[12:13] <jbailey> It's replacing an existing patch already that has the same name.
[12:13] <jbailey> So the load order should be already determined.
[12:14] <jbailey> (And I didn't play with it outside of that environment...  I'd be hard pressed to remember how to build a kernel other than 'debuild' *g*)
[12:17] <BenC> ok :)
[12:20] <BenC> jbailey: is there a bug that this patch fixes?
[12:22] <jbailey> If bugzilla ever comes up, I'll let you know.../
[12:22] <jbailey> *sigh*
[12:38] <jbailey> BenC: 13535
[12:38] <jbailey> It occured to me that I could just search my email. =)
[06:09] <fabbione> morning
[06:17] <fabbione> BenC: did you commit the patch?
[06:17] <fabbione> or do i need to?
[06:18] <fabbione> argh
[06:18] <fabbione> he is not online
[06:18] <fabbione> jbailey: where is your patch?
[06:19] <jbailey> fabbione: My homedir on concordia
[06:19] <fabbione> ok
[06:19] <fabbione> thanks
 BenC: Whenever you're back, the initramfs DSDT patch is in my homedir on concordia and is called drivers-acpi-osl_attach-dsdt-to-initrd.dpatch
[06:20] <fabbione> drivers-acpi-osl_attach-dsdt-to-initrd.dpatch ?
[06:20] <fabbione> ehhe
[06:20] <jbailey> Right.  Replaces the current version.
[06:20] <jbailey> After 3 attempts to figure out how to call it something else, I gave up and gave it the same  name. =)
[06:20] <fabbione> ahaha
[06:20] <jbailey> (I think you generate your patch lists somehow, but the kernel packaging makes gcc's look simple)
[06:21] <fabbione> yes we do..
[06:21] <fabbione> from 00list-$version to 00list
[06:21] <fabbione> because we support patching per arch and per flavour
[06:21] <fabbione> so we need to generate it dynamically
[06:22] <fabbione> dpatch isn't exactly working
[06:22] <fabbione> what do i need to add in the changelog?
[06:26] <fabbione> jb??
[06:26] <fabbione> what do i need to add in the changelog? <-
[06:30] <fabbione> ok nevermind.. i wrote something you lazy canadian :P
[06:42] <jbailey> Eh, I  don't have a nick highlight on 'jb'
[06:42] <jbailey> =)
[06:42] <jbailey> *sleepy* Canadian.
[06:42] <fabbione> yeah right :)
[06:42] <fabbione> lazy lazy...
[06:42] <fabbione> :P
[06:42] <jbailey> Right.
[06:42] <jbailey> If I were less lazy, I'd have gotten up and gone to bed already.
[06:43] <fabbione> ahaha
[06:44] <jbailey> Sleep now
[06:48] <fabbione> night jeff!
[06:54] <fabbione> lamont: ping?
[06:54] <lamont> eep!
[06:55] <fabbione> lamont: i am on f11.. do we have enough pkgs to build the kernel in the breezy chroot?
[06:55] <fabbione> or do i still need to use hoary?
[06:56] <lamont> I think you can use breezy
[06:56] <lamont> ouch.
[06:56] <fabbione> heck i can't remember my passwd on f11
[06:56] <fabbione> grrr...
[06:56] <lamont> need to have gcc-3.4-hppa64 (still universe) installed in the chroot.
[06:56] <fabbione> lamont: can you reset it for me please?
[06:56] <lamont> hppa-hacks/expect-tcl8.3-dev_5.43.0-2_hppa.deb
[06:56] <lamont> hppa-hacks/expect-tcl8.3_5.43.0-2_hppa.deb
[06:56] <lamont> hppa-hacks/expectk-tk8.3_5.43.0-2_hppa.deb
[06:56] <lamont> hppa-hacks/gcc-3.4-hppa64_3.4.4-5ubuntu1_hppa.deb
[06:56] <lamont> hppa-hacks/palo_1.9_hppa.deb
[06:57] <lamont> that's all the universe-should-be-main packages that I have hacked in right now.
[06:57] <fabbione> otherwise i will just use hoary...
[06:57] <lamont> fabbione: sure
[06:57] <fabbione> just put a temp one that i can change on the fly
[06:57] <fabbione> i am already logged in
[06:58] <lamont> yeah, ISTR that sshkeys are the only way to actually login...
[06:59] <fabbione> yeah but i need the passwd for sudo :)
[06:59] <fabbione> so i can install the B-D
[06:59] <fabbione> or you can install them for me..
[06:59] <fabbione> i really don't care tbh 100%
[07:03] <fabbione> lamont: is the chroot reasonably updated?
[07:03] <fabbione> or do i need to upgrade it first?
[07:04] <lamont> hasn't been touched by me in forever
[07:04] <fabbione> ok
[07:04] <lamont> I'll let you fly it
[07:04] <fabbione> deb http://people.ubuntu.com/~lamont/ubuntu-hppa/tree hoary main restricted universe
[07:04] <fabbione> i guess this is not needed anymore..
