[00:06:10] *** OtherJakeSays has joined #chromium-os
[00:06:43] *** JakeSays has quit IRC
[00:06:59] *** OtherJakeSays is now known as JakeSays
[00:07:06] *** JakeSays is now known as OtherJakeSays
[00:07:08] *** OtherJakeSays is now known as JakeSays
[00:07:10] *** JakeSays is now known as OtherJakeSays
[00:07:15] *** OtherJakeSays is now known as JakeSays
[00:14:52] <ellyjones> hm, I broke all the things
[00:15:53] <gsam> "things"?
[00:16:06] <ellyjones> http://gerrit.chromium.org/gerrit/5799
[00:16:11] <sergiu> break _all_ the things! :)
[00:16:25] *** sergiu is now known as Sergiu
[00:16:27] <ellyjones> I caused crosbug.com/18924 and crosbug.com/19045 :P
[00:29:18] *** cros_ has quit IRC
[00:37:11] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot_master" on "TOT Pre-Flight Queue" from 5a1489771842b8566a87d07d7f40ce5fd0e8ab9e: dennisjeffrey at chromium dot org <dennisjeffrey at chromium dot org@0039d316-1c4b-4281-b951-d872f2087c98>)'
[00:41:17] <crosbot> tree became 'Tree is open (yet another chrome sig11)'
[00:41:50] *** petermayo has joined #chromium-os
[00:41:51] *** ChanServ sets mode: +v petermayo
[00:49:06] *** magn3ts has quit IRC
[01:03:02] *** yoshi5534 has joined #chromium-os
[01:04:15] <yoshi5534> hi guys
[01:05:54] <yoshi5534> i'm tryin to set up my chroot and getting some permission denied errors, followed by 'ERROR   : Your chroot version file /etc/cros_chroot_version is bogus:'
[01:09:36] <grundler> yoshi5534, this help? --> http://groups.google.com/a/chromium.org/group/chromium-os-discuss/browse_thread/thread/f7ce192e7cd322ac?pli=1
[01:13:39] <yoshi5534> mine looks almost the same, but there's no fix suggested, isn't it?
[01:14:43] <grundler> yeah, looks like only more questions.
[01:46:36] *** thieule has quit IRC
[01:47:11] <zbehan> yoshi5534: is this 100% reproducible?
[01:47:40] <zbehan> yoshi5534: it's odd because i haven't seen similar errors in testing, i wonder if this could be some odd umask problem
[01:49:51] *** belak has left #chromium-os
[01:51:58] *** D|sT has quit IRC
[01:56:01] *** powderluv has quit IRC
[01:56:41] *** powderluv has joined #chromium-os
[01:56:42] *** ChanServ sets mode: +v powderluv
[01:57:15] *** sbyer has quit IRC
[02:00:08] *** grundler has quit IRC
[02:01:25] *** vmil86 has quit IRC
[02:06:33] <yoshi5534> zbehan: i tried deleting the whole directory and do a clean new sync, but with the same result
[02:07:13] <yoshi5534> i did the same as in the forum, syncing to another folder and symbolic link to my home directory
[02:07:39] <yoshi5534> like it is suggested in the documentation
[02:16:59] *** D|sT has joined #chromium-os
[02:17:29] *** wfrichar has quit IRC
[02:19:01] *** powderluv has quit IRC
[02:24:33] <zbehan> yoshi5534: so, you made a repo checkout somewhere on your disk, then created a symlink in your home to that directory?
[02:24:53] <yoshi5534> exactly
[02:28:00] <zbehan> yoshi5534: well, just to check if this is relevant to your issue, does the same thing happen if you create chroot inside the symlink, or inside the target itself?
[02:28:35] *** lipsinV2 has joined #chromium-os
[02:29:04] <zbehan> yoshi5534: how about filesystem permissions? i suppose you did the repo checkout in the other location, not in your home, so you should be having full permissions to that directory, correct?
[02:31:06] *** bpontes has quit IRC
[02:31:42] <yoshi5534> yes it happened in both cases, but now i mounted an additional disc directly in my home directory
[02:31:53] <yoshi5534> still syncing
[02:48:48] *** D|sT has quit IRC
[02:50:05] *** Keybuk has quit IRC
[02:52:55] <zbehan> yoshi5534: ok, let me know how that goes, i'd like to find out where does that come from, btw. what is your distro?
