Dan Winship [Mon, 14 Jan 2013 22:36:56 +0000 (17:36 -0500)]
2.35.4
Alexander Shopov [Tue, 8 Jan 2013 19:50:14 +0000 (21:50 +0200)]
Updated Bulgarian translation
Balázs Úr [Mon, 31 Dec 2012 18:04:31 +0000 (19:04 +0100)]
Updated Hungarian translation
Antonio Fernandes C. Neto [Tue, 25 Dec 2012 22:45:07 +0000 (20:45 -0200)]
Updated Brazilian Portuguese Translation
Fran Diéguez [Mon, 24 Dec 2012 21:52:12 +0000 (22:52 +0100)]
Updated Galician translations
Dimitris Spingos [Mon, 24 Dec 2012 15:26:37 +0000 (17:26 +0200)]
Updated Greek translation
Robert Ancell [Fri, 21 Dec 2012 02:18:28 +0000 (15:18 +1300)]
Fake being in GNOME to run GNOME proxy settings test
Mattias Põldaru [Wed, 19 Dec 2012 14:36:46 +0000 (16:36 +0200)]
[l10n] Updated Estonian translation
Matej Urbančič [Tue, 18 Dec 2012 07:40:20 +0000 (08:40 +0100)]
Updated Slovenian translation
Dan Winship [Tue, 18 Dec 2012 01:22:47 +0000 (20:22 -0500)]
2.35.3
Dan Winship [Sun, 9 Dec 2012 17:55:33 +0000 (18:55 +0100)]
proxy/gnome: only use this when running under GNOME
The original idea had been that people could just not build/install
this module if they didn't want the GNOME proxy settings, but that
hasn't really worked. So change it to only use this module when in a
GNOME session (falling back to the libproxy module otherwise).
Dan Winship [Fri, 7 Dec 2012 14:37:22 +0000 (15:37 +0100)]
gnutls: fix a hang when closing during handshake
If the app did a synchronous g_io_stream_close() on a connection that
had completed an async handshake but not yet run finish_handshake(),
it would hang forever (because the close would be blocking the main
loop that finish_handshake() was going to run in, and it couldn't
continue until finish_handshake() ran).
Fix this by yielding the handshake op from the end of the handshake
thread, and then in async_handshake_completed(), deal with the
possibility that someone else might have already run
finish_handshake() for us.
Add a test for this to tls/tests/connection. This patch also disables
the refcount checking on test->client_connection for now, because the
new code makes it racy; it's possible we'll run teardown_connection()
before the handshake GTask has dropped its ref on the connection.
https://bugzilla.gnome.org/show_bug.cgi?id=688751
Dan Winship [Wed, 5 Dec 2012 20:05:21 +0000 (15:05 -0500)]
gnutls: move some code from finish_handshake to handshake
verify_peer_certificate() doesn't emit signals, and might block, so it
should happen in the handshake thread. And then with that moved there
there's no reason to not do the other (non-interactive)
certificate-related checks there as well.
Dan Winship [Wed, 5 Dec 2012 20:02:47 +0000 (15:02 -0500)]
tls/tests: fix a race condition
Dan Winship [Wed, 5 Dec 2012 20:02:07 +0000 (15:02 -0500)]
gnutls: deal with G_IO_ERROR_BROKEN_PIPE
glib 2.35 breaks G_IO_ERROR_BROKEN_PIPE out of G_IO_ERROR_FAILED, so
we need to recognize that now as well.
Dan Winship [Tue, 4 Dec 2012 16:37:31 +0000 (11:37 -0500)]
tests: remove an always-false check
Marek Černocký [Sun, 2 Dec 2012 20:49:53 +0000 (21:49 +0100)]
Updated Czech translation
Daniel Mustieles [Sun, 2 Dec 2012 18:32:12 +0000 (19:32 +0100)]
Updated Spanish translation
Piotr Drąg [Sun, 2 Dec 2012 04:58:46 +0000 (05:58 +0100)]
Updated Polish translation
Aurimas Černius [Fri, 30 Nov 2012 19:55:34 +0000 (21:55 +0200)]
Updated Lithuanian translation
Yaron Shahrabani [Fri, 30 Nov 2012 08:54:01 +0000 (10:54 +0200)]
Updated Hebrew translation.
