1 /* database.cc - The database interfaces of the notmuch mail library
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 "database-private.h"
27 #include <glib.h> /* g_free, GPtrArray, GHashTable */
31 #define ARRAY_SIZE(arr) (sizeof (arr) / sizeof (arr[0]))
38 /* Here's the current schema for our database:
40 * We currently have two different types of documents: mail and timestamps.
44 * A mail document is associated with a particular email message file
45 * on disk. It is indexed with the following prefixed terms:
47 * Single terms of given prefix:
51 * id: Unique ID of mail, (from Message-ID header or generated
52 * as "notmuch-sha1-<sha1_sum_of_entire_file>.
54 * thread: The ID of the thread to which the mail belongs
56 * Multiple terms of given prefix:
58 * ref: All unresolved message IDs from In-Reply-To and
59 * References headers in the message. (Once a referenced
60 * message is added to the database and the thread IDs
61 * are linked the corresponding "ref" term is dropped
62 * from the message document.)
64 * tag: Any tags associated with this message by the user.
66 * A mail document also has two values:
68 * TIMESTAMP: The time_t value corresponding to the message's
71 * MESSAGE_ID: The unique ID of the mail mess (see "id" above)
75 * A timestamp document is used by a client of the notmuch library to
76 * maintain data necessary to allow for efficient polling of mail
77 * directories. The notmuch library does no interpretation of
78 * timestamps, but merely allows the user to store and retrieve
79 * timestamps as name/value pairs.
81 * The timestamp document is indexed with a single prefixed term:
83 * timestamp: The user's key value (likely a directory name)
85 * and has a single value:
87 * TIMETAMPS: The time_t value from the user.
90 /* With these prefix values we follow the conventions published here:
92 * http://xapian.org/docs/omega/termprefixes.html
94 * as much as makes sense. Note that I took some liberty in matching
95 * the reserved prefix values to notmuch concepts, (for example, 'G'
96 * is documented as "newsGroup (or similar entity - e.g. a web forum
97 * name)", for which I think the thread is the closest analogue in
98 * notmuch. This in spite of the fact that we will eventually be
99 * storing mailing-list messages where 'G' for "mailing list name"
100 * might be even a closer analogue. I'm treating the single-character
101 * prefixes preferentially for core notmuch concepts (which will be
102 * nearly universal to all mail messages).
105 prefix_t BOOLEAN_PREFIX_INTERNAL[] = {
107 { "ref", "XREFERENCE" },
108 { "timestamp", "XTIMESTAMP" },
111 prefix_t BOOLEAN_PREFIX_EXTERNAL[] = {
117 prefix_t PROBABILISTIC_PREFIX[]= {
120 { "attachment", "XATTACHMENT" },
121 { "subject", "XSUBJECT"}
125 _internal_error (const char *format, ...)
129 va_start (va_args, format);
131 vfprintf (stderr, format, va_args);
139 _find_prefix (const char *name)
143 for (i = 0; i < ARRAY_SIZE (BOOLEAN_PREFIX_INTERNAL); i++)
144 if (strcmp (name, BOOLEAN_PREFIX_INTERNAL[i].name) == 0)
145 return BOOLEAN_PREFIX_INTERNAL[i].prefix;
147 for (i = 0; i < ARRAY_SIZE (BOOLEAN_PREFIX_EXTERNAL); i++)
148 if (strcmp (name, BOOLEAN_PREFIX_EXTERNAL[i].name) == 0)
149 return BOOLEAN_PREFIX_EXTERNAL[i].prefix;
151 for (i = 0; i < ARRAY_SIZE (PROBABILISTIC_PREFIX); i++)
152 if (strcmp (name, PROBABILISTIC_PREFIX[i].name) == 0)
153 return PROBABILISTIC_PREFIX[i].prefix;
