2009-11-10 21:03:05 +01:00
|
|
|
/* notmuch - Not much of an email program, (just index and search)
|
|
|
|
*
|
|
|
|
* Copyright © 2009 Carl Worth
|
|
|
|
*
|
|
|
|
* This program is free software: you can redistribute it and/or modify
|
|
|
|
* it under the terms of the GNU General Public License as published by
|
|
|
|
* the Free Software Foundation, either version 3 of the License, or
|
|
|
|
* (at your option) any later version.
|
|
|
|
*
|
|
|
|
* This program is distributed in the hope that it will be useful,
|
|
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
|
|
* GNU General Public License for more details.
|
|
|
|
*
|
|
|
|
* You should have received a copy of the GNU General Public License
|
|
|
|
* along with this program. If not, see http://www.gnu.org/licenses/ .
|
|
|
|
*
|
|
|
|
* Author: Carl Worth <cworth@cworth.org>
|
|
|
|
*/
|
|
|
|
|
|
|
|
#include "notmuch-client.h"
|
|
|
|
|
2009-11-22 01:44:32 +01:00
|
|
|
#include <unistd.h>
|
|
|
|
|
2010-01-06 00:11:21 +01:00
|
|
|
typedef struct {
|
|
|
|
int output_is_a_tty;
|
|
|
|
int verbose;
|
|
|
|
|
|
|
|
int total_files;
|
|
|
|
int processed_files;
|
|
|
|
int added_messages;
|
|
|
|
struct timeval tv_start;
|
|
|
|
} add_files_state_t;
|
|
|
|
|
2009-11-12 04:50:15 +01:00
|
|
|
static volatile sig_atomic_t do_add_files_print_progress = 0;
|
|
|
|
|
|
|
|
static void
|
|
|
|
handle_sigalrm (unused (int signal))
|
|
|
|
{
|
|
|
|
do_add_files_print_progress = 1;
|
|
|
|
}
|
|
|
|
|
2009-11-13 18:05:13 +01:00
|
|
|
static volatile sig_atomic_t interrupted;
|
|
|
|
|
|
|
|
static void
|
|
|
|
handle_sigint (unused (int sig))
|
|
|
|
{
|
2009-11-23 07:03:35 +01:00
|
|
|
ssize_t ignored;
|
2009-11-13 18:05:13 +01:00
|
|
|
static char msg[] = "Stopping... \n";
|
2009-11-23 07:03:35 +01:00
|
|
|
|
|
|
|
ignored = write(2, msg, sizeof(msg)-1);
|
2009-11-13 18:05:13 +01:00
|
|
|
interrupted = 1;
|
|
|
|
}
|
|
|
|
|
2009-11-10 21:03:05 +01:00
|
|
|
static void
|
|
|
|
tag_inbox_and_unread (notmuch_message_t *message)
|
|
|
|
{
|
|
|
|
notmuch_message_add_tag (message, "inbox");
|
|
|
|
notmuch_message_add_tag (message, "unread");
|
|
|
|
}
|
|
|
|
|
2009-11-12 04:50:15 +01:00
|
|
|
static void
|
|
|
|
add_files_print_progress (add_files_state_t *state)
|
|
|
|
{
|
|
|
|
struct timeval tv_now;
|
|
|
|
double elapsed_overall, rate_overall;
|
|
|
|
|
|
|
|
gettimeofday (&tv_now, NULL);
|
|
|
|
|
|
|
|
elapsed_overall = notmuch_time_elapsed (state->tv_start, tv_now);
|
|
|
|
rate_overall = (state->processed_files) / elapsed_overall;
|
|
|
|
|
|
|
|
printf ("Processed %d", state->processed_files);
|
|
|
|
|
|
|
|
if (state->total_files) {
|
|
|
|
double time_remaining;
|
|
|
|
|
|
|
|
time_remaining = ((state->total_files - state->processed_files) /
|
|
|
|
rate_overall);
|
|
|
|
printf (" of %d files (", state->total_files);
|
|
|
|
notmuch_time_print_formatted_seconds (time_remaining);
|
|
|
|
printf (" remaining). \r");
|
|
|
|
} else {
|
|
|
|
printf (" files (%d files/sec.) \r", (int) rate_overall);
|
|
|
|
}
|
|
|
|
|
|
|
|
fflush (stdout);
|
|
|
|
}
|
|
|
|
|
2010-01-06 02:43:03 +01:00
|
|
|
static int
|
|
|
|
dirent_sort_inode (const struct dirent **a, const struct dirent **b)
|
Read mail directory in inode number order
This gives a rather decent reduction in number of seeks required when
reading a Maildir that isn't in pagecache.
Most filesystems give some locality on disk based on inode numbers.
In ext[234] this is the inode tables, in XFS groups of sequential inode
numbers are together on disk and the most significant bits indicate
allocation group (i.e inode 1,000,000 is always after inode 1,000).
With this patch, we read in the whole directory, sort by inode number
before stat()ing the contents.
