[00:00:46] <rcui> i'm shadowing adlr for build sheriff
[00:00:53] <rcui> dhendrix, can we revert the change?
[00:01:33] <dhendrix> Yeah, let's go ahead and revert for now. Playing with it more, it seems the thing that fixed the ARM build broke the x86 build. Ugh.
[00:03:02] <dianders> autotest-tests appears to still be broken on ARM builds.  Mark Hayter is also seeing this.  Looking into it unless someone else has an answer...
[00:03:15] *** JasonO has joined #chromium-os
[00:04:08] <ferringb> probably a dumb question, but is there a non preflight queue where folks can beat on the builders/testers?
[00:05:00] <ferringb> my understanding, a change passes through gerrit, hits the preflight- if it's failing in preflight/tests and folks can't replicate it, just wondering if there is a set of builders/different pathway they can submit through for trying to run it down
[00:05:01] <kliegs> ferringb: there aren't any try-bots currently.  its being worked on I believe however
[00:05:16] <ferringb> yeah, try bot rings a bell
[00:05:49] <kliegs> ferringb: There are some people who have their own test buildbots running that they can work on, but those are mostly people working on build/release
[00:06:02] <rcui> yea I'm working on local trybots
[00:06:23] <rcui> it's close to being ready...
[00:06:37] <kliegs> rcui: ooh nice. knew you were working on it just had no idea on eta
[00:07:07] <kliegs> will be really nice to have those up and running. could really help get better coverage testing pre-submit as well so we get more greenness
[00:07:28] <rcui> hopefully this week.  i have build sheriff duty tomorrow and thur
[00:15:34] <dhendrix> okee, there's the new ebuild hack in the buildbot. let's hope it works better than the last one...
[00:15:40] <ellyjones> w 2
[00:15:42] <ellyjones> fail
[00:17:30] <ferringb> rcui: what's shadowing the sheriff exactly entail btw?
[00:19:32] *** Adys has joined #chromium-os
[00:22:12] <rcui> ferringb: looking at the problem at the same time as the sheriff. learning how to check logs, closing/opening the tree, etc.
[00:24:32] <ferringb> frequency of folks being sheriff?
[00:24:50] * ferringb needs to get in some shadowing one of these days, although will likely be remote/wfh thurs
[00:29:23] *** behdad has quit IRC
[00:30:07] <rcui> every couple of months?
[00:31:28] <njw> mmm, 0 for 2 today.
[00:33:56] *** micahc has joined #chromium-os
[00:33:56] *** ChanServ sets mode: +v micahc
[00:35:04] *** leavittx has quit IRC
[00:36:35] <quannnum> is there a doc for mirroring chromium.org's portage_binhost? http://commondatastorage.googleapis.com/chromeos-prebuilt/  doesn't look 'wget -m'able
[00:37:26] <quannnum> or is this not supported?
[00:45:23] *** Adys has quit IRC
[00:51:43] *** stalled has quit IRC
[00:57:23] *** Trzaska has quit IRC
[00:57:43] *** Trzaska has joined #chromium-os
[00:58:23] <cmasone> which builder(s) are we still waiting on?  x86 pfq?
[01:03:03] <adlr> sorry i stepped out. i'm glad rcui was able to help
[01:03:46] <cmasone> adlr: isn't there a backup sheriff as well, not just the shadow?
