]> granicus.if.org Git - mutt/log
mutt
12 years agoFix memory leak when multiple x-label or supersedes header fields exist in the header.
Michael Elkins [Thu, 3 Jan 2013 19:04:35 +0000 (19:04 +0000)]
Fix memory leak when multiple x-label or supersedes header fields exist in the header.

In order to keep the current behavior where the value of the last header field present is the one used, free any previous value.

12 years agoBacked out changeset bf84f24e512d
Michael Elkins [Thu, 3 Jan 2013 16:03:49 +0000 (08:03 -0800)]
Backed out changeset bf84f24e512d

12 years agodisplay +/-NCURSES_WIDECHAR when displaying ncurses version to show whether we are...
Michael Elkins [Wed, 2 Jan 2013 23:43:38 +0000 (23:43 +0000)]
display +/-NCURSES_WIDECHAR when displaying ncurses version to show whether we are linked against -lncurses or -lncursesw

12 years agoDrop unneeded variable HEADER_CACHE that causes clang warnings.
Matthias Andree [Tue, 1 Jan 2013 20:39:52 +0000 (21:39 +0100)]
Drop unneeded variable HEADER_CACHE that causes clang warnings.

The variable was unused and clang warned that it was not emitted to the .o file.

12 years agofix xml validation error in manual.xml in auto-generated content due to missing parag...
Michael Elkins [Mon, 31 Dec 2012 23:29:39 +0000 (23:29 +0000)]
fix xml validation error in manual.xml in auto-generated content due to missing paragraph mark in the $editor docstring

12 years agoFix accidental var redeclation with --enable-exact-address
guns [Mon, 31 Dec 2012 23:03:11 +0000 (15:03 -0800)]
Fix accidental var redeclation with --enable-exact-address

closes #3615

12 years agomerge
Michael Elkins [Mon, 31 Dec 2012 22:54:41 +0000 (14:54 -0800)]
merge

12 years agoreset to color normal before clearing to end of line when printing a message/error...
Michael Elkins [Mon, 31 Dec 2012 22:52:22 +0000 (14:52 -0800)]
reset to color normal before clearing to end of line when printing a message/error string

closes #3617

12 years agoupdate regexp to include const keyword. The auto generated documentation of function...
Michael Elkins [Mon, 31 Dec 2012 19:06:05 +0000 (19:06 +0000)]
update regexp to include const keyword.  The auto generated documentation of functions was failing due to a change in the source code format.

12 years agoEnsure that the in-reply-to header field was actually present before deciding it...
Antonio Radici [Mon, 31 Dec 2012 03:48:23 +0000 (19:48 -0800)]
Ensure that the in-reply-to header field was actually present before deciding it was removed.

closes #3378

12 years agoupdate ChangeLog
Michael Elkins [Mon, 31 Dec 2012 03:17:38 +0000 (19:17 -0800)]
update ChangeLog

12 years agoadd __FREE_CHECKED__
Michael Elkins [Mon, 31 Dec 2012 03:14:58 +0000 (19:14 -0800)]
add __FREE_CHECKED__

12 years agodocument the TLS changes since 1.5.21
Michael Elkins [Mon, 31 Dec 2012 03:12:12 +0000 (19:12 -0800)]
document the TLS changes since 1.5.21

12 years agomake the SETCOLOR() macro use bkgdset() on systems which have it, and attrset() when...
Michael Elkins [Mon, 31 Dec 2012 02:49:20 +0000 (18:49 -0800)]
make the SETCOLOR() macro use bkgdset() on systems which have it, and attrset() when we don't.

closes #3005

12 years agoremove m4_translit because it generates a warning with newer versions of autoconf...
Michael Elkins [Sun, 30 Dec 2012 17:26:30 +0000 (09:26 -0800)]
remove m4_translit because it generates a warning with newer versions of autoconf when used in AC_INIT.  Instead, use tr -d to output the version string without a trailing newline.

12 years agorestore ability for version.sh to cope with a repository directory with a space in it.
Michael Elkins [Fri, 28 Dec 2012 05:04:47 +0000 (21:04 -0800)]
restore ability for version.sh to cope with a repository directory with a space in it.

12 years agoUse the directory where the version.sh script is located as the
Aaron Schrab [Fri, 28 Dec 2012 04:33:25 +0000 (20:33 -0800)]
Use the directory where the version.sh script is located as the
directory where it is run.  Since this will be run in a separate shell
anyway, there's no need to worry about changing current directory.

