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"
24 #include <gmime/gmime.h>
25 #include <glib.h> /* GHashTable */
27 struct visible _notmuch_thread {
28 notmuch_database_t *notmuch;
31 GHashTable *authors_hash;
32 GPtrArray *authors_array;
33 GHashTable *matched_authors_hash;
34 GPtrArray *matched_authors_array;
38 /* All messages, oldest first. */
39 notmuch_message_list_t *message_list;
40 /* Top-level messages, oldest first. */
41 notmuch_message_list_t *toplevel_list;
43 GHashTable *message_hash;
51 _notmuch_thread_destructor (notmuch_thread_t *thread)
53 g_hash_table_unref (thread->authors_hash);
54 g_hash_table_unref (thread->matched_authors_hash);
55 g_hash_table_unref (thread->tags);
56 g_hash_table_unref (thread->message_hash);
58 if (thread->authors_array) {
59 g_ptr_array_free (thread->authors_array, TRUE);
60 thread->authors_array = NULL;
63 if (thread->matched_authors_array) {
64 g_ptr_array_free (thread->matched_authors_array, TRUE);
65 thread->matched_authors_array = NULL;
71 /* Add each author of the thread to the thread's authors_hash and to
72 * the thread's authors_array. */
74 _thread_add_author (notmuch_thread_t *thread,
82 if (g_hash_table_lookup_extended (thread->authors_hash,
86 author_copy = talloc_strdup (thread, author);
88 g_hash_table_insert (thread->authors_hash, author_copy, NULL);
90 g_ptr_array_add (thread->authors_array, author_copy);
93 /* Add each matched author of the thread to the thread's
94 * matched_authors_hash and to the thread's matched_authors_array. */
96 _thread_add_matched_author (notmuch_thread_t *thread,
104 if (g_hash_table_lookup_extended (thread->matched_authors_hash,
108 author_copy = talloc_strdup (thread, author);
110 g_hash_table_insert (thread->matched_authors_hash, author_copy, NULL);
112 g_ptr_array_add (thread->matched_authors_array, author_copy);
115 /* Construct an authors string from matched_authors_array and
116 * authors_array. The string contains matched authors first, then
117 * non-matched authors (with the two groups separated by '|'). Within
118 * each group, authors are listed in date order. */
120 _resolve_thread_authors_string (notmuch_thread_t *thread)
124 int first_non_matched_author = 1;
126 /* First, list all matched authors in date order. */
127 for (i = 0; i < thread->matched_authors_array->len; i++) {
128 author = (char *) g_ptr_array_index (thread->matched_authors_array, i);
130 thread->authors = talloc_asprintf (thread, "%s, %s",
134 thread->authors = author;
137 /* Next, append any non-matched authors that haven't already appeared. */
138 for (i = 0; i < thread->authors_array->len; i++) {
139 author = (char *) g_ptr_array_index (thread->authors_array, i);
140 if (g_hash_table_lookup_extended (thread->matched_authors_hash,
143 if (first_non_matched_author) {
144 thread->authors = talloc_asprintf (thread, "%s| %s",
148 thread->authors = talloc_asprintf (thread, "%s, %s",
153 first_non_matched_author = 0;
156 g_ptr_array_free (thread->authors_array, TRUE);
157 thread->authors_array = NULL;
158 g_ptr_array_free (thread->matched_authors_array, TRUE);
159 thread->matched_authors_array = NULL;
162 /* clean up the ugly "Lastname, Firstname" format that some mail systems
163 * (most notably, Exchange) are creating to be "Firstname Lastname"
164 * To make sure that we don't change other potential situations where a
165 * comma is in the name, we check that we match one of these patterns
166 * "Last, First" <first.last@company.com>
167 * "Last, First MI" <first.mi.last@company.com>
170 _thread_cleanup_author (notmuch_thread_t *thread,
171 const char *author, const char *from)
173 char *clean_author,*test_author;
180 clean_author = talloc_strdup(thread, author);
181 if (clean_author == NULL)
183 /* check if there's a comma in the name and that there's a
184 * component of the name behind it (so the name doesn't end with
185 * the comma - in which case the string that strchr finds is just
186 * one character long ",\0").
