1 /* notmuch - Not much of an email program, (just index and search)
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-client.h"
25 typedef struct _filename_node {
27 struct _filename_node *next;
30 typedef struct _filename_list {
31 _filename_node_t *head;
32 _filename_node_t **tail;
38 const char **new_tags;
39 size_t new_tags_length;
44 struct timeval tv_start;
46 _filename_list_t *removed_files;
47 _filename_list_t *removed_directories;
49 notmuch_bool_t synchronize_flags;
50 _filename_list_t *message_ids_to_sync;
53 static volatile sig_atomic_t do_add_files_print_progress = 0;
56 handle_sigalrm (unused (int signal))
58 do_add_files_print_progress = 1;
61 static volatile sig_atomic_t interrupted;
64 handle_sigint (unused (int sig))
67 static char msg[] = "Stopping... \n";
69 ignored = write(2, msg, sizeof(msg)-1);
73 static _filename_list_t *
74 _filename_list_create (const void *ctx)
76 _filename_list_t *list;
78 list = talloc (ctx, _filename_list_t);
83 list->tail = &list->head;
89 _filename_list_add (_filename_list_t *list,
92 _filename_node_t *node = talloc (list, _filename_node_t);
94 node->filename = talloc_strdup (list, filename);
98 list->tail = &node->next;
102 add_files_print_progress (add_files_state_t *state)
104 struct timeval tv_now;
105 double elapsed_overall, rate_overall;
107 gettimeofday (&tv_now, NULL);
109 elapsed_overall = notmuch_time_elapsed (state->tv_start, tv_now);
110 rate_overall = (state->processed_files) / elapsed_overall;
112 printf ("Processed %d", state->processed_files);
114 if (state->total_files) {
115 double time_remaining;
117 time_remaining = ((state->total_files - state->processed_files) /
119 printf (" of %d files (", state->total_files);
120 notmuch_time_print_formatted_seconds (time_remaining);
121 printf (" remaining). \r");
123 printf (" files (%d files/sec.) \r", (int) rate_overall);
130 dirent_sort_inode (const struct dirent **a, const struct dirent **b)
132 return ((*a)->d_ino < (*b)->d_ino) ? -1 : 1;
136 dirent_sort_strcmp_name (const struct dirent **a, const struct dirent **b)
138 return strcmp ((*a)->d_name, (*b)->d_name);
141 /* Test if the directory looks like a Maildir directory.
143 * Search through the array of directory entries to see if we can find all
144 * three subdirectories typical for Maildir, that is "new", "cur", and "tmp".
146 * Return 1 if the directory looks like a Maildir and 0 otherwise.
149 _entries_resemble_maildir (struct dirent **entries, int count)
153 for (i = 0; i < count; i++) {
154 if (entries[i]->d_type != DT_DIR && entries[i]->d_type != DT_UNKNOWN)
157 if (strcmp(entries[i]->d_name, "new") == 0 ||
158 strcmp(entries[i]->d_name, "cur") == 0 ||
159 strcmp(entries[i]->d_name, "tmp") == 0)
170 /* Examine 'path' recursively as follows:
172 * o Ask the filesystem for the mtime of 'path' (fs_mtime)
173 * o Ask the database for its timestamp of 'path' (db_mtime)
175 * o Ask the filesystem for files and directories within 'path'
176 * (via scandir and stored in fs_entries)
177 * o Ask the database for files and directories within 'path'
178 * (db_files and db_subdirs)
180 * o Pass 1: For each directory in fs_entries, recursively call into
181 * this same function.
183 * o Pass 2: If 'fs_mtime' > 'db_mtime', then walk fs_entries
184 * simultaneously with db_files and db_subdirs. Look for one of
185 * three interesting cases:
187 * 1. Regular file in fs_entries and not in db_files
188 * This is a new file to add_message into the database.
190 * 2. Filename in db_files not in fs_entries.
191 * This is a file that has been removed from the mail store.
193 * 3. Directory in db_subdirs not in fs_entries
194 * This is a directory that has been removed from the mail store.