155 INTERNAL_ERROR ("No prefix exists for '%s'\n", name);
161 notmuch_status_to_string (notmuch_status_t status)
164 case NOTMUCH_STATUS_SUCCESS:
165 return "No error occurred";
166 case NOTMUCH_STATUS_OUT_OF_MEMORY:
167 return "Out of memory";
168 case NOTMUCH_STATUS_XAPIAN_EXCEPTION:
169 return "A Xapian exception occurred";
170 case NOTMUCH_STATUS_FILE_ERROR:
171 return "Something went wrong trying to read or write a file";
172 case NOTMUCH_STATUS_FILE_NOT_EMAIL:
173 return "File is not an email";
174 case NOTMUCH_STATUS_DUPLICATE_MESSAGE_ID:
175 return "Message ID is identical to a message in database";
176 case NOTMUCH_STATUS_NULL_POINTER:
177 return "Erroneous NULL pointer";
178 case NOTMUCH_STATUS_TAG_TOO_LONG:
179 return "Tag value is too long (exceeds NOTMUCH_TAG_MAX)";
180 case NOTMUCH_STATUS_UNBALANCED_FREEZE_THAW:
181 return "Unblanced number of calls to notmuch_message_freeze/thaw";
183 case NOTMUCH_STATUS_LAST_STATUS:
184 return "Unknown error status value";
189 find_doc_ids (notmuch_database_t *notmuch,
190 const char *prefix_name,
192 Xapian::PostingIterator *begin,
193 Xapian::PostingIterator *end)
195 Xapian::PostingIterator i;
198 term = talloc_asprintf (notmuch, "%s%s",
199 _find_prefix (prefix_name), value);
201 *begin = notmuch->xapian_db->postlist_begin (term);
203 *end = notmuch->xapian_db->postlist_end (term);
208 static notmuch_private_status_t
209 find_unique_doc_id (notmuch_database_t *notmuch,
210 const char *prefix_name,
212 unsigned int *doc_id)
214 Xapian::PostingIterator i, end;
216 find_doc_ids (notmuch, prefix_name, value, &i, &end);
220 return NOTMUCH_PRIVATE_STATUS_NO_DOCUMENT_FOUND;
223 return NOTMUCH_PRIVATE_STATUS_SUCCESS;
227 static Xapian::Document
228 find_document_for_doc_id (notmuch_database_t *notmuch, unsigned doc_id)
230 return notmuch->xapian_db->get_document (doc_id);
233 static notmuch_private_status_t
234 find_unique_document (notmuch_database_t *notmuch,
235 const char *prefix_name,
237 Xapian::Document *document,
238 unsigned int *doc_id)
240 notmuch_private_status_t status;
242 status = find_unique_doc_id (notmuch, prefix_name, value, doc_id);
245 *document = Xapian::Document ();
249 *document = find_document_for_doc_id (notmuch, *doc_id);
250 return NOTMUCH_PRIVATE_STATUS_SUCCESS;
254 notmuch_database_find_message (notmuch_database_t *notmuch,
255 const char *message_id)
257 notmuch_private_status_t status;
260 status = find_unique_doc_id (notmuch, "id", message_id, &doc_id);
262 if (status == NOTMUCH_PRIVATE_STATUS_NO_DOCUMENT_FOUND)
265 return _notmuch_message_create (notmuch, notmuch, doc_id, NULL);
268 /* Advance 'str' past any whitespace or RFC 822 comments. A comment is
269 * a (potentially nested) parenthesized sequence with '\' used to
270 * escape any character (including parentheses).
272 * If the sequence to be skipped continues to the end of the string,
273 * then 'str' will be left pointing at the final terminating '\0'
277 skip_space_and_comments (const char **str)
282 while (*s && (isspace (*s) || *s == '(')) {
283 while (*s && isspace (*s))
288 while (*s && nesting) {
304 /* Parse an RFC 822 message-id, discarding whitespace, any RFC 822
305 * comments, and the '<' and '>' delimeters.
307 * If not NULL, then *next will be made to point to the first character
308 * not parsed, (possibly pointing to the final '\0' terminator.
310 * Returns a newly talloc'ed string belonging to 'ctx'.