[02:56:30] <yoshi5534> ubuntu 11.04
[03:02:23] <yoshi5534> ok, that seemed to work
[03:02:45] <yoshi5534> but wondering about two messages in the beginning of cros_sdk script
[03:03:50] <yoshi5534>  /home/xxx/.bash_profile: line 6: cd: trunk/src/scripts: No such file or directory
[03:04:37] <yoshi5534>  /home/xxx/.bash_profile: line 7: bash_completion: No such file or directory
[03:06:49] <yoshi5534> but the rest worked fine
[03:07:33] <yoshi5534> its getting silent in this part of the cccamp, i think i'm going to get some sleep
[03:07:50] <yoshi5534> thanks for the help zbehan
[03:07:52] <yoshi5534> cya
[03:13:08] <zbehan> yoshi5534: glad it worked out fine for you
[03:25:28] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot" on "tegra2_seaboard-tangent canary" from None: )'
[03:30:09] *** D|sT has joined #chromium-os
[03:50:40] *** stevenjb has quit IRC
[03:54:30] <falken> I am looking at the build error
[03:55:09] <falken> maybe it is a flake because svn checkout seemed to fail
[03:55:24] <falken> I will try clobbering
[04:00:10] <crosbot> tree became '?'
[04:00:25] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot" on "tegra2_seaboard-tangent canary" from None: )'
[04:02:48] *** gfrog has joined #chromium-os
[04:07:36] *** BThompsonGR has joined #chromium-os
[04:08:34] <crosbot> tree became 'Tree is open (maybe a flake on tegra2_seaboard-tangent canary, error on "svn checkout", clobbered bot)'
[04:11:23] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot" on "arm generic PFQ" from 6db3f80ab78c4cb2d3b09378add10236dd412153: Daniel Erat <derat at chromium dot org>, _third_party_ at chromium dot org, bradnelson at google dot com <bradnelson at google dot com@78cadc50-ecff-11dd-a971-7dbc132099af>)'
[04:13:12] *** jrbarnette has quit IRC
[04:16:37] *** marklarson has joined #chromium-os
[04:23:53] <crosbot> tree became 'Tree is open ("arm generic PFQ" failure: the next build has advanced past where it failed on the previous build)'
[04:31:36] *** petermayo has quit IRC
[04:49:37] *** JasonO has quit IRC
[05:27:34] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot" on "x86 pineview full" from c2a87a2685f5e7da332879655b80f63229e8904f: Daniel Erat <derat at chromium dot org>, _third_party_ at chromium dot org, bradnelson at google dot com <bradnelson at google dot com@78cadc50-ecff-11dd-a971-7dbc132099af>, thakis at chromium dot org <thakis at chromium dot org@268f45cc-cd09-0410-ab3c-d52691b4dbfc>)'
[05:37:16] <crosbot> tree became 'Tree is open (flake in au_test_harness maybe http://crosbug.com/16940)'
[06:32:36] *** borios has joined #chromium-os
[06:49:42] <crosbot> tree became 'Tree is open (chrome sig11)'
[06:53:35] *** borios has quit IRC
[07:04:34] *** BThompsonGR has quit IRC
[07:25:31] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot" on "tegra2_aebl-binary" from bb63a23be1c08f39e3d6d0f457686f61c661b523: kkania at chromium dot org <kkania at chromium dot org@0039d316-1c4b-4281-b951-d872f2087c98>)'
[07:30:26] <falken> That's odd, the latest build was successful
[07:34:41] <crosbot> tree became 'Tree is open (the latest build was successful, not sure why tree closed.)'
[07:35:12] <crosbot> tree became 'Tree is open (the latest build was successful, not sure why tree died)'
[07:38:21] *** ud has joined #chromium-os
[07:38:34] <ud> hi
[07:39:18] <ud> what is the difference between amd64-generic and amd64-host?