Stef Walter [Thu, 29 Nov 2012 12:51:46 +0000 (13:51 +0100)]
gnutls: Don't hang trying to handshake, if early close
When closing a connection before the handshake, the GTlsConnectionGnutls
code would try to do a handshake. Fix this, and test for it.
https://bugzilla.gnome.org/show_bug.cgi?id=689271
Stef Walter [Thu, 29 Nov 2012 10:05:06 +0000 (11:05 +0100)]
tls: Add a test for client certificate auth failure
Add a test where the server requires a certificate from the client
but the server doesn't provide any.
https://bugzilla.gnome.org/show_bug.cgi?id=689260
Stef Walter [Thu, 29 Nov 2012 10:01:45 +0000 (11:01 +0100)]
gnutls: Allow closing a connection cleanly if handshake fails
Only set the internal ever_handshaked flag if the handshake was
successful. Otherwise using g_io_stream_close() after a handshake
failure results in an internal gnutls error.
https://bugzilla.gnome.org/show_bug.cgi?id=689260
Stef Walter [Thu, 29 Nov 2012 10:00:00 +0000 (11:00 +0100)]
gnutls: Appropriate error when server requires cert but none
When the server requires a certificate for auth, but none is
provider, return the appropriate G_TLS_ERROR_CERTIFICATE_REQUIRED
error.
https://bugzilla.gnome.org/show_bug.cgi?id=689260
Stef Walter [Thu, 29 Nov 2012 09:55:23 +0000 (10:55 +0100)]
gnutls: Fix regression where begin_handshake() is not called
The begin_handshake() virtual method is no longer called. This results
in stuff like the GTlsServerConnection authentication-mode not being
passed to gnutls.
Fix issue, and add tests which a future regression in this area
should fall over.
https://bugzilla.gnome.org/show_bug.cgi?id=689259
Stef Walter [Thu, 29 Nov 2012 13:15:09 +0000 (14:15 +0100)]
gnutls: Avoid assertion in finish_handshake()
https://bugzilla.gnome.org/show_bug.cgi?id=689274
Richard Stanislavský [Fri, 2 Nov 2012 19:50:28 +0000 (19:50 +0000)]
Updated Slovak translation
Richard Stanislavský [Fri, 2 Nov 2012 19:49:11 +0000 (19:49 +0000)]
Added Slovak translation
Saleem Abdulrasool [Wed, 31 Oct 2012 17:43:04 +0000 (10:43 -0700)]
build: disable TLS tests unless GNUTLS is enabled
All of the TLS tests hard-depend on GNUTLS. If glib-networking is built without
GNUTLS, then simply disable the TLS tests.
