[00:00:51] *** TW1920 has quit IRC
[00:07:53] <vpalatin> nirnimesh: I have checked my local copy, I have reproduced the login_CryptohomeUnmounted failure on the x86 generic PFQ build 733 : https://sandbox.google.com/storage/?arg=chromeos-image-archive/x86-generic-pre-flight-queue/0.15.1004.0-a1-b733#chromeos-image-archive%2Fx86-generic-pre-flight-queue%2F0.15.1004.0-a1-b733
[00:07:54] <davidjames> nirnimesh: Hmm they fail pretty reliably for the buildbots
[00:08:34] <vpalatin> just using KVM and run_remote_tests smokeSuite
[00:10:01] <ttuttle|work> shit shit shit shit shit
[00:10:20] <adlr> wait, we can swear in here now?
[00:10:38] <ttuttle|work> no
[00:10:48] <ttuttle|work> I just submitted a change that will break the build.
[00:11:36] <davidjames> ttuttle|work: Post a link here and sheriffs will happily revert for you :)(
[00:11:49] <ttuttle|work> fix is faster
[00:12:07] <adlr> reverts are pretty fast...
[00:12:16] <vpalatin> ttuttle|work: so post the link to the fix review
[00:12:32] *** ChanServ sets mode: +v adlr
[00:12:50] <ttuttle|work> working on it
[00:13:10] <gauravsh> adlr: you call saying "shit" swearing? very puritan of you.
[00:13:34] <ttuttle|work> it is swearing
[00:14:23] <ttuttle|work> http://gerrit.chromium.org/gerrit/7363
[00:14:49] <ttuttle|work> davidjames: vpalatin: ^
[00:18:09] <ttuttle|work> anyone? :/
[00:19:02] <ttuttle|work> okay, I am just submitting it.  it is better than what was there.
[00:20:11] <vpalatin> ttuttle|work: I was verifying the types ... but I hit the approved button just in time
[00:20:30] <ttuttle|work> vpalatin: cool, thanks
[00:20:42] <ttuttle|work> it, um
[00:20:45] <ttuttle|work> does let me approve my own CLs :P
[00:20:54] <ttuttle|work> which is perhaps not the best, but can be useful.
[00:21:07] <vpalatin> ttuttle|work: yes it's useful to revert your own CL
[00:21:12] <ttuttle|work> yeah.
[00:21:14] <ttuttle|work> okay, sorry about that.
[00:22:36] <ttuttle|work> okay
[00:22:37] * ttuttle|work -> home
[00:24:01] <vpalatin> benchan, adlr: x86-alex-binary VMTest failed in a strange way : http://chromeos-botmaster.mtv.corp.google.com:8026/builders/x86-alex-binary/builds/1204/steps/VMTest/logs/stdio
[00:24:58] <adlr> looking
[00:25:25] * benchan looking
[00:25:38] *** kinaba_ has quit IRC
[00:27:45] *** kinaba_ has joined #chromium-os
[00:30:00] <davidjames> vpalatin: I think that's leftover because it didn't pick up webui rollback
[00:30:10] <vpalatin> adlr: internal server error (500) on the updater
[00:30:36] <vpalatin> davidjames: it looks pretty different, we get a Omaha error and a code 500
[00:31:16] <davidjames> vpalatin: Oh, yeah you're right. And it would have picked up the rollback
[00:33:23] <adlr> hm
[00:33:23] <benchan> vpalatin: the log doesn't tell much about the problem
[00:34:10] <vpalatin> benchan: the interesting log is test_harness/testUpdateWipeStateful/1_update/update.log
[00:34:24] *** saintlou has joined #chromium-os
[00:34:24] *** ChanServ sets mode: +v saintlou
[00:34:46] <crosbot> tree became 'Tree is closed (Automatic: "update" on "tegra2_kaen-binary")'
[00:34:51] <vpalatin> from http://chromeosbuild9.mtv.corp.google.com/archive/x86-alex-private-bin/0.15.1006.0-a1-b1204/test_results.tgz
[00:35:29] <vpalatin> WTF chromegw is down ?
[00:35:41] <vpalatin> hum transient
[00:36:28] *** reveman has quit IRC
[00:36:28] *** Kyngdom has quit IRC
[00:36:28] *** pastarmovj has quit IRC
[00:36:28] *** Hexxeh has quit IRC
[00:36:29] *** chubb_cloud has quit IRC
[00:36:30] *** dots has quit IRC
[00:36:48] <vpalatin> the tree closer is a strange error : http://chromeos-botmaster.mtv.corp.google.com:8026/builders/tegra2_kaen-binary/builds/1986/steps/gitrevisiondropper/logs/stdio
[00:36:53] <vpalatin> "/bin/sh: cannot create chromite/buildbot/revisions.pfq: Directory nonexistent"
[00:37:14] *** Hexxeh has joined #chromium-os
[00:37:28] <benchan> vpalatin: error: The requested URL returned error: 403 while accessing http://gerrit.chromium.org/gerrit/p/chromiumos/chromite/info/refs
[00:37:50] <vpalatin> yep I saw it afterwards
[00:37:53] <benchan> vpalatin: it failed to clone chromite?
[00:38:09] <vpalatin> as on my workstation, on the chromium sites failed at that time
[00:38:22] <vpalatin> I will retry : I bet on transient network failure
[00:38:23] *** nirnimesh_ has joined #chromium-os
[00:38:24] *** ChanServ sets mode: +v nirnimesh_
[00:38:51] *** saintlou has quit IRC
[00:39:20] <benchan> vpalatin: it returns 403 if I access http://gerrit.chromium.org/gerrit/p/chromiumos/chromite/info/refs
[00:39:35] <davidjames> vpalatin, benchan: Looks like it got past that part now on retry
[00:39:53] <crosbot> tree became 'Tree is open (transient network failure)'
[00:39:57] <vpalatin> davidjames: yes so I have reopened
[00:40:07] *** Hexxeh has quit IRC
[00:41:36] <nirnimesh> Google Storage shows artifacts for the first 100 only, and I don't see a 'next' button
[00:42:15] <vpalatin> nirnimesh: I never found such a feature on the web ui
[00:42:20] <vpalatin> despite looking hard
[00:43:02] <nirnimesh> the builds are sorted in ascending order (with no way to reverse sort). this makes the web ui useless
[00:43:15] <davidjames> nirnimesh: What are you looking for?
[00:43:53] <davidjames> nirnimesh: You can look at buildbot output and check artifacts URL, then download from there
[00:44:01] <nirnimesh> some recent build artifacts for x86-generic-pfq
[00:44:27] <nirnimesh> yeah, doing that.
[00:44:43] *** sergiu has joined #chromium-os
[00:44:44] *** sergiu has joined #chromium-os
[00:45:04] <nirnimesh> buildbot lists only the 4-5 builds. after that I need to change the build-number in the URL
[00:45:27] *** saintlou has joined #chromium-os
[00:45:28] *** ChanServ sets mode: +v saintlou
[00:50:06] <vpalatin> why all the full builders are no longer running ?
[00:51:09] <adlr> they seem to be building to me?
[00:51:12] <adlr> you mean external?
[00:51:32] <benchan> i think they are busy building chrome
[00:51:36] <vpalatin> yes, but there is nothing in the waterfall
[00:52:23] <adlr> vpalatin: which builder exactly isn't building?
[00:53:56] <benchan> anyone knows what to expect after clicking the "Ping Builder" button?
[00:53:57] <vpalatin> x86 generic full
[00:54:32] *** reveman has joined #chromium-os
[00:54:36] <vpalatin> and all the following   aaa bbb full
[00:55:22] <adlr> time to call in a tropper?
[00:55:24] <adlr> trooper
[00:55:53] *** dots has joined #chromium-os
[00:57:25] *** Kyngdom has joined #chromium-os
[00:59:23] *** chubb_cloud has joined #chromium-os
[01:07:04] *** Hexxeh has joined #chromium-os
[01:09:57] <davidjames> adlr: OSError: [Errno 2] No such file or directory: '/var/lib/devserver/static/update.gz'
[01:10:05] <davidjames> adlr: Any idea why that's happening on all the internal bots? see update.log
[01:10:39] <adlr> hm, i'll takea  look
[01:11:06] <davidjames> adlr: stumpy binary, x86 alex binary are good examples
[01:11:33] <adlr> looking
[01:14:17] <adlr> hm, i see the error
[01:16:05] <davidjames> adlr: Also synaptics not compiling on zgb :(
[01:16:31] <adlr> davidjames: which bot?