This eliminates the need for the caller to specify where the source
directory is located, and doesn't require that the code take special
care to use the appropriate directory.  The fallback method of using
`cat VERSION` was broken in this respect, it would not work when run
from a different directory.
---
 configure.ac |    2 +-
 version.sh   |    7 ++++---
 2 files changed, 5 insertions(+), 4 deletions(-)

12 years agouse [] around body of AC_CHECK_HEADER to avoid problems with older versions of
Michael Elkins [Fri, 28 Dec 2012 01:37:42 +0000 (17:37 -0800)]
use [] around body of AC_CHECK_HEADER to avoid problems with older versions of
autoconf

12 years agoavoid using m4_esyscmd_s() since it is not available in older versions of autoconf...
Michael Elkins [Fri, 28 Dec 2012 02:56:02 +0000 (18:56 -0800)]
avoid using m4_esyscmd_s() since it is not available in older versions of autoconf.  use m4_esyscmd() and m4_translit() to whack the newline instead.

see #3613

12 years agoUses SSLv23_client_method() in mutt_ssl_starttls() because TLSv1_2_client_method...
Michael Elkins [Sun, 23 Dec 2012 02:00:18 +0000 (18:00 -0800)]
Uses SSLv23_client_method() in mutt_ssl_starttls() because TLSv1_2_client_method() does not honor the SSL_OP_NO_TLS* flags.

closes #3612

12 years agoIf neither --with-idn nor --without-idn are given, use -I/usr/include/idn when checki...
Michael Elkins [Sat, 22 Dec 2012 23:04:24 +0000 (15:04 -0800)]
If neither --with-idn nor --without-idn are given, use -I/usr/include/idn when checking for the libidn header files, if present (needed for Solaris 11).

Don't bother checking for the idna_* functions if we didn't find -lidn

12 years agoFix VPATH-based generation of non-hg-mq based patchlists.
Matthias Andree [Fri, 21 Dec 2012 10:17:41 +0000 (11:17 +0100)]
Fix VPATH-based generation of non-hg-mq based patchlists.

Note: Makefile.am runs patchlist.sh with $(srcdir)/PATCHES provided on stdin.

12 years agoFix VPATH build outside Mercurial repositories.
Matthias Andree [Fri, 21 Dec 2012 10:07:13 +0000 (11:07 +0100)]
Fix VPATH build outside Mercurial repositories.

12 years agoadd version.sh and check_sec.sh to EXTRA_DIST so they are including in the "make...
Michael Elkins [Fri, 21 Dec 2012 20:37:46 +0000 (12:37 -0800)]
add version.sh and check_sec.sh to EXTRA_DIST so they are including in the "make dist" output.

12 years agofix forgotten RETSIGTYPE change id [18a724e9eba0]
Michael Elkins [Fri, 21 Dec 2012 05:13:12 +0000 (21:13 -0800)]
fix forgotten RETSIGTYPE change id [18a724e9eba0]

12 years agoremove deprecated check for signal() return type
Michael Elkins [Fri, 21 Dec 2012 04:49:08 +0000 (20:49 -0800)]
remove deprecated check for signal() return type

12 years agoupdate configure.ac with autoupdate
Michael Elkins [Fri, 21 Dec 2012 04:45:05 +0000 (20:45 -0800)]
update configure.ac with autoupdate

12 years agochange AM_CONFIG_HEADER to AC_CONFIG_HEADERS and move it after the AM_INIT_AUTOMAKE...
Michael Elkins [Fri, 21 Dec 2012 04:34:21 +0000 (20:34 -0800)]
change AM_CONFIG_HEADER to AC_CONFIG_HEADERS and move it after the AM_INIT_AUTOMAKE.  config.h was not being generated from the dist package.

12 years agomodernize intialization of configure.ac
Michael Elkins [Fri, 21 Dec 2012 01:09:21 +0000 (17:09 -0800)]
modernize intialization of configure.ac

12 years agoFix 'make distcheck': missing txt2c* and po/mutt.pot cleanup
Matthias Andree [Thu, 20 Dec 2012 23:38:12 +0000 (00:38 +0100)]
Fix 'make distcheck': missing txt2c* and po/mutt.pot cleanup

12 years agoremove redundant #include statements
Michael Elkins [Thu, 20 Dec 2012 22:44:33 +0000 (14:44 -0800)]
remove redundant #include statements

12 years agoremove extern decl for Charset_is_utf8 since menu.c includes mbyte.h which has the...
Michael Elkins [Thu, 20 Dec 2012 22:40:41 +0000 (14:40 -0800)]
remove extern decl for Charset_is_utf8 since menu.c includes mbyte.h which has the same decl.