Dan Winship [Mon, 22 Oct 2012 21:38:49 +0000 (23:38 +0200)]
2.35.1
Dan Winship [Mon, 22 Oct 2012 16:57:45 +0000 (18:57 +0200)]
Remove now-unnecessary g_type_init() calls
Dan Winship [Wed, 20 Jun 2012 20:17:57 +0000 (16:17 -0400)]
Port to GTask
Mattias Põldaru [Mon, 22 Oct 2012 19:02:32 +0000 (22:02 +0300)]
[l10n] Updated Estonian translation
Dan Winship [Wed, 10 Oct 2012 21:08:49 +0000 (17:08 -0400)]
post-branch version bump
Dan Winship [Mon, 24 Sep 2012 17:13:50 +0000 (13:13 -0400)]
2.34
YunQiang Su [Sat, 22 Sep 2012 20:06:39 +0000 (04:06 +0800)]
update Simplified Chinese (zh_CN) translation
Unticha Pramgoed [Sat, 22 Sep 2012 15:25:43 +0000 (22:25 +0700)]
Updated Thai translation
Khaled Hosny [Thu, 20 Sep 2012 05:26:44 +0000 (07:26 +0200)]
Updated Arabic translation
Mohammad Alhargan [Thu, 20 Sep 2012 05:23:51 +0000 (07:23 +0200)]
Arabic translation
Alexander Shopov [Thu, 20 Sep 2012 03:11:39 +0000 (06:11 +0300)]
Updated Bulgarian translation
Yoji TOYODA [Wed, 19 Sep 2012 15:22:00 +0000 (00:22 +0900)]
[l10n] Update Japanese translation
Rajesh Ranjan [Wed, 19 Sep 2012 12:09:41 +0000 (17:39 +0530)]
hindi update
Rajesh Ranjan [Wed, 19 Sep 2012 12:05:30 +0000 (17:35 +0530)]
hindi update
Carles Ferrando [Mon, 17 Sep 2012 23:40:26 +0000 (01:40 +0200)]
[l10n]Updated Catalan (Valencian) translation
Gil Forcada [Mon, 17 Sep 2012 23:40:24 +0000 (01:40 +0200)]
[l10n] Updated Catalan translation
Dan Winship [Mon, 17 Sep 2012 20:06:23 +0000 (16:06 -0400)]
2.33.14
Djavan Fagundes [Mon, 17 Sep 2012 14:55:09 +0000 (11:55 -0300)]
Updated Brazilian Portuguese translation by Rafael Ferreira <rafael.f.f1@gmail.com> and myself
Ask H. Larsen [Sun, 16 Sep 2012 11:19:52 +0000 (13:19 +0200)]
Updated Danish translation
A S Alam [Fri, 14 Sep 2012 02:02:26 +0000 (07:32 +0530)]
update Punjabi Translation
Ville-Pekka Vainio [Wed, 12 Sep 2012 19:29:18 +0000 (22:29 +0300)]
Finnish translation update by Jiri Grönroos
Changwoo Ryu [Sat, 8 Sep 2012 20:25:36 +0000 (05:25 +0900)]
Updated Korean translation
Marek Černocký [Sat, 8 Sep 2012 09:20:58 +0000 (11:20 +0200)]
Updated Czech translation
Alexandre Franke [Thu, 6 Sep 2012 15:35:10 +0000 (17:35 +0200)]
Update French translation
Wolfgang Stoeggl [Wed, 5 Sep 2012 22:10:57 +0000 (00:10 +0200)]
Updated German translation
Bruce Cowan [Wed, 5 Sep 2012 18:36:55 +0000 (19:36 +0100)]
Updated British English translation
Dan Winship [Tue, 4 Sep 2012 12:51:56 +0000 (08:51 -0400)]
2.33.12
Gabor Kelemen [Mon, 3 Sep 2012 14:30:07 +0000 (16:30 +0200)]
Updated Hungarian translation
Piotr Drąg [Sun, 2 Sep 2012 23:21:55 +0000 (01:21 +0200)]
Updated Polish translation
Luca Ferretti [Sun, 2 Sep 2012 20:53:56 +0000 (22:53 +0200)]
Updated Italian translation
Rūdolfs Mazurs [Sun, 2 Sep 2012 17:31:33 +0000 (20:31 +0300)]
Updated Latvian translation
Ihar Hrachyshka [Thu, 30 Aug 2012 12:27:28 +0000 (15:27 +0300)]
Updated Belarusian translation.
Nguyễn Thái Ngọc Duy [Wed, 22 Aug 2012 01:54:12 +0000 (08:54 +0700)]
Updated Vietnamese translation
Nguyễn Thái Ngọc Duy [Wed, 22 Aug 2012 01:52:45 +0000 (08:52 +0700)]
po/vi: imported from Damned Lies
Piotr Drąg [Tue, 21 Aug 2012 22:08:45 +0000 (00:08 +0200)]
Updated Polish translation
Stef Walter [Tue, 21 Aug 2012 11:37:10 +0000 (13:37 +0200)]
Revert "gnutls: Implement certificate-bytes and private-key-bytes properties"
* The corresponding commit was reverted in glib due to ABI issues
* Add g_tls_certificate_gnutls_get_bytes() for glib-networking use.
