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 parse_references (void *ctx,
386 ref = _parse_message_id (ctx, refs, &refs);
389 g_hash_table_insert (hash, ref, NULL);
394 notmuch_database_create (const char *path)
396 notmuch_database_t *notmuch = NULL;
397 char *notmuch_path = NULL;
402 fprintf (stderr, "Error: Cannot create a database for a NULL path.\n");
406 err = stat (path, &st);
408 fprintf (stderr, "Error: Cannot create database at %s: %s.\n",
409 path, strerror (errno));
413 if (! S_ISDIR (st.st_mode)) {
414 fprintf (stderr, "Error: Cannot create database at %s: Not a directory.\n",
419 notmuch_path = talloc_asprintf (NULL, "%s/%s", path, ".notmuch");
421 err = mkdir (notmuch_path, 0755);
424 fprintf (stderr, "Error: Cannot create directory %s: %s.\n",
425 notmuch_path, strerror (errno));
429 notmuch = notmuch_database_open (path);
433 talloc_free (notmuch_path);
439 notmuch_database_open (const char *path)
441 notmuch_database_t *notmuch = NULL;
442 char *notmuch_path = NULL, *xapian_path = NULL;
447 if (asprintf (¬much_path, "%s/%s", path, ".notmuch") == -1) {
449 fprintf (stderr, "Out of memory\n");
453 err = stat (notmuch_path, &st);
455 fprintf (stderr, "Error opening database at %s: %s\n",
456 notmuch_path, strerror (errno));
460 if (asprintf (&xapian_path, "%s/%s", notmuch_path, "xapian") == -1) {
462 fprintf (stderr, "Out of memory\n");
466 notmuch = talloc (NULL, notmuch_database_t);
467 notmuch->path = talloc_strdup (notmuch, path);
469 if (notmuch->path[strlen (notmuch->path) - 1] == '/')
470 notmuch->path[strlen (notmuch->path) - 1] = '\0';
473 notmuch->xapian_db = new Xapian::WritableDatabase (xapian_path,
474 Xapian::DB_CREATE_OR_OPEN);
475 notmuch->query_parser = new Xapian::QueryParser;
476 notmuch->term_gen = new Xapian::TermGenerator;
477 notmuch->term_gen->set_stemmer (Xapian::Stem ("english"));
479 notmuch->query_parser->set_default_op (Xapian::Query::OP_AND);
480 notmuch->query_parser->set_database (*notmuch->xapian_db);
481 notmuch->query_parser->set_stemmer (Xapian::Stem ("english"));
482 notmuch->query_parser->set_stemming_strategy (Xapian::QueryParser::STEM_SOME);
484 for (i = 0; i < ARRAY_SIZE (BOOLEAN_PREFIX_EXTERNAL); i++) {
485 prefix_t *prefix = &BOOLEAN_PREFIX_EXTERNAL[i];
486 notmuch->query_parser->add_boolean_prefix (prefix->name,
490 for (i = 0; i < ARRAY_SIZE (PROBABILISTIC_PREFIX); i++) {
491 prefix_t *prefix = &PROBABILISTIC_PREFIX[i];
492 notmuch->query_parser->add_prefix (prefix->name, prefix->prefix);
494 } catch (const Xapian::Error &error) {
495 fprintf (stderr, "A Xapian exception occurred: %s\n",
496 error.get_msg().c_str());
510 notmuch_database_close (notmuch_database_t *notmuch)
512 notmuch->xapian_db->flush ();
514 delete notmuch->term_gen;
515 delete notmuch->query_parser;
516 delete notmuch->xapian_db;
517 talloc_free (notmuch);
521 notmuch_database_get_path (notmuch_database_t *notmuch)
523 return notmuch->path;
526 static notmuch_private_status_t
527 find_timestamp_document (notmuch_database_t *notmuch, const char *db_key,
528 Xapian::Document *doc, unsigned int *doc_id)
530 return find_unique_document (notmuch, "timestamp", db_key, doc, doc_id);
533 /* We allow the user to use arbitrarily long keys for timestamps,
534 * (they're for filesystem paths after all, which have no limit we
535 * know about). But we have a term-length limit. So if we exceed that,
536 * we'll use the SHA-1 of the user's key as the actual key for
537 * constructing a database term.
539 * Caution: This function returns a newly allocated string which the
540 * caller should free() when finished.
