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;
42 struct timeval tv_start;
44 _filename_list_t *removed_files;
45 _filename_list_t *removed_directories;
48 static volatile sig_atomic_t do_add_files_print_progress = 0;
51 handle_sigalrm (unused (int signal))
53 do_add_files_print_progress = 1;
56 static volatile sig_atomic_t interrupted;
59 handle_sigint (unused (int sig))
62 static char msg[] = "Stopping... \n";
64 ignored = write(2, msg, sizeof(msg)-1);
68 static _filename_list_t *
69 _filename_list_create (const void *ctx)
71 _filename_list_t *list;
73 list = talloc (ctx, _filename_list_t);
78 list->tail = &list->head;
84 _filename_list_add (_filename_list_t *list,
87 _filename_node_t *node = talloc (list, _filename_node_t);
89 node->filename = talloc_strdup (list, filename);
93 list->tail = &node->next;
97 tag_inbox_and_unread (notmuch_message_t *message)
99 notmuch_message_add_tag (message, "inbox");
100 notmuch_message_add_tag (message, "unread");
104 add_files_print_progress (add_files_state_t *state)
106 struct timeval tv_now;
107 double elapsed_overall, rate_overall;
109 gettimeofday (&tv_now, NULL);
111 elapsed_overall = notmuch_time_elapsed (state->tv_start, tv_now);
112 rate_overall = (state->processed_files) / elapsed_overall;
114 printf ("Processed %d", state->processed_files);
116 if (state->total_files) {
117 double time_remaining;
119 time_remaining = ((state->total_files - state->processed_files) /
121 printf (" of %d files (", state->total_files);
122 notmuch_time_print_formatted_seconds (time_remaining);
123 printf (" remaining). \r");
125 printf (" files (%d files/sec.) \r", (int) rate_overall);
132 dirent_sort_inode (const struct dirent **a, const struct dirent **b)
134 return ((*a)->d_ino < (*b)->d_ino) ? -1 : 1;
138 dirent_sort_strcmp_name (const struct dirent **a, const struct dirent **b)
140 return strcmp ((*a)->d_name, (*b)->d_name);
143 /* Test if the directory looks like a Maildir directory.
145 * Search through the array of directory entries to see if we can find all
146 * three subdirectories typical for Maildir, that is "new", "cur", and "tmp".
148 * Return 1 if the directory looks like a Maildir and 0 otherwise.
151 _entries_resemble_maildir (struct dirent **entries, int count)
155 for (i = 0; i < count; i++) {
156 if (entries[i]->d_type != DT_DIR && entries[i]->d_type != DT_UNKNOWN)
159 if (strcmp(entries[i]->d_name, "new") == 0 ||
160 strcmp(entries[i]->d_name, "cur") == 0 ||
161 strcmp(entries[i]->d_name, "tmp") == 0)
172 /* Examine 'path' recursively as follows:
174 * o Ask the filesystem for the mtime of 'path' (fs_mtime)
175 * o Ask the database for its timestamp of 'path' (db_mtime)
177 * o Ask the filesystem for files and directories within 'path'
178 * (via scandir and stored in fs_entries)
179 * o Ask the database for files and directories within 'path'
180 * (db_files and db_subdirs)
182 * o Pass 1: For each directory in fs_entries, recursively call into
183 * this same function.
185 * o Pass 2: If 'fs_mtime' > 'db_mtime', then walk fs_entries
186 * simultaneously with db_files and db_subdirs. Look for one of
187 * three interesting cases:
189 * 1. Regular file in fs_entries and not in db_files
190 * This is a new file to add_message into the database.
192 * 2. Filename in db_files not in fs_entries.
193 * This is a file that has been removed from the mail store.
195 * 3. Directory in db_subdirs not in fs_entries
196 * This is a directory that has been removed from the mail store.
198 * Note that the addition of a directory is not interesting here,
199 * since that will have been taken care of in pass 1. Also, we
200 * don't immediately act on file/directory removal since we must
201 * ensure that in the case of a rename that the new filename is
202 * added before the old filename is removed, (so that no
203 * information is lost from the database).