187 * Otherwise just return the copy of the original author name that
189 comma = strchr(author,',');
190 if (comma && strlen(comma) > 1) {
191 /* let's assemble what we think is the correct name */
192 lname = comma - author;
193 fname = strlen(author) - lname - 2;
194 strncpy(clean_author, comma + 2, fname);
195 *(clean_author+fname) = ' ';
196 strncpy(clean_author + fname + 1, author, lname);
197 *(clean_author+fname+1+lname) = '\0';
198 /* make a temporary copy and see if it matches the email */
199 test_author = talloc_strdup(thread,clean_author);
201 blank=strchr(test_author,' ');
202 while (blank != NULL) {
204 blank=strchr(test_author,' ');
206 if (strcasestr(from, test_author) == NULL)
207 /* we didn't identify this as part of the email address
208 * so let's punt and return the original author */
209 strcpy (clean_author, author);
214 /* Add 'message' as a message that belongs to 'thread'.
216 * The 'thread' will talloc_steal the 'message' and hold onto a
220 _thread_add_message (notmuch_thread_t *thread,
221 notmuch_message_t *message,
222 notmuch_string_list_t *exclude_terms)
224 notmuch_tags_t *tags;
226 InternetAddressList *list = NULL;
227 InternetAddress *address;
228 const char *from, *author;
231 _notmuch_message_list_add_message (thread->message_list,
232 talloc_steal (thread, message));
233 thread->total_messages++;
235 g_hash_table_insert (thread->message_hash,
236 xstrdup (notmuch_message_get_message_id (message)),
239 from = notmuch_message_get_header (message, "from");
241 list = internet_address_list_parse_string (from);
244 address = internet_address_list_get_address (list, 0);
246 author = internet_address_get_name (address);
247 if (author == NULL) {
248 InternetAddressMailbox *mailbox;
249 mailbox = INTERNET_ADDRESS_MAILBOX (address);
250 author = internet_address_mailbox_get_addr (mailbox);
252 clean_author = _thread_cleanup_author (thread, author, from);
253 _thread_add_author (thread, clean_author);
254 notmuch_message_set_author (message, clean_author);
256 g_object_unref (G_OBJECT (list));
259 if (! thread->subject) {
261 subject = notmuch_message_get_header (message, "subject");
262 thread->subject = talloc_strdup (thread, subject ? subject : "");
265 for (tags = notmuch_message_get_tags (message);
266 notmuch_tags_valid (tags);
267 notmuch_tags_move_to_next (tags))
269 tag = notmuch_tags_get (tags);
270 /* Mark excluded messages. */
271 for (notmuch_string_node_t *term = exclude_terms->head; term;
273 /* We ignore initial 'K'. */
274 if (strcmp(tag, (term->string + 1)) == 0) {
275 notmuch_message_set_flag (message, NOTMUCH_MESSAGE_FLAG_EXCLUDED, TRUE);
279 g_hash_table_insert (thread->tags, xstrdup (tag), NULL);
284 _thread_set_subject_from_message (notmuch_thread_t *thread,
285 notmuch_message_t *message)
288 const char *cleaned_subject;
290 subject = notmuch_message_get_header (message, "subject");
294 if ((strncasecmp (subject, "Re: ", 4) == 0) ||
295 (strncasecmp (subject, "Aw: ", 4) == 0) ||
296 (strncasecmp (subject, "Vs: ", 4) == 0) ||
297 (strncasecmp (subject, "Sv: ", 4) == 0)) {
299 cleaned_subject = talloc_strndup (thread,
301 strlen(subject) - 4);