* Fix up the tests and other code so it works in the presence of this
reverted commit.
This reverts commit
c30a4bbaf5e6f937f9e15d4c5fba7ab40a52b6d3.
https://bugzilla.gnome.org/show_bug.cgi?id=682081
Andika Triwidada [Tue, 21 Aug 2012 06:46:00 +0000 (13:46 +0700)]
Updated Indonesian translation
Dan Winship [Mon, 20 Aug 2012 21:39:36 +0000 (17:39 -0400)]
2.33.10
Duarte Loreto [Sun, 19 Aug 2012 22:22:45 +0000 (23:22 +0100)]
Updated Portuguese translation
Krishnababu Krothapalli [Fri, 17 Aug 2012 13:53:39 +0000 (19:23 +0530)]
Updated Telugu Translations
Dan Winship [Sat, 11 Aug 2012 13:28:16 +0000 (09:28 -0400)]
gnutls: fix an uninitialized variable
https://bugzilla.gnome.org/show_bug.cgi?id=681636
Stef Walter [Sat, 11 Aug 2012 04:18:53 +0000 (06:18 +0200)]
tests: Fix typo
https://bugzilla.gnome.org/show_bug.cgi?id=681299
Colin Walters [Fri, 10 Aug 2012 22:40:41 +0000 (18:40 -0400)]
build: Fix srcdir != builddir
Stef Walter [Wed, 8 Aug 2012 04:52:12 +0000 (06:52 +0200)]
gnutls: Try to find root certificates locally if not anchored
* If a server erroneously sends us a root certificate, and it
is not anchored, then try to lookup a certificate for the same
issuer in the database.
* Add a test for this case.
https://bugzilla.gnome.org/show_bug.cgi?id=681299
Nilamdyuti Goswami [Wed, 8 Aug 2012 08:47:45 +0000 (14:17 +0530)]
Assamese translation updated
Dan Winship [Tue, 7 Aug 2012 19:10:28 +0000 (15:10 -0400)]
NEWS: update for 2.33.8
(pulled from the 2.33.8 tag, which is not on master because
glib-networking master already depended on post-glib-2.33.8 APIs at
the time I made the tarball.)
Dan Winship [Tue, 7 Aug 2012 19:07:30 +0000 (15:07 -0400)]
Post-release version bump, update glib requirement
Stef Walter [Wed, 1 Aug 2012 12:08:40 +0000 (14:08 +0200)]
gnutls: Fix threading/reference issues in GTlsFileDatabase
* We weren't refing stuff during the lock, and returned data
could have been freed by another thread while in use.
https://bugzilla.gnome.org/show_bug.cgi?id=681122
Stef Walter [Wed, 1 Aug 2012 11:38:25 +0000 (13:38 +0200)]
gnutls: port GTlsFileDatabaseGnutls to GBytes
* Instead of GByteArray, where possible
https://bugzilla.gnome.org/show_bug.cgi?id=681121
Stef Walter [Mon, 6 Aug 2012 20:36:08 +0000 (22:36 +0200)]
gnutls: Add tests for certificate-bytes and private-key-bytes props
* And rearrange the test setup/teardown slightly at the same time.
* Can't actually read out the private-key from the certificate so
just run through the motions of creating it with a private key.
https://bugzilla.gnome.org/show_bug.cgi?id=681319
Stef Walter [Mon, 6 Aug 2012 20:32:32 +0000 (22:32 +0200)]
gnutls: Implement certificate-bytes and private-key-bytes properties
* These were added in GTlsCertificate in glib
https://bugzilla.gnome.org/show_bug.cgi?id=681319
Stef Walter [Fri, 3 Aug 2012 16:46:36 +0000 (18:46 +0200)]
gnutls: Add test for g_tls_certificate_is_same()
* The test can't live in glib since there's no implementation of
GTlsCertificate there.
