5 Copyright (C) 2009-2016
6 Free Software Foundation, Inc.
9 Slava Zanko <slavazanko@gmail.com>, 2009, 2010, 2011, 2013
10 Vitaliy Filippov <vitalif@yourcmc.ru>, 2011
11 Andrew Borodin <aborodin@vmail.ru>, 2013-2015
13 This file is part of the Midnight Commander.
15 The Midnight Commander is free software: you can redistribute it
16 and/or modify it under the terms of the GNU General Public License as
17 published by the Free Software Foundation, either version 3 of the License,
18 or (at your option) any later version.
20 The Midnight Commander is distributed in the hope that it will be useful,
21 but WITHOUT ANY WARRANTY; without even the implied warranty of
22 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
23 GNU General Public License for more details.
25 You should have received a copy of the GNU General Public License
26 along with this program. If not, see <http://www.gnu.org/licenses/>.
33 #include "lib/global.h"
34 #include "lib/strutil.h"
35 #include "lib/search.h"
36 #include "lib/strescape.h"
40 /*** global variables ****************************************************************************/
42 /*** file scope macro definitions ****************************************************************/
44 #define REPLACE_PREPARE_T_NOTHING_SPECIAL -1
45 #define REPLACE_PREPARE_T_REPLACE_FLAG -2
46 #define REPLACE_PREPARE_T_ESCAPE_SEQ -3
48 /*** file scope type declarations ****************************************************************/
52 REPLACE_T_NO_TRANSFORM
= 0,
53 REPLACE_T_UPP_TRANSFORM_CHAR
= 1,
54 REPLACE_T_LOW_TRANSFORM_CHAR
= 2,
55 REPLACE_T_UPP_TRANSFORM
= 4,
56 REPLACE_T_LOW_TRANSFORM
= 8
57 } replace_transform_type_t
;
60 /*** file scope variables ************************************************************************/
62 /*** file scope functions ************************************************************************/
64 #ifndef SEARCH_TYPE_GLIB
66 mc_search__regex_str_append_if_special (GString
* copy_to
, const GString
* regex_str
,
69 const char *special_chars
[] = {
86 const char **spec_chr
;
88 tmp_regex_str
= &(regex_str
->str
[*offset
]);
90 for (spec_chr
= special_chars
; *spec_chr
!= NULL
; spec_chr
++)
94 spec_chr_len
= strlen (*spec_chr
);
96 if (strncmp (tmp_regex_str
, *spec_chr
, spec_chr_len
) == 0
97 && !strutils_is_char_escaped (regex_str
->str
, tmp_regex_str
))
99 if (strncmp ("\\x", *spec_chr
, spec_chr_len
) == 0)
101 if (tmp_regex_str
[spec_chr_len
] != '{')
105 while ((spec_chr_len
< regex_str
->len
- *offset
)
106 && tmp_regex_str
[spec_chr_len
] != '}')
108 if (tmp_regex_str
[spec_chr_len
] == '}')
112 g_string_append_len (copy_to
, tmp_regex_str
, spec_chr_len
);
113 *offset
+= spec_chr_len
;
121 /* --------------------------------------------------------------------------------------------- */
124 mc_search__cond_struct_new_regex_hex_add (const char *charset
, GString
* str_to
,
125 const char *one_char
, gsize str_len
)
130 upp
= mc_search__toupper_case_str (charset
, one_char
, str_len
);
131 low
= mc_search__tolower_case_str (charset
, one_char
, str_len
);
133 for (loop
= 0; loop
< upp
->len
; loop
++)
135 gchar tmp_str
[10 + 1]; /* longest content is "[\\x%02X\\x%02X]" */
138 if (loop
>= low
->len
|| upp
->str
[loop
] == low
->str
[loop
])
140 g_snprintf (tmp_str
, sizeof (tmp_str
), "\\x%02X", (unsigned char) upp
->str
[loop
]);
143 g_snprintf (tmp_str
, sizeof (tmp_str
), "[\\x%02X\\x%02X]",
144 (unsigned char) upp
->str
[loop
], (unsigned char) low
->str
[loop
]);
146 g_string_append_len (str_to
, tmp_str
, tmp_len
);
149 g_string_free (upp
, TRUE
