lib/search/search.c

   1 /*
   2    Search text engine.
   3    Interface functions
   4
   5    Copyright (C) 2009, 2011
   6    The Free Software Foundation, Inc.
   7
   8    Written by:
   9    Slava Zanko <slavazanko@gmail.com>, 2009.
  10
  11    This file is part of the Midnight Commander.
  12
  13    The Midnight Commander is free software: you can redistribute it
  14    and/or modify it under the terms of the GNU General Public License as
  15    published by the Free Software Foundation, either version 3 of the License,
  16    or (at your option) any later version.
  17
  18    The Midnight Commander is distributed in the hope that it will be useful,
  19    but WITHOUT ANY WARRANTY; without even the implied warranty of
  20    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  21    GNU General Public License for more details.
  22
  23    You should have received a copy of the GNU General Public License
  24    along with this program.  If not, see <http://www.gnu.org/licenses/>.
  25  */
  26
  27 #include <config.h>
  28
  29 #include <stdlib.h>
  30 #include <sys/types.h>
  31
  32 #include "lib/global.h"
  33 #include "lib/strutil.h"
  34 #include "lib/search.h"
  35 #ifdef HAVE_CHARSET
  36 #include "lib/charsets.h"
  37 #endif
  38
  39 #include "internal.h"
  40
  41 /*** global variables ****************************************************************************/
  42
  43 /*** file scope macro definitions ****************************************************************/
  44
  45 /*** file scope type declarations ****************************************************************/
  46
  47 /*** file scope variables ************************************************************************/
  48
  49 static const mc_search_type_str_t mc_search__list_types[] = {
  50     {N_("No&rmal"), MC_SEARCH_T_NORMAL},
  51     {N_("Re&gular expression"), MC_SEARCH_T_REGEX},
  52     {N_("He&xadecimal"), MC_SEARCH_T_HEX},
  53     {N_("Wil&dcard search"), MC_SEARCH_T_GLOB},
  54     {NULL, -1}
  55 };
  56
  57 /*** file scope functions ************************************************************************/
  58
  59 static mc_search_cond_t *
  60 mc_search__cond_struct_new (mc_search_t * lc_mc_search, const char *str,
  61                             gsize str_len, const char *charset)
  62 {
  63     mc_search_cond_t *mc_search_cond;
  64     mc_search_cond = g_malloc0 (sizeof (mc_search_cond_t));
  65
  66     mc_search_cond->str = g_string_new_len (str, str_len);
  67     mc_search_cond->charset = g_strdup (charset);
  68
  69     switch (lc_mc_search->search_type)
  70     {
  71     case MC_SEARCH_T_GLOB:
  72         mc_search__cond_struct_new_init_glob (charset, lc_mc_search, mc_search_cond);
  73         break;
  74     case MC_SEARCH_T_NORMAL:
  75         mc_search__cond_struct_new_init_normal (charset, lc_mc_search, mc_search_cond);
  76         break;
  77     case MC_SEARCH_T_REGEX:
  78         mc_search__cond_struct_new_init_regex (charset, lc_mc_search, mc_search_cond);
  79         break;
  80     case MC_SEARCH_T_HEX:
  81         mc_search__cond_struct_new_init_hex (charset, lc_mc_search, mc_search_cond);
  82         break;
  83     default:
  84         break;
  85     }
  86     return mc_search_cond;
  87 }
  88
  89 /* --------------------------------------------------------------------------------------------- */
  90
  91 static void
  92 mc_search__cond_struct_free (mc_search_cond_t * mc_search_cond)
  93 {
  94     if (mc_search_cond->upper)
  95         g_string_free (mc_search_cond->upper, TRUE);
  96
  97     if (mc_search_cond->lower)
  98         g_string_free (mc_search_cond->lower, TRUE);
  99
 100     g_string_free (mc_search_cond->str, TRUE);
 101     g_free (mc_search_cond->charset);
 102
 103 #ifdef SEARCH_TYPE_GLIB
 104     if (mc_search_cond->regex_handle)
 105         g_regex_unref (mc_search_cond->regex_handle);
 106 #else /* SEARCH_TYPE_GLIB */
