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 #define EMPTY_STRING(s) ((s)[0] == '\0')
29 struct visible _notmuch_thread {
30 notmuch_database_t *notmuch;
33 GHashTable *authors_hash;
34 GPtrArray *authors_array;
35 GHashTable *matched_authors_hash;
36 GPtrArray *matched_authors_array;
40 /* All messages, oldest first. */
41 notmuch_message_list_t *message_list;
42 /* Top-level messages, oldest first. */
43 notmuch_message_list_t *toplevel_list;
45 GHashTable *message_hash;
53 _notmuch_thread_destructor (notmuch_thread_t *thread)
55 g_hash_table_unref (thread->authors_hash);
56 g_hash_table_unref (thread->matched_authors_hash);
57 g_hash_table_unref (thread->tags);
58 g_hash_table_unref (thread->message_hash);
60 if (thread->authors_array) {
61 g_ptr_array_free (thread->authors_array, TRUE);
62 thread->authors_array = NULL;
65 if (thread->matched_authors_array) {
66 g_ptr_array_free (thread->matched_authors_array, TRUE);
67 thread->matched_authors_array = NULL;
73 /* Add each author of the thread to the thread's authors_hash and to
74 * the thread's authors_array. */
76 _thread_add_author (notmuch_thread_t *thread,
84 if (g_hash_table_lookup_extended (thread->authors_hash,
88 author_copy = talloc_strdup (thread, author);
90 g_hash_table_insert (thread->authors_hash, author_copy, NULL);
92 g_ptr_array_add (thread->authors_array, author_copy);
95 /* Add each matched author of the thread to the thread's
96 * matched_authors_hash and to the thread's matched_authors_array. */
98 _thread_add_matched_author (notmuch_thread_t *thread,
106 if (g_hash_table_lookup_extended (thread->matched_authors_hash,
110 author_copy = talloc_strdup (thread, author);
112 g_hash_table_insert (thread->matched_authors_hash, author_copy, NULL);
114 g_ptr_array_add (thread->matched_authors_array, author_copy);
117 /* Construct an authors string from matched_authors_array and
118 * authors_array. The string contains matched authors first, then
119 * non-matched authors (with the two groups separated by '|'). Within
120 * each group, authors are listed in date order. */
122 _resolve_thread_authors_string (notmuch_thread_t *thread)
126 int first_non_matched_author = 1;
128 /* First, list all matched authors in date order. */
129 for (i = 0; i < thread->matched_authors_array->len; i++) {
130 author = (char *) g_ptr_array_index (thread->matched_authors_array, i);
132 thread->authors = talloc_asprintf (thread, "%s, %s",
136 thread->authors = author;
139 /* Next, append any non-matched authors that haven't already appeared. */
140 for (i = 0; i < thread->authors_array->len; i++) {
141 author = (char *) g_ptr_array_index (thread->authors_array, i);
142 if (g_hash_table_lookup_extended (thread->matched_authors_hash,
145 if (first_non_matched_author) {
146 thread->authors = talloc_asprintf (thread, "%s| %s",
150 thread->authors = talloc_asprintf (thread, "%s, %s",
155 first_non_matched_author = 0;
158 g_ptr_array_free (thread->authors_array, TRUE);
159 thread->authors_array = NULL;
160 g_ptr_array_free (thread->matched_authors_array, TRUE);
161 thread->matched_authors_array = NULL;
164 /* clean up the ugly "Lastname, Firstname" format that some mail systems
165 * (most notably, Exchange) are creating to be "Firstname Lastname"
166 * To make sure that we don't change other potential situations where a
167 * comma is in the name, we check that we match one of these patterns
168 * "Last, First" <first.last@company.com>
169 * "Last, First MI" <first.mi.last@company.com>
172 _thread_cleanup_author (notmuch_thread_t *thread,
173 const char *author, const char *from)
175 char *clean_author,*test_author;
182 clean_author = talloc_strdup(thread, author);
183 if (clean_author == NULL)
185 /* check if there's a comma in the name and that there's a
186 * component of the name behind it (so the name doesn't end with
187 * the comma - in which case the string that strchr finds is just
188 * one character long ",\0").