205 * o Tell the database to update its time of 'path' to 'fs_mtime'
207 static notmuch_status_t
208 add_files_recursive (notmuch_database_t *notmuch,
210 add_files_state_t *state)
213 struct dirent *entry = NULL;
215 time_t fs_mtime, db_mtime;
216 notmuch_status_t status, ret = NOTMUCH_STATUS_SUCCESS;
217 notmuch_message_t *message = NULL;
218 struct dirent **fs_entries = NULL;
219 int i, num_fs_entries;
220 notmuch_directory_t *directory;
221 notmuch_filenames_t *db_files = NULL;
222 notmuch_filenames_t *db_subdirs = NULL;
224 notmuch_bool_t is_maildir, new_directory;
226 if (stat (path, &st)) {
227 fprintf (stderr, "Error reading directory %s: %s\n",
228 path, strerror (errno));
229 return NOTMUCH_STATUS_FILE_ERROR;
232 /* This is not an error since we may have recursed based on a
233 * symlink to a regular file, not a directory, and we don't know
234 * that until this stat. */
235 if (! S_ISDIR (st.st_mode))
236 return NOTMUCH_STATUS_SUCCESS;
238 fs_mtime = st.st_mtime;
240 directory = notmuch_database_get_directory (notmuch, path);
241 db_mtime = notmuch_directory_get_mtime (directory);
244 new_directory = TRUE;
248 new_directory = FALSE;
249 db_files = notmuch_directory_get_child_files (directory);
250 db_subdirs = notmuch_directory_get_child_directories (directory);
253 /* If the database knows about this directory, then we sort based
254 * on strcmp to match the database sorting. Otherwise, we can do
255 * inode-based sorting for faster filesystem operation. */
256 num_fs_entries = scandir (path, &fs_entries, 0,
258 dirent_sort_inode : dirent_sort_strcmp_name);
260 if (num_fs_entries == -1) {
261 fprintf (stderr, "Error opening directory %s: %s\n",
262 path, strerror (errno));
263 ret = NOTMUCH_STATUS_FILE_ERROR;
267 /* Pass 1: Recurse into all sub-directories. */
268 is_maildir = _entries_resemble_maildir (fs_entries, num_fs_entries);
270 for (i = 0; i < num_fs_entries; i++) {
274 entry = fs_entries[i];
276 /* We only want to descend into directories.
277 * But symlinks can be to directories too, of course.
279 * And if the filesystem doesn't tell us the file type in the
280 * scandir results, then it might be a directory (and if not,
281 * then we'll stat and return immediately in the next level of
283 if (entry->d_type != DT_DIR &&
284 entry->d_type != DT_LNK &&
285 entry->d_type != DT_UNKNOWN)
290 /* Ignore special directories to avoid infinite recursion.
291 * Also ignore the .notmuch directory and any "tmp" directory
292 * that appears within a maildir.
294 /* XXX: Eventually we'll want more sophistication to let the
295 * user specify files to be ignored. */
296 if (strcmp (entry->d_name, ".") == 0 ||
297 strcmp (entry->d_name, "..") == 0 ||
298 (is_maildir && strcmp (entry->d_name, "tmp") == 0) ||
299 strcmp (entry->d_name, ".notmuch") ==0)
304 next = talloc_asprintf (notmuch, "%s/%s", path, entry->d_name);
305 status = add_files_recursive (notmuch, next, state);
306 if (status && ret == NOTMUCH_STATUS_SUCCESS)
312 /* If this directory hasn't been modified since the last
313 * "notmuch new", then we can skip the second pass entirely. */
314 if (fs_mtime <= db_mtime)
317 /* Pass 2: Scan for new files, removed files, and removed directories. */
318 for (i = 0; i < num_fs_entries; i++)
323 entry = fs_entries[i];
325 /* Check if we've walked past any names in db_files or
326 * db_subdirs. If so, these have been deleted. */
327 while (notmuch_filenames_valid (db_files) &&
328 strcmp (notmuch_filenames_get (db_files), entry->d_name) < 0)
330 char *absolute = talloc_asprintf (state->removed_files,
332 notmuch_filenames_get (db_files));
334 _filename_list_add (state->removed_files, absolute);
336 notmuch_filenames_move_to_next (db_files);
339 while (notmuch_filenames_valid (db_subdirs) &&
340 strcmp (notmuch_filenames_get (db_subdirs), entry->d_name) <= 0)
342 const char *filename = notmuch_filenames_get (db_subdirs);
344 if (strcmp (filename, entry->d_name) < 0)
346 char *absolute = talloc_asprintf (state->removed_directories,
347 "%s/%s", path, filename);
349 _filename_list_add (state->removed_directories, absolute);
352 notmuch_filenames_move_to_next (db_subdirs);
355 /* If we're looking at a symlink, we only want to add it if it
356 * links to a regular file, (and not to a directory, say).