[07:04] <fabbione> lamont: that's fine :)
[07:05] <lamont> and should not be present if breezy is
[07:05] <lamont> (it has .debs with diff md5sums, but the same name)
[07:05] <fabbione> i think you did create that chroot during the rebuild of breezy
[07:06] <lamont> so you may have to clone the hoary chroot, nuke /var/cache/apt/archives/*, fix /etc/apt/sources.list, and dist-upgrade
[07:06] <fabbione> for what i need, a dist-upgrade will do
[07:06] <lamont> maybe
[07:07] <fabbione> it's old enough that it's upgrading almost everything :)
[07:07] <lamont> if you get any bitching about md5sums on debs being wrong, remove the afflicted .deb :-)
[07:07] <fabbione> eheh yeah.. have been there before
[07:11] <fabbione> chroot breezy/ apt-get build-dep linux-source-2.6.12
[07:11] <fabbione> Reading package lists... Done
[07:11] <fabbione> Building dependency tree... Done
[07:11] <fabbione> 0 upgraded, 0 newly installed, 0 to remove and 0 not upgraded.
[07:11] <fabbione> i guess that's it :)
[07:17] <fabbione> ok. sparc and hppa building
[07:20] <fabbione> nice.. the latest OCFS2 will compile on sparc and hppa as well..
[07:22] <fabbione> and the others too now...
[07:55] <fabbione> hey Ben
[09:47] <Mithrandir> fabbione: what's the hold-up with 9157?
[09:48] <fabbione> Mithrandir: that the fix we did try, didn't work?
[09:49] <Mithrandir> I can't remember, it's been a whlie.
[09:49] <fabbione> yeah i do remember...
[09:49] <fabbione> that's why.. the patch doesn't help at all...
[09:49] <fabbione> but i don't know why
[09:50] <fabbione> we can ask BenC when he is awake...
[10:18] <Mithrandir> apparently, it should load i2o_core, but not i2o_block.
[10:18] <Mithrandir> what's responsible for loading sd_mod for instance?  I think the same mechanism should be used to load i2o_block
[10:18] <fabbione> hotplug in theory
[10:19] <fabbione> but the point is that it loads correctly the driver for the card, but not the possible associated device manages
[10:19] <fabbione> managers
[10:19] <fabbione> if you load a scsi driver for card foo
[10:19] <fabbione> you don't know what's connected to it
[10:19] <fabbione> it could be a disk or a cdrom
[10:19] <Mithrandir> yeah
[10:20] <Mithrandir> ok, so hotplug should then traverse the device tree and look at what's connected and then load modules as appropriate?
[10:20] <fabbione> that's why initramfs parses /sys after to identify what needs to be loaded
[10:20] <fabbione> yup
[10:24] <Mithrandir> hm, I guess I should see if I can test that machine sometime in the future, then
[10:27] <fabbione> Mithrandir: initramfs atm doesn't check for i2o
[10:27] <fabbione> but it's trivial to implement
[10:27] <fabbione> that would solve your problem i think
[10:28] <fabbione> check in initramfs scripts/functions
[10:28] <fabbione> there is a call to scsi_init_something
[10:29] <fabbione> that's a little bash function that parses /sys in seek of scsi devices and modprobe
[10:29] <fabbione> the same can be done for i2o without any problem...
[10:43] <fabbione> ok ladies
[10:43] <fabbione> 7.11 is ready
[10:53] <Mithrandir> I think I have a fix for i2o, now I just need to get access to the box scheduled..
[12:12] <Mithrandir> fabbione: any idea where Xu's test kernel is?
[12:13] <fabbione> Mithrandir: he did upload already.. it's building mostlikely, waiting for pitti to bless the USN
[12:13] <Mithrandir> ok
[12:13] <fabbione> but i don't have the changelog
[12:13] <fabbione> so i don't know if he added the patch or not
[12:13] <fabbione> it's mostlikely in the archive already
[12:14] <fabbione> because sparc picked it up for building
[12:14] <Mithrandir>   * Fixed sync_page_io bio leak in drivers/md/md.c (Neil Brown).
[12:14] <Mithrandir> yup
[12:14] <fabbione> cool
[12:14] <fabbione> food time
[12:52] <fabbione> hey BenC
[12:52] <fabbione> already awake?
[01:27] <Mithrandir> jbailey: does my i2o patch for initramfs look remotely sane?
[01:27] <jbailey> Mithrandir: I just got here, lemme look. =)
[01:28] <Mithrandir> ok
[01:29] <jbailey> Mithrandir: Where did you send it?
[01:31] <Mithrandir> jbailey: new bug
[01:32] <Mithrandir> 13806
[01:34] <jbailey> Yes, this looks right.
[01:34] <jbailey> Assuming that's where things are in /sys and such.  I don't have i2o systems here.
[01:36] <Mithrandir> well, it's the only thing I saw which resembled a block device.
[01:36] <jbailey> Cool.  Will you have a chance to test it, or should I just apply it?
[01:38] <Mithrandir> I can test it, but not for a while, that box is responsible for all the ads we're sending out here (which means it'll cost about 3000 EUR/day just in lost revenue if we take it down unplanned)
[01:43] <jbailey> Yup, fair enough.
[01:44] <jbailey> I'll just apply it for now then.  It can't be worse than it is now.
[01:44] <Mithrandir> true
[01:44] <jbailey> And I don't see anything harmful in there.
