Mercurial > hg > octave-lyh

changeset 16403:1de4ec2a856d

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Matlab compatibility for strsplit() * scripts/strings/strsplit.m: Matlab compatible version. * NEWS: Mention break in backward compatibility. * scripts/deprecated/javafields.m: Modify call to strsplit(). * scripts/deprecated/javamethods.m: ditto * scripts/general/fieldnames.m: ditto * scripts/general/int2str.m: ditto * scripts/general/methods.m: ditto * scripts/general/num2str.m: ditto * scripts/help/gen_doc_cache.m: ditto * scripts/help/help.m: ditto * scripts/help/lookfor.m: ditto * scripts/io/strread.m: ditto * scripts/java/javaclasspath.m: ditto * scripts/miscellaneous/compare_versions.m: ditto * scripts/miscellaneous/computer.m: ditto * scripts/miscellaneous/fact.m: ditto * scripts/miscellaneous/tar.m: ditto * scripts/miscellaneous/unpack.m: ditto * scripts/miscellaneous/what.m: ditto * scripts/miscellaneous/zip.m: ditto * scripts/pkg/private/configure_make.m: ditto * scripts/pkg/private/fix_depends.m: ditto * scripts/pkg/private/generate_lookfor_cache.m: ditto * scripts/pkg/private/list_forge_packages.m: ditto * scripts/pkg/private/unload_packages.m: ditto * scripts/pkg/private/write_index.m: ditto * scripts/plot/private/__file_filter__.m: ditto * scripts/plot/private/__fltk_file_filter__.m: ditto * scripts/plot/private/__go_draw_axes__.m: ditto * scripts/plot/private/__next_line_style__.m: ditto * scripts/strings/untabify.m: ditto * scripts/testfun/rundemos.m: ditto * scripts/testfun/runtests.m: ditto
author Ben Abbott <bpabbott@mac.com>
date Sun, 31 Mar 2013 19:19:04 -0400
parents 7fa90eb41240
children 24e755668237
files NEWS scripts/deprecated/javafields.m scripts/deprecated/javamethods.m scripts/general/fieldnames.m scripts/general/int2str.m scripts/general/methods.m scripts/general/num2str.m scripts/help/gen_doc_cache.m scripts/help/help.m scripts/help/lookfor.m scripts/io/strread.m scripts/java/javaclasspath.m scripts/miscellaneous/compare_versions.m scripts/miscellaneous/computer.m scripts/miscellaneous/fact.m scripts/miscellaneous/tar.m scripts/miscellaneous/unpack.m scripts/miscellaneous/what.m scripts/miscellaneous/zip.m scripts/pkg/private/configure_make.m scripts/pkg/private/fix_depends.m scripts/pkg/private/generate_lookfor_cache.m scripts/pkg/private/unload_packages.m scripts/pkg/private/write_index.m scripts/plot/private/__file_filter__.m scripts/plot/private/__fltk_file_filter__.m scripts/plot/private/__go_draw_axes__.m scripts/plot/private/__next_line_style__.m scripts/strings/strsplit.m scripts/testfun/rundemos.m scripts/testfun/runtests.m
diffstat 31 files changed, 290 insertions(+), 97 deletions(-) [+]
line wrap: on
line diff
--- a/NEWS
+++ b/NEWS
@@ -17,6 +17,45 @@
 Summary of important user-visible changes for version 3.8:
 ---------------------------------------------------------
 
+ ** strsplit has been modified to be compatible with Matlab.  There
+    are three instances where backward compatibility is broken.
+
+    (1) Delimiters are now string vectors, not scalars.
+
+    Octave's conventioal behavior
+
+      strsplit ("1 2, 3", ", ")
+      ans = 
+      {
+       [1,1] = 1
+       [1,2] = 2
+       [1,3] = 
+       [1,4] = 3
+      }
+
+    Matlab compatible behavior
+
+      strsplit ("1 2, 3", ", ")
+      ans = 
+      {
+       [1,1] = 1 2
+       [1,2] = 3
+      }
+
+    (2) By default, Matlab treats consecutive delimiters are as a single
+    delimiter.  By default, Octave's conventional behavior was to return
+    an empty string.
+
+    (3) Octave's conventional implementation supported splitting 2D
+    character arrays.  The Matlab compatible version requires the input
+    string be a row vector.
+
+    Where the conventional behavior is desired, the in-line function below
+    may be substituted.
+
+    cstrsplit = @(str,del, collapsedelimiters = false) strsplit (strjoin ...
+      (cellstr (str), del(1)), num2cell (del(:)), collapsedelimiters);
+
  ** Octave now supports nested functions with scoping rules that are
     compatible with Matlab.  A nested function is one declared and defined
     within the body of another function.  The nested function is only
--- a/scripts/deprecated/javafields.m
+++ b/scripts/deprecated/javafields.m
@@ -40,7 +40,7 @@
   endif
   
