[00:00:08] <redpig> cmasone: perfect
[00:00:12] <ellyjones> redpig: Q3 OKR: replace cmasone with a perl script.
[00:07:05] <rhashimoto> Sorry, I'm over my head and kmixter is in a meeting.  Should I reopen since the problem seems to be in Chrome?
[00:07:28] <rhashimoto> The Chrome ARM build seems green.
[00:07:43] <cmasone> sorry, what's the problem?
[00:08:05] <rhashimoto> tegra2_aebl canary fails building Chrome.
[00:08:19] <cmasone> rhashimoto: well, arm generic just failed to build the kernel
[00:08:20] <rhashimoto> Compile error.
[00:08:26] <cmasone> redpig: was that you?
[00:09:31] <cmasone> rhashimoto: I don't think we can re-open
[00:09:41] <rhashimoto> Right, not if the kernel is broken.
[00:09:49] <cmasone> rhashimoto: yes, also that :-)
[00:10:58] <gauravsh> chromeos-kernel-0.0.1-r440: /build/arm-generic/tmp/portage/sys-kernel/chromeos-kernel-0.0.1-r440/work/chromeos-kernel-0.0.1/kernel/seccomp.c: In function '__secure_computing':
[00:11:01] <gauravsh> chromeos-kernel-0.0.1-r440: /build/arm-generic/tmp/portage/sys-kernel/chromeos-kernel-0.0.1-r440/work/chromeos-kernel-0.0.1/kernel/seccomp.c:55: error: 'NR_syscalls' undeclared (first use in this function)
[00:11:05] <gauravsh> chromeos-kernel-0.0.1-r440: /build/arm-generic/tmp/portage/sys-kernel/chromeos-kernel-0.0.1-r440/work/chromeos-kernel-0.0.1/kernel/seccomp.c:55: error: (Each undeclared identifier is reported only once
[00:11:09] <gauravsh> chromeos-kernel-0.0.1-r440: /build/arm-generic/tmp/portage/sys-kernel/chromeos-kernel-0.0.1-r440/work/chromeos-kernel-0.0.1/kernel/seccomp.c:55: error: for each function it appears in.)
[00:11:14] <gauravsh> oops
[00:11:22] <gauravsh> redpig: why u no declare NR_syscalls?
[00:11:26] *** jrbarnette has quit IRC
[00:11:43] <ellyjones> y u no, &c
[00:12:00] <ellyjones> also, I like __secure_computing as a function name
[00:13:01] <crosbot> tree became 'Tree is closed (drewry investigating "tegra2_aebl canary")'
[00:13:29] <ellyjones> Breaking the tree: fifty DKP minus.
[00:14:53] <cmasone> ok, there's 2 failures here
[00:14:58] <cmasone> the chrome thing and the kernel thing.
[00:15:12] <cmasone> and, of course, the chrome thing is in the PDF plug in
[00:15:13] <rhashimoto> drewry@ is on the kernel.
[00:15:17] <cmasone> yup
[00:15:19] *** sbyer has joined #chromium-os
[00:15:19] *** ChanServ sets mode: +v sbyer
[00:17:37] *** nick761_ has quit IRC
[00:17:58] *** nick761 has joined #chromium-os
[00:18:22] <crosbot> tree became 'Tree is closed (revert in for arm generic binary, still closed for chrome compile error in tegra2_aebl canary)'
[00:18:24] <sosa> rhashimoto: can you roll back chrome until we figure out the pdf issue?
[00:18:46] <redpig> rhashimoto: just to confirm - rhashimoto is rolling back the config change, then I'll ladn the fix and re-enable :(
[00:19:14] <rhashimoto> sosa:  I will try.  The instructions start with "This process is nightmarish." :)
[00:19:24] <sosa> rhashimoto: it is harsh
[00:19:38] <sosa> rhashimoto: but i'm happy to review your change
[00:19:56] <sosa> rhashimoto: i went through it the other day and it wasn't too terribly bad though
[00:22:57] *** patcito has quit IRC
[00:22:58] *** patcito_ has joined #chromium-os
[00:24:24] *** patcito_ is now known as patcito
[00:24:39] *** sadrul has quit IRC
[00:31:11] *** powderluv has joined #chromium-os
[00:31:11] *** ChanServ sets mode: +v powderluv
[00:33:23] <cmasone> rhashimoto: ihf is investigating the arm * canary failure
[00:33:55] <crosbot> tree became 'Tree is closed (revert in for arm generic binary, rhashimoto backing up chrome version for tegra2_aebl canary, ihf investigating failure)'
[00:33:58] <marcheu> that thing looks like a gcc bug
[00:34:27] <marcheu> unless there's assembly in the file, it shouldn't fail at satisfying constraints
[00:34:39] <cmasone> marcheu: it's possible; I know anush landed some kind of toolchain thing earlier today
[00:34:48] <marcheu> right, that's what I had in mind
[00:35:00] <marcheu> maybe that newer gcc is buggy for that specific stuff
[00:35:35] <marcheu> but then I didn't look at the code, just the message
[00:39:11] <rhashimoto> sosa: Where it says x86-generic-LATEST_RELEASE_CHROME_BINHOST.conf, is x86 what I want or should it be arm?
[00:39:20] <cmasone> marcheu: the toolchain thing was pnly x86_64, so we're probably OK there
[00:39:42] <cmasone> marcheu: ah, there is some asm in that file.  ihf is on it
[00:39:53] <sosa> just follow the binhost instrs as is since the change will affect all platforms
[00:40:00] <marcheu> ah ok. usually it's enough to toy with the constraints until they work, but...
[00:40:09] <marcheu> it'll eventually come back
[00:40:37] <sosa> sorry my comment was meant for rhashimoto
[00:46:40] <rhashimoto> sosa: Uploaded and assigned you as reviewer.  Is it just that one file?  What about those other files I checkout with the git hash?
[00:47:50] <rhashimoto> http://gerrit.chromium.org/gerrit/#change,5258
[00:51:44] *** dospod has joined #chromium-os
[00:52:44] <adlr> anyone else getting connection timeouts to gerrit during repo upload?
[00:53:23] <adlr> there it goes
[00:56:15] <crosbot> tree became '?'
[00:56:30] <crosbot> tree became 'Tree is closed (revert in for arm generic binary, rhashimoto backing up chrome version for tegra2_aebl canary, ihf investigating failure)'
[00:57:51] <sosa> rhashimoto:  i don't see the revert of the chrome ebuild
[00:59:14] <rhashimoto> Or the .conf file, right?  How do I commit them?  I did a git add on each one but git commit didn't pick them up.
[01:00:41] <rhashimoto> I'll try again.
[01:09:27] <sosa> rhashimoto:  this part in the instructions, "Now you must retrieve the ebuild for this Chrome version from before the offending push:"
[01:09:51] <sosa> rhashimoto:  basically you're masking all the new ones as unstable, but you need to restore the old one at the same time.
[01:13:07] <rhashimoto> sosa: I've done that, the file chromeos-chrome-15.0.843.0_rc-r1.ebuild is there, but git doesn't want to commit it even if I explicitly git add it.
[01:14:04] <sosa> git status?
[01:14:07] *** jrbarnette has joined #chromium-os
[01:14:07] *** ChanServ sets mode: +v jrbarnette
[01:14:21] <rhashimoto> sosa: just the package.keywords file.
[01:14:39] <sosa> rhashimoto: are you actually at TOT on that repo?