[01:03:49] <adlr> i guess the autotest problem wasn't llvm/mesa, after all
[01:03:55] <adlr> cmasone: yeah, he had to leave early today
[01:05:06] *** sadrul has quit IRC
[01:08:55] *** stalled has joined #chromium-os
[01:11:33] <rcui> davidjames was a big help too :)
[01:11:43] <rcui> yea we are waiting for x86 to go green
[01:11:48] <rcui> looks good
[01:11:56] <adlr> rcui: you're a natural :)
[01:17:03] <rcui> i'm doing this tomorrow too, so good practice
[01:35:21] *** macjunkie has joined #chromium-os
[01:36:26] *** macjunkie has left #chromium-os
[01:43:54] *** rbyers has joined #chromium-os
[01:58:00] *** rbyers has quit IRC
[01:58:09] *** rbyers has joined #chromium-os
[01:58:10] *** ChanServ sets mode: +v rbyers
[02:01:58] *** sadrul has joined #chromium-os
[02:01:58] *** ChanServ sets mode: +v sadrul
[02:06:46] *** lipsinV2 has joined #chromium-os
[02:07:51] *** wfrichar has quit IRC
[14:10:13] *** echelog-1 has joined #chromium-os
[14:19:06] *** stalled has quit IRC
[14:29:42] *** e_t_ has quit IRC
[14:32:04] *** stalled has joined #chromium-os
[14:33:45] *** flackr has joined #chromium-os
[14:33:45] *** ChanServ sets mode: +v flackr
[15:08:51] *** leavittx has joined #chromium-os
[15:12:37] *** sadrul has joined #chromium-os
[15:12:37] *** ChanServ sets mode: +v sadrul
[15:16:04] *** Adys has joined #chromium-os
[15:23:06] *** rbyers has joined #chromium-os
[15:35:19] *** CrashTest_ has joined #chromium-os
[15:48:40] *** leavittx has quit IRC
[15:51:12] *** leavittx has joined #chromium-os
[15:58:26] *** phil2 has quit IRC
[16:08:17] *** Adys_ has joined #chromium-os
[16:10:42] *** Adys has quit IRC
[16:34:56] *** Adys_ is now known as Adys
[16:44:40] *** rbyers__ has joined #chromium-os
[16:44:40] *** rbyers has quit IRC
[16:44:40] *** rbyers__ is now known as rbyers
[17:18:50] *** CrashTest_ is now known as CrashAFK
[17:21:56] *** jrbarnette has joined #chromium-os
[17:35:58] *** jrbarnette_ has joined #chromium-os
[17:35:59] *** benchan has quit IRC
[17:36:42] *** dmazzoni has quit IRC
[17:37:22] *** saintlou has joined #chromium-os
[17:37:22] *** ChanServ sets mode: +v saintlou
[17:38:32] *** jrbarnette has quit IRC
[17:38:32] *** jrbarnette_ is now known as jrbarnette
[17:38:45] *** wfrichar has joined #chromium-os
[17:38:45] *** ChanServ sets mode: +v wfrichar
[17:42:17] *** Adys has quit IRC
[17:54:07] *** Styx has quit IRC
[17:55:39] *** jujugre has left #chromium-os
[17:55:48] *** borios has joined #chromium-os
[18:01:13] <borios> how can I make branch? 'repo start [name] . ' doesn't work :(
[18:02:55] <cmasone> borios: "doesn't work" isn't really enough info.
[18:03:21] <borios> repo start a .
[18:03:21] <borios> error: project . not found
[18:04:01] <cmasone> are you in a git checkout?
[18:04:54] <borios> umm.. I don't know . I just 'repo sync'
[18:05:30] <cmasone> what project are you trying to create a branch in, then?
[18:07:33] *** behdad has joined #chromium-os
[18:07:48] <borios> I want create a branch and edit some script in 'src/script'
[18:08:08] <cmasone> did you cd to src/script before trying to repo start?
[18:08:46] <borios> oh..
[18:08:50] <borios> Thanks
[18:14:36] *** lipsinV2 has quit IRC
[18:34:30] *** Trzaska has quit IRC
[18:47:07] *** behdad has quit IRC
[19:01:34] *** perezd has joined #chromium-os
[19:02:18] *** behdad has joined #chromium-os
[19:02:19] *** ChanServ sets mode: +v behdad
[19:17:24] *** jrbarnette_ has joined #chromium-os
[19:20:46] *** jrbarnette has quit IRC
[19:21:47] *** jrbarnette_ has quit IRC
[19:23:06] *** jshin has joined #chromium-os
[19:23:06] *** ChanServ sets mode: +v jshin
[19:24:41] *** CrashAFK is now known as CrashTest_
[19:34:02] *** chocobo__ has joined #chromium-os
[19:34:02] *** ChanServ sets mode: +v chocobo__
[19:34:36] *** jrbarnette has joined #chromium-os
[19:34:36] *** ChanServ sets mode: +v jrbarnette
[19:35:42] *** Trzaska has joined #chromium-os
[19:52:54] *** sergiu has joined #chromium-os
[19:59:31] <rochberg_> Anyone around who can help me sort some gerrit ACL issues?