196 * Note that the addition of a directory is not interesting here,
197 * since that will have been taken care of in pass 1. Also, we
198 * don't immediately act on file/directory removal since we must
199 * ensure that in the case of a rename that the new filename is
200 * added before the old filename is removed, (so that no
201 * information is lost from the database).
203 * o Tell the database to update its time of 'path' to 'fs_mtime'
205 static notmuch_status_t
206 add_files_recursive (notmuch_database_t *notmuch,
208 add_files_state_t *state)
211 struct dirent *entry = NULL;
213 time_t fs_mtime, db_mtime;
214 notmuch_status_t status, ret = NOTMUCH_STATUS_SUCCESS;
215 notmuch_message_t *message = NULL;
216 struct dirent **fs_entries = NULL;
217 int i, num_fs_entries;
218 notmuch_directory_t *directory;
219 notmuch_filenames_t *db_files = NULL;
220 notmuch_filenames_t *db_subdirs = NULL;
222 notmuch_bool_t is_maildir, new_directory;
225 if (stat (path, &st)) {
226 fprintf (stderr, "Error reading directory %s: %s\n",
227 path, strerror (errno));
228 return NOTMUCH_STATUS_FILE_ERROR;
231 /* This is not an error since we may have recursed based on a
232 * symlink to a regular file, not a directory, and we don't know
233 * that until this stat. */
234 if (! S_ISDIR (st.st_mode))
235 return NOTMUCH_STATUS_SUCCESS;
237 fs_mtime = st.st_mtime;
239 directory = notmuch_database_get_directory (notmuch, path);
240 db_mtime = notmuch_directory_get_mtime (directory);
243 new_directory = TRUE;
247 new_directory = FALSE;
248 db_files = notmuch_directory_get_child_files (directory);
249 db_subdirs = notmuch_directory_get_child_directories (directory);
252 /* If the database knows about this directory, then we sort based
253 * on strcmp to match the database sorting. Otherwise, we can do
254 * inode-based sorting for faster filesystem operation. */
255 num_fs_entries = scandir (path, &fs_entries, 0,
257 dirent_sort_inode : dirent_sort_strcmp_name);
259 if (num_fs_entries == -1) {
260 fprintf (stderr, "Error opening directory %s: %s\n",
261 path, strerror (errno));
262 ret = NOTMUCH_STATUS_FILE_ERROR;
266 /* Pass 1: Recurse into all sub-directories. */
267 is_maildir = _entries_resemble_maildir (fs_entries, num_fs_entries);
269 for (i = 0; i < num_fs_entries; i++) {
273 entry = fs_entries[i];
275 /* We only want to descend into directories.
276 * But symlinks can be to directories too, of course.
278 * And if the filesystem doesn't tell us the file type in the
279 * scandir results, then it might be a directory (and if not,
280 * then we'll stat and return immediately in the next level of
282 if (entry->d_type != DT_DIR &&
283 entry->d_type != DT_LNK &&
284 entry->d_type != DT_UNKNOWN)
289 /* Ignore special directories to avoid infinite recursion.
290 * Also ignore the .notmuch directory and any "tmp" directory
291 * that appears within a maildir.