);
150 g_string_free (low
, TRUE
);
153 /* --------------------------------------------------------------------------------------------- */
156 mc_search__cond_struct_new_regex_accum_append (const char *charset
, GString
* str_to
,
159 GString
*recoded_part
;
162 recoded_part
= g_string_sized_new (32);
164 while (loop
< str_from
->len
)
168 gboolean just_letters
;
171 mc_search__get_one_symbol (charset
, &(str_from
->str
[loop
]),
172 min (str_from
->len
- loop
, 6), &just_letters
);
173 one_char_len
= strlen (one_char
);
175 if (one_char_len
== 0)
179 loop
+= one_char_len
;
182 mc_search__cond_struct_new_regex_hex_add (charset
, recoded_part
, one_char
,
185 g_string_append_len (recoded_part
, one_char
, one_char_len
);
191 g_string_append_len (str_to
, recoded_part
->str
, recoded_part
->len
);
192 g_string_free (recoded_part
, TRUE
);
193 g_string_set_size (str_from
, 0);
196 /* --------------------------------------------------------------------------------------------- */
199 mc_search__cond_struct_new_regex_ci_str (const char *charset
, const GString
* astr
)
201 GString
*accumulator
, *spec_char
, *ret_str
;
204 ret_str
= g_string_sized_new (64);
205 accumulator
= g_string_sized_new (64);
206 spec_char
= g_string_sized_new (64);
209 while (loop
<= astr
->len
)
211 if (mc_search__regex_str_append_if_special (spec_char
, astr
, &loop
))
213 mc_search__cond_struct_new_regex_accum_append (charset
, ret_str
, accumulator
);
214 g_string_append_len (ret_str
, spec_char
->str
, spec_char
->len
);
215 g_string_set_size (spec_char
, 0);
219 if (astr
->str
[loop
] == '[' && !strutils_is_char_escaped (astr
->str
, &(astr
->str
[loop
])))
221 mc_search__cond_struct_new_regex_accum_append (charset
, ret_str
, accumulator
);
223 while (loop
< astr
->len
&& !(astr
->str
[loop
] == ']'
224 && !strutils_is_char_escaped (astr
->str
,
225 &(astr
->str
[loop
]))))
227 g_string_append_c (ret_str
, astr
->str
[loop
]);
231 g_string_append_c (ret_str
, astr
->str
[loop
]);
238 g_string_append_c (accumulator
, astr
->str
[loop
]);
241 mc_search__cond_struct_new_regex_accum_append (charset
, ret_str
, accumulator
);
243 g_string_free (accumulator
, TRUE
);
244 g_string_free (spec_char
, TRUE
);
248 #endif /* !SEARCH_TYPE_GLIB */
250 /* --------------------------------------------------------------------------------------------- */
252 #ifdef SEARCH_TYPE_GLIB
253 /* A thin wrapper above g_regex_match_full that makes sure the string passed
254 * to it is valid UTF-8 (unless G_REGEX_RAW compile flag was set), as it is a
255 * requirement by glib and it might crash otherwise. See: mc ticket 3449.
256 * Be careful: there might be embedded NULs in the strings. */
258 mc_search__g_regex_match_full_safe (const GRegex
* regex
,
259 const gchar
* string
,
262 GRegexMatchFlags match_options
,
263 GMatchInfo
** match_info
, GError
** error
)
265 char *string_safe
, *p
, *end
;
268 if ((g_regex_get_compile_flags (regex
) & G_REGEX_RAW
)
269 || g_utf8_validate (string
, string_len
, NULL
))
271 return g_regex_match_full (regex
, string
, string_len
, start_position
, match_options
,
277 string_len
= strlen (string
);
280 /* Correctly handle embedded NULs while copying */
281 p
= string_safe
= g_malloc (string_len
);
282 memcpy (string_safe
, string
, string_len
);
283 end
= p
+ string_len
;
287 gunichar c
= g_utf8_get_char_validated (p
, -1);
288 if (c
!= (gunichar
) (-1) && c
!= (gunichar
) (-2))
290 p
= g_utf8_next_char (p
);
294 /* U+FFFD would be the proper choice, but then we'd have to
295 maintain mapping between old and new offsets.