   c_methods = javaMethod ("getFields", "org.octave.ClassHelper", javaobj);
-  method_list = strsplit (c_methods, ';');
+  method_list = strsplit (c_methods, ';', false);
 
   if (nargout == 0)
     if (! isempty (method_list))
--- a/scripts/deprecated/javamethods.m
+++ b/scripts/deprecated/javamethods.m
@@ -40,7 +40,7 @@
   endif
 
   cls_methods = javaMethod ("getMethods", "org.octave.ClassHelper", classname);
-  method_list = strsplit (cls_methods, ';');
+  method_list = strsplit (cls_methods, ';', false);
 
   if (nargout == 0)
     if (! isempty (method_list))
--- a/scripts/general/fieldnames.m
+++ b/scripts/general/fieldnames.m
@@ -52,7 +52,7 @@
       obj = class (obj);
     endif
     names_str = javaMethod ("getFields", "org.octave.ClassHelper", obj);
-    names = strsplit (names_str, ';');
+    names = strsplit (names_str, ';', false);
   else
     error ("fieldnames: Invalid input argument"); 
   endif
--- a/scripts/general/int2str.m
+++ b/scripts/general/int2str.m
@@ -73,7 +73,7 @@
   endif
   tmp = sprintf (fmt, permute (n, [2, 1, 3 : nd]));
   tmp(end) = "";
-  retval = char (strsplit (tmp, "\n"));
+  retval = char (strsplit (tmp, "\n", false));
 
 endfunction
 
--- a/scripts/general/methods.m
+++ b/scripts/general/methods.m
@@ -43,14 +43,14 @@
     mtds_list = __methods__ (obj);
     if (isempty (mtds_list))
       mtds_str = javaMethod ("getMethods", "org.octave.ClassHelper", obj);
-      mtds_list = strsplit (mtds_str, ';');
+      mtds_list = strsplit (mtds_str, ';', false);
     endif
   elseif (isjava (obj))
     ## FIXME: Function prototype that excepts java obj exists, but doesn't
     ##        work if obj is java.lang.String.  Convert obj to classname.
     obj = class (obj);
     mtds_str = javaMethod ("getMethods", "org.octave.ClassHelper", obj);
-    mtds_list = strsplit (mtds_str, ';');
+    mtds_list = strsplit (mtds_str, ';', false);
   else
     error ("methods: Invalid input argument");
   endif
--- a/scripts/general/num2str.m
+++ b/scripts/general/num2str.m
@@ -118,7 +118,7 @@
     fmt = cstrcat (deblank (repmat (fmt, 1, columns (x))), "\n");
     nd = ndims (x);
     tmp = sprintf (fmt, permute (x, [2, 1, 3:nd]));
-    retval = strtrim (char (strsplit (tmp(1:end-1), "\n")));
+    retval = strtrim (char (strsplit (tmp(1:end-1), "\n", false)));
   else   # Complex matrix input
     if (nargin == 2)
       if (ischar (arg))
@@ -164,7 +164,7 @@
     tmp = regexprep (tmp, " +i\n", "i\n");
     tmp = regexprep (tmp, "( +)i", "i$1");
 
-    retval = strtrim (char (strsplit (tmp(1:end-1), "\n")));
+    retval = strtrim (char (strsplit (tmp(1:end-1), "\n", false)));
   endif
 
 endfunction
--- a/scripts/help/gen_doc_cache.m
+++ b/scripts/help/gen_doc_cache.m
@@ -117,7 +117,7 @@
 function cache = gen_doc_cache_in_dir (directory)
 
   ## If 'directory' is not in the current path, add it so we search it
-  dir_in_path = ismember (directory, strsplit (path (), pathsep ()));
+  dir_in_path = ismember (directory, strsplit (path (), pathsep (), false));
 
