[00:00] <Q-FUNK> nope.  none of them appear yet, even in source
[00:06] <bryce> Q-FUNK: what exactly needs to be changed?
[00:07] <bryce> Q-FUNK: do you have a debdiff?
[00:08] <Q-FUNK> I just wanted to double-check whether what was uploaded tonight for all 3 packages was indeed a match for the known-good solution that is already in intrepid
[00:08] <Q-FUNK> too many debdiffs attached to that bug and somewhat confusing discussions for everyone
[00:10] <bryce> yeah, I'm really confused at this point
[00:11] <Q-FUNK> understandable
[00:11] <Q-FUNK> this is why I wanna triple-check everything now, so that we don't miss it this time.
[00:12] <Q-FUNK> it's already nice enough of steve to allow me to ping him back in the morning if I notice anything, before 8.04.1 is done.
[00:12] <Q-FUNK> but that involves me checking the chnages that were commited and testing all 3 packages
[00:13] <Q-FUNK> ogra's also asked various people on #ltsp, some of whom might have old GX1 hardware, to test this ASAP
[00:13] <bryce> Q-FUNK: new -nsc just now uploaded to hardy-proposed
[00:14] <Q-FUNK> thanks!
[00:14] <Q-FUNK> can we get the new -all too, since steve already approved it?
[00:14] <Q-FUNK> I think that -geode was the only one where he'd rather have me finetune the changelog, before we'd move ahead.
[00:14] <bryce> Q-FUNK: what exactly needs to be done? 
[00:15] <Q-FUNK> for -all?  small diff:  make it depend on -geode instead of -amd
[00:15] <Q-FUNK> debdiff attached and revised by mpitt
[00:15] <bryce> link me
[00:16] <Q-FUNK> this is to allow orphaners to drop the transitional -amd package and, afterwards, to allow -geode to Conflicts with older -amd
[00:16] <bryce> (I've got some very pressing libx11 changes that are in the queue to do after geode so I want to get this done asap
[00:16] <Q-FUNK> http://launchpadlibrarian.net/15558838/xorg.219630.debdiff
[00:17] <Q-FUNK> the version in the changelog might be incorrect
[00:17] <Q-FUNK> but the rest of the patch should be good
[00:18] <bryce> yeah 10.1 is already taken
[00:18] <bryce> so is 10.2
[00:20] <bryce> hmm, Q-FUNK, 10.2 is that change 
[00:20] <Q-FUNK> yes, mpitt revised by debdiff with the correct anticipated version, afaik
[00:20] <Q-FUNK> by -> my
[00:20] <bryce> so what do you need me to do precisely?
[00:20] <Q-FUNK> apply it?
[00:21] <bryce> like I said, it seems to already be applied and uploaded to hardy-proposed
[00:21] <Q-FUNK> according to mpitt, it's also needed in intrepid.
[00:21] <bryce> as 1:7.3+10ubuntu10.2
[00:21] <Q-FUNK> ah
[00:21] <Q-FUNK> I don't see it in sources yet
[00:23] <bryce> I just apt-get source'd it
[00:23] <Q-FUNK> steve or pitti probably already uploaded it, but I don't see it yet published on this mirror.
[00:23] <bryce> do you have hardy-proposed in your /etc/apt/sources.lst?
[00:23] <Q-FUNK> yes
[00:24] <Q-FUNK> hang on.  I'll switch to the main one, rather than the finnish mirror
[00:25] <Q-FUNK> ok
[00:25] <Q-FUNK> just seems that the finnish mirror hasn't pulsed yet
[00:25] <bryce> ok
[00:26] <bryce> alright I'll get it changed in intrepid
[00:27] <Q-FUNK> so, ok, -all is done.  now, I just have to wait for -nsc and -geode to enter, then test them
[00:30] <bryce> Q-FUNK: ok -geode is in intrepid
[00:31] <Q-FUNK> hm?!
