	(Fmfilename): Likewise.

2008-07-31  John W. Eaton  <jwe@octave.org>

	* parse.y (assign_lhs): Call force_local_variable on all elements
	of lexer_flags.pending_local_variables here, then clear the set.
	(matrix): Clear lexer_flags.pending_local_variable here.
	* lex.l (lexical_feedback::init): Clear it.
	(force_local_variable): No longer static.
	(is_variable): Also return true for names in the
	lexer_flags.pending_local_variables.
	(handle_identifier): If we are parsing a matrix list, mark
	identifiers as pending local variables rather than forcing them to
	be local variables immediately.
	* lex.h (lexical_feedback::pending_local_variables): New data member.
	(force_local_variable): Provide decl.

2008-07-30  John W. Eaton  <jwe@octave.org>

	* ov-intx.h, ov.cc: Style fixes.

	* gripes.cc (gripe_truncated_conversion): Use the warning ID
	Octave:int-convert-overflow.

2008-07-30  Jaroslav Hajek  <highegg@gmail.com>

	* gripes.cc (gripe_truncated_conversion): New function.
	* gripes.h (gripe_truncated_conversion): Declare it.
	* ops.h (DEFCONVFNX2): New macro.
	(DEFCONVFN, DEFCONVFN2): Call DEFCONVFNX2 instead of DEFCONVFNX.
	* ov-intx.h (OCTAVE_VALUE_INT_MATRIX_T::convert_gripe): New member
	function.
	(OCTAVE_VALUE_INT_MATRIX_T::int*_array_value): Call convert_gripe.
	(OCTAVE_VALUE_INT_SCALAR_T::convert_gripe): New member function.
	(OCTAVE_VALUE_INT_SCALAR_T::int*_scalar_value): Call convert_gripe.
	(OCTAVE_VALUE_INT_SCALAR_T::int*_array_value): Call int*_scalar_value.
	* ov.cc (convert_to_int_array): New function.
	(int_vector_value): Call convert_to_int_array.

2008-07-30  John W. Eaton  <jwe@octave.org>

	* defun-dld.h (DEFUNX_DLD): Eliminate fsname arg.
	* defun-int.h (DEFINE_FUN_INSTALLER_FUN2,
	DEFINE_FUN_INSTALLER_FUN3, DEFINE_FUNX_INSTALLER_FUN2): Delete.
	(DEFINE_FUNX_INSTALLER_FUN): Rename from DEFINE_FUNX_INSTALLER_FUN3.
	Don't emit fsname function.  Don't append cxx_abi to gname.
	(DEFINE_FUN_INSTALLER_FUN): Define in terms of
	DEFINE_FUNX_INSTALLER_FUN, not DEFINE_FUN_INSTALLER_FUN2.
	* dynamic-ld.cc (octave_dynamic_loader::name_mangler,
	octave_dynamic_loader::name_uscore_mangler): New functions.
	(octave_dynamic_loader::mangle_name,
	octave_dynamic_loader::xmangle_name): Delete.
	(octave_dynamic_loader::do_load_oct): Search for function with
	name_mangler then name_uscore_mangler.

	* oct-stream.cc (do_read): New arg, do_NA_conv.
	Perform NA translation if do_NA_conv is true.
	(DO_READ_VAL_TEMPLATE, read_fptr): Include the new arg for do_read
	in decl.
	(octave_stream::read): Pass do_NA_conv to do_read.

2008-07-29  David Bateman  <dbateman@free.fr>

	* data.cc (FNA): Add tests for conversion of single to double NA
	values.

	* ov-flt-re-mat.cc (Fsingle): Documentation fix.

2008-07-28  David Bateman  <dbateman@free.fr>

	* error.cc (last_error_id, last_error_message, last_warning_id,
	last_warning_message): New functions to return the error/warning
	message and id without exposing the internal values.
	* error.h  (last_error_id, last_error_message, last_warning_id,
	last_warning_message): Declare them.
	* DLD-FUNCTIONS/cellfun.cc (Fcellfun): Use them to pass the error
	to the ErrorHandler function.
	
2008-07-28  John W. Eaton  <jwe@octave.org>

	* error.cc (Flasterror, Flasterr): Unwind-protect error_state.

	* DLD-FUNCTIONS/__magick_read__.cc (F__magick_write__, write_image):
	New functions.

2008-07-25  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/det.cc (Fdet): Return calculated determinant for
	numerically singular matrices, not 0.

	* symtab.cc (get_dispatch_type): New function.
	(symbol_table::fcn_info::fcn_info_rep::find): Use it.

	* ov-class.cc (set_class_relationship): Delete static function.
	(Fsuperiorto, Finferiorto): Warn about precedence conflicts.
	Call symbol_table::set_class_relationship instead of local static
	function.
	* symtab.h (symbol_table::class_precedence_table): New data member.
	(symbol_table::set_class_relationship, symbol_table::is_superiorto):
	New static functions.

2008-07-24  John W. Eaton  <jwe@octave.org>

	* load-path.h (load_path::dir_info::class_info): New struct.
	(load_path::dir_info::method_file_map_type): Now a map from
	class name to a to a class_info object.  Change all uses.
	* load-path.cc (load_path::dir_info::get_method_file_map):
	Also look for private functions in the class directory.
	(load_path::add_to_method_map): Also add private functions from
	class directories to private_fcn_map.

	* dirfns.cc (Fmkdir): If directory already exists, return status =
	true, but also set error message.

2008-07-23  John W. Eaton  <jwe@octave.org>

	* ov-usr_fcn.cc (octave_user_function::do_multi_index_op):
	Don't unwind_protect and set curr_parent_function here.
	* toplev.cc (curr_parent_function): Delete definition.
	* toplev.h: (curr_parent_function): Delete declaration.

	* ov-usr-fcn.h (octave_user_function::parent_scope): New data member.
	(octave_user_function::parent_fcn_scope,
	octave_user_function::stash_parent_fcn_scope): New functions.
	* ov-usr_fcn.cc (octave_user_function::octave_user_function):
	Initialize parent_scope.
	* symtab.cc (symbol_table::fcn_info::fcn_info_rep::find):
	Check parent of current function by looking at call stack, not
	global curr_parent_function variable.
	* parse.y (frob_function): If parsing nested function, stash
	current parent function scope.
	* ov-fcn.h (octave_function::parent_fcn_scope): New virtual function.

2008-07-22  Michael Goffioul  <michael.goffioul@gmail.com>

	* graphics.cc (F__go_execute_callback__): New function.

	* DLD-FUNCTIONS/fltk_backend.cc: Undefine min/max after FLTK headers
	inclusion.

	* graphics.h.in (class base_graphics_event, class graphics_event): New
	classes (replace gh_manager::event_data).
	(class gh_manager::event_data): Remove.
	(gh_manager::post_function, gh_manager::do_post_function): Use
	graphics_event::event_fcn type.
	(gh_manager::event_queue, gh_manager::do_post_event): Use
	graphics_event type.
	* graphics.cc (class callback_event, class function_event, class
	set_event): Renamed from xxx_data classes.
	(graphics_event::create_callback_event,
	graphics_event::create_function_event,
	graphics_event::create_set_event): Renamed from gh_manager::xxx
	equivalent methods, removed reference count increment.
	(gh_manager::do_post_event): Likewise.
	(gh_manager::do_post_event): Use graphics_event type.
	(gh_manager::do_post_function): Use graphics_event::event_fcn type.

2008-07-22  John W. Eaton  <jwe@octave.org>

	* version.h (OCTAVE_VERSION): Now 3.1.51+.

2008-07-21  John W. Eaton  <jwe@octave.org>

	* version.h (OCTAVE_VERSION): Now 3.1.51.
	(OCTAVE_RELEASE_DATE): Now 2008-07-21.

	* ov-class.cc (octave_class::subsasgn): Only do internal magic if
	rhs is not an octave_class object.

	* OPERATORS/op-struct.cc: Define concatenation operators for
	struct/matrix concatenation (valid if matrix is empty).
	* OPERATORS/op-cell.cc (install_cell_ops): Likewise, for cells.

	* DLD-FUNCTIONS/fltk_backend.cc: Don't include oct.h.
	Make compilation of entire file conditional on HAVE_FLTK.

	* gl-render.cc: Make compilation of entire file conditional on
	HAVE_OPENGL.

	* Makefile.in (GL_RENDER_SRC, FLTK_BACKEND_SRC): Delete variables.
	(DLD_XSRC): Always include fltk_backend.cc in the list.
	(DIST_SRC): Always include gl-render.cc in the list.

2008-07-19  John W. Eaton  <jwe@octave.org>

	* version.h (OCTAVE_VERSION): Now 3.1.50+.
	(OCTAVE_API_VERSION): Now api-v33+.

2008-07-18  John W. Eaton  <jwe@octave.org>

	* version.h (OCTAVE_VERSION): Now 3.1.50.
	(OCTAVE_API_VERSION): Now api-v33.
	(OCTAVE_RELEASE_DATE): Now 2008-07-18.
	(OCTAVE_COPYRIGHT): Update date to 2008.

2008-07-18  John W. Eaton  <jwe@octave.org>

	* Makefile.in (__magick_read__.df): New rule.
	Append MAGICK_INCFLAGS to CPPFLAGS.

	* Makefile.in (EXTRAS): Add graphics-props.cc to the list.
	List graphics-props.cc as a target generated by genprops.awk.

2008-07-18 Carlo de Falco  <carlo.defalco@gmail.com>

	* gl-render.h: Conditionally include GL/gl.h or OpenGL/gl.h
	and GL/glu.h or OpenGL/glu.h

2008-07-17  John W. Eaton  <jwe@octave.org>

	* symtab.cc (out_of_date_check_internal): New arg, dispatch_type.
	Use it to handle class methods.

	* toplev.h (octave_call_stack::do_caller): New function.
	(octave_call_stack::caller): Use it.

2008-07-17  Jaroslav Hajek  <highegg@gmail.com>

	* Makefile.in (convhulln.oct, __delaunayn__.oct, __voronoi__.oct, 
	regexp.oct, urlwrite.oct, __glpk__.oct, fltk_backend.oct):
	Use OCT_LINK_DEPS instead of DL_LDFLAGS for target-specific
	dependencies.

2008-07-17  David Bateman  <dbateman@free.fr>

	* DLD-FUNCTION/__magick_read__.cc (F__magick_read__): Allow build
	without GraphicsMagick++ installed.

2008-07-16  John W. Eaton  <jwe@octave.org>

	* graphics.h.in (gh_manager::autolock): Delete copy constructor
	and assignment definitions.

2008-07-16  Michael Goffioul  <michael.goffioul@gmail.com>

	* graphics.h.in (callback_property::execute): Remove static version.
	(base_properties::is_hittest, base_properties::is_interruptible,
	base_properties::is_selected, base_properties::is_selectionhighlight):
	New convenience property accessors.
	(base_graphics_object::get_handle, graphics_object::get_handle): Idem.
	(gh_manager::graphics_lock): New global mutex
	(gh_manager::lock, gh_manager::unlock, gh_manager::do_lock,
	gh_manager::do_unlock): Add accessors for it.
	(gh_manager::autolock): New class for easy locking of the graphics
	system.
	(gh_manager::event_data): New class for event management.
	(gh_manager::event_queue): New object to hold pending events.
	(gh_manager::callback_objects): New stack of callback objects.
	(gh_manager::execute_callback, gh_manager:post_callback,
	gh_manager::post_function, gh_manager::post_set,
	gh_manager::process_events, gh_manager::flush_events,
	gh_manager::restore_gcbo): New static methods for event management.
	(gh_manager::do_execute_callback, gh_manager::do_post_callback,
	gh_manager::do_post_function, gh_manager::do_post_set,
	gh_manager::do_process_events, gh_manager::do_post_event,
	gh_manager::do_restore_gcbo): New non-static versions.
	* graphics.cc (xreset_gcbo): Remove.
	(execute_callback): Likewise.
	(base_property::run_listeners, callback_property::execute): Use
	gh_manager::execute_callback.
	(class callback_event_data, class function_event_data, class
	set_event_data): New classes to implement various types of events.
	(gh_manager::event_data::create_callback_event,
	gh_manager::event_data::create_function_event,
	gh_manager::event_data::create_set_event): Implement event factory
	methods.
	(gh_manager::do_restore_gcbo, gh_manager::do_execute_callback,
	gh_manager::do_post_event, gh_manager::do_post_callback,
	gh_manager::do_post_function, gh_manager::do_post_set,
	gh_manager::do_process_events): New methods for event management.
	(Fishandle, Fset, Fget, F__get__, F__go_figure__, F__go_delete__,
	F__go_axes_init__, F__go_handles__, F__go_figure_handles__,
	Favailable_backends, Fdrawnow, Faddlistener, Faddproperty,
	get_property_from_handle, set_property_in_handle): Lock graphics
	system.
	(GO_BODY): Likewise.
	(Fdrawnow): Support single "expose" argument.

2008-07-15  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/__convn__.cc (convn): Cast second arg to
	octave_idx_type in call to std::max.

2008-07-14  John W. Eaton  <jwe@octave.org>

	* Makefile.in (convhulln.oct, __delaunayn__.oct, __voronoi__.oct, 
	regexp.oct, urlwrite.oct, __glpk__.oct, fltk_backend.oct):
	Simplify with target-specific settings for DL_LDFLAGS.

	* DLD-FUNCTIONS/__magick_read__.cc: Delete "using namespace std"
	and "using namespace Magick" directives.  Style fixes.

	* oct-conf.h.in (OCTAVE_CONF_MAGICK_INCFLAGS,
	OCTAVE_CONF_MAGICK_LIBS): Substitute here.
	* toplev.cc (octave_config_info): Include MAGICK_INCFLAGS and
	MAGICK_LIBS in the struct.
	* Makefile.in (DLD_XSRC): Add __magick_read__.cc to the list.
	(OCTAVE_LIBS): Add $(MAGICK_LIBS) to the list for static linking case.
	(__magick_read__.d, __magick_read__.o, pic/__magick_read__.o):
	Append MAGICK_INCFLAGS to CPPFLAGS.
	(__magick_read__.oct): Append MAGICK_LIBS to LDFLAGS.

2008-07-14  Thomas L. Scofield  <scofield@calvin.edu>

	* DLD-FUNCTIONS/__magick_read__.cc: New file from Octave Forge.

2008-07-11  John W. Eaton  <jwe@octave.org>

	* syscalls.cc (const_value): Delete arg NM.  Change all uses.

	* DLD-FUNCTIONS/fft.cc (do_fft): Avoid GCC warning.
	* DLD-FUNCTIONS/fft2.cc (do_fft2): Likewise.
	* DLD-FUNCTIONS/fftn.cc (do_fftn): Likewise.
	* DLD-FUNCTIONS/qr.cc (Fqrshift): Likewise.
	* DLD-FUNCTIONS/lookup.cc (Flookup): Likewise.
	* gl-render.cc (opengl_renderer::draw): Likewise.
	* graphics.cc (axes::update_axis_limits,
	hggroup::update_axis_limits, Favailable_backends):
	Likewise.
	* pt-cmd.cc (tree_no_op_command::dup, tree_function_def::dup):
	Likewise.
	* pt-const.cc (tree_constant::dup): Likewise.
	* pt-id.cc (tree_identifier::dup): Likewise.
	* pt-jump.cc (tree_break_command::dup, tree_continue_command::dup,
	tree_return_command::dup): Likewise.

	* DLD-FUNCTIONS/colamd.cc: Style fixes.
	(tree_postorder): Rename from TreePostorder.
	(tree_postorder, coletree): Avoid GCC warnings.

	* DLD-FUNCTIONS/chol.cc (Fchol): Avoid GCC warnings.
	(Fcholdelete, Fcholshift): Delete unused arg nargout.

	* toplev.cc, toplev.h (octave_call_stack::caller_user_function,
	octave_call_stack::caller_user_script,
	octave_call_stack::do_caller_user_function,
	octave_call_stack::do_caller_user_script):
	Delete functions and decls. 

	* defun.cc (print_usage): Call octave_call_stack::caller_user_code,
	not octave_call_stack::caller_user_function.
	* debug.cc (get_user_code): Likewise.

	* toplev.h (octave_call_stack::difference_type): Delete typedef.
	* toplev.cc, toplev.h (octave_call_stack::caller_user_code_line,
	octave_call_stack::do_caller_user_code_line, 
	octave_call_stack::caller_user_code_column,
	octave_call_stack::do_caller_user_code_column,
	octave_call_stack::caller_script,
	octave_call_stack::do_caller_script,
	octave_call_stack::caller_user_function,
	octave_call_stack::do_caller_user_function,
	octave_call_stack::caller_user_code,
	octave_call_stack::do_caller_user_code):
	Delete unused difference_type arg.

	* ov-float.h, ov-flt-re-mat.cc, ov-range.h, ov-re-mat.h,
	ov-re-sparse.cc, ov-scalar.h: Check for NaN in bool_value and
	bool_array_value member functions to bool.

	* ops.h (DEFSCALARBOOLOP_OP): New macro.
	* OPERATORS/op-s-s.cc, OPERATORS/op-fs-fs.cc: Use it to define
	el_and and el_or ops.

2008-07-10  David Bateman  <dbateman@free.fr>

	* DLD-FUNCTIONS/lookup.cc (assign): Delete.
	(Flookup): Use new NDArray constructors rather than assign.
	* data.cc (Fsort): Use new NDArray constructors rather than 
	directly assigning.
	* besselj.cc (int_array2_to_matrix, int_arrayn_to_array,
	int_array2_to_float_matrix, int_arrayn_to_float_array): Delete.
	(do-bessel): Use new NDArray constructors.
	* max.cc (MINMAX_SPARSE_BODY, MINMAX_INT_BODY, MINMAX_SINGLE_BODY,
	MINMAX_DOUBLE_BODY): Use new NDArray constructors. 
	* ov-bool.h (array_value, float_array_value): Explictly cast
	boolean scalar to double before the assignment to avoid ambiguity.

2008-07-10  David Bateman  <dbateman@free.fr>

	* ls-mat4.cc (read_mat_binary_data, save_mat_binary_data): Add
	loading and saving of sparse matrices.
	
2008-07-10  Michael Goffioul  <michael.goffioul@gmail.com>

	* Makefile.in: Add OPENGL_LIBS to liboctinterp link command. Add
	GRAPHICS_CFLAGS to fltk_backend.cc compilation command.
	* data.cc: Define hypotf to _hypotf when not present.
	* gl-render.h: Replace OCTGRAPHICS_API with OCTINTERP_API.
	* gl-render.cc: Get rid of Array2<vertex_data>.
	* OPERATORS/op-int.h: Use powf instead of pow when invoked
	with float arguments.

2008-07-09  John W. Eaton  <jwe@octave.org>

	* input.cc (get_debug_input): Don't delete global_command here.
	* toplev.cc (main_loop): Undo previous change.

	* toplev.h (octave_call_stack::instance_ok): Push top scope when
	creating instance.

	* DLD-FUNCTIONS/inv.cc (Finv): Avoid GCC warning.

	* DLD-FUNCTIONS/expm.cc (Fexpm): Avoid GCC warning.

	* ov-fcn-inline.cc (octave_fcn_inline::load_ascii): Avoid GCC warning.

2008-07-09  Michael Goffioul  <michael.goffioul@gmail.com>

	* gl-render.cc (opengl_renderer::draw(figure)): Enable line smoothing
	according to __enhanced__ property.
	(opengl_renderer::draw_marker): Only draw markers with valid call
	lists ID.
	(opengl_renderer::make_marker_list): Do not produce filled marker call
	list for non-fillable markers.
	(opengl_renderer::draw(axes)): Do not antialias axes system.

	* gl-render.cc (opengl_renderer::set_polygon_offset): Also enable
	polygon offseting in GL_LINE mode.
	(opengl_renderer::draw_marker): Offset markers foward instead of
	backward (to draw them above lines).

	* Makefile.in: new substituted variables GL_RENDER_SRC and
	FLTK_BACKEND_SRC.

	* gl-render.cc (vertex_data::vertex_data_rep::vertex_data(void),
	vertex_data::vertex_data(void), vertex_data::vertex_data(...)):
	initialize reference counting correctly.

2008-07-09  John W. Eaton  <jwe@octave.org>

	* toplev.cc (main_loop): Unwind-protect global_command.
	* input.cc (get_debug_input): Likewise.

2008-07-09  David Bateman  <dbateman@free.fr>

	* pt-id.cc (octave_value_list tree_identifier::rvalue (int), 
	octave_lvalue tree_identifier::lvalue (void)): Remove 
	MAYBE_DO_BREAKPOINT.

2008-07-08  John W. Eaton  <jwe@octave.org>

	* graphics.h.in (axes::properties::keypos): Declare as
	any_property instead of double_property.

2008-06-28  John W. Eaton  <jwe@octave.org>

	* debug.cc (push_dummy_call_stack_elt): Delete.
	(Fdbstack): Don't push dummy stack elt.  Use nskip instead of
	nframes, and curr_frame instead of idx.  Pass curr_frame to
	octave_call_stack::backtrace.

	* ls-mat5.cc (read_mat5_binary_element): Explicitly pass 0 for
	current_context in call to symbol_table::varref.
	* ov-fcn-handle.cc (octave_fcn_handle::load_ascii): Likewise.
	Fix test.

	* mex.cc (mexGetVariable, mexPutVariable): Use octave_call_stack
	methods to manipulate scope and context.
	* parse.y (source_file, Fassignin, Fevalin): Likewise.

	* ov-usr-fcn.cc (octave_user_function::do_multi_index_op):
	Call octave_call_stack::push after call to symbol_table::push_scope.
	Call unwind_protect::add with octave_call_stack::unwind_pop
	immediately after call to octave_call_stack::push.

	* symtab.cc (symbol_table::xcurrent_caller_scope): Delete definition.
	* symtab.h (symbol_table::xcurrent_caller_scope): Delete decl.
	Delete all uses.
	(xcurrent_context_this_table): Delete.  Delete all uses.
	(symbol_table::current_caller_scope): Delete.

	* toplev.cc (octave_call_stack::do_num_user_code_frames):
	New function.
	(octave_call_stack::do_backtrace): New arg, curr_user_frame.	
	Rename nframes arg to nskip.  Correctly handle curr_frame not at
	end of stack.
	(octave_call_stack::do_goto_frame_relative): Rename n arg to nskip.
	Correctly handle curr_frame not at end of stack.
	(octave_call_stack::do_goto_caller_frame): New function.
	(octave_call_stack::do_goto_base_frame): New function.
	* toplev.h: Provide decls for new functions.
	(octave_call_stack::call_stack_elt::prev): New data member.
	Initialize it in constructor.
	(octave_call_stack::const_reverse_iterator,
	octave_call_stack::reverse_iterator): New typedefs.
	(octave_call_stack::symbol_table::scope_id current_scope,
	octave_call_stack::symbol_table::context_id current_context,
	octave_call_stack::num_user_code_frames,
	octave_call_stack::goto_caller_frame,
	octave_call_stack::goto_base_frame,
	octave_call_stack::do_num_user_code_frames,
	octave_call_stack::do_current_scope,
	octave_call_stack::do_current_context): New functions.

	(octave_call_stack::push): Default value for context arg is
	symbol_table::current_context, not 0.
	(octave_call_stack::do_push): Save previous frame.  Always push
	new frame on back of stack.  Call symbol_table::set_scope_and_context.
	(octave_call_stack::do_pop): Restore previous frame.  Always pop
	frame from back of stack.  Call symbol_table::set_scope_and_context.

2008-06-26  John W. Eaton  <jwe@octave.org>

	* token.h: Omit "class symbol_table::symbol_record" decl;

	* lex.l (grab_block_comment): Use parens around || expression
	within && expression.

2008-06-23  John W. Eaton  <jwe@octave.org>

	* debug.cc (Fdbstack): Avoid shadowed variable warning.

2008-06-23  Jaroslav Hajek  <highegg@gmail.com>

	* genprops.awk (emit_source): Avoid gensub for portability.
	(BEGIN): Ditto.

2008-06-20  Jaroslav Hajek  <highegg@gmail.com>

	* DLD-FUNCTIONS/regexp.cc (octregexp_list): Make "once" an output
	argument.
	(octregexp): Do not use cell arrays when "once" is requested.

	* ov.cc (make_vector_dims): New function.
	(vector_value, complex_vector_value, float_vector_value,
	float_complex_vector_value): Query N-d array values and simplify,
	avoid copying.
	(column_vector_value, complex_column_vector_value,
	float_column_vector_value, float_complex_column_vector_value,
	row_vector_value, complex_row_vector_value,
	float_row_vector_value, float_complex_row_vector_value): 
	Simplify to trivial wrappers.
	(int_vector_value): Avoid conversions if integer type, query N-d array
	value, simplify.

2008-06-17  John W. Eaton  <jwe@octave.org>

	* toplev.h, toplev.cc (class octave_call_stack):
	Push elements on and pop from back of queue.
	(octave_call_stack::do_push, octave_call_stack::do_pop):
	Always adjust curr_frame.
	(octave_call_stack::size, octave_call_stack::do_size): New functions.
	* pt-stmt.cc (tree_statement::eval):
	Also call octave_call_stack::set_statement when debugging.
	* debug.cc (push_dummy_call_stack_elt): New function.
	(Fdbstack): Omit dbstack from call stack by popping element rather
	than adjusting frame number.  Correctly handle arg.

2008-06-12  David Bateman  <dbateman@free.fr>

	* DLD-FUNCTIONS/qr.cc (Fqrupdate, Fqrinsert, Fqrshift, Fqrdelete):
	Allow single precision arguments, add tests for single precision.

2008-06-11  John W. Eaton  <jwe@octave.org>

	* ov-base.cc (octave_base_value::streamoff_value,
	octave_base_value::streamoff_array_value): Delete.
	* ov-base.h: Delete decls.
	(octave_base_value::is_streamoff): Delete.
	* ov-float.cc (octave_float_scalar::streamoff_value,
	octave_float_scalar::streamoff_array_value): Delete.
	* ov-float.h: Delete decls.
	* ov-flt-re-mat.cc (octave_float_matrix::streamoff_array_value): Delete.
	* ov-flt-re-mat.h: Delete decl.
	* ov-re-mat.cc (octave_matrix::streamoff_array_value): Delete.
	* ov-re-mat.h: Delete decl.
	* ov-re-sparse.cc (octave_sparse_matrix::streamoff_array_value): Delete.
	* ov-re-sparse.h: Delete decl.
	* ov-scalar.cc (octave_scalar::streamoff_value,
	octave_scalar::streamoff_array_value): Delete.
	* ov-scalar.h: Delete decls.
	* ov.cc (octave_value::octave_value (const streamoff_array&),
	octave_value::octave_value (const ArrayN<std::streamoff>&),
	octave_value::streamoff_value (void) const,
	octave_value::streamoff_array_value (void) const): Delete.
	* ov.h: Delete decls.
	(octave_value::is_streamoff): Delete.
	* ov.cc (install_types): Don't register streamoff type.
	* ov-streamoff.h, ov-streamoff.cc, OPERATORS/op-streamoff.cc: Delete.
	* Makefile.in: Delete them from the lists.

	* error.cc (verror): Restore newline in msg_string.  Stripping
	"error: " prefix when buffering error messages is no longer neeed.

2008-06-25  David Bateman  <dbateman@free.fr>

	* pr-output.cc (Frats): Print usage if nargin == 0.
	
2008-06-10  John W. Eaton  <jwe@octave.org>

	* mexproto.h (mxCreateLogicalScalar): Declar arg as mxLogical, not int.

2008-06-09  John W. Eaton  <jwe@octave.org>

	* graphics.cc (execute_callback): Pass CB by const reference and
	explicitly make a copy.
	(xget_ancestor): Pass GO by const reference and explicitly make a copy.

	* error.cc (verror): Omit "name: " and "\a" from Vlast_error_msg.
	Save line and column information from user code.
	* toplev.cc (octave_call_stack::do_caller_user_code_line,
	octave_call_stack::do_caller_user_code_column): New functions.
	* toplev.h: Provide decls.
	(octave_call_stack::caller_user_code_line,
	octave_call_stack::caller_user_code_column): New functions.
	(octave_call_stack::current_line, octave_call_stack::current_column): 
	Default return value is -1, not 0.

2008-06-06  John W. Eaton  <jwe@octave.org>

	* ov.h (octave_value::erase_subfunctions):
	New function.
	* ov-base.h (octave_base_value::erase_subfunctions):
	New virtual function.
	* ov-usr-fcn.h (octave_user_function::erase_subfunctions):
	New function.
	* symtab.h (symbol_table::fcn_info::::erase_subfunction,
	symbol_table::fcn_info::fcn_info_rep::erase_subfunction,
	symbol_table::erase_subfunctions_in_scope):
	New functions.
	(symbol_table::fcn_info::fcn_info_rep::clear_user_function):
	Erase subfunctions here.

	* variables.cc (Fmlock): Doc fix.

2008-06-05  John W. Eaton  <jwe@octave.org>

	* gl-render.cc (opengl_renderer::draw): Correctly avoid shadow
	warnings from gcc for xmin, xmax, ymin, ymax, zmin, and zmax.

	* graphics/ChangeLog, graphics/Makefile.in, graphics/Makerules.in,
	graphics/fltk_backend/Makefile.in, graphics/opengl/Makefile.in:
	Delete.

	* gl-render.cc, gl-render.h: Move here from src/graphics/opengl.
	* Makefile.in: Add them to the appropriate lists.
	(octave$(EXEEXT)): Also link with $(OPENGL_LIBS)

	* fltk_backend.cc: Move here from src/graphics/fltk_backend.
	* Makefile.in (DLD_XSRC): Add it to the list
	(fltk_backend.oct): Include special rules for linking with
	$(GRAPHICS_LIBS) and $(FT2_LIBS).

	* dynamic-ld.cc (octave_dynamic_loader::mex_mangler,
	octave_dynamic_loader::mex_uscore_mangler,
	octave_dynamic_loader::mex_f77_mangler): New functions.
	(octave_dynamic_loader::do_load_mex): Use them.
	(octave_dynamic_loader::do_remove_oct): Rename from
	octave_dynamic_loader::do_remove.
	(octave_dynamic_loader::do_remove_mex): New function.
	* dynamic-ld.h: Provide/fix decls.

	* graphics.cc (properties::update_normals): Style fixes.
	* graphics.h.in: Style fixes.

	ChangeLog entries for gl-render.h, gl-render.cc, and
	fltk_backend.cc before the move:

	2008-06-05  John W. Eaton  <jwe@octave.org>

	* opengl/gl-render.cc (xmin): New static function.

	* opengl/gl-render.h (opengl_renderer): Style fixes.
	* fltk_backend/fltk_backend.cc: Style fixes.

	2008-04-26  Michael Goffioul  <michael.goffioul@gmail.com>

	* opengl/gl-render.h (opengl_renderer::draw(hggroup)): New method.
	* opengl/gl-render.cc (opengl_renderer::draw(hggroup)): Likewise.
	(opengl_renderer::draw(graphics_object)): Call it.

	2008-03-17  Shai Ayal  <shaiay@users.sourceforge.net>

	* fltk_backend/fltk_backend.cc (plot_window::resize,
	plot_window::draw): make canvas the size of figure.position

	2008-03-09  Shai Ayal  <shaiay@users.sourceforge.net>

	* fltk_backend/fltk_backend.cc (plot_window::handle): add zoom
	stack

	2008-03-06  Shai Ayal  <shaiay@users.sourceforge.net>

	* fltk_backend/fltk_backend.cc (plot_window::handle): Add handling
	of the 'a' and 'g' keys
	(plot_window: toggle_grid): New helper function
	(plot_window): Add new togglegrid button

	2008-03-01  Shai Ayal  <shaiay@users.sourceforge.net>

	* fltk_backend/fltk_backend.cc (OpenGL_fltk::draw_overlay,
	OpenGL_fltk::zoom, OpenGL_fltk::set_zoom_box): Added to support
	zoom box
	(plot_window::handle): Added zoom box code, B-3 now does autoscale
	(plot_window::axis_auto): New utility function to call axis("auto")
	(plot_window::button_press): "A" button now does autoscale

	* opengl/gl-render.cc (make_marker_list): Add the "+x*.^v><"
	markers

	2008-02-27  Michael Goffioul  <michael.goffioul@gmail.com>

	* opengl/gl-render.cc (opengl_renderer::draw(patch),
	opengl_renderer::draw(surface)): Adapt to type change of facealpha and
	edgealpha, using double_radio_property class.

	2008-02-26  Michael Goffioul  <michael.goffioul@gmail.com>

	* opengl/gl-render.h: Make sure windows.h is included before OpenGL
	headers.
	* fltk_backend/Makefile.in (FLTK_EXTRA_CXXFLAGS): Use $(srcdir) for
	includesion of OpenGL renderer headers.
	(Makeconf): Move inclusion of Makeconf later on, to avoid freezing
	MinGW make.

	* fltk_backend/Makefile.in (GRAPHICS_CFLAGS): New substituted
	variable.

	2008-02-25  Shai Ayal  <shaiay@users.sourceforge.net>

	* fltk_backend/fltk_backend.cc (class plot_window): Many changes
	to use figure::properties instead of figure handle to reference
	the figure
	(class figure_manager): ditto
	(__fltk_redraw__): moved most of functionality into the
	figure_manager class
	(plot_window::pixel2pos): Modified to use axes::pixel2coord
	(plot_window::pixel2staus): Modified to use pixel2pos
	(plot_window::handle): Added zoom with mouse 

	2008-02-24  Shai Ayal  <shaiay@users.sourceforge.net>

	* fltk_backend/fltk_backend.cc (OpenGL_fltk::Draw): removed double
	buffer switch
	(OpenGL_fltk::setup_viewport): removed call to glOrtho --
	gl-render takes care of all the transformations

	2008-02-23  Shai Ayal  <shaiay@users.sourceforge.net>

	* fltk_backend/fltk_backend.cc (plot_window::mark_modifed): mark
	the whole window as damaged (otherwise changing figure.postion
	does not have immediate effect)
	(plot_window::draw): New function, checks for window size 
	(__fltk_maxtime__): New DEFUN to allow tweaking of fltk timeout
	(__fltk_redraw__): Use fltk_maxtime as timeout

	2008-02-21  Michael Goffioul  <michael.goffioul@gmail.com>

	* opengl/gl-render.cc (opengl_renderer::patch_tesselator::combine):
	Protect against NULL vertex data.

	* opengl/gl-render.cc (opengl_renderer::draw(patch)): Add marker
	rendering of patch objects.

	2008-02-21  Shai Ayal  <shaiay@users.sourceforge.net>

	* opengl/gl-render.cc: remove OpenGL includes
	* opengl/gl-render.h: add OpenGL includes
	* fltk_backend/fltk_backend.cc: remove OpenGL includes
	(__fltk_redraw__): put figure handle into the figure's
	__plot_stream__ property for later
	(fltk_backend::close_figure): use argument as figure handle to
	close

	2008-02-20  Shai Ayal  <shaiay@users.sourceforge.net>

	* fltk_backend/Makefile.in: initial import

	* fltk_backend/fltk_backend.cc: initial import
	

	2008-02-20  Michael Goffioul  <michael.goffioul@gmail.com>

	* opengl/gl-render.h (opengl_renderer::draw(patch)): New method to
	render patch objects.
	(class opengl_renderer::patch_tesselator): Forward declaration.
	* opengl/gl-render.cc (opengl_texture::create): Use RGB data format
	instead of RGBA.
	(class opengl_tesselator): New classes to abstract GLU tessellation
	process.
	(class opengl_renderer::patch_tesselator): New class to render opaque
	patch objects.
	(class vertex_data): New class to hold vertex data during tessellation
	of patch objects.
	(opengl_renderer::draw(patch)): New method to render patch objects (no
	transparency, no border, no marker yet).
	(opengl_renderer::draw(graphics_object)): Dispatch to it.

	* opengl/gl-render.cc (opengl_renderer::draw(patch)): Use patch color
	data and support face/vertex single color specification.

	* opengl/gl-render.cc (opengl_tesselator::begin_polygon): Set
	tessellation property also for non-filled polygons.
	(opengl_renderer::patch_tesselator::vertex): Protect against empty
	color matrices.
	(opengl_renderer::draw(patch)): Render patch border (no transparency
	yet).

	2008-02-19  Michael Goffioul  <michael.goffioul@gmail.com>

	* opengl/gl-render.cc (opengl_texture::texture_rep::tex_coord,
	opengl_texture::tex_coord): New wrapper around glTexCoord2d.
	(opengl_renderer::draw(surface)): Use it for texturemap
	implementation.
	(opengl_renderer::draw(surface)): Fix indexing bug when creating clip
	matrix.
	(opengl_texture::operator=): Add assignment operator.
	(opengl_texture::create): New static opengl_texture creator.
	(opengl_texture::is_valid): New accessor.

	2008-02-18  Michael Goffioul  <michael.goffioul@gmail.com>

	* opengl/gl-render.cc (class opengl_texture): New class to wrap
	texture operations in OpenGL.

	2008-02-17  Michael Goffioul  <michael.goffioul@gmail.com>

	* opengl/gl-render.cc (opengl_renderer::draw(surface)): Set material
	color when rendering surface facets.

	* opengl/gl-render.cc (opengl_renderer::draw(surface)): Add rendering
	of mesh and markers.

	2008-02-16  Michael Goffioul  <michael.goffioul@gmail.com>

	* opengl/gl-render.cc (opengl_renderer::draw(figure)): Initialize the
	OpenGL context correctly.
	(opengl_renderer::draw(surface)): Add missing glEnd call.

	2008-02-14  Michael Goffioul  <michael.goffioul@gmail.com>

	* opengl/gl-render.h opengl/gl-render.cc: Add rendering
	interface for surface objects (actual implement still
	missing).

	2008-02-14  Michael Goffioul  <michael.goffioul@gmail.com>

	* Makefile.in Makerules.in: Initial import
	* opengl/Makefile.in: Likewise.
	* opengl/gl-render.h opengl/gl-render.cc: Likewise.

2008-06-04  Shai Ayal  <shaiay@users.sourceforge.net>

	* graphics.cc (axes::properties::set_defaults): Preserve position
	if mode is replace.
	(axes::properties::sync_positions): Use default_axes_postion for
	consistency.

	* graphics.h.in (axes::properties::sync_positions,
	axes::properties::update_position,
	axes::properties::update_outerposition): New functions to sync
	outerposition and position.

	* graphics.h.in (axes::properties::update_xlim,
	axes::properties::update_ylim, axes::properties::update_zlim):
	pass is_logscale flag to axes::properties::calc_ticks_and_lims
	* graphics.cc (axes::properties::calc_ticks_and_lims): Added
	support for log scale

	* graphics.h.in (axes::properities::fix_limits) : New method.
	(axes::properties::update_xlim, axes::properties::update_ylim,
	axes::properties::update_zlim): Use fix_limits.

	* graphics.cc (axes::properties::calc_ticks_and_lims): Make sure
	that lo <= hi.

	* graphics.h.in (axes::pixel2coord, axes::coord2pixel): New functions.

	* graphics.cc (convert_position): No longer static.
	* graphics.h.in: Provide decl.

 	* graphics.cc (Favailable_backends): New function.
  	* graphics.h.in (base_graphics_backend::available_backends_list):
	New function.

	* graphics.cc (axes::properties::calc_ticks_and_lims): Refactor.
	Adjust axes limits & tics together so that the result is consistent.
	(axes::properties::calc_tick_sep): Ditto.
	(axes::properties::get_axis_limits): Ditto.  Now a member of
	axes::properties
	(magform): Now file-scope static function instead of
	axes::properties member function.

	* graphics.h.in (axes::properties::update_xlim): call the new
	axes::properties::calc_ticks_and_lims method
	(axes::properties::update_ylim): ditto
	(axes::properties::update_zlim): ditto

2008-06-04  Michael Goffioul <michael.goffioul@gmail.com>

	* ov-fcn-inline.h (class octave_fcn_inline): Mark with OCTINTERP_API.
	* ov-fcn-handle.h (class Octave_fcn_handle): Likewise.
	* ls-oct-binary.h (save_binary_data, read_binary_data): Likewise.

	* genprops.awk (emit_source): Fix if-then-else statement generation
	when the first property is hidden.
	* graphics.h.in (base_properties::adopt): Call mark_modified.
	(class hggroup): New graphics object class.
	* graphics.cc (lookup_object_name, make_graphics_object_from_type,
	property_list::lookup, property_list::set,
	root_figure::init_factory_properties): Support hggroup as possible
	object type.
	(hggroup::update_axis_limits): New graphics object class.
	(make_graphics_object): Look for "parent" property in the arguments
	and use it as actual parent for the created object.
	(F__go_hggroup__): New function.
	(F__go_delete__): Set Vdrawnow_requested to true.

	* graphics.h.in (base_property::clone, string_property::clone,
	radio_property::clone, color_property::clone, double_property::clone,
	double_radio_property::clone, array_property::clone,
	row_vector_property::clone, bool_property::clone,
	handle_property::clone, callback_property::clone, any_property::clone,
	property::clone): New virtual method.
	(property::create): New static method to create dynamic properties.
	(base_properties::has_property): New method to check the existence of
	a property.
	(any_property::any_property, row_vector_property::row_vector_property,
	double_radio_property::double_radio_property): New copy constructors.
	* graphics.cc (lookup_object_name): New static utility function.
	(make_graphics_object_from_type): Likewise.
	(gh_manager::do_make_graphics_handle): Use it.
	(dprop_obj_map): New static map used for chaching purpose in dynamic
	property creation.
	(property::create): New static method to create dynamic properties.
	(base_properties::has_property): New method to check the existence of
	a property.
	(base_properties::override_defaults): Check parent validity.
	(Faddlistener): Documentation fix.
	(Faddproperty): New function to create dynamic properties.

	* genprops.awk (emit_get_array): Don't produce get_xxx_property
	accessors.
	* graphics.h.in (array_property::get_limits): New method to return the
	array limits.
	(base_properties::get_xdata_property,
	base_properties::get_ydata_property,
	base_properties::get_zdata_property,
	base_properties::get_ldata_property,
	base_properties::get_udata_property,
	base_properties::get_xldata_property,
	base_properties::get_xudata_property,
	base_properties::get_cdata_property): Remove methods, replaced with
	limit-based equivalent.
	(base_graphics_object::get_xdata_property,
	base_graphics_object::get_ydata_property,
	base_graphics_object::get_zdata_property,
	base_graphics_object::get_ldata_property,
	base_graphics_object::get_udata_property,
	base_graphics_object::get_xldata_property,
	base_graphics_object::get_xudata_property,
	base_graphics_object::get_cdata_property): Likewise.
	(graphics_object::get_xdata_property,
	graphics_object::get_ydata_property,
	graphics_object::get_zdata_property,
	graphics_object::get_ldata_property,
	graphics_object::get_udata_property,
	graphics_object::get_xldata_property,
	graphics_object::get_xudata_property,
	graphics_object::get_cdata_property): Likewise.
	(base_properties::get_xlim, base_properties::get_ylim,
	base_properties::get_zlim, base_properties::get_clim,
	base_properties::get_alim, base_properties::is_xliminclude,
	base_properties::is_yliminclude, base_properties::is_zliminclude,
	base_properties::is_climinclude, base_properties::is_aliminclude):
	New methods.
	(base_graphics_object::get_xlim, base_graphics_object::get_ylim,
	base_graphics_object::get_zlim, base_graphics_object::get_clim,
	base_graphics_object::get_alim, base_graphics_object::is_xliminclude,
	base_graphics_object::is_yliminclude,
	base_graphics_object::is_zliminclude,
	base_graphics_object::is_climinclude,
	base_graphics_object::is_aliminclude): Likewise.
	(graphics_object::get_xlim, graphics_object::get_ylim,
	graphics_object::get_zlim, graphics_object::get_clim,
	graphics_object::get_alim, graphics_object::is_xliminclude,
	graphics_object::is_yliminclude, graphics_object::is_zliminclude,
	graphics_object::is_climinclude, graphics_object::is_aliminclude):
	Likewise.
	(base_graphics_object::update_axis_limits): Provide default behavior
	by passing the call to the parent object.
	(line::properties::xdata, line::properties::ydata,
	line::properties::zdata, line::properties::ldata,
	line::properties::udata, line::properties::xldata,
	line::properties::xudata): Replace 'l' modifier with 'u'.
	(line::properties::xlim, line::properties::ylim,
	line::properties::zlim, line::properties::xliminclude,
	line::properties::yliminclude, line::properties::zliminclude): New
	hidden properties.
	(line::properties::update_xdata, line::properties::update_ydata,
	line::properties::update_zdata, line::properties::update_xldata,
	line::properties::update_xudata, line::properties::update_ldata,
	line::properties::update_udata): New updaters that update limit
	properties.
	(line::properties::compute_xlim, line::properties::compute_ylim): New
	method to compute actual limits taking into account x/y data and
	upper/lower bounds.
	(text::properties::position): Add 'u' modifier.
	(text::properties::xlim, text::properties::ylim,
	text::properties::zlim, text::properties::xliminclude,
	text::properties::yliminclude, text::properties::zliminclude): New
	hidden properties.
	(text::properties::update_position): New updater that updates limit
	properties.
	(image::properties::xdata, image::properties::ydata,
	image::properties::cdata): Replace 'l' modifier with 'u'.
	(image::properties::cdatamapping): Add 'l' modifier.
	(image::properties::xlim, image::properties::ylim,
	image::properties::clim, image::properties::xliminclude,
	image::properties::yliminclude, image::properties::climinclude): New
	hidden properties.
	(image::properties::get_climinclude,
	image::properties::is_climinclude): Provide specialized inline
	accessors.
	(image::properties::update_xdata, image::properties::update_ydata,
	image::properties::update_cdata): New updaters that update limit
	properties.
	(patch::properties::xdata, patch::properties::ydata,
	patch::properties::zdata, patch::properties::cdata): Replace 'l'
	modifier with 'u'.
	(patch::properties::cdatamapping): Add 'l' modifier.
	(patch::properties::alphadatamapping): New property.
	(patch::properties::xlim, patch::properties::ylim,
	patch::properties::zlim, patch::properties::clim,
	patch::properties::alim, patch::properties::xliminclude,
	patch::properties::yliminclude, patch::properties::zliminclude,
	patch::properties::climinclude, patch::properties::aliminclude: New
	hidden properties.
	(patch::properties::get_climinclude,
	patch::properties::is_climinclude, patch::properties::get_aliminclude,
	patch::properties::is_aliminclude): Provide specialized inline
	accessors.
	(patch::properties::update_xdata, patch::properties::update_ydata,
	patch::properties::update_zdata, patch::properties::update_cdata): New
	updaters that update limit properties.
	(surface::properties::xdata, surface::properties::ydata,
	surface::properties::zdata, surface::properties::cdata,
	surface::properties::alphadata): Replace 'l' modifier with 'u'.
	(surface::properties::cdatamapping,
	surface::properties::alphadatamapping): Add 'l' modifier.
	(surface::properties::xlim, surface::properties::ylim,
	surface::properties::zlim, surface::properties::clim,
	surface::properties::alim, surface::properties::xliminclude,
	surface::properties::yliminclude, surface::properties::zliminclude,
	surface::properties::climinclude, surface::properties::aliminclude: New
	hidden properties.
	(surface::properties::get_climinclude,
	surface::properties::is_climinclude,
	surface::properties::get_aliminclude,
	surface::properties::is_aliminclude): Provide specialized inline
	accessors.
	(surface::properties::update_xdata, surface::properties::update_ydata,
	surface::properties::update_zdata): Update limit properties.
	(surface::properties::update_cdata,
	surface::properties::update_alphadata): New updaters that update limit
	properties.
	* graphics.cc (base_properties::update_axis_limits): Simply call
	update_axis_limits in owning graphics_object.
	(base_graphics_object::update_axis_limits): Provide default behavior
	by passing the call to the parent object.
	(check_limits_vals(double&,double&,double&,const octave_value&)): New
	static method to work with new limit properties in graphics objects.
	(get_children_limits): New static method to compute the raw limits of
	a set of children, based on the new limit properties.
	(axes::update_axis_limits): Simplify and call get_children_limits. Add
	handling of alpha limits.
	(line::properties::compute_xlim, line::properties::compute_ylim): New
	methods to compute line limits taking into account upper/lower bounds.

