[02:19] <lamont-away> because it'd be nice if it booted...
[03:40] <zul> heylo
[05:32] <fabbione> morning
[05:33] <BenC> good evening
[05:36] <fabbione> hey Ben
[05:36] <fabbione> how did the upload go?
[05:36] <BenC> linux-source went fine
[05:36] <fabbione> what about the others?
[05:36] <BenC> l-r-m is giving me some problems, but I think I got it
[05:37] <fabbione> what problems?
[05:37] <BenC> didn't know that the minor's needed to be bumped in debian/rules
[05:37] <fabbione> oh crap..
[05:37] <fabbione> either i forgot and you can blame me..
[05:37] <BenC> but I fixed that, and linux-meta is also uploaded
[05:37] <fabbione> or it wasn't clear in our talk
[05:37] <BenC> well, even Kamion just said to sed debian/control.stub :)
[05:38] <fabbione> eheh ok
[05:38] <BenC> oh shit, I hope the version numbers in debian/control.stub didn't need exact versions
[05:38] <fabbione> good..
[05:38] <fabbione> BenC: the control.stub is enough you change 2.6.12-7 to 2.6.12-8
[05:38] <BenC> do the minors that are listed as depends need to be bumped too?
[05:38] <BenC> ok
[05:38] <fabbione> no need of versioned build-deps
[05:39] <fabbione> not that i remember...
[05:39] <fabbione> i would have to look at it
[05:39] <fabbione> Maintainer: Ben Collins <ben.collins@ubuntu.com>
[05:39] <fabbione> Changed-By: Ben Collins <bcollins@localhost.localdomain>
[05:39] <BenC> well,hope, none there
[05:39] <fabbione> (linux-meta)
[05:39] <BenC> yeah, I didn't have my env setup right
[05:39] <fabbione> ehhehe
[05:39] <fabbione> no biggie :)
[05:40] <fabbione> nobody cares..really
[05:40] <BenC> been awhile since I've done real package uploads ;)
[05:41] <fabbione> yeah i can understand that
[05:42] <fabbione> kernel is 4 pkgs away in the queue on sparc...
[05:42] <fabbione> good...
[05:42] <BenC> i386, amd64 and ppc are done
[05:42] <fabbione> yup i saw...
[05:43] <BenC> I think ia64 is building
[05:43] <fabbione> ia64 is always slower
[05:44] <BenC> I need to get my U2, G5, i2k and A500 running so I can do builds on everything but amd64
[05:44] <BenC> remote builds are great for lots of test builds, but I need to run these things to
[05:45] <fabbione> BenC: keep in mind that i only have i386/sparc64 here :)
[05:45] <fabbione> and even tho i almost managed...
[05:45] <fabbione> sometimes the rules: the patch applies -> it's ready for stable did work
[05:46] <BenC> yeah, I did "patch applies and it built on i386, and the kernel booted for me"
[05:46] <BenC> I'm going to take mdz's advice and start a wiki page on this whole kernel upload procedure
[05:49] <fabbione> hmm yeah.. that might be a good idea
[05:49] <BenC> your notes are so far right on
[05:49] <BenC> haven't gotten to the baz part yet, but that I've already done, so I just have to follow the commands :)
[05:50] <fabbione> i am pretty sure the notes are ok :)
[05:50] <fabbione> BenC: yes please.. try to commit everything in baz asap
[05:50] <BenC> the minor's was the only thing that caused a hiccup
[05:50] <fabbione> if you get to commit almost at atomic level to baz, your kernel will get more testing..
[05:50] <fabbione> i often build out of baz to do testing
[05:51] <BenC> yeah, I need to get a workflow going
[05:51] <fabbione> like: new patch -> one commit. patch update -> one commit
[05:51] <fabbione> yeah no problem..
