On 23.10.2015 22:02, Rex Dieter wrote:
> Sandro Mani wrote:
>> On 23.10.2015 15:38, Rex Dieter wrote:
>>> Sandro Mani wrote:
>>>> On 20.10.2015 13:31, Rex Dieter wrote:
>>>>> Multiple screens is (still) a bit of an unsolved problem in Qt, not
>>>>> related to driver.
>>>> This patch is looking promising . Been running with it for a couple
>>>> of days and have yet to experience a crash. Remaining issues I
>>>> experience after screen setup change are krunner (when not crashing
>>>> due to baloo) occasionally moving around as you type, and plasmashell
>>>> not removing the screen space it occupies from the available area
>>>> (which is to say: it will cover a portion of maximized windows).
>>>>  https://codereview.qt-project.org/#/c/138201/
>>> Let me know when/if you think that's worth pulling into our builds for
>>> broader testing.
>> There have been some comments on the code, but mostly aesthetic stuff.
>> Depending on how fast things move in codereview, one could also start
>> adding the current patch: the rebased version I'm using is here .
>>  https://smani.fedorapeople.org/138201.patch
> One more question, we currently carry,
> would you recommend keeping that in addition to 138201.patch (or dropping
I'm currently still carrying it, though if I read the proposed 138201
patch correctly, it should avoid that at any moment during screen
configuration change no screens are available, so the first part of
qtbase-opensource-src-5.5-disconnect_displays.patch should not be needed
anymore (though not sure what happens when you start an application
without any screens attached to the system and add screens later on). I
don't know whether the second part of the patch is still needed. But
neither change is in direct conflict with the proposed 138201 patch, so
it should not harm to keep it around until 5.6.
OK, pulled into qt5-qtbase-5.1.1-3
* Sat Oct 24 2015 Rex Dieter <rdieter(a)fedoraproject.org> 5.5.1-3
- pull in more screen connect/disconnect fixes (code review 138201)