	* graphics.cc (axes::properties::sync_positions): Update
	transformation data.

	* graphics.cc (Faddlistener): Rename from Fadd_listener.

	* graphics.h.in (axes::properties::pixel2coord): Center Z coordinate
	on x_zlim instead of 0.
	(axes::properties::zoom, axes::properties::unzoom,
	axes::properties::clear_zoom_stack): New methods to handle zoom stack.
	(axes::properties::zoom_stack): New field to hold zoom stack.
	(axes::properties::update_xlim, axes::properites::update_ylim):
	Additional do_clr_zoom argument to control whether the zoom stack will
	be cleared.
	(axes::properties::update_zlim): Clear zoom stack.
	* graphics.cc (axes::properties::zoom, axes::properties::unzoom,
	axes::properties::clear_zoom_stack): New methods to handle zoom stack.

	* genprops.awk (emit_source): Use all properties in factory defaults.

	* graphics.h.in (base_property::base_property): Set internal counter
	to 1 by default.
	(property::property): Adapt constructors to default counter value in
	base_property.

	* graphics.h.in (base_properties::get_property): Make virtual and
	remove const modifier.
	* graphics.cc (base_properties::get_property): Can return built-in
	properties. Generate an error if the property does not exist.
	(base_properties::add_listener): Use get_property.
	(Fadd_listener): Add documentation.
	* genprops.awk (emit_common_declarations, emit_source): Emit code for
	overloaded get_property method.

	* genprops.awk (emit_declarations): Run listeners in the property
	setter methods.
	* graphics.h.in (enum listener_mode): New enum type for listeners.
	(base_property::set): Make non-virtual and handle listeners execution.
	(base_property::listeners): New field holding a map of listeners.
	(base_property::add_listener, base_property::run_listeners,
	base_property::do_set): New methods to handle listeners.
	(property::add_listener, property::run_listeners): Likewise.
	(base_properties::add_listener): Likewise.
	(base_graphics_object::add_property_listener,
	graphics_object::add_property_listener): Likewise.
	(all property classes): Rename set to do_set and make it protected.
	* graphics.cc (base_property::run_listeners): New method to execute
	listeners.
	(color_property::set, double_radio_property::set,
	handle_property::set): Rename to do_set.
	(base_properties::add_listener): New method to handle listener
	additio. Only handle dynamic properties.
	(Fadd_listener): New octave interface to add property listeners to
	graphics object.

	* genprops.awk (emit_get_data): Remove.
	(emit_declarations): Treat row_vector_property as array_property and
	remove data_property references.
	* graphics.h.in (array_property::array_property(array_property)): Add
	copy constructor.
	(class data_property): Remove (replaced by array_property).
	(class base_graphics_backend, class graphics_backend): Replace
	data_property references with array_property.
	(line::properties::xdata, line::properties::ydata,
	line::properties::zdata, line::properties::ldata,
	line::properties::udata, line::properties::xldata,
	line::properties::xudata): Turn into row_vector_property.
	(image::properties::xdata, image::properties::ydata): Likewise.
	(image::properties::init): Add length constraints for xdata and ydata.
	(patch::properties::xdata, patch::properties::ydata,
	patch::properties::zdata, patch::properties::facevertexalphadata):
	Turn into array_property.
	(surface::properties::xdata, surface::properties::ydata,
	surface::properties::zdata): Likewise.
	(patch::properties::init): Add size constraints for xdata, ydata,
	zdata and facevertexalphadata.
	(surface::properties::init): Add size constraints for xdata, ydata and
	zdata.
	* graphics.cc (check_limit_vals): Remove override with data_property.
	(axes::update_axis_limits): Replace data_property references with
	array_property.

	* graphics.h.in (root_figure::get_default): Use factory defaults when
	no explicit default value exists.
	(figure::properties::set___backend__): Reset __plot_stream__ to empty
	value.
	* graphics.cc (gh_manager::gh_manager): Call
	graphics_backend::default_backend to make sure the default backend is
	initialized.

	* graphics.h.in (patch::properties::edgealpha,
	surface::properties::facealpha, surface::properties::edgealpha): Fix
	typos in property names.

	* graphics.h.in (class double_radio_property): New property class for
	holding face/edge alpha values.
	(patch::properties::facealpha, patch::properties::edgealpha,
	surface::properties::facealpha, surface::properties::edgealpha): Use
	double_radio_property class.
	* graphics.cc (double_radio_property::set): Implement new property
	class.
	* genprops.awk (emit_get_double_radio): New function to emit code for
	double_radio_property.
	(emit_declarations): Use it.

	* graphics.h.in (array_property::array_property(void)): New default
	constructor.
	(array_property::xmin, array_property:xmax, array_property::xminp,
	array_property::min_val, array_property::max_val,
	array_property::min_pos): New fields and accessors to hold min/max
	values of the array data.
	(array_property::get_data_limits): New method to compute min/max
	values of the array data.
	(base_properties::get_cdata_property,
	graphics_object::get_cdata_property): Return array_property.
	(image::properties::cdata, surface::properties::cdata,
	patch::properties::cdata, patch::properties::facevertexcdata): Turn
	into array_property type.
	(image::properties::init, surface::properties::init,
	patch::properties::init): Add constraint initialization for cdata and
	facevertexcdata (patch only).
	* genprops.awk (emit_get_array): New function to emit accessor for
	array_property.
	(emit_declarations): Use it.
	* graphics.cc (get_array_limits): New template function to compute
	min/max values of an Array<T> object.
	(array_property::get_data_limits): New method to compute min/max
	values of the array data, using get_array_limits.
	(check_limit_vals): Overridden function with array_property.
	(axes::update_axis_limits): Turn cdata property into array_property.

	* graphics.h.in (patch::properties::get_color_data): New utility
	function to retrieve actual color data.
	* graphics.cc (patch::properties::get_color_data): Likewise.

	* graphics.h.in (base_scaler::is_linear, lin_scaler::is_linear,
	scaler::is_linear): New method to detect linear scales.
	(graphics_xform::scale(Matrix)): New method to scale 2D/3D coordinates
	directly.
	(patch::properties::facelighting): Set default value to "none".

	* graphics.h.in (axes::axes): Be sure to initialize transformation
	data correctly.

	* graphics.cc (xget_ancestor): Pass graphics_object argument by value
	instead of by reference.

	* graphics.h.in (surface::properties::xdata,
	surface::properties::ydata, surface::properties::zdata,
	surface::properties::normalmode, surface::properties::vertexnormals):
	Add 'u' modifier.
	(surface::properties::update_normals): New method to compute normals
	automatically.
	(surface::properties::update_xdata, surface::properties::update_ydata,
	surface::properties::update_zdata,
	surface::properties::update_normalmode,
	surface::properties::update_vertexnormals): New updaters to update
	normals automatically.
	* graphics.cc (surface::properties::update_normals): Likewise.
	(cross_product): New inlined utility function for cross product
	computation adn accumulation.

	* graphics.h.in (class base_scaler, class lin_scaler, class
	log_scaler, class scaler): Add scale method for NDArray.
	(log_scaler::do_scale): Factorize scaling code.

	* graphics.h.in (figure::properties::update_position): Re-remove.
	(figure::properties::facecolor): Re-add 'texturemap' value.
	(surface::properties::get_color_data): New method to compute actual
	surface color data from cdata.
	* graphics.cc (surface::properties::get_color_data): Likewise.
	(xget_ancestor): New utility function to retrieve an ancestor of a
	given type.
	(convert_cdata): New utility function to convert cdata property into
	actual color data.

	* graphics.h.in (surface::properties::facecolor): Add "texturemap"
	as possible value.
	(class surface::properties): New properties alphadata,
	alphadatmapping, ambientstrength, backfacelighting, diffusestrength,
	edgealpha, edgelighting, erasemode, facelighting, meshstyle,
	normalmode, specularcolorreflectance, specularexponent,
	specularstrength, vertexnormals.
	(surface::properties::init): Add constraints for alphadata,
	vertexnormals and cdata (the latter are commented until cdata
	has changed type).

	* graphics.h.in (base_properties::update_boundingbox): New method
	to handle object resize.
	(figure::properties::set_boundingbox): New method to set figure
	position from backend.
	(figure::properties::update_position): Remove method.
	(figure::properties::position): Remove 'u' modifier and add 'S'
	modifier.
	(axes::properties::update_boundingbox): Overload to recompute
	transformation when axes size changed.
	* graphics.cc (base_properties::update_boundingbox): New method.
	(figure::properties::set_boundingbox,
	figure::properties::set_position): Likewise.

	* genprops.awk: Add 'U' modifier to support extern updaters.
	* graphics.h.in (base_graphics_backend::gripe_invalid): New method
	to simplify error reporting.
	(class base_graphics_backend): Use it.
	(base_graphics_backend::set_figure_position): New backend method.
	(graphics_backend::set_figure_position): Likewise.
	(figure::properties::position): Mark property with updater.
	(figure::properties::update_position): New updater.

	* graphics.h.in (root_figure::properties::callbackobject):
	New root property.
	(root_figure::properties::cbo_stack): New field.
	* graphics.cc (xset_gcbo, xreset_gcbo): New utility functions.
	(execute_callback): Set callbackobject property in root object
	before executing a callback.
	(root_figure::properties::set_callbackobject): Implement accessor.

	* graphics.h.in (root_figure::properties::callbackobject):
	New root property.
	(root_figure::properties::cbo_stack): New field.
	* graphics.cc (xset_gcbo, xreset_gcbo): New utility functions.
	(execute_callback): Set callbackobject property in root object
	before executing a callback.
	(root_figure::properties::set_callbackobject): Implement accessor.

	* graphics.h.in (class root_figure::properties,
	class line::properties, class text::properties,
	class image::properties, class patch::properties,
	class surface::properties): Export classes with
	OCTINTERP_API.

	* graphics.cc (axes::properties::set_defaults): Use correct
	default values for gridlinestyle and minorgridlinestyle.

2008-06-03  Jaroslav Hajek  <highegg@gmail.com>

	* load-save.cc (Fload): Fix "-7" option.

2008-06-02  David Bateman  <dbateman@free.fr>

	* pt.cc (tree:last_break_function): Next pointer to last function
	with a breakpoint.
	* pt.h (tree:last_break_function): ditto.
	* debug.cc (Fdbstep): Use tree::break_function value of zero to
	signal to step into the next function. Set tree::last_break_function
	to indicate position of last breakpoint.
	(Fdbnext):  Set tree::last_break_function to indicate position of
	last breakpoint.
	* pt-bp.h (MAYBE_DO_BREAKPOINT): Check tree::break_function value
	of zero as well. Only check tree::last_line if we are in teh same
	function as the last breakpoint.
	* input.cc (char Vfilemarker): New global variable.
	(Ffilemarker): New function to query and set Vfilemarker.
	* input.h (extern char Vfilemarker): Make Vfilemarker available.
	* util.cc (fcn_file_in_path): If the input argument contains
	Vfilemarker, strip the trailing part of string from this point
	when looking for function file.
	* toplev.cc (octave_call_stack::fo_backtrace): Mark subfunctions
	with the name of the parent function and Vfilemarker.
	* symtab.cc (symbol_table::find_function (const std::string&,
	tree_argument_list *, const string_vector&, octave_value_list*,
	bool&): If function name contains Vfilemarker, check in scope of
	specified parent function.

	* DLD-FUNCTIONS/betainc.cc: Move test code here. Add test code for
	single precision type.
	* DLD-FUNCTIONS/gammainc.cc: Ditto.
	* DLD-FUNCTIONS/gcd.cc (Fgcd): Support single precision. Move test
	code here. Add test code for single precision type.
	* data.cc: Move test code here and add tests for single precision
	type. Add tests for Fislogical.
	(NATIVE_REDUCTION): Support the 'double' argument.
	* mapper.cc: Move test code here and add tests for single precision
	type.
	* ov-float.cc (CD_SCALAR_MAPPER): Replace Complex with
	FloatComplex.
	(erf, erfc, abs, acos, asin, asinh, atan, atanh, ceil, cos, cosh,
	exp, expm1, floor, log10, log1p, sin, sinh, sqrt, tan, tanh): Use
	float version of base functions (ie. ::sinf and not ::sin)
	(lgamma, acosh, log, log2, log10, log1p): Use single precision
	infinity value.
	* ov-flt-complex.cc (erf, erfc, expm1, log1p): Use the float
	version of base functions.
	* ov-flt-cx-mat.cc (DARRY_MAPPER): Replace NDArray with FloatNDArray.
	(erf, erdc, gamma): Use FloatNDArray::dmapper and not
	NDArray::dmapper.
	* ov-flt-cx-mat.h (is_double_type): Delete.
	(is_single_type): New method.
	* ov-flt-re-mat.cc (lgamma,  acosh, log, log2, log10, log1p,
	sqrt): Use single precision infinity value.
	
	* chol.cc (Fcholinv, Fchol2inv, Fcholupdate, Fcholinsert,
	Fcholdelete, Fcholshift): Allow single precision arguments.
	(Fchol): Move test code here. Add test code for single precision.
	(Fcholupdate, Fcholinsert, Fcholdelete, Fcholshift): Add test code
	for single precision.
	* conv2.cc (Fconv2): Add single precision test code.
	* det.cc (Fdet): For single values or empty matrices, return
	single precision arg for single precion input. Move test code
	here. Add single precision test code.
	* fft.cc (do_fft): For empty single precision arguments return a
	single precision value. Add single precision test code. Remove
	fft2 test code.
	* fft2.cc (do_fft2): For empty single precision arguments return a
	single precision value. Add single precision test code. Move fft2
	test code here.
	* fftn.cc (do_fftn): For empty single precision arguments return a
	single precision value.
	* eig.cc (Feig):  Move test code here. Add single precision test
	code.
	* expm.cc (Fexpm): Ditto.
	* find.cc (Ffind): Ditto.
	* hess.cc (Fhess): Ditto.
	* inv.cc (Finc): Ditto.
	* lu.cc (Flu): Ditto.
	* qr.cc (Fqr): Ditto.
	* schur.cc (Fschur): Ditto.
	* svd.cc (Fsvd): Ditto.
	* syl.cc (Fsyl): Ditto.
	
	* op-fcm-fcm.cc, op-fcm-fcs.cc, op-fcm-fm.cc, op-fcm-fs.cc, 
	op-fcs-fcm.cc, op-fcs-fcs.cc, op-fcs-fm.cc, op-fcs-fs.cc, 
	op-fm-fcm.cc, op-fm-fcs.cc, op-fm-fm.cc, op-fm-fs.cc, 
	op-fs-fcm.cc, op-fs-fcs.cc, op-fs-fm.cc, op-fs-fs.cc: Add mixed
	double, single precision concatenation operators.

	* data.cc (Fall, Fany, Fdiag, Fcat, Fismatrix, Fones, Fzeros,
	Finf, FNaN, FNA, Feye, Flinspace, Freshape, Ftranspose,
	Fctranspose, Fsort). Move tests here. Add single precision tests.
	* mappers.cc (Ffinite, Fisinf, Fisna, Fisnan): Ditto.
	* ov-float.cc (octave_float_scalar:resize): single precision
	return value.
	* ov.cc (octave_value::octave_value (const
	FloatComplexDiagMatrix&)): Ditto.
	* data.cc (Fnorm): Add single precision.
	(do_cat): Disable fast return and skipping empty matrices, as they
	play a part in determining the return type.
	* ov.cc (octave_value do_cat_op (const octave_value&, const
	octave_value&, const Array<octave_idx_type>&)): ditto.

2008-06-02  Jaroslav Hajek  <highegg@gmail.com>

	* ov-cell.cc (Fcellstr): For compatibility with Matlab, return {''}
	when given ''.

2008-05-26  Jaroslav Hajek  <highegg@gmail.com>

	* DLD-FUNCTIONS/fsolve.cc (fsolve_user_function,
	fsolve_user_jacobian): Reshape argument to original dims before
	passing. 
	(Ffsolve): Save original dimensions of the starting guess and reshape
	on return. Fix tests.

2008-05-21  David Bateman  <dbateman@free.fr>

	* DLD-FUNCTIONS/quad.cc (quad_float_user_function): New function.
	(Fquad): Handle float type.
	New tests.

2008-05-21  Jaroslav Hajek  <highegg@gmail.com>

	* OPERATORS/op-fcm-fcm.cc (trans_mul, mul_trans, herm_mul, mul_herm):
	New functions.
	(install_fcm_fcm_ops): Install them.
	* OPERATORS/op-fm-fm.cc (trans_mul, mul_trans): New functions.
	(install_fm_fm_ops): Install them.

	* OPERATORS/op-sm-m.cc (trans_mul): New function.
	(install_sm_m_ops): Register it.
	* OPERATORS/op-m-sm.cc (mul_trans): New function.
	(install_m_sm_ops): Register it.
	* OPERATORS/op-scm-cm.cc (trans_mul, herm_mul): New function.
	(install_scm_cm_ops): Register it.
	* OPERATORS/op-cm-scm.cc (mul_trans, mul_herm): New function.
	(install_cm_scm_ops): Register it.

	* dMatrix.cc: Declare DSYRK.
	(xgemm): Call DSYRK if symmetric case detected.
	* CMatrix.cc: Declare ZSYRK, ZHERK.
	(xgemm): Call ZSYRK/ZHERK if symmetric/hermitian case detected.

	* ov.h (octave_value::compound_binary_op): New enum.
	(do_binary_op (octave_value::compound_binary_op, ...), 
	octave_value::binary_op_fcn_name (compound_binary_op),
	octave_value::do_binary_op (compound_binary_op, ...)):
	New declarations.
	(OV_COMP_BINOP_FN): New macro (+ several expansions).
	* ov.cc (octave_value::binary_op_fcn_name (compound_binary_op),
	decompose_binary_op, do_binary_op (compound_binary_op, ...)): 
	New functions.
	* ov-typeinfo.h (octave_value_typeinfo::register_binary_class_op 
	(octave_value::compound_binary_op, ...),
	octave_value_typeinfo::register_binary_op 
	(octave_value::compound_binary_op, ...),
	octave_value_typeinfo::do_register_binary_class_op 
	(octave_value::compound_binary_op, ...),
	octave_value_typeinfo::do_register_binary_op 
	(octave_value::compound_binary_op, ...),
	octave_value_typeinfo::do_lookup_binary_class_op 
	(octave_value::compound_binary_op),
	octave_value_typeinfo::do_lookup_binary_op 
	(octave_value::compound_binary_op, ...)):
	New declarations.
	(octave_value_typeinfo::lookup_binary_class_op
	(octave_value::compound_binary_op),
	(octave_value_typeinfo::lookup_binary_op
	(octave_value::compound_binary_op, ...)):
	New functions.
	(octave_value_typeinfo::compound_binary_class_ops,
	octave_value_typeinfo::compound_binary_ops): 
	New fields.
	* ov-typeinfo.cc (octave_value_typeinfo::register_binary_class_op 
	(octave_value::compound_binary_op, ...),
	octave_value_typeinfo::register_binary_op 
	(octave_value::compound_binary_op, ...),
	octave_value_typeinfo::do_register_binary_class_op 
	(octave_value::compound_binary_op, ...),
	octave_value_typeinfo::do_register_binary_op 
	(octave_value::compound_binary_op, ...),
	octave_value_typeinfo::do_lookup_binary_class_op 
	(octave_value::compound_binary_op),
	octave_value_typeinfo::do_lookup_binary_op 
	(octave_value::compound_binary_op, ...)):
	New functions.
	(octave_value::do_register_type): Resize also compound_binary_ops
	field.
	* pt-exp.h (tree_expression::is_unary_expression): New virtual
	function.
	* pt-unop.h (tree_unary_expression::is_unary_expression): New virtual
	override.
	* pt-cbinop.h, pt-cbinop.cc: New files (implement
	tree_compound_binary_expression class).
	* pt-all.h: Include pt-cbinop.h.
	* Makefile.in (PT_INCLUDES, PT_SRC): Include them in the lists.
	* parse.y (make_binary_op): Call maybe_compound_binary_expression.
	* OPERATORS/op-m-m.cc (trans_mul, mul_trans): New operator handlers.
	(install_m_m_ops): Register them.
	* OPERATORS/op-cm-cm.cc (trans_mul, mul_trans, herm_mul, mul_herm): 
	New operator handlers.
	(install_cm_cm_ops): Register them.

	* DLD-FUNCTIONS/matrix_type.cc: Fix tests relying on the
	older more optimistic hermitian check.

2008-05-21  David Bateman  <dbateman@free.fr>

	* DLD-FUNCTIONS/rcond.cc (Frcond): Add support for single precision.

	* DLD-FUNCTIONS/sqrt.m: Replace DBL_* with FLT_* for single
	precision types.
	* data.cc (static octave_value fill_matrix (const
	octave_value_list&, double, float, const char *)): Add function
	with additional argument to allow for different valid for double
	and single precision.
	(Finf, FNaN, FNA, Frealmax, Frealmin): Use it here.
	(Feps): Modify behavior for a single numerical argument to give
	difference to next largest value in the class of the type passed.

2008-05-21  John W. Eaton  <jwe@octave.org>

	* pt-idx.h (tree_index_expression::tree_index_expression (int, int)): 
	Delete default argument values.

2008-05-20  David Bateman  <dbateman@free.fr>

	* data.cc (Flog2): Handle single precision.
	* ov-float.h, ov.float.cc, ov-flt-complex.h, ov-flt-complex.cc,
	ov-flt-re-mat.h, ov-flt-re-mat.cc, ov-flt-cx-mat.h,
	ov-flt-cx-mat.cc: Provide single precision version of log2 mapper
	function.

	* DLD-FUNCTIONS/__convn__.cc, DLD-FUNCTIONS/__pchip_deriv__.cc,
	DLD-FUNCTIONS/besselj.cc, DLD-FUNCTIONS/betainc.cc,
	DLD-FUNCTIONS/conv2.cc, DLD-FUNCTIONS/gammainc.cc,
	DLD-FUNCTIONS/givens.cc, DLD-FUNCTIONS/kron.cc,
	DLD-FUNCTIONS/lookup.cc, DLD-FUNCTIONS/syl.cc, data.cc:
	Prefer demotion to single precision rather than promotion to double.

	* ov-float.cc, ov-float.h, ov-flt-complex.cc, ov-flt-complex.h,
	ov-flt-cx-mat.cc, ov-flt-cx-mat.h, ov-flt-re-mat.cc,
	ov-flt-re-mat.h (numeric_conversion_function (void) const): 
	Remove method.

	* ov-complex.cc, ov-complex.h, ov-cx-mat.cc, ov-cx-mat.h,
	ov-re-mat.cc, ov-re-mat.h, ov-scalar.cc, ov-scalar.h 
	(numeric_conversion_function (void) const): Add method for
	conversion to single precision.

	* DLD-FUNCTIONS/conv2.cc (Fconv2): Don't access third arg if we
	don't have one.
	
	* DLD-FUNCTIONS/balance.cc, DLD-FUNCTIONS/expm.cc,
	DLD-FUNCTIONS/find.cc, DLD-FUNCTIONS/hess.cc,
	DLD-FUNCTIONS/qr.cc: COnvert for use with single precision.

	* OPERATORS/op-int.h, OPERATORS/op-int-conv.cc,
	OPERATORS/op-int-concat.cc: Adapt for single precision.

	* OPERATORS/op-i8-i8.cc, OPERATORS/op-i16-i16.cc, 
	OPERATORS/op-i32-i32.cc, OPERATORS/op-i64-i64.cc, 
	OPERATORS/op-ui8-ui8.cc, OPERATORS/op-ui16-ui16.cc, 
	OPERATORS/op-ui32-ui32.cc, OPERATORS/op-ui64-ui64.cc:
	Add includes for single precision types.
	
	* OPERATORS/op-b-b.cc, OPERATORS/op-b-bm.cc, OPERATORS/op-bm-b.cc,
	OPERATORS/op-fcm-fs.cc, OPERATORS/op-fcs-fs.cc,
	OPERATORS/op-fm-fs.cc, OPERATORS/op-fs-fcm.cc,
	OPERATORS/op-fs-fcs.cc, OPERATORS/op-fs-fm.cc,
	OPERATORS/op-fs-fs.cc, OPERATORS/op-int.h, ov.cc, ov-scalar.cc,
	ov-float.h, ov-flt-complex.cc, ov-float.cc, ov-flt-re-mat.cc,
	ov-flt-cx-mat.cc: Replace octave_float with octave_scalar_float

	* OPERATORS/op-fm-fm.cc, OPERATORS/op-fm-fs.cc, 
	OPERATORS/op-fm-fcm.cc, OPERATORS/op-fm-fcs.cc, 
	OPERATORS/op-fs-fm.cc, OPERATORS/op-fs-fs.cc, 
	OPERATORS/op-fs-fcm.cc, OPERATORS/op-fs-fcs.cc, 
	OPERATORS/op-fcm-fm.cc, OPERATORS/op-fcm-fs.cc, 
	OPERATORS/op-fcm-fcm.cc, OPERATORS/op-fcm-fcs.cc, 
	OPERATORS/op-fcs-fm.cc, OPERATORS/op-fcs-fs.cc, 
	OPERATORS/op-fcs-fcm.cc, OPERATORS/op-fcs-fcs.cc, 
	OPERATORS/op-m-m.cc, OPERATORS/op-m-s.cc, 
	OPERATORS/op-m-cm.cc, OPERATORS/op-m-cs.cc, 
	OPERATORS/op-s-m.cc, OPERATORS/op-s-s.cc, 
	OPERATORS/op-s-cm.cc, OPERATORS/op-s-cs.cc, 
	OPERATORS/op-cm-m.cc, OPERATORS/op-cm-s.cc, 
	OPERATORS/op-cm-cm.cc, OPERATORS/op-cm-cs.cc, 
	OPERATORS/op-cs-m.cc, OPERATORS/op-cs-s.cc, 
	OPERATORS/op-cs-cm.cc, OPERATORS/op-cs-cs.cc:
	Add mixed single/double assign operators.

	* ov.h (numeric_demotion_function): New method for double to
	single demotion.
	* ov-base.h (numeric_demotion_function): Declare virtual version.

	* ov-complex.cc, ov-complex.h, ov-cx-mat.cc, ov-cx-mat.h,
	ov-re-mat.cc, ov-re-mat.h, ov-scalar.cc, ov-scalar.h 
	(numeric_cdemote_function (void) const): Add method for
	conversion to single precision renamed from the method
	numeric_conversion_function

	* ov.cc (do_binary_op): Use demotion function seperately than the
	numeric conversion function so as to avoid isses like
	a=zeros(2,2);a(1,:)=1:2.

	* OPERATORS/op-fcm-fcm.cc, OPERATORS/op-fcm-fcs.cc,
	OPERATORS/op-fcm-fm.cc, OPERATORS/op-fcm-fs.cc,
	OPERATORS/op-fcs-fcm.cc, OPERATORS/op-fcs-fcs.cc,
	OPERATORS/op-fcs-fm.cc, OPERATORS/op-fcs-fs.cc,
	OPERATORS/op-fm-fcm.cc, OPERATORS/op-fm-fcs.cc,
	OPERATORS/op-fm-fm.cc, OPERATORS/op-fm-fs.cc,
	OPERATORS/op-fs-fcm.cc, OPERATORS/op-fs-fcs.cc,
	OPERATORS/op-fs-fm.cc, OPERATORS/op-fs-fs.cc, ov-float.cc,
	ov-float.h, ov-flt-complex.cc, ov-flt-complex.h, ov-flt-cx-mat.cc,
	ov-flt-cx-mat.h, ov-flt-re-mat.cc, ov-flt-re-mat.h: New files.
	* Makefile.in (OV_INCLUDES, OV_SRC, OP_XSRC. FLOAT_OP_XSRC,
	DOUBLE_OP_XSRC): Add them here.

	* DLD-FUNCTIONS/__convn__.cc, DLD-FUNCTIONS/__lin_interpn__.cc,
	DLD-FUNCTIONS/__pchip_deriv__.cc, DLD-FUNCTIONS/besselj.cc,
	DLD-FUNCTIONS/betainc.cc, DLD-FUNCTIONS/bsxfun.cc,
	DLD-FUNCTIONS/chol.cc, DLD-FUNCTIONS/conv2.cc,
	DLD-FUNCTIONS/det.cc, DLD-FUNCTIONS/eig.cc, DLD-FUNCTIONS/fft.cc,
	DLD-FUNCTIONS/fft2.cc, DLD-FUNCTIONS/fftn.cc,
	DLD-FUNCTIONS/fftw.cc, DLD-FUNCTIONS/filter.cc,
	DLD-FUNCTIONS/gammainc.cc, DLD-FUNCTIONS/givens.cc,
	DLD-FUNCTIONS/inv.cc, DLD-FUNCTIONS/kron.cc,
	DLD-FUNCTIONS/lookup.cc, DLD-FUNCTIONS/lu.cc,
	DLD-FUNCTIONS/matrix_type.cc, DLD-FUNCTIONS/max.cc,
	DLD-FUNCTIONS/pinv.cc, DLD-FUNCTIONS/schur.cc,
	DLD-FUNCTIONS/sqrtm.cc, DLD-FUNCTIONS/svd.cc,
	DLD-FUNCTIONS/syl.cc, DLD-FUNCTIONS/symbfact.cc,
	DLD-FUNCTIONS/typecast.cc, OPERATORS/op-b-b.cc,
	OPERATORS/op-b-bm.cc, OPERATORS/op-bm-b.cc, OPERATORS/op-bm-bm.cc,
	OPERATORS/op-cm-cm.cc, OPERATORS/op-cs-cs.cc, OPERATORS/op-m-m.cc,
	OPERATORS/op-range.cc, OPERATORS/op-s-s.cc, bitfcns.cc, data.cc,
	oct-stream.cc, ov-base.cc, ov-base.h, ov-bool-mat.cc,
	ov-bool-mat.h, ov-bool.h, ov-ch-mat.cc, ov-ch-mat.h,
	ov-complex.cc, ov-complex.h, ov-cx-mat.cc, ov-cx-mat.h, ov-intx.h,
	ov-range.cc, ov-range.h, ov-re-mat.cc, ov-re-mat.h, ov-scalar.h,
	ov.cc, ov.h, pr-output.cc, pr-output.h, pt-mat.cc, utils.cc,
	utils.h, xdiv.cc, xdiv.h, xpow.cc, xpow.h:
	Allow single precision types.

2008-05-20  David Bateman  <dbateman@free.fr>

	* DLD-FUNCTIONS/rcond.cc: New function.
	* Makefile.in (DLD_XSRC): Add it here.

	* debug.cc (Fdbstop): If no line specified assume line 1.
	(Fdbstep, Fdbcont, Fdbnext): Move debugging functions 
	to normal commands.
	(Fdbquit): New command to quit debugging mode and return to the 
	prompt.
	(Fdbstep): Modify the dbstep command for compatibility.
	* input.cc (Vdebugging_current_line): Store current line being
	debugged for use in DEFCMD versions of debug commands.
	(match_sans_spaces_semi): Delete.
	(static void get_debug_input (const std;string&)): New function to
	parse input in debug mode using standard Octave parser.
	(static octave_value_list get_user_input (const
	octave_value_list&, int)): Remove debugging specialization.
	* input.h (Vdebugging_current_line): Store current line being
	debugged for use in DEFCMD versions of debug commands.
	* parse.y (make_return_command): Special handling in debug mode.
	* pt-bp.h (MAYBE_DO_BREAKPOINT): Support break in n lines needed
	to support "dbstep N". 
	* pt.cc (tree::break_next): Convert to a down counter to support
	break in N lines. Breakpoint occure when tree::break_next is zero.
	* toplev.cc (octave_user_script *
	octave_call_stack::do_caller_user_script (difference_type) const):
	Support skipping the first N functions to support "dbstep out".
	(octave_user_function * octave_call_stack::do_caller_user_function
	(difference_type) const): Ditto.
	(octave_user_code * octave_call_stack::do_caller_user_code
	(difference_type) const): Ditto.
	* toplev.h (octave_user_script *
	octave_call_stack::do_caller_user_script (difference_type) const):
	Add difference_type argument.
	(octave_user_function * octave_call_stack::do_caller_user_function
	(difference_type) const): Ditto.
	(octave_user_code * octave_call_stack::do_caller_user_code
	(difference_type) const): Ditto.
	(static octave_user_script *caller_script (difference_type)):
	Ditto.
	(static octave_user_function *caller_user_function
	(difference_type q)): Ditto.
	(static octave_user_code *caller_user_code (difference_type q)):
	Ditto.

2008-05-20  Kim Hansen  <kimhanse@gmail.com>

	* load-path.cc (load_path::do_initialize):
	Include separator when appending sys_path.

2008-05-20  Bill Denney  <bill@denney.ws>

	* file-io.cc (Ffgets, Fgets, Ffputs, Ffscanf): Doc fix (seealso).

2008-05-20  Thomas Weber  <thomas.weber.mail@gmail.com>

	* DLD-FUNCTIONS/rand.cc (Frandn): Doc fix.

2008-05-20  David Bateman  <dbateman@free.fr>

	* load-save.cc (do_load): Treat non verbose list_only output in
	the same manner as Fwho.
	* symtab.h (static void clear_variable_regexp (const
	std::string&): New method.
	(static std::list<symbol_record> regexp (const std::string&)): Ditto.
	(static std::list<symbol_record> regexp_variables (const
	std::string&)): Ditto.
	(static std::list<symbol_record> regexp_global_variables (const
	std::string&)): Ditto,
	(static std::list<symbol_record> regexp_variables (const
	string_vector&)): Ditto.
	(void do_clear_variable_regexp (const std::string&)): Ditto.
	(std::list<symbol_record> do_regexp (const std::string&, bool)
	const): Ditto.
	(do_who): Accept the "-regexp" option. Use regexp versions of
	symbol table functions.
	(static inline bool name_match_any_pattern (const string_vector&,
	int, int, bool): Add regexp argument, and use regexp matching if
	true.
	(do_clear_variables): Add regexp option and pass to
	name_match_any_pattern.
	(Fclear): Accept the -regexp option.

2008-05-07  John W. Eaton  <jwe@octave.org>

	* pt-arg-list.cc, pt-arg-list.h (tree_argument_list::dup):
	New arg, context.
	* pt-assign.cc, pt-assign.h (tree_simple_assignment::dup,
	tree_multi_assignment::dup): Likewise.
	* pt-binop.cc, pt-binop.h (tree_binary_expression::dup,
	tree_boolean_expression::dup): Likewise.
	* pt-cell.cc, pt-cell.h (tree_cell::dup): Likewise.
	* pt-cmd.cc, pt-cmd.h (tree_no_op_command::dup,
	tree_function_def::dup): Likewise.
	* pt-colon.cc, pt-colon.h (tree_colon_expression::dup): Likewise.
	* pt-const.cc, pt-const.h (tree_constant::dup): Likewise.
	* pt-decl.cc, pt-decl.h (tree_decl_elt::dup, tree_decl_init_list::dup,
	tree_global_command::dup, tree_static_command::dup): Likewise.
	* pt-except.cc, pt-except.h (tree_try_catch_command::dup,
	tree_unwind_protect_command::dup): Likewise.
	* pt-exp.h (tree_expression:dup): Likewise.
	* pt-fcn-handle.cc, pt-exp.h (tree_expression:dup): Likewise.
	* pt-fcn-handle.h (tree_fcn_handle::dup,
	tree_anon_fcn_handle::dup): Likewise.
	* pt-id.cc, pt-id.h (tree_identifier::dup): Likewise.
	* pt-idx.cc, pt-idx.h (tree_index_expression::dup): Likewise.
	* pt-jump.cc, pt-jump.h (tree_break_command::dup,
	tree_continue_command::dup, tree_return_command::dup): Likewise.
	* pt-loop.cc, pt-loop.h (tree_while_command::dup,
	tree_do_until_command::dup, tree_simple_for_command::dup,
	tree_complex_for_command::dup): Likewise.
	* pt-mat.cc, pt-mat.h (tree_matrix::dup): Likewise.
	* pt-misc.cc, pt-misc.h (tree_parameter_list::dup,
	tree_return_list::dup): Likewise.
	* pt-select.cc, pt-select.h (tree_if_clause::dup,
	tree_if_command_list::dup, tree_if_command::dup,
	tree_switch_case::dup, tree_switch_case_list::dup,
	tree_switch_command::dup, tree_statement::dup,
	tree_statement_list::dup, tree_prefix_expression::dup,
	tree_postfix_expression::dup): Likewise.

	* ov-fcn-handle.cc (octave_fcn_handle::save_ascii,
	octave_fcn_handle::save_binary, octave_fcn_handle::save_hdf5):
 	Pass context to symbol_table::all_variables.
	* load-save.cc (dump_octave_core): Likewise.
	* ov-fcn-handle.cc (Ffunctions): Don't skip anonymous functions
	with the name set to the text of the function body.
 	Pass context to symbol_table::all_variables and
	symbol_table::symbol_record::varval.

	* symtab.h (symbol_table::varref, symbol_table::do_varref,
	symbol_table::varval, symbol_table::do_varval,
	symbol_table::all_variables, symbol_table::do_all_variables, 
	symbol_table::symbol_record::varref,
	symbol_table::symbol_record::varval,
	symbol_table::symbol_record::is_defined,
	symbol_table::symbol_record::is_variable,
	symbol_table::symbol_record::symbol_record_rep::varref,
	symbol_table::symbol_record::symbol_record_rep::varval,
	symbol_table::symbol_record::symbol_record_rep::is_defined,
	symbol_table::symbol_record::symbol_record_rep::is_variable,
	):
	* symtab.h (symbol_table::do_inherit): New arg, donor_context.
	Look for value in donor_contxt.  Set value in base context.
	* symtab.cc (symbol_table::symbol_record::symbol_record_rep::dump):
	Pass xcurrent_context to varval.

2008-05-06  David Bateman  <dbateman@free.fr>

	* ov-fcn-inline.cc (Finline): Also ignore NaN, Inf, pi, NA and eps.

	* bitfcns.cc (BITOP): Treat octave_bool types and octave_scalar.

2008-05-06  John W. Eaton  <jwe@octave.org>

	* symtab.h (symbol_table::scope_id_cache): New class.  Use it to
	replace scope_ids_in_use and scope_ids_free_list.
	(symbol_table::erase_scope): Call free_scope.
	(symbol_table::free_scope): Call scope_id_cache::free.

	* ov-fcn.h (octave_function::lock_subfunctions,
	octave_function::unlock_subfunctions): New virtual functions.
	(octave_function::lock_subfunctions): Call lock_subfunctions here.
	(octave_function::unlock_subfunctions): Call unlock_subfunctions here.
	* ov-usr-fcn.h (octave_user_function::lock_subfunctions,
	octave_user_function::unlock_subfunctions): New functions.

	* symtab.h (symbol_table::lock_subfunctions,
	symbol_table::lock_subfunctions,
	symbol_table::fcn_info::lock_subfunction,
	symbol_table::fcn_info::unlock_subfunction,
	symbol_table::fcn_info::fcn_info_rep::lock_subfunction, 
	symbol_table::fcn_info::fcn_info_rep::unlock_subfunction):	
	New functions.

	* symtab.h (symbol_table::set_scope, symbol_table::get_instance):
	Don't set instance unless allocation succeeds.
	(symbol_table::print_scope, symbol_table::do_print_scope): Delete.
	(symbol_table::free_scope): Avoid using invalid iterator.
	(symbol_table::erase_scope): Call free_scope here.

	* ov-fcn-handle.cc (octave_fcn_handle::load_ascii,
	octave_fcn_handle::load_binary, octave_fcn_handle::load_hdf5):
	Cache anonymous name here.
	(octave_fcn_handle::octave_fcn_handle): Move here from
	ov-fcn-handle.h.  Cache name if function is user-defined.

	* pt-id.cc (tree_identifier::dup): Avoid shadow warning.

	* symtab.h (symbol_table::cache_name, symbol_table::do_cache_name):
	New functions.
	(symbol_table::get_instance): Cache top-level name here.
	* parse.y (finish_function): Call symbol_table::cache_name here.

	* symtab.cc (F__dump_symtab_info__): New function.
	
	* symtab.cc (symbol_table::dump, symbol_table::dump_global,
	symbol_table::dump_functions, symbol_table::do_dump,
	symbol_table::symbol_record::symbol_record_rep::dump,
	symobl_table::fcn_info::fcn_info_rep::dump): New functions.
	* symtab.h: Provide decls.
	(symtab::scopes, symbol_table::symbol_record::dump,
	symbol_table::fcn_info::dump): New functions.
	(symtab::get_instance): New arg, create; if false throw error if
	instance for given scope is not found.

	* ov-base.cc (octave_base_value::dump): New virtual function.
	* ov-base.h: Proivde decl.
	* ov.h (octave_value::dump): New function.

	* ov.h, ov.c (octave_value::function_value): New const version.
	* ov-base.h, ov-base.cc (octave_base_value::function_value): Likewise.
	* ov-builtin.h (octave_builtin::function_value): Likewise.
	* ov-fcn-handle.h (octave_fcn_handle::function_value): Likewise.
	* ov-mex-fcn.h (octave_mex_function::function_value): Likewise.
	* ov-usr-fcn.h (octave_user_script::function_value,
	octave_user_function): Likewise.
	
	* symtab.h, symtab.cc: Use consistent naming scheme for iterator
	typedefs.  Change all uses.

	* variables.cc (F__print_symtab_info__, F__print_symbol_info__):
	Delete.

	* ov-cell.cc (Fstruct2cell): Handle structure arrays properly.

2008-05-05  David Bateman  <dbateman@free.fr>

	* sysdep.cc (Fputenv): Allow single arg. Alias to setenv.

2008-05-05  John W. Eaton  <jwe@octave.org>

	* input.cc (Fre_read_readline_init_file): New function.

2008-05-04  John W. Eaton  <jwe@octave.org>

	* ov-fcn-handle.cc (Ffunctions): Pass octave_value object instead
	of Cell in structure field assignment.

	* parse.y (frob_function): Don't install subfunctions here.
	(finish_function): Handle subfunctions here.
	Conditionally define tree_function_def object here.

	* symtab.h (symbol_table::fcn_info::fcn_info_rep::find_function):
	Initialize args_evaluated.
	* pt-id.cc (tree_identifier::rvalue): Likewise.
	* variables.cc (symbol_exist): Likewise.

	* pt-id.h (tree_identifier::tree_identifer): Delete useless statement.

	* variables.cc (get_global_value): Use symbol_table::global_varval
	instead of passing scope to symbol_table::varval.
	(set_global_value): Use symbol_table::global_varref
	instead of passing scope to symbol_table::varref.
	(do_who): Use symbol_table::glob_global_variables
	instead of passing scope to symbol_table::glob_variables.
	(do_clear_globals):
	Use symbol_table::global_variable_names and
	symbol_table::clear_global instead of passing scope to
	symbol_table::clear_variables.
	Use symbol_table::global_variable_names instead of passing scope
	to symbol_table::variable_names.

	* unwind-prot.cc (unwind_protect::save_size_t): New function.
	(saved_variable::restore_value): Handle size_t values.
	(saved_variable::saved_variable): New size_t constructor.
	(saved_variable::size_type): New var_type enum value.
	(saved_variable::ptr_to_size_t, saved_variable::size_t_value):
	New union elements.
	* unwind-prot.h (unwind_protect::save_size_t): Provide decl
	(unwind_protect_size_t): New macro.

	* toplev.cc (octave_call_stack::do_goto_frame,
	octave_call_stack::do_goto_frame_relative): New functions.
	(octave_call_stack::do_backtrace): Also return scope and context ids.
	* toplev.h 	(octave_call_stack::do_goto_frame,
	octave_call_stack::do_goto_frame_relative): Provide decls.
	(octave_call_stack::curr_frame): New data member.
	(octave_call_stack::octave_call_stack): Initialize it.
	(octave_call_stack::call_stack_elt::scope,
	(octave_call_stack::call_stack_elt::context): New data members.,
	(octave_call_stack::current_frame,
	(octave_call_stack::do_current_frame
	octave_call_stack::do_goto_frame,
	octave_call_stack::do_goto_frame_relative): New functions.
	(octave_call_stack::push, octave_call_stack::do_push):
	New args, scope and context.
	(octave_call_stack::do_push, octave_call_stack::do_pop): Update
	curr_frame if debugging.