296 So rather do a byte by byte replacement. */
302 g_regex_match_full (regex
, string_safe
, string_len
, start_position
, match_options
,
304 g_free (string_safe
);
307 #endif /* SEARCH_TYPE_GLIB */
309 /* --------------------------------------------------------------------------------------------- */
311 static mc_search__found_cond_t
312 mc_search__regex_found_cond_one (mc_search_t
* lc_mc_search
, mc_search_regex_t
* regex
,
313 GString
* search_str
)
315 #ifdef SEARCH_TYPE_GLIB
316 GError
*mcerror
= NULL
;
318 if (!mc_search__g_regex_match_full_safe
319 (regex
, search_str
->str
, search_str
->len
, 0, G_REGEX_MATCH_NEWLINE_ANY
,
320 &lc_mc_search
->regex_match_info
, &mcerror
))
322 g_match_info_free (lc_mc_search
->regex_match_info
);
323 lc_mc_search
->regex_match_info
= NULL
;
326 lc_mc_search
->error
= MC_SEARCH_E_REGEX
;
327 lc_mc_search
->error_str
=
328 str_conv_gerror_message (mcerror
, _("Regular expression error"));
329 g_error_free (mcerror
);
330 return COND__FOUND_ERROR
;
332 return COND__NOT_FOUND
;
334 lc_mc_search
->num_results
= g_match_info_get_match_count (lc_mc_search
->regex_match_info
);
335 #else /* SEARCH_TYPE_GLIB */
336 lc_mc_search
->num_results
= pcre_exec (regex
, lc_mc_search
->regex_match_info
,
337 search_str
->str
, search_str
->len
, 0, 0,
338 lc_mc_search
->iovector
, MC_SEARCH__NUM_REPLACE_ARGS
);
339 if (lc_mc_search
->num_results
< 0)
341 return COND__NOT_FOUND
;
343 #endif /* SEARCH_TYPE_GLIB */
344 return COND__FOUND_OK
;
348 /* --------------------------------------------------------------------------------------------- */
350 static mc_search__found_cond_t
351 mc_search__regex_found_cond (mc_search_t
* lc_mc_search
, GString
* search_str
)
355 for (loop1
= 0; loop1
< lc_mc_search
->conditions
->len
; loop1
++)
357 mc_search_cond_t
*mc_search_cond
;
358 mc_search__found_cond_t ret
;
360 mc_search_cond
= (mc_search_cond_t
*) g_ptr_array_index (lc_mc_search
->conditions
, loop1
);
362 if (!mc_search_cond
->regex_handle
)
366 mc_search__regex_found_cond_one (lc_mc_search
, mc_search_cond
->regex_handle
,
368 if (ret
!= COND__NOT_FOUND
)
371 return COND__NOT_ALL_FOUND
;
374 /* --------------------------------------------------------------------------------------------- */
377 mc_search_regex__get_max_num_of_replace_tokens (const gchar
* str
, gsize len
)
381 for (loop
= 0; loop
< len
- 1; loop
++)
383 if (str
[loop
] == '\\' && g_ascii_isdigit (str
[loop
+ 1]))
385 if (strutils_is_char_escaped (str
, &str
[loop
]))
387 if (max_token
< str
[loop
+ 1] - '0')
388 max_token
= str
[loop
+ 1] - '0';
391 if (str
[loop
] == '$' && str
[loop
+ 1] == '{')
395 if (strutils_is_char_escaped (str
, &str
[loop
]))
399 loop
+ tmp_len
+ 2 < len
&& (str
[loop
+ 2 + tmp_len
] & (char) 0xf0) == 0x30;
402 if (str
[loop
+ 2 + tmp_len
] == '}')
407 tmp_str
= g_strndup (&str
[loop
+ 2], tmp_len
);
408 tmp_token
= atoi (tmp_str
);
409 if (max_token
< tmp_token
)
410 max_token
= tmp_token
;
418 /* --------------------------------------------------------------------------------------------- */
421 mc_search_regex__get_token_by_num (const mc_search_t