[01:14:54] <sosa> git HEAD^?
[01:15:32] <sosa> oh wait
[01:15:38] <sosa> we are already on 843
[01:15:43] <sosa> you need to go to 842
[01:16:25] <sosa> "First determine what git push brought in the broken version of Chrome"
[01:16:40] *** sadrul has joined #chromium-os
[01:16:40] *** ChanServ sets mode: +v sadrul
[01:16:40] <rhashimoto> git log -p cros/master -M .
[01:16:40] <rhashimoto> commit f1390bf242811d76d4266f44164112c2c135c962
[01:16:41] <rhashimoto> Author: chrome-bot <chrome-bot at chromium dot org>
[01:16:41] <rhashimoto> Date:   Wed Aug 3 12:58:39 2011 -0700
[01:16:41] <rhashimoto>     Marking set of ebuilds as stable
[01:16:41] <rhashimoto>
[01:16:41] <rhashimoto>     Update TOT_CHROME_BINHOST=http://commondatastorage.googleapis.com/chromeos-prebuilt/board/x86-generic/chrome-03.08.11.124537/packages/ in x86-generic-TOT_CHROME_BINHOST.conf
[01:16:42] <rhashimoto>
[01:16:42] <rhashimoto>     Change-Id: Id238a4a0402a0e61765c97f3047a5e8499ec2826
[01:16:43] <rhashimoto>
[01:16:43] <rhashimoto>     Marking tot for chrome ebuild with version 15.0.844.0 as stable.
[01:17:08] <rhashimoto> Whoops, spammy paste.  But it says 15.0.844.0 is stable.  So we should want 843, right?
[01:18:44] <sosa> you're looking at TOT
[01:18:50] <sosa> it's this one http://git.chromium.org/gitweb/?p=chromiumos/overlays/chromiumos-overlay.git;a=commitdiff;h=c8bbdaf22dcb14460d493d9aee39b292382b8ffa
[01:19:11] <rhashimoto> Argh.  Okay, sorry about that.
[01:19:14] <sosa> we don't use TOT chromeos-chrome ebuilds for developers or other builders
[01:19:20] <sosa> no it's my fault, that wasn't clear
[01:19:53] <sosa> i'll add a note in the sheriff FAQ
[01:20:41] <sosa> (or maybe cmasone can since he has the lock for that page ;))
[01:25:49] <rhashimoto> sosa: http://gerrit.chromium.org/gerrit/#change,5260
[01:26:56] <sosa> rhashimoto:  looks good
[01:32:08] <crosbot> tree became 'Tree is open (ihf investigating chrome compile in tegra2_aebl canary, rhashimoto pinned chrome to 15.0.842.0)'
[01:37:31] <rhashimoto> sosa: Thanks for your help!  You're right, wasn't that hard except for picking the proper baseline.
[01:39:57] <sosa> rhashimoto: np
[01:40:55] *** Keybuk has quit IRC
[01:41:49] <crosbot> tree became 'Tree is open (build team looking at x86-pineview, ihf investigating chrome compile in tegra2_aebl canary, rhashimoto pinned chrome to 15.0.842.0)'
[01:42:04] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot" on "tegra2_kaen-binary" from 8e2e74508ddbf52757af5371fb8ad034c72ec53e: sail at chromium dot org <sail at chromium dot org@0039d316-1c4b-4281-b951-d872f2087c98>)'
[01:47:04] <crosbot> tree became 'Tree is closed (rhashimoto looking at "tegra2_kaen-binary")'
[01:49:02] <cmasone> sosa: I have the lock for the sheriff page?  Which one?
[01:53:56] <crosbot> tree became 'Tree is closed (davidjames addressing "tegra2_kaen-binary")'
[01:55:43] <crosbot> tree became 'Tree is closed (davidjames fixing archive build for "tegra2_kaen-binary" and all internal builders/canaries)'
[01:58:56] *** petermayo has quit IRC
[01:58:58] *** lipsinV2 has joined #chromium-os
[01:59:43] *** behdad has quit IRC
[02:02:36] <crosbot> tree became 'Tree is open (archive_build should cycle green -> davidjames)'
[02:03:53] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot" on "x86-zgb-binary" from None: )'
[02:04:54] <crosbot> tree became 'Tree is closed (TOT preflight queue failure -> sheriffs)'
[02:06:41] <crosbot> tree became 'Tree is open (archive_build should cycle green, ellyjones: http://gerrit.chromium.org/gerrit/5214 landed as a stopgap for crosbug.com/18622)'
[02:08:29] *** D|sT has quit IRC
[02:09:22] *** powderluv has quit IRC
[02:09:31] *** rbyers has joined #chromium-os
[02:12:04] *** D|sT has joined #chromium-os
[02:14:50] <crosbot> tree became 'Tree is closed (raymes fixing tegra2_kaen-binary)'
[02:16:38] *** wfrichar has quit IRC
[02:18:46] *** kmixter has joined #chromium-os
[02:19:07] *** ChanServ sets mode: +v kmixter
[02:22:30] <crosbot> tree became 'Tree is open (raymes fixed libc-bench, should cycle green)'
[02:27:38] *** behdad has joined #chromium-os
[02:29:50] *** bpontes has quit IRC
[02:31:44] *** dianders_ has joined #chromium-os
[02:31:56] <crosbot> tree became 'Tree is closed (raymes fixed libc-bench, should cycle green)'
[02:32:26] <crosbot> tree became 'Tree is open (raymes fixed libc-bench, should cycle green)'
[02:36:11] *** dianders_ has quit IRC
[02:43:34] *** BThompsonGR has joined #chromium-os
[02:43:35] *** jrbarnette has quit IRC
[02:43:40] *** rbyers has quit IRC
[02:43:48] *** rbyers has joined #chromium-os
[02:43:49] *** ChanServ sets mode: +v rbyers
[02:53:31] *** jrbarnette has joined #chromium-os
[02:53:31] *** ChanServ sets mode: +v jrbarnette
[02:54:55] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot" on "tegra2_kaen canary" from None: )'
[02:55:25] <crosbot> tree became 'Tree is closed (rhashimoto looking at "tegra2_kaen canary")'
[02:55:40] *** dianders_ has joined #chromium-os
[02:56:11] <crosbot> tree became 'Tree is open (raymes fixed libc-bench, should cycle green)'
[02:57:27] *** Fractal has quit IRC
[02:59:51] *** behdad has quit IRC
[03:00:33] *** zmedico has quit IRC
[03:04:25] *** dianders_ has quit IRC
[03:07:00] *** jrbarnette has quit IRC
[03:07:42] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot" on "tegra2_seaboard-tangent canary" from None: )'
[03:08:28] <crosbot> tree became 'Tree is closed (rhashimoto looking at "tegra2_seaboard-tangent canary")'
[03:08:59] <crosbot> tree became 'Tree is open (raymes fixed libc-bench, should cycle green)'
[03:09:17] <rhashimoto> West coast sheriffs are heading for home.
[03:15:49] *** saintlou has quit IRC
[03:19:08] *** zmedico has joined #chromium-os
[03:22:17] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot" on "stumpy canary" from None: )'
[03:31:46] *** dospod has quit IRC
[03:38:52] *** cmasone_ has joined #chromium-os
[03:42:05] *** Fractal has joined #chromium-os
[03:42:42] <Fractal> Why does Chrome OS use the closed syntp driver instead of synaptics?