 107     g_free (mc_search_cond->regex_handle);
 108 #endif /* SEARCH_TYPE_GLIB */
 109
 110     g_free (mc_search_cond);
 111 }
 112
 113 /* --------------------------------------------------------------------------------------------- */
 114
 115 static void
 116 mc_search__conditions_free (GPtrArray * array)
 117 {
 118     gsize loop1;
 119     mc_search_cond_t *lc_mc_search;
 120
 121     for (loop1 = 0; loop1 < array->len; loop1++)
 122     {
 123         lc_mc_search = (mc_search_cond_t *) g_ptr_array_index (array, loop1);
 124         mc_search__cond_struct_free (lc_mc_search);
 125     }
 126     g_ptr_array_free (array, TRUE);
 127 }
 128
 129 /* --------------------------------------------------------------------------------------------- */
 130
 131 /*** public functions ****************************************************************************/
 132
 133 mc_search_t *
 134 mc_search_new (const gchar * original, gsize str_len)
 135 {
 136     mc_search_t *lc_mc_search;
 137     if (!original)
 138         return NULL;
 139
 140     if ((gssize) str_len == -1)
 141     {
 142         str_len = strlen (original);
 143         if (str_len == 0)
 144             return NULL;
 145     }
 146
 147     lc_mc_search = g_malloc0 (sizeof (mc_search_t));
 148     lc_mc_search->original = g_strndup (original, str_len);
 149     lc_mc_search->original_len = str_len;
 150     return lc_mc_search;
 151 }
 152
 153 /* --------------------------------------------------------------------------------------------- */
 154
 155 void
 156 mc_search_free (mc_search_t * lc_mc_search)
 157 {
 158     if (lc_mc_search == NULL)
 159         return;
 160
 161     g_free (lc_mc_search->original);
 162     g_free (lc_mc_search->error_str);
 163
 164     if (lc_mc_search->conditions)
 165         mc_search__conditions_free (lc_mc_search->conditions);
 166
 167 #ifdef SEARCH_TYPE_GLIB
 168     if (lc_mc_search->regex_match_info)
 169         g_match_info_free (lc_mc_search->regex_match_info);
 170 #else /* SEARCH_TYPE_GLIB */
 171     g_free (lc_mc_search->regex_match_info);
 172 #endif /* SEARCH_TYPE_GLIB */
 173
 174     if (lc_mc_search->regex_buffer != NULL)
 175         g_string_free (lc_mc_search->regex_buffer, TRUE);
 176
 177     g_free (lc_mc_search);
 178 }
 179
 180 /* --------------------------------------------------------------------------------------------- */
 181
 182 gboolean
 183 mc_search_prepare (mc_search_t * lc_mc_search)
 184 {
 185     GPtrArray *ret;
 186     ret = g_ptr_array_new ();
 187 #ifdef HAVE_CHARSET
 188     if (lc_mc_search->is_all_charsets)
 189     {
 190         gsize loop1, recoded_str_len;
 191         gchar *buffer;
 192         for (loop1 = 0; loop1 < codepages->len; loop1++)
 193         {
 194             const char *id = ((codepage_desc *) g_ptr_array_index (codepages, loop1))->id;
 195             if (!g_ascii_strcasecmp (id, cp_source))
 196             {
 197                 g_ptr_array_add (ret,
 198                                  mc_search__cond_struct_new (lc_mc_search, lc_mc_search->original,
 199                                                              lc_mc_search->original_len,
 200                                                              cp_source));
 201                 continue;
 202             }
 203
 204             buffer =
 205                 mc_search__recode_str (lc_mc_search->original, lc_mc_search->original_len,
 206                                        cp_source, id, &recoded_str_len);
 207
 208             g_ptr_array_add (ret,
 209                              mc_search__cond_struct_new (lc_mc_search, buffer,
 210                                                          recoded_str_len, id));
 211             g_free (buffer);
 212         }
 213     }
 214     else
 215     {
 216         g_ptr_array_add (ret,
 217                          (gpointer) mc_search__cond_struct_new (lc_mc_search,
 218                                                                 lc_mc_search->original,