303 cleaned_subject = talloc_strdup (thread, subject);
307 talloc_free (thread->subject);
309 thread->subject = talloc_strdup (thread, cleaned_subject);
312 /* Add a message to this thread which is known to match the original
313 * search specification. The 'sort' parameter controls whether the
314 * oldest or newest matching subject is applied to the thread as a
317 _thread_add_matched_message (notmuch_thread_t *thread,
318 notmuch_message_t *message,
322 notmuch_message_t *hashed_message;
324 date = notmuch_message_get_date (message);
326 if (date < thread->oldest || ! thread->matched_messages) {
327 thread->oldest = date;
328 if (sort == NOTMUCH_SORT_OLDEST_FIRST)
329 _thread_set_subject_from_message (thread, message);
332 if (date > thread->newest || ! thread->matched_messages) {
333 thread->newest = date;
334 if (sort != NOTMUCH_SORT_OLDEST_FIRST)
335 _thread_set_subject_from_message (thread, message);
338 if (!notmuch_message_get_flag (message, NOTMUCH_MESSAGE_FLAG_EXCLUDED))
339 thread->matched_messages++;
341 if (g_hash_table_lookup_extended (thread->message_hash,
342 notmuch_message_get_message_id (message), NULL,
343 (void **) &hashed_message)) {
344 notmuch_message_set_flag (hashed_message,
345 NOTMUCH_MESSAGE_FLAG_MATCH, 1);
348 _thread_add_matched_author (thread, notmuch_message_get_author (hashed_message));
352 _resolve_thread_relationships (notmuch_thread_t *thread)
354 notmuch_message_node_t *node;
355 notmuch_message_t *message, *parent;
356 const char *in_reply_to;
358 for (node = thread->message_list->head; node; node = node->next) {
359 message = node->message;
360 in_reply_to = _notmuch_message_get_in_reply_to (message);
361 if (in_reply_to && strlen (in_reply_to) &&
362 g_hash_table_lookup_extended (thread->message_hash,
365 _notmuch_message_add_reply (parent, message);
367 _notmuch_message_list_add_message (thread->toplevel_list, message);
370 /* XXX: After scanning through the entire list looking for parents
371 * via "In-Reply-To", we should do a second pass that looks at the
372 * list of messages IDs in the "References" header instead. (And
373 * for this the parent would be the "deepest" message of all the
374 * messages found in the "References" list.)
376 * Doing this will allow messages and sub-threads to be positioned
377 * correctly in the thread even when an intermediate message is
378 * missing from the thread.
382 /* Create a new notmuch_thread_t object by finding the thread
383 * containing the message with the given doc ID, treating any messages
384 * contained in match_set as "matched". Remove all messages in the
385 * thread from match_set.
387 * Creating the thread will perform a database search to get all
388 * messages belonging to the thread and will get the first subject
389 * line, the total count of messages, and all authors in the thread.
390 * Each message in the thread is checked against match_set to allow
391 * for a separate count of matched messages, and to allow a viewer to
392 * display these messages differently.
394 * Here, 'ctx' is talloc context for the resulting thread object.
396 * This function returns NULL in the case of any error.