   # dirs not in path
   if (! iscell (directory))
--- a/scripts/help/help.m
+++ b/scripts/help/help.m
@@ -142,7 +142,7 @@
   builtins = sprintf ("*** builtins:\n\n%s\n\n",
                       list_in_columns (__builtins__ ()));
 
-  dirs = strsplit (path, pathsep);
+  dirs = strsplit (path, pathsep, false);
   flist = "";
   for i = 2:numel (dirs)
     files = sort ({dir(fullfile (dirs{i}, "*.m")).name, ...
--- a/scripts/help/lookfor.m
+++ b/scripts/help/lookfor.m
@@ -66,10 +66,10 @@
   endif
 
   ## Search functions in new path dirs.
-  orig_path = strsplit (__pathorig__ (), pathsep ());
+  orig_path = strsplit (__pathorig__ (), pathsep (), false);
 
   ## ditto for path.
-  new_path = strsplit (path (), pathsep ());
+  new_path = strsplit (path (), pathsep (), false);
 
   ## scratch out directories already covered by orig_path.
   if (had_core_cache)
--- a/scripts/io/strread.m
+++ b/scripts/io/strread.m
@@ -621,13 +621,13 @@
                      strrep (words(icol, jptr), fmt_words{ii}, ...
                      [char(255) char(254)]);
                 wrds(2:2:2*numel (words(icol, jptr))-1) = char (255);
-                wrds = strsplit ([wrds{:}], char (255));
+                wrds = strsplit ([wrds{:}], char (255), false);
                 words(icol, jptr) = ...
                   wrds(find (cellfun ("isempty", strfind (wrds, char (254)))));
                 wrds(find (cellfun ("isempty", strfind (wrds, char (254))))) ...
                    = char (255);
                 words(icol+1, jptr) = strsplit (strrep ([wrds{2:end}], ...
-                   char (254), fmt_words{ii}), char (255));
+                   char (254), fmt_words{ii}), char (255), false);
                 ## Former trailing literal may now be leading for next specifier
                 --ii;
                 fwptr = [fwptr(1:ii) (++fwptr(ii+1:end))];
@@ -699,7 +699,7 @@
         case {"%0", "%1", "%2", "%3", "%4", "%5", "%6", "%7", "%8", "%9"}
           sw = regexp (fmt_words{m}, '\d', "once");
           ew = regexp (fmt_words{m}, '[nfudsq]') - 1;
-          nfmt = strsplit (fmt_words{m}(2:ew), ".");
+          nfmt = strsplit (fmt_words{m}(2:ew), ".", false);
           swidth = str2double (nfmt{1});
           switch fmt_words{m}(ew+1)
             case {"d", "u", "f", "n"}
@@ -772,7 +772,7 @@
   endif
 
   ## Split text string along delimiters
-  out = strsplit (text, sep, mult_dlms_s1);
+  out = strsplit (text, num2cell (sep(:)), mult_dlms_s1);
   if (index (sep, eol_char)); out = strrep (out, char (255), ''); endif
   ## In case of trailing delimiter, strip stray last empty word
   if (!isempty (out) && any (sep == text(end)))
--- a/scripts/java/javaclasspath.m
+++ b/scripts/java/javaclasspath.m
@@ -58,11 +58,11 @@
 
   ## dynamic classpath
   dynamic_path = javaMethod ("getClassPath", "org.octave.ClassHelper");
-  dynamic_path_list = strsplit (dynamic_path, pathsep ());
+  dynamic_path_list = strsplit (dynamic_path, pathsep (), false);
 
   ## static classpath
   static_path = javaMethod ("getProperty", "java.lang.System", "java.class.path");
-  static_path_list = strsplit (static_path, pathsep ());
+  static_path_list = strsplit (static_path, pathsep (), false);
   if (numel (static_path_list) > 1)
     ## remove first element (which is .../octave.jar)
     static_path_list(1) = [];
--- a/scripts/miscellaneous/compare_versions.m
+++ b/scripts/miscellaneous/compare_versions.m
@@ -113,8 +113,8 @@
     v2nochar = v2;
   endif
 