 219                                                                 lc_mc_search->original_len,
 220                                                                 cp_source));
 221     }
 222 #else
 223     g_ptr_array_add (ret,
 224                      (gpointer) mc_search__cond_struct_new (lc_mc_search, lc_mc_search->original,
 225                                                             lc_mc_search->original_len,
 226                                                             str_detect_termencoding ()));
 227 #endif
 228     lc_mc_search->conditions = ret;
 229
 230     return (lc_mc_search->error == MC_SEARCH_E_OK);
 231 }
 232
 233 /* --------------------------------------------------------------------------------------------- */
 234
 235 gboolean
 236 mc_search_run (mc_search_t * lc_mc_search, const void *user_data,
 237                gsize start_search, gsize end_search, gsize * found_len)
 238 {
 239     gboolean ret = FALSE;
 240
 241     if (lc_mc_search == NULL || user_data == NULL)
 242         return FALSE;
 243     if (!mc_search_is_type_avail (lc_mc_search->search_type))
 244     {
 245         lc_mc_search->error = MC_SEARCH_E_INPUT;
 246         lc_mc_search->error_str = g_strdup (_(STR_E_UNKNOWN_TYPE));
 247         return FALSE;
 248     }
 249 #ifdef SEARCH_TYPE_GLIB
 250     if (lc_mc_search->regex_match_info)
 251     {
 252         g_match_info_free (lc_mc_search->regex_match_info);
 253         lc_mc_search->regex_match_info = NULL;
 254     }
 255 #endif /* SEARCH_TYPE_GLIB */
 256
 257     lc_mc_search->error = MC_SEARCH_E_OK;
 258     g_free (lc_mc_search->error_str);
 259     lc_mc_search->error_str = NULL;
 260
 261     if ((lc_mc_search->conditions == NULL) && !mc_search_prepare (lc_mc_search))
 262         return FALSE;
 263
 264     switch (lc_mc_search->search_type)
 265     {
 266     case MC_SEARCH_T_NORMAL:
 267         ret = mc_search__run_normal (lc_mc_search, user_data, start_search, end_search, found_len);
 268         break;
 269     case MC_SEARCH_T_REGEX:
 270         ret = mc_search__run_regex (lc_mc_search, user_data, start_search, end_search, found_len);
 271         break;
 272     case MC_SEARCH_T_GLOB:
 273         ret = mc_search__run_glob (lc_mc_search, user_data, start_search, end_search, found_len);
 274         break;
 275     case MC_SEARCH_T_HEX:
 276         ret = mc_search__run_hex (lc_mc_search, user_data, start_search, end_search, found_len);
 277         break;
 278     default:
 279         break;
 280     }
 281     return ret;
 282 }
 283
 284 /* --------------------------------------------------------------------------------------------- */
 285
 286 gboolean
 287 mc_search_is_type_avail (mc_search_type_t search_type)
 288 {
 289     switch (search_type)
 290     {
 291     case MC_SEARCH_T_GLOB:
 292     case MC_SEARCH_T_NORMAL:
 293     case MC_SEARCH_T_REGEX:
 294     case MC_SEARCH_T_HEX:
 295         return TRUE;
 296     default:
 297         break;
 298     }
 299     return FALSE;
 300 }
 301
 302 /* --------------------------------------------------------------------------------------------- */
 303
 304 const mc_search_type_str_t *
 305 mc_search_types_list_get (size_t * num)
 306 {
 307     /* don't count last NULL item */
 308     if (num != NULL)
 309         *num = sizeof (mc_search__list_types) / sizeof (mc_search__list_types[0]) - 1;
 310
 311     return mc_search__list_types;
 312 }
 313
 314 /* --------------------------------------------------------------------------------------------- */
 315
 316 GString *
 317 mc_search_prepare_replace_str (mc_search_t * lc_mc_search, GString * replace_str)
 318 {
 319     GString *ret;
 320
 321     if (lc_mc_search == NULL)
 322         return g_string_new_len (replace_str->str, replace_str->len);
 323
 324     if (replace_str == NULL || replace_str->str == NULL || replace_str->len == 0)
 325         return g_string_new ("");
 326
 327     switch (lc_mc_search->search_type)
 328     {