[07:41:08] *** Sergiu has quit IRC
[08:00:16] *** ud has quit IRC
[08:34:20] *** ghosting has joined #chromium-os
[08:38:40] *** ghosting has quit IRC
[09:05:02] *** Styx has joined #chromium-os
[09:05:49] *** BladeFreak has joined #chromium-os
[09:07:24] *** jujugre has joined #chromium-os
[09:22:43] *** elly has quit IRC
[09:28:14] *** jochen__ has quit IRC
[09:28:19] *** jochen__ has joined #chromium-os
[09:28:19] *** ChanServ sets mode: +v jochen__
[09:32:50] *** vmil86 has joined #chromium-os
[09:44:40] *** elly has joined #chromium-os
[10:29:45] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot_master" on "x86 generic PFQ" from 33782c706aa5a8c299a49e0ef03232c88d145200: _third_party_ at chromium dot org)'
[10:50:10] <crosbot> tree became 'Tree is open (chrome sig11)'
[10:55:44] *** patcito has quit IRC
[11:15:30] *** Styx has quit IRC
[11:18:39] *** Styx has joined #chromium-os
[11:38:37] *** yusukes has quit IRC
[11:53:40] *** BladeFreak has quit IRC
[11:55:16] *** arun_ has quit IRC
[11:55:50] *** arun_ has joined #chromium-os
[12:00:28] *** BladeFreak has joined #chromium-os
[12:25:39] *** rosso has joined #chromium-os
[13:50:37] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot" on "tegra2_kaen-binary" from 44c13de1593cf08ab233699859108c0141075cd8: dubroy at chromium dot org <dubroy at chromium dot org@0039d316-1c4b-4281-b951-d872f2087c98>)'
[14:18:12] *** matjas has left #chromium-os
[14:57:06] *** yoshi5534 has quit IRC
[15:07:17] *** GodoPPL|Away has joined #chromium-os
[15:07:57] *** GodoPPL|Away is now known as GodoPPL
[15:25:59] *** rharrison_chrome has joined #chromium-os
[15:25:59] *** ChanServ sets mode: +v rharrison_chrome
[15:42:59] *** sadrul has quit IRC
[15:47:50] *** jochen__ has quit IRC
[16:05:21] *** varunjain has quit IRC
[16:05:27] *** varunjain has joined #chromium-os
[16:05:28] *** ChanServ sets mode: +v varunjain
[16:31:36] *** sadrul has joined #chromium-os
[16:31:37] *** ChanServ sets mode: +v sadrul
[16:47:00] *** saintlou has joined #chromium-os
[16:47:00] *** ChanServ sets mode: +v saintlou
[16:49:30] <kliegs> crosbot: sherrifs?
[16:49:49] <kliegs> crosbot: sheriffs?
[16:49:49] <crosbot> kliegs: sheriffs: davemoore, gabeblack, kochi, penghuang, falken
[16:56:39] *** cros_ has joined #chromium-os
[16:59:57] <crosbot> tree became 'Tree is closed (build chromeos-minijail failed on x86-alex canary -> penghuang)'
[17:00:32] *** phuang is now known as penghuang
[17:01:48] *** rbyers_ has joined #chromium-os
[17:04:11] *** jrbarnette has joined #chromium-os
[17:16:23] <crosbot> tree became 'Tree is open (x86-alex canary turned green)'
[17:17:01] *** jrbarnette has quit IRC
[17:21:45] *** grundler has joined #chromium-os
[17:21:45] *** ChanServ sets mode: +v grundler
[17:23:17] *** BladeFreak has quit IRC
[17:25:57] *** jrbarnette has joined #chromium-os
[17:35:11] *** jrbarnette_ has joined #chromium-os
[17:38:36] *** jrbarnette has quit IRC
[17:38:36] *** jrbarnette_ is now known as jrbarnette
[17:41:08] *** eggie has quit IRC
[17:44:51] *** Justasic has quit IRC
[17:46:44] <cros_> Are folks using the new cros_sdk script to build chroot? I just updated my depot_tools and am trying to use it but it's failing...one reason is it requires curl
[17:47:37] <ellyjones> no, I'm using make_chroot
[17:47:52] <Stepan> i only used cros_sdk to enter chroots so far
[17:47:53] <ellyjones> basically you want to always do what the buildbots do, because all the other flows might break randomly ;)
[17:48:13] <cros_> ellyjones: Ok, I don't know why they changed from make_chroot but the instructions in the developer guide already switched to say use cros_sdk
[17:48:33] <cros_> ellyjones: And I don't know anything about the buildbot stuff so I wouldn't know how to figure out what buildbot is doing.
