1 /*
2  * Claws Mail -- a GTK+ based, lightweight, and fast e-mail client
3  * Copyright (C) 2012-2014 the Claws Mail team
4  *
5  * This program is free software; you can redistribute it and/or modify
6  * it under the terms of the GNU General Public License as published by
7  * the Free Software Foundation; either version 3 of the License, or
8  * (at your option) any later version.
9  *
10  * This program is distributed in the hope that it will be useful,
11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
13  * GNU General Public License for more details.
14  *
15  * You should have received a copy of the GNU General Public License
16  * along with this program. If not, see <http://www.gnu.org/licenses/>.
17  *
18  */
19 
20 #ifdef HAVE_CONFIG_H
21 # include "config.h"
22 # include "claws-features.h"
23 #endif
24 
25 #include "advsearch.h"
26 
27 #include <glib.h>
28 #include <ctype.h>
29 
30 #include "matcher.h"
31 #include "matcher_parser.h"
32 #include "utils.h"
33 #include "prefs_common.h"
34 #include "timing.h"
35 
36 struct _AdvancedSearch {
37 	struct {
38 		AdvancedSearchType	 type;
39 		gchar			*matchstring;
40 	} request;
41 
42 	MatcherList			*predicate;
43 	gboolean			 is_fast;
44 	gboolean			 search_aborted;
45 
46 	struct {
47 		gboolean (*cb)(gpointer data, guint at, guint matched, guint total);
48 		gpointer data;
49 	} on_progress_cb;
50 	struct {
51 		void (*cb)(gpointer data);
52 		gpointer data;
53 	} on_error_cb;
54 };
55 
advsearch_set_on_progress_cb(AdvancedSearch * search,gboolean (* cb)(gpointer,guint,guint,guint),gpointer data)56 void advsearch_set_on_progress_cb(AdvancedSearch *search, gboolean (*cb)(gpointer, guint, guint, guint), gpointer data)
57 {
58 	search->on_progress_cb.cb = cb;
59 	search->on_progress_cb.data = data;
60 }
61 
advsearch_set_on_error_cb(AdvancedSearch * search,void (* cb)(gpointer data),gpointer data)62 void advsearch_set_on_error_cb(AdvancedSearch* search, void (*cb)(gpointer data), gpointer data)
63 {
64 	search->on_error_cb.cb = cb;
65 	search->on_error_cb.data = data;
66 }
67 
68 static void prepare_matcher(AdvancedSearch *search);
69 static gboolean search_impl(MsgInfoList **messages, AdvancedSearch* search,
70 			    FolderItem* folderItem, gboolean recursive);
71 
72 // --------------------------
73 
advsearch_new()74 AdvancedSearch* advsearch_new()
75 {
76 	AdvancedSearch *result;
77 
78 	result = g_new0(AdvancedSearch, 1);
79 
80 	return result;
81 }
82 
advsearch_free(AdvancedSearch * search)83 void advsearch_free(AdvancedSearch *search)
84 {
85 	if (search->predicate != NULL)
86 		matcherlist_free(search->predicate);
87 
88 	g_free(search->request.matchstring);
89 	g_free(search);
90 }
91 
advsearch_set(AdvancedSearch * search,AdvancedSearchType type,const gchar * matchstring)92 void advsearch_set(AdvancedSearch *search, AdvancedSearchType type, const gchar *matchstring)
93 {
94 	cm_return_if_fail(search != NULL);
95 
96 	search->request.type = type;
97 
98 	g_free(search->request.matchstring);
99 	search->request.matchstring = g_strdup(matchstring);
100 
101 	prepare_matcher(search);
102 }
103 
advsearch_is_fast(AdvancedSearch * search)104 gboolean advsearch_is_fast(AdvancedSearch *search)
105 {
106 	cm_return_val_if_fail(search != NULL, FALSE);
107 
108 	return search->is_fast;
109 }
110 
advsearch_has_proper_predicate(AdvancedSearch * search)111 gboolean advsearch_has_proper_predicate(AdvancedSearch *search)
112 {
113 	cm_return_val_if_fail(search != NULL, FALSE);
114 
115 	return search->predicate != NULL;
116 }
117 
advsearch_search_msgs_in_folders(AdvancedSearch * search,MsgInfoList ** messages,FolderItem * folderItem,gboolean recursive)118 gboolean advsearch_search_msgs_in_folders(AdvancedSearch* search, MsgInfoList **messages,
119 				          FolderItem* folderItem, gboolean recursive)
120 {