 329     case MC_SEARCH_T_REGEX:
 330         ret = mc_search_regex_prepare_replace_str (lc_mc_search, replace_str);
 331         break;
 332     case MC_SEARCH_T_GLOB:
 333         ret = mc_search_glob_prepare_replace_str (lc_mc_search, replace_str);
 334         break;
 335     case MC_SEARCH_T_NORMAL:
 336         ret = mc_search_normal_prepare_replace_str (lc_mc_search, replace_str);
 337         break;
 338     case MC_SEARCH_T_HEX:
 339         ret = mc_search_hex_prepare_replace_str (lc_mc_search, replace_str);
 340         break;
 341     default:
 342         ret = g_string_new_len (replace_str->str, replace_str->len);
 343         break;
 344     }
 345     return ret;
 346 }
 347
 348 /* --------------------------------------------------------------------------------------------- */
 349
 350 char *
 351 mc_search_prepare_replace_str2 (mc_search_t * lc_mc_search, char *replace_str)
 352 {
 353     GString *ret;
 354     GString *replace_str2;
 355
 356     replace_str2 = g_string_new (replace_str);
 357     ret = mc_search_prepare_replace_str (lc_mc_search, replace_str2);
 358     g_string_free (replace_str2, TRUE);
 359     return (ret != NULL) ? g_string_free (ret, FALSE) : NULL;
 360 }
 361
 362 /* --------------------------------------------------------------------------------------------- */
 363
 364 gboolean
 365 mc_search_is_fixed_search_str (mc_search_t * lc_mc_search)
 366 {
 367     if (lc_mc_search == NULL)
 368         return FALSE;
 369     switch (lc_mc_search->search_type)
 370     {
 371     case MC_SEARCH_T_REGEX:
 372     case MC_SEARCH_T_GLOB:
 373         return FALSE;
 374     default:
 375         return TRUE;
 376     }
 377 }
 378
 379 /* --------------------------------------------------------------------------------------------- */
 380
 381 gboolean
 382 mc_search (const gchar * pattern, const gchar * str, mc_search_type_t type)
 383 {
 384     gboolean ret;
 385     mc_search_t *search;
 386
 387     if (str == NULL)
 388         return FALSE;
 389
 390     search = mc_search_new (pattern, -1);
 391     if (search == NULL)
 392         return FALSE;
 393
 394     search->search_type = type;
 395     search->is_case_sensitive = TRUE;
 396
 397     if (type == MC_SEARCH_T_GLOB)
 398         search->is_entire_line = TRUE;
 399
 400     ret = mc_search_run (search, str, 0, strlen (str), NULL);
 401     mc_search_free (search);
 402     return ret;
 403 }
 404
 405 /* --------------------------------------------------------------------------------------------- */
 406
 407 int
 408 mc_search_getstart_result_by_num (mc_search_t * lc_mc_search, int lc_index)
 409 {
 410     if (!lc_mc_search)
 411         return 0;
 412     if (lc_mc_search->search_type == MC_SEARCH_T_NORMAL)
 413         return 0;
 414 #ifdef SEARCH_TYPE_GLIB
 415     {
 416         gint start_pos;
 417         gint end_pos;
 418         g_match_info_fetch_pos (lc_mc_search->regex_match_info, lc_index, &start_pos, &end_pos);
 419         return (int) start_pos;
 420     }
 421 #else /* SEARCH_TYPE_GLIB */
 422     return lc_mc_search->iovector[lc_index * 2];
 423 #endif /* SEARCH_TYPE_GLIB */
 424 }
 425
 426 /* --------------------------------------------------------------------------------------------- */
 427
 428 int
 429 mc_search_getend_result_by_num (mc_search_t * lc_mc_search, int lc_index)
 430 {
 431     if (!lc_mc_search)
 432         return 0;
 433     if (lc_mc_search->search_type == MC_SEARCH_T_NORMAL)
 434         return 0;
 435 #ifdef SEARCH_TYPE_GLIB
 436     {
 437         gint start_pos;
 438         gint end_pos;
 439         g_match_info_fetch_pos (lc_mc_search->regex_match_info, lc_index, &start_pos, &end_pos);
 440         return (int) end_pos;
 441     }
 442 #else /* SEARCH_TYPE_GLIB */
 443     return lc_mc_search->iovector[lc_index * 2 + 1];
 444 #endif /* SEARCH_TYPE_GLIB */
 445 }
 446
 447 /* --------------------------------------------------------------------------------------------- */