[00:31] <bryce> Q-FUNK: in intrepid's -all
[00:31] <Q-FUNK> ah, ok
[00:31] <bryce> Q-FUNK: you should also forward up with debian to make sure it gets added to debian's -all
[00:31] <Q-FUNK> then I think that intrepid has absolutely everything that we neeed
[00:32] <bryce> ok great, then I'll consider geode finally all 100% done now :-)
[00:32] <Q-FUNK> debian refuses to add any of my packages unless I adopt their scripts.
[00:33] <Q-FUNK> intrepid already has -nsc and -geode in sync with debian and you just fixed -all, so I cannot think of anything else that needs to be done there.
[00:33] <Q-FUNK> hardy-proposed is where the last pieces of the puzzle are missing
[00:34] <bryce> oh yeah, they redid one of our other packages that used cdbs (-newport I think)
[00:34] <bryce> they don't care for cdbs for some reason
[00:34] <Q-FUNK> but then it's only a amtter of a few hours of queing in the buildd
[00:34] <Q-FUNK> ah?  newport also was packaged using cdbs?
[00:35] <jcristau> no
[00:36] <jcristau> btw you want to remove newport from ubuntu
[00:36] <Q-FUNK> jcristau: actually, now that the modular debhelper is out, would it be agreedable to everyone to use that framework as a compromise beween cdbs and the current xsf scripts?
[00:37] <jcristau> i'm not going to repackage everything, no
[00:37] <Q-FUNK> since it allegedly introduces a framework similar to cdbs, it's something I might be able to work with.
[00:38] <jcristau> also, dh 7 doesn't do what i want, so meh
[00:38] <bryce> ah, it was -openchrome not -newport
[00:38] <Q-FUNK> ah
[00:38] <jcristau> i tried using it in xcompmgr as an experiment, doesn't really bring much over previous versions afaict
[00:39] <jcristau> (because i want to build everything in a builddir, mostly, so i have to do that by hand anyway)
[00:42] <bryce> jcristau: why do you say we want to remove newport from ubuntu?
[00:42] <jcristau> bryce: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=393709
[00:43] <jcristau> so right now in intrepid it doesn't even build, as it's 'Architecture: mips'
[00:43] <Q-FUNK> jcristau: I would consider debhelper 7 WIP, but it's already a nice compromise between trusting the almight CDBS for everything and the fine-grained control of dh_* populated rules.
[00:44] <bryce> jcristau: on intrepid it's only listed in vars.mips now anyway
[00:44] <jcristau> bryce: right, i dropped it from vars.* before restricting the architecture field :)
[00:49] <jcristau> Q-FUNK: in any case most of the xsf specific stuff isn't either in debhelper or cdbs (mostly because nobody had time to sort out what should be integrated somewhere else and what should just go), so...
[00:51] <Q-FUNK> jcristau: what would really make it easier for me to use the XSF stuff is if it were implemented as a nice debhelper module that I would just need to add to one of the simple debhelper 7 targets.
[00:51] <Q-FUNK> the same way that, with CDBS, someone just adds an include that is specific to e.g. gnome, kde, perl, etc. packages, there would be one for X drivers.
[00:52] <jcristau> Q-FUNK: i'm not a perl guy, so i don't think i'm not going to do that
[00:52] <Q-FUNK> :)
[00:52] <Q-FUNK> same problem here, I'm affraid.
[00:52] <jcristau> i spend too much time on debian-x as it is
[00:53] <Q-FUNK> but that would just be the general idea, to at least make the debian/rules human-readable again,  then, I would gladly use it.
[00:53] <Q-FUNK> understood
[00:53] <jcristau> what's unreadable about the debian/rules?
[00:54] <jcristau> (not talking about xsfbs.{mk,sh})
[00:55] <Q-FUNK> compare it to the simplicity of the cdbs one I use with -geode and you'll see what I mean.  it's a standard automake case, with small additions for versioned X dependencies.