399 _notmuch_thread_create (void *ctx,
400 notmuch_database_t *notmuch,
401 unsigned int seed_doc_id,
402 notmuch_doc_id_set_t *match_set,
403 notmuch_string_list_t *exclude_terms,
406 void *local = talloc_new (ctx);
407 notmuch_thread_t *thread = NULL;
408 notmuch_message_t *seed_message;
409 const char *thread_id;
410 char *thread_id_query_string;
411 notmuch_query_t *thread_id_query;
413 notmuch_messages_t *messages;
414 notmuch_message_t *message;
416 seed_message = _notmuch_message_create (local, notmuch, seed_doc_id, NULL);
418 INTERNAL_ERROR ("Thread seed message %u does not exist", seed_doc_id);
420 thread_id = notmuch_message_get_thread_id (seed_message);
421 thread_id_query_string = talloc_asprintf (local, "thread:%s", thread_id);
422 if (unlikely (thread_id_query_string == NULL))
425 thread_id_query = talloc_steal (
426 local, notmuch_query_create (notmuch, thread_id_query_string));
427 if (unlikely (thread_id_query == NULL))
430 thread = talloc (local, notmuch_thread_t);
431 if (unlikely (thread == NULL))
434 talloc_set_destructor (thread, _notmuch_thread_destructor);
436 thread->notmuch = notmuch;
437 thread->thread_id = talloc_strdup (thread, thread_id);
438 thread->subject = NULL;
439 thread->authors_hash = g_hash_table_new_full (g_str_hash, g_str_equal,
441 thread->authors_array = g_ptr_array_new ();
442 thread->matched_authors_hash = g_hash_table_new_full (g_str_hash,
445 thread->matched_authors_array = g_ptr_array_new ();
446 thread->authors = NULL;
447 thread->tags = g_hash_table_new_full (g_str_hash, g_str_equal,
450 thread->message_list = _notmuch_message_list_create (thread);
451 thread->toplevel_list = _notmuch_message_list_create (thread);
452 if (unlikely (thread->message_list == NULL ||
453 thread->toplevel_list == NULL)) {
458 thread->message_hash = g_hash_table_new_full (g_str_hash, g_str_equal,
461 thread->total_messages = 0;
462 thread->matched_messages = 0;
466 /* We use oldest-first order unconditionally here to obtain the
467 * proper author ordering for the thread. The 'sort' parameter
468 * passed to this function is used only to indicate whether the
469 * oldest or newest subject is desired. */
470 notmuch_query_set_sort (thread_id_query, NOTMUCH_SORT_OLDEST_FIRST);
472 for (messages = notmuch_query_search_messages (thread_id_query);
473 notmuch_messages_valid (messages);
474 notmuch_messages_move_to_next (messages))
478 message = notmuch_messages_get (messages);
479 doc_id = _notmuch_message_get_doc_id (message);
480 if (doc_id == seed_doc_id)
481 message = seed_message;
483 _thread_add_message (thread, message, exclude_terms);
485 if ( _notmuch_doc_id_set_contains (match_set, doc_id)) {
486 _notmuch_doc_id_set_remove (match_set, doc_id);
487 _thread_add_matched_message (thread, message, sort);
490 _notmuch_message_close (message);
493 _resolve_thread_authors_string (thread);
495 _resolve_thread_relationships (thread);
497 /* Commit to returning thread. */
498 talloc_steal (ctx, thread);
506 notmuch_thread_get_toplevel_messages (notmuch_thread_t *thread)
508 return _notmuch_messages_create (thread->toplevel_list);
512 notmuch_thread_get_thread_id (notmuch_thread_t *thread)
514 return thread->thread_id;
518 notmuch_thread_get_total_messages (notmuch_thread_t *thread)
520 return thread->total_messages;
524 notmuch_thread_get_matched_messages (notmuch_thread_t *thread)
526 return thread->matched_messages;
530 notmuch_thread_get_authors (notmuch_thread_t *thread)
532 return thread->authors;
536 notmuch_thread_get_subject (notmuch_thread_t *thread)
538 return thread->subject;
542 notmuch_thread_get_oldest_date (notmuch_thread_t *thread)
544 return thread->oldest;
548 notmuch_thread_get_newest_date (notmuch_thread_t *thread)
550 return thread->newest;
554 notmuch_thread_get_tags (notmuch_thread_t *thread)
556 notmuch_string_list_t *tags;
559 tags = _notmuch_string_list_create (thread);
560 if (unlikely (tags == NULL))
563 keys = g_hash_table_get_keys (thread->tags);
565 for (l = keys; l; l = l->next)
566 _notmuch_string_list_append (tags, (char *) l->data);
570 _notmuch_string_list_sort (tags);
572 return _notmuch_tags_create (thread, tags);
576 notmuch_thread_destroy (notmuch_thread_t *thread)
578 talloc_free (thread);