[00:25] Trevinho: I take it invalidate_for_new_frame isn't called by anything? [00:26] RAOF: it's called by gtkwindow as it's used as function overrider in gtkmirwindowimpl [00:26] Ah, ok. [00:27] RAOF: so at the end in gdkwindow.c at line 3653 [00:27] Why does it bail early if not using GL? You still need to account for previous damage (unless you always do full-window repaints when not using GL?) [00:28] RAOF: not sure I undestood what you mean... [00:29] ah, the function... [00:29] https://github.com/3v1n0/gtk/blob/wip/mir-gdkgl/gdk/mir/gdkmirglcontext.c#L63 [00:29] well, when not using gl it shouldn't be called at all, in that case it uses a 2d surface [00:30] RAOF: basically this function works only when a gdkgl area is used, which happens only when using a gtkglarea, not for all the rest of the widgets [00:32] Fair chop. === chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun [07:43] Wow, src/server/ is 317 files now. No wonder I'm still getting lost [07:46] why so many? [08:14] mlankhorst: Mir does more than you think [08:14] Just not all that you want (yet)\ === alf is now known as alf_ [09:25] alan_g: Any idea why C++ says you can't get to namespace local things in the parameter list, but can in the function body? [09:26] duflu: do you means something different to what I understand by "namespace local things"? [09:27] alan_g: I mean a function parameter list must use full namespace spec, and can't implicitly get to the namespace the function is in [09:27] I guess parameter list is different scope to the body [09:27] I don't believe that is true. [09:27] Give an example [09:28] alan_g: mir::shell::Something::doit(shell::Type x) [09:29] Like mir::logging::Logger::log(Severity, ...)? [09:29] That works [09:30] alan_g: Maybe it's a different issue. No big deal. [09:32] duflu: yeah, I would really like to have support for multiple keyboard/etc [09:33] mlankhorst: If that's your only complaint then we're doing well [09:33] mlankhorst: Find the egl examples? [09:33] We have a few [09:33] well, no [09:33] window management is lacking, makes rootless hard [09:34] mlankhorst: Working on that right this moment [09:34] mlankhorst: You know how to use mir_server_demo_shell? [09:34] no, not at all [09:35] mlankhorst: Very primitive WM: http://unity.ubuntu.com/mir/demo_shell_controls.html [09:35] don't even know the differences between all those mir servers [09:36] oh and support for synaptics would b enice [09:36] working on it [09:36] mlankhorst: I think anpok is working on libinput now [09:37] and the underlying libevdev is causing some trouble with umockdev [09:37] ah [09:37] what about support for dumping keymappings so I can tell X about them? [09:58] alan_g: I didn't realise yesterday that "Shell" is a class already :) [09:58] alan_g: So I'm side-stepping that and making a "WindowManager" to avoid changing too much at once [09:59] I don't know what your trying to solve. I thought I explained why there wasn't a problem yesterday [09:59] alan_g: I'm avoiding large changes suggested yesterday [10:00] These are less invasive [10:00] Not final, but smaller steps === ara_ is now known as ara [10:23] who knows more about the lifetime management of the buffers with mesa? [10:26] mlankhorst: A client is guaranteed to own any buffer the server gives it till the next swap [10:26] That's possibly all you need to know in the client [10:27] duflu: yes, but what happens if I cache the fd's, should that work? [10:28] mlankhorst: Possibly not. The server might resize the surface once it gets it back. And I think that could mean the old one is gone [10:28] Resize means delete buffers when it's safe and replace them with new ones of the new size [10:28] does it reset the buffer age in that case? [10:29] mlankhorst: Don't know. It _should_ I guess but unsure [10:29] mm *checks how xorg-mir does it* [10:31] ok it just does the whole dance over and over [10:47] duflu: but is it really a good idea that the fastpath is gbm_bo_import, eglCreateImageKHR, glGenTextures, glEGLImageTargetTexture2DOES, glDeleteTextures, eglDestroyImageKHR, gbm_bo_destroy ? [10:48] mlankhorst: *shrug* [10:55] bleh [10:57] duflu: what about the drm fd, is that a new one only used by the device? [10:58] mlankhorst: I have lost context, sorry. Would have to dig through source to answer anything like that [10:58] oh oops [10:59] *checks if gbm closes the drm fd or not* [11:08] even dup'ing seems to be affected, great [11:24] Woo, working on-the-fly WM policy [11:24] Just needs a few days to make it nice === chihchun is now known as chihchun_afk === dandrader is now known as dandrader|afk === dandrader|afk is now known as dandrader [13:59] * alan_g wonders if he's the only one to find the log messages now coming from the tests a PITA [14:11] alan_g, I do too... was thinking of logging a bug :) [14:19] https://bugs.launchpad.net/mir/+bug/1394221 [14:19] Launchpad bug 1394221 in Mir "acceptance_tests are too chatty" [Medium,New] [14:26] camako: the AnonymousShmFile tests should pass on kernel 3.13 shouldn't they? https://code.launchpad.net/~alan-griffiths/mir/migrate-more-acceptances-tests/+merge/242094/comments/596315 [14:45] ugh found my issue, I had to RTFM [14:45] "If the EGL rendering context context is not current to any thread, eglDestroyContext destroys it immediately. Otherwise, context is destroyed when it becomes not current to any [14:45] thread." [14:47] x11perf works again now, and I no longer get those weird issues I was having.. [14:53] the basic functionality is now stable, on towards handling resizing events I guess. :P [15:05] alan_g, yes they should [15:06] camako: thanks [15:21] mlankhorst: cool === chihchun_afk is now known as chihchun === dandrader is now known as dandrader|lunch === dandrader|lunch is now known as dandrader [17:58] hi, is there any progress on https://bugs.launchpad.net/mir/+bug/1391976? i can't seem to get unity8 running from trunk because of it [17:58] Launchpad bug 1391976 in Mir "Loading libmircommon.so twice leads to a segfault in protobuf code" [High,Confirmed] === alan_g is now known as alan_g|EOD === dandrader is now known as dandrader|afk === dandrader_ is now known as dandrader