Merge branch '4524_cleanup'
[midnight-commander.git] / lib / strutil / strutil8bit.c
blob468e6ea9912adf7113376bbcf5a6550aa6948a6f
1 /*
2 8bit strings utilities
4 Copyright (C) 2007-2024
5 Free Software Foundation, Inc.
7 Written by:
8 Rostislav Benes, 2007
10 This file is part of the Midnight Commander.
12 The Midnight Commander is free software: you can redistribute it
13 and/or modify it under the terms of the GNU General Public License as
14 published by the Free Software Foundation, either version 3 of the License,
15 or (at your option) any later version.
17 The Midnight Commander is distributed in the hope that it will be useful,
18 but WITHOUT ANY WARRANTY; without even the implied warranty of
19 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
20 GNU General Public License for more details.
22 You should have received a copy of the GNU General Public License
23 along with this program. If not, see <http://www.gnu.org/licenses/>.
26 #include <config.h>
28 #include <ctype.h>
29 #include <stdlib.h>
31 #include "lib/global.h"
32 #include "lib/strutil.h"
34 /* Functions for singlebyte encodings, all characters have width 1
35 * using standard system functions.
36 * There are only small differences between functions in strutil8bit.c
37 * and strutilascii.c.
40 /*** global variables ****************************************************************************/
42 /*** file scope macro definitions ****************************************************************/
45 * Inlines to equalize 'char' signedness for single 'char' encodings.
46 * Instead of writing
47 * isspace ((unsigned char) c);
48 * you can write
49 * char_isspace (c);
51 #define DECLARE_CTYPE_WRAPPER(func_name) \
52 static inline int char_##func_name(char c) \
53 { \
54 return func_name((int)(unsigned char)c); \
57 /*** file scope type declarations ****************************************************************/
59 /*** forward declarations (file scope functions) *************************************************/
61 /*** file scope variables ************************************************************************/
63 static const char replch = '?';
65 /* --------------------------------------------------------------------------------------------- */
66 /*** file scope functions ************************************************************************/
67 /* --------------------------------------------------------------------------------------------- */
69 /* *INDENT-OFF* */
70 DECLARE_CTYPE_WRAPPER (isalnum)
71 DECLARE_CTYPE_WRAPPER (isdigit)
72 DECLARE_CTYPE_WRAPPER (isprint)
73 DECLARE_CTYPE_WRAPPER (ispunct)
74 DECLARE_CTYPE_WRAPPER (isspace)
75 DECLARE_CTYPE_WRAPPER (toupper)
76 DECLARE_CTYPE_WRAPPER (tolower)
77 /* *INDENT-ON* */
79 /* --------------------------------------------------------------------------------------------- */
81 static void
82 str_8bit_insert_replace_char (GString * buffer)
84 g_string_append_c (buffer, replch);
87 /* --------------------------------------------------------------------------------------------- */
89 static gboolean
90 str_8bit_is_valid_string (const char *text)
92 (void) text;
93 return TRUE;
96 /* --------------------------------------------------------------------------------------------- */
98 static int
99 str_8bit_is_valid_char (const char *ch, size_t size)
101 (void) ch;
102 (void) size;
103 return 1;
106 /* --------------------------------------------------------------------------------------------- */
108 static void
109 str_8bit_cnext_char (const char **text)
111 (*text)++;
114 /* --------------------------------------------------------------------------------------------- */
116 static void
117 str_8bit_cprev_char (const char **text)
119 (*text)--;
122 /* --------------------------------------------------------------------------------------------- */
124 static int
125 str_8bit_cnext_noncomb_char (const char **text)
127 if (*text[0] == '\0')
128 return 0;
130 (*text)++;
131 return 1;
134 /* --------------------------------------------------------------------------------------------- */
136 static int
137 str_8bit_cprev_noncomb_char (const char **text, const char *begin)
139 if ((*text) == begin)
140 return 0;
142 (*text)--;
143 return 1;
146 /* --------------------------------------------------------------------------------------------- */
148 static gboolean
149 str_8bit_isspace (const char *text)
151 return char_isspace (text[0]) != 0;
154 /* --------------------------------------------------------------------------------------------- */
