[00:07] ghostcube: is that the windbond-840 module? or the winbond module in staging? [00:10] w83627ehf kernel modul [00:11] W83667HG chipset on an asusu p5q-pro [00:11] loads fine in rc5 but in rc6 is telling me device busy or not ready [00:11] ghostcube: dkms? [00:12] no [00:13] just wanted to mention it its not the biggest bug ever :) but i dont get why its not loading heh [00:13] ghostcube: can you file a bug, and mark it as a regression [00:14] yeah i can do so where to file ? to l aunchpad ? [00:15] ghostcube: yeah if you run ubuntu-bug -p linux [00:15] it will gather all the information (logs, etc), and automate most of the bug filing process [00:17] ok so i will boot up into rc6 and do the command :) can this be tomorrow i dont know if i can file tonight [00:17] thx and see you [00:17] ghostcube: when ever is convenient for you [00:18] ok :) [03:03] hiii [03:03] can anyone tell me how do I find the offset from two given addresses. They both are void* and I can't subtract them [03:25] looking for 7 figure thinkers leadingedgeentrepreneurs.com [05:28] Hi guys. Does anyone know what happened to the gspca driver in the 2.6.31-rc series from http://kernel.ubuntu.com/~kernel-ppa ? [07:05] apw, sparc patches resent, look for [patch] SPARC config patches [08:24] Fjodor: http://ubuntuforums.org/showthread.php?p=7794205 [08:25] it has been confirmed, that the configs are not synced, but not if anyone is fixing it or when :/ [08:26] i did not dare to whine too much though :) [08:44] ameno, i am aware of he issue [08:45] its not that the configs are not sync, but that the wrong series config is being used and mainline has changed the name of a couple of the main top level options [08:45] which triggers whole subsystems to dissappear [08:46] ah ok, someone else here told me the "out of sync" story [08:46] so its not as trivial as i thought [08:46] can we help somehow? [08:50] I guess we just need to find either a way to have a special mainline builds config for older kernels (not using kms by default) or switch to karmic configs instead. But that might prevent testing from Jaunty (and earlier) environments [09:05] yeah the logical first step is to try a build using those karmic configs and see if that fixes the underlying issue [09:05] as smb says that may trigger the kernels not being useful on jaunty without manual configuration [09:06] in that case we would at minimum need hints on what to do when using them on the mainline builds wiki [09:06] apw, bah, shoot me w.r.t. to the SPARC patches, something has gone freaking horribly wrong :-/ [09:06] apw, I don't get why it failed in the DC and not for me, I removed the modules it wanted ... *sigh* [09:21] NCommander, its like that [09:26] apw, ? [10:30] morning! [11:06] NCommander, i was just saying life tends to go wrong [11:06] Q-FUNK, morningg [11:07] apw: don't mean to psuh anything, it'd just be nice to know for planning: will there be another kenrel upload with that AUFS config change before beta? [11:10] the change is commited, so it'll be in the next kernel upload [11:10] that'll likely be whenever the next -rc is out [11:38] apw: weren't we working on some old regression that affects SCx200 geodes, at some point? [11:38] Q-FUNK, couldn't say any more, too long ago [11:40] apw: bug #241307 [11:40] Malone bug 241307 in linux "kernel oops during bootup in LTSP" [High,In progress] https://launchpad.net/bugs/241307 [11:41] still unresolved [11:45] indeed. there are a number of bugs unresolved [11:45] the issue there is that although we know the patch which triggered it, its not trivial to back out [11:45] its the kind of thing where someone with the h/w would be in the best position to fix it [11:46] to work out what in the new code is tripping the oops [11:47] with that info we could then quirk the code for the machine i assume [12:06] apw: I have the hardware, but I lack the knowledge to hack on kernels. [12:06] is that your bug? [12:07] yup [12:52] Q-FUNK, what was the latest kernel you tried on this thing? [12:52] apw: good question. some 2.6.28 IIRC [12:56] migth be worth trying a karmic kernel, as there appear to be yet further a20 updates [12:57] in the latest kernels, which seem to talk about issues on more embeded systems which don't have a keyboard