[03:44:04] *** cmasone_ has quit IRC
[03:46:21] <crosbot> tree became 'Tree is open (raymes fixed libc-bench, should cycle green)'
[03:48:30] *** dospod has joined #chromium-os
[03:49:10] *** dospod has quit IRC
[03:51:08] *** powderluv has joined #chromium-os
[03:51:08] *** ChanServ sets mode: +v powderluv
[03:56:19] *** thieule has quit IRC
[03:59:48] *** oc80z has quit IRC
[03:59:48] *** oc80z has joined #chromium-os
[04:05:24] *** jochen__ has quit IRC
[04:05:29] *** jochen__ has joined #chromium-os
[04:05:29] *** ChanServ sets mode: +v jochen__
[04:07:48] *** oc80z has quit IRC
[04:09:11] *** oc80z has joined #chromium-os
[04:10:48] <elly> Fractal: probably because it's better :)
[04:12:19] *** katier has joined #chromium-os
[04:12:19] *** ChanServ sets mode: +v katier
[04:13:04] <Fractal> elly: Yes probably, but why not contribute to synaptics instead of using a closed source driver which is not even available for the other Linux distributions?
[04:15:35] <elly> Fractal: I can't say, but I suspect the open-source driver didn't support features that were considered important and had no reasonable chance of same
[04:15:44] <elly> I know nothing at all about the trackpad stuff though
[04:15:50] *** oc80z has quit IRC
[04:16:33] *** oc80z has joined #chromium-os
[04:32:04] *** quiche1 has left #chromium-os
[04:32:06] *** BladeFreak has quit IRC
[04:49:21] *** jochen__ has quit IRC
[04:49:26] *** jochen__ has joined #chromium-os
[04:49:27] *** ChanServ sets mode: +v jochen__
[05:04:03] *** rbyers has quit IRC
[05:09:07] *** jochen__ has quit IRC
[05:09:12] *** jochen__ has joined #chromium-os
[05:09:12] *** ChanServ sets mode: +v jochen__
[05:48:22] *** oc80z has quit IRC
[05:48:22] *** oc80z has joined #chromium-os
[06:21:57] *** BThompsonGR has quit IRC
[06:44:09] *** mayday_jay__ has joined #chromium-os
[06:46:15] *** mayday_jay__ has left #chromium-os
[06:50:13] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot_master" on "x86 generic pre flight queue" from ab23eeed5e9ecff4878b076aaa2a3960b84197d5: Dominic Mazzoni <dmazzoni at chromium dot org>, _third_party_ at chromium dot org)'
[06:55:22] *** behdad has joined #chromium-os
[06:55:28] *** mayday_jay__ has joined #chromium-os
[07:00:13] *** kinaba_ has joined #chromium-os
[07:00:25] *** kinaba_ has quit IRC
[07:00:30] *** kinaba_ has joined #chromium-os
[07:06:48] <crosbot> tree became 'Tree is closed (maintenance)'
[07:20:05] <crosbot> tree became 'Tree is Open (sosa: things look good, will continue to watch)'
[07:29:00] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot" on "tegra2_arthur-binary" from dd82e2503a6c3f7ff18ad03e80b5dbaaa69e0214: tkent at chromium dot org <tkent at chromium dot org@268f45cc-cd09-0410-ab3c-d52691b4dbfc>)'
[07:31:41] *** kinaba_ has quit IRC
[07:34:24] *** sergiu has quit IRC
[07:37:20] *** kinaba_ has joined #chromium-os
[07:42:00] <crosbot> tree became 'Tree is closed (hungte watching - caused by cros-devutils race condition, tegra*binary should cycle green soon)'
[07:46:11] *** magn3ts has joined #chromium-os
[07:49:51] *** Sergiu has joined #chromium-os
[07:52:42] <crosbot> tree became 'Tree is open (ARM binary builds passed, should cycle green)'
[07:53:09] *** powderluv has quit IRC
[08:06:32] *** behdad has quit IRC
[08:12:15] *** sadrul has quit IRC
[08:12:57] *** vmil86 has joined #chromium-os
[08:13:53] <crosbot> tree became 'Tree is open (ARM binary builds are good now)'
[08:22:27] *** kinaba_ has quit IRC
[08:24:14] *** stalled has quit IRC
[08:32:59] *** sadrul has joined #chromium-os
[08:34:26] *** stalled has joined #chromium-os
[08:49:17] *** giroro_ has joined #chromium-os
[08:50:43] *** Ruetobas has quit IRC
[08:51:01] *** rush2end has quit IRC
[08:53:27] *** rush2end_ has joined #chromium-os
[08:53:31] *** rush2end_ is now known as rush2end
[09:05:37] *** radhermit has quit IRC
[09:13:09] *** radhermit has joined #chromium-os
[09:19:43] *** powderluv has joined #chromium-os
[09:19:44] *** ChanServ sets mode: +v powderluv
[09:37:48] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot" on "x86-alex-binary" from dd82e2503a6c3f7ff18ad03e80b5dbaaa69e0214: tkent at chromium dot org <tkent at chromium dot org@268f45cc-cd09-0410-ab3c-d52691b4dbfc>)'
[09:38:58] *** magn3ts has quit IRC
[09:47:49] *** Sergiu has quit IRC
[09:49:06] *** SoulRaven has joined #chromium-os
[10:02:30] *** patcito has quit IRC
[10:14:42] <pastarmovj> hi any cros sheriffs around?
[10:15:31] <pastarmovj> I'd like to commit a small patch to the metrics library that shouldn't influence the buildbots color (at least not making them any more red :) )
[10:19:25] *** rosso has joined #chromium-os
[10:25:41] *** gfrog has quit IRC
[10:27:19] *** ser has quit IRC
[10:27:27] *** ser_ has joined #chromium-os
[10:40:04] <pastarmovj> I can imagine there are not many sheriffs in the CEST time zone so I will commit my change and closely observe the tree for any anomalies.
[10:40:32] <crosbot> tree became 'Tree is open (test failure in testUpdateKeepStateful on alex 18679)'
[10:42:17] *** gfrog has joined #chromium-os
[10:46:37] *** gfrog has quit IRC
[11:01:11] <pastarmovj> I will commit a fix for that failing test ASAP
[11:05:58] *** Styx has joined #chromium-os
[11:07:06] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot_master" on "x86 generic PFQ" from a67b4333b6e07cba0c26cd182054be1abe7f7196: Julian Pastarmov <pastarmovj at chromium dot org>)'
[11:20:46] <pastarmovj> fix landed should cycle green I will continue to observe closely
[11:23:20] *** zbehan has quit IRC
[11:24:58] <crosbot> tree became 'Tree is open (test failure in testUpdateKeepStateful on alex 18679, lkgm failure 18681)'
[11:34:16] *** zbehan has joined #chromium-os
[11:52:51] <pastarmovj> All seems good now
[11:53:13] <pastarmovj> ed7daa4cbc53dd559127b976cb93ad57577942cd fixes the buggy test
[12:02:47] *** phil2 has joined #chromium-os
[12:03:13] *** phil2 has quit IRC
[12:03:48] *** phil2 has joined #chromium-os
[12:32:53] <powderluv> which test is this ?
