summaryrefslogtreecommitdiff
AgeCommit message (Collapse)AuthorFilesLines
2010-02-08Fix typo.Jonny Lamb1-2/+2
Signed-off-by: Jonny Lamb <jonny.lamb@collabora.co.uk>
2010-01-28version 0.8.10.1Dafydd Harries1-1/+1
2010-01-28version 0.8.10telepathy-gabble-0.8.10Dafydd Harries1-2/+2
2010-01-28update NEWS fileDafydd Harries1-1/+6
2010-01-27Don't crash on disco requests with no id=''Will Thompson1-0/+5
It's illegal to send <iq/> stanzas without an id='' (per the first rule of IQ Semantics in XMPP Core <http://xmpp.org/rfcs/rfc3920.html#stanzas-semantics-iq>), but neither ejabberd nor Loudmouth filters them out before they reach Gabble. If passed an IQ with no id='', lm_iq_message_make_result() returns NULL; connection_iq_disco_cb() assumed that it returned a valid message, and hence crashed. Other places which call lm_iq_message_make_result() already check if the result was NULL, so this is the only place that needed fixing. Fixes: fd.o#26271 Reviewed-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
2010-01-22Add --enable-is-a-phoneWill Thompson3-0/+18
Reviewed-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk> Reviewed-by: Guillaume Desmottes <guillaume.desmottes@collabora.co.uk>
2010-01-20Merge branch 'caps-cache-0.8' into telepathy-gabble-0.8Dafydd Harries25-100/+935
Reviewed-by: Will Thompson
2010-01-20caps cache: create correct directory when falling back to default pathDafydd Harries1-1/+1
2010-01-13restore per-channel-manager caps from the cacheDafydd Harries2-2/+16
2010-01-13test-caps-file-transfer.py: use a unique client nameDafydd Harries1-1/+1
This avoids problems with pre-existing cache entries.
2010-01-13make caps cache persist to disk using SQLiteDafydd Harries7-19/+556
2010-01-13free shared caps cache when CM exitsDafydd Harries3-0/+14
2010-01-13add initial implementation of CM-wide caps cacheDafydd Harries7-5/+240
This is in memory only, for now.
2010-01-12factor out capability URI extraction from messagesDafydd Harries5-45/+56
Then, make GabbleCapsChannelManager's parse_caps take the resulting gchar **.
2010-01-11make send_disco_reply() take a list of identitiesDafydd Harries7-16/+26
2010-01-11make use of send_disco_reply in testsDafydd Harries7-39/+33
2010-01-11free debugger singleton on exitDafydd Harries3-0/+15
2010-01-11compute_caps_hash(): verify that identities are well-formedDafydd Harries1-0/+5
2010-01-11Grab UNRELEASED-checking distcheck-hook from tp-glibWill Thompson2-0/+17
Fixes: fd.o#23848 Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
2010-01-11Treat <item-not-found/> as the empty vCardWill Thompson3-1/+55
When we want to update our vCard, we first retrieve our vCard from the server. XEP-0054 says that the server MUST return <item-not-found/> if you have no vCard set, but we previously assumed that the server would return success, with an empty <vCard/> node. Prosody respects the XEP. So, let's treat the former case identically to the latter. Fixes: fd.o#25987 Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
2010-01-11Merge branch '0.8-chat-state-negotiation' into telepathy-gabble-0.8Will Thompson4-95/+282
2010-01-11Negotiate chat state support for capsless contactsWill Thompson4-33/+218
XEP-0085 §5.1 defines how to negotiate support for chat states when you don't know a contact's caps. Roughly: • If you don't know whether someone supports them, don't send stand-alone notifications, but include <active/> in messages you send; • If you receive a chat state, mark the contact as supporting chat states; • If you receive a message without a chat state, mark the contact as not supporting chat states. This is complicated slightly by multiple resources, but basically we follow the above rules, resetting whenever we change which resource we're sending to.
2010-01-11Send chat state notifications as type='chat'Will Thompson2-3/+3
This has been a SHOULD since 2003-09-18. From XEP-0085 §5.4: This protocol SHOULD NOT be used with message types other than "chat" or "groupchat".