[17:50:34] <ellyjones> weird, I dunno
[17:54:26] *** jujugre has left #chromium-os
[17:55:01] <ellyjones> penghuang: that failure you linked me to: ???
[17:55:48] *** Justasic has joined #chromium-os
[17:56:26] <penghuang> ellyjones, yeah. do you have any idea?
[17:57:18] *** Justasic_ has joined #chromium-os
[17:57:27] <penghuang> ellyjones, http://chromeos-botmaster.mtv.corp.google.com:8026/builders/x86-alex%20canary/builds/851/steps/Test/logs/stdio
[17:57:35] <ellyjones> penghuang: the error message indicates a discrepancy in whether NDEBUG is used for building libbase and minijail
[17:58:02] <ellyjones> which is disturbing, as is the fact that it built ok before that
[17:59:10] <ellyjones> it failed in the same way at 850
[17:59:10] <ellyjones> hm
[18:00:12] *** Justasic has quit IRC
[18:01:03] <ellyjones> there is no plausible commit that would have fixed it, and 851 shows prior successful builds of it
[18:04:18] *** saintlou has quit IRC
[18:04:38] *** saintlou has joined #chromium-os
[18:04:45] *** rbyers_ has quit IRC
[18:04:46] <penghuang> yeah. but the update_engine is failed on similar issue
[18:05:14] <ellyjones> I don't have any ideas, but I doubt it is minijail-specific
[18:06:12] *** rbyers_ has joined #chromium-os
[18:06:47] *** saintlou has quit IRC
[18:10:46] *** Justasic_ has quit IRC
[18:11:18] *** Justasic has joined #chromium-os
[18:19:35] *** vmil86 has quit IRC
[18:24:19] <cros_> Using cros_sdk to setup your chroot using prebuilt binaries from the servers works once you install curl. Trying to build from your using the --bootstrap option seems broken.
[18:46:40] <grundler> msb__, you need any help with the "crashes not saved" bug ?
[18:46:53] <grundler> I'll have some time to help you with that if I can.
[18:54:24] *** benchan has joined #chromium-os
[18:54:24] *** ChanServ sets mode: +v benchan
[18:57:47] *** lipsinV2 has quit IRC
[19:28:52] *** dnicoara has quit IRC
[19:30:08] <Hexxeh> Is the x86-fruitloop overlay abandoned now?
[19:30:58] <ellyjones> fruitloop? I don't think we've had a buildbot for that for a while
[19:31:05] <ellyjones> any overlay that doesn't get regularly built will break
[19:31:53] <Hexxeh> it's not too different from generic, appears it was added to test nvidia support
[19:32:12] <Hexxeh> considering overhauling it if it's abandoned, and using it for my vanilla builds
[19:32:55] <Hexxeh> disable pae, get nvidia working (i've no idea if it works on fruitloop at the moment, so this might not even be required)
[19:39:11] *** rosso has quit IRC
[19:40:39] *** JasonO has joined #chromium-os
[19:40:39] *** JasonO has joined #chromium-os
[19:46:02] <satorux_> I tried to build chrome for chrome os inside the chroot but failed
[19:46:25] <satorux_> KeyError: 'src/v8/tools/gyp/v8.gyp:js2c#host while trying to load target src/v8/tools/gyp/v8.gyp:v8_nosnapshot#target'
[19:46:25] <satorux_> Error: /usr/bin/python2.6 src/build/gyp_chromium in /home/satorux/chrome_root returned 1
[19:46:30] <satorux_> anyone seeing the same error?