Ideally, directory is sequential and then we make one scan through the
file system stat()ing.
Since the universe is not ideal, we'll probably seek during reading the
directory and a fair bit while reading the inodes themselves.
However... with readahead, and stat()ing in inode order, we should be
in the best place possible to hit the cache.
In a (not very good) benchmark of "how long does it take to find the first
15,000 messages in my Maildir after 'echo 3 > /proc/sys/vm/drop_caches'",
this patch consistently cut at least 8 seconds off the scan time.
Without patch: 50 seconds
With patch: 38-42 seconds.
(I did this in a previous maildir reading project and saw large improvements too)
2009-11-18 02:56:40 +01:00
|
|
|
{
|
2009-11-18 22:27:32 +01:00
|
|
|
return ((*a)->d_ino < (*b)->d_ino) ? -1 : 1;
|
Read mail directory in inode number order
This gives a rather decent reduction in number of seeks required when
reading a Maildir that isn't in pagecache.
Most filesystems give some locality on disk based on inode numbers.
In ext[234] this is the inode tables, in XFS groups of sequential inode
numbers are together on disk and the most significant bits indicate
allocation group (i.e inode 1,000,000 is always after inode 1,000).
With this patch, we read in the whole directory, sort by inode number
before stat()ing the contents.
Ideally, directory is sequential and then we make one scan through the
file system stat()ing.
Since the universe is not ideal, we'll probably seek during reading the
directory and a fair bit while reading the inodes themselves.
However... with readahead, and stat()ing in inode order, we should be
in the best place possible to hit the cache.
In a (not very good) benchmark of "how long does it take to find the first
15,000 messages in my Maildir after 'echo 3 > /proc/sys/vm/drop_caches'",
this patch consistently cut at least 8 seconds off the scan time.
Without patch: 50 seconds
With patch: 38-42 seconds.
(I did this in a previous maildir reading project and saw large improvements too)
2009-11-18 02:56:40 +01:00
|
|
|
}
|
|
|
|
|
2010-01-06 02:43:03 +01:00
|
|
|
static int
|
|
|
|
dirent_sort_strcmp_name (const struct dirent **a, const struct dirent **b)
|
|
|
|
{
|
|
|
|
return strcmp ((*a)->d_name, (*b)->d_name);
|
|
|
|
}
|
|
|
|
|
2009-11-25 22:11:45 +01:00
|
|
|
/* Test if the directory looks like a Maildir directory.
|
|
|
|
*
|
|
|
|
* Search through the array of directory entries to see if we can find all
|
|
|
|
* three subdirectories typical for Maildir, that is "new", "cur", and "tmp".
|
|
|
|
*
|
|
|
|
* Return 1 if the directory looks like a Maildir and 0 otherwise.
|
|
|
|
*/
|
2009-11-28 04:38:46 +01:00
|
|
|
static int
|
2010-01-06 01:42:14 +01:00
|
|
|
_entries_resemble_maildir (struct dirent **entries, int count)
|
2009-11-25 22:11:45 +01:00
|
|
|
{
|
|
|
|
int i, found = 0;
|
|
|
|
|
|
|
|
for (i = 0; i < count; i++) {
|
2010-01-06 01:42:14 +01:00
|
|
|
if (entries[i]->d_type != DT_DIR)
|
|
|
|
continue;
|
|
|
|
|
2009-11-25 22:11:45 +01:00
|
|
|
if (strcmp(entries[i]->d_name, "new") == 0 ||
|
|
|
|
strcmp(entries[i]->d_name, "cur") == 0 ||
|
|
|
|
strcmp(entries[i]->d_name, "tmp") == 0)
|
|
|
|
{
|
|
|
|
found++;
|
2009-11-28 04:38:46 +01:00
|
|
|
if (found == 3)
|
|
|
|
return 1;
|
2009-11-25 22:11:45 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2009-11-12 04:50:15 +01:00
|
|
|
/* Examine 'path' recursively as follows:
|
|
|
|
*
|
2010-01-06 00:31:56 +01:00
|
|
|
* o Ask the filesystem for the mtime of 'path' (fs_mtime)
|
|
|
|
* o Ask the database for its timestamp of 'path' (db_mtime)
|
2009-11-12 04:50:15 +01:00
|
|
|
*
|
2010-01-06 02:43:03 +01:00
|
|
|
* o Ask the filesystem for files and directories within 'path'
|
|
|
|
* (via scandir and stored in fs_entries)
|
|
|
|
* o Ask the database for files and directories within 'path'
|
|
|
|
* (db_files and db_subdirs)
|
2010-01-06 01:35:02 +01:00
|
|
|
*
|
2010-01-06 02:43:03 +01:00
|
|
|
* o Pass 1: For each directory in fs_entries, recursively call into
|
|
|
|
* this same function.
