lib/search/search.c

lib/search/search.c
/* */
This source file includes following definitions.
mc_search__cond_struct_new
mc_search__cond_struct_free
mc_search_new
mc_search_new_len
mc_search_free
mc_search_prepare
mc_search_run
mc_search_is_type_avail
mc_search_types_list_get
mc_search_prepare_replace_str
mc_search_prepare_replace_str2
mc_search_is_fixed_search_str
mc_search
mc_search_getstart_result_by_num
mc_search_getend_result_by_num
mc_search_set_error
   1 /*
   2    Search text engine.
   3    Interface functions
   4 
   5    Copyright (C) 2009-2025
   6    Free Software Foundation, Inc.
   7 
   8    Written by:
   9    Slava Zanko <slavazanko@gmail.com>, 2009
  10    Andrew Borodin <aborodin@vmail.ru>, 2013
  11 
  12    This file is part of the Midnight Commander.
  13 
  14    The Midnight Commander is free software: you can redistribute it
  15    and/or modify it under the terms of the GNU General Public License as
  16    published by the Free Software Foundation, either version 3 of the License,
  17    or (at your option) any later version.
  18 
  19    The Midnight Commander is distributed in the hope that it will be useful,
  20    but WITHOUT ANY WARRANTY; without even the implied warranty of
  21    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  22    GNU General Public License for more details.
  23 
  24    You should have received a copy of the GNU General Public License
  25    along with this program.  If not, see <https://www.gnu.org/licenses/>.
  26  */
  27 
  28 #include <config.h>
  29 
  30 #include <stdarg.h>
  31 #include <stdlib.h>
  32 #include <sys/types.h>
  33 
  34 #include "lib/global.h"
  35 #include "lib/strutil.h"
  36 #include "lib/search.h"
  37 #include "lib/util.h"
  38 #ifdef HAVE_CHARSET
  39 #    include "lib/charsets.h"
  40 #endif
  41 
  42 #include "internal.h"
  43 
  44 /*** global variables ****************************************************************************/
  45 
  46 /*** file scope macro definitions ****************************************************************/
  47 
  48 /*** file scope type declarations ****************************************************************/
  49 
  50 /*** forward declarations (file scope functions) *************************************************/
  51 
  52 /*** file scope variables ************************************************************************/
  53 
  54 static const mc_search_type_str_t mc_search__list_types[] = {
  55     { N_ ("No&rmal"), MC_SEARCH_T_NORMAL },
  56     { N_ ("Re&gular expression"), MC_SEARCH_T_REGEX },
  57     { N_ ("He&xadecimal"), MC_SEARCH_T_HEX },
  58     { N_ ("Wil&dcard search"), MC_SEARCH_T_GLOB },
  59     { NULL, MC_SEARCH_T_INVALID }
  60 };
  61 
  62 /* --------------------------------------------------------------------------------------------- */
  63 /*** file scope functions ************************************************************************/
  64 /* --------------------------------------------------------------------------------------------- */
  65 
  66 static mc_search_cond_t *
  67 mc_search__cond_struct_new (mc_search_t *lc_mc_search, const GString *str, const char *charset)
     /*   */
  68 {
  69     mc_search_cond_t *mc_search_cond;
  70 
  71     mc_search_cond = g_malloc0 (sizeof (mc_search_cond_t));
  72     mc_search_cond->str = mc_g_string_dup (str);
  73     mc_search_cond->charset = g_strdup (charset);
  74 #ifdef HAVE_PCRE2
  75     lc_mc_search->regex_match_info = pcre2_match_data_create (MC_SEARCH__NUM_REPLACE_ARGS, NULL);
  76     lc_mc_search->iovector = pcre2_get_ovector_pointer (lc_mc_search->regex_match_info);
  77 #endif
  78     switch (lc_mc_search->search_type)