156 static gboolean
157 str_8bit_ispunct (const char *text)
159 return char_ispunct (text[0]) != 0;
162 /* --------------------------------------------------------------------------------------------- */
164 static gboolean
165 str_8bit_isalnum (const char *text)
167 return char_isalnum (text[0]) != 0;
170 /* --------------------------------------------------------------------------------------------- */
172 static gboolean
173 str_8bit_isdigit (const char *text)
175 return char_isdigit (text[0]) != 0;
178 /* --------------------------------------------------------------------------------------------- */
180 static gboolean
181 str_8bit_isprint (const char *text)
183 return char_isprint (text[0]) != 0;
186 /* --------------------------------------------------------------------------------------------- */
188 static gboolean
189 str_8bit_iscombiningmark (const char *text)
191 (void) text;
192 return FALSE;
195 /* --------------------------------------------------------------------------------------------- */
197 static int
198 str_8bit_toupper (const char *text, char **out, size_t * remain)
200 if (*remain <= 1)
201 return FALSE;
203 (*out)[0] = char_toupper (text[0]);
204 (*out)++;
205 (*remain)--;
206 return TRUE;
209 /* --------------------------------------------------------------------------------------------- */
211 static gboolean
212 str_8bit_tolower (const char *text, char **out, size_t * remain)
214 if (*remain <= 1)
215 return FALSE;
217 (*out)[0] = char_tolower (text[0]);
218 (*out)++;
219 (*remain)--;
220 return TRUE;
223 /* --------------------------------------------------------------------------------------------- */
225 static int
226 str_8bit_length (const char *text)
228 return strlen (text);
231 /* --------------------------------------------------------------------------------------------- */
233 static int
234 str_8bit_length2 (const char *text, int size)
236 return (size >= 0) ? MIN (strlen (text), (gsize) size) : strlen (text);
239 /* --------------------------------------------------------------------------------------------- */
241 static gchar *
242 str_8bit_conv_gerror_message (GError * mcerror, const char *def_msg)
244 GIConv conv;
245 gchar *ret;
247 /* glib messages are in UTF-8 charset */
248 conv = str_crt_conv_from ("UTF-8");
250 if (conv == INVALID_CONV)
251 ret = g_strdup (def_msg != NULL ? def_msg : "");
252 else
254 GString *buf;
256 buf = g_string_new ("");
258 if (str_convert (conv, mcerror->message, buf) != ESTR_FAILURE)
259 ret = g_string_free (buf, FALSE);
260 else
262 ret = g_strdup (def_msg != NULL ? def_msg : "");
263 g_string_free (buf, TRUE);
266 str_close_conv (conv);
269 return ret;
272 /* --------------------------------------------------------------------------------------------- */
274 static estr_t
275 str_8bit_vfs_convert_to (GIConv coder, const char *string, int size, GString * buffer)
277 estr_t result = ESTR_SUCCESS;
279 if (coder == str_cnv_not_convert)
280 g_string_append_len (buffer, string, size);
281 else
282 result = str_nconvert (coder, string, size, buffer);
284 return result;
287 /* --------------------------------------------------------------------------------------------- */
289 static const char *
290 str_8bit_term_form (const char *text)
292 static char result[BUF_MEDIUM];
293 char *actual;
294 size_t remain;
295 size_t length;
296 size_t pos = 0;
298 actual = result;
299 remain = sizeof (result);
300 length = strlen (text);
302 for (; pos < length && remain > 1; pos++, actual++, remain--)
303 actual[0] = char_isprint (text[pos]) ? text[pos] : '.';
305 actual[0] = '\0';
306 return result;
309 /* --------------------------------------------------------------------------------------------- */
311 static const char *
312 str_8bit_fit_to_term (const char *text, int width, align_crt_t just_mode)
314 static char result[BUF_MEDIUM];
315 char *actual;
316 size_t remain;
317 int ident = 0;
318 size_t length;
319 size_t pos = 0;
321 length = strlen (text);
322 actual = result;
323 remain = sizeof (result);
325 if ((int) length <= width)
327 switch (HIDE_FIT (just_mode))
329 case J_CENTER_LEFT:
330 case J_CENTER:
331 ident = (width - length) / 2;
332 break;
333 case J_RIGHT:
334 ident = width - length;
335 break;
336 default:
337 break;
340 if ((int) remain <= ident)
341 goto finally;
342 memset (actual, ' ', ident);
343 actual += ident;
344 remain -= ident;
346 for (; pos < length && remain > 1; pos++, actual++, remain--)
