1 /* thread.cc - Results of thread-based searches from a notmuch database
3 * Copyright © 2009 Carl Worth
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.
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.
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/ .
18 * Author: Carl Worth <cworth@cworth.org>
21 #include "notmuch-private.h"
22 #include "database-private.h"
26 #include <gmime/gmime.h>
27 #include <glib.h> /* GHashTable */
29 struct _notmuch_thread {
30 notmuch_database_t *notmuch;
33 GHashTable *authors_hash;
34 GPtrArray *authors_array;
35 GHashTable *matched_authors_hash;
36 int has_non_matched_authors;
40 notmuch_message_list_t *message_list;
41 GHashTable *message_hash;
49 _notmuch_thread_destructor (notmuch_thread_t *thread)
51 g_hash_table_unref (thread->authors_hash);
52 g_hash_table_unref (thread->matched_authors_hash);
53 g_hash_table_unref (thread->tags);
54 g_hash_table_unref (thread->message_hash);
56 if (thread->authors_array) {
57 g_ptr_array_free (thread->authors_array, TRUE);
58 thread->authors_array = NULL;
64 /* Add each author of the thread to the thread's authors_hash and the
65 * thread's authors_array. */
67 _thread_add_author (notmuch_thread_t *thread,
75 if (g_hash_table_lookup_extended (thread->authors_hash,
79 author_copy = talloc_strdup (thread, author);
81 g_hash_table_insert (thread->authors_hash, author_copy, NULL);
83 g_ptr_array_add (thread->authors_array, author_copy);
86 /* Add each matched author of the thread to the thread's
87 * matched_authors_hash and to the thread's authors string. */
89 _thread_add_matched_author (notmuch_thread_t *thread,
97 if (g_hash_table_lookup_extended (thread->matched_authors_hash,
101 author_copy = talloc_strdup (thread, author);
103 g_hash_table_insert (thread->matched_authors_hash, author_copy, NULL);
106 thread->authors = talloc_asprintf (thread, "%s, %s",
110 thread->authors = author_copy;
113 /* Copy any authors from the authors array that were not also matched
114 * authors onto the end of the thread's authors string.
115 * We use a '|' to separate matched and unmatched authors. */
117 _complete_thread_authors (notmuch_thread_t *thread)
122 for (i = 0; i < thread->authors_array->len; i++) {
123 author = (char *) g_ptr_array_index (thread->authors_array, i);
124 if (g_hash_table_lookup_extended (thread->matched_authors_hash,
127 if (thread->has_non_matched_authors) {
128 thread->authors = talloc_asprintf (thread, "%s, %s",
132 thread->authors = talloc_asprintf (thread, "%s| %s",
135 thread->has_non_matched_authors = 1;
139 g_ptr_array_free (thread->authors_array, TRUE);
140 thread->authors_array = NULL;
143 /* clean up the uggly "Lastname, Firstname" format that some mail systems
144 * (most notably, Exchange) are creating to be "Firstname Lastname"
145 * To make sure that we don't change other potential situations where a
146 * comma is in the name, we check that we match one of these patterns
147 * "Last, First" <first.last@company.com>
148 * "Last, First MI" <first.mi.last@company.com>
151 _thread_cleanup_author (notmuch_thread_t *thread,
152 const char *author, const char *from)
154 char *clean_author,*test_author;
161 clean_author = talloc_strdup(thread, author);
162 if (clean_author == NULL)
164 /* check if there's a comma in the name and that there's a
165 * component of the name behind it (so the name doesn't end with
166 * the comma - in which case the string that strchr finds is just
167 * one character long ",\0").
168 * Otherwise just return the copy of the original author name that
170 comma = strchr(author,',');
171 if (comma && strlen(comma) > 1) {
172 /* let's assemble what we think is the correct name */
173 lname = comma - author;
174 fname = strlen(author) - lname - 2;
175 strncpy(clean_author, comma + 2, fname);
176 *(clean_author+fname) = ' ';
177 strncpy(clean_author + fname + 1, author, lname);
178 *(clean_author+fname+1+lname) = '\0';
179 /* make a temporary copy and see if it matches the email */
180 test_author = talloc_strdup(thread,clean_author);
182 blank=strchr(test_author,' ');
183 while (blank != NULL) {
185 blank=strchr(test_author,' ');
187 if (strcasestr(from, test_author) == NULL)
188 /* we didn't identify this as part of the email address
189 * so let's punt and return the original author */
190 strcpy (clean_author, author);
195 /* Add 'message' as a message that belongs to 'thread'.
