/srv/irclogs.ubuntu.com/2013/04/15/#ubuntu-release.txt

=== doko_ is now known as doko
apacheloggercould someone please reject the pending konversation upload07:17
dokoapachelogger, done07:51
apacheloggercheers07:51
chrisccoulsonfinal freeze is this week isnt' it?08:13
Laneyyes08:14
seb128chrisccoulson, https://wiki.ubuntu.com/RaringRingtail/ReleaseSchedule says thursday08:14
chrisccoulsonseb128, thanks08:15
chrisccoulsoni wonder what the chances are of me getting 2 firefox uploads in this week? i want to temporarily turn a crash in to a runtime abort to try and get more info out of it ;)08:15
chrisccoulson(and i have another crash fix and a packaging fix)08:15
Riddellchrisccoulson: and fix the kubuntu issue?08:20
chrisccoulsonRiddell, yes, that would be one of them08:20
seb128chrisccoulson, well, I guess you tried other ways to get debug infos without luck?08:23
chrisccoulsonseb128, all i have is crash reports that don't have enough info and with no contact details for the reporters08:23
seb128chrisccoulson, the builders queue seems quite empty, I guess you can try to convince Laney (or whatever other r-t member you find around, which is likely not a lot at this time of the day)08:23
seb128chrisccoulson, is that with the raring version? I should maybe run that rather than the ppa ... I guess you don't have steps to try reproduce the issue?08:25
Laneyseems tight ...08:25
chrisccoulsonseb128, it's actually with the current version in all releases08:26
chrisccoulsonseb128, i'd rather you ran the ppa version though ;)08:26
tseliothi, can any admin please reject nvidia-graphics-drivers-304-updates (304.88-0ubuntu2) and nvidia-graphics-drivers-310-updates (310.44-0ubuntu2) from raring-proposed?09:22
Laneyok09:23
tseliotthanks Laney!09:25
Laneywhat was the problem?09:25
tseliotLaney: there wasn't really a problem, I wanted to include more changes09:28
cjwatsonCould somebody please review my debian-installer upload from Friday?  We need to get new kernels in.09:41
Laneyseb128: is that unity-lens-friends upload alright?09:41
stgrabercjwatson: I'll do that now09:42
stgraber(then go back to pretending I'm not around ;))09:42
Laneylibappindicator could do with being looked at too; I can't because it's my change09:42
stgraberLaney: gah, that one is a sync... we really need to teach LP how to diff those09:43
cjwatsonstgraber: thanks09:44
Laneyyeah ...09:44
LaneyI couldn't figure out how to use LP API to grab them even09:44
cjwatsonDamn, I meant to review doko's gcc-4.7 uploads before the weekend :-/09:44
dokoheh, I did use the time do some ppa builds, until the ppa repo didn't accept any more uploads due to size09:45
seb128Laney, not sure about unity-lens-friends, better to check with Ken why social.scope.in.in has been dropped, I see no corresponding commit in the vcs09:46
stgraberLaney: so I think I'm confused about that changelog entry. I clearly see the change to have the library installed in /usr/lib instead of the multi-arch path, but I'm not seeing anything about getting the pcfile into /usr/share/pkgconfig09:47
stgraberLaney: nevermind, I'm blind09:47
stgraber(I somehow missed the 3rd file in the debdiff)09:47
Laneyphew09:47
stgraberaccepted09:47
Laneymerci09:48
stgraberBitte sehr09:49
Laney\o/09:51
Laneyseb128: should the pot have been dropped too?09:52
Laneyit seems to revert your upload09:52
seb128Laney, yes, dh-translations will generate it at build time09:52
Laneykenvandine: ^ I'll reject unity-lens-friends for safety until you check on the .scope.in.in dropping09:53
cjwatson(rss2irc is a build fix, since I think that's unclear from the changelog)09:53
cjwatsoncould somebody push through the latest set of haskell rebuilds?10:07
cjwatsonalso ideally the haskell syncs in NEW from the end of last week10:07
Laneylooking (at the former)10:07
cjwatsonjust did a batch of armhf removals too; getting close ...10:09
Laneydown to the wire10:09
Laneydid soemone else accept rss2irc?10:10
stgrabercjwatson: speaking of removals, can you remove lttng-tools from raring, the binaries have been taken over by ltt-control (synced from Debian)10:10
stgraberLaney: I did10:10
Laneyalright, great10:10
Laneyotherwise it would have been a queue bug10:10