|
2009-11-12 04:50:15 +01:00
|
|
|
*
|
2010-01-06 02:43:03 +01:00
|
|
|
* o Pass 2: If 'fs_mtime' > 'db_mtime', then walk fs_entries
|
|
|
|
* simultaneously with db_files and db_subdirs. Look for one of
|
|
|
|
* three interesting cases:
|
2009-11-12 04:50:15 +01:00
|
|
|
*
|
2010-01-06 02:43:03 +01:00
|
|
|
* 1. Regular file in fs_entries and not in db_files
|
|
|
|
* This is a new file to add_message into the database.
|
2009-11-12 04:50:15 +01:00
|
|
|
*
|
2010-01-06 02:43:03 +01:00
|
|
|
* 2. Filename in db_files not in fs_entries.
|
|
|
|
* This is a file that has been removed from the mail store.
|
|
|
|
*
|
|
|
|
* 3. Directory in db_subdirs not in fs_entries
|
|
|
|
* This is a directory that has been removed from the mail store.
|
|
|
|
*
|
|
|
|
* Note that the addition of a directory is not interesting here,
|
|
|
|
* since that will have been taken care of in pass 1. Also, we
|
|
|
|
* don't immediately act on file/directory removal since we must
|
|
|
|
* ensure that in the case of a rename that the new filename is
|
|
|
|
* added before the old filename is removed, (so that no
|
|
|
|
* information is lost from the database).
|
|
|
|
*
|
|
|
|
* o Tell the database to update its time of 'path' to 'fs_mtime'
|
2009-11-12 04:50:15 +01:00
|
|
|
*/
|
|
|
|
static notmuch_status_t
|
|
|
|
add_files_recursive (notmuch_database_t *notmuch,
|
|
|
|
const char *path,
|
|
|
|
add_files_state_t *state)
|
|
|
|
{
|
|
|
|
DIR *dir = NULL;
|
Read mail directory in inode number order
This gives a rather decent reduction in number of seeks required when
reading a Maildir that isn't in pagecache.
Most filesystems give some locality on disk based on inode numbers.
In ext[234] this is the inode tables, in XFS groups of sequential inode
numbers are together on disk and the most significant bits indicate
allocation group (i.e inode 1,000,000 is always after inode 1,000).
With this patch, we read in the whole directory, sort by inode number
before stat()ing the contents.
Ideally, directory is sequential and then we make one scan through the
file system stat()ing.
Since the universe is not ideal, we'll probably seek during reading the
directory and a fair bit while reading the inodes themselves.
However... with readahead, and stat()ing in inode order, we should be
in the best place possible to hit the cache.
In a (not very good) benchmark of "how long does it take to find the first
15,000 messages in my Maildir after 'echo 3 > /proc/sys/vm/drop_caches'",
this patch consistently cut at least 8 seconds off the scan time.
Without patch: 50 seconds
With patch: 38-42 seconds.
(I did this in a previous maildir reading project and saw large improvements too)
2009-11-18 02:56:40 +01:00
|
|
|
struct dirent *entry = NULL;
|
2009-11-12 04:50:15 +01:00
|
|
|
char *next = NULL;
|
2010-01-06 00:31:56 +01:00
|
|
|
time_t fs_mtime, db_mtime;
|
2009-11-12 04:50:15 +01:00
|
|
|
notmuch_status_t status, ret = NOTMUCH_STATUS_SUCCESS;
|
|
|
|
notmuch_message_t *message = NULL;
|
2010-01-06 01:15:43 +01:00
|
|
|
struct dirent **fs_entries = NULL;
|
2010-01-06 01:35:02 +01:00
|
|
|
int i, num_fs_entries;
|
2010-01-05 22:29:23 +01:00
|
|
|
notmuch_directory_t *directory;
|
2010-01-06 02:43:03 +01:00
|
|
|
notmuch_filenames_t *db_files = NULL;
|
|
|
|
notmuch_filenames_t *db_subdirs = NULL;
|
2010-01-06 01:06:46 +01:00
|
|
|
struct stat st;
|
2010-01-06 01:42:14 +01:00
|
|
|
notmuch_bool_t is_maildir;
|
2009-11-12 04:50:15 +01:00
|
|
|
|
2010-01-06 01:06:46 +01:00
|
|
|
if (stat (path, &st)) {
|
|
|
|
fprintf (stderr, "Error reading directory %s: %s\n",
|
|
|
|
path, strerror (errno));
|
|
|
|
return NOTMUCH_STATUS_FILE_ERROR;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (! S_ISDIR (st.st_mode)) {
|
|
|
|
fprintf (stderr, "Error: %s is not a directory.\n", path);
|
|
|
|
return NOTMUCH_STATUS_FILE_ERROR;
|
|
|
|
}
|
|
|
|
|
|
|
|
fs_mtime = st.st_mtime;
|
2009-11-12 04:50:15 +01:00
|
|
|
|
2010-01-05 22:29:23 +01:00
|
|
|
directory = notmuch_database_get_directory (notmuch, path);
|
2010-01-06 00:31:56 +01:00
|
|
|
db_mtime = notmuch_directory_get_mtime (directory);
|
2010-01-05 22:29:23 +01:00
|
|
|
|
2010-01-06 02:43:03 +01:00
|
|
|
/* If the database knows about this directory, then we sort based
|
|
|
|
* on strcmp to match the database sorting. Otherwise, we can do
|
|
|
|
* inode-based sorting for faster filesystem operation. */
|
|
|
|
num_fs_entries = scandir (path, &fs_entries, 0,
|
|
|
|
db_mtime ?