[05:51] <BenC> this last one was more or less "crap, which patches do I put in, and shit, how do I test them all"
[05:51] <fabbione> but for ex. i have been working on 8.12 a lo in the last days to debug ocfs2
[05:51] <fabbione> you will never be able to test them all
[05:52] <BenC> yeah, but I like to atleast review them
[05:52] <BenC> but aside from the patches, I need to do some of my own debugging and patching
[05:52] <BenC> just going to take me a little bit longer to get into system
[05:53] <BenC> hopefully get back to how I was with libc6 back when I maintained that (which is close to comparable)
[05:53] <BenC> well, as close as you can come, but not really all that close
[05:54] <fabbione> ehhe
[05:54] <fabbione> i remember when you were maintaing libc6
[05:54] <fabbione> you probably don't recall me bitching about the ipv6 resolver being broken
[05:54] <fabbione> and you said  in an email: "Sure.. patches are welcome"
[05:54] <fabbione> you created a monster
[05:55] <fabbione> because at that time i did never even see a deb or a source code that big
[05:55] <BenC> yeah, I do remember that :)
[05:55] <BenC> ipv6 got a lot of attention because of that
[05:56] <fabbione> eheh
[05:56] <BenC> especially the one major bug that took me about 3 days of testing and a hundred lines of code to fix
[05:56] <BenC> getaddrinfo()
[05:56] <BenC> damn this network has gotten really unstable lately
[05:56] <BenC> lilo needs to chill with the restarts
[05:56] <lamont-away> BenC: I see you've been initiated in the world of stupid-pain that is linux-restricted-modules
[05:57] <fabbione> hey lamont-away 
[05:57] <lamont-away> yo
[05:57] <BenC> lamont-away: it welcomed me with open claws
[05:57] <lamont> hehe
[05:57] <fabbione> BenC: lilo should be either hanged or crossburned
[05:57] <lamont> BenC: I have absolutely NO CLUE why whoever started that felt it was so important to have random version numbers for each package4
[05:57] <BenC> need to redo l-r-m's stuff so that it's fairly automated
[05:57] <BenC> me either
[05:58] <lamont> just make them all have the same version number as the package?
[05:58] <fabbione> we can't
[05:58] <lamont> well, actually, they're tracking upstream version numbers where they exist, so I guess that's not entirely bad
[05:58] <BenC> seems to me that every upload would need those minors bumped, with makes them pointless
[05:58] <fabbione> that's why there is that madness
[05:58] <lamont> fabbione: why can't we?
[05:58] <BenC> well, the remaining ubuntu version needs to just be from the package
[05:58] <fabbione> nvidia-$UPSTREAMMAJOR-crack
[05:58] <fabbione> when i first did the package it was only nvidia
[05:59] <BenC> that's the package name though, right?
[05:59] <fabbione> and i did DOCUMENT it
[05:59] <fabbione> the others started adding all that sort of extra crack
[05:59] <BenC> seems to me it should match the kernel version it was compiled for
[06:00] <fabbione> BenC: the kernel modules are ok.. they do
[06:00] <fabbione> the problem is some userland thing..
[06:03] (BenC/#ubuntu-kernel) fabbione: any reason that xorg-driver-fglrx would need to use a special minor version?
[06:03] (BenC/#ubuntu-kernel) I think the minor version should come from debian/changelog and be generated at build time
[06:03] (fabbione/#ubuntu-kernel) BenC: i have no idea.. i only added nvidia to that thingy.. 
[06:03] (BenC/#ubuntu-kernel) for all those things
[06:04] (fabbione/#ubuntu-kernel) i am sure there is the problem with nvidia $major
[06:04] <fabbione> because some other userland pkgs use it as versioned Depends:
[06:04] <fabbione> i really have no clue about the others
[06:12] <BenC> fabbione: in l-r-m, 2.6.12.2, what does the .2 represent?
[06:13] <fabbione> BenC: it's just a number because they have updated something in the orig.tar.gz
[06:13] <fabbione> nothing fancy
[06:14] <BenC> ok
[06:14] <fabbione> the problem with binary driver is that you can't update them with patches easily ;)
[06:14] <BenC> yeah, so it requires a new .orig.tar.gz upload
[06:15] <fabbione> yup
[06:18] <BenC> fabbione: l-r-m and linux-meta built and are in
[06:18] <fabbione> super
[06:18] <BenC> Kamion is doing d-i
[06:18] <fabbione> congratulation :)
[06:18] <fabbione> he did already
[06:18] <BenC> he did?
[06:18] <fabbione> yes
[06:18] <fabbione> in theory you can upload all of them at once..
[06:19] <fabbione> but there are some drawbacks
[06:19] <fabbione> like you need a buildd admin to give them back due to Dep-Wait and FTBFS
[06:21] <BenC> yeah, ia64 needs a give-back
[06:21] <BenC> after the kernel is done
[06:22] <fabbione> is it dep-wait?
[06:23] <fabbione> because there is an auto giveback.. but i don't remember how often it is called
[06:24] <BenC> yeah, dep-wait
[06:24] <fabbione> than it's no issue.. it will be auto-un-dep-wait 
[06:24] <BenC> all the primary arch's are done, so that's the end of my worry for the evening
[06:25] <fabbione> eeheh
[06:25] <fabbione> yeah
[06:26] <fabbione> i will try to collect the ABI files for sparc since it's going to build soon..
[06:26] <fabbione> but i won't be able to commit them until you finish the baz dance
[06:30] <BenC> I'm going to add the bzip compressed ABI files to the linux-image packages for the next upload, so that I wont have to chase the files around like this
[06:30] <BenC> it's only 54k compressed, so not a major percentage of the package size
[06:31] <fabbione> BenC: we did think about that..
[06:32] <fabbione> you still have the main issue if a kernel is FTBFS on an arch.. you need to get the ABI file..