[01:16:46] <davidjames> adlr: http://chromeos-botmaster.mtv.corp.google.com:8026/builders/x86-zgb-binary/builds/1124/steps/BuildTarget/logs/stdio
[01:17:43] <adlr> hm, i wonder if we're having network issues?
[01:17:44] *** jochen__ has quit IRC
[01:17:50] *** jochen__ has joined #chromium-os
[01:17:50] *** ChanServ sets mode: +v jochen__
[01:18:33] <adlr> ssh: connect to host git.chromium.org port 22: Connection refused
[01:18:39] <adlr> err, oops
[01:20:44] <adlr> can i browser internal repos?
[01:20:47] <adlr> browse
[01:21:01] <adlr> like the synaptics one?
[01:21:44] <sosa> gerrit-int.chromium.org->Admin->projects
[01:22:32] *** sadrul has quit IRC
[01:22:43] *** vmil86 has quit IRC
[01:25:15] <vpalatin> adlr: scp -P 6222 -r synaptics-private at git dot chromium.org:/home/synaptics-private/chromeos-base .
[01:25:22] <marcheu> adlr: there is nothing to browse, synaptics is a binary package that we get over scp
[01:25:24] <vpalatin> stupid brute force but works for me
[01:25:59] <marcheu> (or in that case, which we failed at fetching over scp)
[01:26:32] <adlr> vpalatin: good idea
[01:26:33] <vpalatin> adlr: it seems we have no "synaptics-touchpad-zgb" directory there, only "synaptics-touchpad"
[01:33:55] *** flackr has quit IRC
[01:34:16] <vpalatin> it's late on my special EST schedule, I'm off sheriff duty from now.
[01:37:47] *** cooled_ has joined #chromium-os
[01:38:14] *** lipsinV2 has joined #chromium-os
[01:39:00] *** stevenjb has quit IRC
[01:40:46] *** cooled has quit IRC
[01:47:40] *** Solet has quit IRC
[01:48:24] <adlr> vpalatin: okay
[01:48:38] <adlr> i'm mostly done tracking down the zgb failure
[01:56:55] <adlr> benchan: i'm reverting the synaptics zgb change
[01:58:36] <crosbot> tree became 'Tree is closed (cycling doomed canary builds)'
[01:58:37] <benchan> adlr: i'm still looking at the AU failure in alex full binary
[01:59:01] <adlr> benchan: sosa reverted a change that may have caused it
[01:59:19] <benchan> adlr: I see
[01:59:29] <adlr> sorry, should have been more verbose
[01:59:53] <crosbot> tree became 'Tree is open'
[02:01:20] <adlr> actually, perhaps dparker will fix the syn issue
[02:10:03] <adlr> okay, syn zgb issue fixed
[02:20:24] *** Solet has joined #chromium-os
[02:23:44] *** PuffTheMagic has quit IRC
[02:24:05] *** PuffTheMagic has joined #chromium-os
[02:24:39] *** Guest18939 has quit IRC
[02:27:06] *** |simo| has joined #chromium-os
[02:27:06] *** |simo| is now known as Guest77574
[02:27:57] *** sadrul has joined #chromium-os
[02:27:57] *** ChanServ sets mode: +v sadrul
[02:31:59] <sergiu> umm, so has something changed in the way we commit ?:)
[02:33:31] <crosbot> tree became 'Tree is closed (cycling doomed canary builds again)'
[02:34:02] <crosbot> tree became 'Tree is open'
[02:40:39] *** wfrichar has quit IRC
[02:40:45] *** saintlou has quit IRC
[02:41:13] <crosbot> tree became 'Tree is closed (cycling doomed canary builds again)'
[02:41:44] <crosbot> tree became 'Tree is open'
[02:42:41] <benchan> davidjames: is the canary build failure a known issue?
[02:43:03] <davidjames> benchan: Yeah I'm hoping that once sosa's revert goes in the canaries will go green
[02:43:27] <davidjames> benchan: We'll know once zgb cycles
[02:43:43] <davidjames> benchan: (Fix was blessed by PFQ at 5:05pm)
[02:43:50] <davidjames> benchan: And zgb binary was first builder to pick it up
[02:43:57] <benchan> davidjames: ok, thanks
[02:45:31] <davidjames> benchan: alex-binary has the fix as well
[02:46:04] <benchan> davidjames: nice, so we'd expect the canary to go green this around?
[02:46:26] <davidjames> benchan: Yes, but won't know for sure until alex binary or zgb binary cycles green
[02:48:05] *** mrseb_ has joined #chromium-os
[02:48:38] *** mrseb has quit IRC
[02:49:46] *** saintlou has joined #chromium-os
[02:49:46] *** ChanServ sets mode: +v saintlou
[02:54:08] <davidjames> benchan: x86-alex binary is looking good
[02:54:12] <davidjames> benchan: I think http://gerrit.chromium.org/gerrit/7375 fixed it
[02:54:31] <davidjames> benchan: ( Note that that wasn't 'blessed' until 5:05pm )
[02:55:19] <sergiu> hmm, is it me or did the publish and submit button dissapear?
[02:55:25] <sergiu> disappear *
[02:55:44] <benchan> davidjames: I see
[02:56:17] <davidjames> sergiu: Not just you, see sosa@'s PSA about autotest repo (and his earlier PSA about commit queue)
[02:56:52] <davidjames> sergiu: Commit queue is working on your commit right now, see http://build.chromium.org/p/chromiumos/builders/x86%20generic%20commit%20queue/builds/1878
[03:01:40] <sergiu> davidjames, thanks, where was the PSA sent as I seem to have missed it?
[03:02:22] <sergiu> was it the message to the -team least about the tree task force?
[03:02:32] <sergiu> list even >_<
[03:02:53] <davidjames> sergiu: Search for from:sosa psa
[03:03:10] <davidjames> sergiu: to: Chromium OS dev <chromium-os-dev at chromium dot org>
[03:03:35] <davidjames> sergiu: If you're not on that list, suggest subscribing :)
[03:04:28] <sergiu> davidjames, yeah, I don't know how but I missed that one, thanks
[03:09:40] *** kinaba_ has quit IRC
[03:13:47] <davidjames> sergiu: Congrats! Your change was submitted by chrome-bot!
[03:14:05] <sergiu> yeey :)
[03:17:02] <sergiu> I just have to get used to the new setup, maybe a message in the topic about the change would be useful
[03:17:13] <sergiu> or is it only for some repositories?
[03:18:05] <sergiu> nevermind, reading the messages from the list now :)
[03:19:01] <benchan> adlr, davidjames: I'm taking off
[03:19:23] *** benchan has quit IRC
[03:20:01] <adlr> i will take off soon, too
[03:28:47] * adlr heads to dinner, then home
[03:43:35] *** cooled has joined #chromium-os
[03:44:57] *** cooled_ has quit IRC
[04:00:02] *** gfrog has joined #chromium-os
[04:02:30] *** saintlou has quit IRC
[04:15:43] *** MX80 has joined #chromium-os
[04:46:14] *** jochen__ has quit IRC
[04:46:20] *** jochen__ has joined #chromium-os
[04:46:23] *** ChanServ sets mode: +v jochen__
[05:20:36] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot" on "x86-zgb canary" from None: )'
[05:29:51] <crosbot> tree became 'Tree is open (chromium-os:20034)'
[05:41:11] *** mnissler_ has quit IRC
[05:41:24] *** mnissler_ has joined #chromium-os
[05:41:25] *** ChanServ sets mode: +v mnissler_
[06:17:15] <crosbot> tree became 'Tree is closed (Automatic: "update" on "TOT Pre-Flight Queue")'
[06:19:48] <crosbot> tree became 'Tree is open (git flakiness)'
[06:27:24] *** behdad has joined #chromium-os
[06:55:20] *** sergiu has quit IRC
[07:00:14] *** xc0ffee has joined #chromium-os
[07:03:01] *** behdad has quit IRC
[07:03:18] *** cooled has quit IRC
[07:06:45] *** Hydrogenum has quit IRC
[07:07:00] *** Hydrogenum has joined #chromium-os
[07:18:56] *** cmasone has quit IRC
[07:19:01] *** cmasone_ has joined #chromium-os
[07:19:01] *** ChanServ sets mode: +v cmasone_
[07:19:29] *** dianders has quit IRC
[07:20:07] *** dianders has joined #chromium-os
[07:20:08] *** ChanServ sets mode: +v dianders
[07:30:49] *** amruthraj has joined #chromium-os
[07:46:47] <hru> vpalatin:hi
[07:47:45] <hru> I got below error while running gyp_chromium
[07:47:48] <hru> Exception: Call to 'python build_nacl_irt.py --inputs --platform=x86-64' returned exit status 2. while loading dependencies of chrome/browser/sync/tools/sync_tools.gyp while loading dependencies of build/all.gyp while trying to load build/all.gyp
[07:49:10] <hru> while compiling chromium for chrome os
[08:23:26] *** vmil86 has joined #chromium-os
[08:57:55] *** Styx has joined #chromium-os
[09:03:50] *** rosso has joined #chromium-os
[09:14:47] *** amruthraj has left #chromium-os
[09:17:03] *** Justasic has quit IRC
[09:23:18] *** Justasic has joined #chromium-os
[09:24:11] *** jujugre has joined #chromium-os
[09:28:05] *** petermayo has quit IRC
[09:28:20] *** srikanth has joined #chromium-os
[09:28:51] <srikanth> I need to some help in running layout tests on chrome os
[09:29:12] <srikanth> i have built and chrome os image and installed to my local drive
[09:29:42] <srikanth> Do i need to build any specific chrome os image for layout testing?