|
|
|
|
dirent_sort_strcmp_name : dirent_sort_inode);
|
2009-11-12 04:50:15 +01:00
|
|
|
|
2010-01-06 01:15:43 +01:00
|
|
|
if (num_fs_entries == -1) {
|
2009-11-12 04:50:15 +01:00
|
|
|
fprintf (stderr, "Error opening directory %s: %s\n",
|
|
|
|
path, strerror (errno));
|
|
|
|
ret = NOTMUCH_STATUS_FILE_ERROR;
|
|
|
|
goto DONE;
|
|
|
|
}
|
|
|
|
|
2010-01-06 02:43:03 +01:00
|
|
|
/* Pass 1: Recurse into all sub-directories. */
|
2010-01-06 01:42:14 +01:00
|
|
|
is_maildir = _entries_resemble_maildir (fs_entries, num_fs_entries);
|
|
|
|
|
2010-01-06 01:35:02 +01:00
|
|
|
for (i = 0; i < num_fs_entries; i++) {
|
|
|
|
if (interrupted)
|
2009-11-12 04:50:15 +01:00
|
|
|
break;
|
|
|
|
|
2010-01-06 01:35:02 +01:00
|
|
|
entry = fs_entries[i];
|
Read mail directory in inode number order
This gives a rather decent reduction in number of seeks required when
reading a Maildir that isn't in pagecache.
Most filesystems give some locality on disk based on inode numbers.
In ext[234] this is the inode tables, in XFS groups of sequential inode
numbers are together on disk and the most significant bits indicate
allocation group (i.e inode 1,000,000 is always after inode 1,000).
With this patch, we read in the whole directory, sort by inode number
before stat()ing the contents.
Ideally, directory is sequential and then we make one scan through the
file system stat()ing.
Since the universe is not ideal, we'll probably seek during reading the
directory and a fair bit while reading the inodes themselves.
However... with readahead, and stat()ing in inode order, we should be
in the best place possible to hit the cache.
In a (not very good) benchmark of "how long does it take to find the first
15,000 messages in my Maildir after 'echo 3 > /proc/sys/vm/drop_caches'",
this patch consistently cut at least 8 seconds off the scan time.
Without patch: 50 seconds
With patch: 38-42 seconds.
(I did this in a previous maildir reading project and saw large improvements too)
2009-11-18 02:56:40 +01:00
|
|
|
|
2010-01-06 01:35:02 +01:00
|
|
|
if (entry->d_type != DT_DIR)
|
2009-11-12 04:50:15 +01:00
|
|
|
continue;
|
|
|
|
|
|
|
|
/* Ignore special directories to avoid infinite recursion.
|
2010-01-06 01:42:14 +01:00
|
|
|
* Also ignore the .notmuch directory and any "tmp" directory
|
|
|
|
* that appears within a maildir.
|
2009-11-12 04:50:15 +01:00
|
|
|
*/
|
|
|
|
/* XXX: Eventually we'll want more sophistication to let the
|
|
|
|
* user specify files to be ignored. */
|
|
|
|
if (strcmp (entry->d_name, ".") == 0 ||
|
|
|
|
strcmp (entry->d_name, "..") == 0 ||
|
2010-01-06 01:42:14 +01:00
|
|
|
(is_maildir && strcmp (entry->d_name, "tmp") == 0) ||
|
2009-11-12 04:50:15 +01:00
|
|
|
strcmp (entry->d_name, ".notmuch") ==0)
|
|
|
|
{
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
next = talloc_asprintf (notmuch, "%s/%s", path, entry->d_name);
|
2010-01-06 01:35:02 +01:00
|
|
|
status = add_files_recursive (notmuch, next, state);
|
|
|
|
if (status && ret == NOTMUCH_STATUS_SUCCESS)
|
|
|
|
ret = status;
|
|
|
|
talloc_free (next);
|
|
|
|
next = NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* If this directory hasn't been modified since the last
|
2010-01-06 02:43:03 +01:00
|
|
|
* "notmuch new", then we can skip the second pass entirely. */
|
2010-01-06 01:35:02 +01:00
|
|
|
if (fs_mtime <= db_mtime)
|
|
|
|
goto DONE;
|
|
|
|
|
2010-01-06 02:43:03 +01:00
|
|
|
