Brendan Cully [Tue, 9 Aug 2005 20:13:03 +0000 (20:13 +0000)]
Replace raw latin1 characters with entities, and enable latin output
for manual.txt. I'm not sure this is ideal for non-latin (eg utf8)
environments, but hopefully docbook will resolve this anyway. Closes: #2029.
Remove an unneeded cast, which introduces type-punning. Modified to use FREE
instead of safe_free, since that seems to be the convention. I personally
don't see the point.
Brendan Cully [Mon, 8 Aug 2005 18:59:43 +0000 (18:59 +0000)]
Pause for a moment when a hostname fails to resolve. Closes #1942.
There are a number of cases where mutt cascades error messages. It might
be nice if there were an error stack so that low level functions could
push_error instead of displaying it right away, and the higher level ones
could eg mutt_message(_("Failed to write FCC: %s"), pop_error ())...
Brendan Cully [Mon, 8 Aug 2005 03:30:59 +0000 (03:30 +0000)]
Fix bug just introduced in lsub parsing that could cause mutt to loop indefinitely.
Don't recheck subscriptions at every mailbox change, just the initial connection.
Brendan Cully [Sun, 7 Aug 2005 21:29:23 +0000 (21:29 +0000)]
Allow IMAP tab-completion to complete over defined mailboxes, saving a
bunch of typing for the hostname part of the URL. Should probably be
extended to complete over open connections and possibly also folder/account
hooks.
Certain versions of libc may segfault during regex processing if given
incomplete multibyte characters. Work around this by manually trimming
the display buffer. I modified it somewhat to only check when the
situation is most likely to occur: fgets has read all the way to the
capacity of the buffer. Thanks also to Tamo for his comments.
Brendan Cully [Sat, 6 Aug 2005 21:19:39 +0000 (21:19 +0000)]
Replace pointless imap_force_ssl option with ssl_force_tls, which requires
any connection (POP or IMAP) to a remote server to be encrypted, and attempts
to negotiate encryption even if the capability isn't advertised (this last
bit is of dubious utility, but essentially harmless since the connection
would otherwise fail anyway). Closes: #1284.
Brendan Cully [Fri, 5 Aug 2005 02:16:46 +0000 (02:16 +0000)]
Move some attachment-related prototypes and structures into attach.h
from the global headers. A small step towards sane dependency tracking
and code cleanliness.
Brendan Cully [Tue, 2 Aug 2005 18:15:33 +0000 (18:15 +0000)]
Remove mutt_ssl_nss.c from POTFILES.in. Also hack the makefiles to
build keymap_alldefs.h in $(top_srcdir), since that's where xgettext
looks. Not super clean I guess - the other alternative is to sed
POTFILES.
Brendan Cully [Tue, 2 Aug 2005 05:58:18 +0000 (05:58 +0000)]
More fiddling with PGP results. When rerunning PGP on a message, clear
old security result bits to avoid confusing messages like 'decrypt failed;
PGP signature successfully verified'.
Brendan Cully [Mon, 1 Aug 2005 07:35:04 +0000 (07:35 +0000)]
Cache SSL certificates that have been accepted but not saved until mutt
exits. (closes #643).
Create mutt_add_list_n for adding non-character data to lists, have
mutt_add_list call it.
The builtin pager tries to wrap a line for ever
if:
0) the line contains a space or a tab character
followed by a very long word, and
1) M_PAGER_NSKIP flag is passed to it, and
2) $smart_wrap is set. (default)
Vincent Lefevre [Sat, 30 Jul 2005 04:13:07 +0000 (04:13 +0000)]
When using colors, the status line of the index is displayed in a
>bright color, even when the prefix "bright" is *not* given in the
foreground color. The problem occurs when I have the following additional
line:
color normal brightwhite black
The problem is that BKGDSET's are missing.