[06:32] <fabbione> that sorts of goes back to you as build on all arches before upload
[06:32] <fabbione> but yeah.. that would do in the majority of the case
[06:32] <BenC> how do you get an ABI file for an arch that wont build?
[06:33] <fabbione> that's the whole point of prebuilding :)
[06:33] <fabbione> to avoid FTBFS and get the ABI files if needed :)
[06:39] <infinity> There's no auto-give-back, that's a myth.
[06:40] <infinity> But if it's properly dep-waited, that will work itself out.
[06:40] <infinity> "auto-give-back" means me or lamont occasionally give back everything that's in the "building" state.
[06:41] <infinity> (Usually me, and usually when I'm doing a mass give-back due to a transition or something that I think may have magically made a bunch of things buildable)
[06:41] <infinity> So, if you guys find any kernel-related stuff that's FUBAR, just let me know, and I'll smack it around.
[06:42] <infinity> (auto-give-back in the context of buildd/sbuild, is only triggered on hideously broken chroots or missing/unavailable source... All other give-backs are manual)
[06:53] <BenC> it's in dep-wait, so it should be ok
[07:00] <mdz> fabbione: the d-i upload happened before l-r-m was built
[07:00] <mdz> fabbione: of course, I don't think nic-restricted-modules actually works anymore...
[07:02] <fabbione> mdz: no, because we don't have ld_static in d-i and Kamion isn't happy to add it
[07:04] <BenC> shouldn't d-i build-dep on the proper version it is expecting?
[07:04] <BenC> I would have expected it to go into dep-wait
[07:04] <fabbione> BenC: you can't
[07:04] <fabbione> because you can't Build-Deps on udeb
[07:04] <BenC> it just uses linux-meta?
[07:04] <fabbione> d-i doesn't use anything else
[07:04] <BenC> ah
[07:05] <BenC> doesn't it check versions of things it downloads?
[07:05] <fabbione> yes.. it does, but d-i is allowed to access the net to build
[07:05] <fabbione> but it happens during the build
[07:05] <fabbione> not at Build-Dep installations
[07:05] <BenC> but I mean, can't it check the version of net-restricted-modules that it downloads to make sure it matches the kernel version? :)
[07:06] <BenC> and then ftb
[07:06] <fabbione> hmm i think it does somehow...
[07:06] <fabbione> but we don't really care..
[07:06] <fabbione> that udeb is borked anyway
[07:06] <BenC> ok
[07:33] <mdz> how goes the new kernel?
[07:34] <mdz> l-m and l-r-m seem happy
[07:34] <mdz> I'm about to trigger livefs builds unless someone tells me otherwise
[07:35] <fabbione> mdz: all the packages seems to be in place.. did yo check d-i?
[07:36] <fabbione> d-i did build..
[07:36] <fabbione> mdz: has it been by-handed?
[07:37] <mdz> fabbione: elmo did it before he went to bed
[07:37] <fabbione> ok.. than i think it's all there...
[07:37] <fabbione> i see no reasons to hold it
[07:38] <mdz> livefs builds in progress
[07:38] <mdz> they take about 40m
[07:38] <mdz> if I am still awake when they complete, I will do live CD builds
[07:39] <fabbione> mdz: perfect.. let me know if we get there, so i can rsync and test
[07:39] <fabbione> mdz: btw we trimmed down the OCFS2 problem.. now we are working on a solution
[09:09] <mdz> fabbione: livecd builds started
[09:10] <fabbione> mdz: great
[09:10] <mdz> will be 20050831.1
[09:10] <fabbione> ok..
[11:35] <chmj> BenC: ping 
[11:48] <fabbione> chmj: you are either too early or too late :)
[11:48] <fabbione> pick one
[11:48] <fabbione> ;)
[12:04] <chmj> whatis benc's local time ?
[12:07] <fabbione> US..
[12:17] <chmj> too late then 
[02:05] <zul> heylo
[02:08] <zul> BenC: congrats your first upload
[02:57] <rtcm> he, so, is there a good reason for the vesafb and fbcon modules (and deps) be loaded by default?
[02:58] <rtcm> on the first versions of 2.6.12 this didn't happen...
[03:02] <zul> usplash i think but im not 100% sure
[03:02] <rtcm> usplash uses vga16fb
[03:03] <rtcm> at least vesafb is completely unnecessary AFAICT
[03:32] <zul> hmmm...no preX,13 yet?
[04:35] <BenC> zul: late night, I'll do the repo switch today
[04:37] <zul> cool..no worries
[05:27] <fabbione> hey BenC 
[05:27] <fabbione> hi zul
[05:31] <zul> hey fabbione 
[05:39] <fabbione> BenC: i have the sparc abi files...
[05:40] <fabbione> so as soon as you open 13 let me know and i will commit them
[09:55] <zul> c ya later