[RFC] Enforcing Compositing

[todd rme]

On Sun, Feb 20, 2011 at 4:40 PM, Martin Gräßlin <mgraesslin at kde.org> wrote:
> On Sunday 20 February 2011 22:32:47 Davide Bettio wrote:
>> Hi,
>>
>> On 02/20/11 22:25, Martin Gräßlin wrote:
>> > On Sunday 20 February 2011 22:18:16 Davide Bettio wrote:
>> >> Hi,
>> >>
>> >> On 02/20/11 22:06, Martin Gräßlin wrote:
>> >>>> I can't understand the point about removing enable/disable options.
>> >>>
>> >>> if we don't want to give the user the possibility to disable, we don't
>> >>> need an option to enable/disable.
>> >>
>> >> I don't see anything bad about having enable/disable options on the KCM.
>> >> The only con that I'm aware of is that we have more checkboxes.
>> >
>> > Consider the button would not be there: would you see pro arguments to
>> > add one?
>>
>> Yes, currently at the computer lab in my university for an unknown
>> reason I can't work with compositing enabled, so I had to disable it.
> does it have current drivers or is it completely outdated? I am talking about
> at least Mesa 7.10 which is current at the moment and till 4.7 there will be
> Mesa 7.11.
>>
>> >> The main reason is that buggy software is still around.
>> >
>> > And how do we get the buggy software fixed if we allow workarounds?
>>
>> We can't ask users to wait while the software gets fixed, they need to
>> work so they may have to disable it while in the meantime someone is
>> fixing the bug.
> As long as we don't expose the bugs they won't get fixed. What do you think
> would be the state of the drivers if there would not have been the 4.5
> desaster?

It seems every time someone brings up fixing things in xorg, people
complain about how xorg isn't very responsive to complaints at least
from KDE devs (they seemed to react very quickly when firefox had a
problem).  If xorg developers are not responsive to complaints from
KDE devs, then I don't see how this is going to change anything.

To put it another way, you seem to be portraying as a choice between
people using a working xorg OR relying on a workaround.  But is xorg
really working reliably?  All indications seem to be to the contrary.
And if KDE forces users to use a broken system by removing
workarounds, people are going to blame KDE.

I can sort of understand not implementing new workarounds (although if
KDE won't fix it and xorg won't fix it then all that does is hurt
users).  However, removing existing workarounds does not seem like a
good idea given the history so far.

-Todd