[19:50:19] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot_master" on "TOT Pre-Flight Queue" from 91026204f22a8fa4c0dca3d0dfcaf54d4a1da481: tfarina at chromium dot org <tfarina at chromium dot org@0039d316-1c4b-4281-b951-d872f2087c98>)'
[19:56:33] <satorux_> Changing v8 version from 8902 to 8765 has fixed the problem for me. I'll be sending out mail to chromium-os-dev
[19:58:06] *** Keybuk has joined #chromium-os
[19:58:06] *** ChanServ sets mode: +v Keybuk
[20:00:22] *** oc80z has quit IRC
[20:01:17] *** oc80z has joined #chromium-os
[20:04:10] *** JasonO has quit IRC
[20:04:39] *** JasonO has joined #chromium-os
[20:04:39] *** JasonO has joined #chromium-os
[20:09:29] <crosbot> tree became 'Tree is closed (x86 generic TOT Pre-Flight Queue -> penghuang)'
[20:10:11] <davidjames> penghuang: Notice also stumpy-binary is about to fail due to chromeos-login unit test failure :(
[20:10:42] *** patcito has joined #chromium-os
[20:11:35] *** davemoore__ has joined #chromium-os
[20:11:35] *** ChanServ sets mode: +v davemoore__
[20:12:29] <penghuang> davidjames, :( Did you find out which change caused the problem?
[20:15:00] <davidjames> penghuang: Nope...
[20:18:03] *** wfrichar has joined #chromium-os
[20:18:03] *** ChanServ sets mode: +v wfrichar
[20:20:37] *** jrbarnette_ has joined #chromium-os
[20:23:21] *** jrbarnette has quit IRC
[20:24:59] *** jrbarnette_ has quit IRC
[20:29:58] <satorux_> davidjames: your mention of chromeos-login worries me, as I made a change related to chromeos-login
[20:30:15] <satorux_> what's happening with chromeos-login?
[20:30:34] <davidjames> satorux_: http://chromeos-botmaster.mtv.corp.google.com:8026/builders/x86-zgb-binary/builds/820/steps/Test/logs/stdio
[20:31:05] <davidjames> satorux_:  SessionManagerTest.SetOwnerKeyShouldFail fails
[20:31:59] <satorux_> davidjames: I doubt my change is related to this failure, but fwiw, the change I'm talking about is http://gerrit.chromium.org/gerrit/#change,5854
[20:32:23] <satorux_> I meant: I doubt -> I don't think
[20:33:20] <davidjames> satorux_: It's complaining about the SetOwnerKeyComplete pointer, namely that although it's got the right value it's pointing at a different location in memory
[20:33:27] *** D|sT has quit IRC
[20:34:16] <satorux_> davidjames: i see. My change only touches string constants, so It's unlikely to cause an error like that
[20:34:40] <davidjames> satorux_: Right, but it's complaining about a string constant in the failure message, namely that the string constant changed locations in memory and that's what the test is unhappy about
[20:35:09] <satorux_> let me take a closer look...
[20:35:44] <davidjames> satorux_: I think your change would be likely to move the memory location of the string because it's inlined now
[20:36:05] <satorux_> davidjames: that's a good point
[20:36:28] <satorux_> it's weird that the test depends on memory location of string contants
[20:36:28] <davidjames> satorux_: I think that's the unit tests fault, not yours, but time to fix the unit test? :)
[20:36:52] <satorux_> i'll check if cmasone is here
[20:40:42] *** rginda_home has joined #chromium-os
[20:42:58] <penghuang> kkania reverted some changes. wish TOT PFQ will turn green
[20:43:34] *** jrbarnette has joined #chromium-os
[20:43:35] *** ChanServ sets mode: +v jrbarnette
[20:45:19] *** kmixter has joined #chromium-os
[20:46:03] *** ChanServ sets mode: +v kmixter
[20:46:47] <satorux_> davidjames: not sure what would be the best way to fix this test. one way to do it would be to change the function to take std:string, instead of const char*. then gmock won't compare by pointers
[20:48:06] *** rginda_home has quit IRC
[20:48:11] *** rginda_home has joined #chromium-os
[20:48:41] <quiche> satorux_: can you use StrEq in the EXPECT_CALL ?
[20:49:47] <satorux_> quiche: thanks. that sounds right!
