[12:03] <ricotz> mlankhorst, hi :), are there (precise) packages for wine 1.6.1 yet?
[16:37] <mlankhorst> ricotz: erm scott ritchie does those
[16:45] <ricotz> mlankhorst, ok, thanks, since you are involved too, i was hoping you had uploaded it somewhere already -- haven't found it on launchpad though
[20:31] <ali1234> i've got a question about RetainPermanent and the various atoms used to signal the existence of such on the root window pixmap
[20:33] <ali1234> first: there's a bunch of different ones: XSETROOT, XROOTPMAP, ESETROOT_PIXMAP_ID, _XROOTPMAP_ID... any more?
[20:34] <ali1234> second: how do these pixmaps get freed? if i set and change the background image using one of these tools, it's going to keep allocating new RetainPermanent pixmaps. how are they supposed to get deallocated?
[20:36] <ali1234> RAOF: in my searching on this topic i saw you discussing RetainPermanent in relation to unity-greeter. can you tell me anything about this stuff? i'm hacking on lightdm-gtk-greeter for xubuntu
[21:01] <ali1234> it looks like unity-greeter defines RetainPermanent but doesn't use it. so how does the wallpaper handoff happen between unity-greeter -> compiz?
[21:02] <mlankhorst> ali1234: no idea, magic? :>
[21:02] <ali1234> well all this RetainPermanent stuff sure feels like magic
[21:03] <mlankhorst> I guess until compiz starts compositing the old image will exist
[21:03] <mlankhorst> because nothing new has been drawn to it yet
[21:04] <ali1234> the problem we have is xfwm starts compositing with a empty grey buffer, and there's a delay before the app that draws the wallpaper starts up
[21:04] <ali1234> xfwm can read the RetainPermanent background at startup to avoid this, but the whole thing seems incredibly bad with magic atoms and unmanaged buffers...
[21:17] <mlankhorst> I think you can just read out the uncomposited buffer before starting to composite
[21:17] <mlankhorst> or at least delay drawing until you have a background
[21:18] <mlankhorst> but I haven't put much thought since I mstly worry about driver issues ;)
[21:19] <ali1234> speaking of driver issues, why does nouveau corrupt the hell out of the display when using RetainPermanent pixmaps?
[21:20] <ali1234> "oh, you wanted your wallpaper? here's some windows that were on your screen earlier today"
[21:22] <ali1234> i think "read out the uncomposited buffer" is what compiz does, that's probably going to be our best option
[21:22] <mlankhorst> I guess it uses random vram memory with nothing drawn to it yet
[21:22] <mlankhorst> which happens to match window contents in some cases