	* symtab.cc (symbol_table::global_table): New map for global values.
	(symbol_table::xcurrent_context): New variable for context info.
	* symtab.h: Provide decls.
	(symbol_table::global_varref, symbol_table::global_varval,
	symbol_table::current_context,
	symbol_table::set_scope_and_context,
	symbol_table::glob_global_variables,
	symbol_table::global_variable_names,
	symbol_table::top_leve_variable_names):
	New functions.
	(symbol_table::context_id, const_global_table_iterator,
	global_table_iterator): New typedefs.
	(symbol_table::xcurrent_context_this_table): New variable.
	(symbol_table::symbol_table): Initialize it.
	(symbol_table::set_scope): Use it.
	(symbol_table::symbol_record::find, symbol_table::do_find):
	Use symbol_table::global_varref instead of passing scope to
	symbol_table::varref.
	(symbol_table::symbol_record::symbol_record_rep::value_stack):
	Now a std::deque instead of a std::stack.
	(symbol_table::symbol_record::symbol_record_rep::push_context,
	symbol_table::symbol_record::symbol_record_rep::pop_context):
	Don't push or pop persistent of global variables.
	(symbol_table::symbol_record::symbol_record_rep::varval,
	Handle context here.  Call symbol_table::global_varval instead of
	passing scope to symbol_table::varval.
	(symbol_table::symbol_record::symbol_record_rep::varref):
	Handle context here.  Call symbol_table::global_varref instead of
	passing scope to symbol_table::varref.
	(symbol_table::symbol_record::varref,
	symbol_table::symbol_record::varval): No need to handle global
	values specially here.
	(symbol_table::symbol_record::push_context
	symbol_table::symbol_record::pop_context: No need to handle global
	or persistent values specially here.
	(symbol_table::get_instance): Don't return global scope.
	(symbol_table::do_push_context, symbol_table::do_pop_context):
	Increment/decrement xcurrent_context.
	(symbol_table::do_clear_global, symbol_table::do_clear_global_pattern):
	Use global_table instead of separate scope.	

	* symtab.h, symtab.cc (symbol_table::fcn_info::fcn_info_rep::find,
	symbol_table::fcn_info::fcn_info_rep::find_function,
	symbol_table::fcn_info::find, symbol_table::fcn_info::find_function,
	symbol_table::find, symbol_table::find_function,
	symbol_table::do_find, symbol_table::insert, symbol_table::varref,
	symbol_table::varval, symbol_table::persistent_varref,
	symbol_table::persistent_varval, symbol_table::erase_persistent,
	symbol_table::is_variable, symbol_table::clear,
	symbol_table::clear_variables, symbol_table::clear_global,
	symbol_table::clear_variable, symbol_table::clear_global_pattern,
	symbol_table::clear_variable_pattern, symbol_table::push_context,
	symbol_table::pop_context, symbol_table::mark_hidden,
	symbol_table::mark_global, symbol_table::glob,
	symbol_table::glob_variables, symbol_table::variable_names,
	symbol_table::is_local_variable, symbol_table::is_global):
	Eliminate scope arg.  Change all uses.
	(symbol_table::erase_scope, symbol_table::dup_scope):
	Require scope arg.

	* pt-stmt.cc (tree_statement::eval): Don't update statement info
	in octave_call_stack if debugging.

	* pt-id.h (tree_identifier::xsym): New function.
	(tree_identifier::rvalue, tree_identifier::lvalue,
	tree_identifier::dup, tree_identifier::is_defined,
	tree_identifier::is_variable, tree_identifier::mark_global,
	tree_identifier::mark_as_static, tree_identifier::do_lookup,
	tree_identifier::mark_as_formal_parameter):
	Use xsym to access symbol.
	(tree_identifier::scope): New data member.
	(tree_identifier::tree_identifier): Initialize it.

	* ov-usr-fcn.cc (octave_user_script::octave_user_script,
	octave_user_function::octave_user_function):
	Initialize call_depth to -1.
	(octave_user_script::do_multi_index_op): Fix comparison of
	call_depth with Vmax_recursion_depth.
	(octave_user_function::do_multi_index_op):
	Compare call depth > 0, instead of > 1.
	* ov-usr-fcn.h (octave_user_function::save_args_passed):
	Compare call depth > 0, instead of > 1.

	* input.cc (Vdebugging): New variable.
	(get_user_input): Eliminate DEBUG argument.  Use global Vdebugging
	variable instead.  Change all callers.
	(saved_frame): New static variable.
	(restore_frame): New function.
	(do_keyboard): Unwind-protect Vdebugging here and set it to TRUE.
	Save current frame.  Use unwind_protect to restore it.
	(Fkeyboard): Save current frame.  Use unwind_protect to restore it.
	Move up the call stack one frame before calling do_keyboard.  
	* input.h (Vdebugging): Provide decl.

	* mex.cc (mexGetVariable): Handle global vars with get_global_value.
	(mexPutVariable): Likewise, with set_global_value.

	* octave.cc (intern_argv): Assert that we are at the top level.
	Don't pass scope to symbol_table::varref or symbol_table::mark_hidden.

	* load-save.cc (install_loaded_variable): Use
	symbol_table::global_varref instead of passing global scope to
	symbol_table::varref.

	* help.cc (make_name_list): Call symbol_table::global_variable_names
	and symbol_table::top_level_variable_names instead of passing
	scope to symbol_table::variable_names.

	* debug.cc (do_dbupdown, Fdbup, Fdbdown): New functions.
	(current_stack_frame): Delete static variable.
	(Fdbstack): Rename from F__dbstack__.  Implement all of dbstack
	here instead of partially in dbstack.m.

2008-05-03  John W. Eaton  <jwe@octave.org>

	* pt-arg-list.cc (F__end__): If there are more dimensions than
	indices, smash extra dimensions first.
	(num_indices): New static variable.
	(tree_argument_list::convert_to_const_vector): Save and set it.

	* parse.y (parse_fcn_file): Also temporarily set parser_end_of_input
	and get_input_from_eval_string to false while reading script files.

2008-05-01  John W. Eaton  <jwe@octave.org>

	* parse.y (load_fcn_from_file): Expect
	* dynamic-ld.cc (octave_dynamic_loader::do_load_oct):
	Search currently loaded .oct files by file name.  Don't search
	currently loaded files for functions.
	(octave_shlib_list::find_file, octave_shlib_list::do_find_file):
	New functions.
	(octave_shlib_list::search, octave_shlib_list::do_search): Delete.

	* dynamic-ld.cc (octave_dynamic_loader::do_load_oct): No need to
	do load_path lookups here.
	(octave_shlib_list::iterator, octave_shlib_list::const_iterator):
	New typedefs.
	(octave_shlib_list::do_remove, octave_shlib_list::do_search,
	octave_shlib_list::do_display): Use them.
	(octave_mex_file_list::iterator, octave_mex_file_list::const_iterator):
	New typedefs.
	(octave_mex_file_list::do_remove): Use them.

2008-05-03  Rafael Laboissiere <rafael@debian.org>

	* DLD-FUNCTIONS/time.cc, file_io.cc: Use ischar instead of
	deprecated isstr.

2008-04-30  John W. Eaton  <jwe@octave.org>

	* dynamic-ld.cc (octave_dynamic_loader::do_load_oct):
	If function already exists, reload it anyway.  Clear existing
	oct_file only if reloading a function from the same file.
	(octave_shlib_list::display): New static function.
	(octave_shlib_list::do_display): New function.

	* symtab.cc (out_of_date_check_internal): Better handling for
	functions found in files by relative file names.
	(load_out_of_date_fcn): New function.

	* ls-oct-ascii.cc (extract_keyword): Return early if first char is
	not a comment character.
	(read_ascii_data): Accept .nargin. and .nargout. as valid identifiers.

	* ov-fcn-handle.cc: Combine tests.  Test saving and loading
	handles for built-in, .oct, and .m functions.
	(octave_fcn_handle::save_ascii, octave_fcn_handle::save_binary,
	octave_fcn_handle::save_hdf5, octave_fcn_handle::print):
	Avoid dereferencing invalid pointer.

	* data.cc: Fix tests.  Use "%!assert", not "%! assert" for
	individual assert tests.

2008-04-30  Jaroslav Hajek  <highegg@gmail.com>

	* ov-base.cc, ov-base.h, ov-bool-mat.h, ov-bool-sparse.h,
	ov-bool.h, ov-complex.cc, ov-complex.h, ov-cx-mat.cc, ov-cx-mat.h,
	ov-cx-sparse.cc, ov-cx-sparse.h, ov-intx.h, ov-range.h,
	ov-re-mat.cc, ov-re-mat.h, ov-re-sparse.cc, ov-re-sparse.h,
	ov-scalar.cc, ov-scalar.h, ov.h:
 	Provide log2 mapper function.
	* data.cc (Flog2): New function.

2008-04-25  John W. Eaton  <jwe@octave.org>

	* pt-stmt.h (tree_statement_list::function_body): New data member.
	(tree_statement_list::tree_statement_list): Initialize it.
	(tree_statement_list::mark_as_script_body): New function.
	(tree_statement::maybe_echo_code, tree_statement::eval):
	Rename in_function_body argument to in_function_or_script_body.
	* pt-stmt.cc (tree_statement::eval): Only set statement in call
	stack if in_function_or_script_body is true.

	* pt-stmt.cc (tree_statement_list::eval): Call elt->eval with
	function_body || script_body.
	* ov-usr-fcn.cc (octave_user_script::octave_user_script):
	command list as script body here.
	(octave_user_function::octave_user_function):
	Mark command list as function body here.
	* parse.y (start_function, make_anon_fcn_handle): Not here.

	* toplev.h, toplev.cc (octave_call_stack::backtrace,
	octave_call_stack::do_backtrace): New arg, N.  Skip innermost N
	stack frames.

	* debug.cc (F__dbstack__): New function.
	(current_stack_frame): New static variable.

	* error.cc (verror, pr_where): Use octave_call_stack instead of
	tree_statement stack to get line and column information.
	(pr_where): Use octave_call_stack instead of tree_statement stack
	to get current statement.
	* input.cc (get_user_input): Extract current line number from
	octave_call_stack instead of tree_statement_stack.
	* pt-stmt.cc (tree_statement::eval): Put current statement on
	octave_call_stack instead of tree_statement_stack.
	* pt-stmt.h, pt-stmt.cc (class tree_statement_stack): Delete.

	* toplev.h, toplev.cc (octave_call_stack::call_stack_elt):
	New nested struct.
	(octave_call_stack::cs): Now a deque of call_stack_elts.
	Change all uses.
	(octave_call_stack::current_statement,
	octave_call_stack::current_line,
	octave_call_stack::current_column,
	octave_call_stack::top_statement,
	octave_call_stack::set_statement, octave_call_stack::backtrace):
	New static functions.
	(octave_call_stack::do_current_line,
	octave_call_stack::do_current_column,
	octave_call_stack::do_top_statement,
	octave_call_stack::do_set_statement,
	octave_call_stack::do_backtrace): New member functions.

2008-04-24  John W. Eaton  <jwe@octave.org>

	* toplev.h, toplev.cc (octave_call_stack::unwind_pop_script):
	Delete unused function.

	* ov-usr-fcn.cc: Move constructor definitions here, from ov-usr-fcn.h.
	* ov-usr-fcn.h (octave_user_script::octave_user_script): 
	Also Initialize t_parsed and t_checked.

2008-04-24  Jaroslav Hajek  <highegg@gmail.com>

	* ov-usr-fcn.h (octave_user_script::octave_user_script): 
	Initialize call_depth.

2008-04-24  John W. Eaton  <jwe@octave.org>

	* file-io.cc (do_stream_open): Return -1 for directories.

2008-04-23  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/__qp__.cc (qp): Avoid bounds error when removing
	constraint from active set.

	* lex.l (text_yyinput): New function.  Use it in place of yyinput.
	(next_token_is_sep_op, scan_for_comments, eat_whitespace,
	have_continuation): No need to check for CR or CRLF.
	* parse.y (text_getc): Also return NL for single CR.

2008-04-32  Michael Goffioul  <michael.goffioul@gmail.com>

	* input.cc (get_input_from_file): Open file in binary mode.

2008-04-20  John W. Eaton  <jwe@octave.org>

	* oct-stream.cc (octave_stream::read): Allow single data type
	specification but return double.

2008-04-18  John W. Eaton  <jwe@octave.org>

	* lex.l, lex.h (process_comment): New arg, start_in_block.  Call
	grab_block_comment if start_in_block is true.  Change all uses.
	* lex.l (grab_block_comment): New function.
	(grab_comment_block): New arg, at_bol.  Change all uses.
	Call grab_block_comment if we find the start of a block comment.
	(block_comment_nesting_level): New static variable.
	(^{S}*{CCHAR}\{{S}*{NL}): New rule.
	(<<EOF>>): Warn about open block comments.
	(reset_parser): Set block_comment_nesting_level to zero.
	* parse.y (parse_fcn_file): Stash help text from
	gobble_leading_white_space after calling reset_parser.
	(text_getc): Keep track of input_line_number here.
	(skip_white_sapce): Don't increment input_line_number here.
	* lex.l (grab_comment_block): Or here.

	* lex.l (Vdisplay_tokens): New static variable.
	(F__display_tokens__): New function.
	(display_token): New function.
	(DISPLAY_TOK_AND_RETURN): New macro.
	(COUNT_TOK_AND_RETURN): Use DISPLAY_TOK_AND_RETURN.
	(<COMMAND_START>[\"\'], "'", \"): Move handle_string outside of
	COUNT_TOK_AND_RETURN macro parameter list.
	(handle_identifier): Don't use macros to return token values here.
	(<MATRIX_START>{S}*{COMMENT}{SNLCMT}*|<MATRIX_START>{S}*{NL}{SNLCMT}*):
	Recognize block comments here.
	
	* pr-output.cc (Fdisp): If nargout > 0, produce an sq-string
	unless arg is a dq-string.

2008-04-17  John W. Eaton  <jwe@octave.org>

	* parse.y (looks_like_copyright): Handle leading whitespace.
	(class stdio_stream_reader): New class.
	(skip_white_space): New function.
	(gobble_leading_white_space): New arg, EOF.  Change all uses.
	Use skip_white_space and grab_comment_block to process comments.
	(process_leading_comments): Delete.
	(parse_fcn_file): Call gobble_leading_white_space instead of
	process_leading_comments.
	Skip parsing if gobble_leading_white_space detects EOF.
	* lex.l (process_comment): Delete CCHAR arg.  Change all uses.
	({CCHAR}): Unput comment character before calling process_comment.
	(grab_comment_block): Rename from grab_help_text.  Don't discard
	spaces from comment text.  Update input_line_number and
	current_input_column as characters are read.  Warn only once about
	incompatible comment characters in a given block of comments.
	Use stream_reader class instead of accessing yyinput and yyunput
	directly.
	(class flex_stream_reader): New class.
	(process_comment): Use grab_comment_block to handle all comments.
	Don't call maybe_gripe_matlab_incompatible_comment.
	Always store comment text returned by grab_comment_block.
	* lex.h (grab_comment_block): Provide decl.
	(class stream_reader): New class.

2008-04-16  John W. Eaton  <jwe@octave.org>

	* parse.y (Fautoload, Fmfilename): Call
	octave_call_stack::caller_user_code, not
	octave_call_stack::caller_user_script_or_function.
	* toplev.cc, toplev.h (octave_call_stack::caller_user_code):
	Rename from octave_call_stack::caller_user_script_or_function.
	(octave_call_stack::do_caller_user_code): Rename from
	octave_call_stack::do_caller_user_script_or_function.
	* ov-usr-fcn.h (class octave_user_code): New class, derived from
	octave_fucntion.
	(class octave_user_function, class octave_user_script): Derive
	from octave_user_code, not octave_function.
	(octave_user_script::user_code_value,
	octave_user_function::user_code_value): New functions.
	* ov.cc (octave_value::user_script_value,
	octave_value::user_code_value): New functions.
	* ov.h: Provide decls.
	(octave_value::is_user_code, octave_value::is_user_script):
	New functions.
	* ov-base.cc (octave_base_value::user_script_value,
	octave_base_value::user_code_value): New virutal functions.
	* ov-base.h: Provide decls.
	(octave_base_value::is_user_script, octave_base_value::is_user_code):
	New virtual functions.
	* error.cc (verror, pr_where, error_2, warning_1):
	Call octave_call_stack::caller_user_code instead of
	octave_call_stack::caller_user_script_or_function
	* input.cc (get_user_input): Likewise.
	* debug.h (bp_table::breakpoint_map): Use pointer to
	octave_user_code instead of octave_user_function.
	* debug.cc (get_user_code): Rename from get_user_function.
	Return pointer to octave_user_code instead of octave_user_function.
	Change all uses.
	(bp_table::do_add_breakpoint, bp_table::do_remove_breakpoint,
	bp_table::do_remove_all_breakpoints_in_file,
	bp_table::do_get_breakpoint_list):
	Avoid dereferencing invalid pointers.
	(parse_dbfunction_params): Call
	octave_call_stack::caller_user_code, not
	octave_call_stack::caller_user_script_or_function. 

	* load-save.cc (Fsave): Fix continuation character in doc string.

	* pt-walk.h (tree_walker::visit_function_def): New function.
	* pt-pr-code.cc (tree_print_code::visit_function_def): New function.
	* pt-pr-code.h: Provide decl.
	* pt-bp.cc (tree_breakpoint::visit_function_def): New function.
	* pt-bp.h: Provide decl.
	* pt-check.cc (tree_checker::visit_function_def): New function.
	* pt-check.h: Provide decl.

	* parse.y (finish_function): Return ptr to tree_function_def object
	if curr_fcn_ptr is not defined.  Avoid dereferencing invalid pointers.
	(frob_function): Set curr_fcn_ptr only when reading a function file.
	(function): Call finish function in both cases to generate value
	of production.

	* pt-cmd.h, pt-cmd.cc (tree_function_def): New class.

	* parse.y (parse_and_execute): Delete.
	* parse.h: Delete decl.

	* pt-pb.cc (tree_breakpoint::visit_octave_user_script): New function.
	* pt-pb.h: Provide decl.
	* pt-check.cc (tree_checker::visit_octave_user_script): New function.
	* pt-check.h: Provide decl.
	* pt-pr-code.cc (tree_print_code::visit_octave_user_script):
	New function.
	* pt-pr-code.h: Provide decl.
	* pt-walk.h (visit_octave_user_script): New pure virtual function.

	* parse.y (make_script, process_leading_comments,
	looking_at_function_keyword): New static functions.
	(SCRIPT): New token.
	(script): New non-terminal.
	(command): Handle script here.
	(parse_fcn_file): Eliminate EXEC_SCRIPT argument.  Change all
	callers.  Parse scripts and return an octave_user_script rather
	than executing them here.
	(load_fcn_from_file): Handle script and function files the same.
	(source_file): Call d_multi_index_op on object returned by
	parse_fcn_file.
	(is_function_file): Delete.
	(gobble_leading_white_space): Don't recurse.  Simplify.  Eliminate
	SKIP_CODE, IN_PARTS, and SAVE_COPYRIGHT arguments.  Change all
	callers.
	(octave_function_ptr): Delete unused typedef.
	(get_help_from_file): If help text isn't found by
	gobble_leading_whitespace, parse file to find it.

	* lex.l (SCRIPT_FILE_BEGIN): New exclusive start state.
	(prep_lexer_for_script): New function.
	(grab_help_text): New arg, EOF.
	(process_comment): New function.
	({CCHAR}): Use it.
	* lex.h (prep_lexer_for_script): Provide decl.

	* lex.h, lex.cc, parse.y (lexical_feedback::beginning_of_function):
	Delete data member and all uses.

	* ov-usr-fcn.cc (octave_user_script::~octave_user_script):
	Move destructor here from ov-usr-fcn.h.  Delete cmd_list.
	(octave_user_script::subsref, octave_user_script::accept,
	octave_user_script::traceback_error): New functions.
	(octave_user_script::do_multi_index_op):
	Execute cmd_list instead of calling source_file.
	* ov-usr-fcn.h (octave_usr_script::cmd_list,
	octave_usr_script::t_parsed, octave_usr_script::t_checked,
	octave_usr_script::call_depth): New data members.
	(octave_usr_script::octave_usr_script): Initialize all data members.
	(octave_usr_script::octave_usr_script (const std::string&, const
	std::string&, tree_statement_list *, const std::string&)):
	New constructor.
	(octave_usr_script::function_value,
	octave_usr_script::user_script_value, mark_fcn_file_up_to_date,
	octave_usr_script::stash_fcn_file_time,
	octave_usr_script::time_parsed, octave_usr_script::time_checked,
	octave_usr_script::subsref, octave_usr_script::body): New functions.
	(octave_user_script::subsref, octave_user_script::accept,
	octave_user_script::traceback_error): Provide decls.

2008-04-14  Jaroslav Hajek  <highegg@gmail.com>

	* oct-stream.cc (octave_scan_1): Ensure digit following X is hex
	digit before reading number as hex.

2008-04-14  John W. Eaton  <jwe@octave.org>

	* file-io.cc (Ffread): Allow SKIP arg to be omitted.
	(Ffwrite): Likewise.  Handle args in a way consistent with Ffread.

2008-04-09  Michael Goffioul  <michael.goffioul@gmail.com>

	* DLD-FUNCTIONS/dispatch.cc: Replace system("echo '...'>...") calls
	with real file writing.

2008-04-07  Jaroslav Hajek  <highegg@gmail.com>

	* DLD-FUNCTIONS/qr.cc (Fqrshift): New function.
	* DLD-FUNCTIONS/chol.cc (Fcholinsert, Fcholdelete, Fcholshift):
	New functions.

2008-04-04  John W. Eaton  <jwe@octave.org>

	* parse.y (make_constant): Handle escape sequences in dq-strings.

2008-04-03  John W. Eaton  <jwe@octave.org>

	* parse.y (make_constant): Also stash original text for strings.

	* ov-fcn-handle.cc (octave_fcn_handle::subsref):
	Don't call next_subsref here.

2008-04-01  John W. Eaton  <jwe@octave.org>

	* pt-id.h (do_lookup (bool&, bool)): Delete.
	(do_lookup (tree_argument_list *, const string_vector&,
	octave_value_list&, bool&)): Call MAYBE_DO_BREAKPOINT here.

2008-04-02  David Bateman  <dbateman@free.fr>

	* graphics.cc (void gnuplot_backend::close_figure (const
	octave_value&) const): Allow for an input and output stream.

2008-03-28  Jaroslav Hajek  <highegg@gmail.com>

	* DLD-FUNCTIONS/lookup.cc: New file.
	* Makefile.in (DLD_XSRC): Add it to the list.

2008-03-28  David Bateman  <dbateman@free.fr>

	* ov-complex.cc (SCALAR_MAPPER, CD_SCALAR_MAPPER): New macro for
	complex values with zero imaginary part.
	(erf, erfc, gamma, lgamma): Use the new mappers to define these
	mapper functions.
	* ov-complex.h (erf, erfc, gamma, lgamma): Declare them.
	* ov-cx-mat.cc (any_element_less_than, any_element_greater_than):
	New static functions
	(DARRAY_MAPPER, CD_ARRAY_MAPPER): New macro for complex values
	with zero imaginary part.
	(erf, erfc, gamma, lgamma): Use the new mappers to define these
	mapper functions.
	* ov-cx-mat.h (erf, erfc, gamma, lgamma): Declare them.
	* ov-cx-sparse.cc (any_element_less_than, any_element_greater_than):
	New static functions
	(DSPARSE_MAPPER, CD_SPARSE_MAPPER): New macro for complex values
	with zero imaginary part.
	(erf, erfc, gamma, lgamma): Use the new mappers to define these
	mapper functions.
	* ov-cx-sparse.h (erf, erfc, gamma, lgamma): Declare them.
	* ov-re-sparse.cc (CD_SPARSE_MAPPER): Use correct mapper functors.
	* mapper.cc: Add tests for the above cases.

2008-03-27  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/max.cc: Rename from minmax.cc.
	* Makefile.in (DLD_XSRC): Rename minmax.cc to max.cc.

	* DLD-FUNCTIONS/__convn__.cc (convn): Use traits class and
	typedefs to allow all types to be deduced from argument types.

2008-03-27  Søren Hauberg  <hauberg@gmail.com>

	* DLD-FUNCTIONS/__convn__.cc (Fconvn): Allow convolving real data with
	complex data.

2008-03-26  John W. Eaton  <jwe@octave.org>

	* ov-range.h (octave_range::subsref (const std::string&,
	const std::list<octave_value_list>&, int)):
	Forward to simple subsref.
	* ov-base-sparse.h (octave_base_sparse<T>::subsref (const std::string&,
	const std::list<octave_value_list>&, int)): Likewise.
	* ov-base-scalar.h (octave_base_sparse<T>::subsref (const std::string&,
	const std::list<octave_value_list>&, int)): Likewise.
	* ov-base-matrix.h (octave_base_matrix<T>::subsref (const std::string&,
	const std::list<octave_value_list>&, int)): Likewise.

	* ov-struct.cc, ov-cell.h (octave_cell::subsref (const
	std::string&, const std::list<octave_value_list>&, int)):
	Define nargout version of subsref instead of simple version.
	* ov-struct.cc, ov-struct.h (octave_struct::subsref (const
	std::string&, const std::list<octave_value_list>&, int)):
	Define nargout version of subsref instead of simple version.

	* ov-builtin.h (octave_builtin::subsref (const std::string&,
	const std::list<octave_value_list>&): Forward to nargout subsref.
	* ov-cell.h (octave_cell::subsref (const std::string&,
	const std::list<octave_value_list>&)):
	* ov-class.h (octave_class::subsref (const std::string&,
	const std::list<octave_value_list>&): Likewise.
	* ov-fcn-handle.h (octave_fcn_handle::subsref (const std::string&,
	const std::list<octave_value_list>&): Likewise.
	* ov-list.h (octave_list::subsref (const std::string&,
	const std::list<octave_value_list>&): Likewise.
	* ov-mex-fcn.h (octave_mex_function::subsref (const std::string&,
	const std::list<octave_value_list>&): Likewise.
	* ov-struct.h (octave_struct::subsref (const std::string&,
	const std::list<octave_value_list>&)):
	* ov-usr-fcn.h (octave_user_function::subsref (const std::string&,
	const std::list<octave_value_list>&): Likewise.

	* ov.cc (octave_value::subsref): Use value of nargout instead of
	is_constant method to decide which type of subsref method to call.
	(octave_value::assign): Always call simple subref method here.

	* DLD-FUNCTIONS/chol.cc, DLD-FUNCTIONS/det.cc,
	DLD-FUNCTIONS/getpwent.cc, DLD-FUNCTIONS/inv.cc,
	DLD-FUNCTIONS/qr.cc, DLD-FUNCTIONS/symrcm.cc, file-io.cc):
	Texinfo fixes.

2008-03-26  Søren Hauberg  <hauberg@gmail.com>

	* DLD-FUNCTIONS/__convn__.cc (Fconvn):
	Call complex_array_value to extract N-d array.

2008-03-26  John W. Eaton  <jwe@octave.org>

	* ov-base-sparse.cc (octave_base_sparse<T>::print_raw):
	Also display percentage of elements that are nonzero.

2008-03-25  Søren Hauberg  <hauberg@gmail.com>

	* DLD-FUNCTIONS/__convn__.cc: New file.
	* Makefile.in: Add __convn__.cc

2008-03-25  David Bateman  <dbateman@free.fr>

	* DLD-FUNCTIONS/hex2num.cc: New function
	* Makefile.in (DLD_XSRC): Add hex2num.cc.

2008-03-25  Jaroslav Hajek  <highegg@gmail.com>

	* mappers.cc (Fexpm1, Flog1p): New functions.
	* ov-base.cc, ov-base.h, ov-bool-mat.h, ov-bool-sparse.h,
	ov-bool.h, ov-complex.cc, ov-complex.h, ov-cx-mat.cc, ov-cx-mat.h,
	ov-cx-sparse.cc, ov-cx-sparse.h, ov-range.h, ov-re-mat.cc,
	ov-re-mat.h, ov-re-sparse.cc, ov-re-sparse.h, ov-scalar.cc,
	ov-scalar.h, ov.h:
	Provide expm1 and log1p functions.

	* mappers.cc (Froundb): New functions
	* ov-base.cc, ov-base.h, ov-bool-mat.h, ov-bool-sparse.h,
	ov-bool.h, ov-complex.cc, ov-complex.h, ov-cx-mat.cc, ov-cx-mat.h,
	ov-cx-sparse.cc, ov-cx-sparse.h, ov-intx.h, ov-range.h,
	ov-re-mat.cc, ov-re-mat.h, ov-re-sparse.cc, ov-re-sparse.h,
	ov-scalar.cc, ov-scalar.h, ov.h:
 	Provide roundb mapper function.

2008-03-25  Jaroslav Hajek  <highegg@gmail.com>

	* load-save.cc (save_vars): Handle -struct modifier.
	(save_fields): New function.
	(Fsave): Document new feature.

2008-03-25  John W. Eaton  <jwe@octave.org>

	* lex.h (lexical_feedback::looking_at_initializer_expression):
	New data member.
	* lex.l (lexical_feedback::init): Initialize it.
	(handle_identifier): Don't unconditionally force identifiers to be
	variables in the current scope.  Don't call force_local_variable
	for symbols that appear in parameter initializer expressions.
	* parse.y (decl_param_init): New parser "subroutine".
	(decl2): Use it.  Set lexer_flags.looking_at_initializer_expression
	to false after parsing initializer.

2008-03-24  David Bateman  <dbateman@free.fr>

	* data.cc (map_s_s): Fix for sparse/sparse mappers that resulted
	in an empty sparse matrix being returned.
	(Fhypot): New function based on the libm hypot function.

2008-03-24  Primozz Peterlin  <primozz.peterlin@gmail.com>

	* variables.cc (Fexist): Doc fix.

2008-03-23  David Bateman  <dbateman@free.fr>

	* OPERATORS/op-int.h: Add el_div and el_ldiv operators to the
	binops that were missing them. Added elem_pow functions for mixed
	integer floating point cases. Initialize the mixed integer
	floating point cases.

2008-03-21  John W. Eaton  <jwe@octave.org>

	* ov-cell.h (octave_cell::is_constant): Return true.
	* ov-cell.h, ov-cell.cc (octave_cell::subsref (const std::string&,
	const std::list<octave_value_list>&)): Define.
	(octave_cell::subsref (const std::string&,
	const std::list<octave_value_list>&, int)): Call panic_impossible.

	* ov-struct.h (octave_struct::is_constant): New function.
	* ov-struct.h, ov-struct.cc (octave_struct::subsref (const std::string&,
	const std::list<octave_value_list>&)): Define.
	(octave_struct::subsref (const std::string&,
	const std::list<octave_value_list>&, int)): Call panic_impossible.

2008-03-21  David Bateman  <dbateman@free.fr>

	* DLD-FUNCTIONS/amd.cc: New file.
	* Makefile.in (DLD_XSRC): Add amd.cc.

2008-03-20  David Bateman  <dbateman@free.fr>

	* Cell.cc (Cell Cell::diag (void) const): delete.
	(Cell Cell::diag (octave__idx_type) const):Rewrite in terms of
	template classes function.
	* Cell.h (Cell diag (void) const): delete.

	* ov.h (octave_value diag (octave_idx_type) const): New method.
	* ov-base.h (virtual octave_value diag (octave_idx_type) const):
	New virtual method.
	* ov-base.cc (octave_value octave_base_value::diag
	(octave_idx_type) const): New default method.
	
	* ov-base-mat.h (octave_value diag (octave_idx_type) const): New
	method. 
	* ov-base-sparse.h (octave_value diag (octave_idx_type) const): New
	method. 
	* ov-base-scalar.h (octave_value diag (octave_idx_type) const): New
	method. 
	* ov-range.h (octave_value diag (octave_idx_type) const): New
	method. 

	* data.cc (make_diag, make_spdiag): Delete.
	(Fdiag): Rewrite in terms of octave_value diag function.
		
	* data.cc (static octave_value make_diag (const Cell&,
	octave_idx_type)): New instantiation of template function.
	(static octave_value make_diag (const octave_value&,
	octave_idx_type)): Allow cell arrays.

	* Cell.cc (Cell Cell::diag (void) const, Cell Cell::diag
	(octave__idx_type)): New methods for diagonal matrices.
	* Cell.h (Cell Cell::diag (void) const, Cell Cell::diag
	(octave__idx_type)): Declare them.

2008-03-18  David Bateman  <dbateman@free.fr>

	* ov-re-mat.cc (lgamma): Convert to a allow negative arguments.
	* ov-re-sparse.cc (lgamma): ditto.
	* ov-scalar.cc (lgamma): ditto.

	* DLD-FUNCTIONS/minmax.cc: 64-bit indexing fix.

2008-03-13  John W. Eaton  <jwe@octave.org>

	* ov-usr-fcn.cc (octave_user_function::octave_user_function):
	Handle num_named_args in initialization list instead of functinon body.

	* octave.gperf: Eliminate varargin and varargout keywords.
	* lex.l (is_keyword_token): Eliminate varargin_kw and varargout_kw
	from switch statement.
	* parse.y (return_list): Eliminate special cases for VARARGOUT.
	Call validate on tree_parameter_list object.
	(param_list1): Likewise, for VARARGIN.
	* pt-misc.cc (tree_parameter_list::validate): New function.
	(tree_parameter_list::mark_varargs_only): Now private.
	(tree_parameter_list::mark_varargs): Now private.
	* pt-misc.h (tree_parameter_list::validate): Provide decl.

2008-03-12  John W. Eaton  <jwe@octave.org>

	* variables.cc (Vwhos_line_format): Omit print_dims parameter.
	Fix doc string in Vwhos_line_format DEVAR.
	(symbol_record_name_compare): Delete unused function.
	(whos_parameter::dimensions): Delete struct field.
	(symbol_info_list): New class.
	(dimensions_string_req_first_space, make_dimensions_string,
	dimensions_string_req_total_space): Delete.
	(parse_whos_line_format): Move functionality to new
	symbol_info_list class.
	(print_symbol_info_line): Move functionality to new
	symbol_info_list::struct symbol_info::displaly_line method.
	(do_who): Simplify with new symbol_info_list class.
	Handle index expressions in addition to symbol names.

	* DLD-FUNCTIONS/dlmread.cc (Fdlmread): Fix separator detection.

2008-03-11  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/eig.cc (Feig): Handle possible error from EIG.
	* DLD-FUNCTIONS/qp.cc (qp, Fqp): Likewise.
	* xpow.cc (xpow): Likewise.

2008-03-11  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/dlmread.cc: Style fixes.

2008-03-11  David Bateman  <dbateman@free.fr>

	* DLD-FUNCTIONS/dlmread.cc: Function ported from octave forge. Add
	spreadsheet style ranges.
	* Makefile.in (DLD_XSRC): Add dlmread.cc.

2008-03-10  John W. Eaton  <jwe@octave.org>

	* mex.cc (mxCreateLogicalScalar): Argument is now mxLogical.

	* data.cc (Fisfloat): New function.
	* ov.h (octave_value::is_float_type): Ditto.
	* ov-base.h (octave_base_value::is_float_type): Ditto.
	* ov-complex.h (octave_complex): Ditto.
	* ov-cx-mat.h (octave_complex_matrix): Ditto.
	* ov-cx-sparse.h (octave_sparse_complex_matrix): Ditto.
	* ov-range.h (octave_range): Ditto.
	* ov-re-mat.h (octave_matrix): Ditto.
	* ov-re-sparse.h (octave_sparse_matrix): Ditto.
	* ov-scalar.h (octave_scalar): Ditto.

	* mxarray.h.in (mxLogical): Use unsigned char instead of int.
	From Antwerpen, G. (Gert) van <gert.vanantwerpen@tno.nl>.

2008-03-07  John W. Eaton  <jwe@octave.org>

	* ov-struct.cc (octave_struct::print_raw): Don't print contents fo
	struct arrays that have more than one element.

2008-03-06  David Bateman  <dbateman@free.fr>

	* DLD-FUNCTIONS/matrix_type.cc (Fmatrix_type): Document that
	the initial interpretation of a positive definite return from
	matrix_type is that the matrix is "probably" positive definite and
	not certainly so.

2008-03-06  John W. Eaton  <jwe@octave.org>

	* parse.y: Move tests here from test/test_eval.m.
	* DLD-FUNCTIONS/fft.cc: Move tests here from test/test_signal.m.
	* DLD-FUNCTIONS/dassl.cc: Move tests here from test/test_diffeq.m.
	* DLD-FUNCTIONS/lsode.cc: Move tests here from test/test_diffeq.m.
	* DLD-FUNCTIONS/quad.cc: Move tests here from test/test_quad.m.
	* DLD-FUNCTIONS/time.cc: Move tests here from test/test_system.m.

2008-03-06  Alexander Barth  <barth.alexander@gmail.com>

	* DLD-FUNCTIONS/__lin_interpn__.cc (lookup):
	Handle decreasing coordinate values.

2008-03-05  Jaroslav Hajek  <highegg@gmail.com>

	* DLD-FUNCTIONS/chol.cc (Fcholupdate): Adjust code to meet 
	Octave's coding guidelines.

	* DLD-FUNCTIONS/qr.cc (Fqrupdate, Fqrinsert, Fqrdelete): Adjust 
	code to meet Octave's coding guidelines.
	* DLD-FUNCTIONS/qr.cc (Fqrdelete): Fix incorrect test. 

	* DLD-FUNCTIONS/qr.cc (Fqrinsert, Fqrdelete): Modify to use
	0-based indexing in liboctave's QR classes.

2008-03-04  Jaroslav Hajek  <highegg@gmail.com>

	* DLD-FUNCTIONS/chol.cc (Fcholupdate): New function.

	* DLD-FUNCTIONS/qr.cc (Fqrupdate, Fqrinsert, Fqrdelete):
	New functions.

2008-03-04  Ryan Rusaw  <rrusaw@gmail.com>

	* toplev.h (octave_call_stack::element): New static function.
	(octave_call_stack::cs): Now std::deque instead of std::list.

	* pt-stmt.cc (curr_statement, curr_caller_statement): Delete.
	* pt-stmt.h: Delete decls.

	* pt-stmt.h, pt-stmt.cc (tree_statement_stack): New class.

	* pt-stmt.cc (tree_statement::eval): Use tree_statement_stack
	instead of curr_statement variable.

	* ov-usr-func.cc (octave_user_function::do_multi_index_op):
	Don't set curr_caller_statement.

	* error.cc (verror, pr_where): Call tree_statement_stack::current_line
	and tree_statement_stack::current_column instead of 
	instead of curr_statement->line and curr_statement->column.
	* input.cc (get_user_input): Likewise.

2008-02-27  John P. Swensen  <jpswensen@gmail.com>

	* debug.cc (get_user_function): Call symtab::find_function instead
	of symtab::find_user_function.

2008-02-27  John W. Eaton  <jwe@octave.org>

	* oct-stream.cc (do_read): Stop reading if seek fails.

2008-02-26  John W. Eaton  <jwe@octave.org>

	* ov-base-int.cc (octave_base_int_helper,
	octave_base_int_helper_traits): New templates and specializations.
	(octave_base_int_matrix<T>::convert_to_str_internal,
	octave_base_int_matrix<T>::convert_to_str_internal): Use them.

	* DLD-FUNCTIONS/rand.cc (do_rand): Pass name of calling function
	to octave_rand::state.

	* variables.cc (bind_ans): Handle cs-lists recursively.

	* ov-cs-list.h, ov-cs-list.cc (octave_cs_list::print,
	octave_cs_list::print_raw): Delete.

2008-02-25  John W. Eaton  <jwe@octave.org>

	* Cell.cc (Cell::map): New function.
	* Cell.h (Cell::map): Declare.
	(xisalnum, xisalpha, xisascii, xiscntrl, xisdigit,
	xisgraph, xislower, xisprint, xispunct, xisspace, xisupper,
	xisxdigit, xtoascii, xtolower, xtoupper): New mapper functions.
	(ctype_mapper): New private typedef.

	* ov-cell.h (xisalnum, xisalpha, xisascii, xiscntrl, xisdigit,
	xisgraph, xislower, xisprint, xispunct, xisspace, xisupper,
	xisxdigit, xtoascii, xtolower, xtoupper): New mapper functions.

2008-02-25  Michael Goffioul  <michael.goffioul@gmail.com>

	* ov-scalar.cc (octave_scalar::round): Use xround instead of ::round
	in mapper implementation.

2008-02-25  John W. Eaton  <jwe@octave.org>

	* ov-base.cc (UNDEFINED_MAPPER, STRING_MAPPER): New macros.  Use
	them to define mapper functions.

	* mappers.cc (Fisalnum, Fisalpha, Fisascii, Fiscntrl, Fisdigit,
	Fisgraph, Fislower, Fisprint, Fispunct, Fisspace, Fisupper,
	Fisxdigit, Ftoascii, Ftolower, Ftoupper):
	Use DEFUNX to define ctype mapper functions.
	Use new function names.

	* ov-base.h, ov-base.cc, ov-str-mat.h, ov-str-mat.cc:
	Prepend x to ctype mapper function names.

	* graphics.h.in (row_vector_property::row_vector_property):
	Set default constraints here.
	(row_vector_property::add_constraint): New function.
	(row_vector_property::validate): Delete decl.
	(axes::properties::init): Use single-arg add_constraint function
	for xlim, ylim, zlim, clim, and alim properties.

	* graphics.cc (row_vector_property::validate): Delete.

	* graphics.cc (axes::properties::update_camera,
	axes::properties::update_aspectratios):
	Avoid shadow warnings from GCC.

	* graphics.h.in (base_properties::get_bounding_box):
	Avoid unused argument warning from GCC.

	* graphics.h.in (array_property::add_constraint): Arg is now const
	reference, not value.

	* graphics.h.in, graphics.cc (class row_vector_property): New class.	
	(axes::properties): xlim, ylim, zlim, clim, alim, xtick, ytick,
	ztick properties are now row_vector_property objects instead of
	array_property objects.

	* genprops.awk: Special case row_vector_property in the same way
	as array_property.

2008-02-22  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/ccolamd.cc, DLD-FUNCTIONS/colamd.cc,
	DLD-FUNCTIONS/convhulln.cc, DLD-FUNCTIONS/dmperm.cc,
	DLD-FUNCTIONS/qz.cc, DLD-FUNCTIONS/regexp.cc,
	DLD-FUNCTIONS/symbfact.cc, zfstream.h, zfstream.cc:
	Use 0 instead of NULL.

2008-02-22  David Bateman  <dbateman@free.fr>

	* DLD-FUNCTIONS/chol.c (Fchol, Fcholinv, Fchol2iinv): Treat 
	sparse matrices. Add the "lower" and "vector" flags.
	* DLD-FUNCTIONS/inv.c (Finv): Treat sparse matrices.
	* DLD-FUNCTION/sparse.cc (static bool is_sparse (const
	octave_value&)): Remove and use arg.is_sparse_type () instead.
	(Fspcumprod, Fspcumsum, Fspprod, spsum, spsumsq, spdiag): Remove.
	* DLD-FUNCTIONS/splu.cc: Remove.
	* DLD-FUNCTIONS/lu.cc: Treat sparse matrices, Add "vector" flag.
	* DLD-FUNCTIONS/spchol.cc: Move to symbfact.cc
	* DLD-FUNCTIONS/symbfact.cc: Remove cholesky functions Fspchol,
	Fspcholinv, Fspchol2inv.
	* DLD-FUNCTIONS/luinc.cc: Modify for new sparse LU
	constructors. Ass the 'vector' flag.
	* DLD-FUNCTIONS/spparms.cc: Add the sum_tol flag.

	* Makifile.in (DLD_XSRC): Remove spchol.cc, splu.cc and add 
	symbfact.cc
	* data.cc (NATIVE_REDUCTION): Treat sparse matrices
	(DATA_REDUCTION): Ditto.
	(template <class T> static octave_value make_spdiag (const T&, 
	octave_idx_type)): New template function for sparse diag
	function. Instantiate it.
	(static octave_value make_diag (const octave_value&, 
	octave_idx_type): Use make_spdiag for sparse matrice.
	(Fatan2): Compatibility fixes for mixed full/sparse matrices.
	
2008-02-21  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/fsolve.cc (fsolve_user_jacobian):
	Check dimensions of user-supplied Jacobian matrix.
	(fsolve_user_function): Check for non-square systems.

2008-02-20  John W. Eaton  <jwe@octave.org>

	* data.cc (map_d_m, map_m_d, map_m_m, Fatan2, Ffmod):
	Handle N-d arrays.

	* ov-bool-mat.h (octave_bool_matrix (const Array2<bool>&)): Delete.

2008-02-20  David Bateman  <dbateman@free.fr>

	* DLD-FUNCTIONS/det.cc, DLD-FUNCTIONS/find.cc,
	* DLD-FUNCTIONS/minmax.cc, DLD-FUNCTIONS/qr.cc:
	Treat sparse matrices.
	
	* DLD-FUNCTIONS/sparse.cc (Fspmin, Fspmax, Fatan2): Remove functions.

	* DLD-FUNCTIONS/dmperm.cc: Rename from spqr.cc.
	(Fspqr): Delete function.

	* DLD-FUNCTIONS/spqr.cc, DLD-FUNCTIONS/spdet.cc,
	DLD-FUNCTIONS/spfind.cc: Remove.

	* Makefile.in (DLD_XSRC): Add dmperm.cc to the list.
	Delete det.cc, find.cc, minmax.cc, and qr.cc from the list.

	* Makefile.in (OV_SRC): Remove ov-mapper.cc.
	(OV_INCLUDES): Remove ov-mapper.h.
	(DEFUN_PATTERN): No longer accept DEFUN_MAPPER as valid.
	* ov-mapper.cc, ov-mapepr.h: Delete, remove all includes of
	ov-mapper.h from all files.

	* op-b-sbm.cc, op-bm-sbm.cc, op-smb-b.cc, op-sbm-bm.cc: Include 
	ov-bool-sparse.h.

	* defun-int.h (DEFUN_MAPPER_INTERNAL, install_builtin_mapper):
	Remove.
	* defun.cc (install_builtin_mapper): Ditto.
	* defun.h (DEFUN_MAPPER): Remove.

	* mappers.cc: Rewrite all mapper function using DEFUN and newly
	introduced octave_value mapper functions.
	(dummyp, xabs, xisalnum, xisascii, xiscntrl, xisdigit,
	xisgraph, xislower, xisprint, xispunct, xisspace, xisupper,
	xtoascii, xtolower, xtoupper, xconj, ximag, xreal): Remove
	static wrapper functions.

	* mkbuiltins (XDEFUN_MAPPER_INTERNAL, install_builtin_functions): 
	Remove.
	* mkgendoc (XDEFUN_MAPPER_INTERNAL): Remove.

	* ov.cc (octave_mapper::register_type): Remove.
	
	* ov.h (abs, acos, acosh, angle, arg, asin, asinh, atan, atanh,
	ceil, conj, cos, cosh, erf, erfc, exp, finite, fix, floor, gamma,
	imag, isinf, isna, isnan, lgamma, log, log10, real, round, signum,
	sin, sinh, sqrt, tan, tanh, isalnum, isalpha, isascii, iscntrl,
	isdigit, isgraph, islower, isprint, ispunct, isspace, isupper,
	isxdigit, toascii, tolower, toupper):
	New octave_value mapper	functions.