controller [12:57] and the bit you commented out included that controller setup [12:59] apw: noted. might be worth commenting that on the bug as well. it'll make it easier for me to pass the info around. [13:02] done [13:10] thanks! [13:45] rtg is this valid: [13:45] ifeq ($(disable_d_i),) [13:45] $(DEBIAN)/rules do-binary-udebs; [13:45] endif [13:46] as in .orig.gz + diff upload mode the rules is not going to executable [13:46] normally we run things like perl [13:46] so i'd expect us to be using make -F [13:46] apw, you could try $(MAKE) -f $(DEBIAN)/rules do-binary-udebs; [13:47] right, that would seem appropriate replacement [13:47] I try that immediately [13:47] well with $(DEBIAN) too [13:47] smb, take the one from main debian/rules file [13:48] rtg in karmic we would be ok, as we have that file +x and we are still makeing -sa kernels [13:48] but i think the issue will be there too [13:48] apw, how odd [13:49] i have the feeling that buildd chmod +x debian/rules then run it [13:49] so we have gotten away with it before hand [13:50] apw, how about debian/debian.env ? [13:50] that we generate during the build, so its got whatever perms it has when we make it [13:51] so its unaffected by the -sa !-sa thing [13:51] apw, only if the 'clean' target is run [13:51] otherwise its part of the diff [13:52] hrm yes you are right [13:52] bum [13:52] i wonder if it needs to be -x [13:52] apw, its likely the buildd runs 'clean' first, but I'd have to check for sure [13:52] apw, I replaced that make call and am rerunning the build now [13:52] oh yes they do [13:53] rtg i'll confirm that by uploading netbook [13:53] apw, btw, I run clean first as well [13:53] yes but thats before the diff stage [13:53] the -x on the master version is my fault [13:53] amitk, mjg59: do you know why the touchpad resolution is reported as 1x1 in the kernel? Here's the relevant kernel commit which introduced the resolution: http://git.kernel.org/?p=linux/kernel/git/mingo/linux-2.6-x86.git;a=commitdiff;h=ec20a022aa24fc63d3ab59584cb1e5aa9a21d46c;hp=d7ed5d883c09c5474f842dcb148515dfaef2a567 [13:53] but it rasises an issue that an upload version won't work cause they get lost anyhow [13:54] so it'll be like you are seeing [13:54] even if its fixed in the git tree [13:56] I will get to tun both variants, atm the patches change +x to debain/rules stub and -x on debian.master/rules. Probably debian/rules needs the +x [13:57] i am pretty sure the the buildd process has to chmod +x debian/rules [13:57] due to the orig+diff handling and there being on requiorement on language for debian/rules [13:58] Sounds like a reasonable bet [13:58] hense we have never noticed this internal connection before [13:59] i only thought of it cause i am looking at the whines as i make the netbook upload to test [13:59] dpkg-source: warning: executable mode 0775 of 'debian/debian.env' will not be represented in diff [14:00] yeah, there has been much whining ever [14:12] apw: dpkg-buildpackage will always chmod debian/rules; search for chmod in /usr/bin/dpkg-buildpackage [14:12] That's the only executable file from the diff.gz [14:12] lool, thanks, that meets my expectations [14:26] apw, whats that apt proxy/cacher package that you use? [14:26] i use apt-cacher-ng, seems to be almost painless [14:28] me too [14:31] rtg, i was looking at the imx51 branch on karmic that you rebased [14:31] it seemed you were deleting the whole of debian and debian.master then crteating debian.imx51 [14:31] that seems counter intuitive, is that just a hang over? [14:32] apw, no, it seemed the best way to avoid _any_ conflict in the debian directory [14:32] those commits will always be on top [14:32] but by definition it creates a git conflict on the remove commit [14:33] and i thought the point of the split was that the new stuff would always be in debian. and couldn't conflict [14:33] you know, I remember thinking about that yesterday, and now you've reminded me. [14:34] so if newfiles are created in debian.master, then we'll have issues with theremoval, right? [14:34] or if they change too, i believe you will get a change/remove conflict [14:34] shit, guess I'd better fix that [14:34] on the netbook branch i just left it in there [14:34] and added the change to add DEBIAN=xxx [14:35] I use tab