Mercurial > hg > octave-nkf
view liboctave/data-conv.cc @ 4665:dece11da64ed ss-2-1-52
[project @ 2003-11-25 15:37:32 by jwe]
author | jwe |
---|---|
date | Tue, 25 Nov 2003 15:37:32 +0000 |
parents | 6cb22b9e3942 |
children | 44046bbaa52c |
line wrap: on
line source
/* Copyright (C) 1996, 1997 John W. Eaton This file is part of Octave. Octave is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2, or (at your option) any later version. Octave is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with Octave; see the file COPYING. If not, write to the Free Software Foundation, 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */ #ifdef HAVE_CONFIG_H #include <config.h> #endif #include <cctype> #include <iostream> #include "byte-swap.h" #include "data-conv.h" #include "lo-error.h" #define FIND_SIZED_INT_TYPE(VAL, BITS, TQ, Q) \ do \ { \ int sz = BITS / CHAR_BIT; \ if (sizeof (TQ char) == sz) \ VAL = oct_data_conv::dt_ ## Q ## char; \ else if (sizeof (TQ short) == sz) \ VAL = oct_data_conv::dt_ ## Q ## short; \ else if (sizeof (TQ int) == sz) \ VAL = oct_data_conv::dt_ ## Q ## int; \ else if (sizeof (TQ long) == sz) \ VAL = oct_data_conv::dt_ ## Q ## long; \ else \ VAL = oct_data_conv::dt_unknown; \ } \ while (0) #define FIND_SIZED_FLOAT_TYPE(VAL, BITS) \ do \ { \ int sz = BITS / CHAR_BIT; \ if (sizeof (float) == sz) \ VAL = oct_data_conv::dt_float; \ else if (sizeof (double) == sz) \ VAL = oct_data_conv::dt_double; \ else \ VAL = oct_data_conv::dt_unknown; \ } \ while (0) // I'm not sure it is worth the trouble, but let's use a lookup table // for the types that are supposed to be a specific number of bits // wide. Given the macros above, this should work as long as CHAR_BIT // is a multiple of 8 and there are types with the right sizes. // // The sized data type lookup table has the following format: // // bits // +----+----+----+----+ // | 8 | 16 | 32 | 64 | // +----+----+----+----+ // signed integer | | | | | // +----+----+----+----+ // unsigned integer | | | | | // +----+----+----+----+ // floating point | | | | | // +----+----+----+----+ // // So, the 0,3 element is supposed to contain the oct_data_conv enum // value corresponding to the correct native data type for a signed // 32-bit integer. static void init_sized_type_lookup_table (oct_data_conv::data_type table[3][4]) { int bits = 8; for (int i = 0; i < 4; i++) { FIND_SIZED_INT_TYPE (table[0][i], bits, , ); FIND_SIZED_INT_TYPE (table[1][i], bits, unsigned, u); FIND_SIZED_FLOAT_TYPE (table[2][i], bits); bits *= 2; } } oct_data_conv::data_type oct_data_conv::string_to_data_type (const std::string& str) { data_type retval = dt_unknown; static bool initialized = false; static data_type sized_type_table[3][4]; if (! initialized) { init_sized_type_lookup_table (sized_type_table); initialized = true; } // XXX FIXME XXX -- finish implementing this. int n = str.length (); int k = 0; std::string s (n, ' '); for (int i = 0; i < n; i++) if (! isspace (str[i])) s[k++] = tolower (str[i]); s.resize (k); if (s == "char") retval = dt_char; else if (s == "schar" || s == "signedchar") retval = dt_schar; else if (s == "uchar" || s == "unsignedchar") retval = dt_uchar; else if (s == "short") retval = dt_short; else if (s == "ushort" || s == "unsignedshort") retval = dt_ushort; else if (s == "int") retval = dt_int; else if (s == "uint" || s == "unsignedint") retval = dt_uint; else if (s == "long") retval = dt_long; else if (s == "ulong" || s == "unsignedlong") retval = dt_ulong; else if (s == "float") retval = dt_float; else if (s == "double") retval = dt_double; else if (s == "int8" || s == "char*1" || s == "integer*1") retval = sized_type_table[0][0]; else if (s == "int16" || s == "integer*2") retval = sized_type_table[0][1]; else if (s == "int32" || s == "integer*4") retval = sized_type_table[0][2]; else if (s == "int64" || s == "integer*8") retval = sized_type_table[0][3]; else if (s == "uint8") retval = sized_type_table[1][0]; else if (s == "uint16") retval = sized_type_table[1][1]; else if (s == "uint32") retval = sized_type_table[1][2]; else if (s == "uint64") retval = sized_type_table[1][3]; else if (s == "float32" || s == "real*4") retval = sized_type_table[2][2]; else if (s == "float64" || s == "real*8") retval = sized_type_table[2][3]; else (*current_liboctave_error_handler) ("invalid data type specified"); if (retval == dt_unknown) (*current_liboctave_error_handler) ("unable to find matching native data type for %s", s.c_str ()); return retval; } #define swap_1_bytes(x, y) #define LS_DO_READ(TYPE, swap, data, size, len, stream) \ do \ { \ if (len > 0) \ { \ volatile TYPE *ptr = X_CAST (volatile TYPE *, data); \ stream.read (X_CAST (char *, ptr), size * len); \ if (swap) \ swap_ ## size ## _bytes (ptr, len); \ TYPE tmp = ptr[0]; \ for (int i = len - 1; i > 0; i--) \ data[i] = ptr[i]; \ data[0] = tmp; \ } \ } \ while (0) // Have to use copy here to avoid writing over data accessed via // Matrix::data(). #define LS_DO_WRITE(TYPE, data, size, len, stream) \ do \ { \ if (len > 0) \ { \ char tmp_type = static_cast<char> (type); \ stream.write (&tmp_type, 1); \ TYPE *ptr = new TYPE [len]; \ for (int i = 0; i < len; i++) \ ptr[i] = X_CAST (TYPE, data[i]); \ stream.write (X_CAST (char *, ptr), size * len); \ delete [] ptr ; \ } \ } \ while (0) // Loading variables from files. static void gripe_unrecognized_float_fmt (void) { (*current_liboctave_error_handler) ("unrecognized floating point format requested"); } static void gripe_data_conversion (const char *from, const char *to) { (*current_liboctave_error_handler) ("unable to convert from %s to %s format", from, to); } // But first, some data conversion routines. // Currently, we only handle conversions for the IEEE types. To fix // that, make more of the following routines work. // XXX FIXME XXX -- assumes sizeof (Complex) == 8 // XXX FIXME XXX -- assumes sizeof (double) == 8 // XXX FIXME XXX -- assumes sizeof (float) == 4 static void IEEE_big_double_to_IEEE_little_double (double *d, int len) { swap_8_bytes (d, len); } static void VAX_D_double_to_IEEE_little_double (double * /* d */, int /* len */) { gripe_data_conversion ("VAX D float", "IEEE little endian format"); } static void VAX_G_double_to_IEEE_little_double (double * /* d */, int /* len */) { gripe_data_conversion ("VAX G float", "IEEE little endian format"); } static void Cray_to_IEEE_little_double (double * /* d */, int /* len */) { gripe_data_conversion ("Cray", "IEEE little endian format"); } static void IEEE_big_float_to_IEEE_little_float (float *d, int len) { swap_4_bytes (d, len); } static void VAX_D_float_to_IEEE_little_float (float * /* d */, int /* len */) { gripe_data_conversion ("VAX D float", "IEEE