12 years agobackout out changeset 36b0cc717ecc
Michael Elkins [Thu, 20 Dec 2012 19:20:35 +0000 (11:20 -0800)]
backout out changeset 36b0cc717ecc

12 years agoallow for CRLF line endings when comparing for the OpenPGP block identifiers
Michael Elkins [Wed, 19 Dec 2012 23:26:21 +0000 (15:26 -0800)]
allow for CRLF line endings when comparing for the OpenPGP block identifiers

adapted from http://svnweb.freebsd.org/ports/head/mail/mutt/files/patch-pgp?revision=301308&view=markup

12 years agosupport nbsp in the pager.
Michael Elkins [Wed, 19 Dec 2012 22:52:59 +0000 (14:52 -0800)]
support nbsp in the pager.

imported from http://svnweb.freebsd.org/ports/head/mail/mutt/files/patch-nbsp?revision=301308&view=markup

12 years agodetect newer BDB versions up to 5.2
Michael Elkins [Wed, 19 Dec 2012 22:40:24 +0000 (14:40 -0800)]
detect newer BDB versions up to 5.2

imported from http://svnweb.freebsd.org/ports/head/mail/mutt/files/patch-bdb?revision=301308&view=markup

14 years agohttp://thread.gmane.org/gmane.mail.mutt.devel/15710
Fabian Groffen [Thu, 22 Jul 2010 18:06:33 +0000 (20:06 +0200)]
http://thread.gmane.org/gmane.mail.mutt.devel/15710

Interix doesn't have btowc, but it does have mbtowc.  Linux manpage also
discourages the use of btowc.

12 years agoupdated cs.po
Petr Pisar [Wed, 19 Dec 2012 21:29:29 +0000 (13:29 -0800)]
updated cs.po

12 years agouse WACS_* macros for printing wide versions of the line drawing characters when...
Michael Elkins [Wed, 19 Dec 2012 19:56:25 +0000 (11:56 -0800)]
use WACS_* macros for printing wide versions of the line drawing characters when linking against ncursesw.

This patch is based on http://prefix.gentooexperimental.org:8000/mutt-patches/file/fea851a2808f/solaris-ncurses-chars.patch

12 years agomake configure define _POSIX_C_SOURCE, _XOPEN_SOURCE and _XOPEN_SOURCE_EXTENDED to...
Michael Elkins [Wed, 19 Dec 2012 19:08:48 +0000 (11:08 -0800)]
make configure define _POSIX_C_SOURCE, _XOPEN_SOURCE and _XOPEN_SOURCE_EXTENDED to comply with POSIX.1-2008 and XSI conforming applications.

12 years agoupdate ru.po
Vsevolod Volkov [Wed, 19 Dec 2012 14:57:04 +0000 (06:57 -0800)]
update ru.po

12 years agoremove duplicate const keyword
Vincent Lefevre [Wed, 19 Dec 2012 14:55:25 +0000 (06:55 -0800)]
remove duplicate const keyword

12 years agodisable tls v1.1/1.2 in mutt_ssl_starttls() based on $ssl_use_tlsv1_*
Michael Elkins [Wed, 19 Dec 2012 04:46:33 +0000 (20:46 -0800)]
disable tls v1.1/1.2 in mutt_ssl_starttls() based on $ssl_use_tlsv1_*

see #3571

12 years agoadd support for TLSv1.1/1.2 when using GNUTLS.
Michael Elkins [Wed, 19 Dec 2012 04:38:46 +0000 (20:38 -0800)]
add support for TLSv1.1/1.2 when using GNUTLS.

see #3571

12 years agoadd variables for disabling TLSv1.1/1.2 when compiled against OpenSSL 1.0.1
hncaldwell [Wed, 19 Dec 2012 03:44:57 +0000 (19:44 -0800)]
add variables for disabling TLSv1.1/1.2 when compiled against OpenSSL 1.0.1

see #3571

12 years agorename iswsp() to is_email_wsp()
Michael Elkins [Tue, 18 Dec 2012 21:50:20 +0000 (13:50 -0800)]
rename iswsp() to is_email_wsp()

add new inline funtion skip_email_wsp() to be used in lieu of SKIPWS() when parsing ASCII protocols rather than user input.

change use of SKIPWS() to skip_email_wsp() in places where it is likely to be a problem.