312 * Returns NULL if there is any error parsing the message-id. */
314 parse_message_id (void *ctx, const char *message_id, const char **next)
319 if (message_id == NULL)
324 skip_space_and_comments (&s);
326 /* Skip any unstructured text as well. */
327 while (*s && *s != '<')
338 skip_space_and_comments (&s);
341 while (*end && *end != '>')
350 if (end > s && *end == '>')
355 result = talloc_strndup (ctx, s, end - s + 1);
357 /* Finally, collapse any whitespace that is within the message-id
363 for (r = result, len = strlen (r); *r; r++, len--)
364 if (*r == ' ' || *r == '\t')
365 memmove (r, r+1, len);
371 /* Parse a References header value, putting a (talloc'ed under 'ctx')
372 * copy of each referenced message-id into 'hash'. */
374 parse_references (void *ctx,
384 ref = parse_message_id (ctx, refs, &refs);
387 g_hash_table_insert (hash, ref, NULL);
392 notmuch_database_default_path (void)
396 if (getenv ("NOTMUCH_BASE"))
397 return strdup (getenv ("NOTMUCH_BASE"));
399 if (asprintf (&path, "%s/mail", getenv ("HOME")) == -1) {
400 fprintf (stderr, "Out of memory.\n");
408 notmuch_database_create (const char *path)
410 notmuch_database_t *notmuch = NULL;
411 char *notmuch_path = NULL;
414 char *local_path = NULL;
417 path = local_path = notmuch_database_default_path ();
419 err = stat (path, &st);
421 fprintf (stderr, "Error: Cannot create database at %s: %s.\n",
422 path, strerror (errno));
426 if (! S_ISDIR (st.st_mode)) {
427 fprintf (stderr, "Error: Cannot create database at %s: Not a directory.\n",
432 notmuch_path = talloc_asprintf (NULL, "%s/%s", path, ".notmuch");
434 err = mkdir (notmuch_path, 0755);
437 fprintf (stderr, "Error: Cannot create directory %s: %s.\n",
438 notmuch_path, strerror (errno));
442 notmuch = notmuch_database_open (path);
446 talloc_free (notmuch_path);
454 notmuch_database_open (const char *path)
456 notmuch_database_t *notmuch = NULL;
457 char *notmuch_path = NULL, *xapian_path = NULL;
460 char *local_path = NULL;
464 path = local_path = notmuch_database_default_path ();
466 if (asprintf (¬much_path, "%s/%s", path, ".notmuch") == -1) {
468 fprintf (stderr, "Out of memory\n");
472 err = stat (notmuch_path, &st);
474 fprintf (stderr, "Error opening database at %s: %s\n",
475 notmuch_path, strerror (errno));
479 if (asprintf (&xapian_path, "%s/%s", notmuch_path, "xapian") == -1) {
481 fprintf (stderr, "Out of memory\n");
485 notmuch = talloc (NULL, notmuch_database_t);
486 notmuch->path = talloc_strdup (notmuch, path);
489 notmuch->xapian_db = new Xapian::WritableDatabase (xapian_path,
490 Xapian::DB_CREATE_OR_OPEN);
491 notmuch->query_parser = new Xapian::QueryParser;
492 notmuch->term_gen = new Xapian::TermGenerator;
493 notmuch->term_gen->set_stemmer (Xapian::Stem ("english"));
495 notmuch->query_parser->set_default_op (Xapian::Query::OP_AND);
496 notmuch->query_parser->set_database (*notmuch->xapian_db);
497 notmuch->query_parser->set_stemmer (Xapian::Stem ("english"));
498 notmuch->query_parser->set_stemming_strategy (Xapian::QueryParser::STEM_SOME);
500 for (i = 0; i < ARRAY_SIZE (BOOLEAN_PREFIX_EXTERNAL); i++) {