little endian format"); } static void VAX_G_float_to_IEEE_little_float (float * /* d */, int /* len */) { gripe_data_conversion ("VAX G float", "IEEE little endian format"); } static void Cray_to_IEEE_little_float (float * /* d */, int /* len */) { gripe_data_conversion ("Cray", "IEEE little endian format"); } static void IEEE_little_double_to_IEEE_big_double (double *d, int len) { swap_8_bytes (d, len); } static void VAX_D_double_to_IEEE_big_double (double * /* d */, int /* len */) { gripe_data_conversion ("VAX D float", "IEEE big endian format"); } static void VAX_G_double_to_IEEE_big_double (double * /* d */, int /* len */) { gripe_data_conversion ("VAX G float", "IEEE big endian format"); } static void Cray_to_IEEE_big_double (double * /* d */, int /* len */) { gripe_data_conversion ("Cray", "IEEE big endian format"); } static void IEEE_little_float_to_IEEE_big_float (float *d, int len) { swap_4_bytes (d, len); } static void VAX_D_float_to_IEEE_big_float (float * /* d */, int /* len */) { gripe_data_conversion ("VAX D float", "IEEE big endian format"); } static void VAX_G_float_to_IEEE_big_float (float * /* d */, int /* len */) { gripe_data_conversion ("VAX G float", "IEEE big endian format"); } static void Cray_to_IEEE_big_float (float * /* d */, int /* len */) { gripe_data_conversion ("Cray", "IEEE big endian format"); } static void IEEE_little_double_to_VAX_D_double (double * /* d */, int /* len */) { gripe_data_conversion ("IEEE little endian", "VAX D"); } static void IEEE_big_double_to_VAX_D_double (double * /* d */, int /* len */) { gripe_data_conversion ("IEEE big endian", "VAX D"); } static void VAX_G_double_to_VAX_D_double (double * /* d */, int /* len */) { gripe_data_conversion ("VAX G float", "VAX D"); } static void Cray_to_VAX_D_double (double * /* d */, int /* len */) { gripe_data_conversion ("Cray", "VAX D"); } static void IEEE_little_float_to_VAX_D_float (float * /* d */, int /* len */) { gripe_data_conversion ("IEEE little endian", "VAX D"); } static void IEEE_big_float_to_VAX_D_float (float * /* d */, int /* len */) { gripe_data_conversion ("IEEE big endian", "VAX D"); } static void VAX_G_float_to_VAX_D_float (float * /* d */, int /* len */) { gripe_data_conversion ("VAX G float", "VAX D"); } static void Cray_to_VAX_D_float (float * /* d */, int /* len */) { gripe_data_conversion ("Cray", "VAX D"); } static void IEEE_little_double_to_VAX_G_double (double * /* d */, int /* len */) { gripe_data_conversion ("IEEE little endian", "VAX G"); } static void IEEE_big_double_to_VAX_G_double (double * /* d */, int /* len */) { gripe_data_conversion ("IEEE big endian", "VAX G"); } static void VAX_D_double_to_VAX_G_double (double * /* d */, int /* len */) { gripe_data_conversion ("VAX D float", "VAX G"); } static void Cray_to_VAX_G_double (double * /* d */, int /* len */) { gripe_data_conversion ("VAX G float", "VAX G"); } static void IEEE_little_float_to_VAX_G_float (float * /* d */, int /* len */) { gripe_data_conversion ("IEEE little endian", "VAX G"); } static void IEEE_big_float_to_VAX_G_float (float * /* d */, int /* len */) { gripe_data_conversion ("IEEE big endian", "VAX G"); } static void VAX_D_float_to_VAX_G_float (float * /* d */, int /* len */) { gripe_data_conversion ("VAX D float", "VAX G"); } static void Cray_to_VAX_G_float (float * /* d */, int /* len */) { gripe_data_conversion ("VAX G float", "VAX G"); } void do_double_format_conversion (double *data, int len, oct_mach_info::float_format fmt) { switch (oct_mach_info::native_float_format ()) { case oct_mach_info::flt_fmt_ieee_little_endian: switch (fmt) { case oct_mach_info::flt_fmt_ieee_little_endian: break; case oct_mach_info::flt_fmt_ieee_big_endian: IEEE_big_double_to_IEEE_little_double (data, len); break; case oct_mach_info::flt_fmt_vax_d: VAX_D_double_to_IEEE_little_double (data, len); break; case oct_mach_info::flt_fmt_vax_g: VAX_G_double_to_IEEE_little_double (data, len); break; case oct_mach_info::flt_fmt_cray: Cray_to_IEEE_little_double (data, len); break; default: gripe_unrecognized_float_fmt (); break; } break; case oct_mach_info::flt_fmt_ieee_big_endian: switch (fmt) { case oct_mach_info::flt_fmt_ieee_little_endian: IEEE_little_double_to_IEEE_big_double (data, len); break; case oct_mach_info::flt_fmt_ieee_big_endian: break; case oct_mach_info::flt_fmt_vax_d: VAX_D_double_to_IEEE_big_double (data, len); break; case oct_mach_info::flt_fmt_vax_g: VAX_G_double_to_IEEE_big_double (data, len); break; case oct_mach_info::flt_fmt_cray: Cray_to_IEEE_big_double (data, len); break; default: gripe_unrecognized_float_fmt (); break; } break; case oct_mach_info::flt_fmt_vax_d: switch (fmt) { case oct_mach_info::flt_fmt_ieee_little_endian: IEEE_little_double_to_VAX_D_double (data, len); break; case oct_mach_info::flt_fmt_ieee_big_endian: IEEE_big_double_to_VAX_D_double (data, len); break; case oct_mach_info::flt_fmt_vax_d: break; case oct_mach_info::flt_fmt_vax_g: VAX_G_double_to_VAX_D_double (data, len); break; case oct_mach_info::flt_fmt_cray: Cray_to_VAX_D_double (data, len); break; default: gripe_unrecognized_float_fmt (); break; } break; case oct_mach_info::flt_fmt_vax_g: switch (fmt) { case oct_mach_info::flt_fmt_ieee_little_endian: IEEE_little_double_to_VAX_G_double (data, len); break; case oct_mach_info::flt_fmt_ieee_big_endian: IEEE_big_double_to_VAX_G_double (data, len); break; case oct_mach_info::flt_fmt_vax_d: VAX_D_double_to_VAX_G_double (data, len); break; case oct_mach_info::flt_fmt_vax_g: break; case oct_mach_info::flt_fmt_cray: Cray_to_VAX_G_double (data, len); break; default: gripe_unrecognized_float_fmt (); break; } break; default: (*current_liboctave_error_handler) ("impossible state reached in file `%s' at line %d", __FILE__, __LINE__); break; } } void do_float_format_conversion (float *data, int len, oct_mach_info::float_format fmt) { switch (oct_mach_info::native_float_format ()) { case oct_mach_info::flt_fmt_ieee_little_endian: switch (fmt) { case oct_mach_info::flt_fmt_ieee_little_endian: break; case oct_mach_info::flt_fmt_ieee_big_endian: IEEE_big_float_to_IEEE_little_float (data, len); break; case oct_mach_info::flt_fmt_vax_d: VAX_D_float_to_IEEE_little_float (data, len); break; case oct_mach_info::flt_fmt_vax_g: VAX_G_float_to_IEEE_little_float (data, len); break; case oct_mach_info::flt_fmt_cray: Cray_to_IEEE_little_float (data, len); break; default: gripe_unrecognized_float_fmt (); break; } break; case oct_mach_info::flt_fmt_ieee_big_endian: switch (fmt) { case oct_mach_info::flt_fmt_ieee_little_endian: IEEE_little_float_to_IEEE_big_float (data, len); break; case oct_mach_info::flt_fmt_ieee_big_endian: break; case oct_mach_info::flt_fmt_vax_d: VAX_D_float_to_IEEE_big_float (data, len); break; case oct_mach_info::flt_fmt_vax_g: VAX_G_float_to_IEEE_big_float (data, len); break; case oct_mach_info::flt_fmt_cray: Cray_to_IEEE_big_float (data, len); break; default: gripe_unrecognized_float_fmt (); break; } break; case oct_mach_info::flt_fmt_vax_d: switch (fmt) { case oct_mach_info::flt_fmt_ieee_little_endian: IEEE_little_float_to_VAX_D_float (data, len); break; case oct_mach_info::flt_fmt_ieee_big_endian: IEEE_big_float_to_VAX_D_float (data, len); break; case oct_mach_info::flt_fmt_vax_d: break; case oct_mach_info::flt_fmt_vax_g: VAX_G_float_to_VAX_D_float (data, len); break; case oct_mach_info::flt_fmt_cray: Cray_to_VAX_D_float (data, len); break; default: gripe_unrecognized_float_fmt (); break; } break; case oct_mach_info::flt_fmt_vax_g: switch (fmt) { case oct_mach_info::flt_fmt_ieee_little_endian: IEEE_little_float_to_VAX_G_float (data, len); break; case oct_mach_info::flt_fmt_ieee_big_endian: IEEE_big_float_to_VAX_G_float (data, len); break; case oct_mach_info::flt_fmt_vax_d: VAX_D_float_to_VAX_G_float (data, len); break; case oct_mach_info::flt_fmt_vax_g: break; case oct_mach_info::flt_fmt_cray: Cray_to_VAX_G_float (data, len); break; default: gripe_unrecognized_float_fmt (); break; } break; default: (*current_liboctave_error_handler) ("impossible state reached in file `%s' at line %d", __FILE__, __LINE__); break; } } void read_doubles (std::istream& is, double *data, save_type type, int len, int swap, oct_mach_info::float_format fmt) { switch (type) { case LS_U_CHAR: LS_DO_READ (unsigned char, swap, data, 1, len, is); break; case LS_U_SHORT: LS_DO_READ (unsigned TWO_BYTE_INT, swap, data, 2, len, is); break; case LS_U_INT: LS_DO_READ (unsigned FOUR_BYTE_INT, swap, data, 4, len, is); break; case LS_CHAR: LS_DO_READ (signed char, swap, data, 1, len, is); break; case LS_SHORT: LS_DO_READ (TWO_BYTE_INT, swap, data, 2, len, is); break; case LS_INT: LS_DO_READ (FOUR_BYTE_INT, swap, data, 4, len, is); break; case LS_FLOAT: { volatile float *ptr = X_CAST (float *, data); is.read (X_CAST (char *, data), 4 * len); do_float_format_conversion (X_CAST (float *, data), len, fmt); float tmp = ptr[0]; for (int i = len - 1; i > 0; i--) data[i] = ptr[i]; data[0] = tmp; } break; case LS_DOUBLE: // No conversion necessary. is.read (X_CAST (char *, data), 8 * len); do_double_format_conversion (data, len, fmt); break; default: is.clear (std::ios::failbit|is.rdstate ()); break; } } void write_doubles (std::ostream& os, const double *data, save_type type, int len) { switch (type) { case LS_U_CHAR: LS_DO_WRITE (unsigned char, data, 1, len, os); break; case LS_U_SHORT: LS_DO_WRITE (unsigned TWO_BYTE_INT, data, 2, len, os); break; case LS_U_INT: LS_DO_WRITE (unsigned FOUR_BYTE_INT, data, 4, len, os); break; case LS_CHAR: LS_DO_WRITE (signed char, data, 1, len, os); break; case LS_SHORT: LS_DO_WRITE (TWO_BYTE_INT, data, 2, len, os); break; case LS_INT: LS_DO_WRITE (FOUR_BYTE_INT, data, 4, len, os); break; case LS_FLOAT: LS_DO_WRITE (float, data, 4, len, os); break; case LS_DOUBLE: // No conversion necessary. { char tmp_type = X_CAST (char, type); os.write (&tmp_type, 1); os.write (X_CAST (char *, data), 8 * len); } break; default: (*current_liboctave_error_handler) ("unrecognized data format requested"); break; } } /* ;;; Local Variables: *** ;;; mode: C++ *** ;;; End: *** */