347 actual[0] = char_isprint (text[pos]) ? text[pos] : '.';
349 if (width - length - ident > 0)
351 if (remain <= width - length - ident)
352 goto finally;
353 memset (actual, ' ', width - length - ident);
354 actual += width - length - ident;
357 else if (IS_FIT (just_mode))
359 for (; pos + 1 <= (gsize) width / 2 && remain > 1; actual++, pos++, remain--)
360 actual[0] = char_isprint (text[pos]) ? text[pos] : '.';
362 if (remain <= 1)
363 goto finally;
364 actual[0] = '~';
365 actual++;
366 remain--;
368 pos += length - width + 1;
369 for (; pos < length && remain > 1; pos++, actual++, remain--)
370 actual[0] = char_isprint (text[pos]) ? text[pos] : '.';
372 else
374 switch (HIDE_FIT (just_mode))
376 case J_CENTER:
377 ident = (length - width) / 2;
378 break;
379 case J_RIGHT:
380 ident = length - width;
381 break;
382 default:
383 break;
386 pos += ident;
387 for (; pos < (gsize) (ident + width) && remain > 1; pos++, actual++, remain--)
388 actual[0] = char_isprint (text[pos]) ? text[pos] : '.';
391 finally:
392 if (actual >= result + sizeof (result))
393 actual = result + sizeof (result) - 1;
394 actual[0] = '\0';
395 return result;
398 /* --------------------------------------------------------------------------------------------- */
400 static const char *
401 str_8bit_term_trim (const char *text, int width)
403 static char result[BUF_MEDIUM];
404 size_t remain;
405 char *actual;
406 size_t length;
408 length = strlen (text);
409 actual = result;
410 remain = sizeof (result);
412 if (width > 0)
414 size_t pos;
416 if (width >= (int) length)
418 for (pos = 0; pos < length && remain > 1; pos++, actual++, remain--)
419 actual[0] = char_isprint (text[pos]) ? text[pos] : '.';
421 else if (width <= 3)
423 memset (actual, '.', width);
424 actual += width;
426 else
428 memset (actual, '.', 3);
429 actual += 3;
430 remain -= 3;
432 for (pos = length - width + 3; pos < length && remain > 1; pos++, actual++, remain--)
433 actual[0] = char_isprint (text[pos]) ? text[pos] : '.';
437 actual[0] = '\0';
438 return result;
441 /* --------------------------------------------------------------------------------------------- */
443 static int
444 str_8bit_term_width2 (const char *text, size_t length)
446 return (length != (size_t) (-1)) ? MIN (strlen (text), length) : strlen (text);
449 /* --------------------------------------------------------------------------------------------- */
451 static int
452 str_8bit_term_width1 (const char *text)
454 return str_8bit_term_width2 (text, (size_t) (-1));
457 /* --------------------------------------------------------------------------------------------- */
459 static int
460 str_8bit_term_char_width (const char *text)
462 (void) text;
463 return 1;
466 /* --------------------------------------------------------------------------------------------- */
468 static const char *
469 str_8bit_term_substring (const char *text, int start, int width)
471 static char result[BUF_MEDIUM];
472 size_t remain;
473 char *actual;
474 size_t length;
476 actual = result;
477 remain = sizeof (result);
478 length = strlen (text);
480 if (start < (int) length)
482 size_t pos;
484 for (pos = start; pos < length && width > 0 && remain > 1;
485 pos++, width--, actual++, remain--)
486 actual[0] = char_isprint (text[pos]) ? text[pos] : '.';
489 for (; width > 0 && remain > 1; actual++, remain--, width--)
490 actual[0] = ' ';
492 actual[0] = '\0';
493 return result;
496 /* --------------------------------------------------------------------------------------------- */
498 static const char *
499 str_8bit_trunc (const char *text, int width)
501 static char result[MC_MAXPATHLEN];
502 int remain;
503 char *actual;
504 size_t pos = 0;
505 size_t length;
507 actual = result;
508 remain = sizeof (result);
509 length = strlen (text);
511 if ((int) length > width)
513 for (; pos + 1 <= (gsize) width / 2 && remain > 1; actual++, pos++, remain--)
514 actual[0] = char_isprint (text[pos]) ? text[pos] : '.';
516 if (remain <= 1)
517 goto finally;
518 actual[0] = '~';
519 actual++;
520 remain--;
522 pos += length - width + 1;
523 for (; pos < length && remain > 1; pos++, actual++, remain--)
524 actual[0] = char_isprint (text[pos]) ? text[pos] : '.';
526 else
528 for (; pos < length && remain > 1; pos++, actual++, remain--)
529 actual[0] = char_isprint (text[pos]) ? text[pos] : '.';
532 finally:
533 actual[0] = '\0';
534 return result;
537 /* --------------------------------------------------------------------------------------------- */