[20:00:33] *** CrashTest_ has left #chromium-os
[20:05:30] *** TW1920_ has joined #chromium-os
[20:08:11] *** TW1920 has quit IRC
[20:14:48] <mononoke> Can't build ARM: emerge: there are no ebuilds to satisfy ">=x11-libs/libxcb-1.6" for /build/tegra2_kaen/.
[20:15:01] <mononoke> I just did a repo sync.
[20:18:05] <davidjames> mononoke: Is that in build_packages or build_image ?
[20:20:23] <davidjames> mononoke: build_packages needs to be run after repo sync
[20:27:29] <mononoke> davidjames: build_packages
[20:33:26] *** rcui_ has joined #chromium-os
[20:33:26] *** ChanServ sets mode: +v rcui_
[20:36:19] <rcui_> Looks like package merge failure due to file conflict
[20:38:40] <cmasone> rcui_: that package built successfully on the next cycle...damned if I know why :-/
[20:39:05] <cmasone> kliegs: was this chromeos-bootimage change the thing you were talking about yesterday?
[20:40:44] <cmasone> rcui_: oh, I was looking at the arm tegra2 full failure.  Ignore me
[20:41:24] <cmasone> rcui_: are you sheriffing alone today?
[20:41:51] <chocobo__> i'm here
[20:41:58] *** behdad has quit IRC
[20:42:10] <rcui> cmasone: yea chocobo is sheriff too
[20:42:23] <chocobo__> this is the error:
[20:42:34] <chocobo__> Traceback (most recent call last):
[20:42:35] <chocobo__>   File "/b/build/slave/arm-tegra2_aebl-private-bin/build/chromite/buildbot/cbuildbot_stages.py", line 424, in Run
[20:42:35] <chocobo__>     self._PerformStage()
[20:42:36] <chocobo__>   File "/b/build/slave/arm-tegra2_aebl-private-bin/build/chromite/buildbot/cbuildbot_stages.py", line 763, in _PerformStage
[20:42:36] <chocobo__>     extra_env=env)
[20:42:36] <chocobo__>   File "/b/build/slave/arm-tegra2_aebl-private-bin/build/chromite/buildbot/cbuildbot_commands.py", line 233, in Build
[20:42:36] <chocobo__>     cros_lib.RunCommand(cmd, cwd=cwd, enter_chroot=True, extra_env=env)
[20:42:37] <chocobo__>   File "/b/build/slave/arm-tegra2_aebl-private-bin/build/chromite/lib/cros_build_lib.py", line 154, in RunCommand
[20:42:37] <chocobo__>     raise RunCommandError(msg, cmd, proc.returncode)
[20:42:37] <chocobo__> RunCommandError: Command "['./enter_chroot.sh', 'EXTRA_BOARD_FLAGS=--rebuilt-binaries', '--', './build_packages', '--board=tegra2_aebl', '--fast', '--nowithautotest']" with extra env {'EXTRA_BOARD_FLAGS': '--rebuilt-binaries'}
[20:43:19] <cmasone> chocobo__: that's not a meaningful error, though.  That just says that build_packages failed.  It sounds like rcui has chased it down to a light-sensor file collision, something there's a thread about on chromiumos-dev right now
[20:43:48] <cmasone> chocobo__: rcui I think that's bfreed's thing
[20:43:53] <rcui> where is the light-sensor ebuild located?
[20:44:02] <cmasone> rcui: no idea.  talk to bryan
[20:44:39] <chocobo__> it's weird that the blame list only has an unrelated change
[20:44:41] <kliegs> cmasone: The change of mine that would break the build?  Or the change that dianders was fixing?
[20:44:54] <cmasone> kliegs: the former.  It sounds like it's not
[20:45:11] <kliegs> nope. my change was just removing a file from manifest-internal.
[20:45:17] <rcui> chcobo: yea looks right, bfreed
[20:45:42] <kliegs> cmasone: I thought the buildimage was fixed yesterday.  sheriffs reported it fixed and i saw a tegra2 buildbot go green.  clchiou broke it this morning again and said he was working on a fix when I last talked to him
[20:45:47] <kliegs> but I haven't been actively tracking
[20:46:01] <cmasone> kliegs: no, I think you're fine.  this looks like something else.