[12:33:34] <powderluv> rather which failure on the bot
[12:44:49] *** phil2 has quit IRC
[12:57:28] <pastarmovj> http://build.chromium.org/p/chromiumos/builders/x86%20generic%20PFQ/builds/6/steps/Test/logs/stdio
[12:57:52] <pastarmovj> MetricsLibraryTest.AreMetricsEnabledCaching was failing after my commit
[13:02:11] <crosbot> tree became '?'
[13:02:26] <crosbot> tree became 'Tree is open (test failure in testUpdateKeepStateful on alex 18679, lkgm failure 18681)'
[14:11:14] *** mayday_jay__ has quit IRC
[14:19:21] *** SoulRaven has quit IRC
[14:43:10] *** rbyers has joined #chromium-os
[15:13:42] *** Adys has quit IRC
[15:14:33] *** rharrison_chrome has joined #chromium-os
[15:14:33] *** ChanServ sets mode: +v rharrison_chrome
[15:17:01] <rharrison_chrome> East Coast Sheriff checking in
[15:17:03] *** seventh has joined #chromium-os
[15:18:01] *** Adys has joined #chromium-os
[15:36:29] *** SoulRaven has joined #chromium-os
[15:50:36] *** sadrul has quit IRC
[16:00:27] *** seventh has quit IRC
[16:00:40] *** sadrul has joined #chromium-os
[16:00:40] *** ChanServ sets mode: +v sadrul
[16:06:35] <crosbot> tree became 'Tree is open'
[16:15:37] <ellyjones> yay, tree is open
[16:38:25] *** D|sT has quit IRC
[16:42:57] *** D|sT has joined #chromium-os
[16:46:35] *** varunjain has quit IRC
[16:46:49] *** varunjain has joined #chromium-os
[16:46:49] *** ChanServ sets mode: +v varunjain
[16:51:14] *** varunjain has quit IRC
[16:51:22] *** varunjain has joined #chromium-os
[16:51:24] *** ChanServ sets mode: +v varunjain
[16:51:43] *** varunjain has quit IRC
[16:51:57] *** varunjain has joined #chromium-os
[16:51:57] *** ChanServ sets mode: +v varunjain
[16:52:49] *** varunjain has quit IRC
[16:54:40] *** varunjain has joined #chromium-os
[16:54:41] *** ChanServ sets mode: +v varunjain
[16:54:57] *** varunjain has quit IRC
[16:55:04] *** varunjain has joined #chromium-os
[16:55:05] *** ChanServ sets mode: +v varunjain
[16:55:28] <crosbot> tree became 'Tree is open (rharrison ooo, kliegs is sheriff)'
[16:57:15] *** rosso has quit IRC
[16:57:19] *** Styx has quit IRC
[16:58:00] *** D|sT has quit IRC
[17:06:45] *** rbyers_ has joined #chromium-os
[17:06:46] *** ChanServ sets mode: +v rbyers_
[17:13:58] *** D|sT has joined #chromium-os
[17:16:54] *** jrbarnette has joined #chromium-os
[17:20:32] *** BladeFreak has joined #chromium-os
[17:31:58] *** gfrog has joined #chromium-os
[17:50:40] *** oc80z has quit IRC
[17:53:00] *** Fractal has quit IRC
[17:59:48] *** jrbarnette_ has joined #chromium-os
[17:59:48] *** ChanServ sets mode: +v jrbarnette_
[18:01:58] *** oc80z has joined #chromium-os
[18:02:01] *** behdad has joined #chromium-os
[18:02:55] <kliegs> crosbot: sheriffs
[18:02:56] <crosbot> kliegs: No such command 'sheriffs'. Try 'help'.
[18:02:57] <kliegs> crosbot: sheriff
[18:02:57] <crosbot> kliegs: No such command 'sheriff'. Try 'help'.
[18:02:59] <kliegs> crosbot: sheriff?
[18:03:04] <kliegs> argh
[18:03:08] <kliegs> crosbot: help
[18:03:08] <crosbot> kliegs: admin [[+-]<nick>]*, announce [[+-]<stat>]*, help, ping, quiet, speak, stat <stat-name>
[18:03:09] <cmasone> crosbot: sheriffs?
[18:03:09] <crosbot> cmasone: sheriffs: keybuk, thieule, kinaba, rharrison
[18:03:13] <cmasone> so close :-)
[18:03:25] *** jrbarnette has quit IRC
[18:03:26] *** jrbarnette_ is now known as jrbarnette
[18:03:32] *** saintlou has joined #chromium-os
[18:03:32] *** ChanServ sets mode: +v saintlou
[18:03:32] <kliegs> If only we had some technology here that could do fuzzy matching ons earches
[18:03:42] <kliegs> time to shard crosbot
[18:04:38] <kliegs> hmm. don't see keybuk or thieule on.  wanted to go get lunch
[18:04:40] <kliegs> anyone sit near them?
[18:04:45] <cmasone> kliegs: yes
[18:04:47] <cmasone> they're not here yet
[18:04:52] <kliegs> dang.
[18:05:02] <cmasone> kliegs: I will watch the tree while you get food
[18:05:11] <cmasone> kliegs: lemme know when you're back
[18:05:17] <kliegs> cmasone: yay! thanks.
[18:06:32] <kliegs> rharrison should be back shortly to resume as well. am covering for him. but i'll be quick if he's slow
[18:14:21] *** aaronp_ has joined #chromium-os
[18:26:53] <aaronp_> marcheu: Did you want to review those last-minute Border property changes?
[18:29:00] *** behdad has quit IRC
[18:31:40] *** jrbarnette_ has joined #chromium-os
[18:33:14] <kliegs> cmasone: back, thanks
[18:34:53] <cmasone> kliegs: np
[18:35:16] *** jrbarnette has quit IRC
[18:35:52] *** jrbarnette_ has quit IRC
[18:37:06] *** bpontes has joined #chromium-os
[18:37:07] *** ChanServ sets mode: +v bpontes
[18:49:04] *** lipsinV2 has quit IRC
[18:51:19] *** stevenjb has joined #chromium-os
[18:51:20] *** ChanServ sets mode: +v stevenjb
[18:58:07] *** gfrog has quit IRC
[19:00:12] *** petermayo has joined #chromium-os
[19:00:12] *** ChanServ sets mode: +v petermayo
[19:01:41] <marcheu> aaronp_: let me see
[19:01:53] *** jrbarnette has joined #chromium-os
[19:01:54] *** ChanServ sets mode: +v jrbarnette
[19:07:32] <crosbot> tree became 'Tree is open'
[19:11:05] <Stepan> yay
[19:12:27] *** sergiu has joined #chromium-os
[19:12:28] *** sergiu has joined #chromium-os
[19:12:30] *** jrbarnette has quit IRC
[19:12:41] <marcheu> aaronp_: yeah still reviewed-by, there isn't much that changed
[19:13:35] <aaronp_> Right.
[19:13:35] <aaronp_> Thanks
[19:13:44] <aaronp_> I'll push that and then move forward with the X server piece.
[19:22:21] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot" on "tegra2_kaen-binary" from fbd9d1b26204242b060e56c1b2c15bd492e9d97d: acolwell at chromium dot org <acolwell at chromium dot org@0039d316-1c4b-4281-b951-d872f2087c98>, estade at chromium dot org <estade at chromium dot org@0039d316-1c4b-4281-b951-d872f2087c98>, saintlou at chromium dot org <saintlou at chromium dot org@0039d316-1c4b-4281-b951-d872f2087c98>)'
[19:24:08] <crosbot> tree became 'Tree is closed (investigating -> rharrison)'
[19:25:58] *** JasonO-Web has joined #chromium-os
[19:27:25] <saintlou> sheriffs: I fail to understand how my change (or actually any of the CL above) could have broken the tree.