12 years agofix problem where mutt_substrdup() was called with the start
Michael Elkins [Tue, 18 Dec 2012 10:21:57 +0000 (02:21 -0800)]
fix problem where mutt_substrdup() was called with the start
pointer > end pointer, as a result of using SKIPWS().  This
occurred because the header field had an empty body, so the
trailing newline was skipped over.  The fix is to just skip over
ascii space and htab.

closes #3609

12 years agoif the peer's x.509 cert is not valid, we need to allow the user to inspect it even...
hhorak [Tue, 18 Dec 2012 09:53:08 +0000 (01:53 -0800)]
if the peer's x.509 cert is not valid, we need to allow the user to inspect it even if the digital signatures are valid

closes #3506

12 years agoupdated fr.po
Vincent Lefevre [Tue, 18 Dec 2012 04:42:30 +0000 (20:42 -0800)]
updated fr.po

12 years agofix segfault in pop driver. when rereading the message header, the ENVELOPE* struct...
hhorak [Tue, 18 Dec 2012 01:09:18 +0000 (17:09 -0800)]
fix segfault in pop driver.  when rereading the message header, the ENVELOPE* struct is freed, but ctx->subj_hash retains a pointer to env->real_subj, so the hash table entry must be deleted prior to calling mutt_free_envelope()

http://pkgs.fedoraproject.org/cgit/mutt.git/tree/mutt-1.5.21-pophash.patch?h=f18

12 years agoFix NULL pointer dereference in notation name
Petr Pisar [Tue, 18 Dec 2012 00:47:37 +0000 (16:47 -0800)]
Fix NULL pointer dereference in notation name

closes #3574

12 years agouse TLSv1.2 if available in the system OpenSSL when using starttls
Michael Elkins [Mon, 17 Dec 2012 18:39:46 +0000 (10:39 -0800)]
use TLSv1.2 if available in the system OpenSSL when using starttls

12 years agouse SSL_get_version() to display the actual SSL protocol used rather than the fixed...
Michael Elkins [Mon, 17 Dec 2012 18:09:55 +0000 (10:09 -0800)]
use SSL_get_version() to display the actual SSL protocol used rather than the fixed string SSL.

12 years agoCorrect typos
Andrew Gaul [Mon, 17 Dec 2012 04:23:17 +0000 (20:23 -0800)]
Correct typos

12 years agoupdate documentation for $pgp_getkeys_command to note that %r is replaced with the...
Michael Elkins [Mon, 17 Dec 2012 03:33:29 +0000 (19:33 -0800)]
update documentation for $pgp_getkeys_command to note that %r is replaced with the email address, not the public key ID

closes #1208

12 years agouse new function iswsp() in place of isspace() when testing for whitespace.
Michael Elkins [Sun, 16 Dec 2012 21:28:39 +0000 (13:28 -0800)]
use new function iswsp() in place of isspace() when testing for whitespace.

closes #3608
closes #2956

12 years agomake imap_private.h include message.h for IMAP_HEADER_DATA definition
Michael Elkins [Sat, 15 Dec 2012 15:48:45 +0000 (07:48 -0800)]
make imap_private.h include message.h for IMAP_HEADER_DATA definition

cast &ctx->data to IMAP_HEADER_DATA** to fix compiler warning

imap_free_header_data should check for *data being NULL, rather than "data"

12 years agoMake imap_free_header_data type-safe
Andrew Gaul [Mon, 10 Sep 2012 02:54:46 +0000 (19:54 -0700)]
Make imap_free_header_data type-safe

Also push conditional into function for safety and consistency with
other free functions.

12 years agoFix VPATH build by running txt2c.sh from ${srcdir} rather than ".".
Matthias Andree [Sat, 8 Dec 2012 11:31:11 +0000 (12:31 +0100)]
Fix VPATH build by running txt2c.sh from ${srcdir} rather than ".".