-  v1n = str2num (char (strsplit (v1nochar, ".")));
-  v2n = str2num (char (strsplit (v2nochar, ".")));
+  v1n = str2num (char (strsplit (v1nochar, ".", false)));
+  v2n = str2num (char (strsplit (v2nochar, ".", false)));
   if ((isempty (v1n) && isempty (v1c)) || (isempty (v2n) && isempty (v2c)))
     error ("compare_versions: given version strings are not valid: %s %s",
            v1, v2);
--- a/scripts/miscellaneous/computer.m
+++ b/scripts/miscellaneous/computer.m
@@ -49,7 +49,7 @@
 function [c, maxsize, endian] = computer (a)
 
   if (nargin == 1 && ischar (a) && strcmpi (a, "arch"))
-    tmp = strsplit (octave_config_info ("canonical_host_type"), "-");
+    tmp = strsplit (octave_config_info ("canonical_host_type"), "-", false);
     if (numel (tmp) == 4)
       c = sprintf ("%s-%s-%s", tmp{4}, tmp{3}, tmp{1});
     else
--- a/scripts/miscellaneous/fact.m
+++ b/scripts/miscellaneous/fact.m
@@ -250,7 +250,7 @@
 
 function out = wordwrap (w)
   cols = terminal_size ()(2);
-  wc = strsplit (w, " ");
+  wc = strsplit (w, " ", false);
   out = "\n";
   i = 1;
   numwords = numel (wc);
@@ -265,4 +265,4 @@
     i++;
   endwhile
   out = cstrcat (out, "\n");
-endfunction
\ No newline at end of file
+endfunction
--- a/scripts/miscellaneous/tar.m
+++ b/scripts/miscellaneous/tar.m
@@ -59,7 +59,7 @@
     if (output(end) == "\n")
       output(end) = [];
     endif
-    entries = strsplit (output, "\n");
+    entries = strsplit (output, "\n", false);
     entries = entries';
   endif
 
--- a/scripts/miscellaneous/unpack.m
+++ b/scripts/miscellaneous/unpack.m
@@ -194,7 +194,7 @@
     if (output(length (output)) == "\n")
       output(length (output)) = [];
     endif
-    files = parser (strsplit (output, "\n"))';
+    files = parser (strsplit (output, "\n", false))';
 
     ## Move files if necessary
     if (needmove)
--- a/scripts/miscellaneous/what.m
+++ b/scripts/miscellaneous/what.m
@@ -32,7 +32,7 @@
     d = pwd ();
   elseif (isempty (strfind (d, filesep ())))
     ## Find the appropriate directory on the path.
-    p = strtrim (strsplit (path (), pathsep ()));
+    p = strtrim (strsplit (path (), pathsep (), false));
     d = p{find (cellfun (@(x) ! isempty (strfind (x, d)), p))(end)};
   else
     [status, msg, msgid] = fileattrib (d);
--- a/scripts/miscellaneous/zip.m
+++ b/scripts/miscellaneous/zip.m
@@ -62,7 +62,7 @@
     if (entries(end) == "\n")
       entries(end) = [];
     endif
-    entries = strsplit (entries, "\n");
+    entries = strsplit (entries, "\n", false);
   endif
 
 endfunction
--- a/scripts/pkg/private/configure_make.m
+++ b/scripts/pkg/private/configure_make.m
@@ -91,7 +91,7 @@
       if (filenames(end) == "\n")
         filenames(end) = [];
       endif
-      filenames = strtrim (strsplit (filenames, "\n"));
+      filenames = strtrim (strsplit (filenames, "\n", false));
       delete_idx =  [];
       for i = 1:length (filenames)
         if (! all (isspace (filenames{i})))
--- a/scripts/pkg/private/fix_depends.m
+++ b/scripts/pkg/private/fix_depends.m
@@ -26,7 +26,7 @@
 ## This function returns a cell of structures with the following fields:
 ##   package, version, operator
 function deps_cell = fix_depends (depends)
-  deps = strtrim (strsplit (tolower (depends), ","));
+  deps = strtrim (strsplit (tolower (depends), ",", false));
   deps_cell = cell (1, length (deps));
 
   ## For each dependency.
--- a/scripts/pkg/private/generate_lookfor_cache.m
+++ b/scripts/pkg/private/generate_lookfor_cache.m
@@ -23,7 +23,7 @@
 ## @end deftypefn
 
 function generate_lookfor_cache (desc)
-  dirs = strtrim (strsplit (genpath (desc.dir), pathsep ()));
+  dirs = strtrim (strsplit (genpath (desc.dir), pathsep (), false));
   for i = 1 : length (dirs)
     gen_doc_cache (fullfile (dirs{i}, "doc-cache"), dirs{i});
   endfor
--- a/scripts/pkg/private/unload_packages.m
+++ b/scripts/pkg/private/unload_packages.m
@@ -35,7 +35,7 @@
   endfor
 