293 /* XXX: Eventually we'll want more sophistication to let the
294 * user specify files to be ignored. */
295 if (strcmp (entry->d_name, ".") == 0 ||
296 strcmp (entry->d_name, "..") == 0 ||
297 (is_maildir && strcmp (entry->d_name, "tmp") == 0) ||
298 strcmp (entry->d_name, ".notmuch") ==0)
303 next = talloc_asprintf (notmuch, "%s/%s", path, entry->d_name);
304 status = add_files_recursive (notmuch, next, state);
305 if (status && ret == NOTMUCH_STATUS_SUCCESS)
311 /* If this directory hasn't been modified since the last
312 * "notmuch new", then we can skip the second pass entirely. */
313 if (fs_mtime <= db_mtime)
316 /* Pass 2: Scan for new files, removed files, and removed directories. */
317 for (i = 0; i < num_fs_entries; i++)
322 entry = fs_entries[i];
324 /* Check if we've walked past any names in db_files or
325 * db_subdirs. If so, these have been deleted. */
326 while (notmuch_filenames_valid (db_files) &&
327 strcmp (notmuch_filenames_get (db_files), entry->d_name) < 0)
329 char *absolute = talloc_asprintf (state->removed_files,
331 notmuch_filenames_get (db_files));
333 _filename_list_add (state->removed_files, absolute);
335 notmuch_filenames_move_to_next (db_files);
338 while (notmuch_filenames_valid (db_subdirs) &&
339 strcmp (notmuch_filenames_get (db_subdirs), entry->d_name) <= 0)
341 const char *filename = notmuch_filenames_get (db_subdirs);
343 if (strcmp (filename, entry->d_name) < 0)
345 char *absolute = talloc_asprintf (state->removed_directories,
346 "%s/%s", path, filename);
348 _filename_list_add (state->removed_directories, absolute);
351 notmuch_filenames_move_to_next (db_subdirs);
354 /* If we're looking at a symlink, we only want to add it if it
355 * links to a regular file, (and not to a directory, say).
357 * Similarly, if the file is of unknown type (due to filesytem
358 * limitations), then we also need to look closer.
360 * In either case, a stat does the trick.
362 if (entry->d_type == DT_LNK || entry->d_type == DT_UNKNOWN) {
365 next = talloc_asprintf (notmuch, "%s/%s", path, entry->d_name);
366 err = stat (next, &st);
370 /* Don't emit an error for a link pointing nowhere, since
371 * the directory-traversal pass will have already done
376 if (! S_ISREG (st.st_mode))
378 } else if (entry->d_type != DT_REG) {
382 /* Don't add a file that we've added before. */
383 if (notmuch_filenames_valid (db_files) &&
384 strcmp (notmuch_filenames_get (db_files), entry->d_name) == 0)
386 notmuch_filenames_move_to_next (db_files);
390 /* We're now looking at a regular file that doesn't yet exist
391 * in the database, so add it. */
392 next = talloc_asprintf (notmuch, "%s/%s", path, entry->d_name);
394 state->processed_files++;
396 if (state->verbose) {
397 if (state->output_is_a_tty)
401 state->processed_files,
405 putchar((state->output_is_a_tty) ? '\r' : '\n');
409 status = notmuch_database_add_message (notmuch, next, &message);
412 case NOTMUCH_STATUS_SUCCESS:
413 state->added_messages++;
414 for (tag=state->new_tags; *tag != NULL; tag++)
415 notmuch_message_add_tag (message, *tag);
416 /* Defer sync of maildir flags until after old filenames
417 * are removed in the case of a rename. */
418 if (state->synchronize_flags == TRUE)
419 _filename_list_add (state->message_ids_to_sync,
420 notmuch_message_get_message_id (message));
422 /* Non-fatal issues (go on to next file) */
423 case NOTMUCH_STATUS_DUPLICATE_MESSAGE_ID:
424 /* Defer sync of maildir flags until after old filenames
425 * are removed in the case of a rename. */
426 if (state->synchronize_flags == TRUE)
427 _filename_list_add (state->message_ids_to_sync,
428 notmuch_message_get_message_id (message));
430 case NOTMUCH_STATUS_FILE_NOT_EMAIL:
431 fprintf (stderr, "Note: Ignoring non-mail file: %s\n",
434 /* Fatal issues. Don't process anymore. */
435 case NOTMUCH_STATUS_READ_ONLY_DATABASE:
436 case NOTMUCH_STATUS_XAPIAN_EXCEPTION:
437 case NOTMUCH_STATUS_OUT_OF_MEMORY:
438 fprintf (stderr, "Error: %s. Halting processing.\n",
439 notmuch_status_to_string (status));
443 case NOTMUCH_STATUS_FILE_ERROR:
444 case NOTMUCH_STATUS_NULL_POINTER:
445 case NOTMUCH_STATUS_TAG_TOO_LONG:
446 case NOTMUCH_STATUS_UNBALANCED_FREEZE_THAW:
447 case NOTMUCH_STATUS_LAST_STATUS:
448 INTERNAL_ERROR ("add_message returned unexpected value: %d", status);
453 notmuch_message_destroy (message);
457 if (do_add_files_print_progress) {
458 do_add_files_print_progress = 0;
459 add_files_print_progress (state);
469 /* Now that we've walked the whole filesystem list, anything left
470 * over in the database lists has been deleted. */
471 while (notmuch_filenames_valid (db_files))
473 char *absolute = talloc_asprintf (state->removed_files,
475 notmuch_filenames_get (db_files));
477 _filename_list_add (state->removed_files, absolute);
479 notmuch_filenames_move_to_next (db_files);
482 while (notmuch_filenames_valid (db_subdirs))
484 char *absolute = talloc_asprintf (state->removed_directories,
486 notmuch_filenames_get (db_subdirs));
488 _filename_list_add (state->removed_directories, absolute);
490 notmuch_filenames_move_to_next (db_subdirs);
494 status = notmuch_directory_set_mtime (directory, fs_mtime);
495 if (status && ret == NOTMUCH_STATUS_SUCCESS)
509 notmuch_filenames_destroy (db_subdirs);
511 notmuch_filenames_destroy (db_files);
513 notmuch_directory_destroy (directory);
518 /* This is the top-level entry point for add_files. It does a couple
519 * of error checks, sets up the progress-printing timer and then calls
520 * into the recursive function. */
521 static notmuch_status_t
522 add_files (notmuch_database_t *notmuch,
524 add_files_state_t *state)
526 notmuch_status_t status;
527 struct sigaction action;
528 struct itimerval timerval;
529 notmuch_bool_t timer_is_active = FALSE;
532 if (state->output_is_a_tty && ! debugger_is_active () && ! state->verbose) {
533 /* Setup our handler for SIGALRM */
534 memset (&action, 0, sizeof (struct sigaction));
535 action.sa_handler = handle_sigalrm;
536 sigemptyset (&action.sa_mask);
537 action.sa_flags = SA_RESTART;
538 sigaction (SIGALRM, &action, NULL);
540 /* Then start a timer to send SIGALRM once per second. */
541 timerval.it_interval.tv_sec = 1;
542 timerval.it_interval.tv_usec = 0;
543 timerval.it_value.tv_sec = 1;
544 timerval.it_value.tv_usec = 0;
545 setitimer (ITIMER_REAL, &timerval, NULL);
547 timer_is_active = TRUE;
550 if (stat (path, &st)) {
551 fprintf (stderr, "Error reading directory %s: %s\n",
552 path, strerror (errno));
553 return NOTMUCH_STATUS_FILE_ERROR;
556 if (! S_ISDIR (st.st_mode)) {
557 fprintf (stderr, "Error: %s is not a directory.\n", path);
558 return NOTMUCH_STATUS_FILE_ERROR;
561 status = add_files_recursive (notmuch, path, state);
563 if (timer_is_active) {
564 /* Now stop the timer. */
565 timerval.it_interval.tv_sec = 0;
566 timerval.it_interval.tv_usec = 0;
567 timerval.it_value.tv_sec = 0;
568 timerval.it_value.tv_usec = 0;
569 setitimer (ITIMER_REAL, &timerval, NULL);
571 /* And disable the signal handler. */
572 action.sa_handler = SIG_IGN;
573 sigaction (SIGALRM, &action, NULL);
579 /* XXX: This should be merged with the add_files function since it
580 * shares a lot of logic with it. */
581 /* Recursively count all regular files in path and all sub-directories
582 * of path. The result is added to *count (which should be
583 * initialized to zero by the top-level caller before calling
586 count_files (const char *path, int *count)
588 struct dirent *entry = NULL;
591 struct dirent **fs_entries = NULL;
592 int num_fs_entries = scandir (path, &fs_entries, 0, dirent_sort_inode);
595 if (num_fs_entries == -1) {
596 fprintf (stderr, "Warning: failed to open directory %s: %s\n",
597 path, strerror (errno));
601 while (!interrupted) {
602 if (i == num_fs_entries)
605 entry = fs_entries[i++];
607 /* Ignore special directories to avoid infinite recursion.
