Infinite _pulse_helper callback loop in PackageStatusBar when adding GObject.timeout_add(500, self._pulse_helper)
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
software-center (Ubuntu) |
Won't Fix
|
Medium
|
Unassigned |
Bug Description
There is an infinite loop in the _pulse_helper method, caused by the sentence of GObject.
More details come from this chat with Michael:
(10:03:49 AM) nessita: mvo: the main issue is that self._pulse_helper never returns False
(10:04:02 AM) nessita: so the callback will live in the mainloop for ever
(10:04:58 AM) nessita: mvo: in the tests, this causes the "do_events" helper to process _pulse_helper lots of time... so each test takes longer and longer
(10:05:44 AM) mvo: nessita: ohhhh, good catch!
(10:06:26 AM) mvo: nessita: so we need fix the pulse helper to only start when needed and to actually stop or we disable it for the tests
(10:06:49 AM) nessita: mvo: right, though I wonder if this does not impact real usage of the USC
(10:07:03 AM) nessita: mvo: I don't think that callback is never disabled, no? (even in IRL)
(10:08:02 AM) mvo: nessita: unfortunately hanging around there as a sort of background thread :/
(10:08:23 AM) nessita: mvo: want me to fill a bug about this?
(10:09:19 AM) mvo: nessita: yes please with the background info that it causes the test issues you mentioned please
Changed in software-center (Ubuntu): | |
status: | Triaged → Won't Fix |
Setting as triaged as this is a development issue reported by a USC dev. Thanks!