cjwatsonstgraber: done - I reassigned the one open bug too10:16
Laneytseliot: are the nvidia-graphics-drivers-experimental-{304,310} packages going away?10:39
Laneyah, I see the bugs, yes10:41
tseliotLaney: yep11:06
cjwatsonSigh, more dep-wait11:14
* cjwatson syncs haskell-readargs for haskell-basic-prelude11:14
infinitydoko: Are you still checking rdeps for https://bugs.launchpad.net/ubuntu/+source/ecj/+bug/1165915 ?12:06
ubot2Launchpad bug 1165915 in ecj (Ubuntu Raring) "FFe: update ecj to a version supporting Java7" [Undecided,Incomplete]12:06
dokoinfinity, yes, done, now need to fix openjdk :-/ in progress12:06
infinitydoko: Kay, so we're not good to go on the update then?  Can you update the bug to reflect that, and I'll reject the current sync so no one bothers with it until that's sorted?12:07
cjwatson(Note that you can't resurrect rejected syncs; it'd have to be re-requested.)12:10
infinityYeah, not hard to do a new one.12:11
dokoinfinity, just reject12:12
infinitydoko: ^-- Already done.12:12
* infinity hugs sagari's gcc-4.7 build time, and wishes he could will a few more of those machines magically into the DC.12:14
* cjwatson removes the haskell-dummy binaries, and all of washngo12:30
cjwatsonactually I removed the haskell-dummy source too12:30
=== seb128_ is now known as seb128
jpdsPackages in -proposed go to -updates after 7 days of maturity when verification is done?12:40
seb128jpds, correct12:40
jpdsseb128: Parfait, merci.12:41
seb128jpds, it's a manual process though, so tend to not happen between friday and monday12:41
xnoxjpds: 7 days since publishing in -proposed && verification is done. So e.g. something can be verified on day 3 and published in -updates on day 7. No publishing on fridays, saturday, sundays; unless we can agree and assign people to be online to handle regressions.12:48
xnoxbdmurray: ^^12:56
kenvandineLaney, that scope file shouldn't have been there to start with, so it's fine12:58
kenvandineLaney, and it wasn't installed in the previous version12:58
cjwatsonblimey, removing haskell-dummy took about 400KB (of ~29MB) off the uncompressed size of universe/i386 Packages (60KB of 7MB off gzip)13:00
infinitycjwatson: Impressive...13:01
kenvandineLaney, i'm re-uploading13:08
kenvandinei just noticed seb128 had uploaded another change earlier on friday that wasn't included in my upload13:09
Laneyright, k13:09
Laneyjust re-trigger the copy13:09
Laneyor was it a manual upload? if so I can fish it out13:09
seb128kenvandine, yeah, that change was to discard13:09
seb128kenvandine, it's superseeded by dh-translations use13:09
Laneyindeed13:09
kenvandineah, so i didn't need to?13:09
kenvandinegreat13:09
seb128kenvandine, no, it was just a workaround upload to get the template imported13:09
kenvandineLaney, if you can recover it, that would be great :)13:10
seb128kenvandine, I would have included it in the merge request otherwise ;-)13:10
kenvandineok :)13:10
kenvandineLaney, i can re-upload if that is easier13:13
Laneykenvandine: no, it's already done13:16
kenvandineLaney, thanks13:16
debfxvirtualbox in precise-proposed contains an important bugfix that has been verified and two unverified less important fixes13:43
debfxdo you guys have a suggestion on how to get that fix (compatibility with the lts backport kernel) into -updates?13:43
* cjwatson removes ftphs binaries13:44
debfxre-uploading it with only that one fix?13:44
cjwatsondebfx: 1031217 doesn't seem as though it should be hard for somebody with a 12.04 host system to verify.  I would be uncomfortable with dropping the patch for 1071344 - but for that I don't think we need to verify the fix, just regression-test that 64-bit guests still work, right?13:50
debfxcjwatson: I tried to reproduce 1031217 on precise but failed. It's easily reproducible on quantal so maybe you can trigger the bug with a custom networkmanager config on precise.13:58
cjwatsonfailed> as in the new package broke, or you just couldn't reproduce the original problem?13:59
debfxI couldn't reproduce the origin problem on precise13:59
debfx*original14:01
ogra_======================= Log of livefs.sh output follows =======================14:06
ogra_sh: 1: /usr/sbin/ubuntu-touch-android.sh: not found14:06