   ## Get the current octave path.
-  p = strtrim (strsplit (path (), pathsep ()));
+  p = strtrim (strsplit (path (), pathsep (), false));
 
   if (length (files) == 1 && strcmp (files{1}, "all"))
     ## Unload all.
--- a/scripts/pkg/private/write_index.m
+++ b/scripts/pkg/private/write_index.m
@@ -74,7 +74,7 @@
   if (! isfield (desc, "categories"))
     error ("the DESCRIPTION file must have a Categories field, when no INDEX file is given");
   endif
-  categories = strtrim (strsplit (desc.categories, ","));
+  categories = strtrim (strsplit (desc.categories, ",", false));
   if (length (categories) < 1)
       error ("the Category field is empty");
   endif
--- a/scripts/plot/private/__file_filter__.m
+++ b/scripts/plot/private/__file_filter__.m
@@ -81,7 +81,7 @@
   endswitch
 
   if (isempty (name))
-    extlist = strsplit (filterext, ";");
+    extlist = strsplit (filterext, ";", false);
     extlist = strrep (extlist, "*.", "");
     extlist = toupper (extlist);
     extlist(end+1, :) = repmat ({","}, 1, length (extlist));
--- a/scripts/plot/private/__fltk_file_filter__.m
+++ b/scripts/plot/private/__fltk_file_filter__.m
@@ -34,7 +34,7 @@
   for idx = 1 : r
 
     curr_ext = file_filter{idx, 1};
-    curr_ext = strsplit (curr_ext, ";");
+    curr_ext = strsplit (curr_ext, ";", false);
 
     if (length (curr_ext) > 1)
       curr_ext = regexprep (curr_ext, '\*\.', ',');
--- a/scripts/plot/private/__go_draw_axes__.m
+++ b/scripts/plot/private/__go_draw_axes__.m
@@ -2229,7 +2229,7 @@
   endif
   if (ischar (ticklabel))
     if (rows (ticklabel) == 1 && any (ticklabel == "|"))
-      ticklabel = strsplit (ticklabel, "|");
+      ticklabel = strsplit (ticklabel, "|", false);
     else
       ticklabel = cellstr (ticklabel);
     endif
--- a/scripts/plot/private/__next_line_style__.m
+++ b/scripts/plot/private/__next_line_style__.m
@@ -41,7 +41,7 @@
       elseif (reset || isempty (style_rotation))
         style_rotation = get (gca (), "linestyleorder");
         if (ischar (style_rotation))
-          style_rotation = strsplit (style_rotation, "|");
+          style_rotation = strsplit (style_rotation, "|", false);
         endif
         num_styles = length (style_rotation);
         style_index = 1;
--- a/scripts/strings/strsplit.m
+++ b/scripts/strings/strsplit.m
@@ -17,20 +17,33 @@
 ## <http://www.gnu.org/licenses/>.
 