[09:30:11] <srikanth> any one can please suggest me on how to run layout tests on chrome os
[09:37:07] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot" on "x86-alex-binary" from e79cd6d6a3848104ce814a31acad514eb6c4d908: Hung-Te Lin <hungte at chromium dot org>, _third_party_ at chromium dot org, falken at chromium dot org <falken at chromium dot org@0039d316-1c4b-4281-b951-d872f2087c98>, koz at chromium dot org <koz at chromium dot org@0039d316-1c4b-4281-b951-d872f2087c98>, kristianm at chromium dot org <kristianm at chromium dot org@0039d316-1c4b-4281-b951-d872f2087c98>, shinyak at google dot com <s
[09:45:18] <crosbot> tree became 'Tree is open (cros_au_test_harness timeout, rebuilding)'
[10:00:29] *** arun_ has joined #chromium-os
[10:08:05] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot" on "x86-alex_he canary" from None: )'
[10:19:38] *** patcito has quit IRC
[10:32:08] <crosbot> tree became 'Tree is open (failed in suite_Smoke of VMTest)'
[10:42:15] *** lipsinV2 has quit IRC
[10:45:00] *** lipsinV2 has joined #chromium-os
[10:59:15] <hru> where can I get ./build_chrome.sh  file for compiling chromium.
[11:06:21] *** robe_ has quit IRC
[11:07:17] *** robe_ has joined #chromium-os
[11:11:31] *** rush2end_ has joined #chromium-os
[11:13:33] *** rush2end has quit IRC
[11:13:50] *** rush2end_ is now known as rush2end
[11:19:03] *** robe__ has joined #chromium-os
[11:20:56] *** robe_ has quit IRC
[11:59:59] *** TW1920 has joined #chromium-os
[12:21:39] *** xc0ffee has left #chromium-os
[12:53:07] *** SignedAdam has joined #chromium-os
[12:56:25] *** jujugre has left #chromium-os
[13:02:19] *** McMAGIC--Copy has joined #chromium-os
[13:07:17] <SignedAdam> my netbook keep restarting with the new vels of chrome os,
[13:07:30] <SignedAdam> ChromeOS-Vanilla-0.14.674.2011_06_24_1726-ree5e90b2 works fine on my netbook
[13:07:42] <SignedAdam> anything higher and it does not work
[13:23:59] *** sadrul has quit IRC
[13:28:54] *** t4nk661 has joined #chromium-os
[13:29:17] <t4nk661> Hi does any one know how to port chromium os image(using qemu emulator) to armv7 board?
[13:33:55] *** stalled has quit IRC
[13:51:35] *** sadrul has joined #chromium-os
[13:52:00] *** stalled has joined #chromium-os
[13:58:13] *** vpalatin|h has joined #chromium-os
[13:59:10] * vpalatin|h is now on-duty sheriff
[14:00:00] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot" on "x86-alex-binary" from 9fb225f495122d6d7a4cbb0b62f4ccc5b0fd53e0: chrome-release at google dot com <chrome-release at google dot com@0039d316-1c4b-4281-b951-d872f2087c98>, dgozman at chromium dot org <dgozman at chromium dot org@0039d316-1c4b-4281-b951-d872f2087c98>)'
[14:01:37] *** gfrog has quit IRC
[14:13:08] *** gfrog has joined #chromium-os
[14:13:44] *** FusionX has joined #chromium-os
[14:21:12] *** seventh has quit IRC
[14:30:47] * vpalatin|h is heading toward this office. I will investigate the qlex-binary failure there.
[14:30:52] *** vpalatin|h has quit IRC
[14:32:35] *** gfrog has quit IRC
[14:34:28] *** gfrog has joined #chromium-os
[14:44:10] *** gfrog has quit IRC
[14:50:21] <crosbot> tree became 'Tree is closed (vpalatin investigating "x86-alex-binary")'
[14:50:40] *** Styx has quit IRC
[14:57:19] <crosbot> tree became 'Tree is open (timeout during tests, retrying ...)'
[14:57:27] *** t4nk661 has quit IRC
[15:10:22] *** Clark008 has quit IRC
[15:10:23] *** xiyuan has quit IRC
[15:10:23] *** redpig has quit IRC
[15:10:23] *** MikeW has quit IRC
[15:10:23] *** yoshiki_ has quit IRC
[15:10:23] *** deshantm has quit IRC
[15:10:23] *** msb__ has quit IRC
[15:10:23] *** davidjames has quit IRC
[15:10:24] *** Calchan has quit IRC
[15:10:24] *** jochen__ has quit IRC
[15:10:24] *** jshin has quit IRC
[15:10:24] *** achuith has quit IRC
[15:10:24] *** Flameeyes has quit IRC
[15:10:24] *** vmil86 has quit IRC
[15:10:37] *** xiyuan has joined #chromium-os
[15:10:37] *** ChanServ sets mode: +v xiyuan
[15:10:38] *** MikeW has joined #chromium-os
[15:10:38] *** msb___ has joined #chromium-os
[15:10:38] *** msb___ has quit IRC
[15:10:38] *** msb___ has joined #chromium-os
[15:10:39] *** redpig has joined #chromium-os
[15:10:39] *** ChanServ sets mode: +v msb___
[15:10:40] *** davidjames has joined #chromium-os
[15:10:41] *** ChanServ sets mode: +v redpig
[15:10:44] *** ChanServ sets mode: +v davidjames
[15:10:47] *** Calchan has joined #chromium-os
[15:10:47] *** Calchan has joined #chromium-os
[15:10:49] *** yoshiki_ has joined #chromium-os
[15:10:50] *** jochen__ has joined #chromium-os
[15:10:50] *** ChanServ sets mode: +v jochen__
[15:11:06] *** Flameeyes has joined #chromium-os
[15:11:14] *** deshantm has joined #chromium-os
[16:00:52] *** behdad has joined #chromium-os
[16:10:59] *** wfrichar has joined #chromium-os
[16:11:00] *** ChanServ sets mode: +v wfrichar
[16:13:30] *** petermayo has joined #chromium-os
[16:13:31] *** ChanServ sets mode: +v petermayo
[16:15:25] *** flackr has joined #chromium-os
[16:22:44] *** behdad has quit IRC
[16:23:59] *** SignedAdam has quit IRC
[16:30:35] *** rosso__ has joined #chromium-os
[16:31:15] *** rosso has quit IRC
[16:31:25] <crosbot> tree became 'Tree is open (alex-binary flake -> crosbug.com/19200)'
[16:37:05] *** behdad has joined #chromium-os
[16:37:06] *** ChanServ sets mode: +v behdad
[16:39:04] *** mmocny has quit IRC
[16:41:52] *** rosso__ has quit IRC
[16:43:59] <ellyjones> fff, overnight trybot run ran out of disk
[17:09:45] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot_master" on "x86 generic PFQ" from be15c5f623879671729ff66f0a46760b83a677b1: Elly Jones <ellyjones at chromium dot org>)'
[17:11:07] *** sakoman__ has joined #chromium-os
[17:11:20] *** rochberg_ has joined #chromium-os
[17:11:29] *** ChanServ sets mode: +v rochberg_
[17:11:32] <ellyjones> x86-gen-pfq failure is my fault, I think
[17:12:29] <ellyjones> yeah, the test does not line up with reality quite right
[17:12:35] *** falken_ has joined #chromium-os
[17:12:37] *** Honoome has joined #chromium-os
[17:12:57] *** tty234_ has joined #chromium-os
[17:13:07] *** ser_ has joined #chromium-os
[17:13:12] *** robe_ has joined #chromium-os
[17:13:12] <ellyjones> I will put up a fix shortly, I believe it is a minor problem
[17:14:05] *** stevenjb has joined #chromium-os