[01:45] <jbailey> I won't upload just for this,  though.  I'll probably pull in the LVM fix first.
[02:08] <fabbione> what time is it in BenC land?
[02:10] <zul> heylo
[02:10] <fabbione> hey zul
[02:11] <fabbione> ok time to start uploading the kernel
[02:11] <fabbione> i can't wait forever
[02:12] <zul> you can do it!
[02:12] <mjg59> Woo kernel
[02:15] <fabbione> eheh
[02:17] <fabbione> accepted
[02:17] <fabbione> this time sparc will be the fastest arch to upload :)
[02:19] <zul> because there are only a couple of users? ;)
[02:22] <jbailey> zul: He'
[02:22] <jbailey> s probably primed the ccache. =)
[02:22] <zul> im evil
[02:22] <jbailey> doko: Quick!  Upload gcc!
[02:26] <mjg59> fabbione: That's -7.11?
[02:26] <fabbione> yes
[02:27] <fabbione> jbailey: no.. even better.. i did a full sbuild while waiting.. the kernel was ready for upload this morning
[02:27] <jbailey> Ahahahah
[02:28] <fabbione> but i was waiting Ben for the baz dance and stuff
[02:28] <fabbione> so in the meanwhile i can start
[02:28] <fabbione> it always takes a few hours before it's builded on all 4 arches
[02:42] <TheMuso> c
[02:43] <zul> hey ben
[02:46] <fabbione> hey BenC
[02:46] <fabbione> TheMuso: you are not the only one in this chan :)
[02:46] <TheMuso> Just realised that. :S
[02:49] <BenC> weird
[02:49] <BenC> not sure why I keep getting disconnected
[02:49] <BenC> fabbione: ready to do this kernel?
[02:49] <fabbione> BenC: yes..
[02:50] <fabbione> i did upload the kernel in the meanwhile
[02:50] <fabbione> all the rest needs to be done
[02:50] <fabbione> as it stands now, preX,11 branch is finished
[02:50] <fabbione> so it's like you completed your devel cycle
[02:51] <fabbione> i usually wait for katie to get the ACCEPT message and only after tag/prepare the new devel branch
[02:51] <fabbione> so when you are ready we can start the baz dance
[02:51] <BenC> let's do it
[02:52] <fabbione> ok..
[02:52] <fabbione> the first thing to do is to get mainline...
[02:52] <BenC> bass line
[02:52] <fabbione> ahah
[02:52] <fabbione> kernel-team@ubuntu.com--2005/kernel-debian--mainline--2.6.12
[02:52] <fabbione> since it's the first time:
[02:53] <fabbione> baz get kernel-team@ubuntu.com--2005/kernel-debian--mainline--2.6.12 debian-mainline
[02:53] <fabbione> store it whenever you want..
[02:53] <fabbione> it doesn't need to be in the kernel source
[02:55] <BenC> ok
[02:55] <fabbione> tell me when you have done that..
[02:55] <BenC> it's done
[02:55] <fabbione> ok
[02:55] <fabbione> now we merge from the pre branch
[02:56] <fabbione> baz merge kernel-team@ubuntu.com--2005/kernel-debian--preX,11--2.6.12
[02:56] <fabbione> you do that inside the debian-mainline dir you just got
[02:57] <fabbione> this will take some time...
[02:57] <fabbione> it's a quite slow operation
[02:57] <BenC> ok...
[02:57] <fabbione> in the meantime i suggest you check your enviroment on people.u.c to be sure 100% that your umask is set to 002
[02:57] <fabbione> for sftp operation
[02:58] <BenC> before or after I do this?
[02:58] <fabbione> during...
[02:58] <BenC> ok, did that
[02:58] <fabbione> it's important the umask is set before you commit
[02:59] <BenC> and the merge is done
[02:59] <fabbione> for reading it's no problem
[02:59] <fabbione> ok
[02:59] <fabbione> did you get any conflict?
[02:59] <BenC> nope
[02:59] <fabbione> ok.. now you can commit...
[02:59] <fabbione> baz commit -s'2.6.12-7.11'
[02:59] <fabbione> -s is the comment...
[02:59] <fabbione> or commit log or whatever
[03:00] <fabbione> baz doesn't spawn a nice editor to make commit logs
[03:00] <fabbione> it's sort of retarded that way
[03:00] <BenC> ok, that finished, but the last line was
[03:00] <BenC> arch: no arch user id set
[03:00] <BenC> that a warning?
[03:00] <fabbione> hmmmm
[03:00] <BenC> or did it fail
[03:00] <fabbione> what's the output of baz my-id ?
[03:01] <BenC> arch: no arch user id set
[03:01] <fabbione> ok
[03:01] <fabbione> baz my-id Fabio M. Di Nitto <fabbione@canonical.com>
[03:01] <BenC> some env I need to set?
[03:01] <fabbione> or whatever fits better to you
[03:01] <fabbione> the commit did fail...
[03:01] <fabbione> or it didn't go trough
[03:01] <fabbione> also remmeber your gpg private key
[03:01] <fabbione> you need to sign the commit
[03:02] <fabbione> otherwise it fails even worst
[03:04] <BenC> * committed kernel-team@ubuntu.com--2005/kernel-debian--mainline--2.6.12--patch-10
[03:04] <fabbione> what's your keyid?