121 	if (search == NULL || search->predicate == NULL)
122 		return FALSE;
123 
124 	search->search_aborted = FALSE;
125 	return search_impl(messages, search, folderItem, recursive);
126 }
127 
advsearch_abort(AdvancedSearch * search)128 void advsearch_abort(AdvancedSearch *search)
129 {
130 	search->search_aborted = TRUE;
131 }
132 
advsearch_expand_search_string(const gchar * search_string)133 gchar *advsearch_expand_search_string(const gchar *search_string)
134 {
135 	int i = 0;
136 	gchar term_char, save_char;
137 	gchar *cmd_start, *cmd_end;
138 	GString *matcherstr;
139 	gchar *returnstr = NULL;
140 	gchar *copy_str;
141 	gboolean casesens, dontmatch, regex;
142 	/* list of allowed pattern abbreviations */
143 	struct {
144 		gchar		*abbreviated;	/* abbreviation */
145 		gchar		*command;	/* actual matcher command */
146 		gint		numparams;	/* number of params for cmd */
147 		gboolean	qualifier;	/* do we append stringmatch operations */
148 		gboolean	quotes;		/* do we need quotes */
149 	}
150 	cmds[] = {
151 		{ "a",	"all",				0,	FALSE,	FALSE },
152 		{ "ag",	"age_greater",			1,	FALSE,	FALSE },
153 		{ "al",	"age_lower",			1,	FALSE,	FALSE },
154 		{ "agh","age_greater_hours",		1,	FALSE,	FALSE },
155 		{ "alh","age_lower_hours",		1,	FALSE,	FALSE },
156 		{ "b",	"body_part",			1,	TRUE,	TRUE  },
157 		{ "B",	"message",			1,	TRUE,	TRUE  },
158 		{ "c",	"cc",				1,	TRUE,	TRUE  },
159 		{ "C",	"to_or_cc",			1,	TRUE,	TRUE  },
160 		{ "D",	"deleted",			0,	FALSE,	FALSE },
161 		{ "da", "date_after",			1,	FALSE,	TRUE  },
162 		{ "db", "date_before",			1,	FALSE,	TRUE  },
163 		{ "e",	"header \"Sender\"",		1,	TRUE,	TRUE  },
164 		{ "E",	"execute",			1,	FALSE,	TRUE  },
165 		{ "f",	"from",				1,	TRUE,	TRUE  },
166 		{ "F",	"forwarded",			0,	FALSE,	FALSE },
167 		{ "h",	"headers_part",			1,	TRUE,	TRUE  },
168 		{ "H",	"headers_cont",			1,	TRUE,	TRUE  },
169 		{ "ha",	"has_attachments",		0,	FALSE,	FALSE },
170 		{ "i",	"messageid",			1,	TRUE,	TRUE  },
171 		{ "I",	"inreplyto",			1,	TRUE,	TRUE  },
172 		{ "k",	"colorlabel",			1,	FALSE,	FALSE },
173 		{ "L",	"locked",			0,	FALSE,	FALSE },
174 		{ "n",	"newsgroups",			1,	TRUE,	TRUE  },
175 		{ "N",	"new",				0,	FALSE,	FALSE },
176 		{ "O",	"~new",				0,	FALSE,	FALSE },
177 		{ "r",	"replied",			0,	FALSE,	FALSE },
178 		{ "R",	"~unread",			0,	FALSE,	FALSE },
179 		{ "s",	"subject",			1,	TRUE,	TRUE  },
180 		{ "se",	"score_equal",			1,	FALSE,	FALSE },
181 		{ "sg",	"score_greater",		1,	FALSE,	FALSE },
182 		{ "sl",	"score_lower",			1,	FALSE,	FALSE },
183 		{ "Se",	"size_equal",			1,	FALSE,	FALSE },
184 		{ "Sg",	"size_greater",			1,	FALSE,	FALSE },
185 		{ "Ss",	"size_smaller",			1,	FALSE,	FALSE },
186 		{ "t",	"to",				1,	TRUE,	TRUE  },
187 		{ "tg", "tag",				1,	TRUE,	TRUE  },
188 		{ "T",	"marked",			0,	FALSE,	FALSE },
189 		{ "U",	"unread",			0,	FALSE,	FALSE },
190 		{ "x",	"references",			1,	TRUE,	TRUE  },
191 		{ "X",  "test",				1,	FALSE,  FALSE },
192 		{ "y",	"header \"X-Label\"",		1,	TRUE,	TRUE  },
193 		{ "&",	"&",				0,	FALSE,	FALSE },
194 		{ "|",	"|",				0,	FALSE,	FALSE },
195 		{ "p",	"partial",			0,	FALSE, 	FALSE },
196 		{ NULL,	NULL,				0,	FALSE,	FALSE }
197 	};
198 
199 	if (search_string == NULL)
200 		return NULL;
201 
202 	copy_str = g_strdup(search_string);
203 
204 	matcherstr = g_string_sized_new(16);
205 	cmd_start = copy_str;
206 	while (cmd_start && *cmd_start) {
207 		/* skip all white spaces */
208 		while (*cmd_start && isspace((guchar)*cmd_start))
209 			cmd_start++;
210 		cmd_end = cmd_start;
211 
212 		/* extract a command */
213 		while (*cmd_end && !isspace((guchar)*cmd_end))
214 			cmd_end++;
215 
216 		/* save character */
217 		save_char = *cmd_end;
218 		*cmd_end = '\0';
219 
220 		dontmatch = FALSE;
221 		casesens = FALSE;
222 		regex = FALSE;
223 
224 		/* ~ and ! mean logical NOT */
225 		if (*cmd_start == '~' || *cmd_start == '!')
