lib/search/search.c

   1 /*
   2    Search text engine.
   3    Interface functions
   4
   5    Copyright (C) 2009-2019
   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 <http://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 /*** file scope variables ************************************************************************/
  51
  52 static const mc_search_type_str_t mc_search__list_types[] = {
  53     {N_("No&rmal"), MC_SEARCH_T_NORMAL},
  54     {N_("Re&gular expression"), MC_SEARCH_T_REGEX},
  55     {N_("He&xadecimal"), MC_SEARCH_T_HEX},
  56     {N_("Wil&dcard search"), MC_SEARCH_T_GLOB},
  57     {NULL, MC_SEARCH_T_INVALID}
  58 };
  59
  60 /*** file scope functions ************************************************************************/
  61
  62 static mc_search_cond_t *
  63 mc_search__cond_struct_new (mc_search_t * lc_mc_search, const char *str,
  64                             gsize str_len, const char *charset)
  65 {
  66     mc_search_cond_t *mc_search_cond;
  67
  68     mc_search_cond = g_malloc0 (sizeof (mc_search_cond_t));
  69     mc_search_cond->str = g_string_new_len (str, str_len);
  70     mc_search_cond->charset = g_strdup (charset);
  71
  72     switch (lc_mc_search->search_type)
  73     {
  74     case MC_SEARCH_T_GLOB:
  75         mc_search__cond_struct_new_init_glob (charset, lc_mc_search, mc_search_cond);
  76         break;
  77     case MC_SEARCH_T_NORMAL:
  78         mc_search__cond_struct_new_init_normal (charset, lc_mc_search, mc_search_cond);
  79         break;
  80     case MC_SEARCH_T_REGEX:
  81         mc_search__cond_struct_new_init_regex (charset, lc_mc_search, mc_search_cond);
  82         break;
  83     case MC_SEARCH_T_HEX:
  84         mc_search__cond_struct_new_init_hex (charset, lc_mc_search, mc_search_cond);
  85         break;
  86     default:
  87         break;
  88     }
  89     return mc_search_cond;
  90 }
  91
  92 /* --------------------------------------------------------------------------------------------- */
  93
  94 static void
  95 mc_search__cond_struct_free (mc_search_cond_t * mc_search_cond)
  96 {
  97     if (mc_search_cond->upper)
  98         g_string_free (mc_search_cond->upper, TRUE);
  99
 100     if (mc_search_cond->lower)
 101         g_string_free (mc_search_cond->lower, TRUE);
 102
 103     g_string_free (mc_search_cond->str, TRUE);
 104     g_free (mc_search_cond->charset);
 105
 106 #ifdef SEARCH_TYPE_GLIB
 107     if (mc_search_cond->regex_handle)
 108         g_regex_unref (mc_search_cond->regex_handle);
 109 #else /* SEARCH_TYPE_GLIB */
 110     g_free (mc_search_cond->regex_handle);
 111 #endif /* SEARCH_TYPE_GLIB */
 112
 113     g_free (mc_search_cond);
 114 }
 115
 116 /* --------------------------------------------------------------------------------------------- */
 117
 118 static void
 119 mc_search__conditions_free (GPtrArray * array)
 120 {
 121     g_ptr_array_foreach (array, (GFunc) mc_search__cond_struct_free, NULL);
 122     g_ptr_array_free (array, TRUE);
 123 }
 124
 125 /* --------------------------------------------------------------------------------------------- */
 126 /*** public functions ****************************************************************************/
 127 /* --------------------------------------------------------------------------------------------- */
 128 /* Init search descriptor.
 129  *
 130  * @param original pattern to search
 131  * @param original_charset charset of #original. If NULL then cp_display will be used
 132  *
 133  * @return new mc_search_t object. Use #mc_search_free() to free it.
 134  */
 135
 136 mc_search_t *
 137 mc_search_new (const gchar * original, const gchar * original_charset)
 138 {
 139     if (original == NULL)
 140         return NULL;
 141
 142     return mc_search_new_len (original, strlen (original), original_charset);
 143 }
 144
 145 /* --------------------------------------------------------------------------------------------- */
 146 /* Init search descriptor.
 147  *
 148  * @param original pattern to search
 149  * @param original_len length of #original or -1 if #original is NULL-terminated
 150  * @param original_charset charset of #original. If NULL then cp_display will be used
 151  *
 152  * @return new mc_search_t object. Use #mc_search_free() to free it.