[03:04] <BenC> "Ben Collins <ben.collins@ubuntu.com>"
[03:05] <BenC> 60E80B5B
[03:05] <fabbione> sorry... i mean the 0x..
[03:05] <fabbione> yeah
[03:06] <BenC> so the "--patch-10" is something that baz adds? is that like a revision level?
[03:07] <fabbione> it's like the amount of commits on that branch
[03:07] <fabbione> sort of -rXXXX in svn
[03:07] <BenC> right
[03:07] <fabbione> it's added automatically normaly
[03:07] <fabbione> but you can use it to retrive specific revisions
[03:07] <fabbione> and stuff like that
[03:07] <fabbione> ok.. now
[03:08] <fabbione> we need to tag and prepare the next devel branch
[03:08] <fabbione> do a baz abrowse
[03:08] <fabbione> very useful command to see all the branches in that repo
[03:09] <fabbione> now.. baz as tla sucks hard on how you can name branches..
[03:09] <fabbione> specially when it goes to . - and stuff like that
[03:09] <BenC> lots of them
[03:09] <fabbione> yup
[03:11] <BenC> ok, ready to create a branch
[03:11] <fabbione> baz branch kernel-team@ubuntu.com--2005/kernel-debian--mainline--2.6.12 kernel-team@ubuntu.com--2005/kernel-debian--mainline-2,6,12-7,11-00
[03:11] <fabbione> yes sorry.. connection to uk is slow from here..
[03:11] <fabbione> it was still abrowsing
[03:11] <fabbione> this will create the tag
[03:11] <fabbione> no!
[03:11] <fabbione> there is an weeoe
[03:11] <fabbione> error
[03:11] <BenC> ok
[03:11] <fabbione> baz branch kernel-team@ubuntu.com--2005/kernel-debian--mainline--2.6.12 kernel-team@ubuntu.com--2005/kernel-debian--mainline-2,6,12-7,11--0
[03:11] <fabbione> there
[03:12] <fabbione> note how . - and , change
[03:12] <BenC> that's the tag for the last release, right?
[03:13] <fabbione> yes, that's correct
[03:13] <BenC> done
[03:13] <fabbione> ok
[03:13] <fabbione> now let's create the branch for the new devel
[03:13] <fabbione> baz branch kernel-team@ubuntu.com--2005/kernel-debian--mainline--2.6.12 kernel-team@ubuntu.com--2005/kernel-debian--preX,12--2.6.12
[03:14] <BenC> done
[03:14] <fabbione> perfect
[03:14] <fabbione> now.. go in you kernel source debian/
[03:15] <fabbione> (that i guess you manage with baz)...
[03:15] <fabbione> baz switch kernel-team@ubuntu.com--2005/kernel-debian--preX,12--2.6.12
[03:15] <fabbione> brb.. wife is calling...
[03:19] <fabbione> re
[03:19] <fabbione> tell me when you are done...
[03:19] <BenC> done
[03:20] <fabbione> ok  just one second that i will fix one small thing in the rules file
[03:20] <fabbione> it will make your life simpler :)
[03:21] <fabbione> if you notice at the end of the rules file there are some "I am lazy BOFH" targets to help you do some boring tasks
[03:21] <fabbione> one of them is startnewrelease
[03:21] <fabbione> (that i am fixing now)
[03:22] <fabbione> that's the first command you run on a new devl branch...
[03:22] <fabbione> arch_commit: unable to acquire revision lock (internal error in archive-pfs.c(pfs_lock_revision))
[03:22] <fabbione> GRRRr
[03:23] <BenC> unless .bash_profile isn't getting loaded
[03:23] <fabbione> BenC did you set it at the top or the bottom?
[03:23] <BenC> top
[03:23] <BenC> I just changed what was already there
[03:23] <fabbione> 4 drwxr-xr-x    3 bcollins warthogs 4096 Aug 19 14:14 kernel-debian--preX,12--2.6.12
[03:24] <fabbione> yeah.. it didn't load..
[03:24] <fabbione> let me check where i do it...
[03:24] <fabbione> i do it in .bashrc
[03:24] <fabbione> that's because sftp doesn't load .bash_profile
[03:25] <fabbione> so i need you to do some chmod on people :)
[03:25] <fabbione> cd /home/lamont/public_html/Archives/kernel-team@ubuntu.com--2005
[03:25] <BenC> yeah, doing it now
[03:25] <fabbione> yeah
[03:25] <fabbione> we need it in all the branches/commit you do
[03:25] <BenC> done
[03:26] <BenC> $ find . -user bcollins | xargs chmod g+w
[03:26] <BenC> should get everything
[03:26] <fabbione> it did
[03:26] <fabbione> ok get the change i did to rules and run that makenewrelease as user
[03:27] <fabbione> no need of fakeroot
[03:27] <BenC> added umask to my bashrc too
[03:27] <fabbione> perfect :)
[03:28] <fabbione> it should basically: create the new changelog entry, add debian/patches/00list-* stuff
[03:28] <fabbione> you might want to correct the info in the changelog for the name
[03:28] <fabbione> they depend on your passwd file..