completion, so I try to get rid of duplicates === bjf-afk is now known as bjf [14:35] rtg ahh i see, that is annoying for me too :/ [14:35] apw, ok, thanks for spotting that [14:35] its a work in progress all round [14:36] bjf, that means more mucking about in your dove branch [14:36] You still might call it master.debian and master.whatever [14:36] err [14:36] dumb [14:37] I meant master.debian and soemthing like arm.debian [14:37] smb, indeed, one could. [14:37] rtg, i'm shocked! [14:38] I shocked that you're shocked! [14:38] bjf, Hard to believe :) [14:38] smb, indeed :) [14:38] rtg, how is the imx51 renaming going, looks like i might have a livefs build server later today [14:39] i'd like to build images with the properly named packages ;) [14:39] ogra, the kernel is renamed, but I have been ignoring the headers package naming issues for now. [14:40] ogra, I still need to do the -meta package as well [14:40] well, imx51 will have to replace the existing imx51 headers packages [14:40] else we end up with a massive mess [14:40] currently there exist header packages in the archive due to the fact that imx51 was built in karmic from the main package for quite a while [14:41] ogra, those packages should get reaped eventually [14:41] so we have to be cautious these get properly replaced ... and indeed however yu name them, dont forget upgrades ... [14:42] it will likely be a hell of conflicts/replaces you need [14:42] for the kernel itself meta should take care as long a meta is named like in jaunty (which i think we agreed on) [14:43] so there noting special is needed [14:44] ogra, meta names are likely part of a LiveCD seed, aren't they? [14:44] not seed, but yes, they get used in the liveimage [14:45] armel) [14:45] case "$SUBARCH" in [14:45] imx51) [14:45] LIST="$LIST linux-babbage" [14:45] thats the imx51 snipped from the livefs builde [14:45] r [14:45] *snippet [14:45] its just a matter of renaming it to linux-imx51, one line change [14:45] ogra, didn't we switch from -babbage to -imx51 ? [14:45] the tricky part is debian-installer [14:46] ah, never mind [14:46] for A4 i tried linux-babbage :) [14:46] but that ended up with a mess with the headers [14:58] smb, where am I in your queue? (the FSL patch redo) (I'm not pushing, i'm just asking in case I get asked) [14:58] bjf, Something around #2 or #3 [14:59] smb, timewise, you think this week or next? [14:59] ameno, the latest mainline daily was built with karmics config, perhaps you could see if that has the bits you expect in it [14:59] bjf, I try to get the abstract stuff in place to use it with the arm branch too. Rather this [15:00] smb, ok [15:01] smb, we gonna have as separate source upload for arm in jaunty like we do in karmic? [15:01] apw: yes, looks like drivers/media was compiled in \o/ will test it now... [15:02] apw, Yes, it seems we need that too [15:02] smb the patch stack in the abstracted-debian contains the changes to let the name come fromt he changelog [15:02] bjf, its actually also on my list of things todo, buts its lower down after Karmic and LTS backports stuff. [15:02] smb, is there anything I can help you with w.r.t. FSL [15:02] but you want to get with rtg to make sure you have the latest naming jiggles right [15:02] apw, Yeah, another reason to get that going first [15:03] apw, Yep, my intention. I got an eye on his discussions too [15:03] smb the netbook branch looks good with the new layout [15:03] rtg, what's on your todo? [15:03] bjf, Jaunty FSL [15:03] bjf, Not that I can say now. I likely will run in issues when actually pulling your patches [15:04] bjf, But I will get back to you then [15:04] rtg, oh :-) [15:04] smb, just awaiting the PPA builds of it to confirm its a builder [15:04] smb, ack [15:04] smb, thats with your little fixy applied for the udeb problem [15:04] apw, Ok, mine here seems to run now for quite a while [15:04] apw, So it looks somewhat good, but I wait for it to fionish [15:05] smb, about 10 mins from done i'd say here. and its buildng in a buildd env as a better test [15:05] rtg, If you say Jaunty FSL, are we about to collide with things? [15:05] smb, yep, its gonna take some hackery and git futzing [15:06] I was gonna wait until netbook/master settled [15:06] rtg, Actually it sounds as you are about