 153  */
 154
 155 mc_search_t *
 156 mc_search_new_len (const gchar * original, gsize original_len, const gchar * original_charset)
 157 {
 158     mc_search_t *lc_mc_search;
 159
 160     if (original == NULL || original_len == 0)
 161         return NULL;
 162
 163     lc_mc_search = g_new0 (mc_search_t, 1);
 164     lc_mc_search->original = g_strndup (original, original_len);
 165     lc_mc_search->original_len = original_len;
 166 #ifdef HAVE_CHARSET
 167     lc_mc_search->original_charset =
 168         g_strdup (original_charset != NULL
 169                   && *original_charset != '\0' ? original_charset : cp_display);
 170 #else
 171     (void) original_charset;
 172 #endif
 173
 174     return lc_mc_search;
 175 }
 176
 177 /* --------------------------------------------------------------------------------------------- */
 178
 179 void
 180 mc_search_free (mc_search_t * lc_mc_search)
 181 {
 182     if (lc_mc_search == NULL)
 183         return;
 184
 185     g_free (lc_mc_search->original);
 186 #ifdef HAVE_CHARSET
 187     g_free (lc_mc_search->original_charset);
 188 #endif
 189     g_free (lc_mc_search->error_str);
 190
 191     if (lc_mc_search->conditions != NULL)
 192         mc_search__conditions_free (lc_mc_search->conditions);
 193
 194 #ifdef SEARCH_TYPE_GLIB
 195     if (lc_mc_search->regex_match_info != NULL)
 196         g_match_info_free (lc_mc_search->regex_match_info);
 197 #else /* SEARCH_TYPE_GLIB */
 198     g_free (lc_mc_search->regex_match_info);
 199 #endif /* SEARCH_TYPE_GLIB */
 200
 201     if (lc_mc_search->regex_buffer != NULL)
 202         g_string_free (lc_mc_search->regex_buffer, TRUE);
 203
 204     g_free (lc_mc_search);
 205 }
 206
 207 /* --------------------------------------------------------------------------------------------- */
 208
 209 gboolean
 210 mc_search_prepare (mc_search_t * lc_mc_search)
 211 {
 212     GPtrArray *ret;
 213
 214     ret = g_ptr_array_new ();
 215 #ifdef HAVE_CHARSET
 216     if (lc_mc_search->is_all_charsets)
 217     {
 218         gsize loop1;
 219
 220         for (loop1 = 0; loop1 < codepages->len; loop1++)
 221         {
 222             const char *id;
 223             gsize recoded_str_len;
 224             gchar *buffer;
 225
 226             id = ((codepage_desc *) g_ptr_array_index (codepages, loop1))->id;
 227             if (g_ascii_strcasecmp (id, lc_mc_search->original_charset) == 0)
 228             {
 229                 g_ptr_array_add (ret,
 230                                  mc_search__cond_struct_new (lc_mc_search, lc_mc_search->original,
 231                                                              lc_mc_search->original_len,
 232                                                              lc_mc_search->original_charset));
 233                 continue;
 234             }
 235
 236             buffer =
 237                 mc_search__recode_str (lc_mc_search->original, lc_mc_search->original_len,
 238                                        lc_mc_search->original_charset, id, &recoded_str_len);
 239
 240             g_ptr_array_add (ret,
 241                              mc_search__cond_struct_new (lc_mc_search, buffer,
 242                                                          recoded_str_len, id));
 243             g_free (buffer);
 244         }
 245     }
 246     else
 247     {
 248         g_ptr_array_add (ret,
 249                          mc_search__cond_struct_new (lc_mc_search, lc_mc_search->original,
 250                                                      lc_mc_search->original_len,
 251                                                      lc_mc_search->original_charset));
 252     }
 253 #else
 254     g_ptr_array_add (ret,
 255                      mc_search__cond_struct_new (lc_mc_search, lc_mc_search->original,
 256                                                  lc_mc_search->original_len,
 257                                                  str_detect_termencoding ()));
 258 #endif
 259     lc_mc_search->conditions = ret;
 260
 261     return (lc_mc_search->error == MC_SEARCH_E_OK);
 262 }
 263
 264 /* --------------------------------------------------------------------------------------------- */
 265
 266 /**
 267  * Carries out the search.
 268  *
 269  * Returns TRUE if found.
 270  *
 271  * Returns FALSE if not found. In this case, lc_mc_search->error reveals
 272  * the reason:
 273  *
 274  *   - MC_SEARCH_E_NOTFOUND: the pattern isn't in the subject string.
 275  *   - MC_SEARCH_E_ABORT: the user aborted the search.
 276  *   - For any other reason (but not for the above two!): the description
 277  *     is in lc_mc_search->error_str.