[03:29] <fabbione> but clearly, modify them as you wish
[03:29] <fabbione> these were ok for me.. they might be useless or not enough for you
[03:30] <BenC> startnewrelease?
[03:30] <fabbione> yeah
[03:30] <fabbione> make -f debian/rules startnewrelease
[03:31] <BenC> done
[03:31] <fabbione> ok.. now..
[03:31] <fabbione> baz lint 
[03:31] <fabbione> it will tell you if there are files in the archive that are not in rcs
[03:31] <BenC> no output
[03:31] <fabbione> ok
[03:31] <fabbione> baz diff
[03:31] <fabbione> so you can see what has been done
[03:32] <BenC> changes related startnewrelease
[03:32] <fabbione> perfect
[03:32] <fabbione> now.. we did bump the abi
[03:32] <BenC> already been using that command :)
[03:32] <fabbione> so there is a manual step to do
[03:32] <fabbione> ehe ok.. since i don't know how much you know about baz.. i go step by step..
[03:33] <fabbione> go in debian/abi
[03:33] <BenC> ok
[03:33] <fabbione> baz mv 2.6.12-6.10 2.6.12-7.11
[03:33] <fabbione> once you have done that, commit
[03:33] <BenC> the whole tree?
[03:33] <fabbione> ?
[03:33] <BenC> not just the abi change, right?
[03:33] <fabbione> no no.. the whole tree
[03:34] <fabbione> this is to start a new release after an abi change..
[03:34] <fabbione> the most complex case basically
[03:34] <fabbione> so once you get this one right.. the others will be a breeze..
[03:34] <fabbione> tell me when you are done with the commits...
[03:35] <fabbione> so i can add the new abi files
[03:35] <BenC> done
[03:35] <fabbione> ok gimme a couple of minutes now...
[03:35] <fabbione> in future, when we have abi changes, you will have to take care of i386/amd64/ppc/ia64..
[03:36] <fabbione> that's what you are mandate to :)
[03:36] <fabbione> i will take care of hppa/sparc
[03:36] <fabbione> since it's non-paied porting job
[03:36] <fabbione> but clearly..
[03:36] <fabbione> if you take care of them too.. you will become a Kernel Hero :)
[03:38] <BenC> well, I have both systems here, so I can probably squeeze them out once I settle in to the "have to do" stuff :)
[03:38] <fabbione> well it's simple
[03:38] <fabbione> we officially support only 3 arches
[03:38] <fabbione> i386/amd64/ppc
[03:39] <fabbione> so the rest is work that should come from the community
[03:39] <fabbione> given that i will get pobably ia64 and hppa, and that i love sparc...
[03:39] <fabbione> i took care of them as well
[03:39] <fabbione> it's not high cost to run a build in parallel
[03:39] <fabbione> ia64 is at the DC
[03:39] <fabbione> hppa ping me or lamont
[03:39] <fabbione> sparc we can both do i guess
[03:40] <BenC> amd64 is the only thing I am missing locally
[03:40] <fabbione> BenC: tough.. i only have i386 and sparc :)
[03:40] <fabbione> you build it at the DC
[03:41] <fabbione> somebody will test for you
[03:41] <fabbione> and for what i suggest.. build and work as much as you can on the DC machines
[03:41] <fabbione> they are fast :)
[03:46] <fabbione> ok done
[03:46] <fabbione> BenC: update.. and you are good to go with the next kernel
[03:47] <fabbione> now we need to look at all the surronding of an abi change...
[03:47] <fabbione> BenC: ssh to chinstrap and go to fabbione@chinstrap:~/benc
[03:48] <fabbione> i did prepare all the pkgs for you already signed
[03:48] <fabbione> since i didn't know if you already have upload privileges
[03:48] <fabbione> when changing the abi the sequence to do is:
[03:48] <fabbione> kernel
[03:49] <fabbione> linux-restricted-modules
[03:49] <fabbione> and in no preferred order linux-meta and debian-installer
[03:49] <BenC> ok
[03:49] <fabbione> now.. as soon as the kernel is built on all arches
[03:49] <BenC> what exactly is changed in each of those?
[03:49] <fabbione> http://people.ubuntu.com/~lamont/buildLogs/l/linux-source-2.6.12/ <-
[03:49] <fabbione> linux-restricted-modules is a pain package
[03:50] <fabbione> you need to modify changelog/control.stub/control/rules
[03:50] <fabbione> control and control.stub you need to updated the B-D
[03:50] <fabbione> sed -i -e 's/2.6.12-6/2.6.12-7/g' control.stub
[03:50] <fabbione> will do
[03:50] <fabbione> the REAL pain is rules
[03:51] <fabbione> you need to look at the beginning where there are a bunch of _MINORS=
[03:51] <fabbione> you need to bump these one manually
[03:51] <fabbione> there are 3 or 4 iirc
[03:51] <fabbione> than you just sign and upload
[03:51] <zul>  (NOTE from Fabio: this cool guy did push me something that updates
[03:51] <zul>      drivers-acpi-osl_attach-dsdt-to-initrd.dpatch, but i guess he falled
[03:51] <zul>      over his keyboard before giving me the data for this entry... duh!)