189 * Otherwise just return the copy of the original author name that
191 comma = strchr(author,',');
192 if (comma && strlen(comma) > 1) {
193 /* let's assemble what we think is the correct name */
194 lname = comma - author;
196 /* Skip all the spaces after the comma */
197 fname = strlen(author) - lname - 1;
199 while (*comma == ' ') {
203 strncpy(clean_author, comma, fname);
205 *(clean_author+fname) = ' ';
206 strncpy(clean_author + fname + 1, author, lname);
207 *(clean_author+fname+1+lname) = '\0';
208 /* make a temporary copy and see if it matches the email */
209 test_author = talloc_strdup(thread,clean_author);
211 blank=strchr(test_author,' ');
212 while (blank != NULL) {
214 blank=strchr(test_author,' ');
216 if (strcasestr(from, test_author) == NULL)
217 /* we didn't identify this as part of the email address
218 * so let's punt and return the original author */
219 strcpy (clean_author, author);
224 /* Add 'message' as a message that belongs to 'thread'.
226 * The 'thread' will talloc_steal the 'message' and hold onto a
230 _thread_add_message (notmuch_thread_t *thread,
231 notmuch_message_t *message,
232 notmuch_string_list_t *exclude_terms,
233 notmuch_exclude_t omit_exclude)
235 notmuch_tags_t *tags;
237 InternetAddressList *list = NULL;
238 InternetAddress *address;
239 const char *from, *author;
241 notmuch_bool_t message_excluded = FALSE;
243 if (omit_exclude != NOTMUCH_EXCLUDE_FALSE) {
244 for (tags = notmuch_message_get_tags (message);
245 notmuch_tags_valid (tags);
246 notmuch_tags_move_to_next (tags))
248 tag = notmuch_tags_get (tags);
249 /* Is message excluded? */
250 for (notmuch_string_node_t *term = exclude_terms->head;
254 /* Check for an empty string, and then ignore initial 'K'. */
255 if (*(term->string) && strcmp(tag, (term->string + 1)) == 0) {
256 message_excluded = TRUE;
263 if (message_excluded && omit_exclude == NOTMUCH_EXCLUDE_ALL)
266 _notmuch_message_list_add_message (thread->message_list,
267 talloc_steal (thread, message));
268 thread->total_messages++;
270 g_hash_table_insert (thread->message_hash,
271 xstrdup (notmuch_message_get_message_id (message)),
274 from = notmuch_message_get_header (message, "from");
276 list = internet_address_list_parse_string (from);
279 address = internet_address_list_get_address (list, 0);
281 author = internet_address_get_name (address);
282 /* We treat quoted empty names as if they were empty. */
283 if (author == NULL || author[0] == '\0') {
284 InternetAddressMailbox *mailbox;
285 mailbox = INTERNET_ADDRESS_MAILBOX (address);
286 author = internet_address_mailbox_get_addr (mailbox);
288 clean_author = _thread_cleanup_author (thread, author, from);
289 _thread_add_author (thread, clean_author);
290 _notmuch_message_set_author (message, clean_author);
292 g_object_unref (G_OBJECT (list));
295 if (! thread->subject) {
297 subject = notmuch_message_get_header (message, "subject");
298 thread->subject = talloc_strdup (thread, subject ? subject : "");
301 for (tags = notmuch_message_get_tags (message);
302 notmuch_tags_valid (tags);
303 notmuch_tags_move_to_next (tags))
305 tag = notmuch_tags_get (tags);
306 g_hash_table_insert (thread->tags, xstrdup (tag), NULL);
309 /* Mark excluded messages. */
310 if (message_excluded)
311 notmuch_message_set_flag (message, NOTMUCH_MESSAGE_FLAG_EXCLUDED, TRUE);
315 _thread_set_subject_from_message (notmuch_thread_t *thread,
316 notmuch_message_t *message)
319 const char *cleaned_subject;
321 subject = notmuch_message_get_header (message, "subject");