543 timestamp_db_key (const char *key)
545 int term_len = strlen (_find_prefix ("timestamp")) + strlen (key);
547 if (term_len > NOTMUCH_TERM_MAX)
548 return notmuch_sha1_of_string (key);
554 notmuch_database_set_timestamp (notmuch_database_t *notmuch,
555 const char *key, time_t timestamp)
557 Xapian::Document doc;
559 notmuch_private_status_t status;
560 notmuch_status_t ret = NOTMUCH_STATUS_SUCCESS;
563 db_key = timestamp_db_key (key);
566 status = find_timestamp_document (notmuch, db_key, &doc, &doc_id);
568 doc.add_value (NOTMUCH_VALUE_TIMESTAMP,
569 Xapian::sortable_serialise (timestamp));
571 if (status == NOTMUCH_PRIVATE_STATUS_NO_DOCUMENT_FOUND) {
572 char *term = talloc_asprintf (NULL, "%s%s",
573 _find_prefix ("timestamp"), db_key);
577 notmuch->xapian_db->add_document (doc);
579 notmuch->xapian_db->replace_document (doc_id, doc);
582 } catch (Xapian::Error &error) {
583 fprintf (stderr, "A Xapian exception occurred: %s.\n",
584 error.get_msg().c_str());
585 ret = NOTMUCH_STATUS_XAPIAN_EXCEPTION;
595 notmuch_database_get_timestamp (notmuch_database_t *notmuch, const char *key)
597 Xapian::Document doc;
599 notmuch_private_status_t status;
603 db_key = timestamp_db_key (key);
606 status = find_timestamp_document (notmuch, db_key, &doc, &doc_id);
608 if (status == NOTMUCH_PRIVATE_STATUS_NO_DOCUMENT_FOUND)
611 ret = Xapian::sortable_unserialise (doc.get_value (NOTMUCH_VALUE_TIMESTAMP));
612 } catch (Xapian::Error &error) {
623 /* Find the thread ID to which the message with 'message_id' belongs.
625 * Returns NULL if no message with message ID 'message_id' is in the
628 * Otherwise, returns a newly talloced string belonging to 'ctx'.
631 _resolve_message_id_to_thread_id (notmuch_database_t *notmuch,
633 const char *message_id)
635 notmuch_message_t *message;
636 const char *ret = NULL;
638 message = notmuch_database_find_message (notmuch, message_id);
642 ret = talloc_steal (ctx, notmuch_message_get_thread_id (message));
646 notmuch_message_destroy (message);
651 static notmuch_status_t
652 _merge_threads (notmuch_database_t *notmuch,
653 const char *winner_thread_id,
654 const char *loser_thread_id)
656 Xapian::PostingIterator loser, loser_end;
657 notmuch_message_t *message = NULL;
658 notmuch_private_status_t private_status;
659 notmuch_status_t ret = NOTMUCH_STATUS_SUCCESS;
661 find_doc_ids (notmuch, "thread", loser_thread_id, &loser, &loser_end);
663 for ( ; loser != loser_end; loser++) {
664 message = _notmuch_message_create (notmuch, notmuch,
665 *loser, &private_status);
666 if (message == NULL) {
667 ret = COERCE_STATUS (private_status,
668 "Cannot find document for doc_id from query");
672 _notmuch_message_remove_term (message, "thread", loser_thread_id);
673 _notmuch_message_add_term (message, "thread", winner_thread_id);
674 _notmuch_message_sync (message);
676 notmuch_message_destroy (message);
682 notmuch_message_destroy (message);
688 _my_talloc_free_for_g_hash (void *ptr)
693 static notmuch_status_t
694 _notmuch_database_link_message_to_parents (notmuch_database_t *notmuch,
695 notmuch_message_t *message,
696 notmuch_message_file_t *message_file,
697 const char **thread_id)
699 GHashTable *parents = NULL;
700 const char *refs, *in_reply_to;
701 GList *l, *keys = NULL;
702 notmuch_status_t ret = NOTMUCH_STATUS_SUCCESS;
704 parents = g_hash_table_new_full (g_str_hash, g_str_equal,
705 _my_talloc_free_for_g_hash, NULL);
707 refs = notmuch_message_file_get_header (message_file, "references");
708 parse_references (message, parents, refs);
710 in_reply_to = notmuch_message_file_get_header (message_file, "in-reply-to");
711 parse_references (message, parents, in_reply_to);
712 _notmuch_message_add_term (message, "replyto",
713 _parse_message_id (message, in_reply_to, NULL));
715 keys = g_hash_table_get_keys (parents);
716 for (l = keys; l; l = l->next) {
717 char *parent_message_id;
718 const char *parent_thread_id;
720 parent_message_id = (char *) l->data;
721 parent_thread_id = _resolve_message_id_to_thread_id (notmuch,
725 if (parent_thread_id == NULL) {
726 _notmuch_message_add_term (message, "ref", parent_message_id);
728 if (*thread_id == NULL) {
729 *thread_id = talloc_strdup (message, parent_thread_id);
730 _notmuch_message_add_term (message, "thread", *thread_id);
731 } else if (strcmp (*thread_id, parent_thread_id)) {
732 ret = _merge_threads (notmuch, *thread_id, parent_thread_id);
743 g_hash_table_unref (parents);
748 static notmuch_status_t
749 _notmuch_database_link_message_to_children (notmuch_database_t *notmuch,
750 notmuch_message_t *message,
751 const char **thread_id)
753 const char *message_id = notmuch_message_get_message_id (message);
754 Xapian::PostingIterator child, children_end;
755 notmuch_message_t *child_message = NULL;
756 const char *child_thread_id;
757 notmuch_status_t ret = NOTMUCH_STATUS_SUCCESS;
758 notmuch_private_status_t private_status;
760 find_doc_ids (notmuch, "ref", message_id, &child, &children_end);
762 for ( ; child != children_end; child++) {
764 child_message = _notmuch_message_create (message, notmuch,
765 *child, &private_status);
766 if (child_message == NULL) {
767 ret = COERCE_STATUS (private_status,
768 "Cannot find document for doc_id from query");
772 child_thread_id = notmuch_message_get_thread_id (child_message);
773 if (*thread_id == NULL) {
774 *thread_id = talloc_strdup (message, child_thread_id);
775 _notmuch_message_add_term (message, "thread", *thread_id);
776 } else if (strcmp (*thread_id, child_thread_id)) {
777 _notmuch_message_remove_term (child_message, "ref",
779 _notmuch_message_sync (child_message);
780 ret = _merge_threads (notmuch, *thread_id, child_thread_id);
785 notmuch_message_destroy (child_message);
786 child_message = NULL;
791 notmuch_message_destroy (child_message);
796 /* Given a (mostly empty) 'message' and its corresponding
797 * 'message_file' link it to existing threads in the database.