/* Pass 2: Scan for new files, removed files, and removed directories. */
|
|
|
|
db_files = notmuch_directory_get_child_files (directory);
|
|
|
|
db_subdirs = notmuch_directory_get_child_directories (directory);
|
|
|
|
|
|
|
|
for (i = 0; i < num_fs_entries; i++)
|
|
|
|
{
|
2010-01-06 01:35:02 +01:00
|
|
|
if (interrupted)
|
|
|
|
break;
|
|
|
|
|
|
|
|
entry = fs_entries[i];
|
|
|
|
|
2010-01-06 02:43:03 +01:00
|
|
|
/* Check if we've walked past any names in db_files or
|
|
|
|
* db_subdirs. If so, these have been deleted. */
|
|
|
|
while (notmuch_filenames_has_more (db_files) &&
|
|
|
|
strcmp (notmuch_filenames_get (db_files), entry->d_name) < 0)
|
|
|
|
{
|
|
|
|
printf ("Detected deleted file %s/%s\n", path,
|
|
|
|
notmuch_filenames_get (db_files));
|
|
|
|
|
|
|
|
notmuch_filenames_advance (db_files);
|
|
|
|
}
|
|
|
|
|
|
|
|
while (notmuch_filenames_has_more (db_subdirs) &&
|
|
|
|
strcmp (notmuch_filenames_get (db_subdirs), entry->d_name) <= 0)
|
|
|
|
{
|
|
|
|
if (strcmp (notmuch_filenames_get (db_subdirs), entry->d_name) < 0)
|
|
|
|
printf ("Detected deleted directory %s/%s", path,
|
|
|
|
notmuch_filenames_get (db_subdirs));
|
|
|
|
|
|
|
|
notmuch_filenames_advance (db_subdirs);
|
|
|
|
}
|
|
|
|
|
2010-01-06 01:35:02 +01:00
|
|
|
if (entry->d_type != DT_REG)
|
|
|
|
continue;
|
|
|
|
|
2010-01-06 02:43:03 +01:00
|
|
|
/* Don't add a file that we've added before. */
|
|
|
|
if (notmuch_filenames_has_more (db_files) &&
|
|
|
|
strcmp (notmuch_filenames_get (db_files), entry->d_name) == 0)
|
|
|
|
{
|
|
|
|
notmuch_filenames_advance (db_files);
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* We're not looking at a regular file that doesn't yet exist
|
|
|
|
* in the database, so add it. */
|
2010-01-06 01:35:02 +01:00
|
|
|
next = talloc_asprintf (notmuch, "%s/%s", path, entry->d_name);
|
|
|
|
|
|
|
|
state->processed_files++;
|
|
|
|
|
|
|
|
if (state->verbose) {
|
|
|
|
if (state->output_is_a_tty)
|
|
|
|
printf("\r\033[K");
|
|
|
|
|
|
|
|
printf ("%i/%i: %s",
|
|
|
|
state->processed_files,
|
|
|
|
state->total_files,
|
|
|
|
next);
|
|
|
|
|
|
|
|
putchar((state->output_is_a_tty) ? '\r' : '\n');
|
|
|
|
fflush (stdout);
|
|
|
|
}
|
|
|
|
|
|
|
|
status = notmuch_database_add_message (notmuch, next, &message);
|
|
|
|
switch (status) {
|
|
|
|
/* success */
|
|
|
|
case NOTMUCH_STATUS_SUCCESS:
|
|
|
|
state->added_messages++;
|
|
|
|
tag_inbox_and_unread (message);
|
|
|
|
break;
|
|
|
|
/* Non-fatal issues (go on to next file) */
|
|
|
|
case NOTMUCH_STATUS_DUPLICATE_MESSAGE_ID:
|
|
|
|
/* Stay silent on this one. */
|
|
|
|
break;
|
|
|
|
case NOTMUCH_STATUS_FILE_NOT_EMAIL:
|
|
|
|
fprintf (stderr, "Note: Ignoring non-mail file: %s\n",
|
|
|
|
next);
|
|
|
|
break;
|
|
|
|
/* Fatal issues. Don't process anymore. */
|
|
|
|
case NOTMUCH_STATUS_READONLY_DATABASE:
|
|
|
|
case NOTMUCH_STATUS_XAPIAN_EXCEPTION:
|
|
|
|
case NOTMUCH_STATUS_OUT_OF_MEMORY:
|
|
|
|
fprintf (stderr, "Error: %s. Halting processing.\n",
|
|
|
|
notmuch_status_to_string (status));
|
|
|
|
ret = status;
|
|
|
|
goto DONE;
|
|
|
|
default:
|
|
|
|
case NOTMUCH_STATUS_FILE_ERROR:
|
|
|
|
case NOTMUCH_STATUS_NULL_POINTER:
|
|
|
|
case NOTMUCH_STATUS_TAG_TOO_LONG:
|
|
|
|
case NOTMUCH_STATUS_UNBALANCED_FREEZE_THAW:
|
|
|
|
case NOTMUCH_STATUS_LAST_STATUS:
|
|
|
|
INTERNAL_ERROR ("add_message returned unexpected value: %d", status);
|
|
|
|
goto DONE;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (message) {