358 * Similarly, if the file is of unknown type (due to filesytem
359 * limitations), then we also need to look closer.
361 * In either case, a stat does the trick.
363 if (entry->d_type == DT_LNK || entry->d_type == DT_UNKNOWN) {
366 next = talloc_asprintf (notmuch, "%s/%s", path, entry->d_name);
367 err = stat (next, &st);
371 /* Don't emit an error for a link pointing nowhere, since
372 * the directory-traversal pass will have already done
377 if (! S_ISREG (st.st_mode))
379 } else if (entry->d_type != DT_REG) {
383 /* Don't add a file that we've added before. */
384 if (notmuch_filenames_valid (db_files) &&
385 strcmp (notmuch_filenames_get (db_files), entry->d_name) == 0)
387 notmuch_filenames_move_to_next (db_files);
391 /* We're now looking at a regular file that doesn't yet exist
392 * in the database, so add it. */
393 next = talloc_asprintf (notmuch, "%s/%s", path, entry->d_name);
395 state->processed_files++;
397 if (state->verbose) {
398 if (state->output_is_a_tty)
402 state->processed_files,
406 putchar((state->output_is_a_tty) ? '\r' : '\n');
410 status = notmuch_database_add_message (notmuch, next, &message);
413 case NOTMUCH_STATUS_SUCCESS:
414 state->added_messages++;
415 tag_inbox_and_unread (message);
417 /* Non-fatal issues (go on to next file) */
418 case NOTMUCH_STATUS_DUPLICATE_MESSAGE_ID:
419 /* Stay silent on this one. */
421 case NOTMUCH_STATUS_FILE_NOT_EMAIL:
422 fprintf (stderr, "Note: Ignoring non-mail file: %s\n",
425 /* Fatal issues. Don't process anymore. */
426 case NOTMUCH_STATUS_READ_ONLY_DATABASE:
427 case NOTMUCH_STATUS_XAPIAN_EXCEPTION:
428 case NOTMUCH_STATUS_OUT_OF_MEMORY:
429 fprintf (stderr, "Error: %s. Halting processing.\n",
430 notmuch_status_to_string (status));
434 case NOTMUCH_STATUS_FILE_ERROR:
435 case NOTMUCH_STATUS_NULL_POINTER:
436 case NOTMUCH_STATUS_TAG_TOO_LONG:
437 case NOTMUCH_STATUS_UNBALANCED_FREEZE_THAW:
438 case NOTMUCH_STATUS_LAST_STATUS:
439 INTERNAL_ERROR ("add_message returned unexpected value: %d", status);
444 notmuch_message_destroy (message);
448 if (do_add_files_print_progress) {
449 do_add_files_print_progress = 0;
450 add_files_print_progress (state);
460 /* Now that we've walked the whole filesystem list, anything left
461 * over in the database lists has been deleted. */
462 while (notmuch_filenames_valid (db_files))
464 char *absolute = talloc_asprintf (state->removed_files,
466 notmuch_filenames_get (db_files));
468 _filename_list_add (state->removed_files, absolute);
470 notmuch_filenames_move_to_next (db_files);
473 while (notmuch_filenames_valid (db_subdirs))
475 char *absolute = talloc_asprintf (state->removed_directories,
477 notmuch_filenames_get (db_subdirs));
479 _filename_list_add (state->removed_directories, absolute);
481 notmuch_filenames_move_to_next (db_subdirs);
485 status = notmuch_directory_set_mtime (directory, fs_mtime);
486 if (status && ret == NOTMUCH_STATUS_SUCCESS)
500 notmuch_filenames_destroy (db_subdirs);
502 notmuch_filenames_destroy (db_files);
504 notmuch_directory_destroy (directory);
509 /* This is the top-level entry point for add_files. It does a couple
510 * of error checks, sets up the progress-printing timer and then calls
511 * into the recursive function. */
512 static notmuch_status_t
513 add_files (notmuch_database_t *notmuch,
515 add_files_state_t *state)
517 notmuch_status_t status;
518 struct sigaction action;
519 struct itimerval timerval;
520 notmuch_bool_t timer_is_active = FALSE;
523 if (state->output_is_a_tty && ! debugger_is_active () && ! state->verbose) {
524 /* Setup our handler for SIGALRM */
525 memset (&action, 0, sizeof (struct sigaction));
526 action.sa_handler = handle_sigalrm;
527 sigemptyset (&action.sa_mask);
528 action.sa_flags = SA_RESTART;
529 sigaction (SIGALRM, &action, NULL);
531 /* Then start a timer to send SIGALRM once per second. */
