Re: WIN32Server don't display the auto-completion window

[Fred Kiefer]

Hi German,

a well documented workaround on this level is acceptable. Although I would of 
course prefer to have a clean solution. I myself don't have a GNUstep 
installation on Windows, there still should be my old virtual machine for this, 
but I haven't used it for a long time. It would be great if you or somebody 
else working on Windows could address the issue, but if you don't have the 
time, feel free to commit a workaround into back.

Fred

On the road

Am 26.02.2015 um 02:05 schrieb Germán Arias <address@hidden>:

> El mar, 24-02-2015 a las 23:07 +0100, Fred Kiefer escribió:
>> In this case the code you should be looking into is in Win32CairoSurface. It 
>> may also be worthewhile to check the Testplant branch, whether there is 
>> already a fix for this issue.
>> 
>> Fred
>> 
>> On the road
> 
> They have a solution for this, but I think you will not like that
> solution. See for example the comment at method -initWithDevice: in
> Win32CairoSurface.m:
> 
> // FIXME:
> // Non-retained backing store type on windows doesn't re-display
> correctly
> // after first time (see additional comments in method -[CairoContext
> flushGraphics])
> // Currently handling such windows as buffered store types until a fix
> can be resolved.
> 
> And the comment in -flushGraphics:
> 
>  // FIXME: Why is this here? When is it called?
>  // Comments added 07/20/2013 to address the above question after
>  // further debugging cairo/MSwindows non-retained backing store type:
>  // This is called from NSWindow/flushWindow for the non-retained
> backing
>  // store case.  This code originally seems to have been added due to
> the 
>  // non-retained backing store type NOT showing up on XWindows/Cairo
> combination
>  // and added this XFlush call to handle this case.
> 
>  // For MSWindows a similar problem seems to occur. However, the only
>  // equivalent to XFlush on MSWindows is GdiFlush, which doesn't cause
>  // the window to re-display properly.  So, on MSWindows, the
> non-retained
>  // cairo backend surface is currently excluded from the build and we
>  // handle it as a buffered window expose event for now.
>  // FIXME:  Anyone know how this can be handled/fixed correctly...
> 
> Unfortunately that branch currently doesn't compile and is the version
> 0.23.1. After some changes I'm able to compile it, but the apps crash
> after start. I suppose is incompatible with current gui.
> 
> So I will make a patch for people who wants use GNUstep on Windows. I
> will use my changes because there are small compared with the solution
> in TestPlant branch.
> 
> Germán
> 
> 
> 
> _______________________________________________
> Gnustep-dev mailing list
> address@hidden
> https://lists.gnu.org/mailman/listinfo/gnustep-dev