[17:15:27] <ellyjones> aha, the outer vault dir is expected to be owned by chronos
[17:15:29] <ellyjones> oops :)
[17:15:33] <vpalatin> ellyjones: ok thanks
[17:15:33] <crosbot> tree became 'Tree is closed (ellyjones preparing a fix)'
[17:17:49] * ellyjones taps foot and waits for autotest-tests to re-emerge
[17:19:45] *** behdad has quit IRC
[17:19:45] *** petermayo has quit IRC
[17:19:45] *** wfrichar has quit IRC
[17:19:45] *** Flameeyes has quit IRC
[17:19:45] *** robe__ has quit IRC
[17:19:45] *** arun_ has quit IRC
[17:19:45] *** dianders has quit IRC
[17:19:46] *** nkostylev has quit IRC
[17:19:46] *** dgarrett has quit IRC
[17:19:46] *** seumas has quit IRC
[17:19:46] *** kinaba has quit IRC
[17:19:46] *** BThompson has quit IRC
[17:19:46] *** vbendeb has quit IRC
[17:19:46] *** sbyer has quit IRC
[17:19:46] *** Sarten-X has quit IRC
[17:19:46] *** kliegs has quit IRC
[17:19:46] *** Hamlin has quit IRC
[17:19:46] *** sosa has quit IRC
[17:19:46] *** rochberg has quit IRC
[17:19:46] *** sakoman_ has quit IRC
[17:19:46] *** tty234 has quit IRC
[17:19:46] *** ser has quit IRC
[17:19:46] *** falken has quit IRC
[17:20:02] *** Sarten-X2 has joined #chromium-os
[17:22:34] *** wfrichar has joined #chromium-os
[17:22:34] *** ChanServ sets mode: +v wfrichar
[17:24:01] * ellyjones runs putatively-fixed test
[17:24:43] <ellyjones> hm
[17:24:44] *** benchan has joined #chromium-os
[17:24:44] *** ChanServ sets mode: +v benchan
[17:24:46] *** nkostylev has joined #chromium-os
[17:24:46] *** ChanServ sets mode: +v nkostylev
[17:24:47] *** dgarrett has joined #chromium-os
[17:24:48] *** seumas has joined #chromium-os
[17:24:48] *** ChanServ sets mode: +v dgarrett
[17:24:48] *** ChanServ sets mode: +v seumas
[17:24:53] *** kliegs has joined #chromium-os
[17:24:53] *** ChanServ sets mode: +v kliegs
[17:24:54] <ellyjones> vpalatin: is webui known to be bad or something? I have noticed login tests flaking more for me
[17:25:18] *** BThompson has joined #chromium-os
[17:25:19] *** ChanServ sets mode: +v BThompson
[17:25:28] *** kinaba has joined #chromium-os
[17:25:36] <vpalatin> ellyjones: webui-login has been de-activated again yesterday
[17:25:57] <vpalatin> as it was increasing the VMtest failure rate dramatically
[17:26:06] *** petermayo has joined #chromium-os
[17:26:06] *** ChanServ sets mode: +v petermayo
[17:26:07] <ellyjones> interesting
[17:26:10] <vpalatin> (it's still unclear it's the root cause of the problem)
[17:26:21] *** sbyer has joined #chromium-os
[17:26:21] *** ChanServ sets mode: +v sbyer
[17:26:24] *** dianders has joined #chromium-os
[17:26:24] *** ChanServ sets mode: +v dianders
[17:26:30] *** behdad has joined #chromium-os
[17:26:30] *** ChanServ sets mode: +v behdad
[17:26:35] *** sosa has joined #chromium-os
[17:26:35] *** ChanServ sets mode: +v sosa
[17:27:00] <ellyjones> alright
[17:28:08] *** vbendeb has joined #chromium-os
[17:28:08] *** ChanServ sets mode: +v vbendeb
[17:28:15] *** TW1920 has quit IRC
[17:28:16] *** TheBrokenGLaDOS has quit IRC
[17:28:20] * adlr is wfh for the first part of the AM, then out from 10-noon or so, then in the office all afternoon
[17:28:52] *** behdad has quit IRC
[17:28:52] *** dianders has quit IRC
[17:29:49] *** behdad has joined #chromium-os
[17:29:49] *** ChanServ sets mode: +v behdad
[17:30:16] <ellyjones> vpalatin: http://gerrit.chromium.org/gerrit/#change,7401 fixes the test, I believe
[17:30:59] <ellyjones> I am going to TBR it, and file a bug against myself to fix the test properly, since I introduced some new functionality that is not covered
[17:31:32] <vpalatin> ellyjones: ok I'm looking at it
[17:32:24] <ellyjones> I don't know how to actually _land_ it though - there's no submit button today
[17:32:51] *** Hamlin has joined #chromium-os
[17:33:18] <vpalatin> are you logged with your @chromium.org address ?
[17:33:23] <vpalatin> else I can push it
[17:33:26] <ellyjones> yup
[17:33:35] <ellyjones> you can? o_O
[17:33:47] <vpalatin> oh no , I can't ...
[17:33:50] <vpalatin> strange
[17:33:53] <ellyjones> I have 'publish comments' but no 'publish and submit'
[17:33:56] <ellyjones> today must be a commit-queue day
[17:34:04] <vpalatin> I have the button on another review but not on this one
[17:34:13] <vpalatin> ellyjones: probably
[17:34:13] <benchan> vpalatin, ellyjones: tried, but gerrit still shows "Review in Progress" on 7401
[17:34:41] <ellyjones> benchan: that makes little sense
[17:35:11] <ellyjones> it looks like the commit queue just ran for a bit (http://build.chromium.org/p/chromiumos/builders/x86%20generic%20commit%20queue/builds/2031) but then bailed due to closed tree
[17:35:41] <ellyjones> "WARNING: TREE IS CLOSED.  PLEASE SET TO OPEN OR THROTTLED TO COMMIT
[17:35:42] <ellyjones> "
[17:36:05] <benchan> ellyjones: looks like the commit queue bot was rebooting?
[17:36:34] <ellyjones> I wonder how it will pick up my fix, then
[17:36:59] <kliegs> ellyjones: what's the link to your fix?
[17:37:06] <ellyjones> http://gerrit.chromium.org/gerrit/#change,7401
[17:37:11] <ellyjones> davidjames is on it
[17:37:29] <kliegs> i have a submit button if you want me to push it
[17:37:59] <ellyjones> davidjames bypassed the queue
[17:38:04] <kliegs> ahh ok
[17:38:32] <kliegs> i think i can get around it in some cases too if needed
[17:40:09] <ellyjones> yeah
[17:40:29] <ellyjones> davidjames just pointed out that the unit tests failed too on the bot (!?) so I think a revert is in order
[17:40:46] <ellyjones> I hate that the unit tests do not use the mocked platform thing
[17:40:52] <ellyjones> sigh
[17:41:53] <ellyjones> screw this, I will need to redo the unit tests before I can put this change in after all
[17:41:53] *** arun_ has joined #chromium-os
[17:42:22] <ellyjones> kliegs: can you please push http://gerrit.chromium.org/gerrit/#change,7404 in? I reverted the underlying change
[17:42:27] <ellyjones> so everything will be back to normal
[17:43:36] <kliegs> ellyjones: done. apparently I'm a chump
[17:43:41] <ellyjones> danke
[17:43:44] <kliegs> bitte
[17:43:58] <crosbot> tree became 'Tree is open (everything reverted)'
[17:45:32] <ellyjones> god damn it.
[17:45:51] *** saintlou has joined #chromium-os
[17:45:51] *** ChanServ sets mode: +v saintlou
[17:45:59] * ellyjones sizes up the yak
[17:46:11] <kliegs> gonna be one of those days, eh?