|
|
|
|
notmuch_message_destroy (message);
|
|
|
|
message = NULL;
|
|
|
|
}
|
2009-11-12 04:50:15 +01:00
|
|
|
|
2010-01-06 01:35:02 +01:00
|
|
|
if (do_add_files_print_progress) {
|
|
|
|
do_add_files_print_progress = 0;
|
|
|
|
add_files_print_progress (state);
|
2009-11-12 04:50:15 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
talloc_free (next);
|
|
|
|
next = NULL;
|
|
|
|
}
|
|
|
|
|
2010-01-06 00:23:52 +01:00
|
|
|
if (! interrupted) {
|
2010-01-06 00:31:56 +01:00
|
|
|
status = notmuch_directory_set_mtime (directory, fs_mtime);
|
2010-01-06 00:23:52 +01:00
|
|
|
if (status && ret == NOTMUCH_STATUS_SUCCESS)
|
|
|
|
ret = status;
|
|
|
|
}
|
2009-11-12 04:50:15 +01:00
|
|
|
|
|
|
|
DONE:
|
|
|
|
if (next)
|
|
|
|
talloc_free (next);
|
|
|
|
if (entry)
|
|
|
|
free (entry);
|
|
|
|
if (dir)
|
|
|
|
closedir (dir);
|
2010-01-06 01:15:43 +01:00
|
|
|
if (fs_entries)
|
|
|
|
free (fs_entries);
|
2010-01-06 02:43:03 +01:00
|
|
|
if (db_subdirs)
|
|
|
|
notmuch_filenames_destroy (db_subdirs);
|
|
|
|
if (db_files)
|
|
|
|
notmuch_filenames_destroy (db_files);
|
|
|
|
if (directory)
|
|
|
|
notmuch_directory_destroy (directory);
|
2009-11-12 04:50:15 +01:00
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* This is the top-level entry point for add_files. It does a couple
|
|
|
|
* of error checks, sets up the progress-printing timer and then calls
|
|
|
|
* into the recursive function. */
|
2010-01-06 00:11:21 +01:00
|
|
|
static notmuch_status_t
|
2009-11-12 04:50:15 +01:00
|
|
|
add_files (notmuch_database_t *notmuch,
|
|
|
|
const char *path,
|
|
|
|
add_files_state_t *state)
|
|
|
|
{
|
|
|
|
notmuch_status_t status;
|
|
|
|
struct sigaction action;
|
|
|
|
struct itimerval timerval;
|
2009-11-22 01:44:31 +01:00
|
|
|
notmuch_bool_t timer_is_active = FALSE;
|
2009-11-12 04:50:15 +01:00
|
|
|
|
2009-11-23 01:54:35 +01:00
|
|
|
if (state->output_is_a_tty && ! debugger_is_active () && ! state->verbose) {
|
2009-11-21 01:17:18 +01:00
|
|
|
/* Setup our handler for SIGALRM */
|
2009-11-22 01:44:31 +01:00
|
|
|
memset (&action, 0, sizeof (struct sigaction));
|
|
|
|
action.sa_handler = handle_sigalrm;
|
|
|
|
sigemptyset (&action.sa_mask);
|
|
|
|
action.sa_flags = SA_RESTART;
|
|
|
|
sigaction (SIGALRM, &action, NULL);
|
|
|
|
|
|
|
|
/* Then start a timer to send SIGALRM once per second. */
|
|
|
|
timerval.it_interval.tv_sec = 1;
|
|
|
|
timerval.it_interval.tv_usec = 0;
|
|
|
|
timerval.it_value.tv_sec = 1;
|
|
|
|
timerval.it_value.tv_usec = 0;
|
|
|
|
setitimer (ITIMER_REAL, &timerval, NULL);
|
|
|
|
|
|
|
|
timer_is_active = TRUE;
|
|
|
|
}
|
2009-11-12 04:50:15 +01:00
|
|
|
|
2010-01-06 01:06:46 +01:00
|
|
|
status = add_files_recursive (notmuch, path, state);
|
2009-11-12 04:50:15 +01:00
|
|
|
|
2009-11-22 01:44:31 +01:00
|
|
|
if (timer_is_active) {
|
2009-11-21 01:17:18 +01:00
|
|
|
/* Now stop the timer. */
|
2009-11-22 01:44:31 +01:00
|
|
|
timerval.it_interval.tv_sec = 0;
|
|
|
|
timerval.it_interval.tv_usec = 0;
|
|
|
|
timerval.it_value.tv_sec = 0;
|
|
|
|
timerval.it_value.tv_usec = 0;
|
|
|
|
setitimer (ITIMER_REAL, &timerval, NULL);
|
|
|
|
|
|
|
|
/* And disable the signal handler. */
|
|
|
|
action.sa_handler = SIG_IGN;
|
|
|
|
sigaction (SIGALRM, &action, NULL);
|
|
|
|
}
|
2009-11-12 04:50:15 +01:00
|
|
|
|
|
|
|
return status;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* XXX: This should be merged with the add_files function since it