325 if ((strncasecmp (subject, "Re: ", 4) == 0) ||
326 (strncasecmp (subject, "Aw: ", 4) == 0) ||
327 (strncasecmp (subject, "Vs: ", 4) == 0) ||
328 (strncasecmp (subject, "Sv: ", 4) == 0)) {
330 cleaned_subject = talloc_strndup (thread,
332 strlen(subject) - 4);
334 cleaned_subject = talloc_strdup (thread, subject);
337 if (! EMPTY_STRING(cleaned_subject)) {
339 talloc_free (thread->subject);
341 thread->subject = talloc_strdup (thread, cleaned_subject);
345 /* Add a message to this thread which is known to match the original
346 * search specification. The 'sort' parameter controls whether the
347 * oldest or newest matching subject is applied to the thread as a
350 _thread_add_matched_message (notmuch_thread_t *thread,
351 notmuch_message_t *message,
355 notmuch_message_t *hashed_message;
357 date = notmuch_message_get_date (message);
359 if (date < thread->oldest || ! thread->matched_messages) {
360 thread->oldest = date;
361 if (sort == NOTMUCH_SORT_OLDEST_FIRST)
362 _thread_set_subject_from_message (thread, message);
365 if (date > thread->newest || ! thread->matched_messages) {
366 thread->newest = date;
367 const char *cur_subject = notmuch_thread_get_subject(thread);
368 if (sort != NOTMUCH_SORT_OLDEST_FIRST || EMPTY_STRING(cur_subject))
369 _thread_set_subject_from_message (thread, message);
372 if (!notmuch_message_get_flag (message, NOTMUCH_MESSAGE_FLAG_EXCLUDED))
373 thread->matched_messages++;
375 if (g_hash_table_lookup_extended (thread->message_hash,
376 notmuch_message_get_message_id (message), NULL,
377 (void **) &hashed_message)) {
378 notmuch_message_set_flag (hashed_message,
379 NOTMUCH_MESSAGE_FLAG_MATCH, 1);
382 _thread_add_matched_author (thread, _notmuch_message_get_author (hashed_message));
386 _resolve_thread_relationships (notmuch_thread_t *thread)
388 notmuch_message_node_t *node;
389 notmuch_message_t *message, *parent;
390 const char *in_reply_to;
392 for (node = thread->message_list->head; node; node = node->next) {
393 message = node->message;
394 in_reply_to = _notmuch_message_get_in_reply_to (message);
395 if (in_reply_to && strlen (in_reply_to) &&
396 g_hash_table_lookup_extended (thread->message_hash,
399 _notmuch_message_add_reply (parent, message);
401 _notmuch_message_list_add_message (thread->toplevel_list, message);
404 /* XXX: After scanning through the entire list looking for parents
405 * via "In-Reply-To", we should do a second pass that looks at the
406 * list of messages IDs in the "References" header instead. (And
407 * for this the parent would be the "deepest" message of all the
408 * messages found in the "References" list.)
410 * Doing this will allow messages and sub-threads to be positioned
411 * correctly in the thread even when an intermediate message is
412 * missing from the thread.
416 /* Create a new notmuch_thread_t object by finding the thread
417 * containing the message with the given doc ID, treating any messages
418 * contained in match_set as "matched". Remove all messages in the
419 * thread from match_set.
421 * Creating the thread will perform a database search to get all
422 * messages belonging to the thread and will get the first subject
423 * line, the total count of messages, and all authors in the thread.
424 * Each message in the thread is checked against match_set to allow
425 * for a separate count of matched messages, and to allow a viewer to
426 * display these messages differently.
428 * Here, 'ctx' is talloc context for the resulting thread object.
430 * This function returns NULL in the case of any error.