501 prefix_t *prefix = &BOOLEAN_PREFIX_EXTERNAL[i];
502 notmuch->query_parser->add_boolean_prefix (prefix->name,
506 for (i = 0; i < ARRAY_SIZE (PROBABILISTIC_PREFIX); i++) {
507 prefix_t *prefix = &PROBABILISTIC_PREFIX[i];
508 notmuch->query_parser->add_prefix (prefix->name, prefix->prefix);
510 } catch (const Xapian::Error &error) {
511 fprintf (stderr, "A Xapian exception occurred: %s\n",
512 error.get_msg().c_str());
528 notmuch_database_close (notmuch_database_t *notmuch)
530 notmuch->xapian_db->flush ();
532 delete notmuch->term_gen;
533 delete notmuch->query_parser;
534 delete notmuch->xapian_db;
535 talloc_free (notmuch);
539 notmuch_database_get_path (notmuch_database_t *notmuch)
541 return notmuch->path;
544 static notmuch_private_status_t
545 find_timestamp_document (notmuch_database_t *notmuch, const char *db_key,
546 Xapian::Document *doc, unsigned int *doc_id)
548 return find_unique_document (notmuch, "timestamp", db_key, doc, doc_id);
551 /* We allow the user to use arbitrarily long keys for timestamps,
552 * (they're for filesystem paths after all, which have no limit we
553 * know about). But we have a term-length limit. So if we exceed that,
554 * we'll use the SHA-1 of the user's key as the actual key for
555 * constructing a database term.
557 * Caution: This function returns a newly allocated string which the
558 * caller should free() when finished.
561 timestamp_db_key (const char *key)
563 int term_len = strlen (_find_prefix ("timestamp")) + strlen (key);
565 if (term_len > NOTMUCH_TERM_MAX)
566 return notmuch_sha1_of_string (key);
572 notmuch_database_set_timestamp (notmuch_database_t *notmuch,
573 const char *key, time_t timestamp)
575 Xapian::Document doc;
577 notmuch_private_status_t status;
578 notmuch_status_t ret = NOTMUCH_STATUS_SUCCESS;
581 db_key = timestamp_db_key (key);
584 status = find_timestamp_document (notmuch, db_key, &doc, &doc_id);
586 doc.add_value (NOTMUCH_VALUE_TIMESTAMP,
587 Xapian::sortable_serialise (timestamp));
589 if (status == NOTMUCH_PRIVATE_STATUS_NO_DOCUMENT_FOUND) {
590 char *term = talloc_asprintf (NULL, "%s%s",
591 _find_prefix ("timestamp"), db_key);
595 notmuch->xapian_db->add_document (doc);
597 notmuch->xapian_db->replace_document (doc_id, doc);
600 } catch (Xapian::Error &error) {
601 fprintf (stderr, "A Xapian exception occurred: %s.\n",
602 error.get_msg().c_str());
603 ret = NOTMUCH_STATUS_XAPIAN_EXCEPTION;
613 notmuch_database_get_timestamp (notmuch_database_t *notmuch, const char *key)
615 Xapian::Document doc;
617 notmuch_private_status_t status;
621 db_key = timestamp_db_key (key);
624 status = find_timestamp_document (notmuch, db_key, &doc, &doc_id);
626 if (status == NOTMUCH_PRIVATE_STATUS_NO_DOCUMENT_FOUND)
629 ret = Xapian::sortable_unserialise (doc.get_value (NOTMUCH_VALUE_TIMESTAMP));
630 } catch (Xapian::Error &error) {
641 /* Find the thread ID to which the message with 'message_id' belongs.
643 * Returns NULL if no message with message ID 'message_id' is in the
646 * Otherwise, returns a newly talloced string belonging to 'ctx'.