  79     {
  80     case MC_SEARCH_T_GLOB:
  81         mc_search__cond_struct_new_init_glob (charset, lc_mc_search, mc_search_cond);
  82         break;
  83     case MC_SEARCH_T_NORMAL:
  84         mc_search__cond_struct_new_init_normal (charset, lc_mc_search, mc_search_cond);
  85         break;
  86     case MC_SEARCH_T_REGEX:
  87         mc_search__cond_struct_new_init_regex (charset, lc_mc_search, mc_search_cond);
  88         break;
  89     case MC_SEARCH_T_HEX:
  90         mc_search__cond_struct_new_init_hex (charset, lc_mc_search, mc_search_cond);
  91         break;
  92     default:
  93         break;
  94     }
  95     return mc_search_cond;
  96 }
  97 
  98 /* --------------------------------------------------------------------------------------------- */
  99 
 100 static void
 101 mc_search__cond_struct_free (gpointer data)
     /*   */
 102 {
 103     mc_search_cond_t *mc_search_cond = (mc_search_cond_t *) data;
 104 
 105     if (mc_search_cond->upper != NULL)
 106         g_string_free (mc_search_cond->upper, TRUE);
 107 
 108     if (mc_search_cond->lower != NULL)
 109         g_string_free (mc_search_cond->lower, TRUE);
 110 
 111     g_string_free (mc_search_cond->str, TRUE);
 112     g_free (mc_search_cond->charset);
 113 
 114 #ifdef SEARCH_TYPE_GLIB
 115     if (mc_search_cond->regex_handle != NULL)
 116         g_regex_unref (mc_search_cond->regex_handle);
 117 #else  // SEARCH_TYPE_GLIB
 118     g_free (mc_search_cond->regex_handle);
 119 #endif
 120 
 121     g_free (mc_search_cond);
 122 }
 123 
 124 /* --------------------------------------------------------------------------------------------- */
 125 /*** public functions ****************************************************************************/
 126 /* --------------------------------------------------------------------------------------------- */
 127 /* Init search descriptor.
 128  *
 129  * @param original pattern to search
 130  * @param original_charset charset of #original. If NULL then cp_display will be used
 131  *
 132  * @return new mc_search_t object. Use #mc_search_free() to free it.
 133  */
 134 
 135 mc_search_t *
 136 mc_search_new (const gchar *original, const gchar *original_charset)
     /*   */
 137 {
 138     if (original == NULL)
 139         return NULL;
 140 
 141     return mc_search_new_len (original, strlen (original), original_charset);
 142 }
 143 
 144 /* --------------------------------------------------------------------------------------------- */
 145 /* Init search descriptor.
 146  *
 147  * @param original pattern to search
 148  * @param original_len length of #original or -1 if #original is NULL-terminated
 149  * @param original_charset charset of #original. If NULL then cp_display will be used
 150  *
 151  * @return new mc_search_t object. Use #mc_search_free() to free it.
 152  */
 153 
 154 mc_search_t *
 155 mc_search_new_len (const gchar *original, gsize original_len, const gchar *original_charset)
     /*   */
 156 {
 157     mc_search_t *lc_mc_search;
 158 
 159     if (original == NULL || original_len == 0)
 160         return NULL;
 161 
 162     lc_mc_search = g_new0 (mc_search_t, 1);
 163     lc_mc_search->original.str = g_string_new_len (original, original_len);
 164 #ifdef HAVE_CHARSET
 165     lc_mc_search->original.charset = g_strdup (
 166         original_charset != NULL && *original_charset != '\0' ? original_charset : cp_display);
 167 #else
 168     (void) original_charset;
 169 #endif
 170 
 171     return lc_mc_search;
 172 }
 173 
 174 /* --------------------------------------------------------------------------------------------- */
 175 
 176 void
 177 mc_search_free (mc_search_t *lc_mc_search)
