plasmashell never emits setStage('desktop');

Marco Martin notmart at gmail.com
Sat Sep 13 08:07:09 UTC 2014


On Friday, September 12, 2014, Àlex Fiestas <afiestas at kde.org> wrote:
>
> Looking further in the problem I have been able to check that one of my
> containments is never ready. I tried to debug this myself but Containment
> seems a really complex piece of code, after 2h I have to give up :/

Thanks for looking in the problem, I noticed the same thing during akademy
and had the same diagnosis, but didn't have time to look deeply into it.
I'll look into it/add test later today.

> Could somebody with knowledge please look why Containments are never
ready?
> Also this code could  use some refactoring... It is really really really
> complex.

Not really, at least not more than it needs to be: is pretty simple: it
keeps track of the applets that had the startupcompleted constraint called
on them and on itself, emits the signal when everybody is.
Signaling that the qml is ready and keeping track centrally that all
containments everywhere, and all their applets have indeed their qml ready
as well, has to pass trough quite long hoops, not much to be done around
that.

What it should be done instead, is adding tests: right now there is one
that should catch it, coronatest. Apparently doesn't but maybe the code
coverage thing will help.

When this hard to pinpoint bugs happen, before fiddling with the code to
make them go away, we should try to introduce a new test that reproduces
the problem exactly, at least should be possible on the library code, while
not always easy on the more ui parts.

--
Marco Martin
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/plasma-devel/attachments/20140913/4af77796/attachment.html>


More information about the Plasma-devel mailing list