649 _resolve_message_id_to_thread_id (notmuch_database_t *notmuch,
651 const char *message_id)
653 notmuch_message_t *message;
654 const char *ret = NULL;
656 message = notmuch_database_find_message (notmuch, message_id);
660 ret = talloc_steal (ctx, notmuch_message_get_thread_id (message));
664 notmuch_message_destroy (message);
669 static notmuch_status_t
670 _merge_threads (notmuch_database_t *notmuch,
671 const char *winner_thread_id,
672 const char *loser_thread_id)
674 Xapian::PostingIterator loser, loser_end;
675 notmuch_message_t *message = NULL;
676 notmuch_private_status_t private_status;
677 notmuch_status_t ret = NOTMUCH_STATUS_SUCCESS;
679 find_doc_ids (notmuch, "thread", loser_thread_id, &loser, &loser_end);
681 for ( ; loser != loser_end; loser++) {
682 message = _notmuch_message_create (notmuch, notmuch,
683 *loser, &private_status);
684 if (message == NULL) {
685 ret = COERCE_STATUS (private_status,
686 "Cannot find document for doc_id from query");
690 _notmuch_message_remove_term (message, "thread", loser_thread_id);
691 _notmuch_message_add_term (message, "thread", winner_thread_id);
692 _notmuch_message_sync (message);
694 notmuch_message_destroy (message);
700 notmuch_message_destroy (message);
706 _my_talloc_free_for_g_hash (void *ptr)
711 static notmuch_status_t
712 _notmuch_database_link_message_to_parents (notmuch_database_t *notmuch,
713 notmuch_message_t *message,
714 notmuch_message_file_t *message_file,
715 const char **thread_id)
717 GHashTable *parents = NULL;
718 const char *refs, *in_reply_to;
719 GList *l, *keys = NULL;
720 notmuch_status_t ret = NOTMUCH_STATUS_SUCCESS;
722 parents = g_hash_table_new_full (g_str_hash, g_str_equal,
723 _my_talloc_free_for_g_hash, NULL);
725 refs = notmuch_message_file_get_header (message_file, "references");
726 parse_references (message, parents, refs);
728 in_reply_to = notmuch_message_file_get_header (message_file, "in-reply-to");
729 parse_references (message, parents, in_reply_to);
731 keys = g_hash_table_get_keys (parents);
732 for (l = keys; l; l = l->next) {
733 char *parent_message_id;
734 const char *parent_thread_id;
736 parent_message_id = (char *) l->data;
737 parent_thread_id = _resolve_message_id_to_thread_id (notmuch,
741 if (parent_thread_id == NULL) {
742 _notmuch_message_add_term (message, "ref", parent_message_id);
744 if (*thread_id == NULL) {
745 *thread_id = talloc_strdup (message, parent_thread_id);
746 _notmuch_message_add_term (message, "thread", *thread_id);
747 } else if (strcmp (*thread_id, parent_thread_id)) {
748 ret = _merge_threads (notmuch, *thread_id, parent_thread_id);
759 g_hash_table_unref (parents);
764 static notmuch_status_t
765 _notmuch_database_link_message_to_children (notmuch_database_t *notmuch,
766 notmuch_message_t *message,
767 const char **thread_id)
769 const char *message_id = notmuch_message_get_message_id (message);
770 Xapian::PostingIterator child, children_end;
771 notmuch_message_t *child_message = NULL;
772 const char *child_thread_id;
773 notmuch_status_t ret = NOTMUCH_STATUS_SUCCESS;
774 notmuch_private_status_t private_status;
776 find_doc_ids (notmuch, "ref", message_id, &child, &children_end);
778 for ( ; child != children_end; child++) {
780 child_message = _notmuch_message_create (message, notmuch,
781 *child, &private_status);
782 if (child_message == NULL) {
783 ret = COERCE_STATUS (private_status,
784 "Cannot find document for doc_id from query");
788 child_thread_id = notmuch_message_get_thread_id (child_message);
789 if (*thread_id == NULL) {
790 *thread_id = talloc_strdup (message, child_thread_id);
791 _notmuch_message_add_term (message, "thread", *thread_id);
792 } else if (strcmp (*thread_id, child_thread_id)) {
793 _notmuch_message_remove_term (child_message, "ref",
795 _notmuch_message_sync (child_message);
796 ret = _merge_threads (notmuch, *thread_id, child_thread_id);
801 notmuch_message_destroy (child_message);
802 child_message = NULL;
807 notmuch_message_destroy (child_message);
812 /* Given a (mostly empty) 'message' and its corresponding
813 * 'message_file' link it to existing threads in the database.