12 years agoinvoke gpgme_check_version() as required to initialize the library.
Michael Elkins [Fri, 7 Dec 2012 17:51:46 +0000 (09:51 -0800)]
invoke gpgme_check_version() as required to initialize the library.

move locale code to initialization function.

closes #3300

12 years agoadditional documentation for $editor. closes #3602
Michael Elkins [Thu, 6 Dec 2012 15:56:54 +0000 (07:56 -0800)]
additional documentation for $editor.  closes #3602

12 years agolist-reply documentation should mention its behavior wrt the List-Post header field.
Michael Elkins [Sat, 1 Dec 2012 22:48:27 +0000 (14:48 -0800)]
list-reply documentation should mention its behavior wrt the List-Post header field.

closes #3599

12 years agodon't perform rfc3676 space stuffing when recalling a postponed message where it...
Michael Elkins [Sat, 1 Dec 2012 22:32:48 +0000 (14:32 -0800)]
don't perform rfc3676 space stuffing when recalling a postponed message where it has already been done.

closes #3600

12 years agomake ascii_strcasecmp properly handle unequal length strings with the same prefix
Michael Elkins [Sat, 1 Dec 2012 22:31:42 +0000 (14:31 -0800)]
make ascii_strcasecmp properly handle unequal length strings with the same prefix

closes #3601

12 years agoAdd compiler and configure info to mutt -v output (closes #3537)
David Champion [Tue, 16 Oct 2012 23:13:13 +0000 (18:13 -0500)]
Add compiler and configure info to mutt -v output (closes #3537)

Makefile(.am) updated to produce conststrings.c, which contains
C strings representing:

* the compiler's own version information;
* the CFLAGS value from the Make environment
* the ./configure options

main.c is updated to print them when running 'mutt -v'.

txt2c.sh is added to produce conststrings.c. txt2c.sh uses a compiled
txt2c binary if possible, for complete fidelity to the source strings
in whatever encoding they may use.  If txt2c is not available (could
not be compiled, or was not compiled natively) it falls back on a shell
function to approximate the output using sed and tr.

12 years agomutt_buffer_init: just memset the buffer (closes #3588)
Brendan Cully [Sun, 22 Jul 2012 18:15:30 +0000 (11:15 -0700)]
mutt_buffer_init: just memset the buffer (closes #3588)

This was meant to be part of the previous commit.

12 years agoSplit mutt_buffer_new out of mutt_buffer_init.
Brendan Cully [Sun, 22 Jul 2012 02:51:31 +0000 (19:51 -0700)]
Split mutt_buffer_new out of mutt_buffer_init.

Currently, no callers were providing a non-NULL buffer to
mutt_buffer_init, and splitting it will allow more sane semantics for
buffer allocation, initialization, and destruction in a later patch.

12 years agoDo not use stack-allocated space for BUFFERs
Brendan Cully [Fri, 20 Jul 2012 19:32:44 +0000 (12:32 -0700)]
Do not use stack-allocated space for BUFFERs

It will cause mutt_buffer_* to segfault if it needs to grow the space.
Thanks to Steve Losh for discovering the issue. I have done some simple
grepping to find other cases, but some may remain.

12 years agoSet text_flowed after send-hook has been applied (closes #3550)
Dan Fandrich [Mon, 9 Jul 2012 05:40:06 +0000 (22:40 -0700)]
Set text_flowed after send-hook has been applied (closes #3550)

12 years agosilence some style warnings
Brendan Cully [Mon, 9 Jul 2012 05:27:10 +0000 (22:27 -0700)]
silence some style warnings

12 years agoFix some spelling errors (closes #3493)
Antonio Radici [Mon, 9 Jul 2012 00:34:56 +0000 (17:34 -0700)]
Fix some spelling errors (closes #3493)

12 years agognutls:tls_compare_certificates: check strstr for failure (closes #3547)
Brendan Cully [Mon, 9 Jul 2012 00:25:12 +0000 (17:25 -0700)]
gnutls:tls_compare_certificates: check strstr for failure (closes #3547)

A malformed certificate file could cause strstr to return an unhandled NULL.
Thanks to hhorak for the proposed patch. This one is similar but avoids using
memmem for the first time (I am not sure about its portability).