|
|
|
|
* shares a lot of logic with it. */
|
2009-11-18 00:23:42 +01:00
|
|
|
/* Recursively count all regular files in path and all sub-directories
|
2009-11-12 04:50:15 +01:00
|
|
|
* of path. The result is added to *count (which should be
|
|
|
|
* initialized to zero by the top-level caller before calling
|
|
|
|
* count_files). */
|
|
|
|
static void
|
|
|
|
count_files (const char *path, int *count)
|
|
|
|
{
|
2009-11-18 03:22:20 +01:00
|
|
|
struct dirent *entry = NULL;
|
2009-11-12 04:50:15 +01:00
|
|
|
char *next;
|
|
|
|
struct stat st;
|
2010-01-06 01:15:43 +01:00
|
|
|
struct dirent **fs_entries = NULL;
|
2010-01-06 02:43:03 +01:00
|
|
|
int num_fs_entries = scandir (path, &fs_entries, 0, dirent_sort_inode);
|
2009-11-18 03:22:20 +01:00
|
|
|
int i = 0;
|
2009-11-12 04:50:15 +01:00
|
|
|
|
2010-01-06 01:15:43 +01:00
|
|
|
if (num_fs_entries == -1) {
|
2009-11-12 04:50:15 +01:00
|
|
|
fprintf (stderr, "Warning: failed to open directory %s: %s\n",
|
|
|
|
path, strerror (errno));
|
|
|
|
goto DONE;
|
|
|
|
}
|
|
|
|
|
2009-11-13 18:05:13 +01:00
|
|
|
while (!interrupted) {
|
2010-01-06 01:15:43 +01:00
|
|
|
if (i == num_fs_entries)
|
2009-11-12 04:50:15 +01:00
|
|
|
break;
|
|
|
|
|
2010-01-06 01:15:43 +01:00
|
|
|
entry = fs_entries[i++];
|
2009-11-18 03:22:20 +01:00
|
|
|
|
2009-11-12 04:50:15 +01:00
|
|
|
/* Ignore special directories to avoid infinite recursion.
|
|
|
|
* Also ignore the .notmuch directory.
|
|
|
|
*/
|
|
|
|
/* XXX: Eventually we'll want more sophistication to let the
|
|
|
|
* user specify files to be ignored. */
|
|
|
|
if (strcmp (entry->d_name, ".") == 0 ||
|
|
|
|
strcmp (entry->d_name, "..") == 0 ||
|
|
|
|
strcmp (entry->d_name, ".notmuch") == 0)
|
|
|
|
{
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (asprintf (&next, "%s/%s", path, entry->d_name) == -1) {
|
|
|
|
next = NULL;
|
|
|
|
fprintf (stderr, "Error descending from %s to %s: Out of memory\n",
|
|
|
|
path, entry->d_name);
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
stat (next, &st);
|
|
|
|
|
|
|
|
if (S_ISREG (st.st_mode)) {
|
|
|
|
*count = *count + 1;
|
|
|
|
if (*count % 1000 == 0) {
|
|
|
|
printf ("Found %d files so far.\r", *count);
|
|
|
|
fflush (stdout);
|
|
|
|
}
|
|
|
|
} else if (S_ISDIR (st.st_mode)) {
|
|
|
|
count_files (next, count);
|
|
|
|
}
|
|
|
|
|
|
|
|
free (next);
|
|
|
|
}
|
|
|
|
|
|
|
|
DONE:
|
|
|
|
if (entry)
|
|
|
|
free (entry);
|
2010-01-06 01:15:43 +01:00
|
|
|
if (fs_entries)
|
|
|
|
free (fs_entries);
|
2009-11-12 04:50:15 +01:00
|
|
|
}
|
|
|
|
|
2009-11-10 21:03:05 +01:00
|
|
|
int
|
2009-11-21 01:17:18 +01:00
|
|
|
notmuch_new_command (void *ctx, int argc, char *argv[])
|
2009-11-10 21:03:05 +01:00
|
|
|
{
|
2009-11-12 04:50:15 +01:00
|
|
|
notmuch_config_t *config;
|
2009-11-10 21:03:05 +01:00
|
|
|
notmuch_database_t *notmuch;
|
|
|
|
add_files_state_t add_files_state;
|
|
|
|
double elapsed;
|
|
|
|
struct timeval tv_now;
|
|
|
|
int ret = 0;
|
2009-11-12 04:50:15 +01:00
|
|
|
struct stat st;
|
|
|
|
const char *db_path;
|
|
|
|
char *dot_notmuch_path;
|
2009-11-13 18:05:13 +01:00
|
|
|
struct sigaction action;
|
2009-11-21 01:17:18 +01:00
|
|
|
int i;
|
|
|
|
|
|
|
|
add_files_state.verbose = 0;
|
2009-11-23 01:54:35 +01:00
|
|
|
add_files_state.output_is_a_tty = isatty (fileno (stdout));
|
2009-11-21 01:17:18 +01:00
|
|
|
|
|
|
|
for (i = 0; i < argc && argv[i][0] == '-'; i++) {