	* ov-base.h (abs, acos, acosh, angle, arg, asin, asinh, atan,
	atanh, ceil, conj, cos, cosh, erf, erfc, exp, finite, fix, floor,
	gamma, imag, isinf, isna, isnan, lgamma, log, log10, real, round,
	signum, sin, sinh, sqrt, tan, tanh, isalnum, isalpha, isascii,
	iscntrl, isdigit, isgraph, slower, isprint, ispunct, isspace,
	isupper, isxdigit, toascii, tolower, toupper):
	New virtual mapper functions.
	* ov-base.cc (abs, acos, acosh, angle, arg, asin, asinh, atan,
	atanh, ceil, conj, cos, cosh, erf, erfc, exp, finite, fix, floor,
	gamma, imag, isinf, isna, isnan, lgamma, log, log10, real, round,
	signum, sin, sinh, sqrt, tan, tanh, isalnum, isalpha, isascii,
	iscntrl, isdigit, isgraph, slower, isprint, ispunct, isspace,
	isupper, isxdigit, toascii, tolower, toupper):
	Base versions of mapper functions.

	* ov-bool-mat.h (abs, acos, acosh, angle, arg, asin, asinh, atan,
	atanh, ceil, conj, cos, cosh, erf, erfc, exp, finite, fix, floor,
	gamma, imag, isinf, isna, isnan, lgamma, log, log10, real, round,
	signum, sin, sinh, sqrt, tan, tanh):
	Mapper function recast boolen matrix as double.
	* ov-bool.h (abs, acos, acosh, angle, arg, asin, asinh, atan,
	atanh, ceil, conj, cos, cosh, erf, erfc, exp, finite, fix, floor,
	gamma, imag, isinf, isna, isnan, lgamma, log, log10, real, round,
	signum, sin, sinh, sqrt, tan, tanh): Ditto.
	* ov-bool-sparse.h (abs, acos, acosh, angle, arg, asin, asinh,
	atan, atanh, ceil, conj, cos, cosh, erf, erfc, exp, finite, fix,
	floor, gamma, imag, isinf, isna, isnan, lgamma, log, log10, real,
	round, signum, sin, sinh, sqrt, tan, tanh): Ditto.
	(char_array_value): New method to convert to charNDArray.
	* ov-bool-sparse.cc (char_array_value): New method to convert to
	charNDArray.
	* ov-complex.cc (char_array_value): ditto.
	(xabs, ximag, xreal): Static wrapper functions.
	(abs, acos, acosh, angle, arg, asin, asinh, atan, atanh, ceil,
	conj, cos, cosh, exp, finite, fix, floor, imag, isinf, isna,
	isnan, log, log10, real, round, signum, sin, sinh, sqrt, tan,
	tanh): New mapper methods.
	* ov-complex.h: Provide decls.

	* ov-cx-mat.cc (char_array_value): New method to convert to
	charNDArray.
	(xabs, ximag, xreal): Static wrapper functions.
	(abs, acos, acosh, angle, arg, asin, asinh, atan, atanh, ceil,
	conj, cos, cosh, exp, finite, fix, floor, imag, isinf, isna,
	isnan, log, log10, real, round, signum, sin, sinh, sqrt, tan,
	tanh): New mapper methods.
	* ov-cx-mat.h: Provide decls.
	
	* ov-cx-sparse.cc (char_array_value): New method to convert to
	charNDArray.
	(xabs, ximag, xreal): Static wrapper functions.
	(abs, acos, acosh, angle, arg, asin, asinh, atan, atanh, ceil,
	conj, cos, cosh, exp, finite, fix, floor, imag, isinf, isna,
	isnan, log, log10, real, round, signum, sin, sinh, sqrt, tan,
	tanh): New mapper methods.
	* ov-cx-sparse.h: Provide decls.

	* ov-intx.h (abs, signum, imag, ceil, conj, fix, floor, real,
	round, finite, isinf, isna, isnan): Define for both matrix and
	scalar classes.
	
	* ov-range.h (abs, acos, acosh, angle, arg, asin, asinh, atan,
	atanh, ceil, conj, cos, cosh, erf, erfc, exp, finite, fix, floor,
	gamma, imag, isinf, isna, isnan, lgamma, log, log10, real, round,
	signum, sin, sinh, sqrt, tan, tanh): New mapper functions.

	* ov-re-mat.cc (any_element_less_than): Static function to check if
	any elemet was less than a value,
	(any_element_greater_than): Ditto with greater than.
	(abs, acos, acosh, angle, arg, asin, asinh, atan, atanh, ceil,
	conj, cos, cosh, erf, erfc, exp, finite, fix, floor, gamma, imag,
	isinf, isna, isnan, lgamma, log, log10, real, round, signum, sin,
	sinh, sqrt, tan, tanh): New mapper functions.
	* ov-re-mat.h: Provide decls.
	
	* ov-scalar.cc (any_element_less_than): Static function to check if
	any elemet was less than a value,
	(any_element_greater_than): ditto with greater than.
	(abs acos acosh angle arg asin asinh atan atanh 
	ceil conj cos cosh erf erfc exp finite fix floor gamma imag 
	isinf isna isnan lgamma log log10 real round signum sin sinh 
	sqrt tan tanh): New mapper functions.
	* ov-scalar.h: Provide decls.

	* ov-str-mat.cc (xisalnum, xisascii, xiscntrl, xisdigit,
	xisgraph, xislower, xisprint, xispunct, xisspace, xisupper,
	xtoascii, xtolower, xtoupper): New static wrapper functions.
	(isalnum, isalpha, isascii, iscntrl, isdigit, isgraph, islower,
	isprint, ispunct, isspace, isupper, isxdigit, toascii, tolower,
	toupper): New mapper methods.
	* ov-str-mat.h: Provide decls.

2008-02-18  David Bateman  <dbateman@free.fr>

	* data.cc (Fatan2): Reject arguments that are integer types.

2008-02-19  Jaroslav Hajek  <highegg@gmail.com>

	* DLD-FUNCTIONS/qr.cc: Doc fix.

2008-02-18  John W. Eaton  <jwe@octave.org>

	* symtab.h
	(symbol_table::fcn_info::fcn_info_rep::clear_unlocked_functions):
	symbol_table::fcn_info::fcn_info_rep::clear_cmdline_function,
	symbol_table::fcn_info::fcn_info_rep::clear_autoload_function):
	New functions.
	(symbol_table::fcn_info::fcn_info_rep::clear_user_function):
	Don't clear locked functions.
	(symbol_table::fcn_info::fcn_info_rep::clear_mex_function):
	Call clear_user_function instead of setting function_on_path directly.
	(symbol_table::fcn_info::fcn_info_rep::clear):
	Use new functions to do the real work.

	* ov.h (octave_value::lock, octave_value::unlock,
	octave_value::islocked): New functions.
	* ov-base.cc (octave_base_value::lock, octave_base_value::unlock):
	New functions.
	* ov-base.h Provide decls.
	(octave_base_value::islocked): New function.
	* ov-fcn.h (octave_function::islocked): Now const.

2008-02-15  John W. Eaton  <jwe@octave.org>

	* ov-builtin.cc (octave_builtin::do_multi_index_op):
	Catch possible octave_execution_exception.
	* ov-mex-fcn.cc (octave_mex_function::do_multi_index_op): Likewise.
	* ov.cc (do_binary_op, do_cat_op, do_unary_op,
	octave_value::do_non_const_unary_op): Likewise.

2008-02-14  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/balance.cc, DLD-FUNCTIONS/qz.cc:
	Don't check f77_exception_encountered.

	* sighandlers.cc (user_abort): If interrupting immediately, set
	octave_interrupt_state if it is not already set.

	* pt-stmt.cc (tree_statement::eval): Catch execution exceptions.

	* octave.cc (lo_error_handler): New static function.
	(initialize_error_handlers): Set liboctave_error_handler to
	lo_error_handler, not error.

	* DLD-FUNCTIONS/urlwrite.cc (urlget): Call octave_rethrow_exception
	instead of octave_throw_interrupt_exception.
	* utils.cc (BEGIN_INTERRUPT_IMMEDIATELY_IN_FOREIGN_CODE_FOR_VSNPRINTF): 
	Likewise.

2008-02-12  David Bateman  <dbateman@free.fr>

	* graphics.h.in: Implement the cdatamapping property in patch and
	image objects.

2008-02-12  John W. Eaton  <jwe@octave.org>

	* pt-loop.cc (tree_simple_for_command::eval): Compute range
	element with multiplication.

2008-02-11  John W. Eaton  <jwe@octave.org>

	* Makefile.in ($(MAKEDEPS)): Skip dependencies if omit_deps is defined.

2008-02-09  John W. Eaton  <jwe@octave.org>

	* pr-output.cc (set_range_format): Eliminate sign arg.
	Change all callers.

2008-02-08  John W. Eaton  <jwe@octave.org>

	* ov-struct.cc (octave_struct::subsref): Allow Cell::index to resize.

	* input.cc (interactive_input): Eliminate debug arg.  Change All uses.
	(get_user_input): Don't process input_buf if there is an error.
	Call reset_error_handler instead of setting error_state to 0.

2008-02-08  Michael Goffioul  <michael.goffioul@gmail.com>

	* graphics.h.in (callback_property::execute): New static
	helper method (useful to execute callbacks by name).
	* graphics.cc (callback_property::execute): Likewise.
	(execute_callback): Avoid undefined argument when executing
	callback. Do not use arguments when the callback is a string.

2008-02-07  John W. Eaton  <jwe@octave.org>

	* ov-range.h (octave_range::sort): New functions.

2008-02-07  David Bateman  <dbateman@free.fr>

	* Makefile.in (DLD_XSRC): Delete spkron.cc.
	* DLD-FUNCTIONS/spkron.cc: Delete.
	* DLD-FUNCTIONS/kron.cc: Include here and dispatch to the sparse
	version if either argument is sparse.

2008-02-06  John W. Eaton  <jwe@octave.org>

	* Makefile.in (parse.cc): Also use --defines option for bison.

2008-02-06  Michael Goffioul  <michael.goffioul@gmail.com>

	* graphics.cc (axes::properties::set_defaults): Set default axes
	color to white.

2008-02-06  John W. Eaton  <jwe@octave.org>

	* Makefile.in (parse.cc): Use "-o $@" instead of renaming y.tab.c.
	(maintainer-clean): Don't remove y.tab.c.

	* Makefile.in (parse.cc : parse.y): Use mv instead of move-if-change.
	(stamp-prereq, stamp-liboctave-prereq): Eliminate.
	(clean): Don't remove stamp-prereq and stamp-liboctave-prereq.
	(OPT_BASE, OPT_IN, OPT_INC): New macros.
	(OPT_HANDLERS): Define in terms of OPT_BASE.
	($(OPT_INC) : %.h : %.in): New pattern rule.
	(PREREQ): New macro.
	($(MAKEDEPS)): Simplify with $(PREREQ).

2008-02-05  John W. Eaton  <jwe@octave.org>

	* Makefile.in: Unconditionally include $(MAKEDEPS).
	Mark $(MAKEDEPS) as .PHONY targets if omit_deps is true.

2008-02-05  Michael Goffioul  <michael.goffioul@gmail.com>

	* graphics.h.in (base_properties::get_boundingbox): New method.
	(figure::properties::get_boundingbox): Overload method.
	(axes::properties::get_boundingbox): Likewise.
	* graphics.cc (figure::properties::get_boundingbox): Return a
	left-to-right/top-to-bottom bounding box rectangle.
	(axes::properties::get_boundingbox): Likewise.
	(axes::properties::update_camera): Remove backend access and
	use the new axes bounding box.

2008-02-04  Shai Ayal  <shaiay@users.sourceforge.net>

	* graphics.h.in (axes::properties::update_xlim,
	axes::properties::update_ylim, axes::properties::update_zlim):
	New update methods.

	* graphics.cc (axes::properties::calc_ticks): New function.
	(axes::properties::magform): New function.
	(axes::update_axis_limits): Call update_{x,y,z}lims if
	appropriate.

2008-02-04  Michael Goffioul  <michael.goffioul@gmail.com>

	* graphics.h.in (base_graphics_backend::get_screen_size,
	graphics_backend::get_screen_size): New methods.
	(graphics_backend::available_backends): Export symbol.
	(class figure::properties, class axes::properties): Export classes.
	(figure::properties::get_boundingbox): New utility method.
	(figure::properties::position): Use valid default position.
	(axes::properties::gridlinestyle,
	axes::properties::minorgridlinestyle):
	Use dotted line style as default. 
	* graphics.cc (default_figure_position): New static function.
	(gnuplot_backend::get_screen_size): New method.
	(figure::properties::get_boundingbox): New utility method.

2008-02-02  Shai Ayal  <shaiay@users.sourceforge.org>

        * graphics.h.in (base_scaler::~base_scalar): New virtual destructor.
        * graphics.cc (axes::properties::update_camera): Tag abs with std.

	* graphics.h.in (graphics_backend::find_backend): New function.
	(class figure): Add __backend__ property and set method.
	
2008-02-02  John W. Eaton  <jwe@octave.org>

	* help.cc (do_type): Don't print dyamically loaded function files.

	* ov-fcn.h (octave_function::is_dynamically_loaded_function): Delete.

2008-02-01  John W. Eaton  <jwe@octave.org>

	* symtab.cc (Fset_variable, Fvariable_value):
	New functions, but commented out for now.

2008-01-30  Michael Goffioul  <michael.goffioul@gmail.com>

	* graphics.h.in (axes::properties::get_transform_matrix,
	axes::properties::get_inverse_transform_matrix,
	axes::properties::get_opengl_matrix_1,
	axes::properties::get_opengl_matrix_2,
	axes::properties::get_transform_zlim): New accessors.
	(base_properties::is_clipping): New accessor.
	(class graphics_xform): New class encapsulating axes transformation.
	(axes::properties::get_transform): New method returning a
	graphics_xform object.
	* graphics.cc (class graphics_xform): New class.

2008-01-31  David Bateman  <dbateman@free.fr>

	* ov.cc (octave_value::octave_value (const ArrayN<bool>&),
	octave_value::octave_value (const Sparse<bool>&, const MatrixType &),
	octave_value::octave_value (const ArrayN<std::streamoff>&)): New 
	constructors.
	* ov.h: (octave_value (const ArrayN<bool>&),
	octave_value (const Sparse<bool>&, const MatrixType &),
	octave_value (const ArrayN<std::streamoff>&)): Declare them.
	(octave_value sort (octave_idx_type, sortmode) const, octave_value
	sort (Array<octave_idx_type> &, octave_idx_type, sortmode) const):
	octave_value sort method.
	
	
	* ov-base.cc (sort): Base versions of teh octave_value sort methods.
	* ov-base.h (sort): Declare the octave_value sort methods
	* ov-base-scalar.h (sort): Simple sort methods for scalars.
	* ov-base-mat.h, ov-base-sparse.h (sort): Sort methods calling
	underlying array or sparse sort methods.
	* ov-str-mat.h (sort): String specific sort methods.
	
	* TEMPLATE-INST/Array-tc.cc: Instantiate the array sort methods.
	* ov-streamoff.h (sort): Sort versions returning and error.
	* oct-stream.cc, ov-typeinfo.cc, Array-os.cc: Null instantiation
	of array sort methods.
	
	* Makefile.in (DLD_XSRC): Remove sort.cc
	* DLD-FUNCTIONS/sort.cc: Remove
	* data.cc (Fsort): New function using octave_value sort methods
	for the sorting. Add tests.

2008-01-30  Thomas Weber  <thomas.weber.mail@gmail.com>

	* pager.cc (Fmore): Doc fix.

2008-01-28  Michael Goffioul  <michael.goffioul@gmail.com>

	* genprops.awk: Add update ('u') modifier and document the
	readonly ('r') modifier.
	* graphics.h.in (class base_scaler, class lin_scaler, class
	log_scaler, class scaler): New classes to make abstraction of the
	axis scale.
	(base_graphics_backend::get_screen_resolution,
	graphics_backend::get_screen_resolution): New methods.
	(axes::properties::sx, axes::properties::sy,
	axes::properties::sz): New scaler fields.
	(axes::properties::get_x_scaler, axes::properties::get_y_scaler,
	axes::properties::get_z_scaler): New accessors.
	(axes::properties::x_render, axes::properties::x_render_inv,
	axes::properties::x_gl_mat1, axes::properties::x_gl_mat2,
	axes::properties::x_zlim): New utility Matrix fields.
	(axes::properties::get_boundingbox,
	axes::properties::update_camera,
	axes::properites::update_aspectratios,
	axes::properties::update_transform,
	axes::properties::update_xscale, axes::properties::update_yscale,
	axes::properties::update_zscale, axes::properties::update_view,
	axes::properties::update_xdir, axes::properties::update_ydir,
	axes::properties::update_zdir): New updater methods.
	(axes::properties::init): Initialize sx, sy, sz and x_zlim correctly.
	(axes::properties::position): Use valid default position value.
	(axes::properties::xscale, axes::properties::yscale,
	axes::properties::zscale, axes::properties::xdir,
	axes::properties::ydir, axes::properties::zdir,
	axes::properties::view): Add updater ('u') property modifier.
	* graphics.cc (default_axes_position, default_axes_outerposition):
	New initializers.
	(convert_position): New utility function to convert position
	according to specified units.
	(gnuplot_backend::get_screen_resolution): New method.
	(axes::properties::set_defaults): Initilize recently added properties.
	(xform_matrix, xform_vector, transform, xform_scale,
	xform_translate, scale, translate, xform, normalize, dot, cross,
	unit_cube, cam2xform, xform2cam): New inline transformation
	utility functions.
	(axes::properties::update_camera,
	axes::properties::update_aspectratios,
	axes::properties::get_boundingbox): New updater methods for
	computing transformation matrices.
	(axes::update_axis_limits): Update transformation data.

2008-01-28  John W. Eaton  <jwe@octave.org>

	* oct-stream.cc (BEGIN_CHAR_CLASS_CONVERSION): Handle width properly.
	(OCTAVE_SCAN) [__GNUG__ && ! CXX_ISO_COMPLIANT_LIBRARY]:
	Delete special case.

2008-01-25  David Bateman  <dbateman@free.fr>

	* DLD-FUNCTIONS/rand.cc (Frandp): Relax relative error on randp
	tests.

2008-01-25  Michael Goffioul  <michael.goffioul@gmail.com>

	* graphics.h.in (base_properties::get_backend,
	base_graphics_object::get_backend, graphics_backend::get_backend):
	New methods for convenient access to the backend.
	(figure::get_backend, figure::set_backend): Remove.
	* graphics.cc (base_properties::get_backend): New method.
	(Fdrawnow): Use graphics_object::get_backend to simplify code.

2008-01-24  Pascal Dupuis  <Pascal.Dupuis@worldonline.be>

	* mxarray.h.in: Include <cstring>.

2008-01-22  John W. Eaton  <jwe@octave.org>

	* graphics.cc (clear_drawnow_request): New function.
	(Fdrawnow): Add it to the unwind_protect stack.

	* input.cc (Vdrawnow_requested): No longer static.
	* input.h: Provide decl.
	* graphics.cc (Fdrawnow, Fset, make_graphics_object):
	Use Vdrawnow_requested directly.

	* toplev.cc (octave_add_atexit_function,
	octave_remove_atexit_function): New functions.
	(Fatexit): Use them.
	* graphics.cc (Fdrawnow): Call octave_add_atexit_function instead
	of using eval.
	* toplev.h (octave_add_atexit_function,
	octave_remove_atexit_function): Provide decls.

2008-01-22  Michael Goffioul  <michael.goffioul@gmail.com>

	* graphics.h.in (base_properties::is_visible,
	base_properties::set_modified): New accessors.
	(class base_graphics_backend, class graphics_backend): New classes
	for handling octave/backend interaction.
	(figure::properties::close): Add "pop" argument controlling
	whether the figure should be popped from the list of existing figures.
	(class figure::properties): New backend field and accessors, holding
	the graphics backend associated with the figure.
	* graphics.cc (class gnuplot_backend): New class for the default
	gnuplot backend.
	(figure::properties::close): Add "pop" argument and transfer the
	figure closing to the associated backend.
	(Fdrawnow): New builtin function, converted from drawnow.m.

2008-01-19  John W. Eaton  <jwe@octave.org>

	* genprops.awk (emit_source): Use "pname" for property name argument.

2008-01-19  Shai Ayal  <shaiay@users.sourceforge.net>

	* graphics.h.in (class figure::properties): New properties:
	alphamap, currentcharacter, currentobject, current_point,
	dockcontrols, doublebuffer, filename, integerhandle,
	inverthardcopy, keypressfcn, keyreleasefcn, menubar, mincolormap,
	name, numbertitle, paperunits, paperposition, paperpositionmode,
	papersize, papertype, pointer, pointershapecdata,
	pointershapehotspot, position, renderer, renderermode, resize,
	resizefcn, selectiontype, toolbar, units, windowbuttondownfcn,
	windowbuttonmotionfcn, windowbuttonupfcn, windowbuttonwheelfcn,
	windowstyle, wvisual, wvisualmode, xdisplay, xvisual, xvisualmode,
	buttondownfcn.

2008-01-19  Michael Goffioul  <michael.goffioul@gmail.com>

	* graphics.h.in (base_properties::beingdeleted,
	base_properties::is_beingdeleted,
	base_properties::get_beingdeleted,
	base_properties::set_beingdeleted):
	Add beingdeleted property and accessors.
	* graphics.cc (base_properties::get): Likewise.
	(gh_manager::free): Set beingdeleted to "on" when freeing an object.
	* genprops.awk: Skip C++-style comments in property declaration.

	* graphics.h.in (class axes::properties): New properties: alim,
	alimmode, xminortick, yminortick, zminortick, ambientlightcolor,
	cameraposition, cameratarget, cameraupvector, cameraviewangle,
	camerapositionmode, cameratargetmode, cameraupvectormode,
	cameraviewanglemode, currentpoint, drawmode, fontangle, fontname,
	fontsize, fontweight, fontunits, gridlinestyle,
	minorgridlinestyle, linestyleorder, linewidth, plotboxaspectratio,
	plotboxaspectratiomode, projection, tickdir, tickdirmode,
	ticklength, tightinset, units, x_viewtransform,
	x_projectiontransform, x_viewporttransform, x_normrendertransform,
	x_rendertransform.
	* genprops.awk: Handle new readonly modifier 'r'.

2008-01-18  Juhani Saastamoinen  <juhani@cs.joensuu.fi>

	* genprops.awk: Avoid using "default" as a variable name.

2008-01-17  Michael Goffioul  <michael.goffioul@gmail.com>

	* graphics.h.in (data_property::get_data_limits): Reset min/max values.

2008-01-17  David Bateman  <dbateman@free.fr>

	* ov-fcn-inline.cc: Add tests for Finline.

2008-01-17  Ben Abbott  <bpabbott@mac.com>

	* load-path.cc (F__pathorig__): Rename from Fpathdef.
	(Frestoredefaultpath): New function.

2008-01-16  John W. Eaton  <jwe@octave.org>

	* pt-assign.cc (tree_simple_assignment::rvalue,
	tree_multi_assignment::rvalue): Handle assignment of
	comma-separated lists.

2008-01-15  John W. Eaton  <jwe@octave.org>

	* mappers.cc (install_mapper_functions): Move test for asin here
	from test/test_arith.m.

	* graphics.h.in (class line::properties, class text::properties):
	Rename erase_mode property to erasemode.

2008-01-15  Michael Goffioul  <michael.goffioul@gmail.com>

	* graphics.h.in (base_properties::remove_child,
	base_properties::adopt, base_properties::update_axis_limits):
	Make virtual.
	(base_graphics_object::mark_modified,
	base_graphics_object::override-defaults,
	base_graphics_object::set_from_list, base_graphics_object::set,
	base_graphics_object::get, base_graphics_object::get_parent,
	base_graphics_object::remove_child, base_graphics_object::adopt,
	base_graphics_object::reparent, base_graphics_object::defaults,
	base_graphics_object::type): Add default implementation.
	(class root_figure, class figure, class axes, class line,
	class text, class image, class patch, class surface):
	Remove overloaded virtual methods whose implementation is
	identical to the default one.

	* genprops.awk: Handle 'h' modifier for hidden properties.
	Replace "get(void)" method with "get(bool all = false)" to allow
	access to hidden properties.

	* graphics.h.in (base_properties::get, base_graphics_object::get,
	graphics_object::get, root_figure::get, figure::get, axes::get,
	line::get, text::get, patch::get, surface::get, image::get):
	New arg, ALL, to access hidden properties.
	(gh_manager::do_get_object): Do not look for invalid handles.
	(figure::properties::__plot_stream__,
	figure::properties::__enhanced__, axes::properties::__colorbar__):
	Make properties hidden.
	(axes::properties, line::properties, text::properties,
	patch::properties, surface::properties): Remove obsolete 'a'
	property modifier.
	* graphics.cc (base_properties::get): New arg ALL, to access
	hidden properties.
	(base_properties::mark_modified): Call mark_modified only on valid
	parent object.
	(__get__): New internal function returning all properties,
	including the hidden ones.

2008-01-15  John W. Eaton  <jwe@octave.org>

	* graphics.cc (properties::set_currentfigure,
	properties::set_currentaxes): Don't convert arg to double
	before passing to graphics_handle constructor.
	(base_properties::remove_child): Call mark_modified if list of
	children changes.

2008-01-15  Shai Ayal  <shaiay@users.sourceforge.net>

	* graphics.h.in (class line::properties): New properties:
	displayname, erasemode.
	(class text::properties) New properties: backgroundcolor,
	displayname, edgecolor, erase_mode, editing, fontunits, linestyle,
	linewidth, margin, verticalalignment.

2008-01-14  John W. Eaton  <jwe@octave.org>

	* load-path.cc (load_path::do_initialize): Start with sys_path empty.
	(maybe_add_path_elts): Omit path_sep_str if path is empty.

	* symtab.h (symbol_table::do_pop_context): Remove symbol_records
	which have no more context.
	(symbol_table::symbol_record::pop_context,
	(symbol_table::symbol_record::symbol_record_rep::pop_context):
	Return size of value_stack, or 1 if variable is persistent or global.

2008-01-14  Kai Habel  <kai.habel@gmx.de>

	* graphics.h.in (class patch::properties): New properties:
	cdatamapping, facevertexalphadata, facevertexcdata, vertexnormals,
	normalmode, facelighting, edgealpha, edgelighting,
	backfacelighting, ambientstrength, diffusestrength,
	specularstrength, specularexponent, specularcolorreflectance,
	erasemode.

2008-01-14  Michael Goffioul  <michael.goffioul@gmail.com>

	* graphics.h.in (gh_manager::do_make_graphics_handle,
	gh_manager::make_graphics_handle): New boolean parameter
	do_createfcn.
	* graphics.cc (xcreatefcn): New static function.
	(gh_manager::do_make_graphics_handle): New boolean parameter
	do_createfcn.  Execute createfcn callback accordingly.
	(make_graphics_object, F__go_figure__): Use do_createfcn
	parameter and call xcreatefcn

	* genprops.awk (emit_get_callback): Pass user data to execute method.
	* graphics.cc (execute_callback): New static function.
	(callback_property::validate): Make it work.
	(callback_property::execute): Make it work.
	(gh_manager::do_free): Execute delete function here.
	* graphics.h.in (callback_property::execute): Fix decl.
	(base_properties::buttondownfcn, base_properties::createfcn,
	base_properties::deletefcn, base_properties::userdata): Default
	value is empty Matrix, not undefined octave_value object.
	(base_properties::execute_createfcn): New function.
	(base_properties::execute_deletefcn): New function.

2007-12-13  Shai Ayal  <shaiay@users.sourceforge.net>

	* graphics.h.in, graphics.cc (class base_properties): New
	properties: busyaction, buttondownfcn, clipping, createfcn,
	deletefcn, handlevisibility, hittest, interruptible, selected,
	selectionhighlight, uicontextmenu, userdata, visible.
	(class figure): Delete visible property.

2008-01-13  Michael Goffioul  <michael.goffioul@gmail.com>

	* grahpics.h.in: Sprinkle with OCTINTERP_API as needed.
	(axes::properties::xaxislocation): Allow value of zero.
	(axes::properties::yaxislocation): Likewise.

2008-01-12  Michael Goffioul  <michael.goffioul@gmail.com>

	* Makefile.in (distclean): Delete graphics-props.cc.

	* genprops.awk: Handle new graphics property classes.

	* graphics.h.in, graphics.cc: Adapt for new specific property types.

2008-01-12  John W. Eaton  <jwe@octave.org>

	* toplev.cc (octave_config_info): Add GNUPLOT to the struct.
	* oct-conf.h.in (OCTAVE_CONF_GNUPLOT): New macro.

2008-01-11  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/__qp__.cc (qp): Undo part of change from
	2007-09-04 (compute Y at each iteration).

2008-01-10  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/fsolve.cc (Ffsolve): Doc fix.
	(hybrd_info_to_fsolve_info): Swap return values for -1 and -2 inputs.

	* DLD_FUNCTIONS/__gnuplot_raw__.l: Delete.
	* Makefile.in (DLD_XSRC): Remove it from the list.
	(BUILT_DISTFILES): Remove __gnuplot_raw__.cc from the list.
	(stamp-prereq): Don't depend on __gnuplot_raw__.cc.
	(maintainer-clean): Don't remove __gnuplot_raw__.cc.
	(__gnuplot_raw__.cc): Delete target.
	(__gnuplot_raw__.o, pic/__gnuplot_raw__.o): Delete targets.

	* mex.cc (calc_single_subscript_internal): New static function.
	(mxArray_octave_value::calc_single_subscript): Use it.
	(mxArray_matlab::calc_single_subscript): Use it.

2008-01-07  John W. Eaton  <jwe@octave.org>

	* src/pt-except.cc (tree_try_catch_command::eval):
	Set Vdebug_on_error and Vdebug_on_warning to false while executing
	try block.

	* error.cc (Vdebug_on_error, Vdebug_on_warning): No longer static.
	* error.h: Provide decls.

	* gripes.cc (gripe_wrong_type_arg (const std::string&, const
	octave_value&, bool)): New function.
	* gripes.h: Provide decl.

	* oct-stream.cc (printf_value_cache::printf_value_cache):
	Reject structs, cells, objects, and lists.
	(octave_base_stream::do_printf): Quit early if printf_value_cache
	constructor fails.

	* parse.y (make_anon_fcn_handle): Don't attempt to convert
	expression to assignment to __retval__.

2008-01-07  Michael Goffioul  <michael.goffioul@gmail.com>

	* ov-builtin.h (octave_builtin): Tag with OCTINTERP_API.
	* ov-dld-fcn.h (octave_dld_function): Likewise.

	* ov-dld-fcn.cc (octave_dld_function::create): New function.
	* ov-dld-fcn.h: Provide decl.
	* defun.h (DEFINE_FUNX_INSTALLER_FUN3): Call it instead of
	creating a new octave_dld_function object directly.

2008-01-04  John Swensen  <jpswensen@comcast.net>

	* debug.cc (bp_table::do_remove_all_breakpoints_in_file):
	Avoid calling erase on invalid bp_map iterators.
	(bp_table::do_remove_breakpoint): Only try to delete breakpoints
	if some exist.  Avoid calling erase on invalid bp_map iterators.
	(parse_dbfunction_params): Return early if ARGS is empty.
	New arg, WHO.  Change all uses.
	Accept but do nothing with struct args.

2008-01-04  Thomas Weber  <thomas.weber.mail@gmail.com>

	* variables.cc (Fclear): Doc fix.

2008-01-04  John W. Eaton  <jwe@octave.org>

	* symtab.h (symbol_table:pop_scope): Avoid accessing beyond end of
	scope_stack.

2008-01-04  David Bateman  <dbateman@free.fr>

	* ov-mapper.cc (SPARSE_MAPPER_LOOP_2): Use data method instead of
	elem in the case where F(0) is non-zero.

2007-12-28  John W. Eaton  <jwe@octave.org>

	Merge changes from object branch:

	2007-12-12  John W. Eaton  <jwe@octave.org>

	* load-path.cc (load_path::dir_info::get_file_list,
	load_path::move_method_map, load_path::remove_method_map,
	load_path::do_find_fcn, load_path::do_find_private_fcn,
	load_path::do_find_method, load_path::do_find_file,
	load_path::do_find_first_of, load_path::do_find_all_first_of,
	load_path::do_display, load_path::add_to_method_map, genpath,
	execute_pkg_add_or_del): Use file_ops::concat.

	2007-08-24  John W. Eaton  <jwe@octave.org>

	* variables.cc (Fmislocked): return value for nargin == 0 case.

	2007-07-27  John W. Eaton  <jwe@octave.org>

	* pt-idx.cc (tree_index_expression::rvalue):
	Pass substring of type to subsref when doing partial evaluation.
	(tree_index_expression::lvalue): Likewise.

	2007-06-29  John W. Eaton  <jwe@octave.org>

	* ov-class.cc (is_built_in_class, set_class_relationship,
	Fsuperiorto, Finferiorto): New functions.
	(octave_class::in_class_method): New function.
	(octave_class::subsref, octave_class:subsasgn): Use it.

	2007-06-28  John W. Eaton  <jwe@octave.org>

	* src/ov-class.cc (sanitize): New function.
	(make_idx_args): Use it.

	2007-06-28  John W. Eaton  <jwe@octave.org>

	* src/ov-class.cc (octave_class::subsasgn): Expect and use only
	one return value from feval of subsasgn method.

	* ov.cc (octave_value::assign): Don't convert to struct if
	indexing a class object with ".".

	* pt-idx.cc (tree_index_expression::make_arg_struct):
	Use Cell instead of octave_value_list for subscripts.

	* ov-class.cc (make_idx_args): For "." index, subs field is
	string, not cell.

	2007-06-27  John W. Eaton  <jwe@octave.org>

	* ov-fcn-handle.cc (octave_fcn_handle::save_ascii,
	octave_fcn_handle::load_ascii, octave_fcn_handle::save_binary,
	octave_fcn_handle::load_binary, octave_fcn_handle::save_hdf5,
	octave_fcn_handle::load_hdf5): Adapt to new symbol table objects.

	* symtab.h (symbol_table::all_variables,
	symbol_table::do_all_variables): New arg, defined_only.

	2007-06-26  John W. Eaton  <jwe@octave.org>

	* ls-mat5.cc (read_mat5_binary_element): Adapt to new symbol table
	objects.

	* variables.cc (Vwhos_line_format): New static variable.
	(Fwhos_line_format): New function.
	(symbol_record_name_compare): Delete function.
	(symbol_record_name_compare, whos_parameter): New structs.
	(print_descriptor, dimensions_string_req_first_space,
	dimensions_string_req_total_space, make_dimensions_string,
	parse_whos_line_format, print_symbol_info_line): New static
	functions, adapted from old symbol table class.
	(do_who): Adapt to new symbol table objects.

	* symtab.h (symbol_table::glob_variables): New functions.
	(symbol_table::do_glob): New argument, vars_only.  Change all uses.

	2007-06-22  John W. Eaton  <jwe@octave.org>

	* symtab.cc (symbol_table::fcn_info::fcn_info_rep::help_for_dispatch):
	New function.
	* symtab.h: Provide decl.
	(symbol_table::fcn_info::help_for_dispatch,
	symbol_table::help_for_dispatch): New functions.

	* help.cc (help_from_symbol_table): Call
	symbol_table::help_for_dispatch instead of
	extract_help_from_dispatch.

	* help.cc (extract_help_from_dispatch): Delete.
	* help.h: Delete decl.

	2007-06-21  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/dispatch.cc (class octave_dispatch): Delete class.
	(builtin, any_arg_is_magic_colon, dispatch_record): Delete functions.
	(Fbuiltin, Fdispatch): Adapt to new symbol table objects.

	* symtab.cc (symbol_table::fcn_info::fcn_info_rep::print_dispatch):
	New function.
	* symtab.h: Provide decl.
	(symbol_table::fcn_info::print_dispatch, symbol_table::print_dispatch):
	New functions.

	* symtab.h (symbol_table::fcn_info::fcn_info_rep::clear_dispatch,
	symbol_table::fcn_info::clear_dispatch, symbol_table::clear_dispatch):
	New functions.

	* symtab.h (symbol_table::fcn_info::fcn_info_rep::get_dispatch,
	symbol_table::fcn_info::get_dispatch, symbol_table::get_dispatch):
	New functions.

	* symtab.cc (symbol_table::fcn_info::fcn_info_rep::find):
	Use leftmost class argument as dispatch type, or first argument if
	there are no class arguments.

	2007-06-20  John W. Eaton  <jwe@octave.org>

	* ov-base.cc (Vprint_answer_id_name, Fprint_answer_id_name): Delete.
	(octave_base_value::print_with_name): Always print name.

	* ov-base.cc (Vsilent_functions): No longer static.
	* ov-base.h: Provide decl.

	* ov-class.cc (Fmethods): Define as command.

	* ov-class.cc (octave_class::print): Simply call print_raw.

	* ov-class.cc (octave_class::print_with_name): New function.
	* ov-class.h: Provide decl.

	2007-06-19  John W. Eaton  <jwe@octave.org>

	* ov.cc (octave_value::do_unary_op, octave_value::do_binary_op):
	Handle class methods.

	* ov.cc (octave_value::unary_op_fcn_name,
	octave_value::binary_op_fcn_name): New functions.
	* ov.h: Provide decls.

	* ov-typeinfo.cc (octave_value_typeinfo::register_unary_class_op,
	octave_value_typeinfo::register_binary_class_op,
	octave_value_typeinfo::do_register_unary_class_op,
	octave_value_typeinfo::do_register_binary_class_op,
	octave_value_typeinfo::do_lookup_unary_class_op,
	octave_value_typeinfo::do_lookup_binary_class_op): New functions.
	* ov-typeinfo.h: Provide decls.
	
	* ov-typeinfo.h (octave_value_typeinfo::unary_class_op_fcn,
	octave_value_typeinfo::binary_class_op_fcn): New typedefs.
	(octave_value_typeinfo::unary_class_ops,
	octave_value_typeinfo::binary_class_ops): New data members.
	(octave_value_typeinfo::octave_value_typeinfo): Initialize them.
	(octave_value_typeinfo::lookup_unary_class_op,
	octave_value_typeinfo::lookup_binary_class_op): New functions.

	* OPERATORS/op-class.cc: New file.
	* Makefile.in (OP_XSRC): Add it to the list.

	* ov.cc (install_types): Call octave_class::register_type here.

	* ov-class.h (octave_class): Implement by containing a map object
	instead of deriving from octave_struct.
	* ov-class.cc (octave_class::subsref, octave_class::subsasgn):
	Don't use octave_class implementation as default action.

	* ov.h, ov.cc (octave_value::octave_value (const Octave_map&, const
	std::string&)): New constructor.

	2007-06-18  John W. Eaton  <jwe@octave.org>

	* ov-class.cc (octave_class::subsref, octave_class::subsasgn):
	Handle dispatch to user-defined methods.

	* parse.y (load_fcn_from_file): New arg, dispatch_type.
	Change all uses.

	* ov-fcn.h (octave_function::is_class_constructor,
	octave_function::is_class_method, octave_function::dispatch_class):
	New virtual functions.

	* ov-usr-fcn.h (octave_user_function::dispatch_class,
	octave_usr_function::stash_dispatch_class): New functions.
	(octave_usr_fucntion::xdispatch_class): New data member.
	* parse.y (frob_function): Call stash_dispatch_class here.

	* ov-struct.cc (Ffieldnames): Also handle objects.

	* ov-class.cc (Fmethods): New function.
	* load-path.cc (load_path::do_methods): New function.
	* load-path.h (load_path::methods): New function.

	* ov.h (octave_value::is_object): New function.
	* ov-base.h (octave_base_value::is_object): New virtual function.
	* ov-class.h (octave_class::is_object): New function.
	* ov-class.cc (Fisobject): New function.

	2007-06-15  John W. Eaton  <jwe@octave.org>

	* ov-class.cc (octave_class::print): Call display method if found.

	* symtab.h (symbol_table::fcn_info::find_method): New function.
	(symbol_table::find_method): New function.
	(symbol_table::fcn_info::fcn_info_rep::find_method): Provide decl.
	* symtab.cc (symbol_table::fcn_info::fcn_info_rep::find_method):
	New function.
	(symbol_table::fcn_info::fcn_info_rep::find): Use it.

	2007-06-14  John W. Eaton  <jwe@octave.org>

	* symtab.h (symbol_table::clear_mex_functions): Make it work.
	(symbol_table::fcn_info::fcn_info_rep::clear_mex_function,
	symbol_table::fcn_info::clear_mex_function): New functions.

	2007-06-08  John W. Eaton  <jwe@octave.org>

	* defun.cc (Falias): Delete.

	* variables.cc (load_fcn_try_ctor): New function.
	(lookup): call load_fcn_try_ctor instead of load_fcn_from_file.

	* variables.cc, variables.h (at_top_level, lookup_by_name, lookup,
	initialize_symbol_tables, fcn_out_of_date, symbol_out_of_date,
	lookup_function, lookup_user_function, link_to_global_variable,
	link_to_builtin_or_function, force_link_to_function, Fdocument,
	is_local_variable, do_clear_all, do_clear_functions,
	do_clear_globals, do_clear_variables, do_clear_function,
	do_clear_global, do_clear_variable, do_clear_symbol,
	do_clear_function_pattern, do_clear_global_pattern,
	do_clear_variable_pattern, do_clear_symbol_pattern,
	clear_function, clear_variable, clear_symbol):
	Delete (some functionality moved to the new symtab.cc and some is
	no longer needed).
	(Fignore_function_time_stamp): Move to symtab.cc.

	* lex.l (lookup_identifier): Delete.

	* parse.y (is_function_file): New function.
	(load_fcn_from_file): Return pointer to octave_function, not bool.

	* ov-usr-fcn.h, ov-usr-fcn.cc (octave_user_function::argn_sr,
	octave_user_function::nargin_sr, octave_user_function::nargout_sr,
	octave_user_function::varargin_sr): Delete data members.
	(octave_user_function::install_automatic_vars): Delete.
	(octave_user_script::do_multi_index_op): New function.

	* ov-fcn.h (octave_function::locked, octave_function::my_dir_name):
	New data members.
	(octave_function::dir_name, octave_function::stash_dir_name,
	octave_function::lock, octave_function::unlock,
	octave_function::islocked): New functions.

	* of-fcn-handle.h, ov-fcn-handle.cc (octave_fcn_handle::subsref):
	Call out_of_date_check here to simplify time stamp checking.
	(octave_fcn_handle::reload_warning): Delete.

	* ov.h (octave_value::is_user_script, octave_value::is_user_function):
	New functions.
	* ov-base.h (octave_base_value::is_user_script,
	octave_base_value::is_user_function): New pure virtual functions.
	* ov-fcn.h (octave_function::is_user_script,
	octave_function::is_user_function): Delete.

	* load-save.cc (install_loaded_variable): Greatly simplify.

	* load-path.h, load-path.cc: Change private_fcn_file to private_file.
	(load_path::private_fcn_map): New data member.
	(load_path::do_add, load_path::do_remove, load_path::do_update):
	Also manage private_file_map here.
	(load_path::add_to_private_fcn_map): New function.
	(load_path::remove_private_fcn_map): New function.
	(load_path::do_find_private_fcn): Make it work.
	(get_file_list): New function.
	(load_path::do_display): Use it.  Display private map.
	(load_path::find_method, load_path::find_fcn): Handle directory name.

	* token.cc, token.h, toplev.cc, debug.cc, defun.cc, defun-dld.h,
	defun-int.h, error.cc, help.cc, load-save.h, load-save.cc, mex.cc,
	octave.cc, ov-fcn-handle.cc, ov-usr-fcn.cc, parse.h, parse.y,
	lex.l, variables.cc: Adapt to new symbol table objects (my
	apologies for the lack of detail).

	* unwind-prot.h (unwind_protect::add): Set default value for ptr arg.

	* pt-stmt.cc (tree_statement::eval): Rework method for deciding
	whether to assign value to ans.

	* pt-idx.cc (tree_index_expression::rvalue):
	Looking up symbol may evaluaate first args now.

	* pt-id.h, pt-id.cc (tree_identifier::document,
	tree_identifier::is_defined, tree_identifier::define,
	tree_identifier::is_function, tree_identifier::lookup,
	tree_identifier::link_to_global): Delete.
	(tree_identifier::do_lookup): Simplify.
	(tree_identifier::rvalue): Looking up symbol can't execute script now.

	* pt-misc.cc (tree_parameter_list::initialize_undefined,
	tree_parameter_list::is_defined): Call is_variable for elt, not
	is_defined.

	* pt-fcn-handle.h (tree_anon_fcn_handle::fcn): Now pointer to
	octave_user_function, not value.  Change all uses.

	* pt-decl.h (tree_decl_command::initialized): Delete data member.
	(tree_decl_elt::is_variable): New function.
	* pt-decl.cc: Fix all uses of tree_decl_command::initialized.

	* ls-hdf5.cc, ls-mat-ascii.cc, ls-mat4.cc, ls-mat5.cc,
	ls-oct-ascii.cc, ls-oct-binary.cc, input.cc: Don't include symtab.h.

	* dynamic-ld.h, dynamic-ld.cc (octave_dynamic_loader::load_oct,
	octave_dynamic_loader::load_mex, octave_dynamic_loader::do_load_oct,
	octave_dynamic_loader::do_load_mex): Adapt to new symbol table
	objects.  Return pointer octave_function instead of bool.

	* DLD-FUNCTIONS/dispatch.cc (Fbuiltin): Disable for now.
	Disable code that works with old symbol tables.
	(Fbuiltin): Simply call symbol_table::add_dispatch.

	* pt-arg-list.cc, pt-arg-list.h, pt-assign.cc, pt-assign.h,
	pt-binop.cc, pt-binop.h, pt-bp.h, pt-cell.cc, pt-cell.h,
	pt-cmd.cc, pt-cmd.h, pt-colon.cc, pt-colon.h, pt-const.cc,
	pt-const.h, pt-decl.cc, pt-decl.h, pt-except.cc, pt-except.h,
	pt-exp.h, pt-fcn-handle.cc, pt-fcn-handle.h, pt-id.cc, pt-id.h,
	pt-idx.cc, pt-idx.h, pt-jump.cc, pt-jump.h, pt-loop.cc, pt-loop.h,
	pt-mat.cc, pt-mat.h, pt-misc.cc, pt-misc.h, pt-select.cc,
	pt-select.h, pt-stmt.cc, pt-stmt.h, pt-unop.cc, pt-unop.h:
	Adapt dup functions to use scope instead of symbol_table objects.

	* TEMPLATE-INST/Array-sym.cc: Delete.
	* Makefile.in (TI_XSRC): Remove it from the list.

	* symtab.h, symtab.cc: Replace with new versions.

