summaryrefslogtreecommitdiff
path: root/tests
AgeCommit message (Collapse)AuthorFilesLines
2013-05-30vcard/supported-fields: catch up with fd.o #64319HEADmasterSimon McVittie1-0/+1
2013-05-30Merge branch 'telepathy-gabble-0.16'Simon McVittie3-1/+38
Conflicts: NEWS configure.ac lib/ext/wocky
2013-05-29Add a regression test for fd.o #65036Simon McVittie2-0/+37
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=65036 Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
2013-05-29tests: fix JabberAuthenticator when self.emit_events is FalseSimon McVittie1-1/+1
We don't currently use JabberAuthenticator in this mode, so nobody noticed that it didn't work. I'm about to add a test that does use it.
2013-04-04Merge branch 'telepathy-gabble-0.16'Simon McVittie1-0/+6
2013-04-04Merge with-session-bus.sh changes from telepathy-glibSimon McVittie1-0/+6
We also have local changes, which should go "upstream" to telepathy-glib at some point. Bug: https://bugs.freedesktop.org/show_bug.cgi?id=63119
2013-02-22MUC: stop sending chat states after <error type=wait>Will Thompson1-5/+34
Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=43166#c0
2013-02-22muc/chat-states: fix and uncomment commented chunkWill Thompson1-16/+16
This didn't work as-is because the <presence type='unavailable'> had no <x> child. Using make_muc_presence() does the job nicely.
2013-02-22muc/chat-states: tidy up checking message bodiesWill Thompson1-8/+4
2013-02-22muc: test resource-constraint is exposed reasonablyWill Thompson1-0/+10
https://bugs.freedesktop.org/show_bug.cgi?id=43166#c7
2013-02-22Test Prosody's MUC's rate limit is exposed correctly nowWill Thompson1-12/+59
There were a combination of problems before: * The server's error message was not exposed in the delivery report; * Wocky didn't recognise policy-violation; * There was a bug in Wocky's error parsing code which clobbered the message type. So many bugs for such a small stanza!
2013-02-15muc/send-error: refactorWill Thompson2-27/+50
There will be more cases shortly
2013-02-12tests: don't block on Disconnect() without sending </stream>Will Thompson3-9/+8
These three tests were making blocking calls to Disconnect() rather than using the disconnect_conn() helper. As a result, Gabble was waiting 5 seconds for the test server to send back </stream:stream> before timing out. Using disconnect_conn(), which sends the stream close back properly, shaves between 15 and 18 wasted seconds off the test suite.
2013-01-16Update Wocky snapshot for wocky.h with Jingle headersWill Thompson1-4/+0
2012-12-11Rename most of GabbleJingle* to WockyJingle*Will Thompson1-1/+1
This renames the classes I'm moving to Wocky, and updates all the remaining code to cope. A few things are missed out because it's only reasonable to change them when the code is in Wocky: • GABBLE_TYPE_STUN_SERVER_SOURCE (the GEnum for WockyStunServerSource) can't be renamed easily because it is generated by Gabble's enumtype thing. • GABBLE_DEBUG_MEDIA. Here's the program I used, for posterity. #!/usr/bin/perl -pi use strict; BEGIN { undef $/; } my @Classes = qw(Factory Session Transport_Iface Transport_IceUdp Transport_RawUdp Transport_Google Content Info Media_Rtp Relay); for my $Class (@Classes) { my $FullClass = "Jingle$Class"; $FullClass =~ s/_//g; my $under_class = "jingle_" . lc($Class); my $UNDER_CLASS = uc($under_class); s/Gabble$FullClass/Wocky$FullClass/g; s/gabble_$under_class/wocky_$under_class/g; s/GABBLE_$UNDER_CLASS/WOCKY_$UNDER_CLASS/g; s/GABBLE_TYPE_$UNDER_CLASS/WOCKY_TYPE_$UNDER_CLASS/g; s/GABBLE_IS_$UNDER_CLASS/WOCKY_IS_$UNDER_CLASS/g; } s/GabbleGoogleRelay/WockyGoogleRelay/g; s/gabble_google_relay/wocky_google_relay/g; s/GABBLE_N_JINGLE_RELAY_TYPES/WOCKY_N_JINGLE_RELAY_TYPES/g; s/gabble_stun_server/wocky_stun_server/g; s/GABBLE_STUN_SERVER/WOCKY_STUN_SERVER/g; # Can't rename GABBLE_TYPE_STUN_SERVER_SOURCE because it's generated. Blah. s/GabbleStunServer/WockyStunServer/g; my @Blah_Blah = qw(Candidate State Action Content_Senders Transport_Type Transport_Protocol Candidate_Type Reason Transport_State Dialect Content_State Media_Type Codec Media_Description Feedback_Message Rtp_Header_Extension Media_Profile); for my $Class (@Blah_Blah) { my $FullClass = "Jingle$Class"; $FullClass =~ s/_//g; my $under_class = "jingle_" . lc($Class); my $UNDER_CLASS = uc($under_class); s/\b$FullClass/Wocky$FullClass/g; s/\b_$FullClass/_Wocky$FullClass/g; s/\b$under_class/wocky_$under_class/g; s/\b$UNDER_CLASS/WOCKY_$UNDER_CLASS/g; } s/MAX_JINGLE_STATES/WOCKY_N_JINGLE_STATES/g; s/JINGLE_IS_GOOGLE_DIALECT/WOCKY_JINGLE_DIALECT_IS_GOOGLE/g;
2012-12-07tests: don't crash in verbose mode on unicode in stanzasWill Thompson1-1/+9
We have this hack that sets __repr__ on domish.Element to call its toXml() method, which ends up being used by the verbose logging code when it dumps all of an event's attributes. Unfortunately, this blows up if a stanza contains non-ASCII characters, because repr() tries to convert unicode to str using .encode('ascii'). This made running jingle/test-send-file in verbose mode fail, because the file being sent by the test has a non-ASCII filename. The fix is to make the __repr__ hack escape non-ASCII characters using the unicode-escape codec.