12 years agocheck_charset: check for NULL before calling strtok_r (closes #3326)
Antonio Radici [Sun, 8 Jul 2012 04:57:01 +0000 (21:57 -0700)]
check_charset: check for NULL before calling strtok_r (closes #3326)

Otherwise, if strtok_r cannot handle NULL an empty *charset setting will segfault

12 years agoDo not use SASL for login unless AUTH=LOGIN is advertised (closes #3556)
Brendan Cully [Sat, 7 Jul 2012 22:50:08 +0000 (15:50 -0700)]
Do not use SASL for login unless AUTH=LOGIN is advertised (closes #3556)

12 years agoUpdate Italian translation
Marco Paolone [Fri, 25 May 2012 20:30:47 +0000 (22:30 +0200)]
Update Italian translation

12 years agoRemove AM_C_PROTOTYPES from configure.ac (closes #3579)
Brendan Cully [Fri, 25 May 2012 05:02:57 +0000 (22:02 -0700)]
Remove AM_C_PROTOTYPES from configure.ac (closes #3579)

I doubt we have any users of that macro nowadays. It will be interesting to
hear about it if we do.

12 years agoSupport passwords of up to 127 characters.
Brendan Cully [Thu, 24 May 2012 03:29:37 +0000 (23:29 -0400)]
Support passwords of up to 127 characters.

I received a report on IRC of a failure due to a 64-byte password.

12 years agognutls: catch gnutls_init failures
Brendan Cully [Mon, 30 Apr 2012 05:15:19 +0000 (22:15 -0700)]
gnutls: catch gnutls_init failures

I haven't actually experienced these, but it seems like a good idea.

12 years agogive user time to read SASL init error message
Brendan Cully [Mon, 30 Apr 2012 05:14:22 +0000 (22:14 -0700)]
give user time to read SASL init error message

13 years agoUpdated Japanese translation
TAKAHASHI Tamotsu [Wed, 21 Dec 2011 06:24:35 +0000 (22:24 -0800)]
Updated Japanese translation

13 years agoDeclare many structures const (closes #3552)
Dan Fandrich [Sat, 3 Dec 2011 19:13:10 +0000 (11:13 -0800)]
Declare many structures const (closes #3552)

Many structs used in mutt are actually constant but are defined
without the 'const' keyword. This can slow initialization (slightly)
in some environments due to extra copying and increases the amount of
writable RAM required at run-time, which can be significant on non-MMU
systems. Using const can also increase the opportunities for compiler
optimization.

The attached patch marks many such structures as const. On my test x86
build, this reduces the size of .data by over 50%.

13 years agoInline some small, often-used functions (closes #3551)
Dan Fandrich [Sat, 3 Dec 2011 19:07:41 +0000 (11:07 -0800)]
Inline some small, often-used functions (closes #3551)

When I first sent this patch to the mailing list in 2008, my benchmarking found a
reduction in startup time of 15%.

13 years agobranch merge
Michael Elkins [Mon, 28 Nov 2011 17:47:44 +0000 (09:47 -0800)]
branch merge

13 years agomake kill-eow and kill-word consistent in considering words to consist only of alphan...
Michael Elkins [Mon, 28 Nov 2011 17:46:46 +0000 (09:46 -0800)]
make kill-eow and kill-word consistent in considering words to consist only of alphanumeric characters.  closes #3549.

13 years agoUpdated French translation
Vincent Lefevre [Sun, 27 Nov 2011 04:41:05 +0000 (20:41 -0800)]
Updated French translation

13 years agoCheck that $sendmail is set to avoid potential segfault. Closes #3548.
Michael Elkins [Fri, 25 Nov 2011 20:52:55 +0000 (12:52 -0800)]
Check that $sendmail is set to avoid potential segfault.  Closes #3548.

13 years agoInvalidate header caches when spam rules change. closes #3473
David Champion [Wed, 13 Jul 2011 16:05:37 +0000 (11:05 -0500)]
Invalidate header caches when spam rules change. closes #3473

User 'exg' on #mutt reported a problem when using header cache with IMAP
and spam rules.  Spam tags are retrieved from cache and not updated when
spam rules change.

1. Enable header caching in muttrc
2. Set spam rules in muttrc
3. Use IMAP (or POP or Maildir).  Your spam rules store spam tags into
   your message headers, and these get cached.
4. Quit mutt
5. Change your spam rules in muttrc
6. Relaunch mutt
7. Because of header caching, the rfc822 headers are not reparsed, and
   spam rules are not reapplied.  User has outdated spam tags.