	2007-05-16  John W. Eaton  <jwe@octave.org>

	* oct-lvalue.cc, oct-lvalue.h (octave_lvalue::chg_fcn): Delete.
	Fix all uses.

	2007-05-15  John W. Eaton  <jwe@octave.org>

	* load-path.cc (load_path::do_find_private_function): New function.
	* load-path.h (load_path::do_find_private_function): Provide decl.
        (load_path::find_private_function): New function.

	2007-05-08  John W. Eaton  <jwe@octave.org>

	* pt-idx.cc (tree_index_expression::rvalue): Handle dispatch here.

	* pt-id.cc (tree_identifier::is_variable, tree_identifier::lookup):
	New functions.
	* pt-id.cc: Provide decls.

	* parse.y (current_class_name): New global variable.
	* parse.h: Provide decl.

	* parse.y (load_fcn_from_file, parse_fcn_file):
	New arg, dispatch_type.
	(parse_fcn_file): Protect current_class_name and set it to
	dispatch_type before parsing function.
	(load_fcn_from_file): If dispatch_type is not empty, call
	load_path::find_method instead of load_path::find_fcn.
	* parse.h: Fix decls for extern functions.

	* lex.h (lexical_feedback::parsing_class_method): New data member.
	* lex.l (lexical_feedback::init): Initialize it.
	(lookup_identifier): Check it and set sym_name accordingly.

	* ov-usr-fcn.h (octave_user_function::mark_as_class_constructor,
	octave_user_function::is_class_constructor,
	octave_user_function::mark_as_class_method,
	octave_user_function::is_class_method): New functions.
	(octave_user_function::class_constructor,
	octave_user_function::class_method): New data members.
	* ov-usr-fcn.cc (octave_user_function::octave_user_function):
	Initialize them.

	* load-path.h, load-path.cc: Use typedefs to simplify template decls.
	Use fcn consistently instead of function.
	
	2007-05-03  John W. Eaton  <jwe@octave.org>

	* ov-class.cc (Fclass): Move here.
	* ov-typeinfo.cc: From here.

	* input.cc (octave_gets): Call load_path::update if user input
	contains something other than one of the characters " \t\n\r".

	* ov-class.cc, ov-class.h: New files.
	* Makefile.in: Add them to the appropriate lists.

	* load-path.cc (genpath): Skip directories beginning with "@".
	(load_path::dir_info::get_file_list):  Don't return anything.
	Call get_private_function_map here.
	(load_path::dir_info::initialize): Not here.
	(load_path::dir_info_::get_method_file_map): New function.
	(load_path::method_map): New data member.
	(load_path::do_clear): Also clear method_map.
	(load_path::do_add): Also call add_to_method_map.
	(load_path::do_update): Also clear method_map and call
	(load_path::do_find_method): New function.
	(load_path::do_add_to_method_map): New function.
	(load_path::move_fcn_map, load_path::move_method_map): New functions.
	(load_path::move): Use them.
	(load_path::remove_fcn_map, load_path::remove_method_map):
	New functions.
	(load_path::remove): Use them.
	* load-path.h: Provide/fix decls.
	(load_path::find_method): New static function.

	* Makefile.in (%.df : %.cc): Use mv instead of
	$(simple-move-if-change-rule).

2007-12-21  John W. Eaton  <jwe@octave.org>

	Version 3.0.0 released.

	* version.h (OCTAVE_VERSION): Now 3.0.0.
	(OCTAVE_API_VERSION): Now api-v32.
	(OCTAVE_RELEASE_DATE): Now 2007-12-21.

2007-12-19  Thomas Kasper  <ThomasKasper@gmx.net>

	* OPERATORS/op-scm-cs.cc (DEFBINOP (pow)): Extract complex value
	from second arg.

2007-12-12  David Bateman  <dbateman@free.fr>

	* DLD-FUNCTIONS/sparse.cc (Fsparse): Check for scalar arguments
	for 2 argument version.

2007-12-12  John W. Eaton  <jwe@octave.org>

	* graphics.h.in (class axes) Add the layer property.
	* graphics.cc (class axes) Ditto.

	* graphics.cc (gh_manager::get_handle): Use ceil instead of trunc.

2007-12-11  John W. Eaton  <jwe@octave.org>

	* version.h (OCTAVE_VERSION): Now 2.9.19+.

2007-12-11  John W. Eaton  <jwe@octave.org>

	* version.h (OCTAVE_VERSION): Now 2.9.19.
	(OCTAVE_API_VERSION): Now api-v31.
	(OCTAVE_RELEASE_DATE): Now 2007-12-11.

	* graphics.cc (make_graphics_object):
	If successful, call __request_drawnow__.

	* input.cc (Vgud_mode): New static variable.
	(F__gud_mode__): New function.
	(get_user_input): If debug and Vgud_mode, print location info in
	format for Emacs GUD mode.

2007-12-11  David Bateman  <dbateman@free.fr>

	* OPERATORS/op-bm-sbm.cc, OPERATORS/op-b-sbm.cc, 
	OPERATORS/op-cm-scm.cc, OPERATORS/op-cm-sm.cc, 
	OPERATORS/op-cs-scm.cc, OPERATORS/op-cs-sm.cc, 
	OPERATORS/op-m-scm.cc, OPERATORS/op-m-sm.cc, 
	OPERATORS/op-s-scm.cc, OPERATORS/op-s-sm.cc: Privilege 
	conversion to dense matrices for compatibility.

	* DLD-FUNCTIONS/sparse.cc (FSparse): Remove the mutate flag, as
	default bahavior is now to keep matrix sparse always.
	
	* graphics.cc (axes::properties::properties):
	Initialize xcolor, ycolor, and zcolor to (0, 0, 0).

	* graphics.h.in (gh_manager::next_handle): Now double.
	* graphics.cc (gh_manager::get_handle, gh_manager::gh_manager):
	Set fractional part of next_handle to a random value.

2007-12-10  John W. Eaton  <jwe@octave.org>

	* ov-cell.cc (octave_cell::all_strings): Handle empty elements.
	Handle N-d cell arrays.

	* DLD-FUNCTIONS/fsolve.cc (Ffsolve):
	For compatibility, return [x, fval, info] instead of [x, info, msg].
 	Move tests here from test/test_nonlin.m.

	* OPERATORS/op-cm-s.cc: Define function for el_mul with
	DEFNDBINOP_OP, not DEFBINOP_OP.

	* OPERATORS/op-s-cm.cc: Define functions for el_and and el_or
	operators with DEFNDBINOP_FN.

	* oct-hist.cc (default_history_file): Use file_ops::concat.
	* load-path.cc (dir_info::initialize, dir_info::get_file_list,
	load_path::do_find_fcn, load_path::do_find_file, genpath,
	execute_pkg_add_or_del, load_path::do_find_first_of,
	load_path::do_find_all_first_of): Likewise.

	* help.cc (Flookfor): Avoid doubling directory separator.
	* dirfns.cc (Fmkdir): Likewise.

	* pt-mat.cc (tree_matrix::rvalue): Produce sq_string if any
	strings are sq_string objects.

2007-12-10  David Bateman  <dbateman@free.fr>

	* graphics.h.in (data_property::data): Declare as NDArray instead
	of Matrix.  Change all uses.

2007-12-10  Shai Ayal  <shaiay@users.sourceforge.net>

	* graphics.h.in (class figure) Add the color property.
	* graphics.cc (class figure) Ditto.

2007-12-07  David Bateman  <dbateman@free.fr>

	* data.cc (Fnorm): Don't return a scalar stored as a sparse
	matrix. Convert it to a scalar.

	* graphics.cc (check_limit_val): Delete.
	(check_limit_vals): Simplify and no longer use check_limit_val.

2007-12-05  John W. Eaton  <jwe@octave.org>

	* version.h (OCTAVE_VERSION): Now 2.9.18+.

2007-12-05  John W. Eaton  <jwe@octave.org>

	* version.h (OCTAVE_VERSION): Now 2.9.18.
	(OCTAVE_API_VERSION): Now api-v30.
	(OCTAVE_RELEASE_DATE): Now 2007-12-05.

2007-12-04  John W. Eaton  <jwe@octave.org>

	* Makefile.in (INCLUDES): Add builtins.h to the list.

	* src/oct-hist.cc (default_history_file): Use += instead of
	push_back to append character to std::string object.

	* Makefile.in (LIBDLFCN): Delete all uses.

	* toplev.cc (octave_config_info): Remove LIBDLFCN and
	DLFCN_INCFLAGS from the list.
	* oct-conf.h.in (OCTAVE_CONF_DLFCN_INCFLAGS, OCTAVE_CONF_LIBDLFCN):
	Delete. 

	* error.cc (Ferror): Handle error id.

	* load-save.cc (Fsave, Fload): Doc fixes.
	From Marco Caliari <caliari@sci.univr.it>.

2007-12-04  Shai Ayal  <shaiay@users.sourceforge.net>

	* graphics.h.in (base_properties::get_type
        base_properties::get___modified__,
	base_properties::get___myhandle__): New functions.

2007-12-04  Christoph Mayer  <Christoph.Mayer@dlr.de>.

	* toplev.h (clean_up_and_exit, recover_from_exception,
	do_octave_atexit, global_command, curr_parent_function):
	Tag with OCTINTERP_API. 

2007-12-03  David Bateman  <dbateman@free.fr>

	* DLD-FUNCTIONS/luinc.cc: Make tests conditional on HAVE_UMFPACK.
	* DLD-FUNCTIONS/spqr.cc: Make tests conditional on HAVE_CXSPARSE.

	* DLD-FUNCTIONS/regexp.cc: Use "%!testif" blocks for conditional
	tests on PCRE.

	* graphics.cc (class axes): Add color and activepositionproperty
	properties to axis objects.
	* graphics.h.in (class axes): ditto.

	* DLD-FUNCTIONS/regexp.cc: Also include sys/types.h for regexp.

2007-11-30  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/sort.cc (operator < (const Complex&, const Complex&),
	operator > (const Complex&, const Complex&)):
	Pass args by const reference, not value.

	* src/data.cc, src/matherr.c, src/pr-output.cc, src/sysdep.cc,
	src/DLD-FUNCTIONS/__dsearchn__.cc, src/DLD-FUNCTIONS/minmax.cc,
	src/DLD-FUNCTIONS/qz.cc, src/DLD-FUNCTIONS/sort.cc,
	src/DLD-FUNCTIONS/tsearch.cc: Include lo-math.h instead of cmath
	or math.h.

2007-11-30  Moritz Borgmann  <octave@moriborg.de>

	* ls-mat5.h (mat5_data_type): Delete trailing comma in enum decl.
	* oct-stream.cc (octave_base_stream::do_printf): Use
	OCTAVE_EMPTY_CPP_ARG to avoid annoying Sun compiler warning.
	* OPERATORS/op-sbm-sbm.cc (DEFDBLCONVFN): Delete useless macro
	invocation.

	* DLD-FUNCTIONS/typecast.cc (typecast): No longer static.

2007-11-30  John W. Eaton  <jwe@octave.org>

	* graphics.cc (updating_axis_limits): New static variable.
	(check_limit_val, check_limit_vals, get_axis_limits): New functions.
	(axes::update_axis_limits): Make it work.
	* graphics.h.in (base_properties::get_children): New function.
	(base_properties::get_xdata, base_properties::get_ydata,
	base_properties::get_zdata, base_properties::get_ldata,
	base_properties::get_udata, base_properties::get_xldata,
	base_properties::get_xudata, base_properties::get_cdata,
	base_properties::get_properties (void) const):
	New virtual functions.
	(graphics_object::get_xdata, graphics_object::get_ydata,
	graphics_object::get_zdata, graphics_object::get_ldata,
	graphics_object::get_udata, graphics_object::get_xldata,
	graphics_object::get_xudata, graphics_object::get_cdata,
	graphics_object::get_properties (void) const,
	root_figure::get_properties (void) const,
	figure::get_properties (void) const,
	axes::get_properties (void) const,
	line::get_properties (void) const,
	text::get_properties (void) const,
	image::get_properties (void) const,
	patch::get_properties (void) const,
	surface::get_properties (void) const)): New functions.
	(radio_property::radio_property (const radio_values&)):
	Provide	default value for argument. 
	(radio_property::radio_property (const radio_values&, const
	std::string&)): Delete.
	(axes::xlimmode, axes::ylimmode, axes::zlimmode, axes::climmode,
	axes::xscale, axes::yscale, axes::zscale): Declare as
	radio_property instead of octave_value values.  Adjust all uses.

2007-11-29  John W. Eaton  <jwe@octave.org>

	* pr-output.cc (octave_print_internal_template (std::ostream&,
	const octave_int<T>&, bool)): Rename from octave_print_internal.
	(PRINT_INT_SCALAR_INTERNAL): New macro.  Use it to define
	non-template functions for printing scalar octave_int values.
	(octave_print_internal_template (std::ostream&, const intNDArray<T>&,
	bool, int)): Rename from octave_print_internal.
	(PRINT_INT_ARRAY_INTERNAL): New macro.  Use it to define
	non-template functions for printing scalar octave_int values.
	* pr-output.h: Declare non-template functions for printing
	octave_int scalar and array values.
	Delete declarations of template funtions for printing octave_int
	scalar and array values.

2007-11-28  John W. Eaton  <jwe@octave.org>

	* graphics.cc (base_properties::update_axis_limits,
	axes::update_axis_limits): New functions.
	* graphics.h.in (class data_property): New class.
	(graphics_object::update_axis_limits): New function.
	(base_graphics_object::update_axis_limits): New virtual function.
	(base_properties::update_axis_limits, axes::update_axis_limits):
	Provide decls.
	(class line, class image, class patch, class surface): Use it
	instead of octave_value for data properties.  Tag data properties
	with "l" modifier.
	(class axes): Tag scale properties with "l" modifier.
	* genprops.awk: Handle "l" modifier.

	* mkbuiltins: Emit #include "builtins.h" for the builtins.cc file.
	From Christoph Mayer <Christoph.Mayer@dlr.de>.

	* TEMPLATE-INST/Array-tc.cc (resize_fill_value<octave_value>):
	Delete.

	* Cell.cc (Cell::Cell (const dim_vector&, const string_vector&, bool)):
	Initialize undefined values to resize_fill_value ().

2007-11-27  John W. Eaton  <jwe@octave.org>

	* TEMPLATE-INST/Array-tc.cc (resize_fill_value<octave_value>):
	Tag inline.  From Moritz Borgmann <octave@moriborg.de>.

	* pt-bp.cc (tree_breakpoint::visit_decl_command): Also check line
	number of cmd.
	* pt-decl.cc (tree_global_command::eval, tree_static_command::eval):
	Insert MAYBE_DO_BREAKPOINT here.

	* error.cc (Fwarning): If setting state "all" to "error", leave
	Octave:matlab-incompatible and Octave:single-quote-string warning
	states unchanged.
	(warning_enabled): Allow individual warning states to override
	"warning error all".

	* octave.cc (execute_eval_option_code, execute_command_line_file):
	Handle interrupts.
	* toplev.cc (recover_from_exception): Now extern.
	* toplev.h (recover_from_exception): Provide decl.

	* pt-idx.cc (tree_index_expression::lvalue): Treat object == []
	the same as undefined.

2007-11-26  John W. Eaton  <jwe@octave.org>

	* oct-stream.cc (DO_DOUBLE_CONV): Always use long.

	* oct-stream.cc (do_write): Call float_value and double_value on
	octave_int<T> objects instead of relying on conversion operators.

	* ls-mat5.cc (read_mat5_binary_element, OCTAVE_MAT5_INTEGER_READ): 
	Call double_value on octave_int<T> objects
	instead of relying on conversion operator.
	(read_mat5_binary_file_header): 
	Call char_value on octave_int<T> objects
	instead of relying on conversion operator.
	(read_mat5_binary_element): 
	Call bool_value on octave_int<T> objects
	instead of relying on conversion operator.

	* ov-intx.h (OCTAVE_VALUE_INT_MATRIX_T::double_value, 
	OCTAVE_VALUE_INT_MATRIX_T::scalar_value,
	OCTAVE_VALUE_INT_MATRIX_T::matrix_value,
	OCTAVE_VALUE_INT_MATRIX_T::complex_matrix_value,
	OCTAVE_VALUE_INT_MATRIX_T::array_value,
	OCTAVE_VALUE_INT_MATRIX_T::complex_array_value,
	OCTAVE_VALUE_INT_SCALAR_T::double_value, 
	OCTAVE_VALUE_INT_SCALAR_T::scalar_value,
	OCTAVE_VALUE_INT_SCALAR_T::matrix_value,
	OCTAVE_VALUE_INT_SCALAR_T::complex_matrix_value,
	OCTAVE_VALUE_INT_SCALAR_T::array_value,
	OCTAVE_VALUE_INT_SCALAR_T::complex_array_value,
	Call double_value on octave_int<T> objects
	instead of relying on conversion operator.
	(OCTAVE_VALUE_INT_MATRIX_T::bool_array_value,
	OCTAVE_VALUE_INT_SCALAR_T::bool_array_value):
	Call bool_value on octave_int<T> objects
	instead of relying on conversion operator.
	(OCTAVE_VALUE_INT_MATRIX_T::char_array_value,
	OCTAVE_VALUE_INT_SCALAR_T::char_array_value):
	Call char_value on octave_int<T> objects
	instead of relying on conversion operator.

2007-11-26  David Bateman  <dbateman@free.fr>

	* Sparse-op-defs.h (SPARSE_BASE_REDUCTION_OP): Check that the fill
	value is not zero before creating the matrices.

	* graphics.cc (class axes): Add the {x|y|z}color color property.
	* graphics.h.in (class axes): ditto.

	* ov-base.cc (Vsparse_auto_mutate, Fsparse_auto_mutate): New
	internal variable and built-in function to set it.
	* ov-base.h (extern bool Vsparse_auto_mutate): Export internal
	variable to other functions.
	* ov-re-sparse.cc (octave_sparse_matrix::try_narrowing_conversion
	(void)), ov-cx-sparse.cc 
	(octave_sparse_complex_matrix::try_narrowing_conversion (void)), 
	ov-bool-sparse.cc
	(octave_sparse_bool_matrix::try_narrowing_conversion (void)):
	Use Vsparse_auto_mutate flag to determine whether to convert
	sparse matrices to full matrices if that saves space.

	* DLD-FUNCTIONS/minmax.cc (MINMAX_DOUBLE_BODY): New version of
	MINMAX_BODY macro without the initialization.
	(MINMAX_INT_BODY): Macro for min/max for the integer types
	(MINMAX_BODY): New macro that calls the appropriate instantiation
	of the other two macros.

	* graphics.cc (class axes): Add __colorbar__ property.
	* graphics.h.in (class axes): ditto.

	* data.cc (Fnorm): Document the "inf" string argument for matrice
	and vectors and the "fro" argument for vectors.

	* graphics.h.in (class figure): Add __enhanced__ property to cache
	whether the terminal supports enhanced mode. Add interpreter
	property to all text classes, to determine the renderer for the
	text.
	* graphics.cc (class figure): ditto.

2007-11-26  Shai Ayal  <shaiay@users.sourceforge.net>

	* graphics.cc (string_properties): delete class 
	(property_name): rename class to caseless_str. 
	(radio_values::possible_vals): change type to caseless_str.

2007-11-26  Michael Goffioul  <michael.goffioul@gmail.com>

	* builtins.h (install_builtins): Tag with OCTINTERP_API.
	* toplev.h (octave_interpreter_ready, octave_initialized): Likewise.

2007-11-14  John W. Eaton  <jwe@octave.org>

	* mex.cc (mex::foreign_memlist): New data member.
	(mex::mark_foreign, mex::unmark_foreign): New functions.
	(mex::free): Don't warn about pointers found in foreign_memlist.
	(maybe_mark_foreign): New function.
	(mxArray_octave_value::get_data, mxArray_octave_value::get_ir,
	mxArray_octave_value::get_jc):
	Call maybe_mark_foreign on returned pointer.

2007-11-14  David Bateman  <dbateman@free.fr>

	* mex.cc (mxArray_sparse::mxArray_sparse (const mxArray_sparse&)):
	Also ensure that pr and pi are correctly copied.

2007-11-14  John W. Eaton  <jwe@octave.org>

	* Makefile.in (@bsd_gcc_kluge_targets_frag@): Delete line for
	Makefrag.bsd substitution.

	* graphics.h.in (base_properties::tag): New property.
	(base_properties::get_tag, base_properties::set_tag): New functions.
	(root_figure::set, root_figure::get, figure::set, figure::get,
	axes::set, axes::get, line::set, line::get, text::set, text::get,
	patch::set, patch::get, surface::set, surface::get): Handle tag.
	
2007-11-14  Joseph P. Skudlarek  <Jskud@Jskud.com>

	* DLD-FUNCTIONS/regexp.cc: Undo previous change.

2007-11-14  John W. Eaton  <jwe@octave.org>

	* mex.cc (mex::mark, mex::unmark): Now public.
	(mex::persistent): Delete.
	(mexMakeArrayPersistent): Call maybe_unmark_array instead of
	mex::persistent.
	(mexMakeMemoryPersistent): Call maybe_unmark instead of
	mex::persistent.
	(maybe_unmark (void *)): New function.
	(mxSetDimensions, mxSetPr, mxSetPi, mxSetData, mxSetImagData,
	mxSetIr, mxSetJc): Use it.  From David Bateman.

2007-11-13  Joseph P. Skudlarek  <Jskud@Jskud.com>

	* DLD-FUNCTIONS/rand.cc (Frand): Update URL reference in doc string.

2007-11-12  David Bateman  <dbateman@free.fr>

	* graphics.h.in, graphics.cc (class text): Add the fontangle and
	fontweight properties. Add string_property class to handle the
	string radio values.

2007-11-12  John W. Eaton  <jwe@octave.org>

	* version.h (OCTAVE_VERSION): Now 2.9.17+.

2007-11-12  David Bateman  <dbateman@free.fr>

	* graphics.h.in, graphics.cc (class text): Add the fontname and
	fontsize properties. 

2007-11-10  John W. Eaton  <jwe@octave.org>

	* version.h (OCTAVE_VERSION): Now 2.9.17.
	(OCTAVE_API_VERSION): Now api-v29.
	(OCTAVE_RELEASE_DATE): Now 2007-11-10.

2007-11-09  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/__gnuplot_raw__.l: Don't add atexit ("closeplot")
	to PKG_ADD file.

2007-11-09  David Bateman  <dbateman@free.fr>

	* graphics.cc, graphics.h.in (class patch): Add the field "keylabel".

	* data.cc (Fcputime) [__WIN32__]: Correct scaling of sys and usr times.

	* graphics.cc (is_handle): Handle must be a real scalar.

2007-11-09  Joseph P. Skudlarek  <Jskud@Jskud.com>

	* data.cc (Fislogical): Fix typo in documentation entry.

2007-11-08  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/__gnuplot_raw__.l (F__gnuplot_save_data__):
	Note obsolescence in doc string.
	(WARN_OBSOLETE): New macro
	(F__gnuplot_save_data__, Fgnuplot_command_plot,
	Fgnuplot_command_replot, Fgnuplot_command_splot,
	Fgnuplot_command_using, Fgnuplot_command_with,
	Fgnuplot_command_axes, Fgnuplot_command_title,
	Fgnuplot_command_end, Fgnuplot_use_title_option,
	F__clear_plot_window__, Fcloseplot, Fpurge_tmp_files,
	F__gnuplot_raw__, F__gnuplot_send_inline_data__, F__gnuplot_set__,
	F__gnuplot_show__, F__gnuplot_plot__, F__gnuplot_splot__,
	F__gnuplot_replot__): Use WARN_OBSOLETE.

2007-11-07  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/regexp.cc: Handle HAVE_PCRE_PCRE_H.  Check
	HAVE_PCRE_H instead of HAVE_PCRE to decide whether to include pcre.h.

	* DLD-FUNCTIONS/__gnuplot_raw__.l (Fgnuplot_command_plot,
	Fgnuplot_command_replot, Fgnuplot_command_splot,
	Fgnuplot_command_using, Fgnuplot_command_with,
	Fgnuplot_command_axes, Fgnuplot_command_title,
	Fgnuplot_command_end, Fgnuplot_use_title_option,
	F__clear_plot_window__, Fcloseplot, Fpurge_tmp_files,
	F__gnuplot_raw__, F__gnuplot_send_inline_data__, F__gnuplot_set__,
	F__gnuplot_show__, F__gnuplot_plot__, F__gnuplot_splot__,
	F__gnuplot_replot__): Note obsolescence in doc string.

2007-11-06  Kai Habel  <kai.habel@gmx.de>

	* graphics.h.in (class surface::properties): New properties, 
	cdata, facecolor, facealpha, edgecolor, linestyle,
	linewidth, marker, markeredgecolor, markerfacecolor, markersize.
	* graphics.cc (surface::properties::properties,
	surface::properties::set, surface::properties::get,
	surface::properties::factory_defaults): Handle new properities.

2007-11-06  David Bateman  <dbateman@free.fr>

	* data.cc (DATA_REDUCTION): Handle the 'native' and 'double'
	arguments of the Fsum function.
	* OPERATORS/op-bm-bm.cc (matrix_to_bool_matrix,
	scalar_to_bool_matrix): New type conversion functions.
	(install_bm_bm_ops): Install new type conversions functions.

2007-11-06  Michael Goffioul  <michael.goffioul@gmail.com>

	* toplev.cc (Fsystem) [__WIN32__ && ! __CYGWIN__]:
	Quote the complete command.

2007-11-06  John W. Eaton  <jwe@octave.org>

	* data.cc (Fnorm): New tests.

	* defun-int.h (DEFINE_FUNX_INSTALLER_FUN3): Don't install function
	if check_version produces an error.

2007-11-05  John W. Eaton  <jwe@octave.org>

	* pt-idx.cc (tree_index_expression::lvalue): Try to do a better
	job of computing the number of elements in lvalue expressions when
	the last indexing element is ".".

2007-11-02  John W. Eaton  <jwe@octave.org>

	* file-io.cc (fopen_mode_to_ios_mode): Use std::ios::app instead
	of std::ios::ate.

2007-11-02  Olli Saarela  <Olli.Saarela@kcl.fi>

	* input.cc, pr-output.cc, bitfcns.cc, DLD-FUNCTIONS/time.cc:
        Fix broken @examples in help texts.

2007-10-31  John W. Eaton  <jwe@octave.org>

	* data.cc (Fnorm): Avoid warning about p_val possibly being used
	uninitialized.

	* version.h (OCTAVE_VERSION): Now 2.9.16+.

2007-10-31  John W. Eaton  <jwe@octave.org>

	* version.h (OCTAVE_VERSION): Now 2.9.16.
	(OCTAVE_API_VERSION): Now api-v28.
	(OCTAVE_RELEASE_DATE): Now 2007-10-31.

2007-10-31  Muthiah Annamalai  <muthuspost@gmail.com>

	* pt-assign.h (tree_simple_assignment::op_type,
	tree_multi_assignment::op_type): New functions.
	* pt-unop.h (tree_unary_expression::op_type): New function.

2007-10-31  John W. Eaton  <jwe@octave.org>

	* graphics.cc (line::properties::get): Fix property name
	(markerface -> markerfacecolor).

	* Makefile.in (INCLUDES): Add debug.h to the list.

2007-10-30  John Swensen  <jpswensen@comcast.net>

	* debug.h: New file.
	* debug.cc (parse_dbfunction_params, do_find_bkpt_list,
	intmap_to_ov): New functions.
	(Fdbstop, Fdbclear): Use parse_dbfunction_params.
	Improve compatibility.
	(Fdbstatus): Improve compatibility.

	* help.cc (do_which): No longer static.
	* help.h: Provide decl.

2007-10-30  David Bateman  <dbateman@free.fr>

	* symtab.cc: Doc fixes for small book format.

2007-10-30  John W. Eaton  <jwe@octave.org>

	* file-io.cc (fopen_mode_to_ios_mode): Handle 'W' as 'w' and 'R'
	as 'r', but warn about them.

2007-10-29  Thomas Treichl  <Thomas.Treichl@gmx.net>

	* data.cc: Include sytime.h, sys/types.h, and sys/resource.h.

2007-10-25  John W. Eaton  <jwe@octave.org>

	* graphics.cc (figure::properties::set_currentaxes): 
	Allow currentfigure to be NaN.

2007-10-25  Michael Goffioul  <michael.goffioul@gmail.com>

	* DLD-FUNCTIONS/__contourc__.cc: Use unsigned int instead of uint.
	(drawcn): Use 1 << k instead of pow (2, k).

2007-10-25  John W. Eaton  <jwe@octave.org>

	* symtab.h (symbol_record::TYPE): Delete trailing comma in enum decl.

	* ov-base.h (DECLARE_OV_TYPEID_FUNCTIONS_AND_DATA): Use
	OCTAVE_EMPTY_CPP_ARG to avoid annoying Sun compiler warning.

	* Makefile.in (graphics.h): Use $(AWK) instead of awk.

	* DLD-FUNCTIONS/time.cc (Ftime, Fmktime): Avoid unnecessary cast.

	* data.cc (Ftic, Ftoc): Call double_value on octave_time objects
	instead of relying on conversion operator.
	* ov.cc (octave_value::octave_value (octave_time)): Likewise.

	* variables.cc (symbol_out_of_date): Use explicit conversion to
	time_t instead of relying on conversion operator.
	* ov-fcn-handle.cc (octave_fcn_handle::subsref): Likewise.

	* data.cc (tic_toc_timestamp): Rename from __tic_toc_timestamp__.
	Change all uses.

2007-10-24  David Bateman  <dbateman@free.fr>

	* ov-intx.h (OCTAVE_VALUE_INT_MATRIX_T::OCTAVE_VALUE_INT_MATRIX_T
	(const ArrayN<OCTAVE_INT_T>&)): New constructor.
	* ov.cc (octave_value::octave_value(const ArrayN<T>) with T being
	octave_int8, octave_uint8, octave_int16, octave_uint16,
	octave_int32, octave_uint32, octave_int64, octave_uint64): New
	constructors.
	* ov.h (octave_value::octave_value(const ArrayN<T>) with T being
	octave_int8, octave_uint8, octave_int16, octave_uint16,
	octave_int32, octave_uint32, octave_int64, octave_uint64):
	Declare them.
	* DLD-FUNCTIONS/sort.cc (template class octave_sort<T>,
	template class vec_index<T>, template class
	octave_sort<vec_index<T> *>,  with T being
	octave_int8, octave_uint8, octave_int16, octave_uint16,
	octave_int32, octave_uint32, octave_int64, octave_uint64): New
	instantiations of sort template classes.
	(Fsort): Use them.

2007-10-24  John W. Eaton  <jwe@octave.org>

	* graphics.cc (root_figure::properties::set_currentfigure):
	Allow currentfigure to be NaN.

	* pt-idx.cc (tree_index_expression::lvalue): Correctly compute
	number of elements in lvalue expression when last indexing
	element is ".".

2007-10-23  John W. Eaton  <jwe@octave.org>

	* graphics.cc (is_handle (const graphics_handle&)): New function.
	(gh_manager::do_free, reparent, base_properties::set_parent,
	properties::get_title, properties::get_xlabel,
	properties::get_ylabel, properties::get_zlabel,
	properties::remove_child, make_graphics_object, F__go_figure__,
	F__go_delete__, __go_axes_init__): Call OK on graphics handle
	object instead of relying on implicit conversion operator.
	* graphics.h.in (graphics_handle::operator double ()): Delete.
	(graphics_handle::operator bool ()): Delete.
	(gh_manager::do_handle_list, gh_manager::do_figure_handle_list,
	base_properties::adopt): Call VALUE on graphics handle object
	instead of relying on implicit conversion operator.

	* ov-base.h (DECLARE_OV_TYPEID_FUNCTIONS_AND_DATA): Call
	DECLARE_OV_TYPEID_FUNCTIONS_AND_DATA2 with "( )" as arg list
	instead of "()".

	* mxarray.h.in (mxClassID): Delete trailing comma in enum decl.

	* symtab.h (symbol_table::symbol_table): Reduce default table size
	to 64.

2007-10-22  Kim Hansen  <kimhanse@gmail.com>

	* unwind-prot.cc: Include <cstring>.

2007-10-22  David Bateman  <dbateman@free.fr>

        * data.cc (Ftic, Ftoc, Fcputime): New builtin versions of the
        benchmarking functions for speed.
	* oct-map.cc (Octave_map::squeeze, Octave_map::permute): New methods.
	(Octave_map::index (const octave_value_list&, bool)): Add resize_ok
	argument, define as const and use const_iterator internally.
	(Octave_map::index (idx_vector&, ...), Octave_map::index (Array 
	<idx_vector>&, ...)): New forms of the index function.
	* oct-map.h (squeeze, permute, indx (const octave_value_list&, bool),
	index (idx_vector&, ...), index (Array <idx_vector>&, ...)): Add
	or update declaration.
	* ov-struct.cc (octave_struct::do_index_op (const octave_value_list&,
	bool)): New method.
	* ov-struct.h (do_index_op (const octave_value_list&, bool)): Declare
	it.
	(squeeze (void), permute (const Arra<int>&, bool): New methods.

2007-10-19  Kai Habel  <kai.habel@gmx.de>

	* DLD-FUNCTIONS/__contourc__.cc (add_point): Rename from
	cl_add_point.  Change all uses.
	(end_contour): Rename from cl_end_contour.  Change all uses.
	(start_contour): Rename from cl_start_contour.  Change all uses.
	(drawcn): Rename from cl_drawcn.  New algorithm for locating contours.
	(mark_facets): New function.
	(cntr): Rename from cl_cntr.  Change all uses.  New algorithm for
	locating contours.

2007-10-19  John W. Eaton  <jwe@octave.org>

	* ov-cell.cc (octave_cell::subsasgn): If RHS is cs-list, don't
	fail if shape of LHS is different.
	* ov-struct.cc (octave_struct::subsasgn): Likewise.

2007-10-19  Olli Saarela  <Olli.Saarela@kcl.fi>

	* help.cc (Flookfor): Call print_usage instead of usage.
	* DLD-FUNCTIONS/cellfun.cc (Fmat2cell): Likewise.

2007-10-17  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/spchol.cc (Fsymbfact): Delete special code for METIS.

2007-10-17  Gabriele Pannocchia  <g.pannocchia@ing.unipi.it>

	* DLD-FUNCTIONS/__qp__.cc (qp): Fix check for Wact(j).

2007-10-15  Søren Hauberg  <hauberg@gmail.com>

	* error.cc (Ferror): Make text fit on pages when using smallbook.
	* load-save.cc (Fsave_header_format_string): Ditto.
        * ov-struct.cc (Fcell2struct): Ditto.
        * DLD-FUNCTIONS/besselj.cc (Fairy): Ditto.
        * DLD-FUNCTIONS/urlwrite.cc (Furlwrite, Furlread): Ditto.

2007-10-15  David Bateman  <dbateman@free.fr>

	* graphics.cc (axes::properties::get): Fix typo.

2007-10-13  John W. Eaton  <jwe@octave.org>

	* version.h (OCTAVE_VERSION): Now 2.9.15+.

2007-10-13  John W. Eaton  <jwe@octave.org>

	* version.h (OCTAVE_VERSION): Now 2.9.15.
	(OCTAVE_API_VERSION): Now api-v27.
	(OCTAVE_RELEASE_DATE): Now 2007-10-13.

2007-10-13  David Bateman  <dbateman@free.fr>

	* graphics.h.in (class patch): Add the faces and vertices properties.
	* graphics.cc (patch::properties::properties): Initialize faces
	and vertices.
	(patch::properties::get): Also fetch faces and vertices.
	(patch::properties::factory_defaults): Set faces and vertices.

2007-10-12  John W. Eaton  <jwe@octave.org>

	* Change copyright notices in all files that are part of Octave to
	GPLv3 or any later version.

2007-10-11  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/urlwrite.cc (urlget): Disable use of EPSV mode.

2007-10-11  Brian Gough  <bjg@network-theory.co.uk>

	* DLD-FUNCTIONS/fftn.cc, DLD-FUNCTIONS/regexp.cc, dirfns.cc,
	input.cc: Spelling fixes.

2007-10-10  John W. Eaton  <jwe@octave.org>

	* OPERATORS/op-sbm-sbm.cc (bool_matrix_to_double_matrix):
	New conversion function.
	(install_sbm_sbm_ops): Install it.

	* ov-re-mat.cc (double): Handle sparse as a special case.

2007-10-10  Olli Saarela  <Olli.Saarela@kcl.fi>

	* data.cc, debug.cc, dirfns.cc, error.cc, file-io.cc, help.cc,
	load-save.cc, ov-fcn-inline.cc, parse.y, pr-output.cc, symtab.cc,
	syscalls.cc, toplev.cc, variables.cc, DLD-FUNCTIONS/bsxfun.cc,
	DLD-FUNCTIONS/cellfun.cc, DLD-FUNCTIONS/fft.cc,
	DLD-FUNCTIONS/fft2.cc, DLD-FUNCTIONS/fftw.cc,
	DLD-FUNCTIONS/gcd.cc, DLD-FUNCTIONS/lsode.cc,
	DLD-FUNCTIONS/luinc.cc, DLD-FUNCTIONS/matrix_type.cc,
	DLD-FUNCTIONS/rand.cc, DLD-FUNCTIONS/regexp.cc,
	DLD-FUNCTIONS/sparse.cc, DLD-FUNCTIONS/spchol.cc,
	DLD-FUNCTIONS/splu.cc, DLD-FUNCTIONS/spparms.cc: Spelling fixes.

2007-10-10  Kim Hansen  <kimhanse@gmail.com>

	* DLD-FUNCTIONS/lpsolve.cc: Delete.
	* Makefile.in (DLD_XSRC): Remove it from the list.

        * data.cc (Fsqueeze): Document 2d behaviour.
        * ov-range.h (octave_range::squeeze): New function.

2007-10-09  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/urlwrite.cc (urlget_cleanup): New function.
	(urlget): Protect call to curl_easy_perform with
	BEGIN_INTERRUPT_IMMEDIATELY_IN_FOREIGN_CODE and
	END_INTERRUPT_IMMEDIATELY_IN_FOREIGN_CODE.

2007-10-09  David Bateman  <dbateman@free.fr>

	* input.cc (accept_line): Drop this function and remove automatic
	insertion of closing quotes as the transpose operator confuses it.
	(initialize_command_input): Remove accept_line from here as well.
	
2007-10-09  John W. Eaton  <jwe@octave.org>

	* ov-mapper.cc (octave_mapper::apply): If possible, use
	d_d_map_fcn to handle complex values which have imag(z) == 0.

	* DLD-FUNCTIONS/urlwrite.cc (Furlwrite, Furlread) [! HAVE_CURL]:
	Throw error instead of returning empty string hiding error message
	in third return value.
	(progress_func): Delete.
	(Furlread): Don't set progress callback for curl.
	(write_data, form_query_string, urlget): Now static.

2007-10-08  David Bateman  <dbateman@free.fr>

	*  input.cc (quoting_filename): Function to add a leading quote to
	a string if needed.
	(accept_line): Function to check if a string needs a closing quote
	before calling the rl_newline function.
	(initialize_command_input): Initialize completer_quote_characters,
	filename_quote_characters, quoting_function,
	user_accept_line_function.

	* data.cc (template <class T> static make_diag (...)): New
	template class version of make_diag. Make other make_diag
	functions depend on it.
	( static octave_value make_diag (const octave_value&, 
	octave_idx_type)): Treat all possible internal Octave types.

2007-10-08  John W. Eaton  <jwe@octave.org>

	* load-save.cc 	(save_vars, dump_octave_core):
	Don't pass INFNAN_WARNED to do_save.
	(do_save): Delete unused arg, INFNAN_WARNED.
	Don't pass INFNAN_WARNED to save_ascii_data.
	* ls-oct-ascii.cc (save_ascii_data_for_plotting):
	Don't pass INFNAN_WARNED to save_ascii.
	(save_ascii_data): Delete unused arg, INFNAN_WARNED.

	* ov.h (octave_value::save_ascii): Delete unused arg, INFNAN_WARNED.
	* ov-base-int.cc (octave_base_int_matrix<T>::save_ascii,
	octave_base_int_scalar<T>::save_ascii): Likewise.
	* ov-base-sparse.cc (octave_base_sparse<T>::save_ascii): Likewise.
	* ov-base.cc (octave_base_value::save_ascii): Likewise.
	* ov-bool-mat.cc (octave_bool_matrix::save_ascii): Likewise.
	* ov-bool.cc (octave_bool::save_ascii): Likewise.
	* ov-cell.cc (octave_cell::save_ascii): Likewise.
	* ov-fcn-handle.cc (octave_fcn_handle::save_ascii): Likewise.
	* ov-fcn-inline.cc (octave_fcn_inline::save_ascii): Likewise.
	* ov-list.cc (octave_list::save_ascii): Likewise.
	* ov-struct.cc (octave_struct::save_ascii): Likewise.
	* ov-str-mat.cc (octave_char_matrix_str::save_ascii): Likewise.
	* ov-range.cc (octave_range::save_ascii): Likewise.

	* ov-scalar.cc (octave_scalar::save_ascii): Remove warning about
	reloading inf/nan values.  Delete unused arg, INFNAN_WARNED.
	* ov-complex.cc (octave_complex::save_ascii): Likewise.
	* ov-re-mat.cc (octave_matrix::save_ascii): Likewise.
	* ov-cx-mat.cc (octave_complex_matrix::save_ascii): Likewise.

	* ov-fcn-inline.cc (Finline): Use DEFUNX instead of DEFUN.

2007-10-08  Michael Goffioul  <michael.goffioul@gmail.com>

	* sighandlers.h (can_interrupt): Tag with OCTINERP_API.
	* sysdep.h (raw_mode, octave_popen, octave_pclose,
	same_file_internal): Likewise.

2007-10-06  John W. Eaton  <jwe@octave.org>

	* oct-stream.cc (octave_base_stream::do_scanf):
	Always allow MVAL to re resized if NR > 0.
	(FINISH_CHARACTER_CONVERSION): Likewise.

	* mappers.cc (install_mapper_functions): Undo previous change.
	Fix doc string for lgamma.
	(xzlgamma): Delete.

2007-10-05  John W. Eaton  <jwe@octave.org>

	* graphics.h.in (axes::properties): New property, colororder.
	* graphics.cc (default_colororder): New function.
	(axes::properties::properties, axes::properties::get,
	axes::properties::set_defaults,
	axes::properties::factory_defaults, ): Handle colororder.

	* mappers.cc (xzlgamma): New static function.
	(install_mapper_functions): Pass xzlgamma for c_c_map for lgamma
	mapper.  Pass 1 for can_ret_cmplx_for_real and set hi to
	octave_Inf.

2007-10-04  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/symrcm.cc: Move static functions to top of file to
	avoid forward decls.
	(Q_enq): Delete unused arg QH.  Change all uses.
	(Q_deq): Delete unused arg QT.  Change all uses.
	(find_starting_node): Delete unused local variable J.
	(H_heapify_min, H_insert, find_starting_node, Fsymrcm):
	Move local variable decls to point of first use.

	* OPERATORS/op-streamoff.cc (STREAMOFF_COMP_OP):
	Avoid control-reaches-end-of-non-void-function warning.

	* pt-const.cc (tree_constant::dup): Avoid unused parameter warning.

	* pr-output.cc (set_real_format, set_real_matrix_format,
	set_complex_format, set_complex_matrix_format):
	Delete unused arg, SIGN.  Change uses.

	* oct-map.cc (Octave_map::Octave_map): Avoid shadow warning.

	* load-save.cc (write_header): Use reinterpret_cast to avoid
	old-style cast warning.

	* data.cc (do_permute): Delete unused arg, FNAME.  Change all uses.

	* sysdep.cc (w32_set_octave_home, w32_set_quiet_shutdown,
	MINGW_signal_cleanup):
	Only define if defined (__WIN32__) && !	defined (_POSIX_VERSION).

2007-10-04  Jason Riedy  <jason@acm.org>

        * ls-mat5.cc (read_mat5_binary_data): Map miUTF16 to miUINT16,
        miUTF32 to miUINT32.  The matfile format currently states that
        wide UTF formats are in the same byte order as the integers.
        (read_mat5_binary_element): Replace all UTF16 and UTF32
        characters with the high bit set (value >127) by '?' and warn
        that we've done so.

2007-10-04  Michael Goffioul  <michael.goffioul@gmail.com>

	* sysdep.cc (w32_set_octave_home): Base OCTAVE_HOME location on
	octinterp, not the main exe.

2007-10-03  John W. Eaton  <jwe@octave.org>

	* data.cc (Fnorm): New function.
	(F__vnorm__): Delete.

2007-10-03  Michael Goffioul  <michael.goffioul@gmail.com>

	* DLD-FUNCTIONS/typecast.cc: Include <algorithm>.

2007-10-03  John W. Eaton  <jwe@octave.org>

	* oct-map.cc (Octave_map::Octave_map (const dim_vector&, 
	const string_vector&)): Delete.
	(Octave_map::Octave_map (const dim_vector&, const Cell&)):
	New function.
	* ov-struct.cc (Fstruct): Allow creation of empty struct arrays
	with field names.

2007-10-01  Shai Ayal  <shaiay@users.sourceforge.net>

	* graphics.cc ((color_property::color_property (const
	octave_value& val)): Undo change from 2007-09-26
	(patch::properties::properties): use the "a" modifier to
	face_color & edge_color properties
	(patch::properties::set): Undo change from 2007-09-26
	* graphics.h.in	(color_property::color_property
	(const octave_value& val)): Undo change from 2007-09-26

2007-10-01  Michael Goffioul  <michael.goffioul@gmail.com>

	* octave.cc (octave_main): Add "--line-editing" option to force
	readline line editing.

2007-09-28  David Bateman  <dbateman@free.fr>

	* ov-range.h (int8_array_value, int16_array_value, int32_array_value,
	int64_array_value, uint8_array_value, uint16_array_value, 
	int32_array_value, uint64_array_value): New methods

2007-09-28  Kai Habel  <kai.habel@gmx.de>

	* graphics.h.in (color_property::colormap_property (const Matrix&)):
	Use floating point math in calculation of colormap.

2007-09-26  David Bateman  <dbateman@free.fr>

	* graphics.cc (color_values::str2rgb): accept upper, lower and
	mixed-case versions of the string representing the color. Allow
	black defined as "k" and white as "w".
	(color_property::color_property (const octave_value& val, 
	const radio_values &v)): Modify the constructor to also take a
	radio_values argument. Use it.
	(patch::properties::set): Change set_facecolor calls to initialize
	the color_property argument with the available radio_values.
	* graphics.h.in (color_property::color_property 
	(const octave_value& val, const radio_values &v)): Also pass a 
	radio_values argument.

	* ov-range.h (sparse_matrix_value, sparse_complex_matrix_value):
	New methods.