539 static int
540 str_8bit_offset_to_pos (const char *text, size_t length)
542 (void) text;
543 return (int) length;
546 /* --------------------------------------------------------------------------------------------- */
548 static int
549 str_8bit_column_to_pos (const char *text, size_t pos)
551 (void) text;
552 return (int) pos;
555 /* --------------------------------------------------------------------------------------------- */
557 static char *
558 str_8bit_create_search_needle (const char *needle, gboolean case_sen)
560 (void) case_sen;
561 return (char *) needle;
564 /* --------------------------------------------------------------------------------------------- */
566 static void
567 str_8bit_release_search_needle (char *needle, gboolean case_sen)
569 (void) case_sen;
570 (void) needle;
573 /* --------------------------------------------------------------------------------------------- */
575 static char *
576 str_8bit_strdown (const char *str)
578 char *rets, *p;
580 if (str == NULL)
581 return NULL;
583 rets = g_strdup (str);
585 for (p = rets; *p != '\0'; p++)
586 *p = char_tolower (*p);
588 return rets;
591 /* --------------------------------------------------------------------------------------------- */
593 static const char *
594 str_8bit_search_first (const char *text, const char *search, gboolean case_sen)
596 char *fold_text;
597 char *fold_search;
598 const char *match;
600 fold_text = case_sen ? (char *) text : str_8bit_strdown (text);
601 fold_search = case_sen ? (char *) search : str_8bit_strdown (search);
603 match = g_strstr_len (fold_text, -1, fold_search);
604 if (match != NULL)
606 size_t offset;
608 offset = match - fold_text;
609 match = text + offset;
612 if (!case_sen)
614 g_free (fold_text);
615 g_free (fold_search);
618 return match;
621 /* --------------------------------------------------------------------------------------------- */
623 static const char *
624 str_8bit_search_last (const char *text, const char *search, gboolean case_sen)
626 char *fold_text;
627 char *fold_search;
628 const char *match;
630 fold_text = case_sen ? (char *) text : str_8bit_strdown (text);
631 fold_search = case_sen ? (char *) search : str_8bit_strdown (search);
633 match = g_strrstr_len (fold_text, -1, fold_search);
634 if (match != NULL)
636 size_t offset;
638 offset = match - fold_text;
639 match = text + offset;
642 if (!case_sen)
644 g_free (fold_text);
645 g_free (fold_search);
648 return match;
651 /* --------------------------------------------------------------------------------------------- */
653 static int
654 str_8bit_compare (const char *t1, const char *t2)
656 return strcmp (t1, t2);
659 /* --------------------------------------------------------------------------------------------- */
661 static int
662 str_8bit_ncompare (const char *t1, const char *t2)
664 return strncmp (t1, t2, MIN (strlen (t1), strlen (t2)));
667 /* --------------------------------------------------------------------------------------------- */
669 static int
670 str_8bit_casecmp (const char *s1, const char *s2)
672 /* code from GLib */
674 #ifdef HAVE_STRCASECMP
675 g_return_val_if_fail (s1 != NULL, 0);
676 g_return_val_if_fail (s2 != NULL, 0);
678 return strcasecmp (s1, s2);
679 #else
681 g_return_val_if_fail (s1 != NULL, 0);
682 g_return_val_if_fail (s2 != NULL, 0);
684 for (; *s1 != '\0' && *s2 != '\0'; s1++, s2++)
686 gint c1, c2;
688 /* According to A. Cox, some platforms have islower's that
689 * don't work right on non-uppercase
691 c1 = isupper ((guchar) * s1) ? tolower ((guchar) * s1) : *s1;
692 c2 = isupper ((guchar) * s2) ? tolower ((guchar) * s2) : *s2;
693 if (c1 != c2)
694 return (c1 - c2);
697 return (((gint) (guchar) * s1) - ((gint) (guchar) * s2));
698 #endif
701 /* --------------------------------------------------------------------------------------------- */
703 static int
704 str_8bit_ncasecmp (const char *s1, const char *s2)
706 size_t n;
708 g_return_val_if_fail (s1 != NULL, 0);
709 g_return_val_if_fail (s2 != NULL, 0);
711 n = MIN (strlen (s1), strlen (s2));
713 /* code from GLib */
715 #ifdef HAVE_STRNCASECMP
716 return strncasecmp (s1, s2, n);
717 #else
719 for (; *s1 != '\0' && *s2 != '\0'; s1++, s2++)
721 gint c1, c2;
723 n--;
724 /* According to A. Cox, some platforms have islower's that
725 * don't work right on non-uppercase
727 c1 = isupper ((guchar) * s1) ? tolower ((guchar) * s1) : *s1;
728 c2 = isupper ((guchar) * s2) ? tolower ((guchar) * s2) : *s2;