to go for the same things as I wanted [15:06] smb, yeah sounds like a 100% duplication rather than a clash [15:06] smb, if you've time, then feel free [15:07] I'm still working on Karmic dove, then -meta, then LTS, then .... [15:07] smb, i can work with you based on my netbook experience here in theory [15:07] rtg, Probably not the same. I get confused by the fact that the patchsets for imx51 are from fsl as well [15:12] <_ruben> not sure if this is a -kernel thing, but any clues as to why a partition wouldnt show up under /dev/disk/by-label and by-uuid, it is present under by-id and by-path [15:19] _ruben, it would depend on whether it has a uuid and label, they are often filesystem format specific [15:20] it there needs to be space allocated to it in the fs format not all do [15:21] those files are made by udev [15:21] Greetings Guys. Yesterday my Update Manager downloaded a Kernel Update (uname -r = 2.6.28-15-generic) but now today the Update manager is showing me again that I should update the kernel to... 2.6.28-15? Is that normal or is this a fix of the previous kernel? [15:22] <_ruben> apw: hmm .. it has both a label and uuid according to blkid, its an ext2 partition [15:22] then its most likely an issue with udev [15:25] hrm, can you paste the output of /sbin/blkid -o udev -p /dev/sdX1 [15:25] where X matches your disk [15:26] Frank83: -15 is the ABI version, not the package version [15:26] Frank83: Some updates may keep the same ABI [15:26] <_ruben> Invalid output format udev. Choose from value, device, list, or full [15:26] _ruben, what version of ubuntu are you on [15:26] Frank83, Normal in the sense we had -15 accepted from proposed and today a now version from sercurity came, with the same abi number [15:26] mjg59, Ah, Thanks. That I didn't know. I'll download it, then. [15:27] mjg59, If I made a change to this kernel (Installed Compat-Wireless package to support my Wireless), will this update roll back those changes? [15:28] mjg59, While you are around. Did you get the feedback about acpi_video and _DOD and _DOS? Though there seems to be much change upstream [15:28] <_ruben> apw: fresh intrepid upgraded to jaunty (my pxe env doesnt do jaunty yet) .. but the problem was there in intrepid as well i think [15:28] smb: Yes, there's a discussion in upstream bugzilla [15:28] Frank83, No, it includes it [15:28] Frank83, did you install compat-wireless by installing linux-backports-modules ? [15:29] apw, Negative. I installed it from the source page of the project. [15:29] mjg59: shall I take it as a no? [15:29] Frank83, Oh, in that case it might replace it again [15:30] Frank83, Have you tried backports-modules? [15:30] tseliot: Sorry, I didn't highlight [15:30] smb, Okay. I'll take that into account. No, I haven't tried those, didn't even know of their existence until you mentioned it. What do they do? [15:30] tseliot: At a guess, the hardware doesn't support it? [15:31] Frank83, That package contains the code from compat-wireless put into a separate dir, so it supersedes the normal kernel drivers [15:31] Frank83, And is preserved on updates [15:31] mjg59: is it possible that the kernel doesn't know about the resolution? (just asking) [15:31] smb, my netbook build failed, seems the debug crapola isn't quite right [15:32] working on it now [15:32] smb, Wow, that will save me a lot of time! Thanks! Wonder why I didn't find / saw that in the project page [15:33] apw, My build is still running, odd, as if the concurrency_level is wrong... [15:34] smb, If this kernel update breaks up the compiled compat, then I will install the modules. Installing the modules over the compiled compat doesn't look like a good idea to me. [15:35] rtg, this bit in the karmic stuff that handles the ddebs [15:35] Frank83, You should find it it aptitude. or "sudo apt-get install linux-backports-modules-jaunty" [15:35] mv ../$(dbgpkg)_$(release)-$(revision)_$(arch).deb \ [15:35] ../$(dbgpkg)_$(release)-$(revision)_$(arch).ddeb [15:35] grep -v '^$(dbgpkg)_.