* lc_mc_search
, gsize lc_index
)
423 int fnd_start
= 0, fnd_end
= 0;
425 #ifdef SEARCH_TYPE_GLIB
426 g_match_info_fetch_pos (lc_mc_search
->regex_match_info
, lc_index
, &fnd_start
, &fnd_end
);
427 #else /* SEARCH_TYPE_GLIB */
428 fnd_start
= lc_mc_search
->iovector
[lc_index
* 2 + 0];
429 fnd_end
= lc_mc_search
->iovector
[lc_index
* 2 + 1];
430 #endif /* SEARCH_TYPE_GLIB */
432 if (fnd_end
== fnd_start
)
433 return g_strdup ("");
435 return g_strndup (lc_mc_search
->regex_buffer
->str
+ fnd_start
, fnd_end
- fnd_start
);
439 /* --------------------------------------------------------------------------------------------- */
442 mc_search_regex__replace_handle_esc_seq (const GString
* replace_str
, const gsize current_pos
,
443 gsize
* skip_len
, int *ret
)
445 char *curr_str
= &(replace_str
->str
[current_pos
]);
446 char c
= curr_str
[1];
448 if (replace_str
->len
> current_pos
+ 2)
452 for (*skip_len
= 2; /* \{ */
453 current_pos
+ *skip_len
< replace_str
->len
&& curr_str
[*skip_len
] >= '0'
454 && curr_str
[*skip_len
] <= '7'; (*skip_len
)++)
457 if (current_pos
+ *skip_len
< replace_str
->len
&& curr_str
[*skip_len
] == '}')
460 *ret
= REPLACE_PREPARE_T_ESCAPE_SEQ
;
465 *ret
= REPLACE_PREPARE_T_NOTHING_SPECIAL
;
472 *skip_len
= 2; /* \x */
476 for (*skip_len
= 3; /* \x{ */
477 current_pos
+ *skip_len
< replace_str
->len
478 && g_ascii_isxdigit ((guchar
) curr_str
[*skip_len
]); (*skip_len
)++)
481 if (current_pos
+ *skip_len
< replace_str
->len
&& curr_str
[*skip_len
] == '}')
484 *ret
= REPLACE_PREPARE_T_ESCAPE_SEQ
;
489 *ret
= REPLACE_PREPARE_T_NOTHING_SPECIAL
;
493 else if (!g_ascii_isxdigit ((guchar
) c
))
495 *skip_len
= 2; /* \x without number behind */
496 *ret
= REPLACE_PREPARE_T_NOTHING_SPECIAL
;
502 if (!g_ascii_isxdigit ((guchar
) c
))
503 *skip_len
= 3; /* \xH */
505 *skip_len
= 4; /* \xHH */
506 *ret
= REPLACE_PREPARE_T_ESCAPE_SEQ
;
512 if (strchr ("ntvbrfa", c
) != NULL
)
515 *ret
= REPLACE_PREPARE_T_ESCAPE_SEQ
;
521 /* --------------------------------------------------------------------------------------------- */
524 mc_search_regex__process_replace_str (const GString
* replace_str
, const gsize current_pos
,
525 gsize
* skip_len
, replace_transform_type_t
* replace_flags
)
528 const char *curr_str
= &(replace_str
->str
[current_pos
]);
530 if (current_pos
> replace_str
->len
)
531 return REPLACE_PREPARE_T_NOTHING_SPECIAL
;
535 if (replace_str
->len
> current_pos
+ 2 && curr_str
[0] == '$' && curr_str
[1] == '{'
536 && (curr_str
[2] & (char) 0xf0) == 0x30)
540 if (strutils_is_char_escaped (replace_str
->str
, curr_str
))
543 return REPLACE_PREPARE_T_NOTHING_SPECIAL
;
547 current_pos
+ *skip_len
+ 2 < replace_str
->len
548 && (curr_str
[2 + *skip_len
] & (char) 0xf0) == 0x30; (*skip_len
)++)
551 if (curr_str
[2 + *skip_len
] != '}')
552 return REPLACE_PREPARE_T_NOTHING_SPECIAL
;
554 tmp_str
= g_strndup (curr_str
+ 2, *skip_len
);
556 return REPLACE_PREPARE_T_NOTHING_SPECIAL
;
558 ret
= atoi (tmp_str
);
561 *skip_len
+= 3; /* ${} */
562 return ret
; /* capture buffer index >= 0 */
565 if (curr_str