729 if (c1 != c2)
730 return (c1 - c2);
733 if (n == 0)
734 return 0;
736 return (((gint) (guchar) * s1) - ((gint) (guchar) * s2));
738 #endif
741 /* --------------------------------------------------------------------------------------------- */
743 static int
744 str_8bit_prefix (const char *text, const char *prefix)
746 int result;
748 for (result = 0; text[result] != '\0' && prefix[result] != '\0'
749 && text[result] == prefix[result]; result++);
751 return result;
754 /* --------------------------------------------------------------------------------------------- */
756 static int
757 str_8bit_caseprefix (const char *text, const char *prefix)
759 int result;
761 for (result = 0; text[result] != '\0' && prefix[result] != '\0'
762 && char_toupper (text[result]) == char_toupper (prefix[result]); result++);
764 return result;
767 /* --------------------------------------------------------------------------------------------- */
769 static void
770 str_8bit_fix_string (char *text)
772 (void) text;
775 /* --------------------------------------------------------------------------------------------- */
777 static char *
778 str_8bit_create_key (const char *text, gboolean case_sen)
780 return case_sen ? (char *) text : str_8bit_strdown (text);
783 /* --------------------------------------------------------------------------------------------- */
785 static int
786 str_8bit_key_collate (const char *t1, const char *t2, gboolean case_sen)
788 return case_sen ? strcmp (t1, t2) : strcoll (t1, t2);
791 /* --------------------------------------------------------------------------------------------- */
793 static void
794 str_8bit_release_key (char *key, gboolean case_sen)
796 if (!case_sen)
797 g_free (key);
800 /* --------------------------------------------------------------------------------------------- */
801 /*** public functions ****************************************************************************/
802 /* --------------------------------------------------------------------------------------------- */
804 struct str_class
805 str_8bit_init (void)
807 struct str_class result;
809 result.conv_gerror_message = str_8bit_conv_gerror_message;
810 result.vfs_convert_to = str_8bit_vfs_convert_to;
811 result.insert_replace_char = str_8bit_insert_replace_char;
812 result.is_valid_string = str_8bit_is_valid_string;
813 result.is_valid_char = str_8bit_is_valid_char;
814 result.cnext_char = str_8bit_cnext_char;
815 result.cprev_char = str_8bit_cprev_char;
816 result.cnext_char_safe = str_8bit_cnext_char;
817 result.cprev_char_safe = str_8bit_cprev_char;
818 result.cnext_noncomb_char = str_8bit_cnext_noncomb_char;
819 result.cprev_noncomb_char = str_8bit_cprev_noncomb_char;
820 result.char_isspace = str_8bit_isspace;
821 result.char_ispunct = str_8bit_ispunct;
822 result.char_isalnum = str_8bit_isalnum;
823 result.char_isdigit = str_8bit_isdigit;
824 result.char_isprint = str_8bit_isprint;
825 result.char_iscombiningmark = str_8bit_iscombiningmark;
826 result.char_toupper = str_8bit_toupper;
827 result.char_tolower = str_8bit_tolower;
828 result.length = str_8bit_length;
829 result.length2 = str_8bit_length2;
830 result.length_noncomb = str_8bit_length;
831 result.fix_string = str_8bit_fix_string;
832 result.term_form = str_8bit_term_form;
833 result.fit_to_term = str_8bit_fit_to_term;
834 result.term_trim = str_8bit_term_trim;
835 result.term_width2 = str_8bit_term_width2;
836 result.term_width1 = str_8bit_term_width1;
837 result.term_char_width = str_8bit_term_char_width;
838 result.term_substring = str_8bit_term_substring;
839 result.trunc = str_8bit_trunc;
840 result.offset_to_pos = str_8bit_offset_to_pos;
841 result.column_to_pos = str_8bit_column_to_pos;
842 result.create_search_needle = str_8bit_create_search_needle;
843 result.release_search_needle = str_8bit_release_search_needle;
844 result.search_first = str_8bit_search_first;
845 result.search_last = str_8bit_search_last;
846 result.compare = str_8bit_compare;
847 result.ncompare = str_8bit_ncompare;
848 result.casecmp = str_8bit_casecmp;
849 result.ncasecmp = str_8bit_ncasecmp;
850 result.prefix = str_8bit_prefix;
851 result.caseprefix = str_8bit_caseprefix;
852 result.create_key = str_8bit_create_key;
853 result.create_key_for_filename = str_8bit_create_key;
854 result.key_collate = str_8bit_key_collate;
855 result.release_key = str_8bit_release_key;
857 return result;
860 /* --------------------------------------------------------------------------------------------- */