|
|
|
|
if (STRNCMP_LITERAL (argv[i], "--verbose") == 0) {
|
|
|
|
add_files_state.verbose = 1;
|
|
|
|
} else {
|
|
|
|
fprintf (stderr, "Unrecognized option: %s\n", argv[i]);
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
}
|
2009-11-13 18:05:13 +01:00
|
|
|
|
|
|
|
/* Setup our handler for SIGINT */
|
|
|
|
memset (&action, 0, sizeof (struct sigaction));
|
|
|
|
action.sa_handler = handle_sigint;
|
|
|
|
sigemptyset (&action.sa_mask);
|
|
|
|
action.sa_flags = SA_RESTART;
|
|
|
|
sigaction (SIGINT, &action, NULL);
|
2009-11-10 21:03:05 +01:00
|
|
|
|
2009-11-12 04:50:15 +01:00
|
|
|
config = notmuch_config_open (ctx, NULL, NULL);
|
|
|
|
if (config == NULL)
|
|
|
|
return 1;
|
|
|
|
|
|
|
|
db_path = notmuch_config_get_database_path (config);
|
|
|
|
|
|
|
|
dot_notmuch_path = talloc_asprintf (ctx, "%s/%s", db_path, ".notmuch");
|
|
|
|
|
|
|
|
if (stat (dot_notmuch_path, &st)) {
|
2009-11-13 18:05:13 +01:00
|
|
|
int count;
|
|
|
|
|
|
|
|
count = 0;
|
|
|
|
count_files (db_path, &count);
|
|
|
|
if (interrupted)
|
|
|
|
return 1;
|
|
|
|
|
2009-12-15 01:06:37 +01:00
|
|
|
printf ("Found %d total files (that's not much mail).\n", count);
|
2009-11-12 04:50:15 +01:00
|
|
|
notmuch = notmuch_database_create (db_path);
|
2009-11-13 18:05:13 +01:00
|
|
|
add_files_state.total_files = count;
|
2009-11-12 04:50:15 +01:00
|
|
|
} else {
|
2009-11-21 20:54:25 +01:00
|
|
|
notmuch = notmuch_database_open (db_path,
|
2009-11-22 00:13:24 +01:00
|
|
|
NOTMUCH_DATABASE_MODE_READ_WRITE);
|
2009-11-13 18:05:13 +01:00
|
|
|
add_files_state.total_files = 0;
|
2009-11-10 21:03:05 +01:00
|
|
|
}
|
|
|
|
|
2009-11-12 04:50:15 +01:00
|
|
|
if (notmuch == NULL)
|
|
|
|
return 1;
|
|
|
|
|
|
|
|
talloc_free (dot_notmuch_path);
|
|
|
|
dot_notmuch_path = NULL;
|
|
|
|
|
2009-11-10 21:03:05 +01:00
|
|
|
add_files_state.processed_files = 0;
|
|
|
|
add_files_state.added_messages = 0;
|
|
|
|
gettimeofday (&add_files_state.tv_start, NULL);
|
|
|
|
|
2009-11-12 04:50:15 +01:00
|
|
|
ret = add_files (notmuch, db_path, &add_files_state);
|
2009-11-10 21:03:05 +01:00
|
|
|
|
|
|
|
gettimeofday (&tv_now, NULL);
|
|
|
|
elapsed = notmuch_time_elapsed (add_files_state.tv_start,
|
|
|
|
tv_now);
|
|
|
|
if (add_files_state.processed_files) {
|
|
|
|
printf ("Processed %d %s in ", add_files_state.processed_files,
|
|
|
|
add_files_state.processed_files == 1 ?
|
|
|
|
"file" : "total files");
|
|
|
|
notmuch_time_print_formatted_seconds (elapsed);
|
|
|
|
if (elapsed > 1) {
|
|
|
|
printf (" (%d files/sec.). \n",
|
|
|
|
(int) (add_files_state.processed_files / elapsed));
|
|
|
|
} else {
|
|
|
|
printf (". \n");
|
|
|
|
}
|
|
|
|
}
|
|
|
|
if (add_files_state.added_messages) {
|
2009-12-15 01:06:37 +01:00
|
|
|
printf ("Added %d new %s to the database.\n",
|
2009-11-10 21:03:05 +01:00
|
|
|
add_files_state.added_messages,
|
|
|
|
add_files_state.added_messages == 1 ?
|
|
|
|
"message" : "messages");
|
|
|
|
} else {
|
2009-12-15 01:06:37 +01:00
|
|
|
printf ("No new mail.\n");
|
2009-11-10 21:03:05 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
if (ret) {
|
|
|
|
printf ("\nNote: At least one error was encountered: %s\n",
|
|
|
|
notmuch_status_to_string (ret));
|
|
|
|
}
|
|
|
|
|
2009-11-12 04:50:15 +01:00
|
|
|
notmuch_database_close (notmuch);
|
2009-11-10 21:03:05 +01:00
|
|
|
|
2009-11-13 18:05:13 +01:00
|
|
|
return ret || interrupted;
|
2009-11-10 21:03:05 +01:00
|
|
|
}
|