532 timerval.it_interval.tv_sec = 1;
533 timerval.it_interval.tv_usec = 0;
534 timerval.it_value.tv_sec = 1;
535 timerval.it_value.tv_usec = 0;
536 setitimer (ITIMER_REAL, &timerval, NULL);
538 timer_is_active = TRUE;
541 if (stat (path, &st)) {
542 fprintf (stderr, "Error reading directory %s: %s\n",
543 path, strerror (errno));
544 return NOTMUCH_STATUS_FILE_ERROR;
547 if (! S_ISDIR (st.st_mode)) {
548 fprintf (stderr, "Error: %s is not a directory.\n", path);
549 return NOTMUCH_STATUS_FILE_ERROR;
552 status = add_files_recursive (notmuch, path, state);
554 if (timer_is_active) {
555 /* Now stop the timer. */
556 timerval.it_interval.tv_sec = 0;
557 timerval.it_interval.tv_usec = 0;
558 timerval.it_value.tv_sec = 0;
559 timerval.it_value.tv_usec = 0;
560 setitimer (ITIMER_REAL, &timerval, NULL);
562 /* And disable the signal handler. */
563 action.sa_handler = SIG_IGN;
564 sigaction (SIGALRM, &action, NULL);
570 /* XXX: This should be merged with the add_files function since it
571 * shares a lot of logic with it. */
572 /* Recursively count all regular files in path and all sub-directories
573 * of path. The result is added to *count (which should be
574 * initialized to zero by the top-level caller before calling
577 count_files (const char *path, int *count)
579 struct dirent *entry = NULL;
582 struct dirent **fs_entries = NULL;
583 int num_fs_entries = scandir (path, &fs_entries, 0, dirent_sort_inode);
586 if (num_fs_entries == -1) {
587 fprintf (stderr, "Warning: failed to open directory %s: %s\n",
588 path, strerror (errno));
592 while (!interrupted) {
593 if (i == num_fs_entries)
596 entry = fs_entries[i++];
598 /* Ignore special directories to avoid infinite recursion.
599 * Also ignore the .notmuch directory.
601 /* XXX: Eventually we'll want more sophistication to let the
602 * user specify files to be ignored. */
603 if (strcmp (entry->d_name, ".") == 0 ||
604 strcmp (entry->d_name, "..") == 0 ||
605 strcmp (entry->d_name, ".notmuch") == 0)
610 if (asprintf (&next, "%s/%s", path, entry->d_name) == -1) {
612 fprintf (stderr, "Error descending from %s to %s: Out of memory\n",
613 path, entry->d_name);
619 if (S_ISREG (st.st_mode)) {
621 if (*count % 1000 == 0) {
622 printf ("Found %d files so far.\r", *count);
625 } else if (S_ISDIR (st.st_mode)) {
626 count_files (next, count);
640 upgrade_print_progress (void *closure,
643 add_files_state_t *state = closure;
645 printf ("Upgrading database: %.2f%% complete", progress * 100.0);
648 struct timeval tv_now;
649 double elapsed, time_remaining;
651 gettimeofday (&tv_now, NULL);
653 elapsed = notmuch_time_elapsed (state->tv_start, tv_now);
654 time_remaining = (elapsed / progress) * (1.0 - progress);
656 notmuch_time_print_formatted_seconds (time_remaining);
657 printf (" remaining)");
665 /* Recursively remove all filenames from the database referring to
666 * 'path' (or to any of its children). */
668 _remove_directory (void *ctx,
669 notmuch_database_t *notmuch,
674 notmuch_directory_t *directory;
675 notmuch_filenames_t *files, *subdirs;
676 notmuch_status_t status;
679 directory = notmuch_database_get_directory (notmuch, path);
681 for (files = notmuch_directory_get_child_files (directory);
682 notmuch_filenames_valid (files);
683 notmuch_filenames_move_to_next (files))
685 absolute = talloc_asprintf (ctx, "%s/%s", path,
686 notmuch_filenames_get (files));
687 status = notmuch_database_remove_message (notmuch, absolute);
688 if (status == NOTMUCH_STATUS_DUPLICATE_MESSAGE_ID)
689 *renamed_files = *renamed_files + 1;
691 *removed_files = *removed_files + 1;
692 talloc_free (absolute);
695 for (subdirs = notmuch_directory_get_child_directories (directory);
696 notmuch_filenames_valid (subdirs);
697 notmuch_filenames_move_to_next (subdirs))
699 absolute = talloc_asprintf (ctx, "%s/%s", path,
700 notmuch_filenames_get (subdirs));