     /*   */
 178 {
 179     if (lc_mc_search == NULL)
 180         return;
 181 
 182     g_string_free (lc_mc_search->original.str, TRUE);
 183 #ifdef HAVE_CHARSET
 184     g_free (lc_mc_search->original.charset);
 185 #endif
 186     g_free (lc_mc_search->error_str);
 187 
 188     if (lc_mc_search->prepared.conditions != NULL)
 189         g_ptr_array_free (lc_mc_search->prepared.conditions, TRUE);
 190 
 191 #ifdef SEARCH_TYPE_GLIB
 192     if (lc_mc_search->regex_match_info != NULL)
 193         g_match_info_free (lc_mc_search->regex_match_info);
 194 #else  // SEARCH_TYPE_GLIB
 195     g_free (lc_mc_search->regex_match_info);
 196 #endif
 197 
 198     if (lc_mc_search->regex_buffer != NULL)
 199         g_string_free (lc_mc_search->regex_buffer, TRUE);
 200 
 201     g_free (lc_mc_search);
 202 }
 203 
 204 /* --------------------------------------------------------------------------------------------- */
 205 
 206 gboolean
 207 mc_search_prepare (mc_search_t *lc_mc_search)
     /*   */
 208 {
 209     GPtrArray *ret;
 210 
 211     if (lc_mc_search->prepared.conditions != NULL)
 212         return lc_mc_search->prepared.result;
 213 
 214     ret = g_ptr_array_new_with_free_func (mc_search__cond_struct_free);
 215 #ifdef HAVE_CHARSET
 216     if (!lc_mc_search->is_all_charsets)
 217         g_ptr_array_add (ret,
 218                          mc_search__cond_struct_new (lc_mc_search, lc_mc_search->original.str,
 219                                                      lc_mc_search->original.charset));
 220     else
 221     {
 222         gsize loop1;
 223 
 224         for (loop1 = 0; loop1 < codepages->len; loop1++)
 225         {
 226             const char *id;
 227 
 228             id = ((codepage_desc *) g_ptr_array_index (codepages, loop1))->id;
 229             if (g_ascii_strcasecmp (id, lc_mc_search->original.charset) == 0)
 230                 g_ptr_array_add (ret,
 231                                  mc_search__cond_struct_new (lc_mc_search,
 232                                                              lc_mc_search->original.str,
 233                                                              lc_mc_search->original.charset));
 234             else
 235             {
 236                 GString *buffer;
 237 
 238                 buffer = mc_search__recode_str (lc_mc_search->original.str->str,
 239                                                 lc_mc_search->original.str->len,
 240                                                 lc_mc_search->original.charset, id);
 241                 g_ptr_array_add (ret, mc_search__cond_struct_new (lc_mc_search, buffer, id));
 242                 g_string_free (buffer, TRUE);
 243             }
 244         }
 245     }
 246 #else
 247     g_ptr_array_add (ret,
 248                      mc_search__cond_struct_new (lc_mc_search, lc_mc_search->original.str,
 249                                                  str_detect_termencoding ()));
 250 #endif
 251     lc_mc_search->prepared.conditions = ret;
 252     lc_mc_search->prepared.result = (lc_mc_search->error == MC_SEARCH_E_OK);
 253 
 254     return lc_mc_search->prepared.result;
 255 }
 256 
 257 /* --------------------------------------------------------------------------------------------- */
 258 
 259 /**
 260  * Carries out the search.
 261  *
 262  * Returns TRUE if found.
 263  *
 264  * Returns FALSE if not found. In this case, lc_mc_search->error reveals
 265  * the reason:
 266  *
 267  *   - MC_SEARCH_E_NOTFOUND: the pattern isn't in the subject string.
 268  *   - MC_SEARCH_E_ABORT: the user aborted the search.
 269  *   - For any other reason (but not for the above two!): the description
 270  *     is in lc_mc_search->error_str.