433 _notmuch_thread_create (void *ctx,
434 notmuch_database_t *notmuch,
435 unsigned int seed_doc_id,
436 notmuch_doc_id_set_t *match_set,
437 notmuch_string_list_t *exclude_terms,
438 notmuch_exclude_t omit_excluded,
441 void *local = talloc_new (ctx);
442 notmuch_thread_t *thread = NULL;
443 notmuch_message_t *seed_message;
444 const char *thread_id;
445 char *thread_id_query_string;
446 notmuch_query_t *thread_id_query;
448 notmuch_messages_t *messages;
449 notmuch_message_t *message;
450 notmuch_status_t status;
452 seed_message = _notmuch_message_create (local, notmuch, seed_doc_id, NULL);
454 INTERNAL_ERROR ("Thread seed message %u does not exist", seed_doc_id);
456 thread_id = notmuch_message_get_thread_id (seed_message);
457 thread_id_query_string = talloc_asprintf (local, "thread:%s", thread_id);
458 if (unlikely (thread_id_query_string == NULL))
461 thread_id_query = talloc_steal (
462 local, notmuch_query_create (notmuch, thread_id_query_string));
463 if (unlikely (thread_id_query == NULL))
466 thread = talloc (local, notmuch_thread_t);
467 if (unlikely (thread == NULL))
470 talloc_set_destructor (thread, _notmuch_thread_destructor);
472 thread->notmuch = notmuch;
473 thread->thread_id = talloc_strdup (thread, thread_id);
474 thread->subject = NULL;
475 thread->authors_hash = g_hash_table_new_full (g_str_hash, g_str_equal,
477 thread->authors_array = g_ptr_array_new ();
478 thread->matched_authors_hash = g_hash_table_new_full (g_str_hash,
481 thread->matched_authors_array = g_ptr_array_new ();
482 thread->authors = NULL;
483 thread->tags = g_hash_table_new_full (g_str_hash, g_str_equal,
486 thread->message_list = _notmuch_message_list_create (thread);
487 thread->toplevel_list = _notmuch_message_list_create (thread);
488 if (unlikely (thread->message_list == NULL ||
489 thread->toplevel_list == NULL)) {
494 thread->message_hash = g_hash_table_new_full (g_str_hash, g_str_equal,
497 thread->total_messages = 0;
498 thread->matched_messages = 0;
502 /* We use oldest-first order unconditionally here to obtain the
503 * proper author ordering for the thread. The 'sort' parameter
504 * passed to this function is used only to indicate whether the
505 * oldest or newest subject is desired. */
506 notmuch_query_set_sort (thread_id_query, NOTMUCH_SORT_OLDEST_FIRST);
508 status = notmuch_query_search_messages_st (thread_id_query, &messages);
513 notmuch_messages_valid (messages);
514 notmuch_messages_move_to_next (messages))
518 message = notmuch_messages_get (messages);
519 doc_id = _notmuch_message_get_doc_id (message);
520 if (doc_id == seed_doc_id)
521 message = seed_message;
523 _thread_add_message (thread, message, exclude_terms, omit_excluded);
525 if ( _notmuch_doc_id_set_contains (match_set, doc_id)) {
526 _notmuch_doc_id_set_remove (match_set, doc_id);
527 _thread_add_matched_message (thread, message, sort);
530 _notmuch_message_close (message);
533 _resolve_thread_authors_string (thread);
535 _resolve_thread_relationships (thread);
537 /* Commit to returning thread. */
538 (void) talloc_steal (ctx, thread);
546 notmuch_thread_get_toplevel_messages (notmuch_thread_t *thread)
548 return _notmuch_messages_create (thread->toplevel_list);
552 notmuch_thread_get_messages (notmuch_thread_t *thread)
554 return _notmuch_messages_create (thread->message_list);
558 notmuch_thread_get_thread_id (notmuch_thread_t *thread)
560 return thread->thread_id;
564 notmuch_thread_get_total_messages (notmuch_thread_t *thread)
566 return thread->total_messages;
570 notmuch_thread_get_matched_messages (notmuch_thread_t *thread)
572 return thread->matched_messages;
576 notmuch_thread_get_authors (notmuch_thread_t *thread)
578 return thread->authors;
582 notmuch_thread_get_subject (notmuch_thread_t *thread)
584 return thread->subject;
588 notmuch_thread_get_oldest_date (notmuch_thread_t *thread)
590 return thread->oldest;
594 notmuch_thread_get_newest_date (notmuch_thread_t *thread)
596 return thread->newest;
600 notmuch_thread_get_tags (notmuch_thread_t *thread)
602 notmuch_string_list_t *tags;
605 tags = _notmuch_string_list_create (thread);
606 if (unlikely (tags == NULL))
609 keys = g_hash_table_get_keys (thread->tags);
611 for (l = keys; l; l = l->next)
612 _notmuch_string_list_append (tags, (char *) l->data);
616 _notmuch_string_list_sort (tags);
618 return _notmuch_tags_create (thread, tags);
622 notmuch_thread_destroy (notmuch_thread_t *thread)
624 talloc_free (thread);