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 notmuch_message_list_t *message_list;
39 GHashTable *message_hash;
47 _notmuch_thread_destructor (notmuch_thread_t *thread)
49 g_hash_table_unref (thread->authors_hash);
50 g_hash_table_unref (thread->matched_authors_hash);
51 g_hash_table_unref (thread->tags);
52 g_hash_table_unref (thread->message_hash);
54 if (thread->authors_array) {
55 g_ptr_array_free (thread->authors_array, TRUE);
56 thread->authors_array = NULL;
59 if (thread->matched_authors_array) {
60 g_ptr_array_free (thread->matched_authors_array, TRUE);
61 thread->matched_authors_array = NULL;
67 /* Add each author of the thread to the thread's authors_hash and to
68 * the thread's authors_array. */
70 _thread_add_author (notmuch_thread_t *thread,
78 if (g_hash_table_lookup_extended (thread->authors_hash,
82 author_copy = talloc_strdup (thread, author);
84 g_hash_table_insert (thread->authors_hash, author_copy, NULL);
86 g_ptr_array_add (thread->authors_array, author_copy);
89 /* Add each matched author of the thread to the thread's
90 * matched_authors_hash and to the thread's matched_authors_array. */
92 _thread_add_matched_author (notmuch_thread_t *thread,
100 if (g_hash_table_lookup_extended (thread->matched_authors_hash,
104 author_copy = talloc_strdup (thread, author);
106 g_hash_table_insert (thread->matched_authors_hash, author_copy, NULL);
108 g_ptr_array_add (thread->matched_authors_array, author_copy);
111 /* Construct an authors string from matched_authors_array and
112 * authors_array. The string contains matched authors first, then
113 * non-matched authors (with the two groups separated by '|'). Within
114 * each group, authors are listed in date order. */
116 _resolve_thread_authors_string (notmuch_thread_t *thread)
120 int first_non_matched_author = 1;
122 /* First, list all matched authors in date order. */
123 for (i = 0; i < thread->matched_authors_array->len; i++) {
124 author = (char *) g_ptr_array_index (thread->matched_authors_array, i);
126 thread->authors = talloc_asprintf (thread, "%s, %s",
130 thread->authors = author;
133 /* Next, append any non-matched authors that haven't already appeared. */
134 for (i = 0; i < thread->authors_array->len; i++) {
135 author = (char *) g_ptr_array_index (thread->authors_array, i);
136 if (g_hash_table_lookup_extended (thread->matched_authors_hash,
139 if (first_non_matched_author) {
140 thread->authors = talloc_asprintf (thread, "%s| %s",
144 thread->authors = talloc_asprintf (thread, "%s, %s",
149 first_non_matched_author = 0;
152 g_ptr_array_free (thread->authors_array, TRUE);
153 thread->authors_array = NULL;
154 g_ptr_array_free (thread->matched_authors_array, TRUE);
155 thread->matched_authors_array = NULL;
158 /* clean up the ugly "Lastname, Firstname" format that some mail systems
159 * (most notably, Exchange) are creating to be "Firstname Lastname"
160 * To make sure that we don't change other potential situations where a
161 * comma is in the name, we check that we match one of these patterns
162 * "Last, First" <first.last@company.com>
163 * "Last, First MI" <first.mi.last@company.com>
166 _thread_cleanup_author (notmuch_thread_t *thread,
167 const char *author, const char *from)
169 char *clean_author,*test_author;
176 clean_author = talloc_strdup(thread, author);
177 if (clean_author == NULL)
179 /* check if there's a comma in the name and that there's a
180 * component of the name behind it (so the name doesn't end with
181 * the comma - in which case the string that strchr finds is just
182 * one character long ",\0").
183 * Otherwise just return the copy of the original author name that
185 comma = strchr(author,',');
186 if (comma && strlen(comma) > 1) {
187 /* let's assemble what we think is the correct name */
188 lname = comma - author;
189 fname = strlen(author) - lname - 2;
190 strncpy(clean_author, comma + 2, fname);
191 *(clean_author+fname) = ' ';
192 strncpy(clean_author + fname + 1, author, lname);
193 *(clean_author+fname+1+lname) = '\0';
194 /* make a temporary copy and see if it matches the email */
195 test_author = talloc_strdup(thread,clean_author);
197 blank=strchr(test_author,' ');
198 while (blank != NULL) {
200 blank=strchr(test_author,' ');
202 if (strcasestr(from, test_author) == NULL)
203 /* we didn't identify this as part of the email address
204 * so let's punt and return the original author */
205 strcpy (clean_author, author);
210 /* Add 'message' as a message that belongs to 'thread'.