 271  */
 272 gboolean
 273 mc_search_run (mc_search_t *lc_mc_search, const void *user_data, off_t start_search,
     /*   */
 274                off_t end_search, gsize *found_len)
 275 {
 276     gboolean ret = FALSE;
 277 
 278     if (lc_mc_search == NULL || user_data == NULL)
 279         return FALSE;
 280     if (!mc_search_is_type_avail (lc_mc_search->search_type))
 281     {
 282         mc_search_set_error (lc_mc_search, MC_SEARCH_E_INPUT, "%s", _ (STR_E_UNKNOWN_TYPE));
 283         return FALSE;
 284     }
 285 #ifdef SEARCH_TYPE_GLIB
 286     if (lc_mc_search->regex_match_info != NULL)
 287     {
 288         g_match_info_free (lc_mc_search->regex_match_info);
 289         lc_mc_search->regex_match_info = NULL;
 290     }
 291 #endif
 292 
 293     mc_search_set_error (lc_mc_search, MC_SEARCH_E_OK, NULL);
 294 
 295     if (!mc_search_prepare (lc_mc_search))
 296         return FALSE;
 297 
 298     switch (lc_mc_search->search_type)
 299     {
 300     case MC_SEARCH_T_NORMAL:
 301         ret = mc_search__run_normal (lc_mc_search, user_data, start_search, end_search, found_len);
 302         break;
 303     case MC_SEARCH_T_REGEX:
 304         ret = mc_search__run_regex (lc_mc_search, user_data, start_search, end_search, found_len);
 305         break;
 306     case MC_SEARCH_T_GLOB:
 307         ret = mc_search__run_glob (lc_mc_search, user_data, start_search, end_search, found_len);
 308         break;
 309     case MC_SEARCH_T_HEX:
 310         ret = mc_search__run_hex (lc_mc_search, user_data, start_search, end_search, found_len);
 311         break;
 312     default:
 313         break;
 314     }
 315     return ret;
 316 }
 317 
 318 /* --------------------------------------------------------------------------------------------- */
 319 
 320 gboolean
 321 mc_search_is_type_avail (mc_search_type_t search_type)
     /*   */
 322 {
 323     switch (search_type)
 324     {
 325     case MC_SEARCH_T_GLOB:
 326     case MC_SEARCH_T_NORMAL:
 327     case MC_SEARCH_T_REGEX:
 328     case MC_SEARCH_T_HEX:
 329         return TRUE;
 330     default:
 331         break;
 332     }
 333     return FALSE;
 334 }
 335 
 336 /* --------------------------------------------------------------------------------------------- */
 337 
 338 const mc_search_type_str_t *
 339 mc_search_types_list_get (size_t *num)
     /*   */
 340 {
 341     // don't count last NULL item
 342     if (num != NULL)
 343         *num = G_N_ELEMENTS (mc_search__list_types) - 1;
 344 
 345     return mc_search__list_types;
 346 }
 347 
 348 /* --------------------------------------------------------------------------------------------- */
 349 
 350 GString *
 351 mc_search_prepare_replace_str (mc_search_t *lc_mc_search, GString *replace_str)
     /*   */
 352 {
 353     GString *ret;
 354 
 355     if (replace_str == NULL || replace_str->len == 0)
 356         return g_string_new ("");
 357 
 358     if (lc_mc_search == NULL)
 359         return mc_g_string_dup (replace_str);
 360 
 361     switch (lc_mc_search->search_type)
 362     {
 363     case MC_SEARCH_T_REGEX:
 364         ret = mc_search_regex_prepare_replace_str (lc_mc_search, replace_str);
 365         break;
 366     case MC_SEARCH_T_GLOB:
 367         ret = mc_search_glob_prepare_replace_str (lc_mc_search, replace_str);
 368         break;
 369     case MC_SEARCH_T_NORMAL:
 370         ret = mc_search_normal_prepare_replace_str (lc_mc_search, replace_str);
 371         break;
 372     case MC_SEARCH_T_HEX:
 373         ret = mc_search_hex_prepare_replace_str (lc_mc_search, replace_str);
 374         break;
 375     default:
 376         ret = mc_g_string_dup (replace_str);
 377         break;
 378     }
 379     return ret;
 380 }
 381 
 382 /* --------------------------------------------------------------------------------------------- */
 383 
 384 char *
 385 mc_search_prepare_replace_str2 (mc_search_t *lc_mc_search, const char *replace_str)
     /*   */
 386 {
 387     GString *ret;
 388     GString *replace_str2;
 389 
 390     replace_str2 = g_string_new (replace_str);
 391     ret = mc_search_prepare_replace_str (lc_mc_search, replace_str2);
 392     g_string_free (replace_str2, TRUE);
 393     return (ret != NULL) ? g_string_free (ret, FALSE) : NULL;
 394 }
 395 
 396 /* --------------------------------------------------------------------------------------------- */
 397 
 398 gboolean
 399 mc_search_is_fixed_search_str (const mc_search_t *lc_mc_search)
     /*   */
 400 {
 401     if (lc_mc_search == NULL)
 402         return FALSE;
 403     switch (lc_mc_search->search_type)
 404     {
 405     case MC_SEARCH_T_REGEX:
 406     case MC_SEARCH_T_GLOB:
 407         return FALSE;
 408     default:
 409         return TRUE;
 410     }
 411 }
 412 
 413 /* --------------------------------------------------------------------------------------------- */
 414 /* Search specified pattern in specified string.