Spam rules are applied during rfc822 header parsing because it's the
only time we get a full and unadulterated view of all headers.  Once
header parse is done we have only processed headers and 'user headers'
(X-headers).  We might do spam rule matching against the struct HEADER
copy of headers, except that:

a. the header might not be in HEADER if it's non-standard and not an
   X-header;
b. spam rules are regular expressions matching any arbitrary header
   line, and we would need to map these regexps to the header used to
   create the HEADER fields.  This would change the syntax and behavior
   of spam rules dramatically.

One solution, perhaps stopgap, is to invalidate the header cache when
spam rules change.  That's what this patch does.  It alters the hcache
versioning algorithm so that in addition to using the compiled-in
checksum of mutt's structures definitions, we also mix in the user's
runtime spam rules.  Any time spam rules change the hcachever hash value
changes, causing the hcache to be rewritten.

13 years agoUpdate ChangeLog
Brendan Cully [Sat, 2 Jul 2011 01:21:13 +0000 (18:21 -0700)]
Update ChangeLog

13 years agoversion.sh: fix(?) failure due to bizarre sh quoting rules
Brendan Cully [Sat, 2 Jul 2011 00:23:37 +0000 (17:23 -0700)]
version.sh: fix(?) failure due to bizarre sh quoting rules

I have not checked POSIX, and sh quoting always makes my head hurt, but this
seems to suffice for my sh (actually bash in sh mode) when HGROOT contains a space.

13 years agofolder_file: change type of new from short to int
Olaf Hering [Mon, 23 May 2011 11:22:33 +0000 (13:22 +0200)]
folder_file: change type of new from short to int

As requested by Rocco Rutte in bug #2421 increase the type to int.

Signed-off-by: Olaf Hering <olaf@aepfle.de>
13 years agofix gcc 4.6 warning -Wunused-but-set-variable in pgppubring.c
Olaf Hering [Mon, 23 May 2011 10:59:14 +0000 (12:59 +0200)]
fix gcc 4.6 warning -Wunused-but-set-variable in pgppubring.c

pgppubring.c: In function 'pgp_parse_pgp3_sig':
pgppubring.c:424:17: warning: variable 'hashalg' set but not used [-Wunused-but-set-variable]
pgppubring.c:423:17: warning: variable 'pkalg' set but not used [-Wunused-but-set-variable]

Signed-off-by: Olaf Hering <olaf@aepfle.de>
13 years agoversion.sh: 1003.1 "-k" instead of obsolete "+" syntax (closes: #3528)
David Champion [Wed, 29 Jun 2011 17:01:14 +0000 (12:01 -0500)]
version.sh: 1003.1 "-k" instead of obsolete "+" syntax (closes: #3528)

13 years agoversion.sh: support building outside of source directory
David Champion [Wed, 29 Jun 2011 16:57:12 +0000 (11:57 -0500)]
version.sh: support building outside of source directory

Reported by Emanuele Giaquinta <e.giaquinta@glauco.it>. Thanks also
to Michael Elkins <me@sigpipe.org>.

13 years agoParentheses suggested
Brendan Cully [Sun, 26 Jun 2011 03:07:57 +0000 (20:07 -0700)]
Parentheses suggested

13 years agoDo not subscribe to \noselect mailboxes. Closes #3445
Brendan Cully [Sun, 26 Jun 2011 03:05:37 +0000 (20:05 -0700)]
Do not subscribe to \noselect mailboxes. Closes #3445

13 years agoAvoid creating sparse ctx array when fetch responses are ignored. Closes #3288.
hhorak [Sat, 25 Jun 2011 22:42:52 +0000 (15:42 -0700)]
Avoid creating sparse ctx array when fetch responses are ignored. Closes #3288.

13 years agoSilence some eclipse warnings
Brendan Cully [Sat, 25 Jun 2011 22:23:05 +0000 (15:23 -0700)]
Silence some eclipse warnings

13 years agorfc2047 encode recipient list in resent-to header field. closes #3512
Michael Elkins [Mon, 20 Jun 2011 23:39:27 +0000 (16:39 -0700)]
rfc2047 encode recipient list in resent-to header field.  closes #3512

13 years agocopy idn status when duplicating address object
Michael Elkins [Mon, 20 Jun 2011 23:33:07 +0000 (16:33 -0700)]
copy idn status when duplicating address object

13 years agofix regexp example in documentation
Antonio Radici [Mon, 20 Jun 2011 23:07:27 +0000 (16:07 -0700)]
fix regexp example in documentation