2012-12-06Test all this receipt stuff.Will Thompson4-0/+200
I did write this test as I went along, but I forgot to commit it. Whoops.
2012-12-06Don't crash on <success> after closing auth channel.Will Thompson1-1/+1
2012-12-06tests: add a couple more cases to sasl/abortWill Thompson1-0/+17
There's no reason these should be any different, but I thought it might be worth trying them anyway.
2012-12-06Don't crash on <challenge> after closing auth channelWill Thompson1-1/+1
https://bugs.freedesktop.org/show_bug.cgi?id=52146
2012-12-06sasl-channel: don't crash on Abort() then <challenge>Will Thompson1-1/+1
2012-12-06Add commented-out regression tests for fd.o#52146Will Thompson2-0/+87
test_close_then_challenge() and test_close_then_success() each trigger similar bugs. test_abort_then_challenge() and test_abort_then_success() both trigger the same bugs by a different route, which I don't think my first attempt at fixing the original bug would have caught. test_close_then_failure() and test_abort_then_failure() pass, and are here for completeness.
2012-12-06sasl/abort: clean up common test codeWill Thompson1-29/+15
2012-12-06sasl/abort.py: clean up invoking test casesWill Thompson1-18/+12
2012-12-06gabbletest: exit if we can't connect to the busWill Thompson1-1/+5
Previously, if the bus wasn't there (perhaps I killed it when trying and failing to ^C the test suite?) the test would just sit there forever after the unhandled exception made its way up to the mainloop.
2012-12-06Test ClearAvatar()Will Thompson2-0/+28
I wasn't sure from reading the code that this actually worked any more, and it wasn't tested. I'm happy to say that it does work.
2012-12-06Regression test for malformed privacy list repliesWill Thompson1-5/+15
2012-12-06Tidy up privacy list test very slightlyWill Thompson1-6/+3
2012-12-06Regression test for 57521Will Thompson1-0/+31
2012-12-06Merge branch '57546-there-can-be-only-one-jingletest'Will Thompson10-339/+63
https://bugs.freedesktop.org/show_bug.cgi?id=57546
2012-12-06Merge branch '25961-stun-srv'Will Thompson12-177/+158
https://bugs.freedesktop.org/show_bug.cgi?id=25961 https://bugs.freedesktop.org/show_bug.cgi?id=25385
2012-12-06test-multift: clean up a bitWill Thompson1-11/+1
2012-12-06test-multift: don't rely on q.expect timing outWill Thompson1-22/+23
Previously, this test checked that events didn't happen by waiting for them to happen, asserting if they did, and catching the TimeoutError which occurred if they didn't. But it's more obvious what's happening to forbid the event, take all the actions, and then sync the stream to make sure it hasn't appeared. Plus, it shaves 6 seconds off the test.
2012-12-06Fix file transfer tests for these new interfacesWill Thompson3-9/+24
2012-11-26Remove the olde jingletestWill Thompson3-231/+0
2012-11-26jingle/google-relay: use jingletest2Will Thompson1-13/+5
2012-11-26jingle/test-wait-for-caps-incomplete: don't use jingletestWill Thompson1-5/+3
This test didn't actually use any of the jingle protocol stuff at all—it's just a caps test.
2012-11-26jingle/test-wait-for-caps-incomplete: unskip half the testWill Thompson1-3/+0
The connection doesn't seem to leak any more. I added this in 07822b272 because <https://bugs.freedesktop.org/show_bug.cgi?id=29790>. I haven't investigated fully, but the test passes, sooooo... :)
2012-11-26jingle/test-wait-for-caps: use jingletest2Will Thompson2-16/+18
2012-11-26jingle/stream-handler: use jingletest2Will Thompson1-3/+5
2012-11-26jingle/payload-types: use jingletest2Will Thompson1-16/+9
2012-11-26jingle/payload-types: clean up a bitWill Thompson1-8/+3
2012-11-26jingle/payload-types: remove obsolete FIXMEWill Thompson1-11/+0
2012-11-26jingle/outgoing-many-streams: use jingletest2Will Thompson1-18/+9
2012-11-26jingle/decloak-peer: use jingletest2Will Thompson1-13/+9
2012-11-26jingletest2: make classes new-styleWill Thompson1-2/+2
2012-11-26jingle/stun-server: use jingletest2Will Thompson2-42/+28
In passing, test all dialects. (Though I think this may be a little overzealous.) https://bugs.freedesktop.org/show_bug.cgi?id=25385
2012-11-26tls/server-tls-channel: use with statement when reading certsWill Thompson1-10/+4
2012-11-26jingle-info: discover a STUN server using SRVWill Thompson2-0/+9
Derived from a patch by Mike Ruprecht. https://bugs.freedesktop.org/show_bug.cgi?id=25961
2012-11-26jingle-info: don't discard google:jingleinfo pushes.Will Thompson1-17/+44
Previously, Google's reply to our original google:jingleinfo query would be used correctly, but the act of passing it to gabble_jingle_info_take_stun_server would set get_stun_from_jingle to FALSE, causing any later calls to got_jingle_info_stanza() to ignore the updated STUN server information. We address this by distinguishing between user-set and server-provided STUN servers, as opposed to just between fallback or not.
generated by cgit v1.2.3 (git 2.39.1) at 2025-04-22 15:11:51 +0000