[03:51] <zul> shouldnt that be fell?
[03:51] <zul> :)
[03:52] <fabbione> linux-meta needs a one char change in debian/rules i think
[03:52] <fabbione> it's documented and there is only one place where the abi num is reflected
[03:52] <fabbione> for d-i i usually ping Kamion
[03:53] <fabbione> i don't upload d-i myself.. but he is in holidays now and he gave me the list of things to do
[03:53] <fabbione> in general it's always a grep/sed for the previous ABI
[03:53] <fabbione> now.. the other pain are the seeds
[03:53] <fabbione> the seeds are lists of binary pkgs that are used to move pkgs from universe to main
[03:53] <fabbione> BenC: yeah that's the best :)
[03:54] <fabbione> BenC: we are not allowed (usually) to make arbitrary changes in there
[03:54] <fabbione> that's mdz and Kamion business
[03:54] <fabbione> but..
[03:54] <fabbione> there are a few entries that are names Kernel-Version that are there only for us :)
[03:55] <fabbione> BenC: register the archive i did paste to you
[03:55] <fabbione> that's where the seeds live
[03:55] <fabbione> and get this branch seeds--breezy--0
[03:56] <fabbione> baz get ubuntu-devel@lists.ubuntu.com/seeds--breezy--0 seeds
[03:56] <fabbione> i already updated them for -7 so if you grep for 2.6.12-7 you will see what needs to be changed where
[03:57] <fabbione> so that's basically it
[03:57] <BenC> cool
[03:57] <fabbione> and when you do normal uploads without ABI changes.. you just upload..
[03:57] <fabbione> now one important trick
[03:57] <fabbione> let's make them 2/3 :)
[03:58] <fabbione> when you need to change ABI, bitch always elmo/mdz in advance
[03:58] <fabbione> they need to NEW all the pkgs
[03:58] <fabbione> get them all ready in advance.. the kernel takes hours to build.. even on the buildd...
[03:58] <fabbione> learn the way of baz cacherev :)
[03:59] <BenC> what is that?
[03:59] <fabbione> each time you do 5/6 commits.. run a baz cacherev
[03:59] <BenC> ah, speeds things up?
[03:59] <fabbione> it will create a cached revision both locally and in the archive..
[03:59] <fabbione> yeps..
[03:59] <fabbione> i usually do it every 5/6 commits
[03:59] <fabbione> and at the beginning of each branch
[04:00] <fabbione> on mainline every 2/3 releases...
[04:00] <fabbione> BenC: i would like to stop for an hour or so now and take a nap
[04:00] <BenC> sure
[04:00] <BenC> I've got plenty to do
[04:00] <fabbione> i would like you to keep an eye on the buildd's
[04:00] <BenC> thanks
[04:00] <fabbione> ping elmo to get the kernel newed
[04:00] <BenC> where can I find progress for the buildd's?
[04:01] <fabbione> and if they get in the archive before i wake up.. to upload linux-restricted-.modules from chinstrap
[04:01] <fabbione> you need to use jackass to upload
[04:01] <fabbione> lftp jackass
[04:01] <fabbione> mput linux-restricted*
[04:01] <fabbione> and that's it
[04:01] <BenC> ok
[04:01] <fabbione> there is no progress yet
[04:01] <fabbione> you just keep an eye on the build logs
[04:01] <fabbione> as soon as the 3 main arches are done, you upload l-r-m
[04:02] <BenC> ok
[04:03] <fabbione> there are no l-r-m for ia64/sparc/hppa..
[04:03] <fabbione> so it's pointless to wait :)
[04:07] <fabbione> i am off for a bit..
[04:11] <zul> BenC: in the past i push various bug fixes to fabbione, do you mind registering my archive as well?
[04:12] <zul> http://zulinux.homelinux.net/arch/zulcss@gmail.com--2005
[04:12] <zul> thanks
[04:22] <BenC> sure thing
[04:28] <zul> oh its the quebecker :)
[04:31] <jbailey_> ... and promptly uses it to eat vegan poutine.
[05:07] <zul> yummmy...vegan putone...not..
[05:13] <jbailey_> Vegan gravy can be okay.
[05:13] <jbailey_> Vegan cheese is generally.... not.
[05:22] <BenC> jbailey: do you still have that patch?
[05:23] <jbailey_> BenC: The kernel one?
[05:23] <jbailey_> It's still in my homedir on concordia
[05:23] <BenC> ok
[05:43] <fabbione> re
[05:43] <fabbione> jb, BenC: i did apply that patch already in 7.11
[05:44] <fabbione> how do we look otherwise?
[05:44] <fabbione> missing ppc
[05:51] <fabbione> that would have been jbailey
[05:51] <fabbione> NICKHIGHLIGHT
[05:58] <jbailey_> fabbione: Eh?
[05:58] <jbailey_> fabbione: I'm not sure what respone you're looking for. =)
[05:58] <jbailey_> "Cool, thanks" =)
[05:58] <fabbione> jbailey: your initramfs patch is already part of 7.11
[05:59] <fabbione> you won the last entry in the changelog...