ogra_hmpf14:06
ogra_argh !14:06
ogra_how embarrasing14:10
cjwatsonlooks easy to fix though?  or do you need a BuildLiveCD change ...14:10
ogra_nah14:10
ogra_already uploaded14:10
ogra_just waiting for the bot ...14:11
ogra_there we go ...14:11
ogra_still something that makes me blush ...14:11
ogra_but it seems that BuildLiveCD at least takes the correct path :) thats something14:12
cjwatson^- another missing build-dep of haskell-project-template; I think this is the last of them ...14:20
infinitydoko: What are the odd of getting http://sourceware.org/bugzilla/show_bug.cgi?id=14887 fixed before release (or, if you're too busy, mind if I roll/test/upload for that tomorrow?)15:07
ubot2sourceware.org bug 14887 in gas "Error: ARM register expected -- `str r1,[ r0 ]'" [Normal,Reopened]15:07
infinitydoko: I'm not sure how many build faiures we have (or have fixed) as a result of that, but I know there's a couple.15:07
dokoinfinity, ok, will fix it today15:08
dokoinfinity, I assume you won't target a glibc update anymore?15:08
infinitydoko: If you have an ARM machine to test on, gmp is a good test-case.15:08
infinitydoko: I'm going to do one last conservative glibc upload tomorrow.15:09
dokoinfinity, no, bus errors for everything now15:09
infinitydoko: Oh, that sucks. :/15:09
dokoxnox, yeah for the arm assembler openssl issue ...15:10
infinitydoko: LP bug for that binutils PR is https://bugs.launchpad.net/binutils/+bug/116662815:10
ubot2Launchpad bug 1166628 in binutils (Ubuntu) "New compile error on ARM" [Undecided,New]15:10
ogra_hrm15:13
* ogra_ wonders if he killed kapok.buildd15:13
xnoxbdmurray: are you accepting openssl/precise as well at the same time? or do you want to publish quantal fully first?15:13
ogra_infinity, hulp ...15:14
bdmurrayxnox: I'm getting to the precise one15:14
infinityogra_: Define killed...15:14
xnoxbdmurray: =)))) ok, thanks.15:14
cjwatsonogra_: still responds to HTTP15:14
ogra_cdimage@nusakan:~$ ssh -n -o StrictHostKeyChecking=no -o BatchMode=yes buildd@kapok.buildd /home/buildd/bin/BuildLiveCD -t android mako15:14
ogra_/home/buildd/bin/BuildLiveCD: fork: Cannot allocate memory15:14
infinityHah.15:14
ogra_and the log is definitely not updated15:14
infinityIs ubuntu-touch-android.sh a DoS? :P15:15
ogra_i'm running this script from BuildLiveCD http://paste.ubuntu.com/5710578/ ... the shebang was wrong so i guess the "set -e" has no effect15:15
cjwatsonNot sure how #! and set -e relate15:16
ogra_you mean it would stop BuildLiveCD ?15:16
cjwatsonAh, but IIRC there is some curious interaction between '.' and set -e15:16
cjwatsonI forget the details15:17
cjwatsonDo you mean that the version that was running was with #! /bin/sh?15:17
ogra_nope15:17
ogra_it was with !#/bin/bash15:17
cjwatsonAh15:17
ogra_flipped the chars :(15:17
cjwatsonYou might want to ask IS to figure out what state it's in15:17
infinitywebops, rather.15:17
cjwatsonEr, yeah, that15:18
ogra_theoretically i would expect set -e to stop it15:18
ogra_though it would probably help if BuildLiveCD had set -e set15:18
ogra_oh, it has15:19
ogra_just quite far down15:19
cjwatsonWell, (a) set -e has no effect on things that precede it, (b) not totally sure what !#/bin/bash would do anyway, but if it does nothing then the worst case is probably that the script barrels on with /bin/sh15:19
cjwatsonAnd in any case set -e in BuildLiveCD is irrelevant here15:19
cjwatsonSince it checks the return code of that command explicitly15:20
ogra_"sh: 1: /usr/sbin/ubuntu-touch-android.sh: not found"15:20
ogra_thats the result of my run15:20
cjwatsonFind out what's actually happening on the machine.  Guessing is too hard work.15:20
ogra_so i would have expected set -e to capture that as error and stop BuildLiveCD15:20
cjwatsonset -e is totally irrelevant to anything here.15:20
cjwatsonThe error is inside 'if $LINUX32 sudo chroot ${DIR%/./*} sh -c "cd /${DIR#*/./} && $COMMAND" >> ${LOG} 2>&1; then', which is an explicit check15:21
cjwatsonSo that should have gone into the else case, i.e. exit 115:21
ogra_oh15:21
cjwatsonset -e only applies to commands whose exit status isn't checked15:21
cjwatsonLike I say, find out what's really happening15:21
ogra_oh, indeed15:22
