Art Cancro [Fri, 1 Sep 2006 19:18:22 +0000 (19:18 +0000)]
Fixed bug #173. Multi-select and multi-deselect no longer have
a problem where the internally selected list contains messages
that are not highlighted.
Art Cancro [Fri, 1 Sep 2006 18:53:57 +0000 (18:53 +0000)]
Found the cause of Bug #173 and added a comment to the
code showing where to fix it. (I don't have the time right
now to fix it but will do this later.)
Art Cancro [Fri, 1 Sep 2006 17:13:30 +0000 (17:13 +0000)]
summary view: deleting messages now removes the table rows from
the DOM tree entirely, instead of merely hiding them. This is important
because the next thing we're going to do is traverse the DOM tree to determine
what rows to use for group select.
Art Cancro [Fri, 28 Jul 2006 04:14:15 +0000 (04:14 +0000)]
Bugfix for previous commit: IMAP and POP3 expunge operations
were calling CtdlDeleteMessages() with num_msgs=0 if there was nothing to delete; this
was bad because it causes all msgs in the room to be deleted.
Art Cancro [Fri, 28 Jul 2006 03:30:57 +0000 (03:30 +0000)]
CtdlDeleteMessages() now has a bulk API. Updated all of the
callers to use it. Also updated the IMAP and POP3 expunge functions to actually
make use of the bulk delete. This should significantly improve performance
when used with large folders.
Art Cancro [Thu, 27 Jul 2006 04:21:07 +0000 (04:21 +0000)]
CtdlDeleteMessages() now uses the bulk API CtdlCopyMsgsToRoom()
to move deleted messages to the __CtdlDeletedMessages__ room. The next
step will be to make CtdlDeleteMessages() itself a bulk-move API.
Art Cancro [Mon, 17 Jul 2006 03:44:38 +0000 (03:44 +0000)]
Brought over the new ical_dezonify that handles unspecified
timezones by converting from a default timezone to UTC. This
also has the temporary hack that always uses America/New_York as the
default time zone. Needs to be fixed before we publish.
Art Cancro [Thu, 13 Jul 2006 03:22:51 +0000 (03:22 +0000)]
When finding a non-UTC timestamp with no time
zone specified, convert from the local timezone to UTC. Currently
it is hardcoded for the America/New_York timezone, but we will fix
that.
Art Cancro [Wed, 12 Jul 2006 03:47:47 +0000 (03:47 +0000)]
ical_dezonify: rearranged code and added log output
to determine whether a timestamp is either (1) already UTC, (2) not UTC
but included a time zone to convert from, or (3) not UTC but no timezone
specified. Still need to handle (3) better, but we need to find a way
to determine the current timezone.
Art Cancro [Tue, 11 Jul 2006 17:16:14 +0000 (17:16 +0000)]
When generating Message ID's, pad with leading zeroes
to ten places in order to thwart a b0rken SpamAssassin rule which
assumes that less than five characters to the left of the @ means
a message is spam.
Art Cancro [Thu, 6 Jul 2006 02:44:20 +0000 (02:44 +0000)]
* New HALT command shuts down server but suspends instead of exiting.
* setup.c: fixed a citadel.config path problem that prevented alternative dir installations
from working properly
Art Cancro [Wed, 21 Jun 2006 03:26:57 +0000 (03:26 +0000)]
citadel_ipc.c: comment out block of code that
handles certain OpenSSL errors incorrectly. It was recursing into itself
infinitely until a stack overflow. The condition under which this occurs
has not been located or fixed, but at least now it will exit with an
OpenSSL error message.
Art Cancro [Tue, 23 May 2006 02:01:36 +0000 (02:01 +0000)]
* Fixed some compiler warnings in the new unsignedchar-safe version
of text_to_server_qp()
* Added in an ifdef block around the call to endtls() to make it
safe to compile on non-OpenSSL hosts
Art Cancro [Mon, 22 May 2006 03:06:00 +0000 (03:06 +0000)]
When THE DREADED AUTO-PURGER expires user accounts,
we were previously expiring accounts for which an associated Unix account no longer exists
if running in autologin mode. This is now the ONLY condition under which an account will
be expired in this mode.
Art Cancro [Wed, 17 May 2006 03:49:57 +0000 (03:49 +0000)]
CtdlSaveMsgPointerInRoom() is now a simple wrapper around
the new CtdlSaveMsgPointersInRoom() which can accept a list of
message pointers instead of a single one. This is being written
with IMAP optimization in mind. It has not been extensively tested.
Do not run production on this code yetsvn commit -m "CtdlSaveMsgPointerInRoom() is now a simple wrapper around
the new CtdlSaveMsgPointersInRoom() which can accept a list of
message pointers instead of a single one. This is being written
with IMAP optimization in mind. It has not been extensively tested.
Art Cancro [Tue, 9 May 2006 03:33:21 +0000 (03:33 +0000)]
Fixed the PLAIN auth method in SMTP so that it accepts
the username/password either inline (like Thunderbird does it) or multistate (like Pine
does it).
Art Cancro [Mon, 8 May 2006 03:30:49 +0000 (03:30 +0000)]
Call endtls() when an HTTP session ends normally -- not
just when it sends abnormally. This fixes that huge memory
leak I've been trying to find all this time.
Art Cancro [Wed, 3 May 2006 03:29:12 +0000 (03:29 +0000)]
mime_parser.c: pad the encoded-to-unencoded buffer
a bit more, because certain MIME parts (such as tnef) were blowing the target
buffer. This fixes a problem whose symptom is a server crash during fulltext
indexing.
Art Cancro [Wed, 3 May 2006 03:27:19 +0000 (03:27 +0000)]
mime_parser.c: pad the encoded-to-unencoded buffer
a bit more, because certain MIME parts (such as tnef) were blowing the target
buffer. This fixes a problem whose symptom is a server crash during fulltext
indexing.
Art Cancro [Wed, 26 Apr 2006 02:58:14 +0000 (02:58 +0000)]
Added /static/webcit-tinymce.css which is applied when
TinyMCE is called from WebCit. Configured it to use a 10pt font in
the editor window instead of the tiny font.
Art Cancro [Wed, 26 Apr 2006 02:58:00 +0000 (02:58 +0000)]
Added /static/webcit-tinymce.css which is applied when
TinyMCE is called from WebCit. Configured it to use a 10pt font in
the editor window instead of the tiny font.
Art Cancro [Tue, 25 Apr 2006 15:23:52 +0000 (15:23 +0000)]
IMAP FETCH operations now check before every iteration
of the per-message loop, whether CC->kill_me is set. The loop is aborted
if the session has been marked for cancellation, which usually happens
because the user cancelled out while the FETCH was still in progress.