*$$' $(DEBIAN)/files > $(DEBIAN)/files.new [15:35] mv $(DEBIAN)/files.new $(DEBIAN)/files [15:35] rtg, that looks wrong to me... i think that bit should be debian/ throughout [15:36] if you agree i'll fix it [15:36] smb, I will search for those. Many thanks for the help. :-) [15:36] Frank83, you're welcome :) [15:37] mjg59, Do you happen to have the bugzilla number for that discussion? [15:37] smb: Not offhand [15:37] apw, I think you're right. its the first time the debug packages have been built. [15:38] yep ... will get this bugger built eventually, and fold the changes back to karmic [15:38] mjg59, Ok, np. Just wanted to make sure you got the info that allowing _DOS or _DOD in both places as a valid video device got the backlight working for that one guy with the acer laptop [15:38] apw, test it locally by 'sudo mkdir /CurrretnBuilding ' in your chroot [15:38] but spell it right [15:39] rtg heh, yeah [15:39] it makes mammoth file foreach flavour, roughly 500MB [15:39] yeeks [15:39] which is why I don't do it very often [15:39] smb: http://bugzilla.kernel.org/show_bug.cgi?id=13577 I think [15:40] bugzilla.kernel.org bug 13577 in Power-Video "Broken backlight control on Acer Aspire 8930G due to buggy DSDT/ACPI Video" [Normal,Resolved: code_fix] [15:45] mjg59, sounds similar. Though the reporter here had a 6920G, but Acer seems to be consistent in that respect [15:45] Yes [15:48] mjg59, In the case I had the problem were multiple definitions but only one had _DOD *and* _DOS defined as the code required. All others only had only one. So changing the if cases to allow one or the other in video and video_detect solved that [15:54] apw: the current vanilla image looks ok and the module i missed before works [15:55] the headers and source package seem broken though [15:55] 1,4kB is a bit small for a kernel source package ;) [15:56] its just good compression :P [16:03] ogasawara: bug 388923 needs to be unwound...looks like a lot of unrelated bugs have been marked as duplicates of that bug [16:04] Malone bug 388923 in linux "package linux-image-2.6.28-13-generic 2.6.28-13.44 failed to install/upgrade: subprocess post-installation script returned error exit status 10" [High,Triaged] https://launchpad.net/bugs/388923 [16:04] ogasawara: I've added code to apport to handle the nvidia-common case already [16:04] but there are some others mixed in there [16:04] mdz: ok thanks, I'll take a look [16:05] ogasawara: when I come across an apport-package bug, I generally try to extract the error message from DpkgTerminalLog.txt and put it into a comment, so it's easier to triage onward [16:06] ogasawara: eg, I just did bug 390905 [16:06] Malone bug 390905 in linux "package linux-image-2.6.28-13-generic 2.6.28-13.44 failed to install/upgrade: podproces post-installation script zwrócił kod błędu 1 (dup-of: 388923)" [Undecided,Invalid] https://launchpad.net/bugs/390905 [16:06] Malone bug 388923 in linux "package linux-image-2.6.28-13-generic 2.6.28-13.44 failed to install/upgrade: subprocess post-installation script returned error exit status 10" [High,Triaged] https://launchpad.net/bugs/388923 [16:09] ogasawara: I'm using bug 386042 as the master bug for this one: /usr/share/debconf/confmodule: line 42: printf: write error: Broken pipe [16:09] Malone bug 386042 in grub "package linux-image-2.6.28-13-generic 2.6.28-13.44 failed to install/upgrade: subprocess post-installation script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/386042 [16:12] and 385771 for the update-grub merge failure [16:15] mdz: I'd been using bug 269539 for the update-grub merge failure [16:15] ogasawara: Error: Could not parse data returned by Malone: The read operation timed out [16:17] ogasawara: oh, thanks. sorry about that [16:25] ogasawara: howdy! would you happen to have a kernel for me to test drive? [16:25] Q-FUNK: unfortunately not yet, fails to build and I haven't been able to dig through the log yet [16:26] ogasawara: oh? just disabling that config option makes the build fail? [16:27] Q-FUNK: I had to disable quite a few config options which turned that config option on [16:27] Q-FUNK: but I'm not sure why those would then result in a failed build [16:27] Q-FUNK: hence me needing to look at the build log [16:28] ogasawara: interesting. something tells me that sending your logs to Al Viro might be a good idea, as it seems that his recent changes feature a number of interesting regressions, this panic on Geode being only a small fraction of the lot. [16:29] ogasawara: if nothing else, it could prevent some seriously