Under some circumstances (heavy traffic, slow/dodgy network, etc) disco
requests from gabble time out. Not a big issues as it prevents the disco
pipeline from stalling. The downside is that gabble thinks the reply that
comes in later is an unknown iq and drops it, while it should use it to
learn which features correspond to a certain hash and use that to update the
The point of requests timing out is so that we don't wait indefinitely for
someone's caps to arrive, and so that we can send the disco request to another
contact who might reply sooner. But there are several different levels of
timing out that we could split up something like this:
» after l seconds, see if we can send the same disco request to a different
contact (who might answer sooner);
» after m seconds, if we're waiting on these caps to start a call, give up;
» after n seconds, stop waiting for the reply (to keep memory usage bounded).
This would be similar to how vCard requests were recently made to time out more
-- GitLab Migration Automatic Message --
This bug has been migrated to freedesktop.org's GitLab instance and has been closed from further activity.
You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.freedesktop.org/telepathy/telepathy-gabble/issues/31.