fullscreen windows, magic values & incomplete backends (was: Re: [Libreoffice-commits] .: vcl/inc vcl/source)

[Thorsten Behrens]

Libreoffice Gerrit user wrote:
> commit 8bf500c365e3cf24086a672f63b0b5d5b60ff4a9
> Author: Lubo?? Lu????k <l.lunak at suse.cz>
> Date:   Mon Nov 26 14:54:34 2012 +0100
> 
>     make presentation spanning all screens work (bnc#758138)
>     
>     This reverts an API "improvement" from 6ee5dfa150d408179e20a5525ff7ec46649e0e58
>     that tried to use -1 as the current screen/display but failed. The check
>     for nDisplayScreen being in range broke all-screens (which is
>     represented as last_screen+1 value, which is broken in its own way, as e.g.
>     the screen count can change). The same way -1 as the default value is broken,
>     as everywhere else invalid invalid screen number means spanning all screens.
>     The extra overloads without a default value should take care of that.
>     
>     Change-Id: Ie118038eacacebc007c25860732d5904ad0de2b9
> 
On having a closer look, I don't think this description is entirely
accurate.  Mac, generic unix, and Windows don't seem to know
anything about "default display", rather they tend to use screen #0
- but your fix for that part looks much saner to me.

For the AllScreens case though, sal backends tend to use any
arbitrary out-of-range index, so probably picking -1 as the magic
value & documenting that in wrkwin.hxx sounds good to me.

What's still completely broken though is gtkframe.cxx - it simply
does not handle the allscreens case at all so far, but here simply
creates a very very tiny borderless window somewhere close to the
origin.

Michael, you reworked GtkSalFrame::SetScreen() - pointers
appreciated how to best get that working there.

Cheers,

-- Thorsten
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: not available
URL: <http://lists.freedesktop.org/archives/libreoffice/attachments/20121128/1aecb2e8/attachment.pgp>