2:23 p.m.
I ran into an issue where I would read an email in my inbox that was
sent to me and cc'd a mailing list and it would not remove the 'unread'
tag after syncing the mailbox.
Apparently, when nm_synci'ing the message, the local inbox copy would be
recognized as seen but the duplicate copy in the mailing list folder
would still be in the 'new' directory, so the 'unread' flag would not
be cleared.
The way the 'duplicate' messags where handled in 'rename_filename', it
appears after the add/remove the local copy of 'msg' was not updated
with the new paths. I fixed this by destorying the message and then
re-reading it from the notmuch database.
However, the tags were still off. I had to further call
'update_header_flags' to re-read the tags from the database (using
the updated paths) to clear the 'unread' tag.
Both of these changes seem to work as expected now.
Signed-off-by: Don Zickus <dzickus(a)redhat.com>
--
I am not too familar with this code, so my changes may not be
correct. I do know they seem to work after a couple of days of
testing.
---
mutt_notmuch.c | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/mutt_notmuch.c b/mutt_notmuch.c
index 1a39249..3a25cc1 100644
--- a/mutt_notmuch.c
+++ b/mutt_notmuch.c
@@ -1474,6 +1474,9 @@ static int rename_filename(struct nm_ctxdata *data,
notmuch_database_add_message(db, newpath, NULL);
}
}
+ notmuch_message_destroy(msg);
+ msg = NULL;
+ notmuch_database_find_message_by_filename(db, new, &msg);
st = NOTMUCH_STATUS_SUCCESS;
break;
default:
@@ -1484,6 +1487,7 @@ static int rename_filename(struct nm_ctxdata *data,
if (st == NOTMUCH_STATUS_SUCCESS && h && msg) {
notmuch_message_maildir_flags_to_tags(msg);
+ update_header_tags(h, msg);
update_tags(msg, nm_header_get_tags(h));
}
--
1.8.4
2:23 p.m.
New subject: [PATCH 2/2] mutt, notmuch: Update to use notmuch api version 4
This is a quick and dirty hack to utilize notmuch api version 4.
The main problem is the deprecated apis. I believe those apis work
correctly in version 3, so perhaps just changing the api and leaving
the version definitions alone is sufficient. Or I can further
pollute the code with more #ifdef stuff.
I don't know which way to go here. Let me know and I can respin a
proper patch.
Cheers,
Don
---
configure.ac | 2 +-
mutt_notmuch.c | 25 ++++++++++++++++---------
2 files changed, 17 insertions(+), 10 deletions(-)
diff --git a/configure.ac b/configure.ac
index 6e14662..cd9f864 100644
--- a/configure.ac
+++ b/configure.ac
@@ -189,7 +189,7 @@ AC_ARG_ENABLE(notmuch, AC_HELP_STRING([--enable-notmuch], [Enable
NOTMUCH suppor
[[notmuch_database_open("/path", NOTMUCH_DATABASE_MODE_READ_ONLY,
(notmuch_database_t**)NULL);]]
)],
[notmuch_api_3=yes
- AC_DEFINE([NOTMUCH_API_3], 1, [Define to 1 if you have the notmuch api version 3.])
+ AC_DEFINE([NOTMUCH_API_4], 1, [Define to 1 if you have the notmuch api version 3.])
],
[notmuch_api_3=no]
)
diff --git a/mutt_notmuch.c b/mutt_notmuch.c
index 3a25cc1..966037d 100644
--- a/mutt_notmuch.c
+++ b/mutt_notmuch.c
@@ -230,7 +230,7 @@ static void free_ctxdata(struct nm_ctxdata *data)
dprint(1, (debugfile, "nm: freeing context data %p\n", data));
if (data->db)
-#ifdef NOTMUCH_API_3
+#if defined NOTMUCH_API_3 || defined NOTMUCH_API_4
notmuch_database_destroy(data->db);
#else
notmuch_database_close(data->db);
@@ -454,7 +454,7 @@ static notmuch_database_t *do_database_open(const char *filename,
dprint(1, (debugfile, "nm: db open '%s' %s (timeout %d)\n", filename,
writable ? "[WRITE]" : "[READ]", NotmuchOpenTimeout));
do {
-#ifdef NOTMUCH_API_3
+#if defined NOTMUCH_API_3 || defined NOTMUCH_API_4
st = notmuch_database_open(filename,
writable ? NOTMUCH_DATABASE_MODE_READ_WRITE :
NOTMUCH_DATABASE_MODE_READ_ONLY, &db);
@@ -501,7 +501,8 @@ static int release_db(struct nm_ctxdata *data)
{
if (data && data->db) {
dprint(1, (debugfile, "nm: db close\n"));
-#ifdef NOTMUCH_API_3