2010-01-11Refactor determining whether to send chat statesWill Thompson1-45/+43
2010-01-11Tidy up the chat states test a smidgeWill Thompson1-21/+25
2010-01-08add servicetest.py to the list of Twisted testsDafydd Harries1-2/+2
This is so that its self-tests get run, as with caps_helper.py.
2009-12-15NEWS for fd.o#25341Will Thompson1-0/+7
2009-12-15Merge branch '0.8-fd.o#25341' into telepathy-gabble-0.8Will Thompson3-0/+96
Signed-off-by: Will Thompson <will.thompson@collabora.co.uk>
2009-12-15Test that the fix for fd.o#25341 works.Will Thompson2-0/+48
2009-12-14vcard-manager: Iterate over hashtable instead of usin g_hash_table_find to ↵Andre Moreira Magalhaes (andrunko)1-18/+18
check if vcard changed.
2009-12-14vcard-manager: Do not try to set vcard fields not supported by google servers.Andre Moreira Magalhaes (andrunko)1-2/+22
2009-12-14fd.o#25341: Always sets its own vcard on loginAndre Moreira Magalhaes (andrunko)1-0/+28
Note that gmail only stores vcard fields FN and PHOTO, so when using empathy the nickname (alias) will be updated on every login.
2009-12-08set-set-disconnect.py: fix race when setting the avatarGuillaume Desmottes1-7/+4
Test was relying on the ordering of the vCard IQ reply and the D-Bus method call. Ensure that the first vCard reply has been received by Gabble and then wait for the second vCard query once we have called the SetAvatar method.
2009-12-08Check coding style in test-ns.c not test.nsWill Thompson1-1/+1
2009-12-08Make log-strip.py process its argumentsWill Thompson1-2/+12
2009-12-08Make log-strip.py executableWill Thompson1-1/+1
2009-12-08Tidy some assertions in caps_helper.pyWill Thompson1-11/+16
These triggered while I was experimenting with different places to put the test name, and I thought I'd fix them up.
2009-12-08Shove the test name into 'account' as the resourceWill Thompson1-1/+5
This means the name of (most) tests appears in gabble-testing.log; 'resource' takes priority so the test jid still comes out as test@localhost/Resource.
2009-12-08Make 'resource' override a resource in 'account'Will Thompson1-8/+10
This way round makes more sense to me and (empirically) to #telepathy.
2009-12-07Bump nano-version to 0.8.9.1Will Thompson1-1/+1
2009-12-07Version 0.8.9telepathy-gabble-0.8.9Will Thompson2-3/+5
2009-12-07Clarify no-mo-discoWill Thompson1-3/+5
2009-12-07NEWS for 0.8.9Will Thompson1-0/+32
2009-12-07Make --disable-assumed-ft-cap work.Will Thompson1-4/+4
configure looks for args of the form "--enable-foo-bar" or "--disable-foo-bar", and if if foo-bar is in a big table built from all the AC_ARG_ENABLEs in configure.ac, it sets the variable enable_foo_bar. So then our code would run, and set enable_assumed_ft_cap to yes, trampling over any value it was set to by parsing the arguments to configure. The fourth argument of AC_ARG_ENABLE is the correct place to specify what to do if the user doesn't pass anything about this option to configure, so I moved the defaulting there; I also renamed the variable to avoid the collision, consistent with the Werror arg.
2009-12-07Fix coding style checks for out-of-tree buildsWill Thompson1-2/+2
2009-12-07Move pmuc-v1 #define up with the other bundles'Will Thompson1-2/+1
2009-12-07Fall back to stun.telepathy.im not stun.ccuWill Thompson2-3/+3
2009-12-07test-muc-ownership: Remove magic numbered handlesWill Thompson1-25/+39
This makes the test a bit less write-only.
2009-12-07Add assertFlags{Set,Unset}Will Thompson1-0/+14
2009-12-07Remove obsolete asynchronousity.Will Thompson1-6/+1
generated by cgit v1.2.3 (git 2.39.1) at 2024-11-23 23:12:04 +0000