[02:21] <tjaalton> ari-tczew: eh?
[02:22] <tjaalton> ari-tczew: no, AIUI it's still needed
[02:23] <tjaalton> ari-tczew: hum, so you mean *kernel-common
[02:25] <tjaalton> n-k-c is probably unnecessary, yes
[02:34] <ari-tczew> tjaalton: are you a core developer right?
[02:34] <ari-tczew> could you comment on bug like ACK? then subscribe ubuntu-archive. :)
[02:37]  * ari-tczew is off to bed.
[18:42] <ricotz> bryceh, hi
[18:42] <ricotz> bryceh, i noticed all the build failure mails for the wayland testing ppa
[18:44] <ricotz> bryceh, you need a newer libdrm the api for nouveau changed a bit
[18:49] <bjsnider> the ones that just failed did so because wayland-server-protocol.h is missing
[18:50] <ricotz> bjsnider, nouveau_context.c:132
[18:51] <ricotz> i meant the mesa build
[18:52] <bjsnider> wayland itself failed too
[18:53] <ricotz> yeah, didnt look at this
[18:57] <bryceh> yeah I'm poking at it now
[18:57] <bryceh> looks like a path issue; wayland-server-protocol.h is a generated file but seems to be generated ok; just in a different dir than compositor is looking at
[18:58] <bryceh> ricotz, oh new libdrm for mesa?  mmm ok I'll try that.  know what version in particular?
[18:58] <bryceh> e.g. is the one in xorg-edgers enough?
[18:58] <ricotz> bryceh, you can copy it from the edgers ppa
[18:58] <ricotz> yes
[18:59] <bryceh> ok, does it have any dependencies that need copied over too?
[18:59] <ricotz> bryceh, the error is a bit hidden because of the multithreaded build
[18:59] <ricotz> i dont think so
[18:59] <ricotz> for natty it should work
[18:59] <bryceh> thanks for figuring it out, getting all the bits and pieces sorted is sort of a hair pulling exercise
[19:00] <ricotz> bryceh, yeah, thanks to mark ;-)
[19:05] <bryceh> wish the buildd's built stuff faster
[19:07] <ricotz> bryceh, build is quite fast, but the publishing takes some time
[19:08] <bryceh> yeah
[19:08] <bryceh> libxkbcommon built in <10 min but took 2 hours to publish
[19:08] <bryceh> ricotz, do you know what makes publishing take so long?
[19:09] <ricotz> bryceh, why are you build testing it locally?
[19:09] <ricotz> i dont know sometimes it is fast or sometimes it takes hours ;), might depend on the server load
[19:15] <bryceh> hrm... so the wayland-server-protocol.h issue is tricky
[19:16] <bryceh> since it's a generated file, debian/rules copies it into obj-i686-linux-gnu, not the build tree
[19:16] <bryceh> er not the source tree
[19:16] <bryceh> but the #include looks only in the source tree and so doesn't find it.
[19:17] <ricotz> bryceh, did you generated the tarball with "make distcheck"?
[19:18] <bryceh> ricotz, no, just tarballed up the git tree
[19:19] <bryceh> besides....  $ make distcheck
[19:19] <bryceh> make: *** No rule to make target `distcheck'.  Stop.
[19:19] <ricotz> if wayland has proper autofoo then it is better to create a tarball with that
[19:20] <ricotz> bryceh, ./autogen.sh --prefix=/usr ... , make, make distcheck
[19:22] <bryceh> ricotz, no, that's not generally how we build xorg-edgers packages
[19:23] <ricotz> bryceh, i know, but in my opinion it would be the proper way, and might solve thing like autogenerated files
[19:24] <bjsnider> ricotz, what does that option do?
[19:25] <bryceh> ricotz, no my preference is to put any special package logic into the rules file; else it adds extra steps to whomever packages this in the future
[19:25] <bryceh> ideally I want to make the process compatible with how xorg-edgers packages get built, so the xorg-edges scripts can be used, so we can easily get daily wayland package builds
[19:26] <ricotz> bryceh, the official releases of software will be made with "make dist"!
[19:27] <ricotz> the git tarball snapshot arent the real releases
[19:27] <bryceh> mm hmm
[19:31] <ricotz> bryceh, ok, all they files are created properly and are included in the tarball using make dist
[19:31] <ricotz> just checked it with wayland
[19:32] <ricotz> s/they/the/
[20:21] <ricotz> bryceh, mesa built but still packaging problems :(, could you also upload the xkbcommon,x11proto-kb and cairo to natty pocket?