226 		{
227 			dontmatch = TRUE;
228 			cmd_start++;
229 		}
230 		/* % means case sensitive match */
231 		if (*cmd_start == '%')
232 		{
233 			casesens = TRUE;
234 			cmd_start++;
235 		}
236 		/* # means regex match */
237 		if (*cmd_start == '#') {
238 			regex = TRUE;
239 			cmd_start++;
240 		}
241 
242 		/* find matching abbreviation */
243 		for (i = 0; cmds[i].command; i++) {
244 			if (!strcmp(cmd_start, cmds[i].abbreviated)) {
245 				/* restore character */
246 				*cmd_end = save_char;
247 
248 				/* copy command */
249 				if (matcherstr->len > 0) {
250 					g_string_append(matcherstr, " ");
251 				}
252 				if (dontmatch)
253 					g_string_append(matcherstr, "~");
254 				g_string_append(matcherstr, cmds[i].command);
255 				g_string_append(matcherstr, " ");
256 
257 				/* stop if no params required */
258 				if (cmds[i].numparams == 0)
259 					break;
260 
261 				/* extract a parameter, allow quotes */
262 				while (*cmd_end && isspace((guchar)*cmd_end))
263 					cmd_end++;
264 
265 				cmd_start = cmd_end;
266 				if (*cmd_start == '"') {
267 					term_char = '"';
268 					cmd_end++;
269 				}
270 				else
271 					term_char = ' ';
272 
273 				/* extract actual parameter */
274 				while ((*cmd_end) && (*cmd_end != term_char))
275 					cmd_end++;
276 
277 				if (*cmd_end == '"')
278 					cmd_end++;
279 
280 				save_char = *cmd_end;
281 				*cmd_end = '\0';
282 
283 				if (cmds[i].qualifier) {
284 					if (casesens)
285 						g_string_append(matcherstr, regex ? "regexp " : "match ");
286 					else
287 						g_string_append(matcherstr, regex ? "regexpcase " : "matchcase ");
288 				}
289 
290 				/* do we need to add quotes ? */
291 				if (cmds[i].quotes && term_char != '"')
292 					g_string_append(matcherstr, "\"");
293 
294 				/* copy actual parameter */
295 				g_string_append(matcherstr, cmd_start);
296 
297 				/* do we need to add quotes ? */
298 				if (cmds[i].quotes && term_char != '"')
299 					g_string_append(matcherstr, "\"");
300 
301 				/* restore original character */
302 				*cmd_end = save_char;
303 
304 				break;
305 			}
306 		}
307 
308 		if (*cmd_end)
309 			cmd_end++;
310 		cmd_start = cmd_end;
311 	}
312 
313 	g_free(copy_str);
314 
315 	/* return search string if no match is found to allow
316 	   all available filtering expressions in advanced search */
317 	if (matcherstr->len > 0) {
318 		returnstr = g_string_free(matcherstr, FALSE);
319 	} else {
320 		returnstr = g_strdup(search_string);
321 		g_string_free(matcherstr, TRUE);
322 	}
323 	return returnstr;
324 }
325 
prepare_matcher_extended(AdvancedSearch * search)326 static void prepare_matcher_extended(AdvancedSearch *search)
327 {
328 	gchar *newstr = advsearch_expand_search_string(search->request.matchstring);
329 
330 	if (newstr && newstr[0] != '\0') {
331 		search->predicate = matcher_parser_get_cond(newstr, &search->is_fast);
332 		g_free(newstr);
333 	}
334 }
335 
336 #define debug_matcher_list(prefix, list)					\
337 do {										\
338 	gchar *str = list ? matcherlist_to_string(list) : g_strdup("(NULL)");	\
339 										\
340 	debug_print("%s: %s\n", prefix, str);					\
341 										\
342 	g_free(str);								\
343 } while(0)
344 
prepare_matcher_tag(AdvancedSearch * search)345 static void prepare_matcher_tag(AdvancedSearch *search)
346 {
347 	gchar **words = search->request.matchstring