	* mk-pkg-add: Simplfy the autoload commands.
	* parse.y (Fautoload): Allow bare filename if file is in the same
	directory as the script from where the autoload command is run.

2007-09-25  Matthias Drochner  <m.drochner@fz-juelich.de>

	* syscalls.cc (Fpopen2): Doc fix.
	Use "sort -r" instead of "sort -nr" in test.

2007-09-18  John W. Eaton  <jwe@octave.org>

	* input.cc (input_event_hook, Finput_event_hook): Call
	command_editor::add_event_hook and
	command_editor::remove_event_hook intstead of
	command_editor::set_event_hook and
	command_editor::restore_event_hook.

2007-09-17  John W. Eaton  <jwe@octave.org>

	* version.h (OCTAVE_VERSION): Now 2.9.14+.

2007-09-17  John W. Eaton  <jwe@octave.org>

	* version.h (OCTAVE_VERSION): Now 2.9.14.
	(OCTAVE_API_VERSION): Now api-v26.
	(OCTAVE_RELEASE_DATE): Now 2007-09-17.

	* Makefile.in (DISTFILES): Add genprops.awk to the list.

	* ov-typeinfo.h (cat_op_fcn): Third arg in function pointer
	typedef is now Array<octave_idx_type> instead of Array<int>.

	* file-io.cc (do_stream_open): Use binary mode by default.

2007-09-14  Shai Ayal  <shaiay@users.sourceforge.net>

	* genprop.awk: Handle "a" modifier.

2007-09-14  Kai Habel  <kai.habel@gmx.de>

	* graphics.h.in (radio_values::contains): New function.
	(radio_values::validate): Use it.
	* graphics.cc (color_property::operator =): Call it instead of
	validate here.

2007-09-13  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/__glpk__.cc (glpk): Pass LPX_FX, not LB_DB, to
	lpx_set_col_bnds when lb[i] == ub[i].
	From: Zhi Wang <zcwang@umich.edu>.

	* graphics.h.in (colormap_property::colormap_property):
	Use jet colormap as default.

2007-09-12  John W. Eaton  <jwe@octave.org>

	* graphics.cc (text::properties::properties): Default color is
	black, not white.
	* graphics.h (text::properties::color): Declare as color_property,
	not octave_value.

	* graphics.h.in (root_figure::defaults, figure::defaults,
	axes::defaults, line::defaults, text::defaults, image::defaults,
	patch::defaults, surface::defaults): New functions.
	* gripes.cc (gripe_not_implemented): New function.
	* gripes.h: Provide decl.

2007-09-10  John W. Eaton  <jwe@octave.org>

	* genprops.awk: Add missing newline character at end of file.
	From Alexander Klimov <alserkli@inbox.ru>.

2007-09-10  David Bateman  <dbateman@free.fr>

	* data.cc (do_cat): Avoid invalid indexing into ra_idx vector.

2007-09-06  David Bateman  <dbateman@free.fr>

	* DLD-FUNCTIONS/__delaunay__.cc, DLD-FUNCRIONS/__voronoi__.cc,
	DLD-FUNCTIONS/convhulln.cc: Avoid variables in global scope

2007-09-06  David Bateman  <dbateman@free.fr>

	* Makefile.in (stamp-prereq): Add graphics.h to the dependency list.

2007-08-31  John W. Eaton  <jwe@octave.org>

	* Makefile.in (INCLUDES): Remove graphics.h from the list.
	(DISTFILES): Include graphics.h.in in the list.
	(genprops.awk): New file.
	(graphics.h): New rule.
	(distclean): Remove graphics.h.
	* graphics.h.in: New file, from graphics.h.
	(graphics_handle): Now a class instead of typedef.  Adapt all uses.
	(OCTAVE_GRAPHICS_PROPERTY_INTERNAL, OCTAVE_GRAPHICS_PROPERTY,
	OCTAVE_GRAPHICS_MUTABLE_PROPERTY): Delete macros.
	(class root_figure, class figure, class axes, class line, class
	text, class image, class patch, class surface): Use genprops.awk
	to generate property declarations and set/get functions.
	* graphics.h: Delete.
	* graphics.cc (nan_to_empty, empty_to_nan): Delete.
	(root_figure::properties::set, figure::properties::set,
	axes::properties::set, line::properties::set,
	text::properties::set, image::properties::set,
	patch::properties::set, surface::properties::set): Call
	type-specific set functions to set properties.
	(root_figure::properties::set_currentfigure,
	figure::properties::set_currentaxes, figure::properties::set_visible,
	axes::properties::get_title, axes::properties::get_xlabel,
	axes::properties::get_ylabel,  axes::properties::get_zlabel,
	axes::properties::set_title, axes::properties::set_xlabel,
	axes::properties::set_ylabel,  axes::properties::set_zlabel):
	Define custom versions.

2007-09-06  David Bateman  <dbateman@free.fr>

        * DLD-FUNCTIONS/bsxfun.cc: New function.
        * DLD-FUNCTIONS/typecast.cc: New function.
        * Makefile.in (DLD_XSRC): Add bsxfun.cc and typecast.cc.
        * ov.cc (do_cat_op): Modify use of Array<int> to 
        Array<octave_idx_type> and adjust where necessary.
        * ov.h (do_cat_op): ditto.
        * data.cc (do_cat): ditto.
        * pt-mat.cc (tree_matrix::rvalue): ditto.

2007-09-05  Michael Goffioul  <michael.goffioul@gmail.com>

	* oct-stream.cc (octave_base_stream::do_printf): Ignore precision
	portion of format string if printing Inf, NaN, or NA values.

2007-09-05  David Bateman  <dbateman@free.fr>

        * DLD-FUNCTIONS/sort.cc (mx_sort_sparse, mx_sort_sparse_indexed):
        New template classes for sparse sort functions.
        (Fsort): Use them.
        * ov.h (octave_value (const Sparse<double>&, const MatrixType&),
        octave_value (const Sparse<Complex>&, const MatrixType&)): New
        constructors.
        * ov.cc (octave_value::octave_value (const Sparse<double>&, 
        const MatrixType&), octave_value::octave_value (const 
        Sparse<Complex>&, const MatrixType&)): Define them.
        * ov-re-sparse.h (octave_sparse_matrix (const MSparse<double>&,
        const MatrixType&), octave_sparse_matrix (const Sparse<double>&),
        octave_sparse_matrix (const Sparse<double>&, const MatrixType&)):
        New constructors.
        * ov-cx-sparse.h (octave_sparse_complex_matrix (const MSparse<double>&,
        const MatrixType&), octave_sparse_complex_matrix (const 
        Sparse<double>&), octave_sparse_complex_matrix (const
        Sparse<double>&, const MatrixType&)): ditto.

2007-09-04  Gabriele Pannocchia  <g.pannocchia@ing.unipi.it>

	* DLD-FUNCTIONS/__qp__.cc (qp): Use Wact(j) == i - n_eq when
	checking for blocking constraint.  Don't resize lambda_tmp.  Only
	compute Y = Aact.pseudo_inverse () if it is needed.

2007-08-31  Michael Goffioul  <michael.goffioul@gmail.com>

	* ls-mat-ascii.cc (get_lines_and_columns): Check beg variable for
	NPOS to avoid segmentation fault. 

	* load-path.cc (load_path::do_find_file): Do not assume paths
	use forward slashes.

2007-08-30  John W. Eaton  <jwe@octave.org>

	* sysdep.cc (Fpause): Doc fix.

2007-08-30  Gabriele Pannocchia  <g.pannocchia@ing.unipi.it>

	* DLD-FUNCTIONS/__qp__.cc (qp): Resize Wact to n_act-neq, not n_act.

2007-08-29  John W. Eaton  <jwe@octave.org>

	* graphics.cc (class root_figure::properties):
	Rename from root_figure::root_figure_properties.
	(class figure::properties): Rename from figure::figure_properties.
	(class axes::properties): Rename from axes::axes_properties.
	(class line::properties): Rename from line::line_properties.
	(class text::properties): Rename from text::text_properties.
	(class image::properties): Rename from image::image_properties.
	(class patch::properties): Rename from patch::patch_properties.
	(class surface::properties): Rename from surface::surface_properties.

	* base-list.h (octave_base_list::remove): Implement our own
	remove_if function here.

2007-08-28  John W. Eaton  <jwe@octave.org>

	* graphics.h (base_properties): Move class definition before
	definition of base_graphics_object class.  Provide forward
	declaration of base_graphics_object prior to definition of
	base_properties.
 	(base_graphics_object::get_properties): New virtual function.
	(graphics_object::get_properties, root_figure::get_properties,
	figure::get_properties, axes::get_properties,
	line::get_properties, text::get_properties, image::get_properties,
	patch::get_properties, surface::get_properties): New functions.
	(root_figure::properties, figure::properties, axes::properties,
	line::properties, text::properties, image::properties,
	patch::properties, surface::properties): Data member now private.

2007-08-27  John W. Eaton  <jwe@octave.org>

	* load-path.cc (load_path::do_find_file): Also files with non
	rooted relative names.
	* load-save.cc (find_file_to_load): Likewise.  Also handle
	appending .mat to files with relative names.

	* graphics.cc (base_properties::mark_modified,
	base_properties::override_defaults,
	base_properties::delete_children, figure::get_default,
	axes::get_default): Move definitions here, from graphics.h.
	* graphics.h (class gh_manager): Move decl to end of file.

	* Cell.h (Cell::Cell (const octave_value_list&)): Create row
	vector instead of column vector.

	* pt-idx.cc (tree_index_expression::lvalue): Handle [x.a] =
	... style assignments.
	* ov-struct.cc (octave_struct::subsasgn): Handle case of RHS as
	comma-separated list.

	* ov-cell.cc (gripe_failed_assignment): New function.
	(octave_cell::subsasgn): Call gripe_failed_assignment if assign
	methods fail.

2007-08-24  David Bateman  <dbateman@free.fr>

	* symtab.cc (void symbol_table::clear (void)): If the record in
	the symbol table to clear is a user function that is a sub
	function with a static parent or if the parent is the current
	parent function, don't delete the record.
	(void symbol_table::clear_functions (void)): ditto.
	(void symbol_table::clear (const std::string&)): ditto.
	(void symbol_table::clear_function (const std::string&)): ditto.

	* graphics.h (class text): Add property color.
	* graphics.cc (text::text_properties::text_properties) ditto.
	(text::text_properties::set): ditto.
	(text::text_properties::get): ditto.
	(text::text_properties::factory_defaults): ditto.

2007-08-24  John W. Eaton  <jwe@octave.org>

	* mappers.cc (dummyp): New function.
	(install_mapper_functions): Use it to handle character data in
	finite, isinf, isna, and isnan functions.

	* load-path.cc (load_path::do_remove): Call remove_hook function
	before removing directory from list.

2007-08-24  David Bateman  <dbateman@free.fr>

        * ov-bool-sparse.cc (DEFINE_OV_TYPEID_FUNCTIONS_AND_DATA): Class
        is now logical.
        * ov-re-sparse.cc, ov-cx-sparse.cc 
        (DEFINE_OV_TYPEID_FUNCTIONS_AND_DATA): Class is now double.
        * ov-mapper.cc (octave_mapper::apply): Use is_sparse_type method
        rather than comparing class name.
        * ls-mat5.cc (save_mat5_element_length, save_mat5_binary_element):
        ditto.
        * pt-mat.cc (tree_matrix:rvalue): sparse matrices are now of class
        "double" or "logical", create new single type concat clause for
        them.
        * mex.cc (get_class_id): No longer need to special case sparse
        matrices.

        * DLD-FUNCTIONS/__delaunayn__.cc, DLD-FUNCTIONS/convhulln.cc, 
        DLD-FUNCTIONS/tsearch.cc, DLD-FUNCTIONS/__voronoi__.cc: New
        functions ported from octave-forge.
        * DLD-FUCTIONS/__dsearchn__.cc: New file.
        * DLD-FUNCTIONS/__voronoi__.cc: Return point at infinity and
        include it in the polygrons of the Voronoi diagram for
        compatibility.
        * Makefile.in: Add specific build targets for __delanayn__.cc,
        convhulln.cc and __voronoi__.cc to link to Qhull.
        (DLD_SRC): Add new functions.
        (OCTAVE_LIBS): Add QHULL_LIBS

2007-08-22  David Bateman  <dbateman@free.fr>

	* variables.cc (Fmunlock): Call munlock and not mlock.
	* symtab.cc (symbol_record::mark_as_formal_parameter): Typo. 

2007-08-10  John W. Eaton  <jwe@octave.org>

	* pt-idx.cc (tree_index_expression::get_struct_index): Improve
	error message.
	* ov-struct.cc (Fstruct, Fcell2struct): Trap invalid field names here.

2007-08-10  Peter A. Gustafson  <petegus@umich.edu>

	* graphics.h, graphics.cc (axes::axes_properties): New properties,
	xaxislocation and yaxislocation.

2007-08-10  Kai Habel  <kai.habel@gmx.de>

	* graphics.cc, graphics.h (patch): New class.
	(axes::axes_properties): New properties, clim and climmode.
	(surface::surface_properties::surface_properties): Handle patch.
	(F__go_patch__): New function.

2007-07-30  John W. Eaton  <jwe@octave.org>

	* mex.cc (mxArray_number::mxArray_number (int, const char **)):
	First arg is now mwSize.
	(max_str_len): Return mwSize value, not int.
	* mxarray.h.in (mxArray::mxArray (int, const char **)):
	First arg is now mwSize.
	(mxArray::mxArray (mxClassID, mwSize, mwSize, int, mxComplexity)):
	Third arg is now mwSize.
	(mxArray::get_string (char *, int)): Second arg is now mwSize.

2007-07-26  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/__glpk__.cc: Include glpk/glpk.h if
	HAVE_GLPK_GLPK_H is defined.

2007-07-26  David Bateman  <dbateman@free.fr>

	* pr-output.cc (Frats): Return character array with same number of
	rows as original value.

2007-07-26  John W. Eaton  <jwe@octave.org>

	* pt-bp.h (MAYBE_DO_BREAKPOINT): Rename cur_fcn to xfcn.

	* version.h (OCTAVE_VERSION): Now 2.9.13+.

2007-07-25  John W. Eaton  <jwe@octave.org>

	* version.h (OCTAVE_VERSION): Now 2.9.13
	(OCTAVE_API_VERSION): Now api-v25.
	(OCTAVE_RELEASE_DATE): Now 2007-07-25.

	* pt-bp.h (MAYBE_DO_BREAKPOINT): Rename fcn to curr_fcn.

2007-07-25  David Bateman  <dbateman@free.fr>

	* Makefile.in: 	Adjust DISTFILES to allow out of tree "make dist" 
	to work.

2007-07-24  Shai Ayal  <shaiay@users.sourceforge.net>

	* graphics.cc (color_property::operator = (const octave_value&)):
	New method.
	* graphics.h: Provide decl.

2007-07-24  Rafael Laboissiere  <rafael@debian.org>

	* oct-conf.h.in (OCTAVE_CONF_RUNTEST): Delete definition.
	* toplevel.cc (Foctave_config_info): Remove RUNTEST from struct.

2007-07-23  David Bateman  <dbateman@free.fr>

	* pr-output.cc (rat_format, rat_string_len): Global variable
	controlling behavior of rational approximation. Use throughout.
	(class pr_rational_float): New class for rational approximation of
	floats, specifically with the << operator defined.
	(std::ostream& operator << (std::ostream&, const
	pr_rational_float&)): Operator to print rational approximations of
	double values.
	(std::string rational_approx (double, int)): Function to convert a
	double value to a string of maximum length giving the rational
	approximation.
	(pr_any_float): Include the output of rational approximations.
	(Fformat): Add the "rat" format as an option.
	(Frats): New function.

2007-07-23  Aquil H. Abdullah  <aquil.abdullah@gmail.com>

	* mex.cc (mxCreateStructArray): Declare second arg as const.
	* mexproto.h (mxCreateStructArray): Ditto.

2007-07-20  David Bateman  <dbateman@free.fr>

        * zfstream.cc (BUFSIZE): Increase default buffer size to 256kB
        (gzfilebuf::underflow): Stash the last 16 characters read, so as
        to avoid calling pbackfail as much as possible.

2007-07-18  David Bateman  <dbateman@free.fr>

         * zfstream.cc (int_type gzfilebuf::pbackfail (int_type)): New
         method to putback a character when the putback position in the
         internal buffer doesn't exist.
         * zfstream.h (int_type pbackfail (int_type)): Declaration it.

2007-07-14  Michael Goffioul  <michael.goffioul@swing.be>

	* src/ov-bool-sparse.cc (octave_sparse_bool_matrix:load_hdf5):
	Use OCTAVE_LOCAL_BUFFER for temporary boolean value.
	
2007-06-27  David Bateman  <dbateman@free.fr>

	* DLD-FUNCTIONS/sparse.cc (Fspdiag): Ensure spdiag(zeros(1,0)) returns
	0x0 matrix. Doc update.
	* data.cc (Fdiag): Doc update.

2007-06-28  John W. Eaton  <jwe@octave.org>

	* ov-cell.cc (octave_cell::subsasgn): Given x = {}, convert to
	struct for assignments like x(1).f = val;

	* oct-stream.cc (octave_scan_1): New function
	(octave_scan): Use it.  Handle fmt.width.

	* graphics.h (axes::axes_properties::visible): New data member.
	* graphics.cc (axes::axes_properties::axes_properties, 
	axes::axes_properties::set, axes::axes_properties::get, 
	axes::axes_properties::factory_defaults): Handle visible property.

2007-06-27  Kai Habel  <kai.habel@gmx.de>

	* graphics.h (color_values::color_values): Arg is now std:string
	instead of char.  Call str2rgb, not c2rgb.
	* graphics.h, graphics.cc: (color_values::str2rgb): Rename from
	c2rgb.  Handle long color names, not just single char abbreviations.

2007-06-27  David Bateman  <dbateman@free.fr>
	
	* src/load-save.cc (Fsave): Ensure header is written for non
	existent file with "-append".
	* src/ls-hdf5.h: First steps towards having append work for hdf5.
	
2007-06-26  John W. Eaton  <jwe@octave.org>

	* src/load-save.cc (Fsave): Open files correctly for -append.
	Don't write file headers if appending.  Error for -append -hdf5.

2007-06-25  Olaf Till  <i7tiol@t-online.de>

	* oct-stream.h (octave_stream_list::list): Use std::map.
	(octave_stream_list::curr_len): Delete data member.
	(octave_stream_list::octave_stream_list): Fix initializations.
 
	* oct-stream.cc (octave_steam_list::do_insert,
	octave_steam_list::do_lookup, octave_steam_list::do_remove,
	octave_steam_list::do_clear, octave_steam_list::do_list_open_files, 
	octave_steam_list::do_open_file_numbers,
	octave_steam_list::do_get_file_number):
 	Use new octave_stream_list::list type.
	(octave_stream_list::do_insert): Insert octave_stream with key
	corresponding to file-descriptor.
	(octave_stream_list::do_insert, octave_steam_list::insert):
	Remove const qualifier of argument.

2007-06-18  Søren Hauberg  <hauberg@gmail.com>

        * DLD-FUNCTIONS/__lin_interpn__.cc: Replace octave_NaN with octave_NA.

2007-06-15  Shai Ayal  <shaiay@users.sourceforge.net>

	* graphics.h (OCTAVE_GRAPHICS_PROPERTY_INTERNAL,
	OCTAVE_GRAPHICS_PROPERTY, OCTAVE_GRAPHICS_MUTABLE_PROPERTY):
	New macros. Use them to declare individual properties and define
	accessor methods for each property in the property classes.

2007-06-15  Kai Habel  <kai.habel@gmx.de>

	* graphics.cc (Fget, Fset): Handle vectors of handles.

2007-06-14  John W. Eaton  <jwe@octave.org>

	* sysdep.cc (octave_popen, octave_pclose): New functions.
	* sysdep.h: Provide decls.

	* oct-procbuf.cc (procbuf::open): Use octave_popen.
	(procbuf::close): Use octave_pclose.
	* oct-prcstrm.cc (octave_oprocstream::octave_oprocstream, 
	octave_iprocstream::ictave_oprocstream): Likewise.

	* graphics.h (text::text_properties::rotation): New data member.
	* graphics.cc (text::text_properties::text_properties, 
	text::text_properties::set, text::text_properties::get, 
	text::text_properties::factory_defaults): Handle rotation property.

2007-06-14  Kai Habel  <kai.habel@gmx.de>

	* graphics.cc (color_values::c2rgb): Also accept 'k' for black.

2007-06-14  David Bateman  <dbateman@free.fr>

	* ov-ch-mat.h (idx_vector index_vector (void) const): Remove
	definition.
	* ov-ch-mat.cc (idx_vector index_vector (void) const): Move it
	here. Special case ":" case for compatibility.

2007-06-13  John W. Eaton  <jwe@octave.org>

	* ov-re-mat.cc (octave_matrix::load_ascii):
	Do a better job of handling read errors and empty matrices.
	* ov-cx-mat.cc (octave_complex_matrix::load_ascii): Likewise.
	* ov-bool-mat.cc (octave_bool_matrix::load_ascii): Likewise.
	* ov-str-mat.cc (octave_char_matrix_str::load_ascii): Likewise.

2007-06-13  David Bateman  <dbateman@free.fr>

	* DLD-FUNCTIONS/minmax.cc (MINMAX_BODY): Don't treat as single
	argument if arg2 is empty and nargin=2.

2007-06-13  Shai Ayal  <shaiay@users.sourceforge.net>

	* graphics.h, graphics.cc: Move class declarations to graphics.h.
	Move larger functions outside of class declarations in graphics.cc.

2007-06-12  Benjamin Lindner  <lindnerben@gmx.net>

	* DLD-FUNCTIONS/cellfun.cc: Use fullfile to generate filenames
	instead of assuming / will work as directory separator.

2007-06-12  David Bateman  <dbateman@free.fr>

	* DLD-FUNCTIONS/interpn.cc: Remove it.
	* DLD-FUNCTIONS/__lin_interpn__.cc: Move it. This is now a support
	function of interpn.m.
	* Makefile.in (DLD_XSRC): Remove interpn.cc and add __lin_interpn__.cc.

2007-06-07  David Bateman  <dbateman@free.fr>

	* ov-fcn-handles.cc (octave_fcn_handle::save_hdf5): More care that
	all open HDF5 descriptors are closed.
	(octave_fcn_handle::load_hdf5): Ditto.

2007-06-06  Benjamin Lindner  <lindnerben@gmx.net>

	* utils.cc [__MINGW32__]: Don't define HAVE_C99_VSNPRINTF.

2007-06-06  Michael Goffioul  <michael.goffioul@swing.be>

	* defaults.h.in, ls-hdf5.h, ov-complex.h, ov-cx-mat.h, ov-intx.h,
	sysdep.h: Sprinkle with OCTINTERP_API as needed.

2007-06-05  John W. Eaton  <jwe@octave.org>

	* help.h (raw_help): Tag with OCTINTERP_API.

	* Makefile.in (INCLUDES): Remove mxarray.h from the list so that
	it is not distributed.
	(EXTRAS): Add mxarray.h to the list so that it is installed.

	* sysdep.cc (same_file_internal) [OCTAVE_USE_WINDOWS_API]:
	Avoid leaking file handles.

2007-05-08 Michael Weitzel  <michael.weitzel@uni-siegen.de>

	* DLD-FUNCTIONS/symrcm.cc: Fix for queuing error that might cause
	an infinite loop.

2007-06-04  John W. Eaton  <jwe@octave.org>

	* data.cc (Fislogical): Rename from Fisbool.
	Make isbool an alias for islogical.

2007-06-03  David Bateman  <dbateman@free.fr>

	* Makefile.in (DISTFILES): Add mxarray.h.in
	(install-inc): Modify target so that mxarray.h is installed
	correctly.
	(distclean): Remove mxarray.h on distclean.
	* mex.cc: Use mwSize for dimensions and mwIndex for indexing
	throughout, with the exception of struct key indexes.
	* mexproto.h: ditto.
	* mxarray.h: remove.
	* mxarray.h.in: Copy here and define mwSize, mwIndex, and use
	throughout. 
	* ov-bool-mat.cc (octave_bool_matrix::as_mxArray): Use mwSize and
	mwIndex.
	* ov-bool-sparse (octave_sparse_bool_matrix::as_mxArray): ditto.
	* ov-cell.cc (octave_cell:as_mxArray): ditto.
	* ov-ch-mat.cc (octave_char_matrix:as_mxArray): ditto.
	* ov-cx-mat.cc (octave_complex_matrix::as_mxArray): ditto.
	* ov-cx-sparse.cc (octave_sparse_complex_matrix::as_mxArray): ditto.
	* ov-int.h (as_mxArray): ditto.
	* ov-range.cc (octave_range:as_mxArray): ditto.
	* ov-re-mat.cc (octave_matrix:as_mxArray): ditto.
	* ov-re-sparse.cc (octave_sparse_matrix::as_mxArray): ditto.
	* ov-struct.cc (octave_struct::as_mxArray): ditto.

2007-06-02  David Bateman  <dbateman@free.fr>

	* graphics.cc (color_property class): g++ doesn't like anonymous
	enums. Give type to color_type enum.

2007-05-31  Shai Ayal  <shaiay@users.sourceforge.net>

	* graphics.cc (radio_values, radio_property, color_values):
	New classes.
	(color_property class): Handle both color and radio values.

2007-05-31  John W. Eaton  <jwe@octave.org>

	* toplev.cc (main_loop): Improve bad_alloc error message.

	* octave.cc (execute_command_line_file, execute_eval_option_code):
	Likewise.

2007-05-31  Michael Goffioul  <michael.goffioul@swing.be>

	* toplev.cc (octave_atexit_functions):
	Now std::list instead of std::stack.
	(do_octave_atexit): Adapte to octave_atexit_functions as list.
	(Fatexit): Allow second arg of false to remove element from list.

	* DLD-FUNCTIONS/symrcm.cc: Use ! instead of "not".

	* sysdep.cc (same_file_internal) [OCTAVE_USE_WINDOWS_API]:
	Use INVALID_HANDLE_VALUE, not INVALID_FILE_HANDLE.

2007-05-28  G. D. McBain  <geordie_mcbain@yahoo.com.au>

	* ov-list.cc (append): Doc fix.

2007-05-28  John W. Eaton  <jwe@octave.org>

	* pt-loop.cc (DO_ND_LOOP): Avoid parser problem with obsolete g++.

2007-05-23  John W. Eaton  <jwe@octave.org>

	* version.h (OCTAVE_VERSION): Now 2.9.12+.

2007-05-23  John W. Eaton  <jwe@octave.org>

	* version.h (OCTAVE_VERSION): Now 2.9.12.
	(OCTAVE_RELEASE_DATE): Now 2007-05-23.

	* parse.y (make_anon_fcn_handle): Don't build assignment expression.
	* ov-fcn-handle.cc (octave_fcn_handle::print_raw):
	Don't split assignment expression.
	* ov-usr-fcn.cc (octave_user_function::do_multi_index_op):
	Special case for inline function body evaluation.

2007-05-22  John W. Eaton  <jwe@octave.org>

	* pt-fcn-handle.cc (tree_anon_fcn_handle::rvalue):
	Set current function as parent of anonymous function handle.

	* Makefile.in (uninstall): Also remove
	$(DESTDIR)$(octincludedir)/octave, $(DESTDIR)$(octincludedir),
	$(DESTDIR)$(octlibdir), and $(DESTDIR)$(octfiledir).

2007-05-22  Thomas Weber  <thomas.weber.mail@gmail.com>

	* debug.cc, error.cc, load-save.cc, oct-hist.cc, sighandlers.cc,
	symtab.cc: Fix typos.

2007-05-22  John W. Eaton  <jwe@octave.org>

	* version.h (OCTAVE_VERSION): Now 2.9.11+.

2007-05-22  John W. Eaton  <jwe@octave.org>

	* version.h (OCTAVE_VERSION): Now 2.9.11.
	(OCTAVE_API_VERSION): Now api-v24.
	(OCTAVE_RELEASE_DATE): Now 2007-05-22.

2007-05-21  David Bateman  <dbateman@free.fr>

	* debug.cc (Fdbstop): handle integer, vector and multiple line
	arguments.
	(Fdbclar): ditto. Eliminate extraneous debugging messages.

2007-05-21  Søren Hauberg  <hauberg@gmail.com>

        * load-path.cc (Fpath, Frehash): Replace "LOADPATH" with "load
	path" in doc strings.

        * parse.y (Feval): Add example.

2007-05-21  David Bateman  <dbateman@free.fr>

	* error.cc (rethrow_error): Use NULL name so that "error:" is not
	added to the message
	(Frethrow): Correctly treat empty stack.

2007-05-21  John W. Eaton  <jwe@octave.org>

	* oct-map.h (Octave_map::numel): Return number of array elements,
	not number of map elements.
	(Octave_map::nfields): Rename from length.
	(Octave_map::empty): Delete.
	Change all uses of empty to check nfields () == 0 instead.

2007-05-21  Søren Hauberg  <soren@hauberg.org>

	* help.cc (Fautoload): Doc fix.
	* variables.cc (Fiscommand): Doc fix.

2007-05-19  David Bateman  <dbatemna@free.fr>

	* ov-fcn-inline.cc (Fvectorize): Doc fix.

2007-05-16  Søren Hauberg  <soren@hauberg.org>

	* ov.cc (Fsubsref, Fsubsasgn): Doc fix.

2007-05-16  John W. Eaton  <jwe@octave.org>

	* load-path.h (load_path::sys_path): New static data member.
	(load_path::system_path): New static function.
	(load_path::do_system_path): New function.
	* load-path.cc (Vsystem_path): Delete.
	(load_path::do_initialize): Use sys_path, not Vsystem_path.
	(Fpathdef): Call load_path::system_path instead of using Vsystem_path.
	* ls-mat5.cc (read_mat5_binary_element): Likewise.
	* ov-fcn-handle.cc (octave_fcn_handle::set_fcn): Likewise.

2007-05-16  David Bateman  <dbateman@free.fr>

	* load_pathc.cc (std::string octave_system_path (void)): New
        function.
        * load-path.h (std::string octave_system_path (void)): Declare it.

        * load-save.cc (static load_save_format get_file_format
        (std::istream&, const std::string&)): Add filename argument, and
        pass it to read_mat5_binary_header. Use new format throughout file.
        (Fload): Don't allow gzip of matlab v7 files as the files
        themselves include compression.

        * ls-mat5.cc (arrayclsstype:MAT_FILE_WORKSPACE_CLASS): New class
        type.
        (read_mat5_binary_element): Workspaces, don't have dimensions, so
        don't read them. Implement read of matlab objects, but only use
        them for inline functions. Implement reading of function and
        workspace classes.
        (read_mat5_binary_header): Add filename argument. Read sub-system
        specific data block given as an offset in bytes 117 to 124.
        (save_mat5_binary_element): Include saving of inline functions.

        * ls-mat5.h (read_mat5_binary_header): Include filename.

        * ov-fcn-handle.cc (octave_fcn_handle_save_ascii,
        octave_fcn_handle::load_ascii, octave_fcn_handle::save_binary, 
        octave_fcn_handle::load_binary, octave_fcn_handle::save_hdf5, 
        octave_fcn_handle::load_hdf5): Save and reload the local symbol
        table of the user function associated with anonymous function
        handles. Save and load the absolute path and the exec_prefix for
        normal function handles and use then to find equivalent functions
        between different installations of Octave. Attempt to maintain
        backward and forward compatibility.
        (Ffunctions): Additional outputs, including the workspace of
        anonymous functions, and more compatiable outputs.

        * ov-fcn-handle.h (user_function_value): Expose the user function
        value of a function handle.

        * ov-fcn-inline.cc (Octave_map octave_fcn_inline::map_value
        (void) const): Return a structure compatiable with matlab's class
        implementation of inline functions.

        * ov-fcn-inline.h (map_value): Declare it.

2007-05-14  Bob Weigel  <rweigel@gmu.edu>

	* DLD-FUNCTIONS/svd.cc: Doc fix.

2007-05-14  Thomas Weber  <thomas.weber.mail@gmail.com>

	* DLD-FUNCTIONS/fft.cc (do_fft): Handle empty matrices.  New tests.

2007-05-14  Søren Hauberg  <soren@hauberg.org>

	* toplev.cc (Fatexit): Simplify example in doc string.
	* help.cc (Flookfor): Doc fix.
	* DLD-FUNCTIONS/cellfun.cc (Fcellfun):
	Reformat to avoid long lines in doc string example.

2007-05-13  Søren Hauberg  <soren@hauberg.org>

	* toplev.cc (Fquit): Doc fix.
	* help.cc (Fhelp): Doc fix.
	* oct-hist.cc (Fsaving_history): Doc fix.

2007-05-11  John W. Eaton  <jwe@octave.org>

	* variables.cc (symbol_out_of_date):
	Always check for files that have gone missing.

2007-05-08 Michael Weitzel  <michael.weitzel@uni-siegen.de>

	* DLD-FUNCTIONS/symrcm.cc: New function for Reverse Cuthill-McKee
	permutation.
	
2007-05-07  John W. Eaton  <jwe@octave.org>

	* oct-map.cc (Octave_map::resize): Handle case of no keys.
	(keys_ok): Rename from equiv_keys.  Return value is now status.
	Pass key names as string_vector reference arg.
	(Octave_map::assign (const octave_value_list&, const Octave_map&)):
	Call keys_ok, not equiv_keys.  Handle case of no keys.

2007-04-30  John W. Eaton  <jwe@octave.org>

	* Makefile.in (%.df : %.cc): Use mv instead of
	$(simple-move-if-change-rule).

2007-04-30  David Bateman  <dbateman@free.fr>
	
	* pt-loop.cc (DO_ND_LOOP): New args, CONV and ARG.
	Use octave_idx_type instead of int for loop counters.
	Remove redundant assignments to variable "quit" as it is
	always defined in quit_loop_now.
	Special case rows = 0 and 1 cases in loops over arrays.
	Include some of the code that was separate from the macro
	DO_ND_LOOP in the macro itself
	(tree_simple_for_command::eval): USE DO_ND_LOOP for all loops.
	(DO_LOOP): Delete.

2007-04-30  John W. Eaton  <jwe@octave.org>

	* mex.cc (mex::cleanup): Don't call unmark for elements of the set
	inside the loop.  From Laurent Mazet <laurent.mazet@motorola.com>.

2007-04-30  David Bateman  <dbateman@free.fr>

	* OPERATORS/op-int-conv.cc: Define sq and dq string conversion
	operators.  Delete old char_matrix_str conversions.
	* OPERATORS/op-int-conv.cc (install_int_conv_ops): Install them.

2007-04-27  Benjamin Lindner  <lindnerb@users.sourceforge.net>.

	* octave.cc (execute_startup_files): Call same_file to check for
	already executed init files.

2007-04-27  John W. Eaton  <jwe@octave.org>

	* sysdep.cc (same_file_internal): New function.  POSIX code
	from same_file in utils.cc.  Windows code from
	Benjamin Lindner  <lindnerb@users.sourceforge.net>.
	Don't canonicalize file names.
	Also return false if stat calls fail.
	* sysdep.h: Provide decl.
	* utils.cc (same_file): Use same_file_internal.

2007-04-27  David Bateman  <dbateman@free.fr>

	* graphic.cc (get_property_form_handle, set_property_in_handle):
	New functions.
	* grahics.h: New file.
	* mex.cc (mexGet, mexSet): use the above to implement mexGet
	and mexSet.
	* Makefile.in (INCLUDES): Add graphics.h

2007-04-26  John W. Eaton  <jwe@octave.org>

	* ov-usr-fcn.cc (octave_user_function::do_multi_index_op):
	Only deal with varargout if ret_list->takes_varargs () is true.

2007-04-26  Søren Hauberg  <soren@hauberg.org>

	* DLD-FUNCTIONS/urlwrite.cc: Doc fix.

2007-04-26  David Bateman  <dbateman@free.fr>

	* pt-loop.cc (tree_simple_for_command::eval (void)): Correct
	reshaping of dim_vector in for loop for multi-dimensional array.

2007-04-26  John W. Eaton  <jwe@octave.org>

	* load-save.cc (find_file_to_load): Only consider regular files.

2007-04-24  Shai Ayal  <shaiay@users.sourceforge.net>

	* graphics.cc (color_property): Eliminate alpha channel from RGB
	color spec.

2007-04-23  Shai Ayal  <shaiay@users.sourceforge.net>

	* src/graphics.cc (color_property::color_property):
	New arg A for alpha channel.  Set rgba instead of red, green, blue.
        (color_property::color_property (char)): New constructor.
        (color_propery::rgba): New data member.
	(color_property::red, color_property::green, color_property::blue):
	Delete.
        (color_property::validate): Use rgba.
        (color_property::c2rgba): New function.

2007-04-23  Søren Hauberg  <soren@hauberg.org>

	* data.cc (Fsize_equal): Allow more than two arguments.

2007-04-20  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/__gnuplot_raw__.l (deftypefn): 
	(Vautomatic_replot): Delete static variable.
	(Fautomatic_replot): Delete function.

	* toplev.cc (Fcasesen): Delete obsolete function.

	* DLD-FUNCTIONS/__gnuplot_raw__.l (gnuplot::makeplot): Check
	whether caller is "splot", not "gsplot".

2007-04-19  John W. Eaton  <jwe@octave.org>

	* lex.l (is_keyword_token): Delete all_va_args_kw switch case.
	* octave.gperf: Remove all_va_args_kw from the list.

2007-04-19  Daniel J. Sebald  <daniel.sebald@ieee.org>

	* syscalls.cc: Fix popen2 test to stop trying after 100 times
	throught the loop.

2007-04-16  Geordie McBain  <geordie.mcbain@aeromech.usyd.edu.au>

	* ov-fcn-inline.cc (Fargnames): Doc fix.

2007-04-13  Geordie McBain  <geordie.mcbain@aeromech.usyd.edu.au>

	* DLD-FUNCTIONS/find.cc (Ffind): Doc fix.

2007-04-11  John W. Eaton  <jwe@octave.org>

	* Makefile.in (DOCSTRINGS): Don't echo commands.
	(doc-files): Use mv, not $(simple-move-if-change-rule).

	* data.cc (Fnot, Fuplus, Fuminus, Ftranspose, Fctranspose, Fplus,
	Fminus, Fmtimes, Fmrdivide, Fmpower, Fmldivide, Flt, Fle, Feq,
	Fge, Fgt, Fne, Ftimes, Frdivide, Fpower, Fldivide, Fand, For):
	New functions.

2007-04-09  John W. Eaton  <jwe@octave.org>

	* graphics.cc (line::line_properties::markeredgecolor,
	line::line_properties::markerfacecolor): New properties.

2007-04-06  John W. Eaton  <jwe@octave.org>

	* data.cc (F__vnorm__): New function.

	* pt-fcn-handle.cc (tree_anon_fcn_handle::param_list, 
	tree_anon_fcn_handle::cmd_list,	tree_anon_fcn_handle::ret_list,
	tree_anon_fcn_handle::sym_tab): Delete.  Remove all uses.
	(tree_anon_fcn_handle::fcn): New data member.
	(tree_anon_fcn_handle::tree_anon_fcn_handle): Initialize it.
	(tree_anon_fcn_handle::rvalue, tree_anon_fcn_handle::dup):
	Extract parameter list, return list, function body, and symbol
	table from fcn.
	(tree_anon_fcn_handle::parameter_list, tree_anon_fcn_handle::body):
	Forward request to fcn.

	* ov-usr-fcn.h (octave_user_function::local_sym_tab): Rename from
	sym_tab.  Change all uses.
	(octave_user_function::sym_tab): New function.

	* octave.cc (execute_command_line_file): 

2007-04-05  David Bateman  <dbateman@free.fr>

	* DLD-FUNCTIONS/regexp.cc (Fregexprep): Correct iteration over 
	cell arrays so that the source and pattern are iterated seperately 
	in the same manner as matlab.

2007-04-05  Laurent Mazet  <laurent.mazet@motorola.com>

	* mex.cc (mxArray_octave_value::get_string): Copy nel elements,
	not buflen elements.

2007-04-05  John W. Eaton  <jwe@octave.org>

	* oct-stream.cc (DO_DOUBLE_CONV): New macro.
	(do_printf): Use it.

2007-04-04  John W. Eaton  <jwe@octave.org>

	* input.cc (octave_yes_or_no): Force interactive_input to use readline.

	* octave.cc (execute_eval_option_code): Catch bad::alloc here.

2007-04-03  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/urlwrite.cc (Furlwrite, Furlread):
	Use && for logical, not &.

	* DLD-FUNCTIONS/qr.cc (Fqr): Clarify nargin check.

	* error.cc (Frethrow): Add braces to avoid ambiguous if/else.
	* oct-stream.cc (octave_scan<>): Likewise.
	* DLD-FUNCTIONS/colamd.cc (Fetree): Likewise.
	* DLD-FUNCTIONS/sort.cc (mx_sort, mx_sort_indexed): Likewise.

	* ov-fcn-handle.cc (make_fcn_handle): Pass ultimate parent
	function name to lookup_function.

2007-03-29  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/filter.cc (filter): Fix typo in doc string.
	From Utkarsh Upadhyay <musically.ut@gmail.com>.

2007-03-28  Rafael Laboissiere  <rafael@debian.org>

	* DLD-FUNCTIONS/__glpk__.cc: Fix #ifdef logic around GLPK_PRE_4_14.

2007-03-27  John W. Eaton  <jwe@octave.org>

	* version.h (OCTAVE_VERSION): Now 2.9.10+

2007-03-27  John W. Eaton  <jwe@octave.org>

	* version.h (OCTAVE_VERSION): Now 2.9.10.
	(OCTAVE_API_VERSION): Now api-v23.
	(OCTAVE_RELEASE_DATE): Now 2007-03-27.

	* version.h (OCTAVE_COPYRIGHT): Update for 2007.  Add "and others".
	(X_OCTAVE_WARRANTY_STATEMENT): Rename from OCTAVE_WARRANTY_STATEMENT.
	Accept arg.
	(OCTAVE_WARRANTY_STATEMENT): Define using X_OCTAVE_WARRANTY_STATEMENT.
	(OCTAVE_NAME_VERSION_COPYRIGHT_COPYING_AND_WARRANTY): Define using
	X_OCTAVE_NAME_VERSION_COPYRIGHT_COPYING_AND_WARRANTY.
	(X_OCTAVE_NAME_VERSION_COPYRIGHT_COPYING_AND_WARRANTY): Rename
	from OCTAVE_NAME_VERSION_COPYRIGHT_COPYING_AND_WARRANTY.
	Accept arg and pass to X_OCTAVE_WARRANTY_STATEMENT.
	(OCTAVE_NAME_VERSION_COPYRIGHT_COPYING_AND_WARRANTY): Pass empty
	arg to X_OCTAVE_NAME_VERSION_COPYRIGHT_COPYING_WARRANTY_AND_BUGS.
	(OCTAVE_STARTUP_MESSAGE): Put info about news last.

2007-03-27  John W. Eaton  <jwe@octave.org>

	* Makefile.in (dist, conf-dist): Use ln, not $(LN_S).

2007-03-26  Juhani Saastamoinen  <juhani@cs.joensuu.fi>

	* file-io.cc (do_stream_open) [! HAVE_ZLIB]:
	Call fopen with mode, not tmode.

2007-03-26  John W. Eaton  <jwe@octave.org>

	* OPERATORS/op-str-str.cc (DEFCHARNDBINOP): New macro.  Use it to
	define functions for eq and ne ops.  Also define lt, le, ge, and
	gt ops.

2007-03-23  David Bateman  <dbateman@free.fr>

	* DLD-FUNCTIONS/rand.cc: Make more statistical tests optional.

2007-03-23  John W. Eaton  <jwe@octave.org>

	* bitfcns.cc (Fbitshift): Error if third argument is not a scalar.

2007-03-23  David Bateman  <dbateman@free.fr>

	* DLD-FUNCTIONS/rand.cc: Make statistical tests optional and 
	add tests for fixed sequences.

2007-03-22  John W. Eaton  <jwe@octave.org>

	* graphics.cc (base_graphics_object::mark_modified):
	New virtual function.
	(base_properties::__modified__): New data member.
	(base_properties::base_properties): Initialize it.
	(graphics_object::mark_modified, base_properties::mark_modified, 
	root_figure::mark_modified, figure::mark_modified,
	axes::mark_modified, line::mark_modified, text::mark_modified,
	image::mark_modified, surface::mark_modified,
	root_figure::root_figure_properties::mark_modified,
	figure::figure_properties::mark_modified,
	axes::axes_properties::mark_modified,
	line::line_properties::mark_modified,
	text::text_properties::mark_modified,
	image::image_properties::mark_modified,
	surface::surface_properties::mark_modified): New functions.
	(figure::figure_properties::set, figure::figure_properties::get,
	axes::axes_properties::set, axes::axes_properties::get,
	line::line_properties::set, line::line_properties::get,
	text::text_properties::set, text::text_properties::get,
	image::image_properties::set, image::image_properties::get,
	surface::surface_properties::set, surface::surface_properties::get):
	Handle __modified__ property.

	* parse.y (Fautoload): Use warning_with_id.

2007-03-21  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/__qp__.cc (ABS): Delete.  Use std::abs instead.
	(null): Set elements of retval with magnitudes less than eps to 0.

	* error.cc (Fwarning): Allow setting options with struct.
	If setting options, return previous state.

	* graphics.cc (axes::axes_properties::set_defaults): Reverse sense
	of mode test for setting outerposition property.
	(figure::figure_properties::set): If setting visible property,
	make this figure the current figure.
	(gh_manager::figure_handle_list, gh_manager::do_figure_handle_list):
	New functions.
	(F__go_figure_handles__): New function.

	* sysdep.cc (Fpause): Also call drawnow if nargin == 0.
	(Fkbhit, Fsleep, Fusleep): Also call drawnow here.

2007-03-20  David Bateman  <dbateman@free.fr>

	* DLD-FUNCTIONS/cellfun.cc (Fcellfun): Correct for shape of return
	matrix for the case of UniformOutput being true.

2007-03-20  John W. Eaton  <jwe@octave.org>

	* sysdep.cc (Fpause): Call drawnow.

2007-03-15  John W. Eaton  <jwe@octave.org>

	* variables.cc (F__lock_global__): Delete.

2007-03-14  John W. Eaton  <jwe@octave.org>

	* graphics.cc: New file.
	* Makefile.in (DIST_SRC): Add it to the list.

2007-03-13  John W. Eaton  <jwe@octave.org>

	* file-io.cc (do_stream_open): Use binary mode if 't' is not
	specified in mode, but don't add 'b' if it is already present.

