[Telepathy] ANNOUNCE: telepathy-gabble 0.7.4

Simon McVittie simon.mcvittie at collabora.co.uk
Thu May 1 11:27:47 PDT 2008 

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

The "stop making coredumps, I want to go to the pub" release.

tar: http://telepathy.freedesktop.org/releases/telepathy-gabble/telepathy-gabble-0.7.4.tar.gz
sig: http://telepathy.freedesktop.org/releases/telepathy-gabble/telepathy-gabble-0.7.4.tar.gz.asc
darcs: http://projects.collabora.co.uk/darcs/telepathy/telepathy-gabble
tag: "telepathy-gabble 0.7.4"

Enhancements:

* fd.o #14467: aggregate MembersChanged signals when joining MUCs, so we
  get one big signal rather than lots of single-contact signals, potentially
  reducing client round-trips

* Re-enable some regression tests that were accidentally removed in 0.7.3

* Build documentation for our extensions to XMPP by default

Fixes:

* fd.o #12791, Debian #446259: fix assertion failure with strange roster items

* fd.o #14343: fix race condition that can lose messages between accepting
  a D-Bus tube and actually connecting to it

* fd.o #14602: update the StreamedMedia call API to telepathy-spec 0.17.3

* fd.o #14806: when we add ourselves to the local-pending set for incoming
  calls, set the caller as the actor

* fd.o #15095: fix MUC contacts always appearing to be offline

* fd.o #15198: fix assertion failure when calling GetCapabilities on an invalid
  handle

* When removing a media stream before the initiate is sent, send it if possible

* when blocking gtalk contacts, discard their presence

* disconnect the connection if we don't receive a reply to the discovery
  request for server features

* stop using Automake 1.9 features, and prefer to autoreconf using 1.8 (if
  available) to ensure that we don't accidentally re-add a 1.9 dependency

* make sure teardown of the disco object doesn't crash if one of its callbacks
  causes it to be destroyed (bug exposed by test-disco-no-reply.py regression
  test)

Full changes follow.

Regards,
    Simon

Thu May  1 19:13:43 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  tagged telepathy-gabble 0.7.4

Thu May  1 19:10:23 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Prepare release 0.7.4

Thu May  1 19:04:47 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Release weak ref before calling into user code in GabbleDisco, to avoid crashes if the user code causes the weakly reffed object to be destroyed

Thu May  1 18:31:53 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * GabbleDisco: temporarily ref the disco object while calling out to user code

Thu May  1 18:31:36 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * tests/twisted/Makefile.am: consider it to be a failure if Gabble dumps core

Thu May  1 18:19:28 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Update NEWS

Thu May  1 17:53:00 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Add regression test for fd.o #15198

Thu May  1 17:43:58 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * After returning an error from gabble_connection_get_capabilities, don't happily carry on regardless (fd.o #15198)

Thu May  1 17:34:29 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * When adding ourselves to the local-pending set for an incoming call, make the actor the handle of the caller (fd.o #14806)

Thu May  1 15:24:18 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Improve reference to spec in NEWS

Thu May  1 15:20:54 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Update NEWS

Wed Apr 30 18:46:23 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * ./tests/twisted/vcard/test-muc-alias.py: adjust for #14467 being fixed

Wed Apr 30 18:45:57 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * tests/twisted/tubes/test-muc-*: adjust for #14467 being fixed

Wed Apr 30 18:44:57 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * test-get-available-tubes.py: adjust for #14467 having been fixed

Wed Apr 30 18:40:13 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Adjust sequence of events in test-muc-ownership.py to assert that group membership is aggregated

Wed Apr 30 18:32:37 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Amend muc/test-muc.py to match Alban's fix for fd.o #14467

Fri Apr 25 15:48:30 BST 2008  Alban Crequy <alban.crequy at collabora.co.uk>
  * Better comment about aggregation

Fri Apr 25 15:25:20 BST 2008  Alban Crequy <alban.crequy at collabora.co.uk>
  * Coding style, rename variable to initial_members_aggregator, delete the boolean

Fri Apr 25 13:19:56 BST 2008  Alban Crequy <alban.crequy at collabora.co.uk>
  * Use TpHandleSet instead of TpIntSet

Wed Mar  5 13:20:43 GMT 2008  Alban Crequy <alban.crequy at collabora.co.uk>
  * Fix indentation

Tue Feb 12 14:42:37 GMT 2008  Alban Crequy <alban.crequy at collabora.co.uk>
  * MembersChanged aggregation when joining MUC

