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 * TIMESTAMP: 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 { "replyto", "XREPLYTO" },
109 { "timestamp", "XTIMESTAMP" },
112 prefix_t BOOLEAN_PREFIX_EXTERNAL[] = {
118 prefix_t PROBABILISTIC_PREFIX[]= {
121 { "attachment", "XATTACHMENT" },
122 { "subject", "XSUBJECT"}
126 _internal_error (const char *format, ...)
130 va_start (va_args, format);
132 fprintf (stderr, "Internal error: ");
133 vfprintf (stderr, format, va_args);
141 _find_prefix (const char *name)
145 for (i = 0; i < ARRAY_SIZE (BOOLEAN_PREFIX_INTERNAL); i++)
146 if (strcmp (name, BOOLEAN_PREFIX_INTERNAL[i].name) == 0)
147 return BOOLEAN_PREFIX_INTERNAL[i].prefix;
149 for (i = 0; i < ARRAY_SIZE (BOOLEAN_PREFIX_EXTERNAL); i++)
150 if (strcmp (name, BOOLEAN_PREFIX_EXTERNAL[i].name) == 0)
151 return BOOLEAN_PREFIX_EXTERNAL[i].prefix;
153 for (i = 0; i < ARRAY_SIZE (PROBABILISTIC_PREFIX); i++)
154 if (strcmp (name, PROBABILISTIC_PREFIX[i].name) == 0)
155 return PROBABILISTIC_PREFIX[i].prefix;
157 INTERNAL_ERROR ("No prefix exists for '%s'\n", name);
163 notmuch_status_to_string (notmuch_status_t status)
166 case NOTMUCH_STATUS_SUCCESS:
167 return "No error occurred";
168 case NOTMUCH_STATUS_OUT_OF_MEMORY:
169 return "Out of memory";
170 case NOTMUCH_STATUS_XAPIAN_EXCEPTION:
171 return "A Xapian exception occurred";
172 case NOTMUCH_STATUS_FILE_ERROR:
173 return "Something went wrong trying to read or write a file";
174 case NOTMUCH_STATUS_FILE_NOT_EMAIL:
175 return "File is not an email";
176 case NOTMUCH_STATUS_DUPLICATE_MESSAGE_ID:
177 return "Message ID is identical to a message in database";
178 case NOTMUCH_STATUS_NULL_POINTER:
179 return "Erroneous NULL pointer";
180 case NOTMUCH_STATUS_TAG_TOO_LONG:
181 return "Tag value is too long (exceeds NOTMUCH_TAG_MAX)";
182 case NOTMUCH_STATUS_UNBALANCED_FREEZE_THAW:
183 return "Unblanced number of calls to notmuch_message_freeze/thaw";
185 case NOTMUCH_STATUS_LAST_STATUS:
186 return "Unknown error status value";
191 find_doc_ids (notmuch_database_t *notmuch,
192 const char *prefix_name,
194 Xapian::PostingIterator *begin,
195 Xapian::PostingIterator *end)
197 Xapian::PostingIterator i;
200 term = talloc_asprintf (notmuch, "%s%s",
201 _find_prefix (prefix_name), value);
203 *begin = notmuch->xapian_db->postlist_begin (term);
205 *end = notmuch->xapian_db->postlist_end (term);
210 static notmuch_private_status_t
211 find_unique_doc_id (notmuch_database_t *notmuch,
212 const char *prefix_name,
214 unsigned int *doc_id)
216 Xapian::PostingIterator i, end;
218 find_doc_ids (notmuch, prefix_name, value, &i, &end);
222 return NOTMUCH_PRIVATE_STATUS_NO_DOCUMENT_FOUND;
225 return NOTMUCH_PRIVATE_STATUS_SUCCESS;
229 static Xapian::Document
230 find_document_for_doc_id (notmuch_database_t *notmuch, unsigned doc_id)
232 return notmuch->xapian_db->get_document (doc_id);
235 static notmuch_private_status_t
236 find_unique_document (notmuch_database_t *notmuch,
237 const char *prefix_name,
239 Xapian::Document *document,
240 unsigned int *doc_id)
242 notmuch_private_status_t status;
244 status = find_unique_doc_id (notmuch, prefix_name, value, doc_id);
247 *document = Xapian::Document ();
251 *document = find_document_for_doc_id (notmuch, *doc_id);
252 return NOTMUCH_PRIVATE_STATUS_SUCCESS;
256 notmuch_database_find_message (notmuch_database_t *notmuch,
257 const char *message_id)
259 notmuch_private_status_t status;
262 status = find_unique_doc_id (notmuch, "id", message_id, &doc_id);
264 if (status == NOTMUCH_PRIVATE_STATUS_NO_DOCUMENT_FOUND)
267 return _notmuch_message_create (notmuch, notmuch, doc_id, NULL);
270 /* Advance 'str' past any whitespace or RFC 822 comments. A comment is
271 * a (potentially nested) parenthesized sequence with '\' used to
272 * escape any character (including parentheses).