[19:28:13] <kliegs> saintlou: That's what they all say...
[19:28:23] <saintlou> keybuk, thieule, kinaba, rharrison: let me know what you think.
[19:29:05] <saintlou> kliegs: that is true, but at least I compiled, tested and trybots my code which is more that I can say for some of the CLs I see when I am sheriff ;-)
[19:29:17] <kliegs> saintlou: they all say that too
[19:29:18] <cmasone> saintlou: it wasn't a chrome-side change.
[19:29:24] <rharrison_chrome> Looks like xf86-input-cmt was broken
[19:29:27] * kliegs mutters that they talk to much
[19:29:30] <cmasone> I don't think anyone on that blamelist is at fault
[19:29:37] <saintlou> cmasone: that is what I keep telling kliegs :-)
[19:30:06] *** powderluv has quit IRC
[19:30:28] <kliegs> saintlou, cmasone: I actually agree with cmasone.  Please make sure to go vote up the dory items for "make blame lists accurate"
[19:30:47] <saintlou> :-)
[19:31:24] <cmasone> I wonder if changing the guestures package made us rebuild that input-cmt package, exposing some already-existing failure?
[19:31:30] *** Keybuk has joined #chromium-os
[19:31:30] *** ChanServ sets mode: +v Keybuk
[19:31:33] <cmasone> adlr: ping
[19:33:06] <kliegs> rharrison_chrome, adlr: http://gerrit.chromium.org/gerrit/5127 looks suspicious
[19:33:17] <adlr> uh oh
[19:33:21] * adlr looks
[19:33:51] *** JasonO-Web has quit IRC
[19:34:01] <kliegs> xf86-input-cmt-0.0.1-r17: /build/tegra2_kaen/tmp/portage/x11-drivers/xf86-input-cmt-0.0.1-r17/work/xf86-input-cmt-0.0.1/src/event.h:27: error: 'ABS_MT_DISTANCE' undeclared here (not in a function)
[19:34:35] *** JasonO-Web has joined #chromium-os
[19:34:59] <kliegs> I wish gerrit would change the dates in the git log to when items actually got merged.  running 'git log' shows that change with an Aug 1 date.
[19:35:48] <kliegs> hmm. maybe git log has a flag I can pass
[19:39:48] <cmasone> rharrison_chrome: This failure will probably be interesting to post-mortem :-)
[19:40:08] <rharrison_chrome> The failure is because event.h does not include mt.h where those macros are defined
[19:40:34] <cmasone> rharrison_chrome: sure, but why are we only tripping over this now?  And why is a package with x86 in the name getting build on arm?
[19:40:37] *** patcito has joined #chromium-os
[19:40:46] <cmasone> oh
[19:40:48] <kliegs> xf86 != x86
[19:40:48] <cmasone> xf86
[19:40:51] <cmasone> duuuuuuuh
[19:40:52] *** thieule has joined #chromium-os
[19:40:52] *** ChanServ sets mode: +v thieule
[19:40:53] <marcheu> xfree86 :)
[19:40:56] <cmasone> so strike that
[19:41:01] <cmasone> but, the first thing still stands :-)
[19:41:04] <kliegs> yah.  don't like that name
[19:41:20] *** grundler has joined #chromium-os
[19:41:20] *** ChanServ sets mode: +v grundler
[19:41:33] <rharrison_chrome> I think your assement that the new gesture code exposes a preexisting bug might be correct
[19:41:58] <adlr> i'm setting up a kaen board on my checkout to try to repo, but it'll be a bit
[19:42:23] <cmasone> rharrison_chrome: sure, that's why we tripped over it now, but why did that package build OK last time we built it?  Did something change the headers it depends on, and not cause that package to get built again?
[19:42:26] <adlr> i think the missing names are from linux kernel headers? or whatever populates /usr/include/linux/
[19:42:40] <kliegs> cmasone: that CL added a new include file
[19:43:23] <adlr> feel free to roll it back. that may be the fastest thing to check
[19:43:30] <srao> kliegs: git.chromium.org does show both times... I'm not sure how to get that info via command line
[19:43:39] <kliegs> cmasone: there was a change just submitted to the package - it just didn't show up in the blamelist
[19:43:53] <cmasone> kliegs: to guestures, or to the xf86 package?
[19:43:56] <kliegs> srao: ooh. good to know.  I've been going to the gerrit CL itself to see the "merged now" message but that could help
[19:44:02] <kliegs> cmasone: to the xf86 package - see the CL I linked
[19:44:10] <cmasone> kliegs: ooooh
[19:44:34] <kliegs> rharrison_chrome: I agree with adlr on rollback
[19:44:51] <cmasone> got it
[19:45:18] <kliegs> not sure a postmortem for this is useful if the rollback fixes it. it becomes "change made that wasn't validated on one architecture, binary builder caught it, rolled back" - seems like system worked correctly here
[19:45:27] <cmasone> kliegs: true enough
[19:45:40] <kliegs> although a bug should still be filed for future analysis.
[19:45:47] <kliegs> well. future statistics.
[19:46:05] <cmasone> kliegs: really?  I don't agree...
[19:46:08] <kliegs> and the work being done by rcui on trybots should help catch this type of problem in the future
[19:46:11] <kliegs> on filing the bug?
[19:46:15] <cmasone> kliegs: yeah
[19:46:16] <kliegs> or not needing a postmorem?
[19:46:30] <cmasone> agree on no postmortem, disagree about the bug
[19:46:32] <kliegs> cmasone: I thought there was a policy of "if tree closes file bug with label".
[19:46:42] <cmasone> kliegs: nope.
[19:46:52] <cmasone> kliegs: that policy was mostly generating noise.
[19:47:03] <cmasone> kliegs: if there's an infrastructure failure, then we should file a bug
[19:47:15] <cmasone> kliegs: or something that persists and requires investigation
[19:47:37] <kliegs> oh. i missed that change
[19:47:39] <cmasone> kliegs: new docs express this
[19:48:44] <rharrison_chrome> I have created a revert CL
[19:48:46] <rharrison_chrome> http://gerrit.chromium.org/gerrit/#change,5306
[19:49:02] <rharrison_chrome> adlr can you check this ^
[19:49:14] * adlr checks
[19:49:28] <kliegs> in that case disregard the comment about the bug.
[19:50:48] <adlr> LGTMed
[19:50:50] * kliegs goes back to grumbling about a way to pin this on saintlou
[19:51:10] <crosbot> tree became 'Tree is closed (potential issue identified, rolling back -> rharrison)'
[19:52:35] <srao> kliges: command line way to get commit date:  git log --pretty=fuller
[19:53:04] <kliegs> srao: nice! thanks
[19:53:21] <rharrison_chrome> The fix is in, watching the bot, once it passes the fail point I will open the tree
[19:53:40] *** JasonO-Web has quit IRC
[19:53:55] <adlr> i synced just before the revert, and i'm repro-ing w/ a tegra2_kaen build on my machine
[20:03:26] <rharrison_chrome> adlr: ok, looks like the builder got past that package, any objections to me opening the tree?