broken code from being included in the 2.6.31 final release [16:29] keeping the LKML in the loop might also be appropriate, in this particular case, me thinks. [16:30] Q-FUNK: indeed [16:30] Q-FUNK: I was having some git issues yesterday so I just want to make sure the failed build wasn't a result of that [16:44] with kernels > 2.6.28-12 when I'm done installing them I continue to get `avc_has_perm_no_audit kernel panic not syncing' [16:45] then the keyboard becomes nonresponsive. [16:45] what ought i do [16:46] ret: do you have selinux enabled? [16:46] jjohansen: I do [16:46] ret: my guess is you need updated policy [16:47] jjohansen: how would I go about this? [16:48] ret: have you tried pulling in the latest policy, you could also try audit-to-allow [16:49] hrmm, or something like that its been a while since I looked at that [16:49] ret: ubuntu-hardened is a better channel to ask selinux questions in [16:53] bah, why is it kernel-team@lists instead of ubuntu-kernel@lists like everything else? [16:55] jjohansen: eh, looking at the config shows that its been mucked with [16:55] audit2allow --verbose returns nothing ;\ [16:56] looks like its fixed though. [16:56] thanks [17:04] bjf, amitk: repushed Karmic fsl-imx51 with updated the debian abstraction according to apw. [17:05] hmm, garbled structure sentence [17:06] s/according to/as suggested by' [17:08] so i can blame you for all the breakae now ? cool ;) [17:09] * ogra tickles apw [17:09] * apw falls off his chair [17:12] ogasawara: do you have a master bug for the "missing /boot/grub" errors? [17:12] like bug 394729 [17:12] Malone bug 394729 in linux "package linux-image-2.6.28-13-generic 2.6.28-13.45 failed to install/upgrade: subprocess post-installation script returned error exit status 1 (dup-of: 388923)" [Undecided,New] https://launchpad.net/bugs/394729 [17:12] Malone bug 388923 in linux "package linux-image-2.6.28-13-generic 2.6.28-13.44 failed to install/upgrade: subprocess post-installation script returned error exit status 10" [High,Triaged] https://launchpad.net/bugs/388923 [17:14] mdz: for that one I don't. I've usually been going back and forth with bug reporters to find they've uninstalled grub somehow and mark them invalid [17:15] ogasawara: that's weird, uninstalling grub shouldn't remove /boot/grub I don't think [17:15] mdz: or I've seen the case where they don't have their /boot partition mounted and it fails with the same issue [17:22] mdz: have you started using a master bug? [17:22] ogasawara: no [17:25] ogasawara: are there any others which are common enough that we should add bugpatterns for them? [17:25] mdz: aside from what I have documented in that wiki, I haven't seen any [17:26] ogasawara: only a few of the ones on the wiki have bugpatterns that I see [17:26] ogasawara: thanks for adding 386042, I think that one warrants a bugpattern [17:26] mdz: I can do the bugpattern for that if you haven't already [17:27] ogasawara: doing it now [17:27] mdz: awesome, thanks [17:28] ogasawara: done [17:28] rtg: ack [17:29] ogasawara: I'm not sure what to do with the --fsys-tarfile ones [17:29] those are definitely not kernel bugs, but it's hard to tell where they belong [17:29] I've been putting them on no package for now [17:29] but probably we should bugpattern them [17:29] getting more reports isn't helping [17:30] amitk, bjf: there are some more fsl packaging changes coming down the pipe. Steve L was not happy with libc-dev and headers package names. [17:30] mdz: what's the master bug for that one? [17:30] mdz: at least we can group them together first [17:30] ogasawara: I don't think there is one. I think it means the disk filled up [17:31] ogasawara: we could set up a master bug, or we could just point people to a wiki page [17:31] a lot of these would have been redirected to grub or initramfs-tools if the user were running karmic [17:31] rtg, I assume those will be in your preview branch at some point. Then I can pick those up when I do the dance for Jaunty [17:32] smb, I'll shove them into the core repo later today, as soon as I sort out the naming issues. [17:32] mdz: I'd maybe lean towards pointing to the wiki page [17:32] ogasawara: I've come across a few duplicates of bug 78478, which isn't on the wiki page yet and doesn't have a