FYI: The next thing I was going to try was just cramming everything into the 
one available window level.  Since we are fullscreen and letting X11 manage 
window stacking, this *should* work.

On Dec 10, 2010, at 09:22, Jeremy Huddleston wrote:

> 
> On Dec 10, 2010, at 07:13, doh123 wrote:
> 
>> I know the display capturing issues are complicated...  I'm going to spend 
>> some time on my Christmas vacation trying to dig into Xquartz to learn how 
>> it works some more, and see if I can get it working... but no promises, I'm 
>> far from an expert, or experienced programmer.  If I find a way to get it 
>> working that isn't a hack that just lets it work right for my usage (which 
>> would probably be a miracle) I'll definitely share it.  This is just like 
>> the number 1 thing holding me back from a Wineskin 2.0 release, which I know 
>> isn't your issue... I don't envy you really, as xquartz seemed very 
>> complicated in general to me :-)  Had no idea it needed 5 window levels, I 
>> assumed it was all done on one, and the guard could be put up, then dropped 
>> back a single level and the rest done on the highest.
>> 
>> Thanks though, you do a great job
> 
> Thanks.  Hopefully I can provide some help if you get stuck.  I probably 
> won't be on IRC much, but you can email.  Below you'll find a patch which was 
> my starting point when looking at this last.  It just determines if we should 
> be captured or released and does the capture/release.
> 
> hw/xquartz/xpr/xpr.h has our window levels:
> 
> #define AppleWMWindowLevelNormal        0
> #define AppleWMWindowLevelFloating      1
> #define AppleWMWindowLevelTornOff       2
> #define AppleWMWindowLevelDock          3
> #define AppleWMWindowLevelDesktop       4
> #define AppleWMNumWindowLevels          5
> 
> static const int normal_window_levels[AppleWMNumWindowLevels+1] = {
> 0, 3, 4, 5, INT_MIN + 30, INT_MIN + 29,
> };
> static const int rooted_window_levels[AppleWMNumWindowLevels+1] = {
> 20, 21, 22, 23, 19, 18,
> };
> 
> Our levels are different depending on rootless vs fullscreen mode and are 
> chosen based on the menu bar's level (24), and the levels for normal (0) and 
> floating (3) windows.
> 
> Note that we actually use *6* levels rather than 5.  The 6th one is for the 
> root window.  see the comment in xpr/xprAppleWM.c:
>    // AppleWMNumWindowLevels is allowed, but is only set by the server
>    // for the root window.
> 
> <capture.patch>
> 
>> On Dec 10, 2010, at 2:30 AM, Jeremy Huddleston wrote:
>> 
>>> 
>>> On Dec 9, 2010, at 23:22, doh123 wrote:
>>> 
>>>> Multi Monitor fix is nice...
>>> 
>>> It just fixes the mode-change regression... it doesn't add support for the 
>>> extended resolution changes.
>>> 
>>>> does it actually use display capturing?  
>>> 
>>> No.  Display capturing doesn't quite work right for us.  The guard window 
>>> is at the highest level - 1, but we need 5 window levels ... I tried 
>>> updating the window level, hiding the guard window, and other tricks with 
>>> no success.  I haven't given up, but it is not as simple as "just capture 
>>> the display"
>>> 
>>> 
>>>> I'll try it out in the next couple of days hopefully
>>>> 
>>>> On Dec 10, 2010, at 12:28 AM, Jeremy Huddleston wrote:
>>>> 
>>>>> I just released the first release candidate of XQuartz 2.6.0.  This is 
>>>>> the first pre-release built for Leopard as well as SL, so please give it 
>>>>> a whirl on Leopard to see if there are any Leopard specific issues we may 
>>>>> have missed.
>>>>> 
>>>>> There are still a couple bugs left that I'd like to address before 2.6.0 
>>>>> is released, but they're not worth holding up release if I don't get to 
>>>>> them, so consider this mostly complete.
>>>>> 
>>>>> The changes worth noting from 2.6.0_beta2 are:
>>>>> 
>>>>> Updated freetype which should address font rendering issues introduced in 
>>>>> 2.5.2.
>>>>> The mac OSX screensaver is now disabled while in fullscreen mode.
>>>>> Mode changing works again for multi-monitor configurations.
>>>>> XBell() is now exclusively NSBeep() (ie: the "use system alert" 
>>>>> preference option is now gone)
>>>>> Misc stability fixes in the server
>>>>> libX11-1.4.0 and misc other modules
>>>>> 
>>>>> For the full list of changes, see the draft release notes:
>>>>> http://xquartz.macosforge.org/trac/wiki/X112.6.0
>>>>> 
>>>>> You can install it using automatic updates from a previous beta (or full 
>>>>> version if you enable the beta feed) or directly:
>>>>> Leopard: 
>>>>> http://static.macosforge.org/xquartz/downloads/Leopard/X11-2.6.0_rc1.dmg
>>>>> Snow Leopard: 
>>>>> http://static.macosforge.org/xquartz/downloads/SL/XQuartz-2.6.0_rc1.dmg
>>>>> 
>>>>> At this point, I'm just waiting on the final X11R7.6 release at which 
>>>>> point I'll pull in the rest of the updated modules and call it 2.6.0.
>>>>> 
>>>>> Please test and let me know if there are any regressions.
>>>>> 
>>>>> Thanks,
>>>>> Jeremy
>>>>> 
>>>>> 
>>>>> _______________________________________________
>>>>> Xquartz-dev mailing list
>>>>> [email protected]
>>>>> http://lists.macosforge.org/mailman/listinfo.cgi/xquartz-dev
>>>>> 
>>>> 
>>>> 
>>>> _______________________________________________
>>>> Xquartz-dev mailing list
>>>> [email protected]
>>>> http://lists.macosforge.org/mailman/listinfo.cgi/xquartz-dev
>>> 
>>> 
>>> _______________________________________________
>>> Xquartz-dev mailing list
>>> [email protected]
>>> http://lists.macosforge.org/mailman/listinfo.cgi/xquartz-dev
>>> 
>> 
>> 
>> _______________________________________________
>> Xquartz-dev mailing list
>> [email protected]
>> http://lists.macosforge.org/mailman/listinfo.cgi/xquartz-dev
> 
> 
> _______________________________________________
> Xquartz-dev mailing list
> [email protected]
> http://lists.macosforge.org/mailman/listinfo.cgi/xquartz-dev


_______________________________________________
Xquartz-dev mailing list
[email protected]
http://lists.macosforge.org/mailman/listinfo.cgi/xquartz-dev

Reply via email to