Wed Apr 30 18:17:20 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * test-outgoing-call.py: add a hackish workaround for fd.o#15769 so we don't need to fix it right now

Wed Apr 30 17:23:48 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * test-outgoing-call.py: Don't expect to see a NewSessionHandler until we RequestStreams

Wed Apr 30 17:04:36 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * test-outgoing-call.py: don't add the peer to the group interface, but instead let them appear when called

Wed Apr 30 17:03:38 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * test-outgoing-call-deprecated2.py: instead of requesting a call to foo at bar.com, request a new blank channel and invite foo at bar.com

Wed Apr 30 17:02:31 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Duplicate tests/twisted/jingle/test-outgoing-call-deprecated.py again, it turns out to have been more deprecated than I thought

Wed Apr 30 17:01:40 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Duplicate test-outgoing-call.py as test-outgoing-call-deprecated.py

Wed Apr 30 17:00:26 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * In regression tests, correctly request streams to the other guy, not to ourselves

Wed Apr 30 16:49:36 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * GabbleMediaChannel: when session state changes to indicate we have contacted the peer, stuff them into remote-pending if they're not already there

Wed Apr 30 16:49:26 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Fix variable name

Wed Apr 30 16:41:39 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * gabble_media_channel_request_streams: don't restrict to only creating streams to a member of the channel (fd.o #14602)

Wed Apr 30 16:39:59 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * gabble_media_channel_request_streams: lazily create a session if necessary (also simplify error handling)

Wed Apr 30 16:32:22 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * GabbleMediaSession: ref the peer handle for the duration. We can't rely on it being reffed by the channel's group interface any more, with telepathy-spec 0.17.3 semantics

Wed Apr 30 16:24:55 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Allow adding a member to a Media channel if and only if either it has no session, or it has a session and they are the peer

Wed Apr 30 16:19:28 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Move responsibility for checking VoIP capabilities into create_session()

Wed Apr 30 16:13:18 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * In _gabble_media_channel_dispatch_session_action, assert that we're not about to cause create_session() to get confused

Wed Apr 30 16:12:43 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Have create_session() return a GError; for the moment, assert that it always succeeds

Wed Apr 30 16:12:30 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Whitespace

Mon Apr 28 19:02:16 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Reference Debian #446259 as well as fd.o #12791

Mon Apr 28 18:57:28 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Update NEWS with merged changes since 0.7.3

Mon Apr 28 17:53:42 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Amend #12791 regression test to test with what the submitter actually had on their roster, as well as with a garbled roster that the server shouldn't have sent us

Mon Apr 28 17:53:23 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Skip roster <item>s that contain '/' (fd.o #12791)

Mon Apr 28 17:37:05 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Add fd.o #12791 regression test back as a separate file

Mon Apr 28 17:36:35 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * ./tests/twisted/roster/groups.py: revert addition of #12791 regression test

Wed Apr 23 15:30:54 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * roster: ref handles when they are put in a GroupMembershipUpdate (hopefully fixes fd.o #12791)

Wed Apr 23 15:29:31 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * roster/groups test: include amy's <item> in the roster twice, removing a group the second time (this causes the same crash as fd.o #12791, although it doesn't seem very likely to be the real cause)

Wed Apr 23 15:29:10 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * roster/groups test: if assertions fail, print more useful info

Wed Apr 23 14:36:03 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * roster: amend comments for GroupMembershipUpdate/GroupsUpdateContext to indicate where things are borrowed from

Wed Apr 23 14:19:41 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Add a comment to ./tests/twisted/roster/groups.py explaining how fragile it is

Wed Apr 23 14:17:43 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * _gabble_roster_item_update: eliminate unnecessary temporary, old_groups
  
  old_groups is slightly misleadingly named - it's a direct pointer to the intset
  underlying item->groups, so its contents change whenever item->groups does.
  (In practice we only used it once, before making any changes.)

