src/ecs.c

   1 /*
   2    Basic support for extended character sets.
   3
   4    Written by:
   5    Roland Illig <roland.illig@gmx.de>, 2005.
   6
   7    This file is part of the Midnight Commander.
   8
   9    The Midnight Commander is free software; you can redistribute it
  10    and/or modify it under the terms of the GNU General Public License as
  11    published by the Free Software Foundation; either version 2 of the
  12    License, or (at your option) any later version.
  13
  14    The Midnight Commander is distributed in the hope that it will be
  15    useful, but WITHOUT ANY WARRANTY; without even the implied warranty
  16    of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  17    General Public License for more details.
  18
  19    You should have received a copy of the GNU General Public License
  20    along with this program; if not, write to the Free Software
  21    Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston,
  22    MA 02110-1301, USA.
  23  */
  24
  25 /** \file  ecs.c
  26  *  \brief Source: basic support for extended character sets
  27  */
  28
  29 #include <config.h>
  30
  31 #include <assert.h>
  32 #include <ctype.h>
  33
  34 #include "lib/global.h"
  35 #include "ecs.h"
  36
  37 /*
  38  * String type conversion
  39  */
  40
  41 extern gboolean ecs_mbstr_to_str(ecs_char **ret_str, const char *s)
  42 {
  43 #ifdef EXTCHARSET_ENABLED
  44         size_t maxlen, len;
  45         ecs_char *str;
  46
  47         maxlen = strlen(s);
  48
  49         str = g_new(ecs_char, maxlen + 1);
  50         len = mbstowcs(str, s, maxlen + 1);
  51         if (len == (size_t) -1) {
  52                 g_free(str);
  53                 return FALSE;
  54         }
  55
  56         assert(len <= maxlen);
  57         *ret_str = g_renew(ecs_char, str, len + 1);
  58         return TRUE;
  59 #else
  60         *ret_str = g_strdup(s);
  61         return TRUE;
  62 #endif
  63 }
  64
  65 extern gboolean ecs_str_to_mbstr(char **ret_str, const ecs_char *s)
  66 {
  67 #ifdef EXTCHARSET_ENABLED
  68         size_t maxlen, len;
  69         char *str;
  70
  71         maxlen = ecs_strlen(s) * MB_CUR_MAX;
  72
  73         str = g_new(char, maxlen + 1);
  74         len = wcstombs(str, s, maxlen + 1);
  75         if (len == (size_t) -1) {
  76                 g_free(str);
  77                 return FALSE;
  78         }
  79
  80         assert(len <= maxlen);
  81         *ret_str = g_renew(char, str, len + 1);
  82         return TRUE;
  83 #else
  84         *ret_str = g_strdup(s);
  85         return TRUE;
  86 #endif
  87 }
  88
  89 /*
  90  * Character classification
  91  */
  92
  93 #ifdef EXTCHARSET_ENABLED
  94 # ifdef HAVE_WCTYPE_H
  95 #  include <wctype.h>
  96 #  define ECS_CTYPE(wf, cf, c) \
  97         (wf(c))
  98 # else
  99 #  define ECS_CTYPE(wf, cf, c) \
 100         (((unsigned char) c != c) ? FALSE : (cf(c)))
 101 # endif
 102 #else
 103 # define ECS_CTYPE(wf, cf, c) \
 104         (cf(c))
 105 #endif
 106
 107 extern gboolean ecs_isalnum(ecs_char c)
 108 {
 109         return ECS_CTYPE(iswalnum, isalnum, c);
 110 }
 111
 112 extern gboolean ecs_isalpha(ecs_char c)
 113 {
 114         return ECS_CTYPE(iswalpha, isalpha, c);
 115 }
 116
 117 extern gboolean ecs_iscntrl(ecs_char c)
 118 {
 119         return ECS_CTYPE(iswcntrl, iscntrl, c);
 120 }
 121
 122 extern gboolean ecs_isdigit(ecs_char c)
 123 {
 124         return ECS_CTYPE(iswdigit, isdigit, c);
 125 }
 126
 127 extern gboolean ecs_isgraph(ecs_char c)
 128 {
 129         return ECS_CTYPE(iswgraph, isgraph, c);
 130 }
 131
 132 extern gboolean ecs_islower(ecs_char c)
 133 {
 134         return ECS_CTYPE(iswlower, islower, c);
 135 }
 136
 137 extern gboolean ecs_isprint(ecs_char c)
 138 {
 139         return ECS_CTYPE(iswprint, isprint, c);
 140 }
 141
 142 extern gboolean ecs_ispunct(ecs_char c)
 143 {
 144         return ECS_CTYPE(iswpunct, ispunct, c);
 145 }
 146
 147 extern gboolean ecs_isspace(ecs_char c)
 148 {
 149         return ECS_CTYPE(iswspace, isspace, c);
 150 }
 151
 152 extern gboolean ecs_isupper(ecs_char c)
 153 {
 154         return ECS_CTYPE(iswupper, isupper, c);
 155 }
 156
 157 extern gboolean ecs_isxdigit(ecs_char c)
 158 {
 159         return ECS_CTYPE(iswxdigit, isxdigit, c);
 160 }
 161
 162 #undef ECS_CTYPE
 163
 164 /*
 165  * ISO C90 <string.h> functions
 166  */
 167
 168 /* left out: ecs_strcpy */
 169 /* left out: ecs_strncpy */
 170 /* left out: ecs_strcat */
 171 /* left out: ecs_strncat */
 172
 173 int
 174 ecs_strcmp(const ecs_char *a, const ecs_char *b)
 175 {
 176         size_t i;
 177         unsigned long ca, cb;
 178
 179         for (i = 0; a[i] == b[i]; i++) {
