We have basic tests for registration and presence exchange.

Missing test coverage includes:

* presence probes - don't work right, Bug #26141. I consider that bug to include regression tests.
* decloaking and providing contacts' capabilities (as opposed to the gateway's capabilities) - I'm considering this to be part of Bug #26100
* accepting calls from XMPP to SIP - requires Bug #26100; I'm considering this to be part of Bug #26096
* accepting calls from SIP to XMPP - I'm considering this to be part of Bug #26096
* registering n times in one session
* invalid JIDs anywhere
* generally, the client doing anything outside the normal flow (perhaps maliciously)

So, the things left in this bug are basically "if the client does something unexpected, do we cope?" which is necessary for a production deployment, but not screamingly urgent for a working prototype.

I'll consider this bug closed once we have basic regression tests for calls in both directions, which David is working on.

Let's list obvious missing coverage on this bug.

missing coverage in calls (both ways): codecs with nonzero number of channels; codecs with parameters; needing to ignore video codecs

Some other missing coverage etc.:

* should the initiator be the bare or full JID? Currently it's the bare JID and we assert that

* from-xmpp.py doesn't assert that the candidates have the right ports, components

* from-xmpp.py doesn't assert whether the session-accept has any candidates; should it?

* from-xmpp.py doesn't assert that the PTs in session-accept match the intended PTs

Assigning to myself.

(cd tests
python-coverage -e
make check-TESTS PYTHON='python-coverage -x'
python-coverage -a ../telepathy_fargo/*.py)

combined with a bit of git grep suggests that there are quite a lot of dead functions that can probably be deleted.

Feel free to do some stomping/chopping/hacking/slashing.