[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kwin
Subject:    Re: new info about KWin crashes with XRandR 1.2
From:       "Lucas Murray" <lmurray () undefinedfire ! com>
Date:       2009-01-12 16:32:33
Message-ID: f09827650901120832s6eea33f7gd95c8c8e3c65753f () mail ! gmail ! com
[Download RAW message or body]

On Tue, Jan 13, 2009 at 1:14 AM, Andreas Pakulat <apaku@gmx.de> wrote:
> On 12.01.09 16:16:05, Lubos Lunak wrote:
>> On Monday 12 of January 2009, Lubos Lunak wrote:
>> > On Sunday 11 of January 2009, Guillaume Pothier wrote:
>> > > On Sun, Jan 11, 2009 at 1:11 PM, Andreas Pakulat <apaku@gmx.de> wrote:
>> > > > Heh, let me know if you'd like somebody else to test that. I'm sure
>> > > > this is not quite an uncommon case as IIRC thats the default behaviour
>> > > > of Xorg when adding a second head without a pre-set xorg.conf.
>> > >
>> > > I got that fixed by only reloading modes in RandRScreen::handleEvent
>> > > instead of reloading all settings.
>> > > Here is the newest patch if you want to test it.
>> >
>> >  I'm fine with the KWin changes for 4.2 (i.e. feel free to commit). For
>> > trunk, however, I consider them to be workarounds.
>>
>>  The clientArea() crash from bug #173017 is actually not Kephal related.
>> Screen gets added using xrandr12, Workspace::desktopResized() is called, it
>> calls Workspace::updateClientArea(). Which computes gets new screen sizes,
>> and also workspace sizes as screen sizes minus space takes by struts set by
>> windows. Getting struts results in calling Client::adjustedClientArea(),
>> which calls Workspace::clientArea() to get the screen size, but
>> Workspace::updateClientArea has not updated internal variables yet ->
>> out-of-bounds, crash.
>
> Thats not what happened here for me. As I said in one of my first mails the
> problem was that the screen was always -1, i.e. both activeScreen as well
> as the one calculated by kephal for the cursor position. Thats why it
> crashed with out-of-bounds.
>
> In general that bugreport has so many mixed problems that it should be
> closed nonetheless and users should report their bugs again after trying
> with 4.2. Or somebody from the kwin team goes through the comments and
> duplicate-additions and re-extracts the separate reports.

Yeah, I screwed things up a little by mixing them all together. Unless
someone (Lubos?) would like to go through them all again to see which
ones are not actually duplicates I think the best course of action
would be to close them all and start fresh to see which ones were
fixed and which ones were not.

> Andreas
>
> --
> Your goose is cooked.
> (Your current chick is burned up too!)
> _______________________________________________
> kwin mailing list
> kwin@kde.org
> https://mail.kde.org/mailman/listinfo/kwin
>
_______________________________________________
kwin mailing list
kwin@kde.org
https://mail.kde.org/mailman/listinfo/kwin
[prev in list] [next in list] [prev in thread] [next in thread]