274 * If the sequence to be skipped continues to the end of the string,
275 * then 'str' will be left pointing at the final terminating '\0'
279 skip_space_and_comments (const char **str)
284 while (*s && (isspace (*s) || *s == '(')) {
285 while (*s && isspace (*s))
290 while (*s && nesting) {
306 /* Parse an RFC 822 message-id, discarding whitespace, any RFC 822
307 * comments, and the '<' and '>' delimeters.
309 * If not NULL, then *next will be made to point to the first character
310 * not parsed, (possibly pointing to the final '\0' terminator.
312 * Returns a newly talloc'ed string belonging to 'ctx'.
314 * Returns NULL if there is any error parsing the message-id. */
316 _parse_message_id (void *ctx, const char *message_id, const char **next)
321 if (message_id == NULL)
326 skip_space_and_comments (&s);
328 /* Skip any unstructured text as well. */
329 while (*s && *s != '<')
340 skip_space_and_comments (&s);
343 while (*end && *end != '>')
352 if (end > s && *end == '>')
357 result = talloc_strndup (ctx, s, end - s + 1);
359 /* Finally, collapse any whitespace that is within the message-id
365 for (r = result, len = strlen (r); *r; r++, len--)
366 if (*r == ' ' || *r == '\t')
367 memmove (r, r+1, len);
373 /* Parse a References header value, putting a (talloc'ed under 'ctx')
374 * copy of each referenced message-id into 'hash'.
376 * We explicitly avoid including any reference identical to
377 * 'message_id' in the result (to avoid mass confusion when a single
378 * message references itself cyclically---and yes, mail messages are
379 * not infrequent in the wild that do this---don't ask me why).
382 parse_references (void *ctx,
383 const char *message_id,
393 ref = _parse_message_id (ctx, refs, &refs);
395 if (ref && strcmp (ref, message_id))
396 g_hash_table_insert (hash, ref, NULL);
401 notmuch_database_create (const char *path)
403 notmuch_database_t *notmuch = NULL;
404 char *notmuch_path = NULL;
409 fprintf (stderr, "Error: Cannot create a database for a NULL path.\n");
413 err = stat (path, &st);
415 fprintf (stderr, "Error: Cannot create database at %s: %s.\n",
416 path, strerror (errno));
420 if (! S_ISDIR (st.st_mode)) {
421 fprintf (stderr, "Error: Cannot create database at %s: Not a directory.\n",
426 notmuch_path = talloc_asprintf (NULL, "%s/%s", path, ".notmuch");
428 err = mkdir (notmuch_path, 0755);
431 fprintf (stderr, "Error: Cannot create directory %s: %s.\n",
432 notmuch_path, strerror (errno));
436 notmuch = notmuch_database_open (path);
440 talloc_free (notmuch_path);
446 notmuch_database_open (const char *path)
448 notmuch_database_t *notmuch = NULL;
449 char *notmuch_path = NULL, *xapian_path = NULL;
454 if (asprintf (¬much_path, "%s/%s", path, ".notmuch") == -1) {
456 fprintf (stderr, "Out of memory\n");
460 err = stat (notmuch_path, &st);
462 fprintf (stderr, "Error opening database at %s: %s\n",
463 notmuch_path, strerror (errno));
467 if (asprintf (&xapian_path, "%s/%s", notmuch_path, "xapian") == -1) {
469 fprintf (stderr, "Out of memory\n");
473 notmuch = talloc (NULL, notmuch_database_t);
474 notmuch->path = talloc_strdup (notmuch, path);
476 if (notmuch->path[strlen (notmuch->path) - 1] == '/')
477 notmuch->path[strlen (notmuch->path) - 1] = '\0';
480 notmuch->xapian_db = new Xapian::WritableDatabase (xapian_path,
481 Xapian::DB_CREATE_OR_OPEN);
482 notmuch->query_parser = new Xapian::QueryParser;
483 notmuch->term_gen = new Xapian::TermGenerator;