[20:49:56] <satorux_> I'm no expert of gmock
[20:50:11] <satorux_> let me see how to use that trick
[20:50:32] <quiche> good luck. :)
[20:51:08] <satorux_> ah, adding StrEq() to the parameter in question should suffice
[20:51:22] *** gabeblack has joined #chromium-os
[20:51:22] *** ChanServ sets mode: +v gabeblack
[20:51:24] <satorux_> I'm going to create a ptch
[20:51:36] *** D|sT has joined #chromium-os
[20:52:25] *** sergiu has joined #chromium-os
[20:52:26] *** sergiu has joined #chromium-os
[21:00:22] <cros_> Is there anyway to run the unit tests from the target directly rather than remotely using the run_remote_tests.sh?
[21:00:52] <srao> cros_: yes ssh to the machine and run the control file
[21:01:32] <srao> actually that's not quite right, you don't just run the control file... there's a /usr/local/autotest/bin/ directory let me look
[21:02:14] <cros_> srao: I want to avoid ssh. Can I run from local console on the target?
[21:02:36] <srao> sure
[21:02:47] <srao> you have to have the test over there first obviously
[21:03:01] <satorux_> davidjames, quiche: fixed the issue locally. now under code review by kmixter
[21:03:47] <cros_> srao: That's what I wasn't sure. In reading the developer guide it seems the tests get sent from the host normally?
[21:04:07] <srao> yeah running the run_remote_tests will copy it over there initially
[21:04:45] <cros_> srao: What does a test image actually add to the final image if the tests aren't part of the filesystem?
[21:04:57] *** D|sT has quit IRC
[21:05:21] <srao> cros_: trying to remember off hand, it adds sshd, some utilities needed for the tests as well
[21:05:31] *** D|sT has joined #chromium-os
[21:06:00] <satorux_> davidjames, quinche: submitted
[21:06:07] <srao> cros_: my guess is that it doesn't copy all of them over because that would take a large amount of space
[21:06:26] <cros_> srao: Appreciate it, I'll give it a try.
[21:06:54] <srao> cros_: so once it's been copied over by run_remote tests, you run /usr/local/autotest/bin <test control file>
[21:07:00] <srao> oops
[21:07:10] <srao> /usr/local/autotest/bin/autotest <control file>
[21:07:22] *** patcito has quit IRC
[21:07:25] *** patcito_ has joined #chromium-os
[21:08:53] *** patcito_ is now known as patcito
[21:08:57] <cros_> srao: What all would I need to copy manually if I wanted to avoid running the tests remotely?
[21:09:11] <cros_> srao: I'm looking in my build directory and I see a bunch of stuff under usr/local/autotest
[21:13:37] *** sbyer has joined #chromium-os
[21:13:38] *** ChanServ sets mode: +v sbyer
[21:17:04] *** dianders has left #chromium-os
[21:23:37] <ellyjones> hm
[21:23:45] <ellyjones> I think I just double-scumbag-ellyjonesd
[21:23:49] <ellyjones> 1) committed on closed tree
[21:23:57] <ellyjones> 2) landed change to boot configuration 45m before I leave work on a friday
[21:27:52] <grundler> ellyjones, you could up the ante by taking vacation all of next week :)
[21:28:13] <ellyjones> I could
[21:28:24] <ellyjones> these vacation days are burning a hole in my metaphorical pocket :P
[21:30:21] * davidjames wishes at least one builder would run tests, but it's not possible until PFQ pushes satorux_'s change...
[21:31:39] <davidjames> Bingo, PFQ pushed satorux_'s change! Any builders that start now should cycle green...
[21:31:44] *** dianders has joined #chromium-os
[21:31:45] *** ChanServ sets mode: +v dianders
[21:31:48] *** dianders has quit IRC
[21:32:08] *** dianders has joined #chromium-os
[21:32:09] *** ChanServ sets mode: +v dianders
[21:32:44] <crosbot> tree became 'Tree is closed (canaries and full builders cycling red, then green -> penghuang)'
[21:35:47] *** bpontes has joined #chromium-os
[21:35:47] *** ChanServ sets mode: +v bpontes
[21:45:58] <davidjames> ellyjones: As predicted, tree broke with your change :)
[21:46:11] <ellyjones> please tell me you are joking
[21:46:21] <davidjames> ellyjones: Nope. /usr/lib/installer/chromeos-common.sh: line 368: sudo: unbound variable
[21:46:25] <ellyjones> FFFFFFFFFFFFFFFf
[21:46:27] <ellyjones> are you serious
[21:46:28] <davidjames> ellyjones: http://build.chromium.org/p/chromiumos/builders/x86%20generic%20PFQ/builds/256/steps/BuildTarget/logs/stdio
[21:46:52] <ellyjones> I'll put a fix up quick
[21:46:52] <davidjames> ellyjones: Might be time to revert, tree was closed when you committed anyway :)
[21:47:00] <ellyjones> yeah
[21:47:04] <ellyjones> well, the fix is trivial
[21:47:13] <ellyjones> and a lesson was learned: even a one-character change is not too small to test
[21:47:16] <ellyjones> sigh
[21:48:32] <jrbarnette> yup, i just notice the CL, and was worried it was going to break things
[21:48:57] <jrbarnette> i'd just started up trying to test it...