608 * Also ignore the .notmuch directory.
610 /* XXX: Eventually we'll want more sophistication to let the
611 * user specify files to be ignored. */
612 if (strcmp (entry->d_name, ".") == 0 ||
613 strcmp (entry->d_name, "..") == 0 ||
614 strcmp (entry->d_name, ".notmuch") == 0)
619 if (asprintf (&next, "%s/%s", path, entry->d_name) == -1) {
621 fprintf (stderr, "Error descending from %s to %s: Out of memory\n",
622 path, entry->d_name);
628 if (S_ISREG (st.st_mode)) {
630 if (*count % 1000 == 0) {
631 printf ("Found %d files so far.\r", *count);
634 } else if (S_ISDIR (st.st_mode)) {
635 count_files (next, count);
649 upgrade_print_progress (void *closure,
652 add_files_state_t *state = closure;
654 printf ("Upgrading database: %.2f%% complete", progress * 100.0);
657 struct timeval tv_now;
658 double elapsed, time_remaining;
660 gettimeofday (&tv_now, NULL);
662 elapsed = notmuch_time_elapsed (state->tv_start, tv_now);
663 time_remaining = (elapsed / progress) * (1.0 - progress);
665 notmuch_time_print_formatted_seconds (time_remaining);
666 printf (" remaining)");
674 /* Recursively remove all filenames from the database referring to
675 * 'path' (or to any of its children). */
677 _remove_directory (void *ctx,
678 notmuch_database_t *notmuch,
683 notmuch_directory_t *directory;
684 notmuch_filenames_t *files, *subdirs;
685 notmuch_status_t status;
688 directory = notmuch_database_get_directory (notmuch, path);
690 for (files = notmuch_directory_get_child_files (directory);
691 notmuch_filenames_valid (files);
692 notmuch_filenames_move_to_next (files))
694 absolute = talloc_asprintf (ctx, "%s/%s", path,
695 notmuch_filenames_get (files));
696 status = notmuch_database_remove_message (notmuch, absolute);
697 if (status == NOTMUCH_STATUS_DUPLICATE_MESSAGE_ID)
698 *renamed_files = *renamed_files + 1;
700 *removed_files = *removed_files + 1;
701 talloc_free (absolute);
704 for (subdirs = notmuch_directory_get_child_directories (directory);
705 notmuch_filenames_valid (subdirs);
706 notmuch_filenames_move_to_next (subdirs))
708 absolute = talloc_asprintf (ctx, "%s/%s", path,
709 notmuch_filenames_get (subdirs));
710 _remove_directory (ctx, notmuch, absolute, renamed_files, removed_files);
711 talloc_free (absolute);
714 notmuch_directory_destroy (directory);
718 notmuch_new_command (void *ctx, int argc, char *argv[])
720 notmuch_config_t *config;
721 notmuch_database_t *notmuch;
722 add_files_state_t add_files_state;
724 struct timeval tv_now;
728 char *dot_notmuch_path;
729 struct sigaction action;
731 int renamed_files, removed_files;
732 notmuch_status_t status;
735 add_files_state.verbose = 0;
736 add_files_state.output_is_a_tty = isatty (fileno (stdout));
738 for (i = 0; i < argc && argv[i][0] == '-'; i++) {
739 if (STRNCMP_LITERAL (argv[i], "--verbose") == 0) {
740 add_files_state.verbose = 1;
742 fprintf (stderr, "Unrecognized option: %s\n", argv[i]);
747 config = notmuch_config_open (ctx, NULL, NULL);
751 add_files_state.new_tags = notmuch_config_get_new_tags (config, &add_files_state.new_tags_length);
752 add_files_state.synchronize_flags = notmuch_config_get_maildir_synchronize_flags (config);
753 add_files_state.message_ids_to_sync = _filename_list_create (ctx);
754 db_path = notmuch_config_get_database_path (config);
756 dot_notmuch_path = talloc_asprintf (ctx, "%s/%s", db_path, ".notmuch");
758 if (stat (dot_notmuch_path, &st)) {
762 count_files (db_path, &count);