=== smartboyhw_ is now known as smartboyhw
cjwatsongrr, another haskell-yesod build-dep16:57
cjwatsonLaney: too slow old man :)17:04
Laneyoho17:04
cjwatsontantalisingly close now, I think it's just those two syncs above17:05
cjwatsonI removed the haskell-platform binaries17:06
infinityogra_: Erm, hold on.  You realise that you're doing things in a chroot that doesn't get deleted, right?17:07
infinityogra_: So, you're adding a bunch of cruft that never gets removed.17:07
LaneyTo an extremely well tested Haskell suite in 13.04 *ahem*17:07
cjwatsonHey, it builds.17:08
cjwatsonMostly.17:08
infinityThat's about as tested as it usually is.17:08
Laneyspoken like a true functional programmer17:08
ogra_infinity, hrm, so should i debootstrap my own ?17:08
ogra_i thought that one comes from a tarball freshly unpacked on every build17:09
infinityogra_: No, they definitely don't unpack fresh ones.17:09
ogra_ok17:10
infinityogra_: The other curiosity is, if the build depends on i386 packages, why don't you build it on an i386 buildd?17:10
ogra_i'll debootstrap then17:10
infinityogra_: Instead of focing multiarch on amd64?17:10
infinitys/focing/forcing/17:10
ogra_infinity, because the android tools expect an amd64 builder17:10
infinity...17:10
ogra_yeah17:10
ogra_its all pretty awful17:10
infinityThey expect an amd64 builder and i386 binaries?17:10
ogra_ask rsalveti :)17:10
infinityThat sounds positively insane.17:11
rsalvetiyup, it's indeed17:11
ogra_the tree also ships its own toolchain etc ...17:11
ogra_its all pretty insane17:11
rsalvetiandroid way of doing it17:11
doko?17:11
dokopff17:11
ogra_android :017:12
doko"we only trust our own bugs"17:12
ogra_infinity, so reject that upload, chrooting all that stuff will take a bit of a rewrite17:13
ogra_sigh ...17:13
* ogra_ notes that testing all that stuff locally gained him exactly nothing ... 17:14
=== Ursinha is now known as Ursinha-afk
ogra_i know i'm not supposed to test in production ... but replicating the setup seems pretty impossible17:15
infinityReplicating a livefs builder is easy.17:15
=== Ursinha-afk is now known as Ursinha
infinitydebootstrap a chroot in ~/build-raring-live/chroot-raring/ ; install livecd-rootfs in it, copy BuildLiveCD out to somewhere you can call it, edit BuildLiveCD to not mail logs all over the place, done.17:16
ogra_infinity, well, set up an android mirror somewhere in my LAN to pull the 15G tree from, have a local package mirror as well etc etc17:23
ogra_there is followup work :)17:23
ogra_hmm17:28
ogra_ chroot $builddir bash -c ". build/envsetup.sh && brunch $codename"17:28
ogra_would that work ?17:28
ogra_(to source the needed bits before running the brunch command)17:28
ogra_oh, crap ... sigh ...17:30
ogra_i guess i should just inject a script into the chroot ... that will work better17:31
ogra_since the android bits will likely choke on absolute paths17:31
ogra_infinity, mind taking a look at http://paste.ubuntu.com/5710991/ before i waste another version number ...17:40
ogra_hmm, line 14 could just drop the check i suppose ... since i know already i'm on amd6417:47
infinityogra_: You might want proc and devpts mounted (and unmounted) in that chroot.18:04
infinityogra_: And to start your builds by cleaning out builddir, in case of previous cruft.18:05
infinityogra_: Otherwise, if that passes a local test, go nuts.  I'm not actually here today (and about to be a lot less here), so...18:06
ogra_ok18:06
ogra_ah, i just added proc and sys mounting ... forgot devpts :)18:07
ogra_thanks !18:07
mdeslaurcan I upload a security fix for haproxy?18:16
ScottKmdeslaur: Yes18:16
mdeslaurScottK: thanks18:16
phillwwith the dropping of alternate from all but lubuntu and server, is it still flagging up an issue where sudo /cdrom/cdromupgrade still then attempts to download over the internet instead of using the CD image?20:05
phillw*is still worth flagging up*20:05
cjwatsonIf it exists it should work20:06
cjwatsonThere's the question of whether it's worth keeping that script on the images if it's only on a small number of images and thus not well-tested20:06
cjwatsonBut that doesn't mean it isn't a problem as it stands, based on that report20:08