197 * The 'thread' will talloc_steal the 'message' and hold onto a
201 _thread_add_message (notmuch_thread_t *thread,
202 notmuch_message_t *message)
204 notmuch_tags_t *tags;
206 InternetAddressList *list = NULL;
207 InternetAddress *address;
208 const char *from, *author;
211 _notmuch_message_list_add_message (thread->message_list,
212 talloc_steal (thread, message));
213 thread->total_messages++;
215 g_hash_table_insert (thread->message_hash,
216 xstrdup (notmuch_message_get_message_id (message)),
219 from = notmuch_message_get_header (message, "from");
221 list = internet_address_list_parse_string (from);
224 address = internet_address_list_get_address (list, 0);
226 author = internet_address_get_name (address);
227 if (author == NULL) {
228 InternetAddressMailbox *mailbox;
229 mailbox = INTERNET_ADDRESS_MAILBOX (address);
230 author = internet_address_mailbox_get_addr (mailbox);
232 clean_author = _thread_cleanup_author (thread, author, from);
233 _thread_add_author (thread, clean_author);
234 notmuch_message_set_author (message, clean_author);
236 g_object_unref (G_OBJECT (list));
239 if (! thread->subject) {
241 subject = notmuch_message_get_header (message, "subject");
242 thread->subject = talloc_strdup (thread, subject ? subject : "");
245 for (tags = notmuch_message_get_tags (message);
246 notmuch_tags_valid (tags);
247 notmuch_tags_move_to_next (tags))
249 tag = notmuch_tags_get (tags);
250 g_hash_table_insert (thread->tags, xstrdup (tag), NULL);
255 _thread_set_subject_from_message (notmuch_thread_t *thread,
256 notmuch_message_t *message)
259 const char *cleaned_subject;
261 subject = notmuch_message_get_header (message, "subject");
265 if ((strncasecmp (subject, "Re: ", 4) == 0) ||
266 (strncasecmp (subject, "Aw: ", 4) == 0) ||
267 (strncasecmp (subject, "Vs: ", 4) == 0) ||
268 (strncasecmp (subject, "Sv: ", 4) == 0)) {
270 cleaned_subject = talloc_strndup (thread,
272 strlen(subject) - 4);
274 cleaned_subject = talloc_strdup (thread, subject);
278 talloc_free (thread->subject);
280 thread->subject = talloc_strdup (thread, cleaned_subject);
284 _thread_add_matched_message (notmuch_thread_t *thread,
285 notmuch_message_t *message,
289 notmuch_message_t *hashed_message;
291 date = notmuch_message_get_date (message);
293 if (date < thread->oldest || ! thread->matched_messages)
294 thread->oldest = date;
296 if (date > thread->newest || ! thread->matched_messages)
297 thread->newest = date;
299 thread->matched_messages++;
301 if (g_hash_table_lookup_extended (thread->message_hash,
302 notmuch_message_get_message_id (message), NULL,
303 (void **) &hashed_message)) {
304 notmuch_message_set_flag (hashed_message,
305 NOTMUCH_MESSAGE_FLAG_MATCH, 1);
308 _thread_add_matched_author (thread, notmuch_message_get_author (hashed_message));
310 if ((sort == NOTMUCH_SORT_OLDEST_FIRST && date <= thread->newest) ||
311 (sort != NOTMUCH_SORT_OLDEST_FIRST && date == thread->newest))
313 _thread_set_subject_from_message (thread, message);
318 _resolve_thread_relationships (unused (notmuch_thread_t *thread))
320 notmuch_message_node_t **prev, *node;
321 notmuch_message_t *message, *parent;
322 const char *in_reply_to;
324 prev = &thread->message_list->head;
325 while ((node = *prev)) {
326 message = node->message;
327 in_reply_to = _notmuch_message_get_in_reply_to (message);
328 if (in_reply_to && strlen (in_reply_to) &&
329 g_hash_table_lookup_extended (thread->message_hash,
334 if (thread->message_list->tail == &node->next)
335 thread->message_list->tail = prev;
337 _notmuch_message_add_reply (parent, node);
339 prev = &((*prev)->next);
343 /* XXX: After scanning through the entire list looking for parents
344 * via "In-Reply-To", we should do a second pass that looks at the
345 * list of messages IDs in the "References" header instead. (And
346 * for this the parent would be the "deepest" message of all the
347 * messages found in the "References" list.)
349 * Doing this will allow messages and sub-threads to be positioned
350 * correctly in the thread even when an intermediate message is
351 * missing from the thread.
355 /* Create a new notmuch_thread_t object for the given thread ID,
356 * treating any messages matching 'query_string' as "matched".
358 * Creating the thread will trigger two database searches. The first
359 * is for all messages belonging to the thread, (to get the first
360 * subject line, the total count of messages, and all authors). The
361 * second search is for all messages that are in the thread and that
362 * also match the given query_string. This is to allow for a separate
363 * count of matched messages, and to allow a viewer to display these
364 * messages differently.
366 * Here, 'ctx' is talloc context for the resulting thread object.
368 * This function returns NULL in the case of any error.
371 _notmuch_thread_create (void *ctx,
372 notmuch_database_t *notmuch,
373 const char *thread_id,
374 const char *query_string,
377 notmuch_thread_t *thread;
378 const char *thread_id_query_string;
379 notmuch_query_t *thread_id_query;
381 notmuch_messages_t *messages;
382 notmuch_message_t *message;
383 notmuch_bool_t matched_is_subset_of_thread;
385 thread_id_query_string = talloc_asprintf (ctx, "thread:%s", thread_id);
386 if (unlikely (query_string == NULL))
389 /* Under normal circumstances we need to do two database
390 * queries. One is for the thread itself (thread_id_query_string)
391 * and the second is to determine which messages in that thread
392 * match the original query (matched_query_string).