[20:46:30] <kliegs> cmasone: I didn't check my change in.  tree's been red every time I look
[20:46:40] <kliegs> so i can still threaten sheriffs with future redness
[20:46:45] <cmasone> kliegs: ha
[20:46:57] <kliegs> I feel like ellyjones
[20:49:18] <ellyjones> weirdest hilight ever
[20:50:28] <kliegs> :)
[20:50:48] *** bfreed has joined #chromium-os
[20:50:48] *** ChanServ sets mode: +v bfreed
[20:54:53] <cmasone> updates?  Is bfreed here to help investigate?  is someone talking to the infrastructure guys to try and figure this out?
[20:55:11] *** benchan has joined #chromium-os
[20:55:12] *** ChanServ sets mode: +v benchan
[20:56:06] *** behdad has joined #chromium-os
[20:56:06] *** ChanServ sets mode: +v behdad
[20:58:26] <adlr> bfreed is looking
[20:58:35] <adlr> he and rcui are talking
[20:58:51] <adlr> he said he committed a lot of stuff, in (he thinks) the proper order
[20:59:00] <adlr> but some CLs may have landed in a different order (?)
[20:59:12] <rcui> the cl's landed close to the buildbot sync stage
[20:59:15] <adlr> and a poorly-timed build thus would have broken
[20:59:29] <rcui> so some changes maybe didn't make it
[21:06:04] *** behdad has quit IRC
[21:10:38] *** powderluv_ has joined #chromium-os
[21:10:38] *** ChanServ sets mode: +v powderluv_
[21:19:04] *** Justasic has quit IRC
[21:20:25] *** Justasic has joined #chromium-os
[21:23:31] <bfreed> The preflight queue (PFQ) is a double edged sword.
[21:24:56] <rcui> so issue was that bfreed's light-sensor change depended on an earlier powermanager change by him.  But since internal binaries don't uprev, he needed to wait for the PFQ to make his change live.
[21:25:27] <rcui> *wait for PFQ to make his powermanager change live before checking in light-sensor changes.
[21:29:53] <bfreed> And then make the light sensor ebuild depend on that particular revision of the power_manager.
[21:32:21] <rcui> hmm more package file conflicts....
[21:33:49] *** rochberg_ has quit IRC
[21:36:26] *** patcito has joined #chromium-os
[21:36:45] <cmasone> bfreed: the PFQ isn't a double-edged sword, so much as the mechanism that is intended to protect us against issues like this...it sounds like waiting for power_manager to uprev and then setting the dependencies in the light-sensor ebuild correctly would have worked, right?
[21:36:54] <bfreed> rcui: Yeah, those are mine, too.  :-(
[21:37:16] <bfreed> cmasone: right.
[21:37:57] <cmasone> bfreed: where is this light-sensor ebuild, anyhow?
[21:40:04] <bfreed> cmasone: The light-sensor ebuilds are in various private overlays/chromeos-base/light-sensor.
[21:41:18] <cmasone> bfreed: so it's a package we authored here?
[21:41:45] <rcui> i restarted the build, perhaps the bsp ebuild missed the uprev train
[21:43:24] <bfreed> rcui: My best guess is the alex failure can be fixed with a similar negative dependancy of chromeos-bsp-alex on "!board-devices".
[21:52:35] <ellyjones> hm, lack of policykit breaks enewgroup/enewuser... is that intentional?
[21:52:39] <ellyjones> or consolekit or whatever
[21:55:08] <sosa> rcui:  you can check the commits in the manifest printed out in the sync stage to verify
[21:55:09] <sosa> the manifest don't lie
[21:58:49] <rcui> sosa: it's a different problem than before
[21:59:03] <sosa> (12:45:50) rcui: i restarted the build, perhaps the bsp ebuild missed the uprev train
[21:59:27] <rcui> this time bfreed manually upreved the package
[21:59:27] <sosa> missed the uprev train meaning it missed hte uprev commit?  the uprev commit refers to some git commit which can be checked in the manifest
[22:00:50] <powderluv_> ellyjones: what prompted the breakage
[22:02:50] <ellyjones> trying to install a package that uses 'enewuser' onto my target
[22:02:59] <ellyjones> it complains about missing getent and then the install bails
[22:03:20] <cmasone> getent was installed by that package??