[17:46:32] <ellyjones> well
[17:46:46] <ellyjones> the security_ProfilePermissions thing was my fault all the way - didn't use a trybot, didn't even think to run that particular test
[17:47:01] <ellyjones> but the unit test thing is just that the unit tests use the real fs for about half their ops and fake one for the other half
[17:47:16] <ellyjones> and so whether they pass or not is a matter of luck and initial state
[17:47:20] <kliegs> oi
[17:47:36] <kliegs> better find some good shears
[17:47:44] <ellyjones> I've been putting off fixing it for a while but I guess now is the time
[17:47:58] <ellyjones> cryptohome$ g 'file_util::' * | wc -l
[17:47:58] <ellyjones> 223
[17:59:59] *** Adys has quit IRC
[18:01:31] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot" on "x86-alex-binary" from 9336ac3b22efce861456ff011230b03b907a37ab: _third_party_ at chromium dot org, timurrrr at chromium dot org <timurrrr at chromium dot org@0039d316-1c4b-4281-b951-d872f2087c98>)'
[18:02:10] *** Adys has joined #chromium-os
[18:03:35] <crosbot> tree became 'Tree is closed (desktopui_WindowManagerFocusNewWindows test fails on "x86-alex-binary" / vpalatin looking at it)'
[18:04:28] <vpalatin> benchan, adlr: I think we got this one once yesterday. Did you remember anything ?
[18:05:28] <adlr> hm
[18:05:37] <adlr> i don't recall it, personally
[18:05:49] <adlr> perhaps it was while I was at the meeting?
[18:05:54] <crosbot> tree became 'Tree is closed (chrome-wm crash during desktopui_WindowManagerFocusNewWindows test on "x86-alex-binary" / vpalatin looking at it)'
[18:06:37] <vpalatin> adlr: I suspect we never investigate it as it was in the middle of other failures
[18:07:55] <benchan> does anyone know if dhcpd has been recently updated?
[18:10:05] <adlr> vpalatin: have you talked to derat about the wm crash?
[18:10:14] <adlr> He wrote it, so he should probably look
[18:10:54] <vpalatin> adlr: no I don't
[18:11:04] <adlr> i'll IM him
[18:11:59] *** dianders has joined #chromium-os
[18:11:59] *** ChanServ sets mode: +v dianders
[18:13:23] *** magn3ts_ has joined #chromium-os
[18:13:35] *** magn3ts_ has quit IRC
[18:16:10] <davidjames> vpalatin, adlr: Did you spot the canary fails as well?
[18:16:26] <vpalatin> davidjames: not yet :(
[18:17:26] <vpalatin> davidjames, adlr: seems to be the same updater failure as yesterday
[18:17:51] <davidjames> vpalatin: Well, alex had trouble starting a VM
[18:18:13] <davidjames> vpalatin: alex_he has a FAIL in platform_FilePerms
[18:22:31] <benchan> davidjames:  platform_FilePerms shows a warning "fd /opt gid wrong"
[18:23:04] <davidjames> benchan: Yeah, wonder why -- could you file a bug for that? I think it's new flakiness
[18:23:16] <benchan> davidjames
[18:23:19] <benchan> davidjames: ok
[18:23:48] <benchan> davidjames: is alex_he any different from alex?
[18:24:14] *** derat has joined #chromium-os
[18:24:14] *** ChanServ sets mode: +v derat
[18:24:16] *** lipsinV2 has quit IRC
[18:25:23] <davidjames> benchan: Yes those are different platforms (and the builders also have different failures)
[18:28:11] <benchan> davidjames: apparently crosbug.com/17798 filed on Jul 18 mentions the same issue with platform_FilePerm
[18:31:46] <davidjames> benchan: Yeah that looks similar, I guess update that bug?
[18:32:50] <benchan> davidjames: I'm looking into the test history and see if it fails intermittently on certain platforms
[18:33:32] <adlr> Hey guys, I'm sorry, I need to step out a bit earlier. I'll be in the office asap (likely around noon) and be ona duty for the rest of the day
[18:36:32] <derat> the window manager crash in x86-alex-binary b1217 is the usual thing, where xcb did something stupid because we shut down the X server in the middle of initialization.
[18:37:35] <vpalatin> derat: do we already have a bug for that ?
[18:37:46] <vpalatin> so I can merge the new one into it
[18:37:46] <derat> vpalatin: a bunch of them.  tree should be reopened in the meantime
[18:37:52] <vpalatin> I will do it
[18:38:02] <derat> i haven't seen a bug for the new instance, but i can reopen one of the old ones.
[18:38:29] <crosbot> tree became 'Tree is open (known chromeos-wm shutdown issue)'
[18:39:10] <vpalatin> derat: or just comment on 20185 : http://code.google.com/p/chromium-os/issues/detail?id=20185
[18:43:00] <davidjames> derat: Should we whitelist sig 6 again?
[18:44:03] <davidjames> vpalatin: Looks like ellyjones' changes didn't compile
[18:45:33] <crosbot> tree became 'Tree is closed (PFQ can't compile cryptohome -> ellyjones)'
[18:45:45] <derat> davidjames: yes, i think so.  i can probably rework things in the window manager so that we exit with 1 instead of sigabrt-ing when something breaks during initialization, but i think i care more about getting crash reports if this ever happens in production than about working around brokenness that only happens in tests... :-/
[18:46:08] <ellyjones> what? I reverted them
[18:46:36] <ellyjones> the revert is http://gerrit.chromium.org/gerrit/#change,7403 and says it's merged
[18:46:38] <davidjames> ellyjones: Hmm, maybe an incomplete revert?
[18:46:56] <ellyjones> I only had one change, and I hit 'revert' on gerrit for it
[18:47:00] <ellyjones> lemme look
[18:47:42] <vpalatin> ellyjones: revert then submit ?
[18:48:01] <ellyjones> yes
[18:48:19] <ellyjones> the revert is http://gerrit.chromium.org/gerrit/7403 for cryptohome and http://gerrit.chromium.org/gerrit/7404 for the test, both submitted
[18:48:27] <davidjames> derat: So now the tests wait for Chrome to start, I guess the problem is that when Chrome is started, it doesn't necessarily mean that chromeos-wm is started. Should tests maybe wait for chromeos-wm to finish starting as well?
[18:48:29] <ellyjones> and I did not touch cryptohomed.c at all
[18:49:27] * ellyjones sees if she can reproduce the compile failure locally...
[18:50:43] <ellyjones> it does not build for me either, but I blame glotov's 6f32f9faf1d5cced36b5678d15d5502c5dc92861
[18:51:09] <ellyjones> with that reverted it builds
[18:51:30] <davidjames> glotov: Can we revert that?
[18:52:28] <glotov> ok, since it fixes the tree. Looking...
[18:55:31] <derat> davidjames: i have a hard time figuring out what goes on in the testing code.  the status file says that desktopui_WindowManagerFocusNewWindows started at 08:37:05 and ended at 08:37:20.  the session_manager log shows the wm getting started at 083709, 083709, and 083718.  the test doesn't do any logging in or out besides initializing cros_ui_test.
[18:57:54] <crosbot> tree became 'Tree is closed (PFQ can't compile cryptohome -> glotov)'
[18:58:15] *** mmocny has joined #chromium-os
[18:58:16] *** ChanServ sets mode: +v mmocny
[18:58:30] <glotov> davidjames, will you revert?
[18:59:27] <davidjames> glotov: Sure, just did, http://gerrit.chromium.org/gerrit/#change,7408
[18:59:28] <vpalatin> glotov: I will do it unless you have a fix handy
[18:59:36] <vpalatin> davidjames: thanks
[19:00:28] <crosbot> tree became 'Tree is open (cryptohome CL reverted)'
[19:01:49] <gauravsh> davidjames: when I emerge-<board> a package repeatedly, is there a way to speed up the calculating dependencies step - maybe use cached information? It seems to be the longest step whenever I emerge a package.
[19:02:18] <davidjames> gauravsh: It does use cached information
[19:02:47] <davidjames> gauravsh: It's even slower without the cache -- takes many minutes. With the cache it's just a few seconds, but yeah that can be the slowest step
[19:03:05] <davidjames> gauravsh: --nodeps speeds it up but note that that literally means it ignores all dependencies
[19:03:50] <gauravsh> yeah, sometimes when I am just making changes/testing on a specific package, I have to do this repeatedly. Most of the time, this dependency information hasn't changed since the last time I ran the emerge.