394 * But under two circumstances, we use only the
395 * thread_id_query_string:
397 * 1. If the original query_string *is* just the thread
400 * 2. If the original query_string matches all messages ("" or
403 * In either of these cases, we can be more efficient by running
404 * just the thread_id query (since we know all messages in the
405 * thread will match the query_string).
407 * Beyond the performance advantage, in the second case, it's
408 * important to not try to create a concatenated query because our
409 * parser handles "" and "*" as special cases and will not do the
410 * right thing with a query string of "* and thread:<foo>".
412 matched_is_subset_of_thread = 1;
413 if (strcmp (query_string, thread_id_query_string) == 0 ||
414 strcmp (query_string, "") == 0 ||
415 strcmp (query_string, "*") == 0)
417 matched_is_subset_of_thread = 0;
420 thread_id_query = notmuch_query_create (notmuch, thread_id_query_string);
421 if (unlikely (thread_id_query == NULL))
424 thread = talloc (ctx, notmuch_thread_t);
425 if (unlikely (thread == NULL))
428 talloc_set_destructor (thread, _notmuch_thread_destructor);
430 thread->notmuch = notmuch;
431 thread->thread_id = talloc_strdup (thread, thread_id);
432 thread->subject = NULL;
433 thread->authors_hash = g_hash_table_new_full (g_str_hash, g_str_equal,
435 thread->authors_array = g_ptr_array_new ();
436 thread->matched_authors_hash = g_hash_table_new_full (g_str_hash,
439 thread->authors = NULL;
440 thread->has_non_matched_authors = 0;
441 thread->tags = g_hash_table_new_full (g_str_hash, g_str_equal,
444 thread->message_list = _notmuch_message_list_create (thread);
445 if (unlikely (thread->message_list == NULL))
448 thread->message_hash = g_hash_table_new_full (g_str_hash, g_str_equal,
451 thread->total_messages = 0;
452 thread->matched_messages = 0;
456 notmuch_query_set_sort (thread_id_query, NOTMUCH_SORT_OLDEST_FIRST);
458 for (messages = notmuch_query_search_messages (thread_id_query);
459 notmuch_messages_valid (messages);
460 notmuch_messages_move_to_next (messages))
462 message = notmuch_messages_get (messages);
464 _thread_add_message (thread, message);
466 if (! matched_is_subset_of_thread)
467 _thread_add_matched_message (thread, message, sort);
469 _notmuch_message_close (message);
472 notmuch_query_destroy (thread_id_query);
474 if (matched_is_subset_of_thread)
476 const char *matched_query_string;
477 notmuch_query_t *matched_query;
479 matched_query_string = talloc_asprintf (ctx, "%s AND (%s)",
480 thread_id_query_string,
482 if (unlikely (matched_query_string == NULL))
485 matched_query = notmuch_query_create (notmuch, matched_query_string);
486 if (unlikely (matched_query == NULL))
489 for (messages = notmuch_query_search_messages (matched_query);
490 notmuch_messages_valid (messages);
491 notmuch_messages_move_to_next (messages))
493 message = notmuch_messages_get (messages);
494 _thread_add_matched_message (thread, message, sort);
495 _notmuch_message_close (message);
498 notmuch_query_destroy (matched_query);
501 _complete_thread_authors (thread);
503 _resolve_thread_relationships (thread);
509 notmuch_thread_get_toplevel_messages (notmuch_thread_t *thread)
511 return _notmuch_messages_create (thread->message_list);
515 notmuch_thread_get_thread_id (notmuch_thread_t *thread)
517 return thread->thread_id;
521 notmuch_thread_get_total_messages (notmuch_thread_t *thread)
523 return thread->total_messages;
527 notmuch_thread_get_matched_messages (notmuch_thread_t *thread)
529 return thread->matched_messages;
533 notmuch_thread_get_authors (notmuch_thread_t *thread)
535 return thread->authors;
539 notmuch_thread_get_subject (notmuch_thread_t *thread)
541 return thread->subject;
545 notmuch_thread_get_oldest_date (notmuch_thread_t *thread)
547 return thread->oldest;
551 notmuch_thread_get_newest_date (notmuch_thread_t *thread)
553 return thread->newest;
557 notmuch_thread_get_tags (notmuch_thread_t *thread)
559 notmuch_tags_t *tags;
562 tags = _notmuch_tags_create (thread);
563 if (unlikely (tags == NULL))
566 keys = g_hash_table_get_keys (thread->tags);
568 for (l = keys; l; l = l->next)
569 _notmuch_tags_add_tag (tags, (char *) l->data);
573 _notmuch_tags_prepare_iterator (tags);
579 notmuch_thread_destroy (notmuch_thread_t *thread)
581 talloc_free (thread);