 ## -*- texinfo -*-
-## @deftypefn  {Function File} {[@var{cstr}] =} strsplit (@var{s}, @var{sep})
-## @deftypefnx {Function File} {[@var{cstr}] =} strsplit (@var{s}, @var{sep}, @var{strip_empty})
-## Split the string @var{s} using one or more separators @var{sep} and return
-## a cell array of strings.  Consecutive separators and separators at
-## boundaries result in empty strings, unless @var{strip_empty} is true.
-## The default value of @var{strip_empty} is false.
+## @deftypefn  {Function File} {[@var{cstr}] =} strsplit (@var{s})
+## @deftypefnx {Function File} {[@var{cstr}] =} strsplit (@var{s}, @var{del})
+## @deftypefnx {Function File} {[@var{cstr}] =} strsplit (@var{s}, @var{del}, @var{collapsedelimiters})
+## @deftypefnx {Function File} {[@var{cstr}] =} strsplit (@dots{}, @var{name}, @var{value})
+## @deftypefnx {Function File} {[@var{cstr}, @var{matches}] =} strsplit (@dots{})
+## Split the string @var{s} using the delimiters specified by @var{del} and return
+## a cell array of strings.  For a single delimiter, @var{del} may be a string,
+## or a scalar cell-string.  For multible delimiters, @var{del} must be a cell-string
+## array.  Unless @var{collapsedelimiters} is specified to be @var{false}, consecutive
+## delimiters are collapsed into one.
 ##
-## 2-D character arrays are split at separators and at the original column
-## boundaries.
+## The second output, @var{matches}, returns the delmiters which were matched
+## in the original string.  The matched delimiters are uneffected by the
+## @var{collapsedelimiters}.
 ##
 ## Example:
 ##
 ## @example
 ## @group
+## strsplit ("a b c")
+##       @result{}
+##           @{
+##             [1,1] = a
+##             [1,2] = b
+##             [1,3] = c
+##           @}
+##
 ## strsplit ("a,b,c", ",")
 ##       @result{}
 ##           @{
@@ -39,80 +52,221 @@
 ##             [1,3] = c
 ##           @}
 ##
-## strsplit (["a,b" ; "cde"], ",")
+## strsplit ("a foo b,bar c", @{"\s", "foo", "bar"@})
+##       @result{}
+##           @{
+##             [1,1] = a
+##             [1,2] = b
+##             [1,3] = c
+##           @}
+##
+## strsplit ("a,,b, c", @{",", " "@}, false)
+##       @result{}
+##           @{
+##             [1,1] = a
+##             [1,2] = 
+##             [1,3] = b
+##             [1,4] = 
+##             [1,5] = c
+##           @}
+##
+## @end group
+## @end example
+##
+## Supported @var{name}/@var{value} pair arguments are;
+##
+## @itemize
+## @item @code{collapsedelimiters} may take the value of @var{true} or @var{false}
+## with the default being @var{false}.
+## @item @code{delimitertype} may take the value of @code{simple} or @code{regularexpression},
+## with the default being @code{simple}.
+## @end itemize
+## 
+## Example:
+##
+## @example
+## @group
+## strsplit ("a foo b,bar c", ",|\\s|foo|bar", "delimitertype", "regularexpression")
+##       @result{}
+##           @{
+##             [1,1] = a
+##             [1,2] = b
+##             [1,3] = c
+##           @}
+## 
+## strsplit ("a,,b, c", "[, ]", false, "delimitertype", "regularexpression")
+##       @result{}
+##           @{
+##             [1,1] = a
+##             [1,2] = 
+##             [1,3] = b
+##             [1,4] = 
+##             [1,5] = c
+##           @}
+## 
+## strsplit ("a,\t,b, c", @{',', '\s'@}, "delimitertype", "regularexpression")
 ##       @result{}
 ##           @{
 ##             [1,1] = a
 ##             [1,2] = b
-##             [1,3] = cde
+##             [1,3] = c
 ##           @}
 ## @end group
 ## @end example
-## @seealso{strtok}
+## 
+## @seealso{strtok, regexp}
 ## @end deftypefn
 
-function cstr = strsplit (s, sep, strip_empty = false)
+function [result, matches] = strsplit (str, del, varargin)
 
-  if (nargin < 2 || nargin > 3)
+  args.collapsedelimiters = true;
+  args.delimitertype = "simple";
+
+  [reg, params] = parseparams (varargin);
+
+  if (numel (reg) > 1)
     print_usage ();
-  elseif (! ischar (s) || ! ischar (sep))
-    error ("strsplit: S and SEP must be string values");
-  elseif (! isscalar (strip_empty))
-    error ("strsplit: STRIP_EMPTY must be a scalar value");
+  elseif (numel (reg) == 1)
+    if (islogical (reg{1}) || isnumeric (reg{1}))
+      args.collapsedelimiters = reg{1};
+    else
+      print_usage ();
+    endif
+  endif
+  fields = fieldnames (args);
+  for n = 1:2:numel(params)
+    if (any (strcmpi (params{n}, fields)))
+      args.(lower(params{n})) = params{n+1};
+    elseif (ischar (varargin{n}))
+      error ("strsplit:invalid_parameter_name",
+        sprintf ("strsplit: Invalid parameter name, `%s'", varargin{n}))
+    else
+      print_usage ();
+    endif
+  endfor
+
+  # Save the length of the "delimitertype" parameter
+  length_deltype = numel (args.delimitertype);
+
+  if (nargin == 1 || (nargin > 1 && (islogical (del) || isnumeric (del))))
+    if (nargin > 1)
+      ## Second input is the "collapsedelimiters" parameter
+      args.collapsedelimiters = del;
+    endif
+    ## Set proper default for the delimiter type
+    if (strncmpi (args.delimitertype, "simple", numel (args.delimitertype)))
+      del = {" ","\f","\n","\r","\t","\v"};
+    else
+      del = "\\s";
+    endif
+  endif
+
+  if (nargin < 1)
+    print_usage ();
+  elseif (! ischar (str) || (! ischar (del) && ! iscellstr (del)))
+    error ("strsplit: S and DEL must be string values");
+  elseif (rows (str) > 1)
+    error ("strsplit: S must be a string value");
+  elseif (! isscalar (args.collapsedelimiters))
+    error ("strsplit: COLLAPSEDELIMITERS must be a scalar value");
+  endif
+
+  if (strncmpi (args.delimitertype, "simple", length_deltype))
+    if (iscellstr (del))
+      del = cellfun (@(x) regexp2simple (x, false), del, "uniformoutput",
+        false);
+    else
+      del = regexp2simple (del, false);
+    endif
   endif
 