[19:04:20] *** jochen__ has quit IRC
[19:04:36] <gauravsh> so I am fine even with it ignoring it.
[19:04:38] <gauravsh> glotov: platform.h:30: error: forward declaration of 'const struct cryptohome::base::Time'
[19:04:55] *** mnissler_ has quit IRC
[19:05:09] <davidjames> gauravsh: Note that the cache depends on timestamps so it often unnecessarily spends time regenerating the cache because a timestamp changed
[19:10:01] <gauravsh> davidjames: i see, i think --nodeps is good enough.
[19:10:51] *** jochen__ has joined #chromium-os
[19:10:52] *** ChanServ sets mode: +v jochen__
[19:10:54] *** mnissler has joined #chromium-os
[19:10:55] *** ChanServ sets mode: +v mnissler
[19:13:50] <gauravsh> glotov: it prepends the cryptohome namespace to that forward declaration.
[19:14:10] <gauravsh> glotov: if you do it outside namespace cryptohome {}, it compiles fine.
[19:14:55] <gauravsh> might as well just get rid of forward declaration, and just #include the header.
[19:15:19] <gauravsh> forward declaration syntax for a class in a different namespace is just ugly.
[19:17:00] <davidjames> derat: Take a look at cros_ui_test and you'll see a lot of logging in and out :P
[19:19:58] <davidjames> derat: First step in any test is to run 'stop ui'. I guess it doesn't check how long the ui has been running for ?
[19:21:37] *** dianders has left #chromium-os
[19:25:55] *** mnissler has quit IRC
[19:26:31] *** jshin has joined #chromium-os
[19:26:32] *** ChanServ sets mode: +v jshin
[19:26:49] *** mnissler has joined #chromium-os
[19:26:50] *** ChanServ sets mode: +v mnissler
[19:29:01] <davidjames> derat: So I have a potential fix, maybe we should add login.wait_for_browser() after we logout at the end of each test? That ensures that Chrome has started up before we declare a test as 'finished'
[19:29:25] *** Solet has quit IRC
[19:29:29] <davidjames> derat: That way, when we cros_ui.stop() in the next test, that isn't killing a chrome and window manager that aren't completely started up. Thoughts on that?
[19:31:11] <bryeung> after a repo sync today, I started seeing these errors
[19:31:13] <bryeung> /usr/libexec/gcc/armv7a-cros-linux-gnueabi/ld: /build/tegra2_seaboard/usr/lib/libcrosapi.a(load.o): Unknown mandatory EABI object attribute 44
[19:31:17] <bryeung> anyone know what is going on?
[19:32:07] <bryeung> specifically for power_manager and chrome builds
[19:32:12] <davidjames> bryeung: Sounds like you need to do a build_packages ?
[19:32:20] <gsam> two finger scroll broke w/ latest tot build on alex
[19:32:22] <bryeung> davidjames: after a build_packages is when this started
[19:32:42] <bryeung> davidjames: is there some prior step that I might need?  re-setup_board or something?
[19:33:02] <davidjames> bryeung: Hmm, well setup_board  --force will probably fix it, just doesn't explain why you ran into it
[19:33:19] <davidjames> bryeung: Some package needs to be recompiled, which package is it? Maybe Chrome?
[19:33:33] <gsam> and chrome is busted too; scrolling a view draws only part of the screen
[19:33:48] <bryeung> davidjames: chrome and power_manager are both getting recompiled and failing with these errors
[19:33:53] <gsam> other than that life is good
[19:34:27] <davidjames> bryeung: Hmm, well power manager is the easier one
[19:35:17] <davidjames> bryeung: Probably running into trouble with a recent change to libcros, not sure which one triggered that.
[19:35:40] <kliegs> davidjames, bryeung: I'm seeing the error now as well after running build_packages
[19:35:42] <kliegs> My usage:
[19:35:45] <kliegs> 1)  Ran repo sync
[19:35:52] <kliegs> 2) emerged power_manager, no error
[19:35:58] <kliegs> 3) ran build_packages, power_manager failed
[19:36:28] <davidjames> kliegs: And what happens if you emerge power_manager after that?
[19:36:31] <kliegs> my last repo sync was early yesterday I believe. so something in the host environment has either broken or is not backwards compatible
[19:36:42] <bryeung> davidjames: there are a bunch of other libraries that are also giving this error when building chrome:
[19:37:05] <kliegs> davidjames: fails
[19:37:21] <davidjames> kliegs: Right, so you probably picked up a change to libcrosapi.a
[19:37:38] <davidjames> kliegs: Try emerging libcros then merging power_manager
[19:37:53] <kliegs> davidjames: host or target emerge?
[19:37:57] <davidjames> kliegs: Target
[19:37:58] <benchan> adlr: the platform_FilePerms failure on x86-alex_he is probably due to how Synaptics installs stuff into /opt, seems like /opt is being chown to owned by the chronos group
[19:38:25] <benchan> adlr: I can reproduce it locally
[19:38:48] <gsam> benchan, that would explain my loss of 2-finger scroll too probably
[19:38:52] <bryeung> all these are failing me too:  libgtk-x11-2.0.so, libgdk-x11-2.0.so, libgdk_pixbuf-2.0.so, libXtst.so, libXi.so
[19:39:13] <kliegs> davidjames: worked
[19:39:16] <bryeung> (for a chrome compile)
[19:39:31] <davidjames> kliegs: Aha, yeah, something is borked about the libcros pulled in by build_packages
[19:39:41] <kliegs> davidjames: dependency failure?
[19:40:06] <davidjames> kliegs: Well try emerge-board -1g libcros; then emerge -1 power_manager
[19:40:14] <davidjames> kliegs: Er, emerge-board -1g libcros; then emerge-board -1 power_manager
[19:40:25] <davidjames> kliegs: That should replicate the failure maybe?
[19:40:46] <kliegs> davidjames: sorry I need to bolt. bryeung has the same problem so working on it
[19:40:53] <kliegs> i'll be back in ~90ish
[19:41:15] *** nkostylev has quit IRC
[19:41:17] <bryeung> davidjames: an emerge of libcros fixed it for power_manager for me
[19:44:18] *** cooled has joined #chromium-os
[19:46:16] *** Solet has joined #chromium-os
[19:57:25] *** mnissler has quit IRC
[19:57:27] *** mnissler_ has joined #chromium-os
[19:57:28] *** ChanServ sets mode: +v mnissler_
[20:00:57] <bryeung> davidjames: yes, indeed.  the commands you sent kliegs reproduce the issue.
[20:02:11] <davidjames> bryeung: Can you rev the libcros ebuild? That'd fix the issue
[20:07:05] *** patcito has joined #chromium-os
[20:09:00] <bryeung> davidjames: sorry, one moment...my repo is a little busted elsewhere
[20:12:56] <bryeung> davidjames: ebuild rev landed
[20:13:10] <bryeung> davidjames: thanks for helping me figure out what was going on
[20:17:41] *** patcito has quit IRC
[20:24:31] *** sergiu has joined #chromium-os
[20:24:59] *** BladeFreak has joined #chromium-os
[20:26:49] *** mikael__ has joined #chromium-os
[20:29:26] *** mikael__ has quit IRC
[20:31:45] *** patcito has joined #chromium-os
[20:33:28] *** achuith has joined #chromium-os
[20:33:28] *** ChanServ sets mode: +v achuith
[20:34:20] *** patcito has quit IRC
[20:34:57] *** patcito has joined #chromium-os
[20:35:08] *** m1k3l_ has joined #chromium-os
[20:39:24] *** chocobo__ has joined #chromium-os
[20:39:24] *** ChanServ sets mode: +v chocobo__
[20:41:52] *** patcito has quit IRC
[20:42:42] *** patcito has joined #chromium-os
[20:47:50] * adlr back on sheriff duty
[20:47:55] <adlr> (and in the office)
[20:48:16] <adlr> I'll be looking into the file permissions autotest failure
[20:51:30] <m1k3l_> is there a way to build an image that is not EFI eg ext3 or ext4?
[20:52:22] <vpalatin> m1k3l_, : ext3/ext4 are filesystem format, EFI is partition table format
[20:53:21] *** GodoPPL has joined #chromium-os
[20:55:09] <m1k3l_> right
[20:55:18] *** patcito has quit IRC
[20:56:25] *** patcito has joined #chromium-os
[20:56:35] <m1k3l_> my board doesn't support uboot so I wanted to mount part_3 from EFI image
[20:57:04] <m1k3l_> but it doesn't work
[20:58:19] <m1k3l_> so I wonder if there is a way to get a rootfs compatible with a typical initrd-type bootloader?