[05:59] <fabbione> or almost ;)
[05:59] <jbailey_> As in, I got in by the skin of my teeth? =)
[05:59] <fabbione> as in "you didn't answer to me this morning for a changelong entry"
[06:00] <jbailey_> Oh, whups. =)
[06:00] <fabbione> but the patch made 7.11
[06:02] <fabbione> oh speaking of bugs..
[06:03] <fabbione> who wants to close them in bugzilla
[06:03] <fabbione> ?
[06:03] <fabbione> zul: i know you love that ;)
[06:03] <jbailey_> "Make initramfs and dsdt not suck.  Side effect:  initrd and dsdt will suck, but mjg59 said that it's okay"
[06:04] <fabbione> too late :)
[06:04] <fabbione> the kernel is up already
[06:07] <fabbione>   Changes by Jeff Bailey:
[06:07] <fabbione>   * (NOTE from Fabio: this cool guy did push me something that updates
[06:07] <fabbione>     drivers-acpi-osl_attach-dsdt-to-initrd.dpatch, but i guess he falled
[06:07] <fabbione>     over his keyboard before giving me the data for this entry... duh!)
[06:07] <fabbione>   (Closes: #13535)
[06:07] <fabbione> you won this entry...
[06:08] <jbailey_> Cool.  My next glibc upload should have The "Hey Fabio, It's fell, not falled" release. on it. =)
[06:08] <zul> lol!!
[06:08] <fabbione> ahha
[06:08] <fabbione> yeah.. sure!
[06:08] <fabbione> i like changelog "war" ;)
[06:09] <mjg59> fabbione: x86 seems to have made the archive. Thanks!
[06:09] <fabbione> mjg59: so does amd64
[06:09] <jbailey_> Yeah, but ever since mailirat, everthing is is just small skirmishes. =)
[06:09] <fabbione> only ppc/ia64/hppa are still building
[06:09] <fabbione> marillat is french.. that says it all
[06:10] <jbailey_> What, that he's really just a gtk bug? =)
[06:10] <fabbione> worst.... ;)
[06:14] <mjg59> fabbione: Excellent
[07:12] <jbailey> mjg59: Does the -7 contain the acpi hacking that might fix STR that used to work?  You had said you didn't think you had time for it, so suspected no.
[08:12] <mjg59> jbailey: It should do
[08:12] <mjg59> But acpi-support needs to be updated as well
[08:15] <mjg59> Hmm. Interesting.
[08:15] <mjg59> Something's changed and breaks bootup on this Toshiba now
[08:15] <mjg59> It probes ide0 and then hangs
[08:15] <mjg59> Oh, no, it just takes FOREVER
[08:16] <mjg59> So it's the same problem it had before. Never mind
[08:16] <jbailey> Joy.  At least it's not another initramfs bug. =)
[08:22] <zul> i didnt do it
[08:31] <fabbione> me neither
[08:31] <zul> i blame foo
[08:33] <fabbione> BenC: are you back online? 
[08:52] <zul> i guess not
[08:59] <zul> holy lag batman
[09:15] <fabbione> BenC: dude???
[09:15] <BenC> damn this connection sucks
[09:15] <BenC> I keep typing and no one is responding
[09:15] <fabbione> yeah we noticed..
[09:15] <BenC> then it finally cycles
[09:15] <fabbione> specially my wife that would like to have a word or 2 with you disappearing :)
[09:16] <fabbione> we were supposed to be at a concert now...
[09:16] <fabbione> forced here by the ABI bump :)
[09:16] <fabbione> not that i mind.. it was a danish crappy group
[09:16] <fabbione> but she did :)
[09:16] <zul> abba?
[09:16] <fabbione> BenC: anyway.. everything went into place..
[09:17] <fabbione> zul: no.. i can't recall the name.. some solo retarded singer..
[09:17] <fabbione> female kind of singers...
[09:17] <BenC> hehe, sorry
[09:18] <zul> heh scandanavians have crappy pop bands
[09:20] <zul> abba, aqua...all crap
[09:20] <fabbione> zul: SafriDuo..
[09:20] <fabbione> should we talk about them???
[09:22] <fabbione> so i guess Benc did fall from the net again
[09:22] <zul> never heard of them thankfully
[09:23] <fabbione> i am pretty sure you did
[09:23] <fabbione> i am ready to bet...
[09:24] <zul> nope..
[09:24] <zul> i wouldnt listen to crap
[09:25] <fabbione> http://www.safriduo.dk/
[09:28] <zul> oh darn i need flash...i dont have flash at work ;)
[09:35] <fabbione> night guys
[09:35] <fabbione> i am crashing
[09:35] <fabbione> cya monday
[09:35] <zul> c ta
[09:38] <zul> i off as well c ya
[10:40] <mdz> where is BenC?
[10:40] <crimsun> his client timed out.
[10:40] <mdz> 2.6.12-7 has broken unionfs
[10:41] <mdz> I am going to need to roll back the whole thing unless someone can give me a hint where it went wrong
[11:04] <BenC> fabbione: any ideas on the unionfs regression between 6.10 and 7.11?
[11:07] <mdz> fabbione collapsed in a heap
[11:07] <BenC> not surprising, he's been pulling some late hours
[11:07] <mdz> I see nothing in the changelog or a naive diff which looks related
[11:07] <mdz> but then, I have no way to even see which files changed from one rev to another
[11:07] <mdz> do you guys have any better tools for that?