212 * The 'thread' will talloc_steal the 'message' and hold onto a
216 _thread_add_message (notmuch_thread_t *thread,
217 notmuch_message_t *message,
218 notmuch_string_list_t *exclude_terms)
220 notmuch_tags_t *tags;
222 InternetAddressList *list = NULL;
223 InternetAddress *address;
224 const char *from, *author;
227 _notmuch_message_list_add_message (thread->message_list,
228 talloc_steal (thread, message));
229 thread->total_messages++;
231 g_hash_table_insert (thread->message_hash,
232 xstrdup (notmuch_message_get_message_id (message)),
235 from = notmuch_message_get_header (message, "from");
237 list = internet_address_list_parse_string (from);
240 address = internet_address_list_get_address (list, 0);
242 author = internet_address_get_name (address);
243 if (author == NULL) {
244 InternetAddressMailbox *mailbox;
245 mailbox = INTERNET_ADDRESS_MAILBOX (address);
246 author = internet_address_mailbox_get_addr (mailbox);
248 clean_author = _thread_cleanup_author (thread, author, from);
249 _thread_add_author (thread, clean_author);
250 notmuch_message_set_author (message, clean_author);
252 g_object_unref (G_OBJECT (list));
255 if (! thread->subject) {
257 subject = notmuch_message_get_header (message, "subject");
258 thread->subject = talloc_strdup (thread, subject ? subject : "");
261 for (tags = notmuch_message_get_tags (message);
262 notmuch_tags_valid (tags);
263 notmuch_tags_move_to_next (tags))
265 tag = notmuch_tags_get (tags);
266 /* Mark excluded messages. */
267 for (notmuch_string_node_t *term = exclude_terms->head; term;
269 /* We ignore initial 'K'. */
270 if (strcmp(tag, (term->string + 1)) == 0) {
271 notmuch_message_set_flag (message, NOTMUCH_MESSAGE_FLAG_EXCLUDED, TRUE);
275 g_hash_table_insert (thread->tags, xstrdup (tag), NULL);
280 _thread_set_subject_from_message (notmuch_thread_t *thread,
281 notmuch_message_t *message)
284 const char *cleaned_subject;
286 subject = notmuch_message_get_header (message, "subject");
290 if ((strncasecmp (subject, "Re: ", 4) == 0) ||
291 (strncasecmp (subject, "Aw: ", 4) == 0) ||
292 (strncasecmp (subject, "Vs: ", 4) == 0) ||
293 (strncasecmp (subject, "Sv: ", 4) == 0)) {
295 cleaned_subject = talloc_strndup (thread,
297 strlen(subject) - 4);
299 cleaned_subject = talloc_strdup (thread, subject);
303 talloc_free (thread->subject);
305 thread->subject = talloc_strdup (thread, cleaned_subject);
308 /* Add a message to this thread which is known to match the original
309 * search specification. The 'sort' parameter controls whether the
310 * oldest or newest matching subject is applied to the thread as a
313 _thread_add_matched_message (notmuch_thread_t *thread,
314 notmuch_message_t *message,
318 notmuch_message_t *hashed_message;
320 date = notmuch_message_get_date (message);
322 if (date < thread->oldest || ! thread->matched_messages) {
323 thread->oldest = date;
324 if (sort == NOTMUCH_SORT_OLDEST_FIRST)
325 _thread_set_subject_from_message (thread, message);
328 if (date > thread->newest || ! thread->matched_messages) {
329 thread->newest = date;
330 if (sort != NOTMUCH_SORT_OLDEST_FIRST)
331 _thread_set_subject_from_message (thread, message);
334 if (!notmuch_message_get_flag (message, NOTMUCH_MESSAGE_FLAG_EXCLUDED))
335 thread->matched_messages++;
337 if (g_hash_table_lookup_extended (thread->message_hash,
338 notmuch_message_get_message_id (message), NULL,
339 (void **) &hashed_message)) {
340 notmuch_message_set_flag (hashed_message,
341 NOTMUCH_MESSAGE_FLAG_MATCH, 1);
344 _thread_add_matched_author (thread, notmuch_message_get_author (hashed_message));
348 _resolve_thread_relationships (unused (notmuch_thread_t *thread))
350 notmuch_message_node_t **prev, *node;
351 notmuch_message_t *message, *parent;
352 const char *in_reply_to;
354 prev = &thread->message_list->head;
355 while ((node = *prev)) {
356 message = node->message;
357 in_reply_to = _notmuch_message_get_in_reply_to (message);
358 if (in_reply_to && strlen (in_reply_to) &&
359 g_hash_table_lookup_extended (thread->message_hash,
364 if (thread->message_list->tail == &node->next)
365 thread->message_list->tail = prev;
367 _notmuch_message_add_reply (parent, node);
369 prev = &((*prev)->next);
373 /* XXX: After scanning through the entire list looking for parents
374 * via "In-Reply-To", we should do a second pass that looks at the
375 * list of messages IDs in the "References" header instead. (And
376 * for this the parent would be the "deepest" message of all the
377 * messages found in the "References" list.)
379 * Doing this will allow messages and sub-threads to be positioned
380 * correctly in the thread even when an intermediate message is
381 * missing from the thread.
385 /* Create a new notmuch_thread_t object by finding the thread
386 * containing the message with the given doc ID, treating any messages
387 * contained in match_set as "matched". Remove all messages in the
388 * thread from match_set.
