[Bug 69885] MC tests don't simulate Contacts interface

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=69885

--- Comment #9 from Simon McVittie <simon.mcvittie at collabora.co.uk> ---
Created attachment 86929
  --> https://bugs.freedesktop.org/attachment.cgi?id=86929&action=edit
09/11] Implement Contacts properly on SimulatedConnection, and  fix tests

(I should --amend this text into a long commit message when I commit this)

The SimulatedConnection used in the regression tests didn't previously list
Contacts in its supported interfaces, even though it implemented at least part
of the interface.

account-manager/avatar-*.py: stop expecting GetKnownAvatarTokens,
since implementing Contacts will mean telepathy-glib uses
GetContactAttributes or possibly GetContactByID.
This actually unmasks a bug - GetContactAttributes is the wrong thing
for MC to be using, because of some subtleties in the spec of
GetContactAttributes and GetKnownAvatarTokens - but that was a bug that
already existed for all Contacts CMs, and I'll address it in a
subsequent commit.

account-manager/irc.py, account-manager/nickname.py: stop expecting
GetAliases, likewise

Because the Contacts interface is more variant-based, this exposed
a bug in dbus-python (<https://bugs.freedesktop.org/show_bug.cgi?id=69967>)
which I worked around.

I didn't add test coverage for CMs *without* Contacts (the sort we were
previously simulating) because any such CM is considered to be unusably
broken these days; the interface has been mandatory since 2009!

---

I'll take a positive review of this as also being a positive review for "copy
the extra constants into Gabble".

-- 
You are receiving this mail because:
You are the QA Contact for the bug.