484 notmuch->term_gen->set_stemmer (Xapian::Stem ("english"));
486 notmuch->query_parser->set_default_op (Xapian::Query::OP_AND);
487 notmuch->query_parser->set_database (*notmuch->xapian_db);
488 notmuch->query_parser->set_stemmer (Xapian::Stem ("english"));
489 notmuch->query_parser->set_stemming_strategy (Xapian::QueryParser::STEM_SOME);
491 for (i = 0; i < ARRAY_SIZE (BOOLEAN_PREFIX_EXTERNAL); i++) {
492 prefix_t *prefix = &BOOLEAN_PREFIX_EXTERNAL[i];
493 notmuch->query_parser->add_boolean_prefix (prefix->name,
497 for (i = 0; i < ARRAY_SIZE (PROBABILISTIC_PREFIX); i++) {
498 prefix_t *prefix = &PROBABILISTIC_PREFIX[i];
499 notmuch->query_parser->add_prefix (prefix->name, prefix->prefix);
501 } catch (const Xapian::Error &error) {
502 fprintf (stderr, "A Xapian exception occurred: %s\n",
503 error.get_msg().c_str());
517 notmuch_database_close (notmuch_database_t *notmuch)
519 notmuch->xapian_db->flush ();
521 delete notmuch->term_gen;
522 delete notmuch->query_parser;
523 delete notmuch->xapian_db;
524 talloc_free (notmuch);
528 notmuch_database_get_path (notmuch_database_t *notmuch)
530 return notmuch->path;
533 static notmuch_private_status_t
534 find_timestamp_document (notmuch_database_t *notmuch, const char *db_key,
535 Xapian::Document *doc, unsigned int *doc_id)
537 return find_unique_document (notmuch, "timestamp", db_key, doc, doc_id);
540 /* We allow the user to use arbitrarily long keys for timestamps,
541 * (they're for filesystem paths after all, which have no limit we
542 * know about). But we have a term-length limit. So if we exceed that,
543 * we'll use the SHA-1 of the user's key as the actual key for
544 * constructing a database term.
546 * Caution: This function returns a newly allocated string which the
547 * caller should free() when finished.
550 timestamp_db_key (const char *key)
552 int term_len = strlen (_find_prefix ("timestamp")) + strlen (key);
554 if (term_len > NOTMUCH_TERM_MAX)
555 return notmuch_sha1_of_string (key);
561 notmuch_database_set_timestamp (notmuch_database_t *notmuch,
562 const char *key, time_t timestamp)
564 Xapian::Document doc;
566 notmuch_private_status_t status;
567 notmuch_status_t ret = NOTMUCH_STATUS_SUCCESS;
570 db_key = timestamp_db_key (key);
573 status = find_timestamp_document (notmuch, db_key, &doc, &doc_id);
575 doc.add_value (NOTMUCH_VALUE_TIMESTAMP,
576 Xapian::sortable_serialise (timestamp));
578 if (status == NOTMUCH_PRIVATE_STATUS_NO_DOCUMENT_FOUND) {
579 char *term = talloc_asprintf (NULL, "%s%s",
580 _find_prefix ("timestamp"), db_key);
584 notmuch->xapian_db->add_document (doc);
586 notmuch->xapian_db->replace_document (doc_id, doc);
589 } catch (Xapian::Error &error) {
590 fprintf (stderr, "A Xapian exception occurred: %s.\n",
591 error.get_msg().c_str());
592 ret = NOTMUCH_STATUS_XAPIAN_EXCEPTION;
602 notmuch_database_get_timestamp (notmuch_database_t *notmuch, const char *key)
604 Xapian::Document doc;
606 notmuch_private_status_t status;
610 db_key = timestamp_db_key (key);
613 status = find_timestamp_document (notmuch, db_key, &doc, &doc_id);
615 if (status == NOTMUCH_PRIVATE_STATUS_NO_DOCUMENT_FOUND)
618 ret = Xapian::sortable_unserialise (doc.get_value (NOTMUCH_VALUE_TIMESTAMP));
619 } catch (Xapian::Error &error) {
630 /* Find the thread ID to which the message with 'message_id' belongs.
632 * Returns NULL if no message with message ID 'message_id' is in the
635 * Otherwise, returns a newly talloced string belonging to 'ctx'.
