- User Since
- Feb 25 2013, 1:50 AM (447 w, 2 d)
Jun 6 2017
ooops... fixing the wrong ticket
It seems wrong. It shouldn't set the body geometry twice.
I'll provide a patch. My apologies.
Mar 24 2017
Mar 23 2017
Feb 2 2017
Thanks for reporting it, @Jaehyun_Cho
I'll take a look on that soon
Dec 22 2016
Closed by commit 8a42514d71195ca28028346a338e4db2b2f4faf9
Dec 21 2016
Make changes suggested by @iscaro
answering missed inline question
Dec 20 2016
Inline comments replied
Dec 19 2016
OK, name changed to 'custom' instead of 'system' and note added to documentation.
Dec 15 2016
Just tried to reproduce it here, and no success.
It's working as expected.
Dec 14 2016
Pretty much all required changes were done.
(see inline replies)
Dec 13 2016
Dec 12 2016
I'm using it on Edje and it works fine
Dec 6 2016
Include changes discussed on RFC and review
Dec 5 2016
Dec 2 2016
Closed by commit 5fdc93c02e0743c8750cd1a03d2f3a6b4df61278
Nov 30 2016
We're getting closer =D
But there is still an issue that you can easily reproduce using evas_multiseat_events example.
This PR was merged already,
on commit fc5cc218095dfc3cc772c7ab664358a66ef64379
pointer,in and pointer,out logic seems to be wrong.
Please test the following scenario with the recently added evas_multiseat_events example (using two seats with pointers):
Nov 28 2016
Not sure if related to the inline comment, but the pointer in / out logic seems buggy at the moment.
I've done a few tests here with two seats with mouses going over a few different evas objects and changing focus between them on click.
At some points when clicking on other object with a seat generates a pointer,out for the other seat.
I'll try to post a comprehensible log later.
Nov 24 2016
Nov 18 2016
No reasons for apologies.
Even if it wasn't caused by recent changes on focus, we as community need to fix the issue you're facing anyway.
Not being able to reproduce the issue makes it a bit harder to debug, but with your help we can do it =)
I tried to reproduce it running it on X11 on Ubuntu.
Nov 16 2016
Code seems better now.
Handle all unref in a single callback
It seems good overall. Just a few inline comments.
Indeed I failed to preview such a scenario.
Nov 9 2016
Nov 8 2016
Closed by commit 2a96097a658e0817c95f72b56d37879720886fa5
Nov 7 2016
Hey Iscaro, I've got a few questions / comments related to this series that eventually would require more changes.
Nov 4 2016
Nov 1 2016
Closed by commit rEFL3ac1812a1f99
Oct 31 2016
Seems ok to me.
Also it fixes issues I've been facing on multi seat wayland support branch.
Oct 10 2016
Sep 30 2016
Actually we found some issues with the second patch, that shouldn't be creating new buffers everytime w / h diverges from ee->w, ee->h.
@iscaro is going to push a new version soon.
Sep 29 2016
Applied on commit 50f5704cc6b95ed30b25c668e2fc1679a4131279
It seems correct.
Just adding Dave since he spotted the issue initially.
Sep 28 2016
This approach sounds right to me.
I've tested it using E, terminology, ephoto. No more deadlocks.
Ref/unref seems to be the way to avoid memory duplication.
Sep 26 2016
Merged on commit 0566abeee85f899dffd1b58260ada696d8a86af7
Sep 20 2016
Made all required changes on previous revision
Sep 16 2016
Long time I don't code something on top of ConnMan API (last time was a connman module about 3 years ago? =) ), but it seems really good overall.
Sep 9 2016
I believe raster merged something fixing it already.
Sep 8 2016
I believe you're missing an environment variable, USER_SESSION_DIR
Have you tried to do something like that?