[20:03:39] <adlr> no
[20:04:03] *** ChanServ sets mode: +v adlr
[20:04:25] <rharrison_chrome> I am re-opening the tree, nobody commit CLs that will cause redness, k? :-)
[20:04:57] <adlr> i never have and i never will
[20:05:10] <adlr> ;)
[20:05:12] <crosbot> tree became 'Tree is open (xf86-input-cmt issue fixed -> rharrison)'
[20:06:53] *** stalled has quit IRC
[20:11:44] *** zmedico_ has joined #chromium-os
[20:12:15] *** zmedico has quit IRC
[20:13:16] <adlr> hm, so i did an internel sync, setup board, and build packages on tegra2_kaen and it worked
[20:13:22] <adlr> internal, that is
[20:15:39] *** behdad has joined #chromium-os
[20:25:20] <kliegs> adlr: did you try emerge-tegra2_kaen xf86-input-cmt ?
[20:25:29] <kliegs> build_packages might have fetched the prebuilt version
[20:25:37] <kliegs> well. or cros_workon start xf86-input-cmt
[20:25:38] <adlr> yeah
[20:26:00] <adlr> i think i'm workon-ing it, but my local repos are all checked out to m/master
[20:26:16] <kliegs> equery-tegra2_kaen l xf86-input-cmt ?
[20:26:51] <adlr> [I-O] [  ] x11-drivers/xf86-input-cmt-9999:0
[20:27:37] <kliegs> ok.  i've been bit by that too many times so wanted to check
[20:27:55] <adlr> i'll do a clean checkout/build
[20:28:08] *** kmixter has quit IRC
[20:29:18] <kliegs> adlr: its possible its also a dependency issue.  which could explain why it worked on one but not another
[20:29:27] <petermayo> plea/reminder:  please fill in a username when forcing a build.
[20:31:13] <adlr> kliegs: ah
[20:31:48] <adlr> so if i have a totally clean checkout, maybe i should try to build cmt by itself before running build_packages, then?
[20:32:03] <kliegs> adlr: yah. its always worth trying that
[20:32:09] <adlr> will do
[20:32:41] <kliegs> technically every package should build that way, so catching cases where it doesn't is usually worth it even if not applicable to the current bug. less overall flakiness=good
[20:32:50] *** patcito has quit IRC
[20:32:50] *** patcito_ has joined #chromium-os
[20:32:59] <adlr> yup
[20:33:01] <adlr> i agree
[20:33:14] <kliegs> plus its quicker :)
[20:33:25] <ellyjones> that would be easy to automate testing of
[20:34:15] *** patcito_ is now known as patcito
[20:34:23] *** BladeFreak has quit IRC
[20:34:34] <kliegs> easy but expensive.  i think it'd been mentioned in the past but probably wasn't a priority
[20:35:25] <kliegs> I think there are ~300+ packages.  so you'd need to a setup_board --force; emerge foo 300+ times
[20:39:13] <kliegs> so the missing defines come from /usr/include/linux/input.h - odd
[20:39:22] <kliegs> too bad we did a fresh run on the bot.  kinda want to poke
[20:39:30] <kliegs> rharrison_chrome, cmasone: this one may need a post morterm after all
[20:41:36] <njw> one bot that did that per-package test would probably take a few days to cycle over everything, but dependancies don't change very fast. Would definitely be better than nothing.
[20:42:22] *** rcui_ has joined #chromium-os
[20:43:30] <kliegs> njw: agreed.  like a weekly bot that goes on weekends could be good.  even if things changed on it, it could still catch some stuff
[20:49:02] <crosbot> tree became 'Tree is open'
[20:49:56] *** BladeFreak has joined #chromium-os
[20:54:36] *** sideone has joined #chromium-os
[20:55:18] *** sideone has left #chromium-os
[20:56:30] *** BladeFreak has quit IRC
[21:04:19] <kliegs> adlr: I forgot -g, I take back the "faster" part of my previous comment
[21:18:22] <gspencer> Whoa.  That's a lot of green up there.  Cool!  I'll have to submit something to get a little variety in the buildbot display. :)
[21:19:09] <rharrison_chrome> please wait until 4:30 EDT :-)
[21:20:40] <gspencer> rharrison_chrome: OK, will do. :)
[21:21:01] <kliegs> adlr chromeos-base/gestures just failed to build for me of a clean environment
[21:22:41] <thieule> hmm...did we just lose the x86 pfq?
[21:23:37] <thieule> nvm
[21:26:20] *** rcui_ has left #chromium-os
[21:26:55] *** rcui_ has joined #chromium-os
[21:27:35] *** rcui__ has joined #chromium-os
[21:27:36] *** ChanServ sets mode: +v rcui__
[21:31:13] *** rcui_ has quit IRC
[21:33:40] <kliegs> adlr: don't think it was the cause of the break but sent you a cleanup CL - http://gerrit.chromium.org/gerrit/5314
[21:34:41] <rharrison_chrome> thieule, thieule: I have meetings to attend and then I am done for the day, is one of you going to be watching the tree from now?
[21:34:50] <rharrison_chrome> keybuk ^
[21:35:01] <rharrison_chrome> Keybuk^
[21:35:54] <Keybuk> I am currently watching the tree while thieule is at lunch
[21:36:22] <rharrison_chrome> ok, thanks, just wanted to make sure that someone would be here when I am gone
[21:36:25] <sergiu> anyone know how I can make the alex crash on the second cpu? msb said something with (.*)asset(.*) but I didn't remember what and he's out for lunch :)
[21:36:33] *** rharrison_chrome has quit IRC
[21:36:39] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot" on "x86 generic full" from cabb5849aea2e02869a2a5b8ac9e6689b7fadad1: Taylor Hutt <thutt at chromium dot org>, _third_party_ at chromium dot org, macpherson at chromium dot org <macpherson at chromium dot org@268f45cc-cd09-0410-ab3c-d52691b4dbfc>, mseaborn at chromium dot org <mseaborn at chromium dot org@fcba33aa-ac0c-11dd-b9e7-8d5594d729c2>, pfeldman at chromium dot org <pfeldman at chromium dot org@268f45cc-cd09-0410-ab3c-d52691b4dbfc>)'
[21:37:03] <Keybuk> oh, wow, timing
[21:38:42] <crosbot> tree became 'investigating -> keybuk'
[21:40:14] <crosbot> tree became 'Tree is closed (investigating -> keybuk)'
[21:42:32] <crosbot> tree became 'Tree is open (keybuk: x86-generic-full failed repo sync after 2h30, let it try again)'
[21:45:21] <crosbot> tree became 'Tree is open (keybuk: x86-generic-full failed repo sync; issue 18250)'
[21:46:44] *** Styx has joined #chromium-os
[21:50:12] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot" on "x86-alex-binary" from f94f03235d9b3f79e809fe352aac4e16598816dd: jrt at chromium dot org <jrt at chromium dot org@0039d316-1c4b-4281-b951-d872f2087c98>)'
[21:52:21] *** zmedico_ is now known as zmedico
[21:54:02] <crosbot> tree became 'Tree is open (keybuk: chrome sig 11 in login_CrytohomeMounted)'
[21:54:05] <Keybuk> ok, I go to lunch now
[21:54:08] <Keybuk> thieule: you have the conn
[22:19:09] *** stalled has joined #chromium-os
[22:25:46] <adlr> kliegs: cmasone: i did a new checkout, patched in kliegs' CL that I LGTMed (gestures's deps fix) and built gestures and CMT w/ the failing patch. All without issue
[22:26:33] <adlr> at this point, it seems it would be okay to check in the change we reverted then
[22:29:04] <cmasone> adlr: i 'unno...give it a shot, i guess :-)
[22:30:15] <adlr> i guess partly i'm curious why this could be. are the linux headers treated specially?