638 _resolve_message_id_to_thread_id (notmuch_database_t *notmuch,
640 const char *message_id)
642 notmuch_message_t *message;
643 const char *ret = NULL;
645 message = notmuch_database_find_message (notmuch, message_id);
649 ret = talloc_steal (ctx, notmuch_message_get_thread_id (message));
653 notmuch_message_destroy (message);
658 static notmuch_status_t
659 _merge_threads (notmuch_database_t *notmuch,
660 const char *winner_thread_id,
661 const char *loser_thread_id)
663 Xapian::PostingIterator loser, loser_end;
664 notmuch_message_t *message = NULL;
665 notmuch_private_status_t private_status;
666 notmuch_status_t ret = NOTMUCH_STATUS_SUCCESS;
668 find_doc_ids (notmuch, "thread", loser_thread_id, &loser, &loser_end);
670 for ( ; loser != loser_end; loser++) {
671 message = _notmuch_message_create (notmuch, notmuch,
672 *loser, &private_status);
673 if (message == NULL) {
674 ret = COERCE_STATUS (private_status,
675 "Cannot find document for doc_id from query");
679 _notmuch_message_remove_term (message, "thread", loser_thread_id);
680 _notmuch_message_add_term (message, "thread", winner_thread_id);
681 _notmuch_message_sync (message);
683 notmuch_message_destroy (message);
689 notmuch_message_destroy (message);
695 _my_talloc_free_for_g_hash (void *ptr)
700 static notmuch_status_t
701 _notmuch_database_link_message_to_parents (notmuch_database_t *notmuch,
702 notmuch_message_t *message,
703 notmuch_message_file_t *message_file,
704 const char **thread_id)
706 GHashTable *parents = NULL;
707 const char *refs, *in_reply_to, *in_reply_to_message_id;
708 GList *l, *keys = NULL;
709 notmuch_status_t ret = NOTMUCH_STATUS_SUCCESS;
711 parents = g_hash_table_new_full (g_str_hash, g_str_equal,
712 _my_talloc_free_for_g_hash, NULL);
714 refs = notmuch_message_file_get_header (message_file, "references");
715 parse_references (message, notmuch_message_get_message_id (message),
718 in_reply_to = notmuch_message_file_get_header (message_file, "in-reply-to");
719 parse_references (message, notmuch_message_get_message_id (message),
720 parents, in_reply_to);
722 /* Carefully avoid adding any self-referential in-reply-to term. */
723 in_reply_to_message_id = _parse_message_id (message, in_reply_to, NULL);
724 if (strcmp (in_reply_to_message_id,
725 notmuch_message_get_message_id (message)))
727 _notmuch_message_add_term (message, "replyto",
728 _parse_message_id (message, in_reply_to, NULL));
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 look 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 = NULL;
856 notmuch_status_t ret = NOTMUCH_STATUS_SUCCESS;
857 notmuch_private_status_t private_status;
859 const char *date, *header;
860 const char *from, *to, *subject;
866 message_file = notmuch_message_file_open (filename);
867 if (message_file == NULL) {
868 ret = NOTMUCH_STATUS_FILE_ERROR;
872 notmuch_message_file_restrict_headers (message_file,
883 /* Before we do any real work, (especially before doing a
884 * potential SHA-1 computation on the entire file's contents),
885 * let's make sure that what we're looking at looks like an
886 * actual email message.
888 from = notmuch_message_file_get_header (message_file, "from");
889 subject = notmuch_message_file_get_header (message_file, "subject");
890 to = notmuch_message_file_get_header (message_file, "to");
896 ret = NOTMUCH_STATUS_FILE_NOT_EMAIL;
900 /* Now that we're sure it's mail, the first order of business
901 * is to find a message ID (or else create one ourselves). */
903 header = notmuch_message_file_get_header (message_file, "message-id");
905 message_id = _parse_message_id (message_file, header, NULL);
906 /* So the header value isn't RFC-compliant, but it's
907 * better than no message-id at all. */
908 if (message_id == NULL)
909 message_id = talloc_strdup (message_file, header);
911 /* No message-id at all, let's generate one by taking a
912 * hash over the file's contents. */
913 char *sha1 = notmuch_sha1_of_file (filename);
915 /* If that failed too, something is really wrong. Give up. */
917 ret = NOTMUCH_STATUS_FILE_ERROR;
921 message_id = talloc_asprintf (message_file,
922 "notmuch-sha1-%s", sha1);
926 /* Now that we have a message ID, we get a message object,
927 * (which may or may not reference an existing document in the
930 message = _notmuch_message_create_for_message_id (notmuch,
934 talloc_free (message_id);
939 /* Is this a newly created message object? */
940 if (private_status == NOTMUCH_PRIVATE_STATUS_NO_DOCUMENT_FOUND) {
941 _notmuch_message_set_filename (message, filename);
942 _notmuch_message_add_term (message, "type", "mail");
944 ret = NOTMUCH_STATUS_DUPLICATE_MESSAGE_ID;
948 ret = _notmuch_database_link_message (notmuch, message, message_file);
952 date = notmuch_message_file_get_header (message_file, "date");
953 _notmuch_message_set_date (message, date);
955 _notmuch_message_index_file (message, filename);
957 _notmuch_message_sync (message);
958 } catch (const Xapian::Error &error) {
959 fprintf (stderr, "A Xapian exception occurred: %s.\n",
960 error.get_msg().c_str());
961 ret = NOTMUCH_STATUS_XAPIAN_EXCEPTION;
967 if (ret == NOTMUCH_STATUS_SUCCESS && message_ret)
968 *message_ret = message;
970 notmuch_message_destroy (message);
974 notmuch_message_file_close (message_file);