bugpattern [17:32] Malone bug 78478 in grub2 "Package: linux-image-2.6.20-5-generic fails to install" [Undecided,Fix released] https://launchpad.net/bugs/78478 [17:33] rtg, Ok cool. I make sure I closely follow that lead [17:33] mdz: whoa, 2.6.20 :) I actually hadn't seen that one yet. [17:34] mdz: I'll add it to the wiki and get the bugpattern [17:34] ogasawara: I believe that's the case where someone installs grub2 (pre-karmic) [17:34] and update-grub is /usr/sbin/update-grub [17:34] but /etc/kernel-img.conf still points to /sbin/update-grub [17:37] ogasawara: I think we need to write some code to automatically extract the error message from DpkgTerminalLog.txt before submitting the bug [17:37] mdz: that would be awesome to have it injected in the bug description [17:41] ogasawara: the fsys-tarfile error is a nasty one, because it can be legitimate [17:41] you get that error if a package is trying to overwrite a file in another package, e.g. [17:41] mdz, flash-kernel on armel uses the symlinks [17:41] mdz: so we should have a master bug then [17:42] (just seeing your mail) [17:42] ogasawara: the actual error depends on what precedes that in the log [17:42] in this weird case we're seeing, there is no error message preceding it [17:42] just [17:42] Unpacking linux-image-2.6.28-13-generic (from .../linux-image-2.6.28-13-generic_2.6.28-13.44_i386.deb) ... [17:42] Done. [17:42] dpkg-deb: subprocess paste killed by signal (Broken pipe) [17:42] dpkg: error processing /var/cache/apt/archives/linux-image-2.6.28-13-generic_2.6.28-13.44_i386.deb (--unpack): [17:42] subprocess dpkg-deb --fsys-tarfile returned error exit status 2 [17:43] hrm, not exactly helpful [17:45] that "Done." indicates the preinst finished successfully (I don't know why the kernel preinst does that but it does) [17:58] mdz: I pushed a bugpattern for 78478 but used "Could not find postinst hook script [update-grub]." as the reg exp [17:58] mdz: but it seems it could also use "The provided postinst hook script [/sbin/update-grub] could not be run." [17:59] ogasawara: I've not had much luck making a bugpattern for the above...it's finicky about multiline patterns [17:59] ogasawara: did you test that pattern? [18:00] it contains square brackets, which are regex metacharacters [18:00] ./test-local will tell you if it matches a particular bug [18:00] mdz: I'm going to uncommit it [18:02] ameno, apw: Thank you for explaining [18:03] apw: Is there any timeline for when it is resolved? [18:04] Fjodor, for what to be resolved? [18:05] Fjodor, too many threads going on [18:05] apw: Oh sorry, I should have mentioned :-$ The problem with 2.6.31-kernels from kernel-ppa not having a lot of media/video modules [18:05] try the latest daily [18:06] that i build with the karmic configs, ameno seemed to think it was better [18:06] apw: Ok, great! Thank you :-) [18:06] better yes, but did you read my two lines afterwards too? :) [18:06] apw: ^ [18:07] you'll ahve to repeat them [18:07] apw: the headers and source package seem broken though [18:07] apw: 1,4kB is a bit small for a kernel source package ;) [18:07] ogasawara: I've escalated bug 269539 with the foundations team because it has so many duplicates [18:07] Malone bug 269539 in ucf "package linux-image-2.6.27-3-generic failed to install/upgrade: "Conflicts found! Please edit `/var/run/grub/menu.lst' and sort them out manually."" [Undecided,Confirmed] https://launchpad.net/bugs/269539 [18:08] mdz: thanks [18:08] ameno, yep aware of those, thats a systemic problem in karmic builds, am working on testing the fixes for that right now [18:08] ogasawara: bug 391519 is another of those weird "failed to run depmod" with no error messages. do you have a master for that? [18:08] Malone bug 391519 in linux "package linux-image-2.6.28-13-generic 2.6.28-13.44 failed to install/upgrade: subprocess post-installation script returned error exit status 1 (dup-of: 388923)" [Undecided,Invalid] https://launchpad.net/bugs/391519 [18:08] Malone bug 388923 in linux "package linux-image-2.6.28-13-generic 2.6.28-13.44 failed to install/upgrade: subprocess post-installation script returned error exit status 10" [High,Triaged] https://launchpad.net/bugs/388923 [18:08] we saw it during the sprint [18:08] cjwatson was looking at it [18:09]