[22:30:23] <adlr> should we have clobbered that bot?
[22:31:16] <Keybuk> adlr: which set of linux headers? :-)
[22:31:51] <adlr> for one, /build/$BOARD/usr/include/linux/input.h
[22:32:37] <Keybuk> those are copied from /usr/i686-pc-linux-gnu/usr/include/linux by setup_board
[22:32:59] <Keybuk> thus they're actually part of the host, not the target
[22:33:08] <adlr> wow
[22:33:34] <adlr> is there a version of input.h for the target?
[22:33:40] <Keybuk> no
[22:34:04] <adlr> is this considered broken?
[22:34:27] *** rbyers has quit IRC
[22:34:32] <Keybuk> well, toolchain has to come from outside of the target chroot
[22:34:55] <adlr> sure
[22:35:14] <adlr> but, we have an xinput driver that uses constants from input.h
[22:35:28] <adlr> it seems weird to use the host's version of that
[22:35:49] <Keybuk> well, you have to have a consistent API
[22:36:34] <Keybuk> so to update linux/input.h, you update the linux-headers ebuild
[22:36:50] <Keybuk> to get it used, you need to update the version pin for it and the hardcode in setup_board
[22:36:59] <Keybuk> which generally needs a toolchain respin (though there was work to add a builder to do that)
[22:37:34] <adlr> hm, so is this why a builder may have not had a recent version of input.h?
[22:37:42] <Keybuk> exactly
[22:38:14] <adlr> i see. i don't suppose there's a way to put something in the ebuild for the xinput driver, eh? We need to clobber the bots?
[22:38:36] <Keybuk> well, you could copy the linux/input.h header into the driver and include it locally instead
[22:38:51] *** rbyers has joined #chromium-os
[22:38:54] <adlr> hm
[22:39:21] <adlr> that seems like trading one broken solution for another... let me walk over to the kernel folks and ask
[22:39:35] <Keybuk> davidjames has probably slightly more up to date knowledge
[22:40:52] <Keybuk> I'm trying to remember what the difference between linux-headers and kernel-headers was
[22:41:05] <Keybuk> the host package is linux-headers, and the target package is kernel-headers
[22:41:11] <Keybuk> but I can't remember which gets used for what
[22:41:16] <adlr> hm, neither davidjames nor kernel folks at their desks
[22:41:20] <marcheu> yeah the host headers are stuck at 2.6.32
[22:41:31] <Keybuk> marcheu: I bumped them all to 2.6.38 a while back
[22:41:37] <marcheu> but I don't see why you'd want to use those
[22:41:49] <marcheu> Keybuk: they weren't a month ago, is that more recent?
[22:41:58] <Keybuk> marcheu: the dump was early June
[22:42:02] <Keybuk> bump I mean
[22:42:12] <marcheu> for the target yeah, but I'm talking about the host
[22:42:19] <marcheu> AFAIK the host uses old headers still
[22:42:38] <Keybuk> you mean in /usr/include/linux on the host?
[22:42:51] <Keybuk> [ebuild   R   ] cross-i686-pc-linux-gnu/linux-headers-2.6.38
[22:42:57] <Keybuk> that's certainly 2.6.38
[22:43:24] <marcheu> well, look at what's in /usr/include/linux vs in /build/x86-generic/usr/include/linux
[22:43:27] <marcheu> it's obviously different
[22:43:30] <Keybuk> right, sure
[22:43:34] <Keybuk> but /usr/include/linux isn't used for anything
[22:43:56] <Keybuk> the ones in /usr/i686-pc-linux-gnu/usr/include/linux are the important ones
[22:44:09] <marcheu> right, but
[22:44:12] <Keybuk> and the ones in /build/x86-generic/usr/include/linux too
[22:44:18] <marcheu> when working on amd64-generic I found a lot of pachages which use those in /usr/include/
[22:44:33] <Keybuk> marcheu: so amd64 is going to have other problems
[22:44:49] <Keybuk> because we mask things out on amd64 assuming that amd64 means "the host chroot"
[22:44:49] <marcheu> well, it happens on x86 also, it's just not always noticed
[22:45:06] <marcheu> I know, I fixed a ton of those over the last month
[22:45:58] <marcheu> so anyway, my point is that there are some bogus packages which still grab stuff from /usr/include/
[22:46:36] <Keybuk> adlr: so, I think, the one you want to look at is chromeos-base/kernel-headers
[22:46:56] <Keybuk> which has a GIT SHA in it of a version of our linux git tree that it's matched to
[22:47:10] <Keybuk> dumping that GIT SHA (or cros_workon'ing it) should give you updated headers to build against
[22:47:19] <Keybuk> bumping, not dumping - I keep making that typo today
[22:47:30] <adlr> hm
[22:48:05] <adlr> so this issue is not that i can't build locally. it's that i can, but the one bot did not
[22:48:07] <Keybuk> marcheu: right, so to bump that, it's just a matter of package.keywords I think
[22:48:54] <Keybuk> marcheu: since:
[22:48:56] <Keybuk> (cros-chroot) keybuk@angrybirds ~/trunk/src/scripts $ ACCEPT_KEYWORDS="~amd64" emerge -pqv sys-kernel/linux-headers
[22:48:59] <Keybuk> [ebuild     U ] sys-kernel/linux-headers-2.6.38 [2.6.30-r1]
[22:49:08] <Keybuk> it's just masked
[22:49:10] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot" on "x86-zgb-binary" from 9e1d761e758547ca4ffe5d180a4cd55d13c92ecb: estade at chromium dot org <estade at chromium dot org@0039d316-1c4b-4281-b951-d872f2087c98>, hnguyen at chromium dot org <hnguyen at chromium dot org@0039d316-1c4b-4281-b951-d872f2087c98>)'
[22:50:29] <Keybuk> adlr: which bot ooi?
[22:50:39] <adlr> tegra2_kaen binary
[22:50:58] <crosbot> tree became 'Tree is closed (Investigating -> thieule)'
[22:51:57] <marcheu> Keybuk: yeah, I think old things depend on it, that'd mean update the chroot as well
[22:53:41] <Keybuk> adlr: that does, to me, look like an outdated builder
[22:53:55] <Keybuk> adlr: I don't see any evidence that it's grabbing the header from the wrong place
[22:55:33] <Keybuk> Copying file:///b/cbuild/arm-tegra2_kaen-private-bin/chroot/build/tegra2_kaen/packages/chromeos-base/kernel-headers-0.0.1-r5.tbz2...
[22:55:39] <Keybuk> adlr: found from another build log ^
[22:55:48] <Keybuk> it's definitely got an older 2.,6.30-era kernel-headers on there
[22:55:51] <Keybuk> r6 is 2.6.38
[22:55:59] <Keybuk> sounds like someone needs to setup_board that bot
[23:00:25] <adlr> i see
[23:00:46] <adlr> rcui pointed me to another issue on x86-generic i'm going to look into
[23:00:52] <Keybuk> adlr: this looks more suspicious actually
[23:00:58] <adlr> oh?
[23:02:06] <Keybuk> adlr: wander over here for a second ;)
[23:02:25] <adlr> Keybuk: what's your ldap?