390 * Creating the thread will perform a database search to get all
391 * messages belonging to the thread and will get the first subject
392 * line, the total count of messages, and all authors in the thread.
393 * Each message in the thread is checked against match_set to allow
394 * for a separate count of matched messages, and to allow a viewer to
395 * display these messages differently.
397 * Here, 'ctx' is talloc context for the resulting thread object.
399 * This function returns NULL in the case of any error.
402 _notmuch_thread_create (void *ctx,
403 notmuch_database_t *notmuch,
404 unsigned int seed_doc_id,
405 notmuch_doc_id_set_t *match_set,
406 notmuch_string_list_t *exclude_terms,
409 notmuch_thread_t *thread;
410 notmuch_message_t *seed_message;
411 const char *thread_id;
412 char *thread_id_query_string;
413 notmuch_query_t *thread_id_query;
415 notmuch_messages_t *messages;
416 notmuch_message_t *message;
418 seed_message = _notmuch_message_create (ctx, notmuch, seed_doc_id, NULL);
420 INTERNAL_ERROR ("Thread seed message %u does not exist", seed_doc_id);
422 thread_id = notmuch_message_get_thread_id (seed_message);
423 thread_id_query_string = talloc_asprintf (ctx, "thread:%s", thread_id);
424 if (unlikely (thread_id_query_string == NULL))
427 thread_id_query = notmuch_query_create (notmuch, thread_id_query_string);
428 if (unlikely (thread_id_query == NULL))
431 talloc_free (thread_id_query_string);
433 thread = talloc (ctx, notmuch_thread_t);
434 if (unlikely (thread == NULL))
437 talloc_set_destructor (thread, _notmuch_thread_destructor);
439 thread->notmuch = notmuch;
440 thread->thread_id = talloc_strdup (thread, thread_id);
441 thread->subject = NULL;
442 thread->authors_hash = g_hash_table_new_full (g_str_hash, g_str_equal,
444 thread->authors_array = g_ptr_array_new ();
445 thread->matched_authors_hash = g_hash_table_new_full (g_str_hash,
448 thread->matched_authors_array = g_ptr_array_new ();
449 thread->authors = NULL;
450 thread->tags = g_hash_table_new_full (g_str_hash, g_str_equal,
453 thread->message_list = _notmuch_message_list_create (thread);
454 if (unlikely (thread->message_list == NULL))
457 thread->message_hash = g_hash_table_new_full (g_str_hash, g_str_equal,
460 thread->total_messages = 0;
461 thread->matched_messages = 0;
465 /* We use oldest-first order unconditionally here to obtain the
466 * proper author ordering for the thread. The 'sort' parameter
467 * passed to this function is used only to indicate whether the
468 * oldest or newest subject is desired. */
469 notmuch_query_set_sort (thread_id_query, NOTMUCH_SORT_OLDEST_FIRST);
471 for (messages = notmuch_query_search_messages (thread_id_query);
472 notmuch_messages_valid (messages);
473 notmuch_messages_move_to_next (messages))
477 message = notmuch_messages_get (messages);
478 doc_id = _notmuch_message_get_doc_id (message);
479 if (doc_id == seed_doc_id)
480 message = seed_message;
482 _thread_add_message (thread, message, exclude_terms);
484 if ( _notmuch_doc_id_set_contains (match_set, doc_id)) {
485 _notmuch_doc_id_set_remove (match_set, doc_id);
486 _thread_add_matched_message (thread, message, sort);
489 _notmuch_message_close (message);
492 notmuch_query_destroy (thread_id_query);
494 _resolve_thread_authors_string (thread);
496 _resolve_thread_relationships (thread);
502 notmuch_thread_get_toplevel_messages (notmuch_thread_t *thread)
504 return _notmuch_messages_create (thread->message_list);
508 notmuch_thread_get_thread_id (notmuch_thread_t *thread)
510 return thread->thread_id;
514 notmuch_thread_get_total_messages (notmuch_thread_t *thread)
516 return thread->total_messages;
520 notmuch_thread_get_matched_messages (notmuch_thread_t *thread)
522 return thread->matched_messages;
526 notmuch_thread_get_authors (notmuch_thread_t *thread)
528 return thread->authors;
532 notmuch_thread_get_subject (notmuch_thread_t *thread)
534 return thread->subject;
538 notmuch_thread_get_oldest_date (notmuch_thread_t *thread)
540 return thread->oldest;
544 notmuch_thread_get_newest_date (notmuch_thread_t *thread)
546 return thread->newest;
550 notmuch_thread_get_tags (notmuch_thread_t *thread)
552 notmuch_string_list_t *tags;
555 tags = _notmuch_string_list_create (thread);
556 if (unlikely (tags == NULL))
559 keys = g_hash_table_get_keys (thread->tags);
561 for (l = keys; l; l = l->next)
562 _notmuch_string_list_append (tags, (char *) l->data);
566 _notmuch_string_list_sort (tags);
568 return _notmuch_tags_create (thread, tags);
572 notmuch_thread_destroy (notmuch_thread_t *thread)
574 talloc_free (thread);