Thomas Adam [Sat, 8 Jul 2017 16:17:16 +0000 (17:17 +0100)]
compile-time output: use two lists
When asking NeoMutt for a list of compile-time features, the current
implementation assumes a mixture of options which can be enabled through
./configure, and some which default to on, without compile-time
exclusion.
It's not clear to anyone why this variable existed in the first place,
so it was decided that it should be removed.
Most of the time, AC_CHECK_LIB and AC_SEARCH_LIBS will already alter
LIBS to prepend to it the found library, so there is no real reason to
duplicate this mechanism.
Some manual assignments have been kept because either the
action-if-found was set to no-op, or because new found libraries were
appended to MUTTLIBS, so this keeps the diff minimal with the same
"behaviour".
Richard Russon [Fri, 14 Jul 2017 11:08:04 +0000 (12:08 +0100)]
merge: upstream fixes (mutt/default)
* fix signed/unsigned comparison in longest_common_prefix
* Fix crash when $postponed is on another server.
* bcache: cast to avoid implicit signed/unsigned comparison in bcache_path
* drop unused flags argument from imap_access
Kevin McCarthy [Wed, 12 Jul 2017 19:38:22 +0000 (12:38 -0700)]
Fix crash when $postponed is on another server.
imap_mxcmp() translates NULL to "INBOX". When $postponed points to a
URL with an empty or "INBOX" path, this will end up matching against a
NULL idata->mailbox in imap_status(). This resulted in a crash
because idata->ctx is also NULL.
Thanks to Olaf Hering for the detailed bug report and suggested fix.
Richard Russon [Thu, 6 Jul 2017 14:15:31 +0000 (15:15 +0100)]
merge: upstream fixes (mutt/default)
* Add more description for the %S and %Z $index_format characters.
* Add L10N comments to the GNUTLS certificate prompt.
* Add ~<() and ~>() immediate parent/children patterns. (closes #3144)
Stefan Bühler [Tue, 4 Jul 2017 09:57:24 +0000 (11:57 +0200)]
fix gpgme S/MIME non-detached signature handling
- gpgsm ends the session after the decrypt error, so restart gpgsm when
retrying with "maybe_signed".
- evolution sends encrypted + signed S/MIME mails with non-detached
signatures
Fabrice Bellet [Wed, 10 May 2017 16:05:17 +0000 (18:05 +0200)]
menu: fix the thread tree color
The background tree color is really meaningless as it mixes with the
background color of indexes subject in unpleasant ways. We prefer to
create a new color based on the tree foreground color only.
Richard Russon [Sun, 2 Jul 2017 20:44:55 +0000 (21:44 +0100)]
merge: build fixes for Solaris
* Handle native Solaris GSSAPI.
* Fix compilation error on Solaris.
* Include term.h for tigetstr except for slang build.
* Look for wide curses in libcurses itself.
* Remove unneeded forward declarations.
Thomas Klausner [Sat, 1 Jul 2017 09:39:14 +0000 (11:39 +0200)]
Fix compilation error on Solaris.
resize.c: In function 'mutt_resize_screen':
resize.c:38:18: error: storage size of 'w' isn't known
struct winsize w;
^
resize.c:47:19: error: 'TIOCGWINSZ' undeclared (first use in this function)
if (ioctl(fd, TIOCGWINSZ, &w) != -1)
^
If `mark_old` is false, unread mails will not be marked old after
they've been viewed. But they will be marked old if found in the `/cur`
subdirectory of the Maildir. This is inconsistent behaviour, since a
user with disabled `mark_old` will expect no old flags at all.
Pietro Cerutti [Fri, 9 Jun 2017 12:53:24 +0000 (12:53 +0000)]
More sophisticated check for BDB version + support for DB6 (non default)
This commit enhances configure.ac to actually parse (as in CPP) and
extract the values of DB_VERSION_MAJOR and DB_VERSION_MINOR from the
Berkeley DB include files. The previous approach to just use grep failed
on some system where the includes are just wrappers around the real bdb
headers.
Also, add support for the current db6 version (6.2), although db5 is
still the first one looked for.
Antonio Radici [Sun, 25 Jun 2017 22:47:11 +0000 (23:47 +0100)]
revert part of 40f81d1 Add debug_level/debug_file options
This patch moves the set_default/restore_default couple after the defaults have
been set by the init function; this was correctly placed in 1.7.2 and somehow it
got moved to the wrong place, which means that tmpdir (and many other variable)
are not correctly set with their new default value determined from environment
variables (or other logic), so when they get resetted with a 'reset all' call in
.muttrc, they will never have the correct default to fall back to and they will
use the wrong one in init.h
This patch fixes:
- http://bugs.debian.org/865822
- http://bugs.debian.org/865842
Richard Russon [Sat, 24 Jun 2017 16:44:15 +0000 (17:44 +0100)]
merge: upstream fixes (mutt/default)
* Change the compose menu fields to be dynamically padded.
* Rename 'sign as' to 'Sign as'; makes compose menu more consistent.
* Make GPGME key selection behavior the same as classic-PGP. (see #3950)
* Force hard redraw after $sendmail instead of calling mutt_endwin. (closes #3952) (see #3948)
* Enable all header fields in the compose menu to be translated.
* Auto-pad translation for the GPGME key selection "verify key" headers.
* Improve the L10N comment about Sign as.