 278  */
 279 gboolean
 280 mc_search_run (mc_search_t * lc_mc_search, const void *user_data,
 281                gsize start_search, gsize end_search, gsize * found_len)
 282 {
 283     gboolean ret = FALSE;
 284
 285     if (lc_mc_search == NULL || user_data == NULL)
 286         return FALSE;
 287     if (!mc_search_is_type_avail (lc_mc_search->search_type))
 288     {
 289         mc_search_set_error (lc_mc_search, MC_SEARCH_E_INPUT, "%s", _(STR_E_UNKNOWN_TYPE));
 290         return FALSE;
 291     }
 292 #ifdef SEARCH_TYPE_GLIB
 293     if (lc_mc_search->regex_match_info != NULL)
 294     {
 295         g_match_info_free (lc_mc_search->regex_match_info);
 296         lc_mc_search->regex_match_info = NULL;
 297     }
 298 #endif /* SEARCH_TYPE_GLIB */
 299
 300     mc_search_set_error (lc_mc_search, MC_SEARCH_E_OK, NULL);
 301
 302     if ((lc_mc_search->conditions == NULL) && !mc_search_prepare (lc_mc_search))
 303         return FALSE;
 304
 305     switch (lc_mc_search->search_type)
 306     {
 307     case MC_SEARCH_T_NORMAL:
 308         ret = mc_search__run_normal (lc_mc_search, user_data, start_search, end_search, found_len);
 309         break;
 310     case MC_SEARCH_T_REGEX:
 311         ret = mc_search__run_regex (lc_mc_search, user_data, start_search, end_search, found_len);
 312         break;
 313     case MC_SEARCH_T_GLOB:
 314         ret = mc_search__run_glob (lc_mc_search, user_data, start_search, end_search, found_len);
 315         break;
 316     case MC_SEARCH_T_HEX:
 317         ret = mc_search__run_hex (lc_mc_search, user_data, start_search, end_search, found_len);
 318         break;
 319     default:
 320         break;
 321     }
 322     return ret;
 323 }
 324
 325 /* --------------------------------------------------------------------------------------------- */
 326
 327 gboolean
 328 mc_search_is_type_avail (mc_search_type_t search_type)
 329 {
 330     switch (search_type)
 331     {
 332     case MC_SEARCH_T_GLOB:
 333     case MC_SEARCH_T_NORMAL:
 334     case MC_SEARCH_T_REGEX:
 335     case MC_SEARCH_T_HEX:
 336         return TRUE;
 337     default:
 338         break;
 339     }
 340     return FALSE;
 341 }
 342
 343 /* --------------------------------------------------------------------------------------------- */
 344
 345 const mc_search_type_str_t *
 346 mc_search_types_list_get (size_t * num)
 347 {
 348     /* don't count last NULL item */
 349     if (num != NULL)
 350         *num = G_N_ELEMENTS (mc_search__list_types) - 1;
 351
 352     return mc_search__list_types;
 353 }
 354
 355 /* --------------------------------------------------------------------------------------------- */
 356
 357 GString *
 358 mc_search_prepare_replace_str (mc_search_t * lc_mc_search, GString * replace_str)
 359 {
 360     GString *ret;
 361
 362     if (replace_str == NULL || replace_str->len == 0)
 363         return g_string_new ("");
 364
 365     if (lc_mc_search == NULL)
 366         return g_string_new_len (replace_str->str, replace_str->len);
 367
 368     switch (lc_mc_search->search_type)
 369     {
 370     case MC_SEARCH_T_REGEX:
 371         ret = mc_search_regex_prepare_replace_str (lc_mc_search, replace_str);
 372         break;
 373     case MC_SEARCH_T_GLOB:
 374         ret = mc_search_glob_prepare_replace_str (lc_mc_search, replace_str);
 375         break;
 376     case MC_SEARCH_T_NORMAL:
 377         ret = mc_search_normal_prepare_replace_str (lc_mc_search, replace_str);
 378         break;
 379     case MC_SEARCH_T_HEX:
 380         ret = mc_search_hex_prepare_replace_str (lc_mc_search, replace_str);
 381         break;
 382     default:
 383         ret = g_string_new_len (replace_str->str, replace_str->len);
 384         break;
 385     }
 386     return ret;
 387 }
 388
 389 /* --------------------------------------------------------------------------------------------- */
 390
 391 char *
 392 mc_search_prepare_replace_str2 (mc_search_t * lc_mc_search, const char *replace_str)
 393 {
 394     GString *ret;
 395     GString *replace_str2;
 396
 397     replace_str2 = g_string_new (replace_str);
 398     ret = mc_search_prepare_replace_str (lc_mc_search, replace_str2);
 399     g_string_free (replace_str2, TRUE);
 400     return (ret != NULL) ? g_string_free (ret, FALSE) : NULL;
 401 }
 402
 403 /* --------------------------------------------------------------------------------------------- */
 404
 405 gboolean
 406 mc_search_is_fixed_search_str (mc_search_t * lc_mc_search)
 407 {
 408     if (lc_mc_search == NULL)
 409         return FALSE;
 410     switch (lc_mc_search->search_type)
 411     {
 412     case MC_SEARCH_T_REGEX:
 413     case MC_SEARCH_T_GLOB:
 414         return FALSE;
 415     default:
 416         return TRUE;
 417     }
 418 }
 419
 420 /* --------------------------------------------------------------------------------------------- */
 421 /* Search specified pattern in specified string.