815 * We first looke at 'message_file' and its link-relevant headers
816 * (References and In-Reply-To) for message IDs. We also look in the
817 * database for existing message that reference 'message'.p
819 * The end result is to call _notmuch_message_add_thread_id with one
820 * or more thread IDs to which this message belongs, (including
821 * generating a new thread ID if necessary if the message doesn't
822 * connect to any existing threads).
824 static notmuch_status_t
825 _notmuch_database_link_message (notmuch_database_t *notmuch,
826 notmuch_message_t *message,
827 notmuch_message_file_t *message_file)
829 notmuch_status_t status;
830 const char *thread_id = NULL;
832 status = _notmuch_database_link_message_to_parents (notmuch, message,
838 status = _notmuch_database_link_message_to_children (notmuch, message,
843 if (thread_id == NULL)
844 _notmuch_message_ensure_thread_id (message);
846 return NOTMUCH_STATUS_SUCCESS;
850 notmuch_database_add_message (notmuch_database_t *notmuch,
851 const char *filename,
852 notmuch_message_t **message_ret)
854 notmuch_message_file_t *message_file;
855 notmuch_message_t *message;
856 notmuch_status_t ret = NOTMUCH_STATUS_SUCCESS;
858 const char *date, *header;
859 const char *from, *to, *subject, *old_filename;
865 message_file = notmuch_message_file_open (filename);
866 if (message_file == NULL) {
867 ret = NOTMUCH_STATUS_FILE_ERROR;
871 notmuch_message_file_restrict_headers (message_file,
882 /* The first order of business is to find/create a message ID. */
884 header = notmuch_message_file_get_header (message_file, "message-id");
886 message_id = parse_message_id (message_file, header, NULL);
887 /* So the header value isn't RFC-compliant, but it's
888 * better than no message-id at all. */
889 if (message_id == NULL)
890 message_id = talloc_strdup (message_file, header);
892 /* No message-id at all, let's generate one by taking a
893 * hash over the file's contents. */
894 char *sha1 = notmuch_sha1_of_file (filename);
896 /* If that failed too, something is really wrong. Give up. */
898 ret = NOTMUCH_STATUS_FILE_ERROR;
902 message_id = talloc_asprintf (message_file,
903 "notmuch-sha1-%s", sha1);
907 /* Now that we have a message ID, we get a message object,
908 * (which may or may not reference an existing document in the
911 /* Use NULL for owner since we want to free this locally. */
912 message = _notmuch_message_create_for_message_id (NULL,
917 talloc_free (message_id);
922 /* Has a message previously been added with the same ID? */
923 old_filename = notmuch_message_get_filename (message);
924 if (old_filename && strlen (old_filename)) {
925 ret = NOTMUCH_STATUS_DUPLICATE_MESSAGE_ID;
928 _notmuch_message_set_filename (message, filename);
929 _notmuch_message_add_term (message, "type", "mail");
932 ret = _notmuch_database_link_message (notmuch, message, message_file);
936 date = notmuch_message_file_get_header (message_file, "date");
937 _notmuch_message_set_date (message, date);
939 from = notmuch_message_file_get_header (message_file, "from");
940 subject = notmuch_message_file_get_header (message_file, "subject");
941 to = notmuch_message_file_get_header (message_file, "to");
947 ret = NOTMUCH_STATUS_FILE_NOT_EMAIL;
951 _notmuch_message_index_file (message, filename);
953 _notmuch_message_sync (message);
954 } catch (const Xapian::Error &error) {
955 fprintf (stderr, "A Xapian exception occurred: %s.\n",
956 error.get_msg().c_str());
957 ret = NOTMUCH_STATUS_XAPIAN_EXCEPTION;
963 if (ret == NOTMUCH_STATUS_SUCCESS && message_ret)
964 *message_ret = message;
966 notmuch_message_destroy (message);
970 notmuch_message_file_close (message_file);