348 			? g_strsplit(search->request.matchstring, " ", -1)
349 			: NULL;
350 	gint i = 0;
351 
352 	if (search->predicate == NULL) {
353 		search->predicate = g_new0(MatcherList, 1);
354 		search->predicate->bool_and = FALSE;
355 		search->is_fast = TRUE;
356 	}
357 
358 	while (words && words[i] && *words[i]) {
359 		MatcherProp *matcher;
360 
361 		g_strstrip(words[i]);
362 
363 		matcher = matcherprop_new(MATCHCRITERIA_TAG, NULL,
364 					  MATCHTYPE_MATCHCASE, words[i], 0);
365 
366 		search->predicate->matchers = g_slist_prepend(search->predicate->matchers, matcher);
367 
368 		i++;
369 	}
370 	g_strfreev(words);
371 }
372 
prepare_matcher_header(AdvancedSearch * search,gint match_header)373 static void prepare_matcher_header(AdvancedSearch *search, gint match_header)
374 {
375 	MatcherProp *matcher;
376 
377 	if (search->predicate == NULL) {
378 		search->predicate = g_new0(MatcherList, 1);
379 		search->predicate->bool_and = FALSE;
380 		search->is_fast = TRUE;
381 	}
382 
383 	matcher = matcherprop_new(match_header, NULL, MATCHTYPE_MATCHCASE,
384 			search->request.matchstring, 0);
385 
386 	search->predicate->matchers = g_slist_prepend(search->predicate->matchers, matcher);
387 }
388 
prepare_matcher_mixed(AdvancedSearch * search)389 static void prepare_matcher_mixed(AdvancedSearch *search)
390 {
391 	prepare_matcher_tag(search);
392 	debug_matcher_list("tag matcher list", search->predicate);
393 
394 	/* we want an OR search */
395 	if (search->predicate)
396 		search->predicate->bool_and = FALSE;
397 
398 	prepare_matcher_header(search, MATCHCRITERIA_SUBJECT);
399 	debug_matcher_list("tag + subject matcher list", search->predicate);
400 	prepare_matcher_header(search, MATCHCRITERIA_FROM);
401 	debug_matcher_list("tag + subject + from matcher list", search->predicate);
402 	prepare_matcher_header(search, MATCHCRITERIA_TO);
403 	debug_matcher_list("tag + subject + from + to matcher list", search->predicate);
404 	prepare_matcher_header(search, MATCHCRITERIA_CC);
405 	debug_matcher_list("tag + subject + from + to + cc matcher list", search->predicate);
406 }
407 
prepare_matcher(AdvancedSearch * search)408 static void prepare_matcher(AdvancedSearch *search)
409 {
410 	const gchar *search_string;
411 
412 	cm_return_if_fail(search != NULL);
413 
414 	if (search->predicate) {
415 		matcherlist_free(search->predicate);
416 		search->predicate = NULL;
417 	}
418 
419 	search_string = search->request.matchstring;
420 
421 	if (search_string == NULL || search_string[0] == '\0')
422 		return;
423 
424 	switch (search->request.type) {
425 		case ADVANCED_SEARCH_SUBJECT:
426 			prepare_matcher_header(search, MATCHCRITERIA_SUBJECT);
427 			debug_matcher_list("subject search", search->predicate);
428 			break;
429 
430 		case ADVANCED_SEARCH_FROM:
431 			prepare_matcher_header(search, MATCHCRITERIA_FROM);
432 			debug_matcher_list("from search", search->predicate);
433 			break;
434 
435 		case ADVANCED_SEARCH_TO:
436 			prepare_matcher_header(search, MATCHCRITERIA_TO);
437 			debug_matcher_list("to search", search->predicate);
438 			break;
439 
440 		case ADVANCED_SEARCH_TAG:
441 			prepare_matcher_tag(search);
442 			debug_matcher_list("tag search", search->predicate);
443 			break;
444 
445 		case ADVANCED_SEARCH_MIXED:
446 			prepare_matcher_mixed(search);
447 			debug_matcher_list("mixed search", search->predicate);
448 			break;
449 
450 		case ADVANCED_SEARCH_EXTENDED:
451 			prepare_matcher_extended(search);
452 			debug_matcher_list("extended search", search->predicate);
453 			break;
454 
455 		default:
456 			debug_print("unknown search type (%d)\n", search->request.type);
457 			break;
458 	}
459 }
460 
search_progress_notify_cb(gpointer data,gboolean on_server,guint at,guint matched,guint total)461 static gboolean search_progress_notify_cb(gpointer data, gboolean on_server, guint at,
462 		guint matched, guint total)
463 {
464 	AdvancedSearch *search = (AdvancedSearch*) data;
465 
466 	if (search->search_aborted)
467 		return FALSE;
468 
469 	if (on_server || search->on_progress_cb.cb == NULL)
470 		return TRUE;
471 
472 	return search->on_progress_cb.cb(search->on_progress_cb.data, at, matched, total);
473 }
474 
search_filter_folder(MsgNumberList ** msgnums,AdvancedSearch * search,FolderItem * folderItem,gboolean onServer)475 static gboolean search_filter_folder(MsgNumberList **msgnums, AdvancedSearch *search,
476 					  FolderItem *folderItem, gboolean onServer)
477 {
478 	gint matched;
479 	gboolean tried_server = onServer;
480 
481 	matched = folder_item_search_msgs(folderItem->folder,
482 		folderItem,
483 		msgnums,
484 		&onServer,
485 		search->predicate,
486 		search_progress_notify_cb,
487 		search);
488 
489 	if (matched < 0) {
490 		if (search->on_error_cb.cb != NULL)
491 			search->on_error_cb.cb(search->on_error_cb.data);
492 		return FALSE;
493 	}
494 
495 	if (folderItem->folder->klass->supports_server_search && tried_server && !onServer) {
496 		return search_filter_folder(msgnums, search, folderItem, onServer);
497 	} else {
498 		return TRUE;
499 	}
500 }
501 
search_impl(MsgInfoList ** messages,AdvancedSearch * search,FolderItem * folderItem,gboolean recursive)502 static gboolean search_impl(MsgInfoList **messages, AdvancedSearch* search,
503 			    FolderItem* folderItem, gboolean recursive)
504 {
505 	if (recursive) {
506 		START_TIMING("recursive");
507 		if (!search_impl(messages, search, folderItem, FALSE)) {
508 			END_TIMING();
509 			return FALSE;
510 		}
511 		if (folderItem->node->children != NULL && !search->search_aborted) {
512 			GNode *node;
513 			for (node = folderItem->node->children; node != NULL; node = node->next) {
514 				FolderItem *cur = FOLDER_ITEM(node->data);
515 				debug_print("in: %s\n", cur->path);
516 				if (!search_impl(messages, search, cur, TRUE)) {
517 					END_TIMING();
518 					return FALSE;
519 				}
520 			}
521 		}
522 		END_TIMING();
523 	} else if (!folderItem->no_select) {
524 		MsgNumberList *msgnums = NULL;
525 		MsgNumberList *cur;
526 		MsgInfoList *msgs = NULL;
527 		gboolean can_search_on_server = folderItem->folder->klass->supports_server_search;
528 		START_TIMING("folder");
529 		if (!search_filter_folder(&msgnums, search, folderItem,
530 					  can_search_on_server)) {
531 			g_slist_free(msgnums);
532 			END_TIMING();
533 			return FALSE;
534 		}
535 
536 		for (cur = msgnums; cur != NULL; cur = cur->next) {
537 			MsgInfo *msg = folder_item_get_msginfo(folderItem, GPOINTER_TO_UINT(cur->data));
538 
539 			msgs = g_slist_prepend(msgs, msg);
540 		}
541 
542 		while (msgs != NULL) {
543 			MsgInfoList *front = msgs;
544 
545 			msgs = msgs->next;
546 
547 			front->next = *messages;
548 			*messages = front;
549 		}
550 
551 		g_slist_free(msgnums);
552 		END_TIMING();
553 	}
554 
555 	return TRUE;
556 }
557