2007-03-09  David Bateman  <dbateman@free.fr>

	* data.cc (do_cat):  Also treat case of all empty matrices.

2007-03-08  John W. Eaton  <jwe@octave.org>

	* mex.cc (mxArray_octave_value::set_dimensions,
	mxArray_octave_value::set_m, mxArray_octave_value::set_n, 
	mxArray_octave_value::set_class_name,
	mxArray_octave_value::set_ir, mxArray_octave_value::set_jc,
	mxArray_octave_value::remove_field,
	mxArray_octave_value::set_field_by_number):
	Don't panic; request mutation instead.
	(class mxArray_octave_value): 

	* mxarray.h (mxArray::set_m, mxArray::set_n,
	mxArray::set_dimensions): Wrap method call call with
	DO_VOID_MUTABLE_METHOD.

2007-03-08  David Bateman  <dbateman@free.fr>

	* data.cc (do_cat): Ignore leading empty matrices.

2007-03-07  Bob Weigel  <rweigel@gmu.edu>

	* urlwrite.cc (urlget): Allow URL redirects.

2007-03-05  David Bateman  <dbateman@free.fr>

	* DLD-FUNCTIONS/md5sum.cc (Fmd5sum): Treat both files and strings.

2007-03-05  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/__glpk__.cc (F__glpk__): Check GLPK_PRE_4_14, not
	GLPK_PRE_4_15.

2007-03-02  John W. Eaton  <jwe@octave.org>

	* parse.y (Fautoload): Undo previous change.
	Warn if FILE is not an absolute file name.

	* utils.cc (make_absolute): Make copy of arg before 

2007-03-01  John W. Eaton  <jwe@octave.org>

	* ov-base-mat.h	(octave_base_matrix::octave_base_matrix (const MT&), 
	(octave_base_matrix::octave_base_matrix (const MT&, const MatrixType&)):
	Use common definition with default argument value.
	* ov-base-scalar.h (octave_base_scalar<T>::typ): New data member.
	Initialize in constructors.
	(octave_base_scalar<T>::matrix_type): New funtions.

2007-03-01  David Bateman  <dbateman@free.fr>

        * DLD-FUNCTIONS/md5sum.cc: New file.
        * Makefile.in (DLD_XSRC): Add md5sum.cc

2007-03-01  Olli Saarela  <Olli.Saarela@kcl.fi>

	* input.cc (FPS1): Fix @seealso.

2007-03-01  David Bateman  <dbateman@free.fr>

	* DLD-FUNCTIONS/matrix_type.cc (Fmatrix_type): Special case scalar
	types and always return "Full" matrix type.

2007-02-28  John W. Eaton  <jwe@octave.org>

	* input.cc (interactive_input): Also call flush_octave_stdout
	after calling drawnow.

2007-02-27  John W. Eaton  <jwe@octave.org>

	* Makefile.in (uninstall): Also remove octave-$(version)$(EXEEXT), 
	oct-gperf.h, and all .oct files.  Remove PKG_ADD.
	From Thomas Treichl <Thomas.Treichl@gmx.net>.

	* load-path.h, (load_path::initialize, load_path::do_initialize):
	New arg, set_initial_path.
	* load-path.cc (load_path::do_initialize): Don't add system
	directories to apth if set_initial_path is false.
	* octave.cc (set_initial_path): New static variable.
	(NO_INITIAL_PATH_OPTION): New define.
	(usage_string): Include --no-initial-path in message.
	(long_opts): Include no-initial-path/NO_INITIAL_PATH_OPTION here.
	(octave_main): Handle NO_INITIAL_PATH_OPTION.
	Pass set_initial_path to load_path::initialize.

	* parse.y (Fautoload): Warn about duplicate entries.  Only insert
	the first found.

2007-02-27  David Bateman  <dbateman@free.fr>

	* error.cc (Vlast_error_file, Vlast_erro_name, Vlast_error_row,
	Vlast_error_column): New static variables.
	(verror): Use them to store the location of last error.
	(rethrow_error, Frethrow, Flasterror): New functions.

        * DLD-FUNCTIONS/regexp.cc (octcellregexp): Wrapper to octregexp
        function for cases when string or pattern are cell arrays
        (Fregexp, Fregexpi): Use them.
        (octregexprep): New function with functionality of old Fregexprep.
        (Fregexprep): Treat cell arguments.

2007-02-26  Michael Goffioul  <michael.goffioul@swing.be>

	* Makefile.in: Use $(LN_S) instead of ln or ln -s.

	* DLD-FUNCTIONS/getrusage.cc: Undef min after including windows.h.

2007-02-25  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/interpn.cc: Include dNDArray.h, not dMatrix.cc.

	* error.h: Include cstdarg.
	From Juhani Saastamoinen <juhani@cs.joensuu.fi>.

2007-02-23  John W. Eaton  <jwe@octave.org>

	* variables.cc (lookup_function): Don't dereference NULL
	curr_parent_fucntion pointer.
	* ov-fcn-handle.cc (make_fcn_handle): Call lookup_function with
	parent set to empty string if call_stack is empty.

	* DLD-FUNCTIONS/dispatch.cc (Fbuiltin): Also adjust argument list
	in calls to functions that are not overloaded.  Call
	lookup_by_name to find function instead of searching fbi_symtab
	directly.

	* help.cc (do_which): Return empty string if file not found.

2007-02-22  John W. Eaton  <jwe@octave.org>

	* mex.cc (mxArray_cell::mxArray_cell (const mxArray_cell&),
	mxArray_struct::mxArray_struct (const mxArray_struct&)):
	Avoid calling clone on 0 elements.

	* variables.cc (symbol_out_of_date): If checking nested function,
	look for file using parent function name.  Delete unused
	variable NAMES.

	* src/oct-stream.cc (octave_stream::do_gets): If no limit or not
	at limit, read ahead one character at end of line for compatibility.
	(octave_stream::gets, octave_stream::getl): Set max_len to -1 if
	tc_max_len is not defined.
	* file-io.cc (Ffgets, Ffgetl): If no limit specified, pass
	undefined octave_value object as max_len in call to
	octave_stream::gets.

2007-02-21  John W. Eaton  <jwe@octave.org>

	* mex.cc (mexErrMsgIdAndTxt, mexWarnMsgIdAndTxt): Handle second
	arg as format and accept variable number of arguments.
	* mexproto.h: Fix decls.

	* error.h, error.cc (vmessage, vusage, vwarning, verror,
	vparse_error, vmessage_with_id, vusage_with_id, vwarning_with_id,
	verror_with_id, vparse_error_with_id): Provide va_list versions of
	variadic error, warning, and message functions.
	(message, usage, warning, error, parse_error, message_with_id,
	usage_with_id, warning_with_id, error_with_id,
	parse_error_with_id): Call va_list versions.

	* DLD-FUNCTIONS/urlwrite.cc (Furlwrite, Furlread): Return error
	code and message instead of throwing error if functionality is
	missing.

	* oct-obj.h (octave_value_list::splice): Set default for
	replacement list argument.

2007-02-20  Rafael Laboissiere  <rafael@laboissiere.net>

	* DLD-FUNCTIONS/__glpk__.cc: Adapt code for changes in the GLPK
	API for version 4.15 or later.

2007-02-20  John W. Eaton  <jwe@octave.org>

	* mxarray.h (mxArray::get_scalar): New function.
	* mex.cc (mxArray_base::get_scalar): New pure virtual function.
	(mxArray_octave_value::get_scalar, mxArray_matlab::get_scalar,
	mxArray_number::get_scalar): New functions.
	(mxGetScalar): Call get_scalar here.

	* mex.cc (mxArray_octave_value::get_dimensions): Cache ndims here.
	(mxArray_octave_value::get_number_of_dimensions):
	Call get_dimensions here to cache both ndims and dims.

2007-02-17  John W. Eaton  <jwe@octave.org>

	* variables.cc (symbol_out_of_date): Don't exit early if looking
	at nested function.

2007-02-16  John W. Eaton  <jwe@octave.org>

	* dynamic-ld.cc (octave_dynamic_loader::do_load_oct): Clear
	function if original was loaded from relative path and the name
	can no longer be found in path.  Mark files found from relative
	path as relative.
	(clear): Only warn if there is more than one function to clear.

	* variables.cc (symbol_out_of_date): Don't ignore return value in
	call to octave_env::make_absolute.
	(symbol_out_of_date): Clear symbol if original was loaded from
	relative path and name can no longer be found in path.

	* dynamic-ld.cc (octave_dynamic_loader::do_load_oct):
	Also check whether new file is same as the old file.

	* utils.cc (same_file): Move here from variables.cc.
	* utils.h: (same_file): Provide decl.

	* parse.y (frob_function): Stash parent function name if parsing
	nested function.

	* ov-fcn-handle.cc (make_fcn_handle): Pass current function name
	as parent in call to lookup_function.

	* ov-fcn.h (octave_function::parent_fcn_name): New virtual function.

	* ov-usr-fcn.h (octave_user_function::parent_name): New data member.
	(octave_user_function::stash_parent_function_name,
	octave_user_function::parent_function_name): New methods.
	* ov-usr-fcn.cc (octave_user_function::octave_user_function):
	Initialize parent_name.

	* variables.h, variables.cc (lookup_function): New arg, parent.
	If not empty, try this first for lookup.

	* dynamic-ld.cc (octave_dynamic_loader::do_load_mex): If doing
	path lookup, check relative status. Pass relative to oct_file_in_path.
	(octave_dynamic_loader::do_load_mex): Likewise, for mex_file_in_path

	* defun-int.h, defun.cc (install_mex_function): New arg, relative.
	(install_dld_function): Likewise.
	(octave_dld_fcn_installer): Likewise.
	(DEFINE_FUNX_INSTALLER_FUN3): Pass relative to install_dld_function.

	* dynamic-ld.h (octave_dynamic_loader::load_oct,
	octave_dynamic_loader::load_mex,
	octave_dynamic_loader::do_load_oct
	octave_dynamic_loader::do_load_mex): New arg, relative.

	* dirfns.h (Vcurrent_directory): Delete unused variable.

	* variables.cc (symbol_out_of_date): Also compare function time
	stamp to Vlast_chdir_time if function is from relative lookup.

	* dirfns.cc (Vlast_chdir_time): New variable.
	(octave_change_to_directory): Update it if chdir succeeds.
	* dirfns.h (Vlast_chdir_time): Provide decl.

	* ov-fcn.h (octave_function::relative): New data member.
	(octave_function::mark_relative, octave_function::is_relative):
	New functions.

	* parse.y (fcn_file_from_relative_lookup): New static variable.
	(load_fcn_from_file): Note whether function file was found from
	relative path element.
	(frob_function): Maybe mark function as relative.

	* parse.y (lookup_autoload): Don't call octave_env::make_absolute
	on return value.
	* variables.cc (symbol_out_of_date): Make name absolute after call
	to lookup_autoload.

	* input.cc (interactive_input): New arg, DEBUG.  Don't call
	drawnow if debugging.
	(get_user_input): Pass DEBUG to interactive_input.

2007-02-16  Michael Goffioul  <michael.goffioul@swing.be>

	* syscalls.cc (Fpopen2): New function.
	(pipe): Modify to return input and output file descriptor
	seperately rather than in a list.
	
2007-02-16  Muthiah Annamalai  <muthuspost@gmail.com>

	* debug.cc (Fdbtype): Improve compatibility.

2007-02-16  John W. Eaton  <jwe@octave.org>

	* toplev.cc (wait_for_input): New function.
	(run_command_and_return_output): Use it instead of napping.

	* oct-procbuf.h (octave_procbuf::pid): Now const.
	* procstream.h (procstreambase::pid): Now const.
	(procstreambase::file_number): New function.

2007-02-15  John W. Eaton  <jwe@octave.org>

	* mxarray.h (mxChar): Use char instead of unsigned short.

	* toplev.cc (Foctave_config_info): Remove
	OCTAVE_CONF_MKOCTFILE_INCFLAGS and OCTAVE_CONF_MKOCTFILE_LFLAGS
	from the list.
	* oct-conf.h.in (OCTAVE_CONF_MKOCTFILE_INCFLAGS,
	OCTAVE_CONF_MKOCTFILE_LFLAGS): Don't define.
	(OCTAVE_CONF_INCLUDEDIR, OCTAVE_CONF_OCTINCLUDEDIR,
	OCTAVE_CONF_OCTLIBDIR, OCTAVE_CONF_PREFIX): New definitions.

2007-02-14  Alexander Barth  <abarth@marine.usf.edu>

	* DLD-FUNCTIONS/interpn.cc: New file.
	* Makefile.in (DLD_XSRC): Add it to the list.

2007-02-14  John W. Eaton  <jwe@octave.org>

	* input.cc (interactive_input): Check error_state after call to feval.

2007-02-10  John W. Eaton  <jwe@octave.org>

	* oct-stream.cc (octave_stream::rewind): Call seek (0, SEEK_SET)
	instead of rep->rewind.
	(octave_base_stream::rewind): Delete Function.
	* oct-stream.h (octave_base_stream::rewind): Delete decl.

2007-02-09  John W. Eaton  <jwe@octave.org>

	* ls-mat5.cc (PAD): Adjust to change in write_mat5_tag.
	(TAGLENGTH): Delete unused macro.

	* ov-struct.cc (octave_struct::load_ascii,
	octave_struct::load_binary, octave_struct::load_hdf5):
	Delete obsolete attempt at backward compatibility.

	* ls-mat5.cc (read_mat5_binary_element): Don't attempt to read
	fieldnames if there are no fields.
	(write_mat5_tag): Don't use small data element format if bytes == 0.

	* ls-mat5.cc (read_mat5_binary_element): Always create a structure
	that is at least 1x1.
	* ov-struct.cc (octave_struct::load_ascii,
	octave_struct::load_binary, octave_struct::load_hdf5): Likewise.

2007-02-08  Michael Goffioul  <michael.goffioul@swing.be>

	* DLD-FUNCTIONS/__glpk__.cc: Include glplib.h if glpk.h does not.
	Provide definitions for lib_set_fault_hook and lib_set_print_hook
	if they are not defined.

2007-02-07  John W. Eaton  <jwe@octave.org>

	* defaults.cc (subst_octave_home):
	Only substitute at beginning of string.

	* ls-hdf5.cc (save_hdf5_empty): Use OCTAVE_LOCAL_BUFFER.
	* ov-bool-mat.cc (octave_bool_matrix::save_hdf5,
	octave_bool_matrix::load_hdf5): Likewise.
	* ov-bool-sparse.cc (octave_sparse_bool_matrix::save_hdf5,
	octave_sparse_bool_matrix::load_hdf5): Likewise.

2007-02-07  Michael Goffioul  <michael.goffioul@swing.be>

	* ov-cell.cc (octave_cell::save_hdf5): Correct test for H5Dwrite
	return value.

2007-02-07  John W. Eaton  <jwe@octave.org>

	* zfstream.cc (gzfilebuf::open_mode): Always append "b" to c_mode.

	* toplev.cc (Foctave_config_info): Use struct for conf_info.
	Call subst_octave_home on selected values when initializing map.
	* defaults.cc (subst_octave_home): Now extern.
	* defaults.h.in: Provide decl.

2007-02-05  John W. Eaton  <jwe@octave.org>

	* mex.cc (mxArray_number::as_octave_value): Fake mxSINGLE_CLASS
	by returning double-precision values.
	(mxArray_sparse::as_octave_value): Clarify error message.

	* ov-complex.h (octave_complex): Use std instead of OCTAVE_STD
	since the latter was unconditionally defined to be std anyway.

2007-02-05  Michael Goffioul  <michael.goffioul@swing.be>

	* ov-complex.h: Tag octave_complex class with OCTINTERP_API.

2007-01-30  John W. Eaton  <jwe@octave.org>

	* Merge of changes from graphics-branch:

	2007-01-26  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/__contourc__.cc: New function.
	* Makefile.in (DLD_XSRC): Add it to the list.

	2007-01-25  John W. Eaton  <jwe@octave.org>

	* input.cc (Vdrawnow_requested): New static variable.
	(F__request_drawnow__): New function.
	(interactive_input): New function.  Update Vlast_prompt_time here.
	(octave_gets, get_user_input, octave_yes_or_no):
	Call interactive_input instead of gnu_readline.

	* symtab.h (symbol_record::eternal): New data member.
	(symbol_record::symbol_record): Initialize it.
	(symbol_record::make_eternal, symbol_record::is_eternal):
	Don't forward to symbol_def functions.
	(symbol_record::symbol_def::make_eternal,
	symbol_record::symbol_def::is_eternal): Delete.
	(symbol_record::symbol_def::eternal): Delete data member.
	(symbol_record::symbol_def::symbol_def): Delete initialization.

	* pt-id.cc (tree_identifier::link_to_global): Include variable
	name in warning message.

	* variables.cc (F__lock_global__): New function.

	2007-01-11  John W. Eaton  <jwe@octave.org>

	* ls-oct-ascii.cc (save_ascii_data_for_plotting, save_three_d):
	Set precision to 6 instead of 4.

2007-01-29  David Bateman  <dbateman@free.fr>

	* DLD-FUNCTIONS/find.cc (find_nonzero_elem_idx): Don't panic if
	nargout is greater than 3.

2007-01-24  John W. Eaton  <jwe@octave.org>

	* pt-assgn.cc (former_built_in_variables): New static data.
	(maybe_warn_former_built_in_variable): New static function.
	(tree_simple_assignment::tree_simple_assignment,
	tree_multi_assignment::tree_multi_assignment):
	Move definition here from pt-assign.h
	Maybe warn about assignment to obsolete built-in variables.

	* version.h (OCTAVE_STARTUP_MESSAGE): Mention "news" function.

	* pt-stmt.cc (tree_statement::eval): Use dynamic_cast, not
	static_cast.

	* help.cc (help_from_file): Show .oct or .mex file name if one
	exists in the same directory as the .m file.

2007-01-23  Luis F. Ortiz  <lortiz@interactivesupercomputing.com>

	* strfns.cc (Fstrncmp): New function.

2007-01-17  John W. Eaton  <jwe@octave.org>

	* help.cc (help_from_file, help_from_symbol_table, help_from_list):
	Rewrite using raw_ versions.

2007-01-17  David Bateman  <dbateman@free.fr>

	* help.cc (raw_help, raw_help_from_file,
	raw_help_from_symbol_table, raw_help_from_list): New functions.
	* help.h (raw_help): Provide decl.

2007-01-16  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/__pchip_deriv__.cc: Fix dpchim decl for --enable-64.
	(F__pchip_deriv__): Fix call to dpchim for --enable-64.

	* DLD-FUNCTIONS/fftw.cc: Delete decl for fftw_planner.

2007-01-11  Michael Goffioul  <michael.goffioul@swing.be>

	* load-path.h (class load_path): Tag with OCTINTERP_API.

	* ls-oct-binary.cc (read_binary_data): Use unsigned char, not
	char, for reading flags and data types.

2007-01-10  Luis F. Ortiz  <lortiz@interactivesupercomputing.com>

	* parse.y (load_fcn_from_file): Delete unused variable NAMES.
	If NM is absolute, strip directory and extension parts.

2007-01-10  Michael Goffioul  <michael.goffioul@swing.be>

	* DLD-FUNCTIONS/getrusage.cc: Use #ifdef __WIN32__, not __MINGW32__.

2007-01-10  John W. Eaton  <jwe@octave.org>

	* parse.y (load_fcn_from_file): Only compare last two characters
	when looking for ".m".

2007-01-06  John W. Eaton  <jwe@octave.org>

	* version.h (OCTAVE_CONFIG_STATEMENT): New macro
	(OCTAVE_NAME_AND_VERSION): Remove config info.
	(OCTAVE_NAME_VERSION_COPYRIGHT_COPYING_AND_WARRANTY):
	Use OCTAVE_CONFIG_STATEMENT here.

2007-01-05  David Bateman  <dbateman@free.fr>

	* Makefile.in (DLD_XSRC): Add fftw.cc and remove fftw_wisdom.cc
	* DLD-FUNCTIONS/fftw.cc: New file.
	* DLD-FUNCTIONS/fftw_wisdom.cc: Delete.
	* defaults.cc (Vfftw_wisdom_program): Delete variable.
	(set_default_fftw_wisdom_prog): Delete function that sets it.
	(install_defaults): Delete set_default_fftw_prog from defaults.
	(Ffftw_wisdom_program): Delete.

2007-01-04  David Bateman  <dbateman@free.fr>

	* ov-fcn-handle.cc (octave_fcn_handle::load_ascii,
	octave_fcn_handle::load_binary):
	Throw error if handle can't be created.

2007-01-04  Luis F. Ortiz  <lortiz@interactivesupercomputing.com>

	* mex.cc (mxArray_number::mxArray_number (int, const char **)):
	Correctly index LHS in assignment.

2007-01-03  John W. Eaton  <jwe@octave.org>

	* data.cc (Fisinteger): New function.
	* ov-intx.h (OCTAVE_VALUE_INT_MATRIX_T::is_integer_type,
	OCTAVE_VALUE_INT_SCALAR_T::is_integer_type): New function.
	* ov.h (octave_value::is_integer_type): New function.
	* ov-base.h (octave_base_value::is_integer_type): New virtual function.

2007-01-03  Michael Goffioul  <michael.goffioul@swing.be>

	* toplev.cc (Fsystem): Handle async calls on Windows systems.

2007-01-03  David Bateman  <dbateman@free.fr>

	* (OPERATORS/op-cm-scm.cc, OPERATORS/op-cm-sm.cc, 
	OPERATORS/op-cs-scm.cc, OPERATORS/op-cs-sm.cc, 
	OPERATORS/op-m-scm.cc, OPERATORS/op-m-sm.cc, 
	OPERATORS/op-scm-cm.cc, OPERATORS/op-scm-cs.cc, 
	OPERATORS/op-scm-m.cc, OPERATORS/op-scm-s.cc, 
	OPERATORS/op-scm-scm.cc, OPERATORS/op-scm-sm.cc, 
	OPERATORS/op-sm-cm.cc, OPERATORS/op-sm-cs.cc, 
	OPERATORS/op-sm-m.cc, OPERATORS/op-sm-s.cc, 
	OPERATORS/op-sm-scm.cc, OPERATORS/op-sm-sm.cc, 
	OPERATORS/op-s-scm.cc, OPERATORS/op-s-sm.cc):
	Modify div and ldiv functions so that scalars stored as sparse 
	matrices are special cased.

	* ov-re-sparse.cc (double_value, complex_value): Scalar can be
	stored as a sparse matrix and so don't warn on implicit conversion
	to a scalar.
	* ov-cx-sparse.cc (double_value, complex_value): ditto.
	* ov-bool-sparse.cc (double_value, complex_value): ditto.

2007-01-03  John W. Eaton  <jwe@octave.org>

	* dynamic-ld.cc (octave_dynamic_loader::do_load_mex): Also check
	for _mexFunction.

2006-12-30  John W. Eaton  <jwe@octave.org>

	* ov-fcn-inline.cc: For compatibility, class id is
	"function_handle", not "inline function".

2006-12-27  John W. Eaton  <jwe@octave.org>

	* Makefile.in (%.df : %.cc): Insert a "do not edit" notice in the
	.df files.

	* mex.cc (mxArray_matlab::get_class_name,
	mxArray_octave_value::get_class_id): Use "function_handle", not
	"function handle".

2006-12-14  John W. Eaton  <jwe@octave.org>

	* pt-decl.cc (tree_decl_elt::eval): New function.
	* pt-decl.h (tree_decl_elt::eval): Provide decl.
	(tree_decl_elt::is_defined, tree_decl_elt::lvalue_ok, 
	tree_decl_elt::mark_as_formal_parameter, tree_decl_elt::rvalue,
	tree_decl_elt::lvalue): New functions.

	* pt-misc.h (class tree_parameter_list): Derive from
	octave_base_list<tree_decl_elt *> instead of
	octave_base_list<tree_identifier *>.
	(tree_parameter_list::tree_parameter_list (tree_identifier *)): Delete.
	(tree_parameter_list::tree_parameter_list (tree_decl_elt *)):
	New function.
	* pt-misc.cc (tree_parameter_list::mark_as_formal_parameters,
	tree_parameter_list::initialize_undefined_elements,
	tree_parameter_list::undefine, tree_parameter_list::dup,
	tree_parameter_list::convert_to_const_vector,
	tree_parameter_list::is_defined): Handle argument list elements
	as tree_decl_list objects instead of tree_identifier objects.
	(tree_parameter_list::define_from_arg_vector): Likewise.
	Always process entire list, setting default values if possible.
	Accept ":" to mean "use default argument".

	* parse.y (param_list2): Use decl2 to recognize
	"identifier '=' expression" in addition to "identifier".

	* parse.y (return_list1, make_anon_fcn_handle, finish_function):
	Adapt to new definition of tree_parameter_list.
	* pt-bp.cc (tree_breakpoint::visit_parameter_list): Likewise.
	* pt-check.cc (tree_checker::visit_parameter_list): Likewise.
	* pt-pr-code.cc (tree_print_code::visit_parameter_list): Likewise.

2006-12-08  John W. Eaton  <jwe@octave.org>

	* ov-intx.h (OCTAVE_VALUE_INT_MATRIX_T::array_value, 
	OCTAVE_VALUE_INT_MATRIX_T::complex_array_value, 
	OCTAVE_VALUE_INT_MATRIX_T::bool_array_value, 
	OCTAVE_VALUE_INT_MATRIX_T::char_array_value):
	Use fortran_vec to avoid expensive indexing operator.

2006-12-08  David Bateman  <dbateman@free.fr>

	* ov-intx.h (OCTAVE_VALUE_INT_SCALAR_T::matrix_value,
	OCTAVE_VALUE_INT_SCALAR_T::complex_matrix_value, 
	OCTAVE_VALUE_INT_MATRIX_T::matrix_value, 
	OCTAVE_VALUE_INT_MATRIX_T::comlex_matrix_value): New functions.
	(OCTAVE_VALUE_INT_MATRIX_T::array_value, 
	OCTAVE_VALUE_INT_MATRIX_T::comlex_array_value):
	Use octave_idx_type instead of int for indexing.

2006-12-04  David Bateman  <dbateman@free.fr>

	* xpow.cc (xpow (const Matrix&, double)): Add matrix type probing
        to matrix inverse.
        (xpow (const ComplexMatrix&, double)): ditto.
        * DLD-FUNCTIONS/inv.cc (Finv): Add matrix type probing.

2006-12-06  John W. Eaton  <jwe@octave.org>

	* sysdep.cc: Include "Cell.h" here.
	* input.h: Include "oct-obj.h", not "ov-list.h".

2006-12-06  Michael Goffioul  <michael.goffioul@swing.be>

	* mappers.cc (install_mapper_functions): Undefine isascii before
	the DEFUN_MAPPER for it.

	* input.cc (get_user_input): Prevent out of bounds array access
	when checking for EOL.

2006-12-05  John W. Eaton  <jwe@octave.org>

	* ls-oct-ascii.cc (extract_keyword): If no match and looking for
	more, skip to end of line before trying another match.

	* pt-mat.cc (tm_row_const::empty): New function.
	(tm_const::init): Don't append anything if tmp tm_row_const object
	is empty.
	(tree_matrix::rvalue): Default return value is Matrix().
	Don't do anything if tmp tm_const object is empty.

	* dirfns.cc (Fmkdir): Fix thinko in previous change.

2006-12-05  Paul Kienzle  <pkienzle@users.sf.net>

	* DLD-FUNCTIONS/rand.cc: Test for Poisson generator with lambda > 1e8.

2006-12-04  David Bateman  <dbateman@free.fr>

	* data.cc (Fdata): If ndims is greater than nargout and
	nargout is greater than 1, then collect size of trailing
	dimensions into retval(end).

	* load-path.cc (load_path::do_find_fcn): Return empty string if
	tests fail.

        * ov-base-mat.cc (void octave_base_matrix<MT>::assign (const
	octave_value_list&, const MT&)): Invalidate matrix type on
	assignment.

2006-11-30  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/fftw_wisdom.cc (Ffftw_wisdom):
	Accept "r" or "w" for second argument.

2006-11-29  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/fftw_wisdom.cc (Ffftw_wisdom): Look in load-path
	if reading wisdom file, but not if writing.

2006-11-28  John W. Eaton  <jwe@octave.org>

	* mex.cc (mxArray_struct::get_field_by_number):
	Return 0 if key_num is out of range.
	(mxArray_struct::set_field_by_number):
	Do nothing if key_num is out of range.
	(mxArray_cell::get_cell, mxArray_cell::set_cell):
	Avoid out-of-bounds indexing

2006-11-28  Luis F. Ortiz  <lortiz@interactivesupercomputing.com>

	* mex.cc (mxArray_matlab::get_n, mxArray_octave_value::get_n):
	Return product of last N-1 dims.

2006-11-28  John W. Eaton  <jwe@octave.org>

	* lex.l (eat_whitespace): Also handle CRLF as EOL.

	* dirfns.cc (Fmkdir): Handle "mkdir (parent, dir)".

2006-11-21  John W. Eaton  <jwe@octave.org>

	* load-path.cc (load_path::do_find_file,
	load_path::do_find_first_of, load_path::do_find_all_first_of): 
	Call rooted_pathname instead of absolute_pathname.
	* utils.cc (fcn_file_in_path, oct_file_in_path, mex_file_in_path):
	Likewise.

2006-11-16  Michael Goffioul  <michael.goffioul@swing.be>

	* oct-hist.cc (default_history_file): Instead of appending
	"/.octave_hist", append directory separator (but only if
	necessary), then ".octave_hist".

2006-11-16  John W. Eaton  <jwe@octave.org>

	* data.cc (Fresize): Fix doc string.

2006-11-15  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/__gnuplot_raw__.l (write_data, write_inline_data):
	New functions.
	(save_in_tmp_file): Use write_data.
	(gnuplot::send_inline_data, gnuplot_do_send_inline_data):
	New functions.
	(F__gnuplot_send_inline_data__, F__gnuplot_save_data__):
	New functions.

	* ls-oct-ascii.cc (save_ascii_data_for_plotting):
	Call save_ascii_data with precision = 4.
	(save_three_d): Temporarily set precision to 4 for output stream.

2006-11-15  Michael Goffioul  <michael.goffioul@swing.be>

	* Cell.h (Cell): Tag class with OCTINTERP_API.

2006-11-13  John W. Eaton  <jwe@octave.org>

	* octave.cc (maximum_braindamage): Disable
	Octave:fopen-file-in-path and Octave:load-file-in-path warnings.

	* load-save.cc (find_file_to_load): New function.
	(Fload): Call find_file_to_load to search load path for file.

	* file-io.cc (Ffopen): Search load path for file.

	* load-path.cc (path::do_find_first_of, path::do_find_file):
	Break out of all loops once file is found, not just innermost one.

	* data.cc (Fsize_equal): New function.

2006-11-13  Michael Goffioul  <michael.goffioul@swing.be>

	* ov.cc (check_subsref_elements, Fsubsref, Fsubsasgn):
	New functions.

	* ov-re-mat.h, ov-scalar.h, pr-output.h:
	Sprinkle with OCTINTERP_API as needed.

2006-11-11  John W. Eaton  <jwe@octave.org>

	* Makefile.in (parse.cc): Avoid creating empty files.
	(%.df : %cc, builtins.cc, mkbuiltins, PKG_ADD, DOCSTRINGS,
	doc-files, gendoc.cc, ops.cc, $(OPT_HANDLERS), oct-errno.cc,
	oct-gperf.h): Use $(simple-move-if-change-rule) here.
	(lex.cc, __gnuplot_raw__.cc):
	Use $(destdir-move-if-change-rule) here.
	(ifndef omit_deps): Don't include stamp-prereq here.
	($(DEF_FILES), $(MAKEDEPS)): Also depend on stamp-prereq.

2006-11-10  John W. Eaton  <jwe@octave.org>

	* ov-str-mat.cc (octave_char_matrix_str::load_ascii,
	octave_char_matrix_str::load_binary):
	Use chMatrix as buffer instead of C string.

2006-11-09  John W. Eaton  <jwe@octave.org>

	* ov-usr-fcn.h (octave_user_function::inline_function):
	 New data member.
	(octave_user_function::mark_as_inline_function): Set it.
	(octave_user_function::is_inline_function): Check it.
	* ov-usr-fcn.cc (octave_user_function::do_multi_index_op):
	Also skip setting curr_parent_fucntion if evaluating an inline
	function.
	(octave_user_function::octave_user_function):
	 Initialize inline_function.
	* pt-fcn-handle.cc (tree_anon_fcn_handle::rvalue):
	Mark user function as inline here.

2006-11-09  Michael Goffioul  <michael.goffioul@swing.be>

	* load-path.cc (load_path::move): Don't use reference to file_info
	object that will be erased.

2006-11-07  John W. Eaton  <jwe@octave.org>

	* utils.cc (file_in_path): Don't unconditionally return "".

	* pager.cc (default_pager): Don't append options here.
	(pager_command): New function.
	(do_sync): Use it.
	(VPAGER_FLAGS): New variable.
	(FPAGER_FLAGS): New function.

2006-11-06  John W. Eaton  <jwe@octave.org>

	* oct-hist.cc (default_history_file): If env_file is not empty,
	just accept it rather than checking to see if it exists.

2006-11-03  Bill Denney  <denney@seas.upenn.edu>

	* help.cc (keywords): Document try and unwind_protect.

2006-11-03  John W. Eaton  <jwe@octave.org>

	* Makefile.in (DLL_CXXDEFS): Rename from XTRA_CXXDEFS.
	(DLL_CDEFS): Rename from XTRA_CDEFS.
	Substitute OCTINTERP_DLL_DEFS, not XTRA_OCTINTERP_DEFS.

2006-11-02  Michael Goffioul  <michael.goffioul@swing.be>

	* sysdep.cc (MSVC_init): Also cal w32_sigint_init and
	w32_set_quiet_shutdown.

2006-11-01  John W. Eaton  <jwe@octave.org>

	* ov.h (octave_value::idx_type_value): New function.
	* data.cc (fill_matrix, Flinspace, Freshape): Use idx_type_value
	instead of int_value to extract size args.

2006-10-31  John W. Eaton  <jwe@octave.org>

	* ov-range.h (octave_range::empty_clone): Return octave_matrix
	object instead of octave_range.

2006-10-29  Juhani Saastamoinen  <juhani@cs.joensuu.fi>

	* ls-mat5.cc (read_mat5_tag): Declare bytes as int32_t, not just int. 

2006-10-28  John W. Eaton  <jwe@octave.org>

	* toplev.cc (Fatexit): Push function names on the stack even if we
	don't have atexit or on_exit.

2006-10-28  Michael Goffioul  <michael.goffioul@swing.be>

	* mappers.cc (install_mapper_functions): Undefine toascii before
	the DEFUN_MAPPER for it.

2006-10-27  John W. Eaton  <jwe@octave.org>

	* mexproto.h: Include oct-dlldefs.h.

	* pr-output.cc (SPECIALIZE_UABS): New macro.
	Use it to generate specializations of abs for unsigned int types.
	Instantiate abs for signed int types.

	* load-path.cc (load_path::do_initialize):
	Use dir_path::path_sep_str instead of ":".
	Don't append ":::" to xpath when calling do_set.

	* dirfns.cc (Fls, cleanup_iprocstream): Delete.

	* sysdep.cc (Ftilde_expand): If arg is cellstr, return cellstr.

	* ov.h (octave_value::is_cellstr): New function.
	* ov-base.h (octave_base_value::is_cellstr): New function.
	* ov-cell.h (octave_cell::is_cellstr): New function.
	* ov-cell.cc (Fiscellstr): Implement with is_cellstr.
	* Cell.cc (Cell::Cell (const dim_vector&, const string_vector&, bool)): 
	New constructor.
	(Cell::is_cellstr): New function.
	* Cell.h: Provide decls.

	* defaults.cc (subst_octave_home): If dir_sep_char is not '/',
	replace before returning.
	From Michael Goffioul <michael.goffioul@swing.be>.

2006-10-26  John W. Eaton  <jwe@octave.org>

	* cutils.c (octave_strcasecmp, octave-strncasecmp):
	Move to liboctave/lo-cutils.c.
	* utils.h: Delete decls.
	* strcasecmp.c: Move to liboctave/strcasecmp.c.
	* strncase.c: Move to liboctave/strncase.c.
	* Makefile.in (DIST_SRC): Delete them from the list.

	* bitfcns.cc (bitshift): If A < 0, return -bitshift (-A, N, MASK).

2006-10-26  Michael Goffioul  <michael.goffioul@swing.be>

	* help.cc (display_help_text): Quote sed patterns with ".

	* file-io.cc: Include file-io.h.

	* TEMPLATE-INST/Array-os.cc, TEMPLATE-INST/Array-sym.cc,
	TEMPLATE-INST/Array-tc.cc, defun-int.h, error.h, file-io.h,
	gripes.h, ls-oct-ascii.h, mexproto.h, oct-map.h, oct-obj.h,
	oct-stream.cc, oct-stream.h, octave.h, ov-base.h,
	ov-bool-sparse.cc, ov-bool-sparse.h, ov-cx-sparse.cc,
	ov-cx-sparse.h, ov-fcn.h, ov-re-sparse.cc, ov-re-sparse.h,
	ov-str-mat.h, ov-typeinfo.cc, ov-typeinfo.h, ov.h, pager.h,
	parse.h, pr-output.cc, pr-output.h, procstream.h, sighandlers.h,
	symtab.h, unwind-prot.h, utils.h, variables.h:
	Sprinkle with OCTINTERP_API as needed.

2006-10-26  John W. Eaton  <jwe@octave.org>

	* ov-bool.h (octave_bool::sparse_matrix_value,
	octave_bool::sparse_complex_matrix_value
	octave_bool::sparse_bool_matrix_value): New functions.

2006-10-25  John W. Eaton  <jwe@octave.org>

	* sighandlers.cc: Check defined (RETSIGTYPE_IS_VOID) instead of
	RETSIGTYPE == void.

	* oct-procbuf.cc (BUFSIZ): Define if not already defined.
	(octave_procbuf::open): Pass BUFSIZ as size argument to setvbuf.

2006-10-25  Michael Goffioul  <michael.goffioul@swing.be>

	* Makefile.in (XTRA_CDEFS, XTRA_CXXDEFS): Substitute here.

	* oct-procbuf.cc [_MSC_VER]: Define W32POPEN and W32PCLOSE the
	same as for __MINGW32__.
	(octave_procbuf::open, octave_procbuf::close) [_MSC_VER]:
	Use the same code as for __MINGW32__ and __CYGWIN__.

	* oct-prcstrm.cc [_MSC_VER]: Define popen and pclose.

2006-10-25  John W. Eaton  <jwe@octave.org>

	* sysdep.cc (w32_set_octave_home): Correctly initialize bin_dir.
	Fill it with '\0' instead of ' '.

2006-10-24  Michael Goffioul  <michael.goffioul@swing.be>

	* Makefile.in: Filter out $(XTRA_CXXDEFS) from $(ALL_CXXFLAGS) for
	$(DLD_PICOBJ).

2006-10-24  John W. Eaton  <jwe@octave.org>

	* sysdep.cc (MSC_init): New function.
	(sysdep_init): Call it.
	(w32_set_octave_home): New function, based on code from Michael
	Goffioul <michael.goffioul@swing.be>.
	(MINGW_init): Call w32_set_octave_home here too.

2006-10-23  John W. Eaton  <jwe@octave.org>

	* symtab.cc (symbol_table::clear_mex_functions): New function.
	* symtab.h: Provide decl.
	* vriables.cc (clear_mex_functions): New function.
	* variables.h Provide decl.
	* toplev.cc: Call clear_mex_functions instead of delete_symbol_tables.

	* variables.cc (delete_symbol_tables): Delete.
	* variables.h: Delete decl.

	* mex.cc (mex::unmark_array): New function.
	(mex::persistent): Define with unmark_array.
	(maybe_unmark_array): New function.
	(mxArray_struct::set_field_by_number, mxArray_cell::set_cell):
	Call maybe_unmark_array on val to avoid freeing val twice on exit
	from mex function.
	(mxFree): Call xfree, not free.

2006-10-21  John W. Eaton  <jwe@octave.org>

	* ov-intx.h
	(OCTAVE_VALUE_INT_MATRIX_T::OCTAVE_TYPE_PREDICATE_FUNCTION,
	(OCTAVE_VALUE_INT_SCALAR_T::OCTAVE_TYPE_PREDICATE_FUNCTION):
	Function is now const, so it properly overloads method in base
	class.

	* mex.cc (mxArray_octave_value::is_uint32):
	Call val.is_uint32_type, not val.is_int32_type.
	(mxArray_octave_value::is_uint64):
	Call val.is_uint64_type, not val.is_int64_type.
	(mxArray_octave_value::is_uint8):
	Call val.is_uint8_type, not val.is_int8_type.

2006-10-20  Paul Kienzle  <pkienzle@users.sf.net>

	* ov-mex-fcn.h (octave_mex_function::atexit): New function.
	(octave_mex_function::exit_fcn_ptr): New data member.
	* ov-mex-fcn.cc (octave_mex_function::exit_fcn_ptr): New data member.
	(octave_mex_function::octave_mex_function): Initialize it.
	(octave_mex_function::~octave_mex_function):
	If we have an exit function, call it.

2006-10-20  John W. Eaton  <jwe@octave.org>

	* variables.cc (delete_symbol_tables): New function.
	* variables.h: Provide decl.
	* toplev.cc (do_octave_atexit): Call it.

	* mex.cc (mex::mex): New arg, a pointer to the current mex function.
	(mex::curr_mex_fcn): New data member.
	(mex::current_mex_function): New function.
	(mexAtExit): Set exit function pointer in current mex file object.

2006-10-20  Paul Kienzle  <pkienzle@users.sf.net>

	* mex.cc (enum callstyle): Delete enum definition.
	(Fortran_mex, C_mex): Delete functions.
	(call_mex): First arg is now bool.
	* ov-mex-fcn.cc (call_mex): Fix decl to match new definition.
	(Fortran_mex, C_mex): Delete decls.
	(octave_mex_function::do_multi_index_op): Simplify call to call_mex.

2006-10-20  John W. Eaton  <jwe@octave.org>

	* lex.l (handle_identifier): If a command name is found, skip
	starting command mode if parsing an object index.

2006-10-20  David Bateman  <dbateman@free.fr>

	* DLD-FUNCTION/spqr.cc (dmperm_internal): New function with core
	of Fdmperm.
	(Fdmperm): Call dmperm_internal rather then calculating loally.
	(Fsprank): New function to calculate the strutural rank also using
	dmperm_internal.

2006-10-19  John W. Eaton  <jwe@octave.org>

	* ov-struct.cc (octave_struct::as_mxArrary):
	Use OCTAVE_LOCAL_BUFFER to allocate tempoarary array of field names.

	* mxarray.h (mxArray::persistent): Delete var and all uses.
	(mxArray::mark_persistent, mxArray::unmark_persistent,
	mxArray::is_persistent): Delete functions.
	* mex.cc (mex::function_name): Use mxArray::strsave, not strsave.
	(mex::mark_array): New function.
	(mex::make_value): Use it.
	(mex::free_value): Return true if we found ptr in arraylist.
	(mex::persistent (mxArray *)): Remove ptr from arraylist instead
	of marking it as persistent.
	(mxArray::malloc): Call ::malloc instead of malloc.
	(mxArray::calloc): Call ::calloc instead of calloc.
	(maybe_mark_array): New function.
	(mxCreateCellArray, mxCreateCellMatrix, mxCreateCharArray,
	mxCreateCharMatrixFromStrings, mxCreateDoubleMatrix,
	mxCreateDoubleScalar, mxCreateLogicalArray, mxCreateLogicalMatrix,
	mxCreateLogicalScalar, mxCreateNumericArray,
	mxCreateNumericMatrix, mxCreateSparse,
	mxCreateSparseLogicalMatrix, mxCreateString, mxCreateStructArray,
	mxCreateStructMatrix, mxDuplicateArray): Use it.
	(mxDestroyArray): No need to check persistence now.
	Also delete ptr if mex_context->free_value does not.
	(call_mex): No need to delete elements of argout now.

2006-10-18  John W. Eaton  <jwe@octave.org>

	* dynamic-ld.cc (octave_shlib_list::remove,
	octave_shlib_list::do_remove, octave_mex_file_list::remove,
	octave_mex_file_list::do_remove): New arg, cl_hook.
	(octave_shlib_list::do_remove): Pass cl_hook to octave_shlib close
	function.
	(octave_dynamic_loader::do_load_oct): Don't call close on shl
	directly.  Pass do_clear_function to octave_shlib_list::remove.  

	* mex.cc (mexUnlock): Don't warn if unlocking a function that is
	not locked.

	* pt-fcn-handle.cc (tree_anon_fcn_handle::dup):
	Correctly duplicate symbol table info.

2006-10-17  Michael Goffioul  <michael.goffioul@swing.be>

	* oct-map.h: Include <algorithm>.

2006-10-16  John W. Eaton  <jwe@octave.org>

	* oct-stream.cc (octave_stream_list::do_remove): Handle "all" as a
	special case.

2006-10-13  Michael Goffioul  <michael.goffioul@swing.be>

	* Makefile.in: Adapt rules to use $(LIBPRE).

2006-10-11  John W. Eaton  <jwe@octave.org>

	* mex.h (UINT64_T, uint64_T, INT64_T, int64_T, UINT32_T, uint32_T,
	INT32_T, int32_T, UINT16_T, uint16_T, INT16_T, int16_T, UINT8_T,
	uint8_T, INT8_T, int8_T): Conditionally define.
	From Andy Adler <adler@site.uottawa.ca>.

2006-10-09  John W. Eaton  <jwe@octave.org>

	* oct-conf.h.in (OCTAVE_CONF_CURL_LIBS): Substitute.
	* toplev.cc (octave_config_info): Add CURL_LIBS to the list.

2006-10-09  Alexander Barth  <abarth@marine.usf.edu>

	* DLD-FUNCTIONS/urlwrite.cc: New file providing urlwrite and urlread.

2006-10-09  John W. Eaton  <jwe@octave.org>

	* pt-mat.cc (tree_matrix::dup): Append new elements to new matrix.
	* pt-cell.cc (tree_cell::dup): Append new elements to new cell array.

2006-10-04  John W. Eaton  <jwe@octave.org>

	* version.h (OCTAVE_VERSION): Now 2.9.9+.

	* DLD-FUNCTIONS/__gnuplot_raw__.l (F__clear_plot_window__):
	Rename from Fclearplot.  Only clear plot window.

2006-10-03  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/luinc.cc (Fluinc): Avoid crash if SparseLU or
	SparseComplexLU constructor fails.

	* DLD-FUNCTIONS/find.cc (find_nonzero_elem_idx):
	Delete unused variable.