[00:55] <jcristau> i know what a cdbs rules files looks like, and i don't find that readable
[00:56] <jcristau> as in i have no idea what happens in there
[00:56] <Q-FUNK> I don't find it necessary to know what happens at the atomic level.
[00:56] <jcristau> i find it necessary to be able to fix shit :)
[00:57] <Q-FUNK> so do I and handmade debhelper rules prevent me from dooing that, while cdbs rules help me do that.
[00:58] <Q-FUNK> or rather, I avoid situations that need constant fixing and cdbs is a good tool for that
[01:00] <jcristau> i've seen cdbs break too many times
[01:01] <Q-FUNK> but really, just for the heck of it, try comparing the exact debian/rules I use for -geode with, let's say, the one used by the team for -nsc.
[01:02] <Q-FUNK> I haven't seen it break in ages.
[01:02] <jcristau> yeah, i find one of them readable and the other not. now guess which is which ;)
[01:03] <Q-FUNK> I see more breakages resulting from GCC transitions exposing upstream bugs that I've ever seen CDBS break.
[01:03] <Q-FUNK> tetu, le mec :)
[01:04] <Q-FUNK> but whatever. I'm happy with my choices and so are you. that's what matters the most. :)
[01:05] <Q-FUNK> bryce: nsc was already synced with debian since a few days.  also geode is synced with debian 2.10
[01:08] <bryce> Q-FUNK: ok.  you may want to update the state for the -geode/hardy task
[01:09] <Q-FUNK> geode/hrady is still to be done
[01:09] <Q-FUNK> I'd need to clena up the chnagelog as steve suggested
[01:09] <Q-FUNK> geode/intrepid is what's done (transition in -all, plus latest upstream package via debian)
[01:10]  * bryce nods
[01:12] <jcristau> Q-FUNK: btw you have some stuff duplicated in debian/rules, and you should use /usr/share/xserver-xorg/serverminver instead of dpkg-query -W xserver-xorg-dev; also use /usr/share/xserver-xorg/videoabiver for the Provides
[01:13] <Q-FUNK> jcristau: the duplicated stuff was recently removed.
[01:13] <jcristau> was looking at 2.10.0-3
[01:14] <Q-FUNK> as for servermin and videoabiver, not backward-compatible, unfortunately, and I need this to be portable to Etch until the debian-olpc guys have upgraded to lenny.
[01:14] <jcristau> your debian/control isn't portable to etch anyway
[01:15] <jcristau> because Provides: xserver-xorg-video-2
[01:15] <Q-FUNK> wtf?!  I had removed that duplicated stuff.  I'm beginging to wonder what my sponsors are smoking.
[01:16] <Q-FUNK> jcristau: true, but then it's a one-line change
[01:19] <Q-FUNK> I'd really need to be on the DM keyring to upload my own stuff, to avoid this.
[01:24] <Q-FUNK> bryce: was it to hardy-proposed that you uploaded -nsc ?  
[01:28] <bryce> yes
[01:28] <bryce> Q-FUNK: slangasek's not allowed it through yet though
[01:28] <Q-FUNK> ah
[01:28] <Q-FUNK> that would explain it
[01:29] <Q-FUNK> ok, really heading to bed, now
[02:24] <wgrant> Hmm... Since I rebooted yesterday, my brightness gradually decreases to 0 over a couple of minutes. Running Intrepid with an i915, and I had a similar problem early in Hardy, but it went the other direction. Shall I file a bug?
[21:42] <bryce> tjaalton: are you around?
[23:41] <Q-FUNK> re
[23:42] <Q-FUNK> seems that the nsc version stated in geode 2.2 Conflicts is incorrect
[23:44] <Q-FUNK> other than that, it looks good
[23:53] <Q-FUNK> with the current version missing a 1: and having the wrong version to conflict against, we wouldn't be removing older nsc releases.