This is an automatically generated e-mail. To reply, visit: http://git.revie= wboard.kde.org/r/100893/ |
On March 20th, 2011, 4:31 p.m., Martin Gr= =C3=A4=C3=9Flin wrote:
Very nice= functionality, I like it. Probably we should extend it to * announce support in _NET_SUPPORTED * add functionality to netwm/KWindowSystem * bring it to NETWM spec (though I doubt we have success given that GNOME S= hell and Unity require compositing all the time)
Afaik unity= will explicitly NOT require compositing (at all) - therefore the move towa= rds Qt. However i'm not too optimistic that it will be accepted (outside KDE) e= ither. We might have to think about a (partial) logic inversion. In that case, every fullscreen client would block compositing unless explic= itly permitting it. So _KDE_NET_WM_BLOCK_COMPOSITING would turn into _KDE_NET_WM_WANT_COMPOSITI= NG 0 means "no", everything else "yes" and an absent prope= rty triggers the default behavior (not fs -> compositing good, fs -> = compositing bad) This way we'd catch "legal" fs clients (games, video players)= out of the box, nasty (tm ;-) gkt+ applications (office/browser) look &quo= t;ugly" in fs while fs prone KDE applications (<strike>koffice&l= t;/strike> calligra, konqueror/rekonq i assume? - whatever) could keep c= ompositing active despite the fs mode. I'd rather dislike this way, but ultimately it's the result on the = user desktop that matters... :-\ =3D> I'll branch off the netwm stuff, we try and on failure can stil= l think about the more aggressive push.
On March 20th, 2011, 4:31 p.m., Martin Gr= =C3=A4=C3=9Flin wrote:
=
kwin/client.cpp (Diff revision 1) = = void Client::updateCursor()2114 void Client::updateComposit= eBlocking(boolreadProperty)2115 {
2116 const bool usedToBlock = =3D blocks_compositing;= span>2117 if (readProperty)= {Why the b= ool readProperty? If I read the code correctly it's a no-op if readProp= erty is false
Code body. If we don't (find a way to) do it state dependent (fs= mode) it's completely superfluous. Otherwise just lacks "else blocks_compositing =3D rules()->checkBlo= ckCompositing(blocks_compositing);"
- Thomas
On March 20th, 2011, 3:11 p.m., Thomas L=C3=BCbking wrote:
Review request for kwin.
By Thomas L=C3=BCbking.
Updated March 20, 2011, 3:11 p.m. Descripti= on
Testing <= /h1>
Diffs=
|