[22:03:39] <ellyjones> the package was trying to use enewuser, and enewuser tries to use getent
[22:03:41] <cmasone> I remember at some point AAAAAAAGES ago discovering that we didn't have getent on the image
[22:03:45] <ellyjones> getent was installed by some random package
[22:04:08] <cmasone> ellyjones: yes, it's supposed to be installed by the toolchain or something, but it never has been
[22:04:27] <ellyjones> ok, that seems like a bug?
[22:04:31] <cmasone> ellyjones: and so, at some point in like...2009, I added a script that implemented it and installed it in usr/bin
[22:04:56] <ellyjones> that seems to be gone now though
[22:05:08] <cmasone> ellyjones: yes, it was bug 1570
[22:05:10] <rcui> sque is taking a look at the power manager failures on x86-generic-pfq family
[22:05:14] <cmasone> http://code.google.com/p/chromium-os/issues/detail?id=1570
[22:05:21] <cmasone> it was supposedly fixed in may
[22:05:58] <ellyjones> "until the next compiler roll-out"
[22:06:03] <ellyjones> when's that?
[22:06:07] <cmasone> ellyjones: dunno :-/
[22:06:11] <cmasone> davidjames: ping ^^^
[22:06:13] <ellyjones> oh :\
[22:09:06] *** sergiu is now known as ergiu
[22:09:09] *** ergiu is now known as Sergiu
[22:13:01] <powderluv_> raymes/asharif may know
[22:13:20] <powderluv_> ofcourse they aren't on here :)
[22:20:24] *** e_t_ has joined #chromium-os
[22:26:37] *** Trzaska has quit IRC
[22:26:59] <gauravsh> didn't asharif send out a recent email to cros-dev about testing out the new gcc 4.6 based cross-compiler
[22:27:28] <gauravsh> he did say that it was going to be the default "soon"...
[22:28:01] <ellyjones> yeah, and I wish I knew what 'soon' meant
[22:30:36] *** tfarina has joined #chromium-os
[22:30:49] *** micahc has quit IRC
[22:31:09] <tfarina> is derat around?
[22:33:35] <rcui> seeing pci_add_option_rom: failed to find romfile "pxe-virtio.bin" in the x86-pineview test log
[22:33:40] <rcui> "pci_add_option_rom: failed to find romfile "pxe-virtio.bin""
[22:33:47] <rcui> any idea what's going on?
[22:33:53] *** tfarina has left #chromium-os
[22:36:17] <cmasone> rcui: irrelevant error output from VMs
[22:41:17] <cmasone> Once this tree breakage gets sorted out, can we get a post-mortem, so that we can figure out what we can fix/change so that this kind of problem doesn't trip us up again?
[22:41:30] <davidjames> cmasone, ellyjones: Yeah we should have getent by now
[22:42:00] <cmasone> davidjames: we don't :-/  At least, not on the build I did yesterday.  I did modify for test, and maybe that horked it
[22:42:44] <davidjames> cmasone: Hmm, no, I'm not seeing getent either on my machine, I'll check and see what's going on
[22:45:06] <ellyjones> danke
[22:48:26] *** CrashTest_ has joined #chromium-os
[22:49:14] *** CrashTest_ has left #chromium-os
[22:51:01] <chocobo__> some bots are failing on mesa-7.11_rc1. stephane says we need to clobber those bots
[22:51:04] <chocobo__> i will do them now
[22:53:12] <mononoke> From a fresh repo sync:   fatal: pack has 1370 unresolved deltas --- fatal: index-pack failed --- error: Cannot fetch chromiumos/overlays/portage
[22:53:28] <mononoke> Every time I take a step back, something else is failing.
[22:53:29] <sosa> why?
[22:53:35] <sosa> chocobo?
[22:53:49] <sosa> does this means devs also have to clobber their existing builds?
[22:55:27] <chocobo__> stephane said you only have to do it once
[22:55:48] <davidjames> chocobo__: Full builders are already building from source
[22:55:56] <davidjames> chocobo__: So clobber won't make a difference
[22:56:16] <davidjames> chocobo__: Same for chromiumos-sdk -- it's a full build from scratch, already rm -rf's the chroot prior to building
[22:56:20] <marcheu> davidjames: hmm interesting, why does it fail now though?