 422  *
 423  * @param pattern string to search
 424  * @param pattern_charset charset of #pattern. If NULL then cp_display will be used
 425  * @param str string where search #pattern
 426  * @param search type (normal, regex, hex or glob)
 427  *
 428  * @return TRUE if found is successful, FALSE otherwise.
 429  */
 430
 431 gboolean
 432 mc_search (const gchar * pattern, const gchar * pattern_charset, const gchar * str,
 433            mc_search_type_t type)
 434 {
 435     gboolean ret;
 436     mc_search_t *search;
 437
 438     if (str == NULL)
 439         return FALSE;
 440
 441     search = mc_search_new (pattern, pattern_charset);
 442     if (search == NULL)
 443         return FALSE;
 444
 445     search->search_type = type;
 446     search->is_case_sensitive = TRUE;
 447
 448     if (type == MC_SEARCH_T_GLOB)
 449         search->is_entire_line = TRUE;
 450
 451     ret = mc_search_run (search, str, 0, strlen (str), NULL);
 452     mc_search_free (search);
 453     return ret;
 454 }
 455
 456 /* --------------------------------------------------------------------------------------------- */
 457
 458 int
 459 mc_search_getstart_result_by_num (mc_search_t * lc_mc_search, int lc_index)
 460 {
 461     if (lc_mc_search == NULL)
 462         return 0;
 463     if (lc_mc_search->search_type == MC_SEARCH_T_NORMAL)
 464         return 0;
 465 #ifdef SEARCH_TYPE_GLIB
 466     {
 467         gint start_pos;
 468         gint end_pos;
 469
 470         g_match_info_fetch_pos (lc_mc_search->regex_match_info, lc_index, &start_pos, &end_pos);
 471         return (int) start_pos;
 472     }
 473 #else /* SEARCH_TYPE_GLIB */
 474     return lc_mc_search->iovector[lc_index * 2];
 475 #endif /* SEARCH_TYPE_GLIB */
 476 }
 477
 478 /* --------------------------------------------------------------------------------------------- */
 479
 480 int
 481 mc_search_getend_result_by_num (mc_search_t * lc_mc_search, int lc_index)
 482 {
 483     if (lc_mc_search == NULL)
 484         return 0;
 485     if (lc_mc_search->search_type == MC_SEARCH_T_NORMAL)
 486         return 0;
 487 #ifdef SEARCH_TYPE_GLIB
 488     {
 489         gint start_pos;
 490         gint end_pos;
 491
 492         g_match_info_fetch_pos (lc_mc_search->regex_match_info, lc_index, &start_pos, &end_pos);
 493         return (int) end_pos;
 494     }
 495 #else /* SEARCH_TYPE_GLIB */
 496     return lc_mc_search->iovector[lc_index * 2 + 1];
 497 #endif /* SEARCH_TYPE_GLIB */
 498 }
 499
 500 /* --------------------------------------------------------------------------------------------- */
 501 /**
 502  * Replace an old error code and message of an mc_search_t object.
 503  *
 504  * @param mc_search mc_search_t object
 505  * @param code error code, one of mc_search_error_t values
 506  * @param format format of error message. If NULL, the old error string is free'd and become NULL
 507  */
 508
 509 void
 510 mc_search_set_error (mc_search_t * lc_mc_search, mc_search_error_t code, const gchar * format, ...)
 511 {
 512     lc_mc_search->error = code;
 513
 514     MC_PTR_FREE (lc_mc_search->error_str);
 515
 516     if (format != NULL)
 517     {
 518         va_list args;
 519
 520         va_start (args, format);
 521         lc_mc_search->error_str = g_strdup_vprintf (format, args);
 522         va_end (args);
 523     }
 524 }
 525
 526 /* --------------------------------------------------------------------------------------------- */