799 * We first look at 'message_file' and its link-relevant headers
800 * (References and In-Reply-To) for message IDs. We also look in the
801 * database for existing message that reference 'message'.p
803 * The end result is to call _notmuch_message_add_thread_id with one
804 * or more thread IDs to which this message belongs, (including
805 * generating a new thread ID if necessary if the message doesn't
806 * connect to any existing threads).
808 static notmuch_status_t
809 _notmuch_database_link_message (notmuch_database_t *notmuch,
810 notmuch_message_t *message,
811 notmuch_message_file_t *message_file)
813 notmuch_status_t status;
814 const char *thread_id = NULL;
816 status = _notmuch_database_link_message_to_parents (notmuch, message,
822 status = _notmuch_database_link_message_to_children (notmuch, message,
827 if (thread_id == NULL)
828 _notmuch_message_ensure_thread_id (message);
830 return NOTMUCH_STATUS_SUCCESS;
834 notmuch_database_add_message (notmuch_database_t *notmuch,
835 const char *filename,
836 notmuch_message_t **message_ret)
838 notmuch_message_file_t *message_file;
839 notmuch_message_t *message = NULL;
840 notmuch_status_t ret = NOTMUCH_STATUS_SUCCESS;
841 notmuch_private_status_t private_status;
843 const char *date, *header;
844 const char *from, *to, *subject;
850 message_file = notmuch_message_file_open (filename);
851 if (message_file == NULL) {
852 ret = NOTMUCH_STATUS_FILE_ERROR;
856 notmuch_message_file_restrict_headers (message_file,
867 /* Before we do any real work, (especially before doing a
868 * potential SHA-1 computation on the entire file's contents),
869 * let's make sure that what we're looking at looks like an
870 * actual email message.
872 from = notmuch_message_file_get_header (message_file, "from");
873 subject = notmuch_message_file_get_header (message_file, "subject");
874 to = notmuch_message_file_get_header (message_file, "to");
880 ret = NOTMUCH_STATUS_FILE_NOT_EMAIL;
884 /* Now that we're sure it's mail, the first order of business
885 * is to find a message ID (or else create one ourselves). */
887 header = notmuch_message_file_get_header (message_file, "message-id");
889 message_id = _parse_message_id (message_file, header, NULL);
890 /* So the header value isn't RFC-compliant, but it's
891 * better than no message-id at all. */
892 if (message_id == NULL)
893 message_id = talloc_strdup (message_file, header);
895 /* No message-id at all, let's generate one by taking a
896 * hash over the file's contents. */
897 char *sha1 = notmuch_sha1_of_file (filename);
899 /* If that failed too, something is really wrong. Give up. */
901 ret = NOTMUCH_STATUS_FILE_ERROR;
905 message_id = talloc_asprintf (message_file,
906 "notmuch-sha1-%s", sha1);
910 /* Now that we have a message ID, we get a message object,
911 * (which may or may not reference an existing document in the
914 message = _notmuch_message_create_for_message_id (notmuch,
918 talloc_free (message_id);
923 /* Is this a newly created message object? */
924 if (private_status == NOTMUCH_PRIVATE_STATUS_NO_DOCUMENT_FOUND) {
925 _notmuch_message_set_filename (message, filename);
926 _notmuch_message_add_term (message, "type", "mail");
928 ret = NOTMUCH_STATUS_DUPLICATE_MESSAGE_ID;
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 _notmuch_message_index_file (message, filename);
941 _notmuch_message_sync (message);
942 } catch (const Xapian::Error &error) {
943 fprintf (stderr, "A Xapian exception occurred: %s.\n",
944 error.get_msg().c_str());
945 ret = NOTMUCH_STATUS_XAPIAN_EXCEPTION;
951 if (ret == NOTMUCH_STATUS_SUCCESS && message_ret)
952 *message_ret = message;
954 notmuch_message_destroy (message);
958 notmuch_message_file_close (message_file);