2006-10-03  David Bateman  <dbateman@free.fr>

	* Makefile.in (OCT_LINK_DEPS) Include $(CAMD_LIBS) in the list.
	(octave$(EXEEXT)): Include $(CAMD_LIBS) in link command.

2006-10-02  John W. Eaton  <jwe@octave.org>

	* version.h (OCTAVE_VERSION): Now 2.9.9.
	(OCTAVE_API_VERSION): Now api-v22.
	(OCTAVE_RELEASE_DATE): Now 2006-10-02.

	* pr-output.cc (pr_plus_format): Use "inline" instead of "static
	inline" for template functions to avoid problems when not
	compiling with 	g++.

	* mex.cc (call_mex): Delete elements of argout.
	From Kai Labusch <labusch@inb.uni-luebeck.de>.

2006-09-29  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/spfind.cc (sparse_find_non_zero_elem_idx):
	Leading dimension is NR, not NC.

	* Makefile.in (DEP_5): Include oct-errno.cc in the list.

2006-09-27  John W. Eaton  <jwe@octave.org>

	* pr-output.cc (abs): New template function.  Instantiate for
	unsigned types.

2006-09-27  David Bateman  <dbateman@free.fr>

	* file-io.cc (Fsprintf): If fmt is an sq_string, return an sq_string.

	* DLD-FUNCTIONS/sparse.cc (spfind, sparse_find): Delete.
	* DLD-FUNCTIONS/spfind.cc: New file implementating compatible
	sparse find function.
	* Makefile.in (DLD_XSRC): Add spfind.cc.
	
2006-09-26  Bill Denney  <bill@givebillmoney.com>

	* DLD-FUNCTIONS/find.cc (find_nonzero_elem_idx, Ffind):
	Handle direction and limit on number of elements to find.

2006-09-26  John W. Eaton  <jwe@octave.org>

	* error.cc (warning_1): Call error_2, not error.

2006-09-15  John W. Eaton  <jwe@octave.org>

	* ops.h (DEFSTRDBLCONVFN): New arg, TFROM.
	* OPERATORS/op-double-conv.cc: Declare and install sq_string
	conversions too.

	* ov-base-int.cc (octave_base_int_scalar<T>::convert_to_str_internal,
	octave_base_int_matrix<T>::convert_to_str_internal):
	New functions.
	* ov-base-int.h: Provide decls.

2006-09-15  Søren Hauberg  <soren@hauberg.org>.

	* data.cc (Fsize): If nargout > ndims, fill with 1.

2006-09-15  John W. Eaton  <jwe@octave.org>

	* octave.cc: Fix xerbla decl.

2006-09-13  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/cellfun.cc: Improve error message for uniform
	output case when results are not scalars.  Avoid shadow decls.

2006-09-13  David Bateman  <dbateman@free.fr>

	* ov-scalar.h (octave_scalar::sparse_matrix_value,
	octave_scalar::sparse_complex_matrix_value): New extractor functions.
	* ov-complex.h (octave_complex::sparse_matrix_value,
	octave_complex::sparse_complex_matrix_value): Ditto.
	* DLD-FUNCTIONS/spkron.cc (Fspkron): Change example in help.

2006-09-12  David Bateman  <dbateman@free.fr>

	* DLD-FUNCTIONS/cellfun.cc (Fcellfun): Allow matlab compatiable
	'UniformOutput' and 'ErrorHandler' options.  Change output when
	called with function handler or inline function to by default have
	'UniformOutput' set to true.  Allow functions with multiple inputs
	and outputs. Add test code.  Replace some int with octave_idx_type.
	(Fnum2cell): Replace some int with octave_idx_type.  Add test code.
	
2006-09-11  Yozo Hida  <yozo@cs.berkeley.edu>

	* DLD-FUNCTIONS/gcd.cc (Fgcd): Extend range by using std::floor
	instead of converting to int.

2006-09-05  John W. Eaton  <jwe@octave.org>

	* mex.cc (mxArray_sparse::as_octave_value): Cast nzmax to
	octave_idx_type for sparse matrix constructors.

2006-09-01  John W. Eaton  <jwe@octave.org>

	* dirfns.cc: Don't handle nargout == 0 as a special case.
	(octave_change_to_directory): Perform tilde expansion on directory
	name here.

2006-08-30  John W. Eaton  <jwe@octave.org>

	* load-save.cc (get_file_format): Fix misplaced #endif.

2006-08-29  John W. Eaton  <jwe@octave.org>

	* load-path.cc (execute_pkg_add_or_del):
	Source PKG_ADD or PKG_DEL in base workspace.
	* parse.y (source_file): New optional arg, context.
	* parse.h (source_file): Fix decl.

2006-08-25  John W. Eaton  <jwe@octave.org>

	* version.h (OCTAVE_VERSION): Now 2.9.8+.

2006-08-24  John W. Eaton  <jwe@octave.org>

	* version.h (OCTAVE_VERSION): Now 2.9.8.
	(OCTAVE_API_VERSION): Now api-v21.
	(OCTAVE_RELEASE_DATE): Now 2006-08-24.

2006-08-23  John W. Eaton  <jwe@octave.org>

	* ov.h (octave_value::save_ascii): Delete strip_nan_and_inf arg.
	* ov-base.h, ov-base.cc (octave_base_value::save_ascii): Likewise.
	* ov-base-int.h, ov-base-int.cc	(octave_base_int_matrix<T>::save_ascii,
	octave_base_int_scalar<T>::save_ascii, ): Likewise.
	* ov-base-sparse.cc, ov-base-sparse.h
	(octave_base_sparse<T>::save_ascii): Likewise.
	* ov-bool-mat.cc, ov-bool-mat.h (octave_bool_matrix::save_ascii):
	Likewise.
	* ov-bool.cc, ov-bool.h (octave_bool::save_ascii): Likewise.
	* ov-cell.cc, ov-cell.h (octave_cell::save_ascii): Likewise.
	* ov-complex.cc, ov-complex.h (octave_complex::save_ascii): Likewise.
	* ov-fcn-handle.cc, ov-fcn-handle.h (octave_fcn_handle::save_ascii):
	Likewise.
	* ov-fcn-inline.cc, ov-fcn-inline.h (octave_fcn_inline::save_ascii):
	Likewise.
	* ov-list.cc, ov-list.h (octave_list::save_ascii): Likewise.
	* ov-range.cc, ov-range.h (octave_range::save_ascii): Likewise.
	* ov-scalar.cc, ov-scalar.h (octave_scalar::save_ascii): Likewise.
	* ov-str-mat.cc, ov-str-mat.h (octave_char_matrix_str::save_ascii):
	Likewise.
	* ov-struct.cc, ov-struct.h (octave_struct::save_ascii): Likewise.
	* ov-re-mat.cc, ov-re-mat.cc (octave_matrix::save_ascii): Likewise.
	* ov-cx-mat.cc, ov-cx-mat.cc (octave_complex_matrix::save_ascii):
	Likewise.
	* ov-cx-mat.cc, ov-cx-mat.cc (octave_complex_matrix::save_ascii):
	Likewise.
	* ov-re-mat.cc, ov-re-mat.cc (octave_matrix::save_ascii): Likewise.

	* ls-oct-ascii.cc, ls-oct-ascii.h (save_ascii_data):
	Delete strip_nan_and_inf arg.
	(save_ascii_data_for_plotting): Delete strip_nan_and_inf arg from
	call to save_ascii_data.

	* DLD-FUNCTIONS/__gnuplot_raw__.l (handle_using): Accept "(EXPR)"
	as components of using clauses.
	(enum _toktype): New element DOLLAR.
	Accept "$" in lexer.

	* ls-oct-ascii.cc (save_ascii_data): Delete arg strip_nan_and_inf.
	Change all uses.

	* ls-oct-ascii.h (save_three_d): Provide decl.
	* load-save.h (save_ascii_data_for_plotting, save_three_d):
	Delete decls.

2006-08-22  John W. Eaton  <jwe@octave.org>

	* ov.h (octave_value::save_ascii): strip_nan_and_inf is now int,
	not bool.
	* ov-base.h, ov-base.cc (octave_base_value::save_ascii): Likewise.
	* ov-base-int.h, ov-base-int.cc	(octave_base_int_matrix<T>::save_ascii,
	octave_base_int_scalar<T>::save_ascii, ): Likewise.
	* ov-base-sparse.cc, ov-base-sparse.h
	(octave_base_sparse<T>::save_ascii): Likewise.
	* ov-bool-mat.cc, ov-bool-mat.h (octave_bool_matrix::save_ascii):
	Likewise.
	* ov-bool.cc, ov-bool.h (octave_bool::save_ascii): Likewise.
	* ov-cell.cc, ov-cell.h (octave_cell::save_ascii): Likewise.
	* ov-complex.cc, ov-complex.h (octave_complex::save_ascii): Likewise.
	* ov-fcn-handle.cc, ov-fcn-handle.h (octave_fcn_handle::save_ascii):
	Likewise.
	* ov-fcn-inline.cc, ov-fcn-inline.h (octave_fcn_inline::save_ascii):
	Likewise.
	* ov-list.cc, ov-list.h (octave_list::save_ascii): Likewise.
	* ov-range.cc, ov-range.h (octave_range::save_ascii): Likewise.
	* ov-scalar.cc, ov-scalar.h (octave_scalar::save_ascii): Likewise.
	* ov-str-mat.cc, ov-str-mat.h (octave_char_matrix_str::save_ascii):
	Likewise.
	* ov-struct.cc, ov-struct.h (octave_struct::save_ascii): Likewise.
	* ov-re-mat.cc, ov-re-mat.cc (octave_matrix::save_ascii): Likewise.
	* ov-cx-mat.cc, ov-cx-mat.cc (octave_complex_matrix::save_ascii):
	Likewise.

	* ov-cx-mat.cc, ov-cx-mat.cc (octave_complex_matrix::save_ascii):
	Don't strip Inf and NaN here.  Call ComplexMatrix::save_ascii to
	do the real work.
	* ov-re-mat.cc, ov-re-mat.cc (octave_matrix::save_ascii):
	Don't strip Inf and NaN here.  Call Matrix::save_ascii to do the
	real work.

	* ov-re-mat.cc, ov-cx-mat.cc (strip_infnan): Delete.
	* ls-oct-ascii.cc, ls-oct-ascii.h (save_ascii_data):
	strip_nan_and_inf is now int, not bool.
	(strip_infnan): Delete.
	(save_ascii_data_for_plotting): Call save_ascii_data with
	strip_nan_and_inf = 2.

	* Makefile.in (INCLUDES): Remove matrix.h from the list.

2006-08-22  David Bateman  <dbateman@free.fr>

	* sparse-xpow.cc: Replace all uses of pow by std::pow.
	(elem_pow): Simplify for two sparse matrices arguments.

2006-08-22  John W. Eaton  <jwe@octave.org>

	* ls-oct-ascii.cc: Increase default value of save_precision to 16.

	* ls-mat-ascii.cc (save_mat5_ascii_data): Use scientific format.

2006-08-21  John W. Eaton  <jwe@octave.org>

	* ls-mat5.cc (read_mat5_binary_data, read_mat5_integer_data,
	write_mat5_array): Unconditionally enable code for 64-bit int types.

	* ov-fcn-handle.cc (DEFINE_OV_TYPEID_FUNCTIONS_AND_DATA):	
	For compatibility, set class to function_handle.

2006-08-18  John W. Eaton  <jwe@octave.org>

	* pr-output.cc (set_real_matrix_format): Also always include space
	for sign if values are int, inf, or nan.
	(set_complex_format, set_complex_matrix_format): Always include
	space for sign in real format, but never in imaginary format.

	* gripes.cc (gripe_logical_conversion): New function.
	* gripes.h: Provide decl.
	* ov.h (octave_value::bool_value, octave_value::bool_matrix_value,
	octave_value::bool_array_value): New arg, warn.
	* ov-base.cc, ov-base.h (octave_base_value::bool_value,
	octave_base_value::bool_matrix_value, 
	octave_base_value::bool_array_value): Likewise.
	* ov-intx.h (OCTAVE_VALUE_INT_MATRIX_T::bool_array_value,
	OCTAVE_VALUE_INT_SCALAR_T::bool_array_value): Likewise.
	* ov-range.h (octave_range::bool_array_value): Likewise.
	* ov-re-mat.cc, ov-re-mat.h (octave_matrix::bool_array_value):
	Likewise.
	* ov-re-sparse.cc, ov-re-sparse.h
	(octave_sparse_matrix::bool_array_value): Likewise.
	* ov-scalar.h (octave_scalar::bool_value,
	octave_scalar::bool_array_value): Likewise.
	* OPERATORS/op-bm-b.cc (oct_assignop_conv_and_assign):
	Ask for warning from bool_array_value.
	* OPERATORS/op-bm-bm.cc (oct_assignop_conv_and_assign): Likewise.
	* ov-bool.h (octave_bool::bool_value,
	octave_bool::bool_matrix_value, octave_bool::bool_array_value):
	Likewise.
	* ov-bool-mat.h (octave_bool_matrix::bool_matrix_value,
	octave_bool_matrix::bool_array_value): Likewise.
	* ov-bool-sparse.cc, ov-bool-sparse.h
	(octave_sparse_bool_matrix::bool_matrix_value,
	octave_sparse_bool_matrix::bool_array_value): Likewise.

2006-08-18  Benjamin Lindner  <lindnerben@gmx.net>

	* ls-mat5.cc (read_mat5_tag): Declare type as int32_t, not int.
	(read_mat5_binary_element): Likewise, for len and element_length.

2006-08-18  John W. Eaton  <jwe@octave.org>

	* load-save.h (enum load_save_format): New element, LS_MAT_ASCII_LONG.
	* load-save.cc (Fload, Fsave): Make -ascii Matlab compatible.
	(do_save): Handle LS_MAT_ASCII.
	* ls-mat-ascii.cc (save_mat_ascii_data): New function.
	* ls-mat-ascii.h: Provide decl.

2006-08-17  John W. Eaton  <jwe@octave.org>

	* ls-mat5.cc (save_mat5_element_length): Correctly compute element
	length for character data.  Handle N-d character data
	(save_mat5_binary_element): Handle N-d character data.

2006-08-16  John W. Eaton  <jwe@octave.org>

	* parse.y (gobble_leading_white_space): New arg, SKIP_CODE.
	Change all uses.

2006-08-15  John W. Eaton  <jwe@octave.org>

	* help.cc (help_from_file): Call get_help_from_file with new file
	arg.  Print file info here.

	* parse.y (get_help_from_file): Delete include_file_info arg.
	Provide two versions, one that returns the file found in a
	reference arg and one that does not.

	* variables.cc (do_isglobal): New function.
	(Fisglobal): Use it.
	(Fis_global): New function.

2006-08-14  John W. Eaton  <jwe@octave.org>

	* variables.cc (symbol_record_name_compare): Fix casts.

	* ov-cell.cc (octave_cell::subsasgn): Call empty_conv for 0x0
	objects, not just empty objects.

	* oct-map.h (Octave_map::clear): Also clear key_list.

	* load-path.cc (load_path::find_dir_info, load_path::do_add,
	load_path::do_remove): Perform tilde expansion on directory here.

2006-07-29  John W. Eaton  <jwe@octave.org>

	* matrix.h: Delete to avoid conflict with liboctave/Matrix.h on
	case-insensitive filesystems.

	* version.h (OCTAVE_VERSION): Now 2.9.7+.

2006-07-28  John W. Eaton  <jwe@octave.org>

	* version.h (OCTAVE_VERSION): Now 2.9.7.
	(OCTAVE_API_VERSION): Now api-v20.
	(OCTAVE_RELEASE_DATE): Now 2006-07-28.

	* load-path.cc (path::do_add): Warn for trailing "//".

	* strfns.cc: Comment out some tests for snapshot.

2006-07-27  John W. Eaton  <jwe@octave.org>

	* defaults.cc (Vlocal_api_arch_lib_dir): New variable.
	(set_default_local_api_arch_lib_dir): New function.
	(install_defaults): Call it.
	(set_exec_path): Put Vlocal_api_arch_lib_dir in the list.
	* toplev.cc (localapiarchlibdir): Include it in the list.
	* defaults.h.in (OCTAVE_LOCALAPIARCHLIBDIR): Substitute here.

2006-07-26  John W. Eaton  <jwe@octave.org>

	* mex.cc (mxArray::as_octave_value (mxArray *)): New static function.
	(call_mex, mexPutVariable, mexCallMATLAB,
	mxArray_cell::as_octave_value, mxArray_struct::as_octave_value):
	Use static version of as_octave_value here.
	(mxArray_sparse::mxArray_sparse (const mxArray_sparse&)):
	Check pr, ir, and jc before calling memcpy.
	(mxArray_number::mxArray_number): Check pr before calling memcpy

	* mxarray.h (mxArray::as_octave_value (mxArray *)): Provide decl.
	(mxArray::as_octave_value): Now protected.

	* file-io.cc (Ffeof, Fferror): Use DEFUNX instead of DEFUN.

	* mex.cc (xfree): New function.
	(mex::free): Use it.
	(mxArray_struct::~mxArray_struct, mxArray_cell::~mxArray_cell):
	Delete elements with delete, not mxDestroyArray.
	(mex::cleanup): Don't call mex::free or mex::free_value.
	(mex::free_value): Add debug warning.
	(mex::mark, mex::unmark): Fix debug warning.
	(call_mex): Use unwind_protect frame.
	(mexUnlock): Use iterator to remove item from mex_lock_count.

	* octave.cc (maximum_braindamage): Use disable_warning instead of
	bind_internal_variable to disable function-name-clash warning.
	* error.cc (disable_warning): No longer static.
	* error.h: Provide decl.

2006-07-25  David Bateman  <dbateman@free.fr>

	* mex.cc (mxArray_octave_value::get_class_id): Handle sparse.
	(class mxArray_sparse): Derive from mxArray_matlab, not
	mxArray_number.
	(mxArray_sparse::as_octave_value): Implement function.
	* ov-bool-sparse.cc (octave_sparse_bool_matrix::as_mxArray):
	Implement function.
	* ov-cx-sparse.cc (octave_sparse_complex_matrix::as_mxArray):
	Implement function.	
	* ov-re-sparse.cc (octave_sparse_matrix::as_mxArray):
	Implement function.

2006-07-25  John W. Eaton  <jwe@octave.org>

	* mex.cc (mxArray_struct::as_octave_value, call_mex,
	mexCallMATLAB, mxArray_cell::as_octave_value):
	Convert NULL mxArray* pointers to empty matrix values.
	(mxArray_octave_value:get_ir): Fix typo.

2006-07-22  John W. Eaton  <jwe@octave.org>

	* mxarray.h: New file.
	* mex.h, mexproto.h, mex.cc: New implementation of mxArray and MEX
	interface.

	* ov.h (octave_value::mex_get_ir, octave_value::mex_get_jc):
	New functions.
	* ov-base.h (octave_base_value::mex_get_ir,
	octave_base_value::mex_get_jc): New virtual functions.
	* ov-base-sparse.h (octave_base_sparse<T>::mex_get_ir,
	octave_base_sparse<T>::mex_get_jc): New functions.

	* ov-intx.h (OCTAVE_VALUE_INT_SCALAR_T::as_mxArray,
	OCTAVE_VALUE_INT_MATRIX_T::as_mxArray): New functions.
	* ov-int8.h, ov-int16.h, ov-int32.h, ov-int64.h, ov-uint8.h,
	ov-uint16.h, ov-uint32.h, ov-uint64.h: Define OCTAVE_INT_MX_CLASS
	before including ov-intx.h.  Undef it after.
	* ov-range.cc (octave_range::as_mxArray): New function.
	* ov-range.h: Provide decl.
	* ov-scalar.cc (octave_scalar::as_mxArray): New function.
	* ov-scalar.h: Provide decl.
	* ov-complex.cc (octave_complex::as_mxArray): New function.
	* ov-complex.h: Provide decl.
	* ov-re-mat.cc (octave_matrix::as_mxArray): New function.
	* ov-re-mat.h: Provide decl.
	* ov-cx-mat.cc (octave_complex_matrix::as_mxArray): New function.
	* ov-cx-mat.h: Provide decl.
	* ov-ch-mat.cc (octave_char_matrix::as_mxArray): New function.
	* ov-ch-mat.h: Provide decl.
	* ov-bool-mat.cc (octave_bool_matrix::as_mxArray): New function.
	* ov-bool-mat.h: Provide decl.
	* ov-bool.cc (octave_bool::as_mxArray): New function.
	* ov-bool.h: Provide decl.

	* ov-struct.h (octave_struct::numel): New function.

	* ls-mat5.cc (arrayclasstype): Use "MAT_FILE_" instead of "mx" as
	prefix for element names.

	* ov.h (octave_value::as_mxArray): New function.
	* ov-base.cc (octave_base_value::as_mxArray): New function.
	* ov-base.h: Provide decl.

	* ov.h (octave_value::mex_get_data): New function.
	* ov-base.h (octave_base_value::mex_get_data): New function.
	* ov-base-scalar.h (octave_base_scalar<ST>::mex_get_data): New function.
	* ov-base-matrix.h (octave_base_matrix<MT>::mex_get_data): New function.
	* ov-intx.h (OCTAVE_VALUE_INT_SCALAR_T::mex_get_data): New function.

	* ov.h (octave_value::nfields): New function.
	* ov-base.cc (octave_base_value::nfields): New virtual function.
	* ov-base.h: Provide decl.
	* ov-struct.h (octave_struct::nfields): New function.

2006-07-19  John W. Eaton  <jwe@octave.org>

	* OPERATORS/op-bm-bm.cc (oct_assignop_conv_and_assign): New function.
	(install_bm_bm_ops): Install it for various types.
	* OPERATORS/op-bm-b.cc (oct_assignop_conv_and_assign): New function.
	(install_bm_b_ops): Install it for various types.
	* ov-intx.h (OCTAVE_VALUE_INT_MATRIX_T::bool_array_value,
	OCTAVE_VALUE_INT_SCALAR_T::bool_array_value): New functions.
	* ov-range.h (octave_range::bool_array_value): New function.
	* ov-re-sparse.cc (octave_sparse_matrix::bool_array_value):
	New function.
	* ov-re-sparse.h: Provide decl.
	* ov-re-mat.cc (octave_matrix::bool_array_value): New function.
	* ov-re-mat.h: Provide decl.

	* ov-base.cc (octave_base_value::numeric_assign):
	Avoid memory leak when converting LHS.

2006-07-18  John W. Eaton  <jwe@octave.org>

	* ov.h (octave_value::is_int8_type, octave_value::is_int16_type,
	octave_value::is_int32_type, octave_value::is_int64_type,
	octave_value::is_uint8_type, octave_value::is_uint16_type,
	octave_value::is_uint32_type, octave_value::is_uint64_type):
	New functions. 
	* ov-base.h (octave_base_value::is_int8_type,
	octave_base_value::is_int16_type,
	octave_base_value::is_int32_type,
	octave_base_value::is_int64_type,
	octave_base_value::is_uint8_type,
	octave_base_value::is_uint16_type,
	octave_base_value::is_uint32_type,
	octave_base_value::is_uint64_type): New functions.
	* ov-int8.h, ov-int16.h, ov-int32.h, ov-int16.h, ov-uint8.h,
	ov-uint16.h, ov-uint32.h, ov-uint16.h: Define OCTAVE_BASE_INT_T.
	* ov-intx.h: Use OCTAVE_BASE_INT_T to define type-specific predicate.

	* ov.h (octave_value::is_double_type, octave_value::is_single_type):
	New functions.
	* ov-base.h (octave_base_value::is_double_type,
	octave_base_value::is_single_type): New functions.
	* ov-re-mat.h (octave_matrix::is_double_type): New function.
	* ov-cx-mat.h (octave_matrix::is_double_type): New function.
	* ov-scalar.h (octave_scalar::is_double_type): New function.
	* ov-complex.h (octave_complex::is_double_type): New function.
	* ov-range.h (octave_range::is_double_type): New function.
	* ov-re-sparse.h (octave_sparse_matrix::is_double_type): New function.
	* ov-cx-sparse.h (octave_sparse_complex_matrix::is_double_type):
	New function.
	
2006-07-15  John W. Eaton  <jwe@octave.org>

	* ov-typeinfo.cc: Also instantiate arrays of assignany_ops.

	* oct-errno.cc.in (octave_errno::octave_errno): VALUE field in
	errno_struct no longer const.

	* DLD-FUNCTIONS/dispatch.cc (octave_dispatch::do_index_op):
	Declare	resize_ok arg as bool, not int.
	Provide default value for resize_ok.

2006-07-14  John W. Eaton  <jwe@octave.org>

	* oct-stream.cc (do_write): Include explicit instantiations.

	* ov-str-mat.h (octave_char_matrix_sq_str::resize):
	Include fill arg to match base class.
	* ov-struct.h (octave_struct::resize): Likewise.

	* ov-base-mat.cc, ov-base-mat.h, ov-base-sparse.cc,
	ov-base-sparse.h, ov-base.cc, ov-base.h, ov-bool.cc,
	ov-bool.h, ov-complex.cc, ov-complex.h, ov-intx.h, ov-list.cc,
	ov-list.h, ov-range.cc, ov-range.h, ov-scalar.cc, ov-scalar.h,
	ov-str-mat.cc, ov-str-mat.h, ov.h (do_index_op):
	Declare	resize_ok arg as bool, not int.
	Provide default value for resize_ok.

	* c-file-ptr-stream.h (c_file_ptr_stream::c_file_ptr_stream):
	Qualify init.

2006-07-13  John W. Eaton  <jwe@octave.org>

	* oct-map.h (Octave_map::del): Add missing std:: qualifier to use
	of find template.

2006-07-07  John W. Eaton  <jwe@octave.org>

	* ov.h (octave_value::is_bool_scalar): New function.
	* ov-base.h (octave_base_value::is_bool_scalar): New function.
	* ov-bool.h (octave_bool::is_bool_scalar): New function.

	* oct-map.cc (Octave_map::keys): Use assert.
	(Octave_map::assign): Avoid inserting new key in map unless
	assignment to cell succeeds.

	* oct-map.h (Octave_map::del): Only touch key_list if map contains key.
	Assert that key_list contains key.

	* oct-map.h (Octave_map::maybe_add_to_key_list): For efficiency,
	check map, not key_list.  From Paul Kienzle  <pkienzle@users.sf.net>.

2006-07-06  John W. Eaton  <jwe@octave.org>

	* DLD-FUNCTIONS/__glpk__.cc (F__glpk__): Declare mrowsc volatile.

	* ov-struct.cc (octave_struct::print_raw): Print keys in key_list
	order.

	* oct-map.h (Octave_map::key_list): New data member.
	(Octave_map::maybe_add_to_key_list): New function.
	(Octave_map::key_list_iterator, Octave_map::const_key_list_iterator):
	New typedefs.
	(Octave_map::del): Also delete key from key_list.
	* oct-map.cc (Octave_map::keys): Compute return value from
	key_list instead of map.
	(Octave_map::assign, Octave_map::contents): Call maybe_add_to_key_list.
	* oct-map.h, oct-map.cc (Octave_map::Octave_map):
	Insert keys in key_list.

	* oct-map.h (Octave_map::Octave_map (const dim_vector&,
	const string_vector&)): New arg, key_list_arg.
	* oct-map.cc: Move definition here.

2006-07-05  John W. Eaton  <jwe@octave.org>

	* mex.cc (mexGetVariable): Rename from mexGetArray, swap arguments.
	(mexGetVariablePtr): Rename from mexGetArrayPtr, swap arguments.
	(mexErrMsgIdAndTxt, mexWarnMsgIdAndTxt): New functions.
	* mexproto.h: Provide decls.
	Include decls for as yet unsupported functions.
	* mex.h: Add enum for class ids.
	(mexGetArray): Define using mexGetVariable.
	(mexGetArrayPtr): Define using mexGetVariablePtr.

2006-07-02  John W. Eaton  <jwe@octave.org>

	* ov-str-mat.cc (CHAR_MATRIX_CONV): Omit "warn-" from label.
	* parse.y (maybe_warn_variable_switch_label): Likewise.

2006-07-01  John W. Eaton  <jwe@octave.org>

	* parse.y (finish_cell): Use finish_matrix to do constant folding.

2006-06-30  John W. Eaton  <jwe@octave.org>

	* ov-base.cc (octave_base_value::numeric_assign):
	Avoid memory leak when converting LHS.

	* oct-hist.cc (do_history, initialize_history, Fhistory_file):
	Don't perform tilde expansion on history file name.

	* syscalls.cc (Flstat, Fstat): Don't perform tilde expansion on arg.
	* input.cc (Fread_readline_init_file): Likewise.
	* dirfns.cc (Fcd, Freaddir): Likewise.

	* load-path.cc (genpath): Don't perform tilde expansion on name.
	(load_path::do_add): Don't warn about moving "." to front of list.

2006-06-28  John W. Eaton  <jwe@octave.org>

	* load-path.cc (Faddpath): Don't treat "." specially here.
	Don't check directory status here.
	(Fpath): Handle all args.  Don't treat "." specially here.
	(Faddpath, Frmpath): Delete unused variable xpath.
	(load_path::do_add): New function.
	(load_path::do_prepend, load_path::do_append): Use it.
	(load_path::do_remove): Really prevent removal of ".".
	(load_path::do_clear): Clearing doesn't remove ".".
	(load_path::append, load_path::prepend, load_path::do_append,
	load_path::do_prepend): New arg, warn.

	* load-path.h: Fix decls.

	* DLD-FUNCTIONS/regexp.cc (octregexp_list):
	Avoid bug in older versions of g++.

2006-06-27  John W. Eaton  <jwe@octave.org>

	* Makefile.in (clean): Also remove $(DLD_OBJ).

	* sighandlers.cc (install_signal_handlers): Don't handle SIGPROF.
	From Geoffrey Knauth <geoff@knauth.org>.

2006-06-25  John W. Eaton  <jwe@octave.org>

	* mk-pkg-add: Avoid apparently non-portable sed patterns.

2006-06-21  John W. Eaton  <jwe@octave.org>

	* variables.cc (do_who): Handle mex.

	* symtab.cc (SYMBOL_DEF::type_as_string, SYMBOL_DEF::which,
	symbol_table::clear_functions, table::clear_function,
	symbol_table::clear_function_pattern): 
	Handle mex.
	* symtab.h (symbol_record::is_mex_function,
	symbol_record::symbol_def::is_mex_function): New functions.
	(symbol_record::TYPE): New value, MEX_FUNCTION.
	(symbol_record::is_function, symbol_table::user_function_name_list,
	SYMTAB_ALL_TYPES): Handle MEX_FUNCTION.

	* ov-fcn.h (octave_function::octave_function):
	Provide default for doc string arg.

	* defun.cc (install_mex_function): New function.
	* defun-int.h: Provide decl.

	* ov-builtin.cc (any_arg_is_magic_colon): Delete.
	(octave_builtin::do_multi_index_op): Call has_magic_colon method
	for args instead of any_arg_is_magic_colon.

	* ov-base.h (octave_base_value::is_mex_function): New function.
	* ov.h (octave_value::is_mex_function): New function.

	* ov-mex-fcn.h, ov-mex-fcn.cc: New files.

	* parse.y (load_fcn_from_file): Also handle .mex files.

	* dynamic-ld.cc (octave_dynamic_loader::load_oct): Rename from load.
	(octave_dynamic_loader::do_load_oct): Rename from do_load.
	* dynamic-ld.h: Fix decls.

	* utils.cc (mex_file_in_path): New function.
	* utils.h: Provide decl.

	* variables.cc (symbol_out_of_date): Also handle mex files.

	* load-path.cc (load_path::dir_info::get_file_list,
	load_path::dir_info::get_private_function_map,
	load_path::do_find_fcn, load_path::add_to_fcn_map,
	load_path::do_display): Also handle mex files.
	* load-path.h (load_path::find_mex_file): New static function.
	(load_path::MEX_FILE): New static data member.
	(load_path::do_find_fcn): By default, also look for mex files.

	* matrix.h, mex.h, mex.cc: New files from Octave Forge
	* mexproto.h: New file, extracted from mex.h.
	* Makefile.in: Add them to the appropriate lists.

2006-06-20  John W. Eaton  <jwe@octave.org>

	* ov-re-mat.cc (octave_matrix::convert_to_str_internal):
	Don't lose empty dimensions on conversion to char.

	* strfns.cc (Fstrcmp): Handle comparison of cellstr and empty string.

2006-06-16  John W. Eaton  <jwe@octave.org>

	* parse.y (%union): New type, anon_fcn_handle_type.
	(anon_fcn_handle): Now anon_fcn_handle_type, not tree_constant_type.
	(make_anon_fcn_handle):	Return tree_anon_fcn_handle, not tree_constant.

	* comment-list.cc (octave_comment_list::dup): New function.
	* comment-list.h: Provide decl.

	* pt-cmd.cc (tree_no_op_command::dup): New function.
	* pt-cmd.h: Provide decl.
	(tree_command::dup): New pure virtual function.

	* pt-loop.cc (tree_while_command::dup, tree_do_until_command::dup,
	tree_simple_for_command::dup, tree_complex_for_command::dup):
	New functions.
	* pt-loop.h: Provide decls.

	* pt-jump.cc (tree_break_command::dup, tree_continue_command::dup,
	tree_return_command::dup): New functions.
	* pt-jump.h: Provide decls.

	* pt-except.cc (tree_try_catch_command::dup,
	tree_unwind_protect_command::dup): New functions.
	* pt-except.h: Provide decls.

	* pt-select.cc (tree_if_clause::dup, tree_if_command_list::dup,
	tree_if_command::dup, tree_switch_case::dup,
	tree_switch_case_list::dup, tree_switch_command::dup):
	New functions.
	* pt-select.h: Provide decls.

	* pt-decl.cc (tree_decl_elt::dup, tree_decl_init_list::dup,
	tree_global_command::dup, tree_static_command::dup): New functions.
	* pt-decl.h: Provide decls.
	
	* pt-exp.h (tree_expression::copy_base): New function.
	(tree_expression::dup): New pure virtual function.
	
	* pt-fcn-handle.cc, pt-fcn-handle.h (tree_anon_fcn_handle): New class.

	* pt-fcn-handle.cc (tree_function_handle::dup): New function.
	* pt-fcn-handle.h: Provide decl.

	* pt-colon.cc (tree_colon_expression::dup): New function.
	* pt-colon.h: Provide decl.
	(tree_colon_expression::tree_colon_expression (tree_expression *,
	tree_expression *, tree_expression *, int, int)): New constructor.
	
	* pt-const.cc (tree_constant::dup): New function.
	(tree_constant::tree_constant (const octave_value&, const
	std::string&, int, int)): New constructor.
	* pt-const.h: Provide decls.

	* pt-idx.cc (tree_index_expression::dup): New functoin.
	(tree_index_expression::tree_index_expression (int, int)):
	New constructor.
	* pt-idx.h: Provide decls.

	* pt-arg-list.cc (tree_argument_list::dup): New function.
	* pt-arg-list.h: Provide decl.

	* pt-id.cc (tree_identifier::dup): New function.
	* pt-id.h: Provide decl.

	* pt-cell.cc (tree_cell::dup): New function.
	* pt-cell.h: Provide decl.

	* pt-mat.cc (tree_matrix::dup): New function.
	* pt-mat.h: Provide decl.

	* pt-assign.cc (tree_simple_assignment::dup,
	tree_multi_assignment::dup): New functions.
	* pt-assign.h: Provide decls.

	* pt-binop.cc (tree_binary_expression::dup,
	tree_boolean_expression::dup): New functions.
	* pt-binop.h: Provide decls.
	
	* pt-unop.cc (tree_prefix_expression::dup,
	tree_postfix_expression::dup): New functions.
	* pt-unop.h: Provide decls.

	* pt-stmt.cc (tree_statement::dup, tree_statement_list::dup):
	New functions.
	* pt-stmt.h: Provide decls.

	* pt-misc.cc (tree_parameter_list::dup,
	tree_return_list::dup): New functions.
	* pt-misc.h: Provide decls.

	* symtab.cc (symbol_table::dup, symbol_table::inherit,
	symbol_record:is_automatic_variable): New functions.
	* symtab.h: Provide decls.
	(symbol_record::automatic_variable): New data member.
	(symbol_record::symbol_record): Initialize it.
	(symbol_record::mark_as_automatic_variable): New function.

	* pt-walk.h (tree_walker::visit_anon_fcn_handle):
	New pure virtual function.
	* pt-pr-code.cc (tree_print_code::visit_anon_fcn_handle): New function.
	* pt-pr-code.h: Provide decl.
	* pt-bp.cc (tree_breakpoint::visit_anon_fcn_handle): New function.
	* pt-bp.h: Provide decl.
	* pt-check.cc (tree_checker::visit_anon_fcn_handle): New function.
	* pt-check.h: Provide decl.

2006-06-13  John W. Eaton  <jwe@octave.org>

	* pt-stmt.cc (tree_statement_list::eval): Revise previous change
	to preserve return-last-value-computed semantics.

	* DLD-FUNCTIONS/fsolve.cc (hybrd_info_to_fsolve_info):
	Warn about invalid values of INFO from MINPACK instead of calling
	panic_impossible.

	* ov-usr-fcn.cc (octave_user_function::do_multi_index_op):
	Don't assign values returned by evaluation of cmd_list to
	temporary variable.  Delete unused variable last_computed_value.  

	* pt-stmt.cc (tree_statement_list::eval):
	Clear retval before each statement is evaluated.

2006-06-12  John W. Eaton  <jwe@octave.org>

	* unwind-prot.h (unwind_protect_fptr): New macro.
	* load-path.cc (load_path::do_set): Use it instead of
	unwind_protect_ptr when protecting add_hook function pointer.
	* dynamic-ld.cc (octave_dynamic_loader::do_load): Use FCN_PTR_CAST
	here instead of reinterpret_cast.

2006-06-09  John W. Eaton  <jwe@octave.org>

	* version.h (OCTAVE_VERSION): Now 2.9.6+.

2006-06-09  John W. Eaton  <jwe@octave.org>

	* version.h (OCTAVE_VERSION): Now 2.9.6.
	(OCTAVE_API_VERSION): Now api-v19.
	(OCTAVE_RELEASE_DATE): Now 2006-06-09.

	* ov-cell.cc (cell::save_hdf5): Use leading zeros to keep elements
	correctly sorted in the HDF file.
	From Stephen Fegan <sfegan@astro.ucla.edu>.

2006-06-08  John W. Eaton  <jwe@octave.org>

	* ov-usr-fcn.cc (function::do_multi_index_op): Append contents of
	varargout to ret_list before calling convert_to_const_vector.

	* pt-arg-list.cc (tree_argument_list::convert_to_const_vector):
	No need to handle all_va_args here.

	* ov-va-args.h, ov-va-args.cc: Delete files.
	* Makefile.in (OV_INCLUDES, OV_SRC): Delete from the lists.

	* ov.cc (install_types): Don't register octave_all_va_args type.
	* ov.h, ov.cc (octave_value::octave_value (octave_value::all_va_args)):
	Delete.
	* ov.h (octave_value::is_all_va_args): Delete.
	(octave_value::enum all_va_args): Delete.

	* pt-pr-code.cc	(tree_print_code::visit_octave_user_function_header):
	Print "varargout" and "varargin" instead of "...".

	* pt-misc.h, pt-misc.cc	(tree_parameter_list::convert_to_const_vector):
	Arg is now const Cell& instead of tree_va_return_list*.

	* ov-usr-fcn.h, ov-usr-fcn.cc (octave_user_function::vr_list,
	octave_user_function::curr_va_arg_number):
	Delete variables and all uses.
	(octave_function::octave_va_start, octave_function::octave_va_arg,
	octave_function::octave_vr_val,	octave_function::has_varargout):
	Delete functions and all uses.

	* ov-fcn.h (octave_function::octave_va_start,
	octave_function::octave_va_arg, octave_function::octave_vr_val,
	octave_function::has_varargout): Delete.

	* ov-usr-fcn.cc (Fva_arg, Fva_start, Fvr_val): Delete.

	* lex.l ({EL}): Don't handle "..." as varargin or varargout.
	* octave.gperf (all_va_args, ALL_VA_ARGS, all_va_args_kw): Delete.
	* parse.y (ALL_VA_ARGS): Delete token and all uses.

	* defaults.h.in (OCTAVE_DATAROOTDIR): Substitute here.
	* toplev.cc (octave_config_info): Add datarootdir to the struct.

2006-06-07  John W. Eaton  <jwe@octave.org>

	* ov-cell.cc (octave_cell::subsasgn): Handle c-s list object on
	RHS if last index is '{'.
	* pt-assign.cc (tree_multi_assignment::rvalue): Compute all LHS
	lvalues before evaluating RHS.  If LHS lvalue expects more than
	one output, pass octave_cs_list object to assign method.
	* pt-idx.cc (tree_index_expression::lvalue): Compute expected
	number of values if '{' index appears last.
	* pt-arg-list.cc (tree_argument_list::lvalue_list): New function.
	* pt-arg-list.h (tree_argument_list::lvalue_list): Provide decl.
	* oct-lvalue.h (octave_lvalue::nel): New data member.
	(octave_lvalue::numel): New functions
	* oct-obj.cc (octave_value_list::has_magic_colon): New function.
	* oct-obj.h (octave_value_list::has_magic_colon): Provide decl.
	* pt-arg-list.cc (tree_argument_list::nargout_count): Delete function.
	* pt-arg-list.h (tree_argument_list::nargout_count): Delete decl.

2006-06-06  John W. Eaton  <jwe@octave.org>

	* sighandlers.cc (generic_sig_handler, sigint_handler):
	Use strsignal instead of accessing sys_siglist directly.
	* siglist.h: Use HAVE_DECL_SYS_SIGLIST instead of SYS_SIGLIST_DECLARED.
	* siglist.c: Update sys_siglist.
	(init_signals, strsignal): New functions.
	* octave.cc (octave_main): Call init_signals here.

	* toplev.cc (Foctave_config_info): Delete LIBPLPLOT from struct.

2006-06-05  John W. Eaton  <jwe@octave.org>

	* pt-assign.cc (Vprint_rhs_assign_val): Delete variable.
	(Fprint_rhs_assign_val): Delete function.
	tree_simple_assignment::rvalue, tree_multi_assignment::rvalue):
	No special case for Vprint_rhs_assign_val.

2006-06-02  John W. Eaton  <jwe@octave.org>

	* pt-arg-list.h (tree_argument_list::mark_as_simple_assign_lhs,
	tree_argument_list::is_simple_assign_lhs): New functions.
	(tree_argument_list::simple_assign_lhs): New data member.
	* parse.y (assign_lhs): Classify LHS here.
	(make_assign_op): Create simple or multi assign op based on
	classification of LHS, not its length.

2006-06-01  John W. Eaton  <jwe@octave.org>

	* Makefile.in (DLD_XSRC): Add __pchip_deriv__.cc to the list.

2006-06-01  David Bateman  <dbateman@free.fr>

	* DLD-FUNCTIONS/__pchip_deriv__.cc: New file.

2006-05-31  John W. Eaton  <jwe@octave.org>

	* load-path.h (load_path::set_command_line_path): Make it
	additive.

2006-05-26  John W. Eaton  <jwe@octave.org>

	* load-path.cc (genpath, maybe_add_path_elts, Fgenpath, Frehash,
	Fpath, Fpathdef): Move here from defaults.cc.
	* load-path.cc, load-path.h: New files.
	* Makefile.in (INCLUDES, DIST_SRC): Add them to the lists.
	* help.cc (make_name_list, simple_help): Use load_path instead of
	octave_fcn_file_name_cache.
	(Flookfor): Use load_path instead of Vload_path_dir_path.
	* octave.cc (octave_main):  Likewise.
	* parse.y (load_fcn_from_file): Likewise.
	* utils.cc (Ffile_in_loadpath, file_in_path, fcn_file_in_path,
	oct_file_in_path): Likewise.
	* variables.cc (symbol_exist, symbol_out_of_date): Likewise.
	* DLD-FUNCTIONS/fftw_wisdom.cc (Ffftw_wisdom): Likewise.
	* utils.cc (Ffind_first_of_in_loadpath): Delete.
	* input.cc (octave_gets): Call load_path::update here.
	* dirfns.cc (octave_change_to_directory): Likewise.
	* defaults.cc (VLOADPATH, VDEFAULT_LOADPATH, Vload_path_dir_path,
	maybe_add_or_del_packages, update_load_path_dir_path,
	execute_default_pkg_add_files, set_load_path): Delete.
	* defaults.h.in (Vload_path_dir_path,
	execute_default_pkg_add_files, maybe_add_default_load_path,
	set_load_path): Delete decls.

	* fn-cache.h, fn-cache.cc: Delete.
	* Makefile.in (INCLUDES, DIST_SRC): Remove them from the lists.

	* input.cc (octave_gets): Only update Vlast_prompt_time if we are
	interactive and actually printing a prompt.  Don't print prompt if
	reading startup files or command line files.  Initialize
	Vlast_prompt_time to 0.

	* pr-output.cc (set_format): Always include space for sign.
	
2006-05-23  John W. Eaton  <jwe@octave.org>

	* load-save.cc (Fsave): Use tellp instead of pubseekoff to
	determine whether we are at beginning of file.

	* ov-intx.h (OCTAVE_VALUE_INT_SCALAR_T::increment, 
	OCTAVE_VALUE_INT_SCALAR_T::decrement,
	OCTAVE_VALUE_INT_MATRIX_T::increment,
	OCTAVE_VALUE_INT_MATRIX_T::decrement): New functions.
	* OPERATORS/op-int.h (OCTAVE_INSTALL_M_INT_UNOPS,
	OCTAVE_INSTALL_S_INT_UNOPS, OCTAVE_M_INT_UNOPS,
	OCTAVE_S_INT_UNOPS): Uncomment increment and decrement ops.

2006-05-19  John W. Eaton  <jwe@octave.org>

	* symtab.cc (SYMBOL_DEF::document): If definition is a function,
	also pass doc string to function object.
	* ov-fcn.h (octave_function::document): New function.
	* defun.cc (print_usage (octave_function *)): New function.
	(print_usage (void)): New function.
	(print_usage (const std::string&)): Provide for backward
	compatibility.  Ignore arg.
	(Fprint_usage): Don't accept arg.
	* defun-int.h (print_usage (void)): Provide decl.
	(print_usage (const std::string&)): Tag with GCC_ATTR_DEPRECATED.

2006-05-04  David Bateman  <dbateman@free.fr>

	* DLD-FUNCTIONS/conv2.cc: New file from Octave Forge.
	* Makefile.in (DLD_XSRC): Add it to the list

2006-05-17  Bill Denney  <bill@givebillmoney.com>