[22:56:33] <marcheu> I pushed this like a week ago
[22:56:40] <davidjames> marcheu: Yeah it wouldn't be a change from last week
[22:56:49] <marcheu> I can't see another reason
[22:56:55] <davidjames> marcheu: Something changed today, chromiumos-sdk was working earlier today and so did other full builders
[22:57:16] <marcheu> where is the log for the mesa build faile?
[22:57:28] <davidjames> marcheu: http://build.chromium.org/p/chromiumos/builders/chromiumos%20sdk/builds/220/steps/BuildBoard/logs/stdio
[22:57:57] <marcheu> oh right that has nothing to do with the issue that requires clobber
[22:58:10] <marcheu> I take back everything I said about needing to clobber :)
[22:58:52] <marcheu> missing stddef is bothering though :)
[22:59:15] <davidjames> So at 10:30am today the build was working, and at 11:57am the build was broken, so something changed in between those two times
[22:59:52] <chocobo__> ok, will investigate mesa failure
[23:11:36] *** leavittx has quit IRC
[23:15:03] <rcui> so the x86-pineview image crashes when i try to load it in kvm
[23:15:30] <ellyjones> oh, gr, you can't use 'fowners' in ebuilds because target uids != host uids
[23:31:48] <rcui> mesa failure - No rule to make target `/usr/lib/gcc/x86_64-redhat-linux/4.6.0/include/stddef.h', needed by `main/api_arrayelt.o'.  Stop.
[23:31:58] <rcui> why is it looking for redhad linux headers?
[23:33:07] *** oc80z has quit IRC
[23:33:07] *** oc80z has joined #chromium-os
[23:33:53] *** oc80z has joined #chromium-os
[23:36:21] *** micahc has joined #chromium-os
[23:36:22] *** ChanServ sets mode: +v micahc
[23:36:52] *** rbyers has quit IRC
[23:38:10] *** arun__ has quit IRC
[23:38:10] *** arun__ has joined #chromium-os
[23:38:53] *** arun__ is now known as arun_
[23:44:38] <rcui> marcheu: i was able to reproduce the mesa failure in my environment.  Could you try to see if you can reproduce it as well?
[23:45:00] <rcui> it's on the host
[23:45:05] <rcui> run 'emerge -1 mesa'
[23:45:08] <marcheu> it's not mesa though, it's a header update from the compiler
[23:45:16] <nirnimesh> So if you rebase and reupload to gerrit after getting an LGTM, are you supposed to ask for LGTM again?
[23:46:56] <davidjames> marcheu: When was compiler / flags changed?
[23:47:28] <marcheu> no idea
[23:48:11] *** tfarina has joined #chromium-os
[23:48:20] <marcheu> davidjames: so it looks like mesa doesn't build with gcc 4.6, so whoever updated to 4.6 didn't test
[23:50:27] <marcheu> davidjames: should we roll back to the previous gcc in the meantime?
[23:50:31] <davidjames> marcheu: Did we update to gcc 4.6?
[23:51:02] <marcheu> let me see
[23:51:02] <davidjames> marcheu: SDK builder seems like it's using gcc 4.3.4-r4
[23:51:15] <marcheu> then why does it say /usr/lib/gcc/x86_64-redhat-linux/4.6.0/include/stddef.h
[23:54:20] <davidjames> marcheu: No idea, but rcui reproduced same problem without changing gcc
[23:54:20] *** tfarina has quit IRC
[23:54:36] <marcheu> and, why wasn't it an issue until today?
[23:55:06] <davidjames> marcheu: That's what we're trying to figure out :)
[23:55:23] <marcheu> for what it's worth I'm building mesa fine here, but I never tried 4.6.0
[23:55:45] <davidjames> marcheu: You're building on the host? It fails for me
[23:56:00] <marcheu> hmm host, lets see
[23:59:55] <davidjames> marcheu: There's a depend file in /var/tmp/portage/media-libs/mesa-7.11_rc1/work/Mesa-7.11-rc1/src/mesa that has references to 4.6.0 everywhere. I think that depend file might be coming from the tarball we downloaded





top