766 printf ("Found %d total files (that's not much mail).\n", count);
767 notmuch = notmuch_database_create (db_path);
768 add_files_state.total_files = count;
770 notmuch = notmuch_database_open (db_path,
771 NOTMUCH_DATABASE_MODE_READ_WRITE);
775 if (notmuch_database_needs_upgrade (notmuch)) {
776 printf ("Welcome to a new version of notmuch! Your database will now be upgraded.\n");
777 gettimeofday (&add_files_state.tv_start, NULL);
778 notmuch_database_upgrade (notmuch, upgrade_print_progress,
780 printf ("Your notmuch database has now been upgraded to database format version %u.\n",
781 notmuch_database_get_version (notmuch));
784 add_files_state.total_files = 0;
790 /* Setup our handler for SIGINT. We do this after having
791 * potentially done a database upgrade we this interrupt handler
793 memset (&action, 0, sizeof (struct sigaction));
794 action.sa_handler = handle_sigint;
795 sigemptyset (&action.sa_mask);
796 action.sa_flags = SA_RESTART;
797 sigaction (SIGINT, &action, NULL);
799 talloc_free (dot_notmuch_path);
800 dot_notmuch_path = NULL;
802 add_files_state.processed_files = 0;
803 add_files_state.added_messages = 0;
804 gettimeofday (&add_files_state.tv_start, NULL);
806 add_files_state.removed_files = _filename_list_create (ctx);
807 add_files_state.removed_directories = _filename_list_create (ctx);
809 ret = add_files (notmuch, db_path, &add_files_state);
813 for (f = add_files_state.removed_files->head; f; f = f->next) {
814 status = notmuch_database_remove_message (notmuch, f->filename);
815 if (status == NOTMUCH_STATUS_DUPLICATE_MESSAGE_ID)
821 for (f = add_files_state.removed_directories->head; f; f = f->next) {
822 _remove_directory (ctx, notmuch, f->filename,
823 &renamed_files, &removed_files);
826 talloc_free (add_files_state.removed_files);
827 talloc_free (add_files_state.removed_directories);
829 /* Now that removals are done (hence the database is aware of all
830 * renames), we can synchronize maildir_flags to tags for all
831 * messages that had new filenames appear on this run. */
832 if (add_files_state.synchronize_flags) {
833 _filename_node_t *node;
834 notmuch_message_t *message;
835 for (node = add_files_state.message_ids_to_sync->head;
839 message = notmuch_database_find_message (notmuch, node->filename);
840 notmuch_message_maildir_flags_to_tags (message);
841 notmuch_message_destroy (message);
845 talloc_free (add_files_state.message_ids_to_sync);
846 add_files_state.message_ids_to_sync = NULL;
848 gettimeofday (&tv_now, NULL);
849 elapsed = notmuch_time_elapsed (add_files_state.tv_start,
852 if (add_files_state.processed_files) {
853 printf ("Processed %d %s in ", add_files_state.processed_files,
854 add_files_state.processed_files == 1 ?
855 "file" : "total files");
856 notmuch_time_print_formatted_seconds (elapsed);
858 printf (" (%d files/sec.). \n",
859 (int) (add_files_state.processed_files / elapsed));
865 if (add_files_state.added_messages) {
866 printf ("Added %d new %s to the database.",
867 add_files_state.added_messages,
868 add_files_state.added_messages == 1 ?
869 "message" : "messages");
871 printf ("No new mail.");
875 printf (" Removed %d %s.",
877 removed_files == 1 ? "message" : "messages");
881 printf (" Detected %d file %s.",
883 renamed_files == 1 ? "rename" : "renames");
889 printf ("\nNote: At least one error was encountered: %s\n",
890 notmuch_status_to_string (ret));
893 notmuch_database_close (notmuch);
895 return ret || interrupted;