[21:01:35] <vpalatin> m1k3l_: the rootfs is just standard ext2 + signing
[21:01:56] <vpalatin> with rootfs verification disabled, you can mount it as ext2
[21:03:44] <m1k3l_> ok
[21:04:30] <m1k3l_> I did a similar thing: I extracted all the files from rootfs and put them on an ext4 partition
[21:05:21] <m1k3l_> then booted from there
[21:05:29] *** McMAGIC--Copy has quit IRC
[21:05:56] <m1k3l_> but /sbin/init doesn't work... kernel panic
[21:06:01] <m1k3l_> any idea?
[21:06:31] <vpalatin> m1k3l_: maybe your kernel command line is not ok
[21:07:15] <m1k3l_> I thought about that too but it works with a debian-based system
[21:07:35] <m1k3l_> what would you recommend for chromium?
[21:08:52] <adlr> vpalatin: (back)
[21:09:08] *** McMAGIC--Copy has joined #chromium-os
[21:09:21] <vpalatin> adlr: bon appetit
[21:10:45] <vpalatin> m1k3l_: your kernel panic probably contains the answer about what is your issue. btw for the kernel cmdline,  "root=<right/dev> rootwait" might be useful depending on your boot device
[21:11:10] <vpalatin> else a bunch of setting for dm-verity but it seems you are trying not to use it
[21:14:05] <m1k3l_> didn't think about rootwait as it boots quickly with debian
[21:14:10] *** chocobo__ has quit IRC
[21:14:57] <vpalatin> m1k3l_: the real issue is probably visible in the panic log...
[21:15:18] <m1k3l_> seems like something to do with IPI
[21:16:51] *** ihf_ has joined #chromium-os
[21:24:25] *** rbyers_ has joined #chromium-os
[21:24:25] *** ChanServ sets mode: +v rbyers_
[21:26:04] *** sbyer has quit IRC
[21:32:59] <derat> davidjames: sorry, was afk.  that sounds great (which you probably inferred from my LGTM) :-P
[21:33:36] *** sbyer has joined #chromium-os
[21:33:36] *** ChanServ sets mode: +v sbyer
[21:34:12] *** derat has quit IRC
[21:37:23] <nirnimesh> davidjames: I'm not sure your fix in http://gerrit.chromium.org/gerrit/#patch,sidebyside,7415,1,client/cros/cros_ui_test.py is going to help. self.pyauto.Logout() kinda already ensures that the browser is up after logout
[21:37:46] *** chocobo__ has joined #chromium-os
[21:37:46] *** ChanServ sets mode: +v chocobo__
[21:38:37] <nirnimesh> davidjames: login.wait_for_browser() does a pgrep, and is probably just buying time
[21:39:08] <m1k3l_> is chromium booting the same way gentoo? or are there any modifs for chromiumos?
[21:39:47] <ttuttle|work> it is entirely different from gentoo.
[21:40:14] <m1k3l_> is there any doc? or can you detail?
[21:42:34] <vpalatin> m1k3l_: we are using upstart and some custom bash scripts
[21:46:40] <m1k3l_> ok but first process is still /sbin/init?
[21:50:02] <vpalatin> m1k3l_: yes
[21:51:54] *** Hydrogenum has quit IRC
[21:51:54] *** Hydrogenum has joined #chromium-os
[21:54:55] <davidjames> nirnimesh: Does it ensure it fully launches?
[21:55:50] <nirnimesh> davidjames: what does "fully" mean? It doesn't wait for the pixels on the screen, but it does wait for the automation to connect back to the browser process -- which implies that chrome process is up and running
[21:56:11] <davidjames> nirnimesh: I see. Well fully launched means that if you stop Chrome it won't crash
[21:56:31] <davidjames> nirnimesh: Chrome crashes if you stop it 'too early' -- how do you know when it's OK to kill Chrome?
[21:56:49] <nirnimesh> isn't that a bug that needs to be fixed?
[21:57:31] <davidjames> nirnimesh: Ideally, but it's not fixed, and we get a lot of flakiness with those sort of crashes
[21:57:53] <davidjames> nirnimesh: Those crashes aren't useful to developers because nobody would stop Chrome before it puts up its first window
[21:58:36] <davidjames> nirnimesh: So should I do wait_for_initial_chrome_window instead ?
[21:58:37] <nirnimesh> wait_for_browser() doesn't ensure any of that
[21:59:04] <davidjames> nirnimesh: Or wait_for_login_prompt ?
[21:59:49] <nirnimesh> the future of wait_for_login_prompt is uncertain with webui login
[22:00:11] <davidjames> nirnimesh: Well, it needs to continue working, else tests become flaky
[22:00:13] <nirnimesh> wait_for_initial_chrome_window might be helpful
[22:00:30] <davidjames> nirnimesh: What's the diff between the two as far as the login prompt?
[22:00:49] *** rbyers_ has quit IRC
[22:01:15] <nirnimesh> they are 2 different hacks, piled on at different times :)
[22:01:41] <davidjames> nirnimesh: So they would both ensure that a window has shown up at least? And in theory user could start clicking stuff?
[22:01:48] <nirnimesh> I'm hoping all this won't be required with webui login, since the automation there ensures that login window is up
[22:02:24] <nirnimesh> davidjames: there's no more clicking code left at the login screen. autox has been replaced with pyauto login calls, which does not do any clicking
[22:02:39] <nirnimesh> For now, I think you should use wait_for_initial_chrome_window
[22:02:51] <davidjames> nirnimesh: Ah ok sure I'll use that
[22:03:11] <davidjames> nirnimesh: By the way, the reason why I want the window to be there isn't for any code that emulates clicking
[22:03:44] <davidjames> nirnimesh: It's for ensuring that we are emulating something the user could actually do (i.e. that the window is there so it would only be realistic to close chrome once the 'X' shows up that allows them to close the browser)
[22:04:39] <nirnimesh> davidjames: I understand that
[22:05:40] <davidjames> nirnimesh: Should we also make sure that Chrome has started up before calling login in pyauto ?
[22:05:58] <nirnimesh> the login automation does that
[22:06:15] <davidjames> nirnimesh: Do we still need wait_for_browser with the pgrep ?
[22:06:40] <nirnimesh> no. it's redundant after self.pyauto.Logout()
[22:06:55] <davidjames> nirnimesh: No I meant after         cros_ui.start()
[22:07:14] <davidjames> nirnimesh: After that it calls login.wait_for_browser(), then initializes pyauto
[22:07:46] <nirnimesh> No. not needed there. I haven't seen problems there. login automation handles it correctly
[22:08:10] <davidjames> nirnimesh: Ah so the code there can be removed I guess
[22:09:02] <nirnimesh> davidjames: errr.. sorry I misread. Yes wait_for_browser() is needed after cros_ui.start().
[22:09:19] <nirnimesh> wait_for_initial_chrome_window is not necessary there
[22:39:49] <kliegs> davidjames: i'm still seeing errors building chromeos-chrome. sounds like bryeung is too.  with the "unknown mandatory EABI object attribute 44" error
[22:47:09] <vpalatin> benchan, adlr: the x86-alex-canary is going to fail once again
[22:47:51] <benchan> vpalatin: due to the AU error?
[22:48:25] <vpalatin> looks like but we don't have the log yet
[22:49:00] <ellyjones> is there anything special that needs to be done when one adds a kernel config entry? I 1) added a line to kernel/Makefile, 2) added a line to chromeos/config/config.common.chromeos setting my config option to m, and 3) emerged
[22:49:05] <ellyjones> and ended up with my code not built
[22:49:44] <vpalatin> ellyjones: Kconfig might be wrong
[22:50:15] <vpalatin> try  ./chromeos/scripts/kernelconfig oldconfig ?
[22:50:50] <vpalatin> else it's my usual method to do it
[22:55:40] <ellyjones> vpalatin: hm, that does not seem to have helped
[22:55:51] * ellyjones greps for changes to chromeos/config to use as models
[22:58:22] <vpalatin> ellyjones: do you diff your config file after this oldconfig ?