Wed Apr 23 12:36:47 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Add a regression test for roster groups

Wed Apr 23 12:32:02 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * docs/Makefile.am: projects.ccu is now people.ccu

Wed Apr 23 12:04:38 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Update xep.xsl to r1806. Stop pasting in the revision number, it's error-prone if there is a <?xml?> declaration

Wed Apr 23 11:35:10 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * roster: misc stylistic fixes

Wed Apr 23 11:31:34 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * roster: Annotate GroupsUpdateContext, GroupMembershipUpdate with ownership

Tue Apr 22 18:39:50 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Update xep.xsl to r1803 with an even more lenient license

Tue Apr 22 18:20:57 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Put HTML docs in $(htmldir)

Tue Apr 22 18:18:13 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * COPYING: explain the status of tools/xep.xsl

Tue Apr 22 18:10:17 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Build documentation by default

Tue Apr 22 18:09:13 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Add xep.xsl from xmpp.org, and a maintainer-only Makefile target to resync it

Tue Apr 22 12:05:21 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * autogen.sh: fix bashism "&>"

Mon Apr 21 18:43:17 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Remove ls-interfaces.xsl from tools/ (no longer used or needed)

Mon Apr 21 17:45:44 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Add INSTALL to _boring

Mon Apr 21 17:44:30 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Remove unnecessary rule from data/Makefile.am

Mon Apr 21 17:44:14 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Use slightly nicer sed syntax in data/Makefile.am

Mon Apr 21 17:43:50 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Add comments explaining why the .service.in file has a shorter name than you might expect

Thu Apr 17 15:30:11 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Prefer to use automake-1.8, to check that our build will still work in Scratchbox

Thu Apr 17 15:29:55 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Remove INSTALL from Darcs, autogen.sh will make one for us

Thu Apr 17 15:29:24 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * tests/twisted/tools/: also rename .service.in file here

Fri Apr 11 18:37:21 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Turn off tar-ustar in configure.ac - it's not supported by automake 1.8, as used on Maemo

Thu Apr 17 14:54:29 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Don't ship as-ac-expand.m4, we no longer need it

Thu Apr 17 14:53:27 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Don't AC_DEFINE, AC_SUBST or AS_AC_EXPAND the D-Bus services directory - there's no point, and AS_AC_EXPAND is Bad And Wrong

Thu Apr 17 14:49:50 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Simplify data/Makefile.am and rename .service.in filename to be shorter.
  
  Shortening the filename should mean we no longer need to use tar-ustar,
  since we're much less likely to hit the old 99-character tarball filename
  limit, even if we do a datestamped prerelease like
  `make dist distdir=telepathy-gabble-0.7.3.1-prerelease-20080416` (prereleases
  for laptop.org were the original motivation for using tar-ustar).

Mon Apr 21 14:48:42 BST 2008  Guillaume Desmottes <guillaume.desmottes at collabora.co.uk>
  * tests/twisted/Makefile.am: sort TWISTED_TESTS

Mon Apr 21 14:43:08 BST 2008  Guillaume Desmottes <guillaume.desmottes at collabora.co.uk>
  * connection: lower-case disco_reply_timeout and make it static

Mon Apr 21 14:29:38 BST 2008  Guillaume Desmottes <guillaume.desmottes at collabora.co.uk>
  * test-disco-no-reply.py: no need to extend the timeout anymore as Gabble will timeout after 3 seconds

Mon Apr 21 14:28:49 BST 2008  Guillaume Desmottes <guillaume.desmottes at collabora.co.uk>
  * run a tweaked version of Gabble with tests

Mon Apr 21 14:26:29 BST 2008  Guillaume Desmottes <guillaume.desmottes at collabora.co.uk>
  * libtoolize main (renamed gabble_main) and call it from main.c

Mon Apr 21 14:12:09 BST 2008  Guillaume Desmottes <guillaume.desmottes at collabora.co.uk>
  * gabble-connection: allow to tweak the timeout of the disco reply (should be used only for tests)

Mon Apr 21 14:04:29 BST 2008  Guillaume Desmottes <guillaume.desmottes at collabora.co.uk>
  * slow tests was a bad idea. All tests should be run by default

Wed Apr 16 16:03:02 BST 2008  Guillaume Desmottes <guillaume.desmottes at collabora.co.uk>
  * gabble_debug_set_log_file_from_env: trunc the file when opening

Mon Apr 14 16:51:18 BST 2008  Guillaume Desmottes <guillaume.desmottes at collabora.co.uk>
  * we don't need AC_DEFINE and AC_SUBST for --enable-slow-tests

Mon Apr 14 16:34:37 BST 2008  Guillaume Desmottes <guillaume.desmottes at collabora.co.uk>
  * run slow tests only if build with --enable-slow-tests

Mon Apr 14 16:19:26 BST 2008  Guillaume Desmottes <guillaume.desmottes at collabora.co.uk>
  * move test-disco-no-reply.py to the slow directory