[0] == '\\' && replace_str
->len
> current_pos
+ 1)
567 if (strutils_is_char_escaped (replace_str
->str
, curr_str
))
570 return REPLACE_PREPARE_T_NOTHING_SPECIAL
;
573 if (g_ascii_isdigit (curr_str
[1]))
575 ret
= g_ascii_digit_value (curr_str
[1]); /* capture buffer index >= 0 */
576 *skip_len
= 2; /* \\ and one digit */
580 if (!mc_search_regex__replace_handle_esc_seq (replace_str
, current_pos
, skip_len
, &ret
))
583 ret
= REPLACE_PREPARE_T_REPLACE_FLAG
;
589 *replace_flags
|= REPLACE_T_UPP_TRANSFORM
;
590 *replace_flags
&= ~REPLACE_T_LOW_TRANSFORM
;
593 *replace_flags
|= REPLACE_T_UPP_TRANSFORM_CHAR
;
596 *replace_flags
|= REPLACE_T_LOW_TRANSFORM
;
597 *replace_flags
&= ~REPLACE_T_UPP_TRANSFORM
;
600 *replace_flags
|= REPLACE_T_LOW_TRANSFORM_CHAR
;
603 *replace_flags
= REPLACE_T_NO_TRANSFORM
;
606 ret
= REPLACE_PREPARE_T_NOTHING_SPECIAL
;
613 /* --------------------------------------------------------------------------------------------- */
616 mc_search_regex__process_append_str (GString
* dest_str
, const char *from
, gsize len
,
617 replace_transform_type_t
* replace_flags
)
622 if (len
== (gsize
) (-1))
625 if (*replace_flags
== REPLACE_T_NO_TRANSFORM
)
627 g_string_append_len (dest_str
, from
, len
);
631 for (loop
= 0; loop
< len
; loop
+= char_len
)
633 GString
*tmp_string
= NULL
;
636 tmp_str
= mc_search__get_one_symbol (NULL
, from
+ loop
, len
- loop
, NULL
);
637 char_len
= strlen (tmp_str
);
639 if ((*replace_flags
& REPLACE_T_UPP_TRANSFORM_CHAR
) != 0)
641 *replace_flags
&= ~REPLACE_T_UPP_TRANSFORM_CHAR
;
642 tmp_string
= mc_search__toupper_case_str (NULL
, tmp_str
, char_len
);
643 g_string_append_len (dest_str
, tmp_string
->str
, tmp_string
->len
);
644 g_string_free (tmp_string
, TRUE
);
646 else if ((*replace_flags
& REPLACE_T_LOW_TRANSFORM_CHAR
) != 0)
648 *replace_flags
&= ~REPLACE_T_LOW_TRANSFORM_CHAR
;
649 tmp_string
= mc_search__tolower_case_str (NULL
, tmp_str
, char_len
);
650 g_string_append_len (dest_str
, tmp_string
->str
, tmp_string
->len
);
651 g_string_free (tmp_string
, TRUE
);
653 else if ((*replace_flags
& REPLACE_T_UPP_TRANSFORM
) != 0)
655 tmp_string
= mc_search__toupper_case_str (NULL
, tmp_str
, char_len
);
656 g_string_append_len (dest_str
, tmp_string
->str
, tmp_string
->len
);
657 g_string_free (tmp_string
, TRUE
);
659 else if ((*replace_flags
& REPLACE_T_LOW_TRANSFORM
) != 0)
661 tmp_string
= mc_search__tolower_case_str (NULL
, tmp_str
, char_len
);
662 g_string_append_len (dest_str
, tmp_string
->str
, tmp_string
->len
);
663 g_string_free (tmp_string
, TRUE
);
670 /* --------------------------------------------------------------------------------------------- */
673 mc_search_regex__process_escape_sequence (GString
* dest_str
, const char *from
, gsize len
,
674 replace_transform_type_t
* replace_flags
,
681 if (len
== (gsize
) (-1))
694 if (i
< len
&& from
[i
] == '{')
698 if (from
[i
] >= '0' && from
[i
] <= '9')
699 c
= c
* 16 + from
[i
] - '0';
700 else if (from
[i
] >= 'a' && from
[i
] <= 'f')
701 c
= c
* 16 + 10 + from
[i
] - 'a';
702 else if (from
[i
] >= 'A' && from
[i
] <= 'F')
703 c