[22:58:50] <ellyjones> I'm not sure which file to diff, really
[22:59:32] <vpalatin> "git diff chromeos/config" is enough probably
[22:59:38] * adlr tries to catch up on the canary failure
[23:00:19] <sergiu> ellyjones, are you adding new entries or just activating some existent ones
[23:00:30] <vpalatin> adlr: the build has not failed yet, it will soon
[23:01:54] <adlr> vpalatin:  you mean "x86-alex canary"?
[23:02:08] <vpalatin> adlr: yes
[23:02:25] <adlr> looks red to me
[23:03:06] <ellyjones> adding new entries, sergiu
[23:03:15] <ellyjones> mutating existing ones seems to work fine :)
[23:03:22] <vpalatin> adlr: the previous one was also red, but this one is not archived yet. So the tree closing has not happened
[23:03:46] <sergiu> ellyjones, maybe you forgot to add them to the Kconfig file as well?
[23:03:50] <adlr> ah
[23:03:53] <sergiu> http://gerrit.chromium.org/gerrit/#patch,sidebyside,5427,3,drivers/platform/Kconfig as an example
[23:05:33] <ellyjones> sergiu: which Kconfig file is appropriate if my config option is in kernel/Makefile?
[23:06:23] <vpalatin> ellyjones: it depends on the kind of feature
[23:06:35] <sergiu> ellyjones, umm, what file does it affect?
[23:07:11] <ellyjones> vpalatin: kernel/keystash.c, which I am about to add :)
[23:07:17] <ellyjones> sergiu: ^
[23:07:34] <ellyjones> I stole half of the ramoops space and am about to add a character device to access it from userspace
[23:08:41] <adlr> hm, seems the VM isn't reachable by the host?
[23:08:52] <adlr> or, semi not reachable?
[23:08:57] <ellyjones> hrm, I can just stick it into arch-specific kconfig or something
[23:09:03] <sergiu> ellyjones, since it's a char driver, maybe drivers/char would be a better place to put it
[23:09:13] <sergiu> there's a Kconfig file you could modify as well there
[23:09:19] <sergiu> ramoops is put there as well
[23:09:22] <ellyjones> alright, I'll do that
[23:09:23] <ellyjones> thanks, sergiu :)
[23:09:40] <sergiu> sure, no problem
[23:14:20] <kliegs> sosa: If I want to kick off the tegra2 chrome PFQ, I do that by starting the chrome pfq?
[23:14:31] <sosa> yup
[23:14:31] <kliegs> I noticed that the tegra2 chrome pfq hasn't run since 2 am
[23:14:38] <sosa> that is correct
[23:14:50] <sosa> the non-tot pfq's only kick off when there's a DEPS change
[23:14:58] <sosa> non-tot Chrome pfq's *
[23:15:07] <sosa> because they only respond to DEPS changes
[23:15:14] <kliegs> sosa: ahh.. that makes sense
[23:15:20] <kliegs> sosa: wait - its fine that I just started it right?
[23:15:29] <sosa> no
[23:15:32] <sosa> :)
[23:16:26] <kliegs> sosa: oh. crap.
[23:16:37] <kliegs> sosa: I mean I started the x86 generic one?
[23:18:19] <bryeung> good news for anyone seeing the EABI errors (maybe just kliegs and me)
[23:18:31] <bryeung> emerging the packages fixes it
[23:18:36] *** Solet has quit IRC
[23:18:55] *** Solet has joined #chromium-os
[23:26:01] <ellyjones> sergiu: is ramoops on in our config?
[23:26:27] <sergiu> yes, but it is not upstream
[23:27:14] *** behdad has quit IRC
[23:27:23] <sosa> kliegs: sorry i read your comment wrong, it's fine to start them up
[23:27:59] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot" on "x86-alex-binary" from ef27540edd8f0e91863e2b4e26538c5376b05888: _third_party_ at chromium dot org, rolandsteiner at chromium dot org <rolandsteiner at chromium dot org@268f45cc-cd09-0410-ab3c-d52691b4dbfc>, willchan at chromium dot org <willchan at chromium dot org@0039d316-1c4b-4281-b951-d872f2087c98>)'
[23:28:21] <adlr> uh oh
[23:28:46] *** lipsinV2 has joined #chromium-os
[23:29:12] <adlr> oh, looks like machine is down
[23:29:18] <adlr> should we open the tree?
[23:29:59] <vpalatin> adlr: we need a trooper probably
[23:30:03] <vpalatin> yes I think so
[23:30:20] <adlr> crosbot: troopers?
[23:30:20] <crosbot> adlr: troopers: djmm, scottz, bradnelson, maruel (EST), kliegs (EST)
[23:30:30] <kliegs> busted by crosbot
[23:30:31] <kliegs> what's up?
[23:30:50] <sergiu> ellyjones, yes, ramoops is in our config files
[23:31:01] <sergiu> but it is not in the upstream ones
[23:31:08] <ellyjones> alright, my tree must be out of date
[23:31:09] <adlr> kliegs: seems x86-alex-binary is rebooting/rebooted?
[23:31:17] <vpalatin> kliegs: x86-alex-binary builder is doing weird filesystem errors
[23:31:23] * adlr reopens tree
[23:31:48] <vpalatin> kliegs: http://chromeos-botmaster.mtv.corp.google.com:8026/builders/x86-alex-binary/builds/1222/steps/CleanUp/logs/stdio
[23:32:08] <crosbot> tree became 'Tree is open (bot was rebooted?)'
[23:32:46] *** achuith has quit IRC
[23:32:57] *** behdad has joined #chromium-os
[23:33:57] <kliegs> bot was rebooted mid-run?
[23:34:01] *** elly has quit IRC
[23:34:02] <kliegs> they tend to reboot after each build automatically
[23:34:24] <kliegs> oi. that log does look unhealthy
[23:35:00] *** behdad has quit IRC
[23:35:16] *** elly has joined #chromium-os
[23:35:24] <sosa> kliegs: ....
[23:35:29] <sosa> kliegs: it does
[23:35:43] *** chocobo__ has quit IRC
[23:36:06] <sosa> kliegs:  oh it's because ofthe force stop.  maybe a bug with the master.cfg but the bots aren't restarting after a force stop it seems
[23:36:52] <kliegs> sosa: ahh. gotcha.  so it should be fine on this next cycle?
[23:37:00] *** saintlou has quit IRC
[23:40:32] *** achuith has joined #chromium-os
[23:40:32] *** ChanServ sets mode: +v achuith
[23:41:28] <ellyjones> hm, something is super wrong... I just added a syntax error to the drivers/char/ramoops.c and no compilation failure
[23:44:53] <ellyjones> sergiu: are you sure ramoops is getting built?
[23:44:54] <kliegs> cros_workoned?
[23:44:57] <ellyjones> yep
[23:45:03] <ellyjones> it's building the -9999 kernel
[23:46:58] <sergiu> ellyjones, well, it would be very strange if it didn't do that, since the tests should fail.. lemme check
[23:49:13] <sergiu> are the .config files up to date?
[23:49:49] <sergiu> ellyjones, http://gerrit.chromium.org/gerrit/#patch,sidebyside,5428,5,chromeos/config/config.common.chromeos enables CONFIG_RAMOOPS, which should trigger building it in the Makefile in /drivers/char
[23:49:49] <kliegs> ellyjones: check in /build/foo/boot/config
[23:49:55] <kliegs> see if its got hte right config option set
[23:50:14] <ellyjones> it doesn't look like they are, sergiu; build/tegra2_kaen/.config has CONFIG_RAMOOPS commented out
[23:50:21] <ellyjones> is there something I need to run after syncing my kernel tree?
[23:51:43] <sergiu> ellyjones, hmm.. try ./chromeos/scripts/kernelconfig oldconfig maybe?
[23:52:08] <sergiu> chromeos/config/config.common.chromeos should affect all platforms afaik
[23:52:38] <ellyjones> there we go!
[23:52:57] <ellyjones> I feel like that should've happened automatically, though
[23:54:31] <sergiu> yeah, apparently it should, don't know if it does
[23:56:14] <kliegs> i've got a very bad feeling right about now
[23:56:24] <kliegs> however while I run one final test I'll let everyone panic
[23:58:55] <kliegs> Ok - it looks like the buildbots are using a different toolchain version than we're using locally
[23:59:11] <kliegs> so binary packages built from the toolchain aren't compatible with files we're building local.  at least for arm-tegra2
[23:59:50] *** patcito has quit IRC





top