[21:49:01] <ellyjones> that wasn't even the main part of my CL, that was part of my cleanup
[21:49:03] <ellyjones> sigh
[21:49:58] <jrbarnette> word to the wise:  you have *no idea* how pernicious the sharing of things like partoffset() has gotten :-(
[21:50:36] <ellyjones> jrbarnette: http://gerrit.chromium.org/gerrit/5908
[21:50:51] <ellyjones> in lieu of a revert
[21:50:55] <redpig> I still wish we were doing image building, updates, etcv via a library
[21:51:03] <redpig> libextfs, libcgpt, etc :/
[21:51:10] <ellyjones> redpig: hm, that is a good idea
[21:51:23] <jrbarnette> i've been breaking up build_image along those lines
[21:52:05] <jrbarnette> it's crossed my mind that we could do a better separation of responsibility for the installer
[21:52:12] <ellyjones> filed crosbug.com/19118 so I don't forget about the cleanup
[21:52:14] <redpig> jrbarnette: nice! I have skeleton code I wrote just before the cr48 launch but never had the chance to make it real
[21:52:25] <ellyjones> jrbarnette: fix landed
[21:52:34] <redpig> jrbarnette: ah yeah. there are a lot of pieces that should be done differently and in a way we can test
[21:52:39] <davidjames> ellyjones: ${sudo:-sudo} would also work
[21:52:52] <ellyjones> davidjames: yeah
[21:53:01] <ellyjones> davidjames: but undoing what I did is the lowest-risk way to do it :P
[21:53:25] <jrbarnette> that makes me nervous; any old program that says "local sudo=<something" could suddenly change its behavior
[21:53:44] <ellyjones> isn't that the intent?
[21:54:08] *** cwolfe has quit IRC
[21:54:16] <jrbarnette> only if you've examined every client of partoffset and partsize, and demonstrated that it's a better behavior
[21:54:36] <jrbarnette> even then, it's risky, since we don't know what new clients we might pick up :-(
[21:54:43] <ellyjones> sigh
[21:54:53] <ellyjones> shell, y u no have proper encapsulation support
[21:54:58] <jrbarnette> i feel your pain
[21:55:29] <ellyjones> doing 'libraries' by textual inclusion is insane :P
[21:55:41] <ellyjones> hey redpig can we use python instead? ;)
[21:56:07] <jrbarnette> yeah, although the real issue is overdependency on global variables
[21:56:44] <jrbarnette> ... and dynamic scoping of local variables, just for funsies
[21:57:11] <ellyjones> dynamic scoping of locals was realized to be a good idea just slightly too late
[21:57:55] <jrbarnette> um, do you mean "bad idea"?  and the LISP community understood the issue back in the '70s
[21:58:05] <ellyjones> er, yes, bad idea
[21:58:11] <ellyjones> and yes, but not soon enough to save elisp
[21:58:38] <jrbarnette> ... or any other LISP outside of Scheme, but i digress
[21:58:50] <redpig> ellyjones: bash can be namespaces, posix shell, not so much
[21:59:14] <jrbarnette> namespaces in bash, what?
[21:59:21] <ellyjones> jrbarnette: wait, CL totally has lexical scope, doesn't it?
[21:59:24] <jrbarnette> that's a trick i'd like to learn more of...
