Thomas Roessler [Mon, 21 Dec 1998 11:27:23 +0000 (11:27 +0000)]
Mail folder synchronization-related fixes:
- Another situation under which mutt would lose track of the current
message when maildir folders changed has been removed.
- While synchronizing mbox-type folders, a write error when
constructing the draft version of the new mail folder would leave
behind a copy ofthe file containing the draft. This is most
annoying if the write error is due to lack of disk space.
Thomas Roessler [Mon, 21 Dec 1998 10:30:04 +0000 (10:30 +0000)]
Here is a small patch to simplify m4/Makefile.am.in. This came out
of a suggestion I made to Jim Meyering (maintainer of GNU fileutils,
where I snarfed the code from :). (From: Lars Hecking
<lhecking@nmrc.ucc.ie>)
Thomas Roessler [Thu, 17 Dec 1998 16:17:43 +0000 (16:17 +0000)]
I've been told that the FreeBSD shell has been fixed meanwhile.
Additionally, add the note from Vikas on patching Solaris to get
mutt work properly on that system.
Thomas Roessler [Fri, 11 Dec 1998 20:32:25 +0000 (20:32 +0000)]
The distdir target creates hard links for the source files. This
has the consequence that the us-dist hook will demolish files in the
main source tree. We have to un-link the various instances of the
source code.
Thomas Roessler [Tue, 8 Dec 1998 09:50:31 +0000 (09:50 +0000)]
More fixes to the maildir resynchronization code. This should now
also smoothly handle the case of messages _vanishing_ from maildir's
new/ subdirectory.
Thomas Roessler [Wed, 2 Dec 1998 10:05:54 +0000 (10:05 +0000)]
The attached patch fixes yet another redraw related bug. When I call
OP_SAVE from the Compose menu and use the file-browser and 'q' out
of the file-browser, the screen still shows the browser, but the
control is really in the Compose menu. A Ctrl-L verifies this.
Basically, a screen redraw is missing.
Thomas Roessler [Wed, 2 Dec 1998 10:04:59 +0000 (10:04 +0000)]
From the Compose menu, invoke attach-message. Tag a whole bunch of
messages. Say, for example, 100. When you exit the folder, the
Compose menu's status line has disappeared.
Thomas Roessler [Mon, 23 Nov 1998 10:13:31 +0000 (10:13 +0000)]
From Byrial:
I don't think that the match-all pattern modifier works as can be
expected from a logical point of view.
The attached patch will make it work as I expect it to work:
- It works on all patterns which explicit or implicit use address
lists: ~c, ~C, ~e, ~f, ~l, ~L, ~p, ~P, and ~t. (It didn't work
with ~l before).
- It works such that a modified pattern always match, unless there
is at least one address in at least one of the referenced lists
that it doesn't match.
Thomas Roessler [Wed, 18 Nov 1998 22:56:57 +0000 (22:56 +0000)]
This patch simplifies the generation of content type parameters
significantly by introducing a funciton mutt_set_parameter().
Additionally, we re-use code from parse.c for parsing user-input
content type headers on the compose screen.
Thomas Roessler [Tue, 17 Nov 1998 13:05:11 +0000 (13:05 +0000)]
Configuration fixes from Lars Hecking, with some more fixes and
modifications from myself. With some luck, you don't need a gettext
installation any more.
Thomas Roessler [Sun, 15 Nov 1998 09:34:15 +0000 (09:34 +0000)]
I can elaborate a little more on the last paragraph. It is deletion
of the top level part direct under a message/rfc822 part which
gives problems because the headers of the message/rfc822 part have
to be changed in this case. It is hard to tell exactly how many
bytes and lines will be removed and added in this case.
If we can refrain from deletion of such parts and thus only allow
deletion of parts direct under a multipart part, then it is
relative easy always to make correct Content-Length: and Lines:
headers in the copy. (The correctness of the Lines count is
dependent of the correctness of the initial value, though).