 415  *
 416  * @param pattern string to search
 417  * @param pattern_charset charset of #pattern. If NULL then cp_display will be used
 418  * @param str string where search #pattern
 419  * @param search type (normal, regex, hex or glob)
 420  *
 421  * @return TRUE if found is successful, FALSE otherwise.
 422  */
 423 
 424 gboolean
 425 mc_search (const gchar *pattern, const gchar *pattern_charset, const gchar *str,
     /*   */
 426            mc_search_type_t type)
 427 {
 428     gboolean ret;
 429     mc_search_t *search;
 430 
 431     if (str == NULL)
 432         return FALSE;
 433 
 434     search = mc_search_new (pattern, pattern_charset);
 435     if (search == NULL)
 436         return FALSE;
 437 
 438     search->search_type = type;
 439     search->is_case_sensitive = TRUE;
 440 
 441     if (type == MC_SEARCH_T_GLOB)
 442         search->is_entire_line = TRUE;
 443 
 444     ret = mc_search_run (search, str, 0, strlen (str), NULL);
 445     mc_search_free (search);
 446     return ret;
 447 }
 448 
 449 /* --------------------------------------------------------------------------------------------- */
 450 
 451 int
 452 mc_search_getstart_result_by_num (mc_search_t *lc_mc_search, int lc_index)
     /*   */
 453 {
 454     if (lc_mc_search == NULL)
 455         return 0;
 456     if (lc_mc_search->search_type == MC_SEARCH_T_NORMAL)
 457         return 0;
 458 #ifdef SEARCH_TYPE_GLIB
 459     {
 460         gint start_pos;
 461         gint end_pos;
 462 
 463         g_match_info_fetch_pos (lc_mc_search->regex_match_info, lc_index, &start_pos, &end_pos);
 464         return (int) start_pos;
 465     }
 466 #else  // SEARCH_TYPE_GLIB
 467     return lc_mc_search->iovector[lc_index * 2];
 468 #endif
 469 }
 470 
 471 /* --------------------------------------------------------------------------------------------- */
 472 
 473 int
 474 mc_search_getend_result_by_num (mc_search_t *lc_mc_search, int lc_index)
     /*   */
 475 {
 476     if (lc_mc_search == NULL)
 477         return 0;
 478     if (lc_mc_search->search_type == MC_SEARCH_T_NORMAL)
 479         return 0;
 480 #ifdef SEARCH_TYPE_GLIB
 481     {
 482         gint start_pos;
 483         gint end_pos;
 484 
 485         g_match_info_fetch_pos (lc_mc_search->regex_match_info, lc_index, &start_pos, &end_pos);
 486         return (int) end_pos;
 487     }
 488 #else  // SEARCH_TYPE_GLIB
 489     return lc_mc_search->iovector[lc_index * 2 + 1];
 490 #endif
 491 }
 492 
 493 /* --------------------------------------------------------------------------------------------- */
 494 /**
 495  * Replace an old error code and message of an mc_search_t object.
 496  *
 497  * @param mc_search mc_search_t object
 498  * @param code error code, one of mc_search_error_t values
 499  * @param format format of error message. If NULL, the old error string is free'd and become NULL
 500  */
 501 
 502 void
 503 mc_search_set_error (mc_search_t *lc_mc_search, mc_search_error_t code, const gchar *format, ...)
     /*   */
 504 {
 505     lc_mc_search->error = code;
 506 
 507     MC_PTR_FREE (lc_mc_search->error_str);
 508 
 509     if (format != NULL)
 510     {
 511         va_list args;
 512 
 513         va_start (args, format);
 514         lc_mc_search->error_str = g_strdup_vprintf (format, args);
 515         va_end (args);
 516     }
 517 }
 518 
 519 /* --------------------------------------------------------------------------------------------- */
/* */
root/lib/search/search.c

DEFINITIONS
