Re: Freeze break: mutter: switching monitors loses workspaces
- From: Frederic Crozat <fred crozat net>
- To: Matthias Clasen <matthias clasen gmail com>
- Cc: Owen Taylor <otaylor redhat com>, alexl redhat com, release-team gnome org
- Subject: Re: Freeze break: mutter: switching monitors loses workspaces
- Date: Sat, 26 Mar 2011 14:01:09 +0100
On Sat, Mar 26, 2011 at 12:51 PM, Matthias Clasen
<matthias clasen gmail com> wrote:
> On Sat, Mar 26, 2011 at 12:24 AM, Owen Taylor <otaylor redhat com> wrote:
>> With Mutter 2.91.92 there are two separate bugs that cause your
>> workspaces to collapse to a single workspace
>>
>> * If the primary monitor (the laptop screen, say) is on the left, then
>> there's a logic bug that causes the workspace assignments to be lost.
>>
>> * If the primary monitor (the laptop screen, say) is on the right, then
>> all the workspaces are lost for a more fundamental reason - the
>> behavior of Metacity/Mutter for a long time is that if you hot
>> plug a new monitor, all your windows stay at the same X/Y positions
>> and hop to the other monitor. Since that's now a secondary monitor
>> outside your workspaces, Mutter sees your workspaces as empty
>> and deletes everything but the first workspace.
>>
>> https://bugzilla.gnome.org/show_bug.cgi?id=645408 has patches that
>> address both of these. I like both patches, and think they are correct
>> and safe, but I'm much more confident in that judgment for the one-line
>> patch that fixes the case of primary monitor on the left than the bigger
>> patch that fixes both cases. So, my instinct is that it makes most sense
>> to apply only the small patch for 3.0.0 and see if we can get sufficient
>> testing for 3.0.1 for the larger patch.
>>
>> I'm however willing to listen to people who think that handling both
>> cases is really important for 3.0.0, especially if they give the patch
>> some extensive testing. :-)
>
> The one-line patch looks obviously correct. +1/2 for that. For the
> more extensive fix, I'll defer to your instincts as the module
> maintainer. If you feel better having it in 3.0.1, so be it...
Same reasoning : r-t approval 2/2
--
Frederic Crozat
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]