 180                 if (a[i] == ECS_CHAR('\0'))
 181                         return 0;
 182         }
 183         ca = (unsigned long) a[i];
 184         cb = (unsigned long) b[i];
 185         return (ca < cb) ? -1 : (ca > cb) ? 1 : 0;
 186 }
 187
 188 /* left out: ecs_strcoll */
 189 /* left out: ecs_strncmp */
 190 /* left out: ecs_strxfrm */
 191
 192 ecs_char *
 193 ecs_strchr(const ecs_char *s, ecs_char c)
 194 {
 195         size_t i;
 196
 197         for (i = 0; s[i] != c; i++) {
 198                 if (s[i] == ECS_CHAR('\0'))
 199                         return NULL;
 200         }
 201         return (ecs_char *) s + i;
 202 }
 203
 204 size_t
 205 ecs_strcspn(const ecs_char *haystack, const ecs_char *needles)
 206 {
 207         size_t i, j;
 208
 209         for (i = 0; haystack[i] != ECS_CHAR('\0'); i++) {
 210                 for (j = 0; needles[j] != ECS_CHAR('\0'); j++) {
 211                         if (haystack[i] == needles[j])
 212                                 return i;
 213                 }
 214         }
 215         return i;
 216 }
 217
 218 /* left out: ecs_strpbrk */
 219
 220 ecs_char *
 221 ecs_strrchr(const ecs_char *s, ecs_char c)
 222 {
 223         ecs_char *pos;
 224         size_t i;
 225
 226         for (i = 0, pos = NULL;; i++) {
 227                 if (s[i] == c)
 228                         pos = (ecs_char *) s + i;
 229                 if (s[i] == ECS_CHAR('\0'))
 230                         return pos;
 231         }
 232 }
 233
 234 size_t
 235 ecs_strspn(const ecs_char *s, const ecs_char *chars)
 236 {
 237         size_t i;
 238
 239         for (i = 0; s[i] != ECS_CHAR('\0'); i++) {
 240                 if (ecs_strchr(chars, s[i]) == NULL)
 241                         break;
 242         }
 243         return i;
 244 }
 245
 246 ecs_char *
 247 ecs_strstr(const ecs_char *s, const ecs_char *sub)
 248 {
 249         size_t i, j;
 250
 251         for (i = 0; s[i] != ECS_CHAR('\0'); i++) {
 252                 for (j = 0; sub[j] != ECS_CHAR('\0'); j++) {
 253                         if (s[i + j] != sub[j])
 254                                 goto next_i;
 255                 }
 256                 return (ecs_char *) s + i;
 257         next_i:
 258                 continue;
 259         }
 260         return NULL;
 261 }
 262
 263 /* left out: ecs_strtok */
 264
 265 size_t
 266 ecs_strlen(const ecs_char *s)
 267 {
 268         size_t i;
 269
 270         for (i = 0; s[i] != ECS_CHAR('\0'); i++)
 271                 continue;
 272         return i;
 273 }
 274
 275 /*
 276  * Other functions
 277  */
 278
 279 ecs_char *ecs_xstrdup(const ecs_char *s)
 280 {
 281         ecs_char *retval;
 282         size_t len;
 283
 284         len = ecs_strlen(s);
 285         retval = g_new(ecs_char, len + 1);
 286         memcpy(retval, s, (len + 1) * sizeof(ecs_char));
 287         return retval;
 288 }
 289
 290 size_t
 291 ecs_strlcpy(ecs_char *dst, const ecs_char *src, size_t dstsize)
 292 {
 293         size_t n = 0;           /* number of copied characters */
 294
 295         if (dstsize >= 1) {
 296                 while (n < dstsize - 1 && *src != ECS_CHAR('\0')) {
 297                         *dst++ = *src++;
 298                         n++;
 299                 }
 300                 *dst = ECS_CHAR('\0');
 301         }
 302
 303         while (*src != ECS_CHAR('\0')) {
 304                 n++;
 305                 src++;
 306         }
 307
 308         return n;
 309 }
 310
 311 size_t
 312 ecs_strlcat(ecs_char *dst, const ecs_char *src, size_t dstsize)
 313 {
 314         size_t di = 0;
 315
 316         while (di < dstsize && dst[di] != ECS_CHAR('\0'))
 317                 di++;
 318         return di + ecs_strlcpy(dst + di, src, dstsize - di);
 319 }
 320
 321 gboolean
 322 ecs_strbox(const ecs_char *s, size_t *ret_width, size_t *ret_height)
 323 {
 324         size_t nlines = 0, ncolumns = 0, colindex = 0, i;
 325
 326         for (i = 0; s[i] != ECS_CHAR('\0'); i++) {
 327                 if (s[i] == ECS_CHAR('\n')) {
 328                         nlines++;
 329                         colindex = 0;
 330                 } else {
 331                         if (!ecs_isprint(s[i]))
 332                                 return FALSE;
 333
 334                         /* FIXME: This code assumes that each printable
 335                          * character occupies one cell on the screen. */
 336                         colindex++;
 337                         if (colindex > ncolumns)
 338                                 ncolumns = colindex;
 339                 }
 340         }
 341         *ret_width  = ncolumns;
 342         *ret_height = nlines;
 343         return TRUE;
 344 }