701 _remove_directory (ctx, notmuch, absolute, renamed_files, removed_files);
702 talloc_free (absolute);
705 notmuch_directory_destroy (directory);
709 notmuch_new_command (void *ctx, int argc, char *argv[])
711 notmuch_config_t *config;
712 notmuch_database_t *notmuch;
713 add_files_state_t add_files_state;
715 struct timeval tv_now;
719 char *dot_notmuch_path;
720 struct sigaction action;
722 int renamed_files, removed_files;
723 notmuch_status_t status;
726 add_files_state.verbose = 0;
727 add_files_state.output_is_a_tty = isatty (fileno (stdout));
729 for (i = 0; i < argc && argv[i][0] == '-'; i++) {
730 if (STRNCMP_LITERAL (argv[i], "--verbose") == 0) {
731 add_files_state.verbose = 1;
733 fprintf (stderr, "Unrecognized option: %s\n", argv[i]);
738 config = notmuch_config_open (ctx, NULL, NULL);
742 db_path = notmuch_config_get_database_path (config);
744 dot_notmuch_path = talloc_asprintf (ctx, "%s/%s", db_path, ".notmuch");
746 if (stat (dot_notmuch_path, &st)) {
750 count_files (db_path, &count);
754 printf ("Found %d total files (that's not much mail).\n", count);
755 notmuch = notmuch_database_create (db_path);
756 add_files_state.total_files = count;
758 notmuch = notmuch_database_open (db_path,
759 NOTMUCH_DATABASE_MODE_READ_WRITE);
763 if (notmuch_database_needs_upgrade (notmuch)) {
764 printf ("Welcome to a new version of notmuch! Your database will now be upgraded.\n");
765 gettimeofday (&add_files_state.tv_start, NULL);
766 notmuch_database_upgrade (notmuch, upgrade_print_progress,
768 printf ("Your notmuch database has now been upgraded to database format version %u.\n",
769 notmuch_database_get_version (notmuch));
772 add_files_state.total_files = 0;
778 /* Setup our handler for SIGINT. We do this after having
779 * potentially done a database upgrade we this interrupt handler
781 memset (&action, 0, sizeof (struct sigaction));
782 action.sa_handler = handle_sigint;
783 sigemptyset (&action.sa_mask);
784 action.sa_flags = SA_RESTART;
785 sigaction (SIGINT, &action, NULL);
787 talloc_free (dot_notmuch_path);
788 dot_notmuch_path = NULL;
790 add_files_state.processed_files = 0;
791 add_files_state.added_messages = 0;
792 gettimeofday (&add_files_state.tv_start, NULL);
794 add_files_state.removed_files = _filename_list_create (ctx);
795 add_files_state.removed_directories = _filename_list_create (ctx);
797 ret = add_files (notmuch, db_path, &add_files_state);
801 for (f = add_files_state.removed_files->head; f; f = f->next) {
802 status = notmuch_database_remove_message (notmuch, f->filename);
803 if (status == NOTMUCH_STATUS_DUPLICATE_MESSAGE_ID)
809 for (f = add_files_state.removed_directories->head; f; f = f->next) {
810 _remove_directory (ctx, notmuch, f->filename,
811 &renamed_files, &removed_files);
814 talloc_free (add_files_state.removed_files);
815 talloc_free (add_files_state.removed_directories);
817 gettimeofday (&tv_now, NULL);
818 elapsed = notmuch_time_elapsed (add_files_state.tv_start,
821 if (add_files_state.processed_files) {
822 printf ("Processed %d %s in ", add_files_state.processed_files,
823 add_files_state.processed_files == 1 ?
824 "file" : "total files");
825 notmuch_time_print_formatted_seconds (elapsed);
827 printf (" (%d files/sec.). \n",
828 (int) (add_files_state.processed_files / elapsed));
834 if (add_files_state.added_messages) {
835 printf ("Added %d new %s to the database.",
836 add_files_state.added_messages,
837 add_files_state.added_messages == 1 ?
838 "message" : "messages");
840 printf ("No new mail.");
844 printf (" Removed %d %s.",
846 removed_files == 1 ? "message" : "messages");
850 printf (" Detected %d file %s.",
852 renamed_files == 1 ? "rename" : "renames");
858 printf ("\nNote: At least one error was encountered: %s\n",
859 notmuch_status_to_string (ret));
862 notmuch_database_close (notmuch);
864 return ret || interrupted;