[11:08] <mdz> the naive diff is 400k lines
[11:08] <BenC> can you diff between branches?
[11:08] <mdz> yes
[11:08] <BenC> run it through diffstat
[11:08] <mdz> but that's only going to show you which dpatches changed
[11:08] <mdz> not which files in the kernel tree
[11:08] <mdz> I can tell that from the changelog
[11:08] <BenC> true
[11:08] <mdz> the other possibility is a gcc issue, since doko just updated that
[11:09] <BenC> yeah, I think gcc is at fault for atleast one other bug (k7 optimized, crash in inotify)
[11:09] <mdz> 3.4.4-6ubuntu4 is the version used for the build
[11:09] <mdz> that's the old one
[11:09] <mdz> so it isn't that
[11:11] <BenC> only thing I see related to unionfs in the changelog is
[11:11] <BenC>   * Add unionfs-modules udeb:
[11:12] <mdz> yes, which is a packaging change, and doesn't affect the linux-image .deb
[11:13] <BenC> right
[11:16] <mdz> if we're going to track 2.6.12.x, at the very least we need to release a new Ubuntu version for each upstream version
[11:16] <mdz> including this high volume of changes in each upload makes regressions difficult to isolate
[11:17] <BenC> yeah
[11:17] <BenC> is there an updated unionfs?
[11:17] <BenC> not even sure where the patches came from
[11:18] <mdz> no, there isn't
[11:18] <mdz> they come from ftp://ftp.fsl.cs.sunysb.edu/pub/unionfs/
[11:18] <mdz> as should be documented in external-drivers
[11:19] <doko> mdz: no code changes in today's gcc-3.4 upload, and the added include dirs are still empty
[11:19] <BenC> ah, good file
[11:20] <mdz> doko: what about in ubuntu3?
[11:21] <mdz> 2.6.12-6.10 was built with _3.4.4-6ubuntu2
[11:21] <mdz> 2.6.12-7.11 was built with 3.4.4-6ubuntu4
[11:21] <mdz> today's upload was ubuntu5
[11:21] <mdz> so definitely today's upload is unrelated
[11:21] <mdz> hmm, there is no ubuntu3 in the changelog
[11:23] <doko> no, ubuntu3 was something internal preparing the biarch uploads. never hit the archives
[11:26] <lamont> GSI 42 (level, low) -> CPU 0 (0x0000) vector 52
[11:26] <lamont> ACPI: PCI Interrupt 0000:a0:03.0[A]  -> GSI 42 (level, low) -> IRQ 52
[11:26] <lamont> hrm... something tells me there should be more output after that...  :-(
[11:26] <doko> mdz: do we keep the uploads somewhere? I'm currently handicaped with a defect computer in repair
[11:27] <mdz> doko: what uploads?
[11:27] <doko> source uploads to breezy
[11:27] <mdz> morgue
[11:31] <BenC> no compiler warnings from unionfs compile
[11:31] <mdz> I'm testing with the latest unionfs snapshot just for kicks
[11:31] <mdz> they do a snapshot after every commit
[11:33] <BenC> where did the externa-fs-unionfs-fixups.dpatch come from?
[11:33] <BenC> those pulled from unionfs CVS?
[11:34] <mdz> no change
[11:34] <mdz> BenC: yes
[11:34] <mdz> in order to fix a regression
[11:34] <BenC>  and it's odd that it is still needed after so many updates to the main unionfs
[11:34] <mdz> unionfs is not exactly solid at present
[11:34] <BenC> it was pulled in 2.6.11.91-1.1
[11:34] <mdz> but they are building a pretty good regression test suite as the ygo
[11:35] <mdz> they go
[11:35] <BenC> pulled in
[11:40] <mdz> BenC: where do we go from here to debug this?
[11:46] <BenC> why haven't the tracing the kernel
[11:46] <BenC> oops
[11:46] <BenC> tracing the kernel
[11:47] <doko> argh, I pulled the Debian cvs update in 3.4.4-6ubuntu4. there are three fixes, which touch code generation (fixing regressions). does it help you to make a build with these reverted?
[11:47] <BenC> I have it reproducing in my vmware install
[11:47] <BenC> hmm
[11:47] <BenC> yeah, do you have the ubuntu2 packages still?
[11:48] <doko> no, my computer is in repair, I don't know, if I do have them. I cannot find them on morgue.u.c
[11:50] <mdz> 2005-08-12/gcc-3.4_3.4.4-6ubuntu2_i386.deb
[11:54] <mdz> BenC: http://people.ubuntu.com/~mdz/gcc-3.4/
[11:56] <doko> mdz: where did you get these? locate couldn't find that on morgue
[11:56] <mdz> doko: see #-devel
 I stopped morgue syncing because it would run rookery out of sapce
[11:59] <BenC> mdz: I'll have this figured out by tomorrow
[11:59] <BenC> I'll try the older gcc first, if that isn't it, I'll start tracing the driver
[12:01] <mdz> ok, I'm going to work on getting a fallback bind-mount solution in place for ltsp