+#if defined NOTMUCH_API_3 || defined NOTMUCH_API_4
+
notmuch_database_destroy(data->db);
#else
notmuch_database_close(data->db);
@@ -930,10 +931,13 @@ static void nm_progress_update(CONTEXT *ctx, notmuch_query_t *q)
if (!data->progress_ready && q) {
static char msg[STRING];
+ unsigned int count;
+
snprintf(msg, sizeof(msg), _("Reading messages..."));
+ notmuch_query_count_messages_st(q, &count);
mutt_progress_init(&data->progress, msg, M_PROGRESS_MSG,
- ReadInc, notmuch_query_count_messages(q));
+ ReadInc, count);
data->progress_ready = 1;
}
@@ -1082,7 +1086,8 @@ static void read_mesgs_query(CONTEXT *ctx, notmuch_query_t *q, int
dedup)
limit = get_limit(data);
- for (msgs = notmuch_query_search_messages(q);
+ notmuch_query_search_messages_st(q, &msgs);
+ for (;
notmuch_messages_valid(msgs) &&
(limit == 0 || ctx->msgcount < limit);
notmuch_messages_move_to_next(msgs)) {
@@ -1101,7 +1106,8 @@ static void read_threads_query(CONTEXT *ctx, notmuch_query_t *q, int
dedup, int
if (!data)
return;
- for (threads = notmuch_query_search_threads(q);
+ notmuch_query_search_threads_st(q, &threads);
+ for (;
notmuch_threads_valid(threads) &&
(limit == 0 || ctx->msgcount < limit);
notmuch_threads_move_to_next(threads)) {
@@ -1605,7 +1611,7 @@ static unsigned count_query(notmuch_database_t *db, const char
*qstr)
if (q) {
apply_exclude_tags(q);
- res = notmuch_query_count_messages(q);
+ notmuch_query_count_messages_st(q, &res);
notmuch_query_destroy(q);
dprint(1, (debugfile, "nm: count '%s', result=%d\n", qstr, res));
}
@@ -1672,7 +1678,7 @@ int nm_nonctx_get_count(char *path, int *all, int *new)
rc = 0;
done:
if (db) {
-#ifdef NOTMUCH_API_3
+#if defined NOTMUCH_API_3 || defined NOTMUCH_API_4
notmuch_database_destroy(db);
#else
notmuch_database_close(db);
@@ -1779,7 +1785,8 @@ int nm_check_database(CONTEXT *ctx, int *index_hint)
limit = get_limit(data);
- for (i = 0, msgs = notmuch_query_search_messages(q);
+ notmuch_query_search_messages_st(q, &msgs);
+ for (i = 0;
notmuch_messages_valid(msgs) && (limit == 0 || i < limit);
notmuch_messages_move_to_next(msgs), i++) {
--
1.8.4
8:19 a.m.
New subject: [PATCH 2/2] mutt, notmuch: Update to use notmuch api version 4
On Fri, Dec 11, 2015 at 03:22:51PM -0500, Don Zickus wrote:
This is a quick and dirty hack to utilize notmuch api version 4.
I think it's too early, notmuch 0.21 was released in 2015-10-29 (or
we need patch with #ifdefs to support the old API too).
IMHO we can wait a few month before we switch to API 4.
Karel
--
Karel Zak <kzak(a)redhat.com>
http://karelzak.blogspot.com
11:02 a.m.
New subject: [PATCH 2/2] mutt, notmuch: Update to use notmuch api version 4
On Tue, Dec 15, 2015 at 03:19:58PM +0100, Karel Zak wrote:
On Fri, Dec 11, 2015 at 03:22:51PM -0500, Don Zickus wrote:
> This is a quick and dirty hack to utilize notmuch api version 4.
I think it's too early, notmuch 0.21 was released in 2015-10-29 (or
we need patch with #ifdefs to support the old API too).
IMHO we can wait a few month before we switch to API 4.
No problem. I git clone'd notmuch to get a later copy to hack on (as this
doesn't work out of the box on RHEL-6 I believe). I just foolishly started
using API 4 without much thought. I figured I hacked something up, it might
be useful for someone, so I posted my dirty hack.
It makes no difference for me right now as I am compiling everything locally
to work around my awful RHEL-6 limitations. :-)
Thanks for accepting the other patch!
Cheers,
Don
8:06 a.m.
New subject: [PATCH 1/2] mutt, notmuch: Update 'unread' tag for all files
mapped to msg
On Fri, Dec 11, 2015 at 03:22:50PM -0500, Don Zickus wrote:
Both of these changes seem to work as expected now.
OK :-) Applied, thanks.
--
Karel Zak <kzak(a)redhat.com>
http://karelzak.blogspot.com
3055
days inactive
3059
days old
mutt-kz@lists.fedoraproject.org
4 comments
2 participants
participants (2)
-
Don Zickus -
Karel Zak