[Bug 28866] Allow channel requesters to associate metadata with requests for the benefit of other clients

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=28866

--- Comment #14 from Simon McVittie <simon.mcvittie at collabora.co.uk> 2010-09-03 09:29:19 PDT ---
I've stolen this branch:

http://git.collabora.co.uk/?p=user/smcv/telepathy-spec-smcv.git;a=shortlog;h=refs/heads/request-channel-25018
http://people.freedesktop.org/~smcv/telepathy-spec-request_channel_25018/spec/

The *implementation* of this branch turns out to be quite a can of worms (not
Guillaume's fault, it touches MC code that contains lurking horrors) so I think
we need to get it merged relatively soon (before the spec is undrafted) and
refactor until it actually makes sense.

Accordingly, I've moved the new API to .FUTURE pseudo-interfaces, so we can do
the refactoring without needing to compile MC against a patched telepathy-glib.
I opened Bug #30000 (eek, 30k Freedesktop bugs) for the MC implementation.

I've also made some other misc changes, for which see my git repository.

------------------------------------

Insta-stable changes
====================

Remaining changes to stable API in my version of the branch:

* Client.I.Requests: Hints SHOULD be included among the ChannelRequest's
immutable properties. Uncontroversial, because every immutable property SHOULD
be included anyway :-)

* request-properties added to ObserveChannels' dict of misc. Uncontroversial, I
think.

* Likewise for HandleChannels' dict of misc.

I don't think any of those should block merging this.

------------------------------------

To-do list
==========

Should SucceededWithChannel have the immutable properties of the channel? I
think it probably should, but I haven't added this yet.

Should SucceededWithChannel be plural? In principle, a ChannelRequest could
operate on a hypothetical plural request, and return many channels... although
plural requests are a tower of cans, each containing worms.

Should SucceededWithChannel give you properties of the Connection? Probably
not, IMO: connections don't have many (useful) properties that are actually
immutable!

Should we explicitly say that the CD will never interpret Hints? I think we
should, and I've added a commit that says so.

While we're adding new ways to request channels, should we add an a{sv}
placeholder for extra arguments that the channel dispatcher *does* interpret?
(Obviously, this is unnecessary if you convince me that my previous assertion
was wrong.)

-- 
Configure bugmail: https://bugs.freedesktop.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.