Mon Apr 14 13:54:52 BST 2008  Guillaume Desmottes <guillaume.desmottes at collabora.co.uk>
  * add test-disco-no-reply.py to tests set

Mon Apr 14 13:54:34 BST 2008  Guillaume Desmottes <guillaume.desmottes at collabora.co.uk>
  * add test-disco-no-reply.py (re-recorded)

Mon Apr 14 13:45:56 BST 2008  Guillaume Desmottes <guillaume.desmottes at collabora.co.uk>
  * gabble-connection: disconnect the connection if we don't receive a reply to the discovery request

Mon Apr 14 13:32:01 BST 2008  Guillaume Desmottes <guillaume.desmottes at collabora.co.uk>
  * allow user to define his own event timeout

Mon Apr 14 13:22:22 BST 2008  Guillaume Desmottes <guillaume.desmottes at collabora.co.uk>
  * gabbletest.py: allow user to pass his own protocol to exec_test

Tue Mar  4 21:58:41 GMT 2008  Senko Rasic <senko.rasic at collabora.co.uk>
  * fix function comment

Thu Mar  6 21:59:22 GMT 2008  Senko Rasic <senko.rasic at collabora.co.uk>
  * when blocking gtalk contacts, discard their presence

Tue Mar 11 14:54:40 GMT 2008  Senko Rasic <senko.rasic at collabora.co.uk>
  * don't try to gracefully avoid a bug if it's there; assert what can't happen, and fix the bug if and when it happens

Tue Mar 11 12:34:52 GMT 2008  Senko Rasic <senko.rasic at collabora.co.uk>
  * _got_self_avatar_for_get_known_avatar_tokens: in the corner case of us starting disconnect right after GetKnownAvatarTokens call but before our own token is known, don't report bogus handle 0 to the user

Tue Apr 15 15:43:24 BST 2008  Alban Crequy <alban.crequy at collabora.co.uk>
  * Fix Copyright

Tue Apr 15 12:01:17 BST 2008  Alban Crequy <alban.crequy at collabora.co.uk>
  * Coding style

Tue Apr 15 11:52:12 BST 2008  Alban Crequy <alban.crequy at collabora.co.uk>
  * Only limit the queue size with MAX_QUEUE_SIZE bytes. No limit on number of D-Bus messages

Tue Apr 15 11:45:37 BST 2008  Alban Crequy <alban.crequy at collabora.co.uk>
  * Free priv->dbus_msg_queue in _dispose

Tue Apr 15 11:45:17 BST 2008  Alban Crequy <alban.crequy at collabora.co.uk>
  * Code style

Mon Apr 14 18:46:23 BST 2008  Alban Crequy <alban.crequy at collabora.co.uk>
  * Bug #14343: limit the queue to 16 messages and 4MB

Mon Apr 14 17:47:49 BST 2008  Alban Crequy <alban.crequy at collabora.co.uk>
  * Fix coding style

Mon Apr 14 17:27:27 BST 2008  Alban Crequy <alban.crequy at collabora.co.uk>
  * print debug message "deliver" only if we deliver the message to the application

Mon Apr 14 17:16:35 BST 2008  Alban Crequy <alban.crequy at collabora.co.uk>
  * Bug #14343: Do not queue more than 16 messages

Mon Apr 14 17:15:19 BST 2008  Alban Crequy <alban.crequy at collabora.co.uk>
  * Bug #14343: Use GSList instead of GList. Do not leak the list.

Fri Apr 11 16:38:40 BST 2008  Alban Crequy <alban.crequy at collabora.co.uk>
  * Fix race in Bug #14343

Mon Apr 14 12:35:49 BST 2008  Guillaume Desmottes <guillaume.desmottes at collabora.co.uk>
  * re-enable tests which aren't in any sub directory

Sat Apr 12 10:30:59 BST 2008  Sjoerd Simons <sjoerd at luon.net>
  * When removing a media stream before the initiate is sent, send it if possible

Fri Apr 11 15:57:36 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  * Nano version 0.7.3.1

Fri Apr 11 15:29:42 BST 2008  Simon McVittie <simon.mcvittie at collabora.co.uk>
  tagged telepathy-gabble 0.7.3
-----BEGIN PGP SIGNATURE-----

iD8DBQFIGgujWSc8zVUw7HYRAlgVAKCUm9if3GUMJkm5v+mIcurtxzupMwCg3Imy
7CX3cMW+PSYdusjvVGHdmg8=
=o+ot
-----END PGP SIGNATURE-----

More information about the Telepathy mailing list