[23:03:34] <Keybuk> adlr: keybuk ;-)
[23:03:47] <Keybuk> US-MTV-1950-373-J :-)
[23:03:50] <adlr> ah ha!
[23:03:54] <adlr> coming over
[23:06:10] <kliegs> adlr: I built fine earlier without the error. so something odd
[23:07:15] <Keybuk> kliegs: right now it looks like tegra2_kaen-binary is randomly building with different versions of the kernel-headers package
[23:07:32] <kliegs> keybuk: that's not good.  i'm logging into it now to see if I spot anything funky
[23:07:50] <kliegs> looks like this was a deeper issue and not adlr's fault - he just happened to trigger it
[23:08:42] <Keybuk> all the bad builds are coming from chromeosbuild16
[23:09:10] <kliegs> oh.  kaen binary uses multiple builders
[23:09:17] * adlr steps out for a shave. i'll check in the CL later this afternoon and i'll coordinate w/ sheriffs
[23:09:45] <Keybuk> yes
[23:10:23] <kliegs> Keybuk: ok. can definitely confirm input.h on it is bad
[23:10:33] <kliegs> is it ok for me to start a chroot on it and poke?
[23:14:05] <gspencer> cmasone: did you already fix the page_cycler_tests problem?
[23:15:01] <cmasone> gspencer: I already updated teh ebuild to remove the obsolete targets and add the new one, yeah
[23:15:09] <gspencer> cmasone: so what do I have to do to kick it so it works for me?  I just synced, and I still get "No rule to make target `page_cycler_tests'" when I build.
[23:15:42] <cmasone> ohhhhhh
[23:15:47] <cmasone> I know what's up
[23:15:58] <cmasone> you're using LOCAL_SOURCE?
[23:16:03] <gspencer> Yep
[23:16:13] <gspencer> (why would I do anything else? :)
[23:16:39] <cmasone> and the "version" it's building is 842 or something, right?
[23:16:43] <cmasone> (not 9999)
[23:16:51] <gspencer> 842.0, yes.
[23:17:12] <cmasone> we had to pin chrome to tuesday's version yesterday due to an unrelated failure.  So you're using an ebuild that's out of sync with your source tree
[23:17:21] <gspencer> Ahh.
[23:17:46] <cmasone> There was a thread on chromium-os-dev earlier today about various ways to get it to use the 9999 ebuild for chrome
[23:17:55] <gspencer> OK, I'll look there.
[23:17:56] <cmasone> which will get you unblocked for today
[23:18:15] <cmasone> and then we should be able to roll chrome forward this evening sometime and everything will be fine tomorrow
[23:19:24] <gspencer> OK, cool.  Was that because the version numbers went backwards or something?
[23:20:11] <cmasone> nah, it's just because the LOCAL_SOURCE thing has always been kinda hacky...it just uses whatever source tree you have, which whatever the most recent stable ebuild is.
[23:20:37] <gspencer> OK.
[23:20:38] <cmasone> if the stable ebuild is a day or two old because we had to pin it
[23:20:45] <cmasone> then you can run into this problem
[23:21:08] <cmasone> when building the actual package, it checks out the source at the appropriate revision for the ebuild
[23:21:24] <cmasone> but when forcing it to use a local tree that's checked out at...wherever...yoyu can run into this problem.
[23:21:36] <cmasone> The right thing is to make LOCAL_SOURCE be tied to using the 9999 ebuild somehow
[23:21:42] <cmasone> which is something we hope to get to
[23:21:45] <cmasone> in the next quarter or so
[23:22:06] *** giroro_ has quit IRC
[23:22:12] <gspencer> OK, thanks.  I understand now.
[23:23:17] <kliegs> kernel-headers is installed special as part of setup_board right?
[23:23:42] <Keybuk> kinda yeah
[23:23:48] <Keybuk> just spoken to davidjames
[23:23:56] <kliegs> so running build_packages won't update it?
[23:23:56] <Keybuk> the problem is nothing under chromeos actually depends on kernel-headers
[23:24:01] <Keybuk> (or build-depends)
[23:24:04] <kliegs> yah. that's what its looking like
[23:24:05] <Keybuk> so nothing is actually making it be upgraded
[23:24:22] <kliegs> i did an emerge -pv kernel-headers on buidl16 and it said it would do -r6 which is the correct one
[23:24:27] <Keybuk> yeah
[23:24:31] <Keybuk> we just did the same
[23:24:40] <Keybuk> so I'll fix this by adding a build-depends on k-h somewhere appropriate
[23:24:40] <kliegs> ahh - didn't realize you were on the box too.
[23:25:08] <kliegs> sounds good.
[23:25:19] <kliegs> a short term fix to clobber build16 might work.
[23:25:38] <kliegs> and its probably worth filing a bug to examine all packages installed by setup_board to ensure there aren't other stale packages orphaned
[23:25:52] *** achuith_ has joined #chromium-os
[23:25:52] *** ChanServ sets mode: +v achuith_
[23:25:57] *** achuith_ has quit IRC
[23:26:01] <Keybuk> well, I think I'd like to watch a CL to add the bdepend actually drag the upgrade to build16
[23:26:35] <Keybuk> kliegs: kernel-headers is the only one
[23:26:49] *** Ruetobas has joined #chromium-os
[23:26:56] <kliegs> oh good. easy to fix then :)
[23:27:13] *** achuith has quit IRC
[23:27:22] *** rbyers has quit IRC
[23:30:33] <sergiu> anyone know how I can force the crash reporter to collect my kernel dumps?
[23:30:49] *** achuith has joined #chromium-os
[23:30:50] *** ChanServ sets mode: +v achuith
[23:30:59] <sergiu> it works in the autotests but gives me a no consent log when trying directly
[23:33:29] <srao> sergiu: you need to create a file let me find it
[23:33:47] <sergiu> I created that file
[23:34:01] <srao> touch /var/lib/crash_sender_paused && echo <your ldap> /home/chronos/Consent\ To\ Send\ Stats
[23:34:14] <srao> two files I guess
[23:34:32] <srao> one to collect the crash and one to stop the crash sender from uploading and deleting it
[23:34:36] <sergiu> does that work even when running echo panic > /proc/breakme (as root) on the book?
[23:34:44] <sergiu> chromebook*
[23:34:51] <srao> it should.. this all happens on reboot time
[23:34:57] <srao> so it should be created beforehand
[23:35:03] <srao> there's an autotest I thought ?
[23:35:38] <sergiu> there is,  platform_KernelErrorPaths but I have to run a modified version
[23:38:49] <sergiu> grr, still "ignoring - no consent"
[23:41:06] *** satorux_ has left #chromium-os
[23:45:32] *** SoulRaven has quit IRC
[23:54:01] *** rush2end_ has joined #chromium-os
[23:55:45] *** rush2end has quit IRC
[23:55:48] *** rush2end_ is now known as rush2end
[23:57:03] <crosbot> tree became 'Tree is open (thieule: filing flaky test bug against testInterruptedUpdate)'
[23:57:20] <ellyjones> Keybuk: why does upstart need ptrace?
[23:57:32] <cmasone> ellyjones: to follow forks, I thunk
[23:57:38] <Keybuk> ellyjones: what he said
[23:58:14] <ellyjones> gross
[23:58:14] <ellyjones> ok
[23:58:40] <Keybuk> ellyjones: patches welcome ;-)
[23:59:20] *** ser_ is now known as ser





top