https://bugzilla.gnome.org/show_bug.cgi?id=681154
Stef Walter [Fri, 3 Aug 2012 15:59:25 +0000 (17:59 +0200)]
gnutls: Fix invalid printf format
https://bugzilla.gnome.org/show_bug.cgi?id=681148
Stef Walter [Wed, 1 Aug 2012 10:46:07 +0000 (12:46 +0200)]
gnutls: More tests for GTlsFileDatabaseGnutls
* Check the properties
* Check looking up certificates by issuer
https://bugzilla.gnome.org/show_bug.cgi?id=681120
Stef Walter [Tue, 31 Jul 2012 15:49:12 +0000 (17:49 +0200)]
tls: Split the database tests out from the certificate tests
* For now these test GTlsFileDatabase
https://bugzilla.gnome.org/show_bug.cgi?id=681119
Dan Winship [Thu, 2 Aug 2012 13:06:40 +0000 (09:06 -0400)]
gnutls: allow overriding the priority string with an environment variable
Based on a patch from Josh Rickmar
https://bugzilla.gnome.org/show_bug.cgi?id=677717
Chao-Hsiung Liao [Thu, 2 Aug 2012 02:11:52 +0000 (10:11 +0800)]
Updated Traditional Chinese translation(Hong Kong and Taiwan)
Alexsey Nadtochey [Wed, 1 Aug 2012 19:54:07 +0000 (23:54 +0400)]
Updated Russian translation
Fran Diéguez [Sun, 29 Jul 2012 23:53:13 +0000 (01:53 +0200)]
Updated Galician translations
Мирослав Николић [Wed, 25 Jul 2012 06:58:12 +0000 (08:58 +0200)]
Updated Serbian translation
Yaron Shahrabani [Tue, 24 Jul 2012 20:14:03 +0000 (23:14 +0300)]
Updated Hebrew translation.
Tom Tryfonidis [Tue, 24 Jul 2012 08:29:34 +0000 (11:29 +0300)]
Updated Greek translation
Kjartan Maraas [Mon, 23 Jul 2012 09:43:39 +0000 (11:43 +0200)]
Updated Norwegian bokmål translation
Aurimas Černius [Sat, 21 Jul 2012 21:19:08 +0000 (00:19 +0300)]
Updated Lithuanian translation
Daniel Mustieles [Sat, 21 Jul 2012 16:44:32 +0000 (18:44 +0200)]
Updated Spanish translation
Matej Urbančič [Thu, 19 Jul 2012 07:11:51 +0000 (09:11 +0200)]
Updated Slovenian translation
Dan Winship [Mon, 2 Jul 2012 16:58:02 +0000 (12:58 -0400)]
gnutls: handle simultaneous ops, do handshaking in a thread
Keep separate internal read/write states so that you can do
simultaneous reads/writes (either simultaneous async ops in the same
thread, or simultaneous sync ops in different threads). Add
locking/blocking so that this works correctly even in the presence of
rehandshakes (and add test cases for this).
As part of this, change handshaking so that the I/O part of it always
happens in a separate thread, which has three advantages:
1. It simplifies GTlsConnectionGnutlsSource by making it not have to
worry about flipping back and forth between read and write.
2. (Future) It will let the caller asynchronously handle
certificate-related functionality that is implemented via
synchronous callbacks in gnutls.
3. (Future) We can use g_task_set_return_on_cancel() to allow
cancellation even during uncancellable PKCS#11 operations.
https://bugzilla.gnome.org/show_bug.cgi?id=656343
https://bugzilla.gnome.org/show_bug.cgi?id=660252
Dan Winship [Tue, 17 Jul 2012 22:47:26 +0000 (18:47 -0400)]
gnutls: reorganize certificate handling
Move some of the code in handshake_internal() out into helper
functions, and move the code from the client and server verify_peer()
implementations into a helper function in gtlsconnection-gnutls.c as
well, to reduce duplication between them.
Dan Winship [Sun, 12 Feb 2012 16:30:22 +0000 (11:30 -0500)]
gnutls: preserve handshake errors
If an error occurs during handshake, remember it for later, since the
connection is now essentially broken.
Dan Winship [Thu, 28 Jun 2012 15:35:46 +0000 (11:35 -0400)]
gnutls: use g_clear_object()