-  if (isempty (s))
-    cstr = cell (size (s));
-  else
-    if (rows (s) > 1)
-      ## For 2-D arrays, add separator character at line boundaries
-      ## and transform to single string
-      s(:, end+1) = sep(1);
-      s = reshape (s.', 1, numel (s));
-      s(end) = []; 
+  if (isempty (str))
+    result = {str};
+  elseif (strncmpi (args.delimitertype, "regularexpression", length_deltype)
+          || strncmpi (args.delimitertype, "simple", length_deltype))
+    if (iscellstr (del))
+      del = sprintf ('%s|', del{:});
+      del(end) = [];
     endif
-
-    ## Split s according to delimiter
-    if (isscalar (sep))
-      ## Single separator
-      idx = find (s == sep);
-    else
-      ## Multiple separators
-      idx = strchr (s, sep);
+    [result, ~, ~, ~, matches] = regexp (str, del, "split");
+    if (args.collapsedelimiters)
+      result(cellfun (@isempty, result)) = [];
+    endif
+    if (strncmpi (args.delimitertype, "simple", length_deltype))
+      matches = cellfun (@(x) regexp2simple (x, true), matches,
+        "uniformoutput", false);
     endif
+  else
+    error ("strsplit:invalid_delimitertype", 
+      sprintf ("strsplit: Invalid DELIMITERTYPE"))
+  endif
+endfunction
 
-    ## Get substring lengths.
-    if (isempty (idx))
-      strlens = length (s);
-    else
-      strlens = [idx(1)-1, diff(idx)-1, numel(s)-idx(end)];
-    endif
-    ## Remove separators.
-    s(idx) = [];
-    if (strip_empty)
-      ## Omit zero lengths.
-      strlens = strlens(strlens != 0);
-    endif
-
-    ## Convert!
-    cstr = mat2cell (s, 1, strlens);
+function str = regexp2simple (str, reverse = false)
+  rep = {'\', '[', ']', '{', '}', '$', '^', '(', ')', '*', '+', '.', '?', '|'};
+  if (reverse)
+    ## backslash must go last
+    for r = numel(rep):-1:1
+      str = strrep (str, [char(92), rep{r}], rep{r});
+    endfor
+  else
+    ## backslash must go first
+    for r = 1:numel(rep)
+      str = strrep (str, rep{r}, [char(92), rep{r}]);
+    endfor
   endif
-
 endfunction
 
+%!shared str
+%! str = "The rain in Spain stays mainly in the plain.";
+% Split on all whitespace.
+%!assert (strsplit (str), {"The", "rain", "in", "Spain", "stays", ...
+%! "mainly", "in", "the", "plain."})
+% Split on "ain".
+%!assert (strsplit (str, "ain"), {"The r", " in Sp", " stays m", ...
+%!  "ly in the pl", "."})
+% Split on " " and "ain" (treating multiple delimiters as one).
+%!test
+%! s = strsplit (str, '\s|ain', true, "delimitertype", "r");
+%! assert (s, {"The", "r", "in", "Sp", "stays", "m", "ly", "in", "the", "pl", "."})
+%!test
+%! s = strsplit (str, "\\s|ain", true, "delimitertype", "r");
+%! assert (s, {"The", "r", "in", "Sp", "stays", "m", "ly", "in", "the", "pl", "."})
+%!test
+%! [s, m] = strsplit (str, {"\\s", "ain"}, true, "delimitertype", "r");
+%! assert (s, {"The", "r", "in", "Sp", "stays", "m", "ly", "in", "the", "pl", "."})
+%! assert (m, {" ", "ain", " ", " ", "ain", " ", " ", "ain", " ", " ", " ", "ain"})
+% Split on " " and "ain", and treat multiple delimiters separately.
+%!test
+%! [s, m] = strsplit (str, {" ", "ain"}, "collapsedelimiters", false);
+%! assert (s, {"The", "r", "", "in", "Sp", "", "stays", "m", "ly", "in", "the", "pl", "."})
+%! assert (m, {" ", "ain", " ", " ", "ain", " ", " ", "ain", " ", " ", " ", "ain"})
 