phillwthanks cjwatson It's been a while since I've used an alternate image. I followed the instructions at http://linuxpoison.blogspot.co.uk/2011/06/how-to-upgrade-ubuntu-using-alternate.html and all went well until It decided to go and download stuff instead of using the ISO image that was mounted.20:09
phillwI'd guess it is more an issue for server, where they may want to do a mass upgrade. I've not tried the update ISO yet.20:10
cjwatsonRealistically I can't imagine server users wanting to upgrade from a CD20:10
cjwatsoncdromupgrade was always a desktop-user use case20:10
cjwatsonIndeed I believe that cdromupgrade is not on the server image20:11
cjwatsonBecause it can't work off the squashfs-base system20:11
phillwcjwatson: that's fine. I'll try using the upgrade iso :) As ever, thank you for your time.20:13
cjwatsonthere's no separate upgrade ISO ...20:14
phillwcjwatson: There are lots of them at http://iso.qa.ubuntu.com/qatracker/milestones/243/builds :)20:15
phillwlubuntu has not yet tested our set, I was just curious as to if the alternate still worked :)20:16
slangasekno, there aren't; those are the tests for upgrading the OS, there is no "upgrade ISO" for any of these20:16
phillwslangasek: sorry, you are entirely correct! there is no ISO!20:17
Laneyghc.html looking nice20:17
phillwHmm, okies. So cannot upgrade using iso. not to worry :)20:17
stgraberLaney: is that a good thing? It felt like the goal with ghc was to have it in some kind of constant transition ;)20:18
slangasekstgraber: "looking nice" is subjective, of course; he didn't say it was looking empty ;)20:19
Laneytrue, I'd better upload a new ghc :-)20:19
LaneyAlso looks likely that the autohinter will manage to transition it20:20
stgraberslangasek: oh indeed ;)20:20
cjwatsonLaney: Yep, it should do it this run20:20
cjwatsonJust waiting for the publisher20:20
* Laney is impressed20:20
cjwatsonIf you don't want that you'd better block it quick ;-)20:21
Laneyand slightly sad that we don't get to construct a monster hint20:21
cjwatsonI think I can do without that20:21
Laneyseems that the gitit workaround didn't do the trick though20:21
Laneyho hum20:21
cjwatsonLaney: That's just because I'm an idiot - I'm uploading a fixed version to experimental as I type20:22
cjwatsonForgot that $? is spelled $$? in make20:22
Laneyoh, nice20:22
cjwatsonfinal: agda,agda-stdlib,cpphs,darcs,ghc,[...]20:41
cjwatsonCLANG20:41
cjwatsonSUCCESS (552/0)20:41
dokoa beer for every migrated package \o/20:43
Laneyglorious20:43
cjwatsonNot sure Laney's and my liver combined can handle 552 beers20:44
LaneyMight prevent us having to do this ever again if we try20:44
* cjwatson watches ackee try to keep up with the copies20:45
cjwatsonEven promote-to-release hasn't finished yet, and it's asynchronous!20:49
dokointeresting times for debian unfreezing ...20:49
* Laney starts receiving emails20:54
cjwatsonAnd I think it's done20:55
balloonsmaybe it's better asked in here.. I'm helping get ubuntu touch added to the isotracker. So you should see a new family 'ubuntu touch' and products being listed.20:56
balloonsthis doesn't affect the raring manifest, but I did want to make sure we got or have a product owner lined up for these20:57
cjwatsonShockingly, this publisher run might take a while21:07
skellatScottK: Mark appears to have spoken: https://bugs.launchpad.net/ubuntu/+source/unity/+bug/116923821:16
ubot2Launchpad bug 1169238 in Unity "[UIFe] BFB icon swirl should run clockwise not anti-clockwise" [Undecided,Triaged]21:16
phillwskellat: why was that a surprise?21:26
phillwI actually agree with him (and that is a rare comment from me :P )21:27
sarnoldhello; in the course of preparing a security update for curl for our supported distributions, I have also prepared a new package for raring, containing the back-ported security fix from upstream.21:50
sarnoldis it okay to push this package to raring?21:50
Laneyyes21:53
sarnoldthanks Laney21:53
Laneynp!21:53
cjwatson... and update_excuses is readable again, yay22:00
Laneysurprised I only got 21 copies apparently22:00
* Laney attempts to bootstrap a codesearch instance22:01
doko$ wc -l update_output.txt22:07
doko22 update_output.txt22:07
dokomy bad ...23:39

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!