= c
* 16 + 10 + from
[i
] - 'A';
708 else if (from
[i
] >= '0' && from
[i
] <= '7')
709 for (; i
< len
&& from
[i
] >= '0' && from
[i
] <= '7'; i
++)
710 c
= c
* 8 + from
[i
] - '0';
737 mc_search_regex__process_append_str (dest_str
, from
, len
, replace_flags
);
742 if (c
< 0x80 || !is_utf8
)
743 g_string_append_c (dest_str
, (char) c
);
747 g_string_append_c (dest_str
, b
);
748 b
= 0x80 | (c
& 0x3F);
749 g_string_append_c (dest_str
, b
);
751 else if (c
< 0x10000)
753 b
= 0xE0 | (c
>> 12);
754 g_string_append_c (dest_str
, b
);
755 b
= 0x80 | ((c
>> 6) & 0x3F);
756 g_string_append_c (dest_str
, b
);
757 b
= 0x80 | (c
& 0x3F);
758 g_string_append_c (dest_str
, b
);
760 else if (c
< 0x10FFFF)
762 b
= 0xF0 | (c
>> 16);
763 g_string_append_c (dest_str
, b
);
764 b
= 0x80 | ((c
>> 12) & 0x3F);
765 g_string_append_c (dest_str
, b
);
766 b
= 0x80 | ((c
>> 6) & 0x3F);
767 g_string_append_c (dest_str
, b
);
768 b
= 0x80 | (c
& 0x3F);
769 g_string_append_c (dest_str
, b
);
773 /* --------------------------------------------------------------------------------------------- */
775 * Get regex flags for compilation of expressions.
776 * @param charset the charset
778 * @return regex flags
781 static GRegexCompileFlags
782 mc_search__regex_get_compile_flags (const char *charset
, gboolean is_case_sensitive
)
784 GRegexCompileFlags g_regex_options
= G_REGEX_OPTIMIZE
| G_REGEX_DOTALL
;
786 if (!(mc_global
.utf8_display
&& str_isutf8 (charset
)))
787 g_regex_options
|= G_REGEX_RAW
;
789 if (!is_case_sensitive
)
790 g_regex_options
|= G_REGEX_CASELESS
;
792 return g_regex_options
;
795 /* --------------------------------------------------------------------------------------------- */
796 /*** public functions ****************************************************************************/
797 /* --------------------------------------------------------------------------------------------- */
800 mc_search__cond_struct_new_init_regex (const char *charset
, mc_search_t
* lc_mc_search
,
801 mc_search_cond_t
* mc_search_cond
)
803 if (lc_mc_search
->whole_words
&& !lc_mc_search
->is_entire_line
)
805 /* NOTE: \b as word boundary doesn't allow search
806 * whole words with non-ASCII symbols.
807 * Update: Is it still true nowadays? Probably not. #2396, #3524 */
808 g_string_prepend (mc_search_cond
->str
, "(?<![\\p{L}\\p{N}_])");
809 g_string_append (mc_search_cond
->str
, "(?![\\p{L}\\p{N}_])");
813 #ifdef SEARCH_TYPE_GLIB
814 GError
*mcerror
= NULL
;
816 mc_search_cond
->regex_handle
=
817 g_regex_new (mc_search_cond
->str
->str
,
818 mc_search__regex_get_compile_flags (charset
,
819 lc_mc_search
->is_case_sensitive
), 0,
824 lc_mc_search
->error
= MC_SEARCH_E_REGEX_COMPILE
;
825 lc_mc_search
->error_str
=
826 str_conv_gerror_message (mcerror
, _("Regular expression error"));
827 g_error_free (mcerror
);
830 #else /* SEARCH_TYPE_GLIB */
833 int pcre_options
= PCRE_EXTRA
| PCRE_MULTILINE
;
835 if (str_isutf8 (charset
) && mc_global
.utf8_display
)
837 pcre_options
|= PCRE_UTF8
;
838 if (!lc_mc_search
->is_case_sensitive
)
839 pcre_options
|= PCRE_CASELESS
;