+%!assert (strsplit ("road to hell"), {"road", "to", "hell"})
 %!assert (strsplit ("road to hell", " "), {"road", "to", "hell"})
-%!assert (strsplit ("road to^hell", " ^"), {"road", "to", "hell"})
-%!assert (strsplit ("road   to--hell", " -", true), {"road", "to", "hell"})
-%!assert (strsplit (["a,bc";",de"], ","), {"a", "bc", char(ones(1,0)), "de "})
-%!assert (strsplit (["a,bc";",de"], ",", true), {"a", "bc", "de "})
-%!assert (strsplit (["a,bc";",de"], ", ", true), {"a", "bc", "de"})
+%!assert (strsplit ("road to^hell", {" ","^"}), {"road", "to", "hell"})
+%!assert (strsplit ("road   to--hell", {" ","-"}, true), {"road", "to", "hell"})
+%!assert (strsplit (["a,bc,,de"], ",", false), {"a", "bc", "", "de"})
+%!assert (strsplit (["a,bc,de"], ",", true), {"a", "bc", "de"})
+%!assert (strsplit (["a,bc,de"], {","," "}, true), {"a", "bc", "de"})
+%!test
+%! [s, m] = strsplit ("hello \t world", 1);
+%! assert (s, {"hello", "world"});
+%! assert (m, {" ", "\t", " "});
+
+%!assert (strsplit ("road to hell", " ", "delimitertype", "r"), {"road", "to", "hell"})
+%!assert (strsplit ("road to^hell", '\^| ', "delimitertype", "r"), {"road", "to", "hell"})
+%!assert (strsplit ("road to^hell", "[ ^]", "delimitertype", "r"), {"road", "to", "hell"})
+%!assert (strsplit ("road   to--hell", "[ -]", false, "delimitertype", "r"), {"road", "", "", "to", "", "hell"})
+%!assert (strsplit (["a,bc,de"], ",", "delimitertype", "r"), {"a", "bc", "de"})
+%!assert (strsplit (["a,bc,,de"], ",", false, "delimitertype", "r"), {"a", "bc", "", "de"})
+%!assert (strsplit (["a,bc,de"], ",", true, "delimitertype", "r"), {"a", "bc", "de"})
+%!assert (strsplit (["a,bc,de"], "[, ]", true, "delimitertype", "r"), {"a", "bc", "de"})
+%!assert (strsplit ("hello \t world", 1, "delimitertype", "r"), {"hello", "world"});
 
 %% Test input validation
 %!error strsplit ()
-%!error strsplit ("abc")
 %!error strsplit ("abc", "b", true, 4)
-%!error <S and SEP must be string values> strsplit (123, "b")
-%!error <S and SEP must be string values> strsplit ("abc", 1)
-%!error <STRIP_EMPTY must be a scalar value> strsplit ("abc", "def", ones (3,3))
+%!error <S and DEL must be string values> strsplit (123, "b")
+%!error <COLLAPSEDELIMITERS must be a scalar value> strsplit ("abc", "def", ones (3,3))
 
--- a/scripts/testfun/rundemos.m
+++ b/scripts/testfun/rundemos.m
@@ -30,7 +30,7 @@
 function rundemos (directory)
 
   if (nargin == 0)
-    dirs = strsplit (path (), pathsep ());
+    dirs = strsplit (path (), pathsep (), false);
   elseif (nargin == 1)
     if (is_absolute_filename (directory))
       dirs = {directory};
--- a/scripts/testfun/runtests.m
+++ b/scripts/testfun/runtests.m
@@ -30,7 +30,7 @@
 function runtests (directory)
 
   if (nargin == 0)
-    dirs = strsplit (path (), pathsep ());
+    dirs = strsplit (path (), pathsep (), false);
   elseif (nargin == 1)
     if (is_absolute_filename (directory))
       dirs = {directory};