843 if (!lc_mc_search
->is_case_sensitive
)
847 tmp
= mc_search_cond
->str
;
848 mc_search_cond
->str
= mc_search__cond_struct_new_regex_ci_str (charset
, tmp
);
849 g_string_free (tmp
, TRUE
);
853 mc_search_cond
->regex_handle
=
854 pcre_compile (mc_search_cond
->str
->str
, pcre_options
, &error
, &erroffset
, NULL
);
855 if (mc_search_cond
->regex_handle
== NULL
)
857 lc_mc_search
->error
= MC_SEARCH_E_REGEX_COMPILE
;
858 lc_mc_search
->error_str
= g_strdup (error
);
861 lc_mc_search
->regex_match_info
= pcre_study (mc_search_cond
->regex_handle
, 0, &error
);
862 if (lc_mc_search
->regex_match_info
== NULL
)
866 lc_mc_search
->error
= MC_SEARCH_E_REGEX_COMPILE
;
867 lc_mc_search
->error_str
= g_strdup (error
);
868 g_free (mc_search_cond
->regex_handle
);
869 mc_search_cond
->regex_handle
= NULL
;
873 #endif /* SEARCH_TYPE_GLIB */
876 lc_mc_search
->is_utf8
= str_isutf8 (charset
);
879 /* --------------------------------------------------------------------------------------------- */
882 mc_search__run_regex (mc_search_t
* lc_mc_search
, const void *user_data
,
883 gsize start_search
, gsize end_search
, gsize
* found_len
)
885 mc_search_cbret_t ret
= MC_SEARCH_CB_ABORT
;
886 gsize current_pos
, virtual_pos
;
890 if (lc_mc_search
->regex_buffer
!= NULL
)
891 g_string_free (lc_mc_search
->regex_buffer
, TRUE
);
893 lc_mc_search
->regex_buffer
= g_string_sized_new (64);
895 virtual_pos
= current_pos
= start_search
;
896 while (virtual_pos
<= end_search
)
898 g_string_set_size (lc_mc_search
->regex_buffer
, 0);
899 lc_mc_search
->start_buffer
= current_pos
;
901 if (lc_mc_search
->search_fn
!= NULL
)
905 int current_chr
= '\n'; /* stop search symbol */
907 ret
= lc_mc_search
->search_fn (user_data
, current_pos
, ¤t_chr
);
909 if (ret
== MC_SEARCH_CB_ABORT
)
912 if (ret
== MC_SEARCH_CB_INVALID
)
917 if (ret
== MC_SEARCH_CB_SKIP
)
922 g_string_append_c (lc_mc_search
->regex_buffer
, (char) current_chr
);
924 if ((char) current_chr
== '\n' || virtual_pos
> end_search
)
930 /* optimization for standard case (for search from file manager)
931 * where there is no MC_SEARCH_CB_INVALID or MC_SEARCH_CB_SKIP
932 * return codes, so we can copy line at regex buffer all at once
938 current_chr
= ((char *) user_data
)[current_pos
];
939 if (current_chr
== '\0')
944 if (current_chr
== '\n' || current_pos
> end_search
)
948 /* use virtual_pos as index of start of current chunk */
949 g_string_append_len (lc_mc_search
->regex_buffer
, (char *) user_data
+ virtual_pos
,
950 current_pos
- virtual_pos
);
951 virtual_pos
= current_pos
;
954 switch (mc_search__regex_found_cond (lc_mc_search
, lc_mc_search
->regex_buffer
))
957 #ifdef SEARCH_TYPE_GLIB
958 g_match_info_fetch_pos (lc_mc_search
->regex_match_info
, 0, &start_pos
, &end_pos
);
959 #else /* SEARCH_TYPE_GLIB */
960 start_pos
= lc_mc_search
->iovector
[0];
961 end_pos
= lc_mc_search
->iovector
[1];
962 #endif /* SEARCH_TYPE_GLIB */
963 if (found_len
!= NULL
)
964 *found_len
= end_pos
- start_pos
;
965 lc_mc_search
->normal_offset
= lc_mc_search
->start_buffer
+ start_pos
;
967 case COND__NOT_ALL_FOUND
:
970 g_string_free (lc_mc_search
->regex_buffer
, TRUE
);
971 lc_mc_search
->regex_buffer
= NULL
;
975 if ((lc_mc_search
->update_fn
!= NULL
) &&
976 ((lc_mc_search
->update_fn
) (user_data
, current_pos
) == MC_SEARCH_CB_ABORT
))
977 ret
= MC_SEARCH_CB_ABORT
;
979 if (ret
== MC_SEARCH_CB_ABORT
)
983 g_string_free (lc_mc_search
->regex_buffer
, TRUE
);
984 lc_mc_search
->regex_buffer
= NULL
;
985 lc_mc_search
->error
= MC_SEARCH_E_NOTFOUND
;
987 if (ret
!= MC_SEARCH_CB_ABORT
)
988 lc_mc_search
->error_str
= g_strdup (_(STR_E_NOTFOUND
));
990 lc_mc_search
->error_str
= NULL
;
995 /* --------------------------------------------------------------------------------------------- */
998 mc_search_regex_prepare_replace_str (mc_search_t
* lc_mc_search
, GString
* replace_str
)
1002 int num_replace_tokens
;
1005 replace_transform_type_t replace_flags
= REPLACE_T_NO_TRANSFORM
;
1007 num_replace_tokens
=
1008 mc_search_regex__get_max_num_of_replace_tokens (replace_str
->str
, replace_str
->len
);
1010 if (lc_mc_search
->num_results
< 0)
1011 return g_string_new_len (replace_str
->str
, replace_str
->len
);
1013 if (num_replace_tokens
> lc_mc_search
->num_results
- 1
1014 || num_replace_tokens
> MC_SEARCH__NUM_REPLACE_ARGS
)
1016 lc_mc_search
->error
= MC_SEARCH_E_REGEX_REPLACE
;
1017 lc_mc_search
->error_str
= g_strdup (_(STR_E_RPL_NOT_EQ_TO_FOUND
));
1021 ret
= g_string_sized_new (64);
1023 for (loop
= 0; loop
< replace_str
->len
- 1; loop
++)
1029 lc_index
= mc_search_regex__process_replace_str (replace_str
, loop
, &len
, &replace_flags
);
1031 if (lc_index
== REPLACE_PREPARE_T_NOTHING_SPECIAL
)
1035 mc_search_regex__process_append_str (ret
, replace_str
->str
+ prev
, loop
- prev
,
1037 mc_search_regex__process_append_str (ret
, replace_str
->str
+ loop
+ 1, len
- 1,
1040 loop
= prev
- 1; /* prepare to loop++ */
1046 if (lc_index
== REPLACE_PREPARE_T_REPLACE_FLAG
)
1049 mc_search_regex__process_append_str (ret
, replace_str
->str
+ prev
, loop
- prev
,
1052 loop
= prev
- 1; /* prepare to loop++ */
1056 /* escape sequence */
1057 if (lc_index
== REPLACE_PREPARE_T_ESCAPE_SEQ
)
1059 mc_search_regex__process_append_str (ret
, replace_str
->str
+ prev
, loop
- prev
,
1061 /* call process_escape_sequence without starting '\\' */
1062 mc_search_regex__process_escape_sequence (ret
, replace_str
->str
+ loop
+ 1, len
- 1,
1063 &replace_flags
, lc_mc_search
->is_utf8
);
1065 loop
= prev
- 1; /* prepare to loop++ */
1069 /* invalid capture buffer number */
1070 if (lc_index
> lc_mc_search
->num_results
)
1072 g_string_free (ret
, TRUE
);
1073 lc_mc_search
->error
= MC_SEARCH_E_REGEX_REPLACE
;
1074 lc_mc_search
->error_str
= g_strdup_printf (_(STR_E_RPL_INVALID_TOKEN
), lc_index
);
1078 tmp_str
= mc_search_regex__get_token_by_num (lc_mc_search
, lc_index
);
1081 mc_search_regex__process_append_str (ret
, replace_str
->str
+ prev
, loop
- prev
,
1084 mc_search_regex__process_append_str (ret
, tmp_str
, -1, &replace_flags
);
1088 loop
= prev
- 1; /* prepare to loop++ */
1091 mc_search_regex__process_append_str (ret
, replace_str
->str
+ prev
, replace_str
->len
- prev
,