[21:59:34] <ellyjones> redpig: how can bash be namespaced? oO
[21:59:40] <redpig> jrbarnette: still a hack - lemme dust off some stuff and see how wrong i am :)
[22:00:15] <jrbarnette> 'local' creates variables with dynamic, not lexical scope
[22:00:24] <davidjames> Perl supports dynamic scoping via local too: http://perldoc.perl.org/perlsub.html#Temporary-Values-via-local()
[22:00:45] <redpig> jrbarnette: so in bash you can use :: in function names to create working namespaces
[22:00:52] <redpig> like
[22:00:56] <redpig> function cryptohome::is_mounted() { return 1; }
[22:01:16] <redpig> I built a small library management system using that mechanism a few years ago
[22:01:20] <ellyjones> but you alway shave to reference them with fully qualified names, right?
[22:01:29] <redpig> yeah but you can alias in bash too :p
[22:01:44] <jrbarnette> yeesh, but how does that help with variable?
[22:01:59] <redpig> well you have to use local and not write code outside a function scope
[22:02:22] <jrbarnette> ... but 'local' isn't lexically scoped...
[22:02:39] <ellyjones> hrm
[22:02:45] <redpig> hrm does it bleed into sub-calls?
[22:02:46] <ellyjones> so you can do this, using what drewry said:
[22:03:00] <ellyjones> foo::bar::lol() { return 1; } x=foo::bar; $x::lol
[22:03:09] <ellyjones> which mitigates the total loss of having to write foo::bar::lol each time
[22:03:34] <redpig> and bash let's you even use variable values ot name another variable :p
[22:03:36] <redpig> which is good times
[22:03:53] <jrbarnette> : custard.mtv.corp ~; function what_fun() { echo $fubar ; }
[22:03:54] <jrbarnette> : custard.mtv.corp ~; function no_fun () { local fubar="no fun" ; what_fun ; }
[22:03:54] <jrbarnette> : custard.mtv.corp ~; fubar="this is fun"
[22:03:54] <jrbarnette> : custard.mtv.corp ~; no_fun
[22:03:54] <jrbarnette> no fun
[22:03:57] <redpig> like ${!othervar}  iirc
[22:04:06] <ellyjones> redpig: yeah, but foo::bar::baz=5 does not work
[22:04:18] *** cwolfe has joined #chromium-os
[22:04:18] *** ChanServ sets mode: +v cwolfe
[22:04:56] <redpig> jrbarnette: for that scoping requires a subshell ;)
[22:05:21] <redpig> bash-3.2$ fubar="this is fun"
[22:05:21] <redpig> bash-3.2$ (no_fun)
[22:05:21] <redpig> no fun
[22:05:21] <redpig> bash-3.2$ echo $fubar
[22:05:21] <redpig> this is fun
[22:05:26] <redpig> which if not fun :)
[22:05:28] <redpig> err is
[22:06:22] <redpig> jrbarnette: however
[22:06:31] <redpig> I think functions still isolate
[22:06:37] <redpig> bash-3.2$ function a() { local x=1; echo "I call b; x=$x"; b; echo "x=$x"; }
[22:06:41] <redpig> bash-3.2$ function b() { local x=2; echo "B sez 2"; }
[22:07:40] <Keybuk> the trick to remember is that bash doesn't use scope
[22:08:00] <Keybuk> instead it's more like "change environ[] from here on"
[22:09:13] <redpig> fair enough - that makes the functions logical too since they get treated like new invocations to each other even if they aren't subshells
[22:16:57] *** achuith has joined #chromium-os
[22:16:57] *** ChanServ sets mode: +v achuith
[22:19:13] *** grundler has quit IRC
[22:24:27] *** GodoPPL has quit IRC
[22:28:38] <jrbarnette> redpig: n.b. this example
[22:28:48] <jrbarnette> function a() { local x=1; echo "I call b; x=$x"; b; echo "x=$x"; }
[22:29:00] <jrbarnette> function b() { x=2; echo "B sez 2"; }
[22:29:13] <jrbarnette> a
[22:29:14] <jrbarnette> I call b; x=1
[22:29:14] <jrbarnette> B sez 2
[22:29:14] <jrbarnette> x=2
[22:32:15] *** benchan has quit IRC
[22:36:10] <davidjames> crosbot: stat sheriffs
[22:36:11] <crosbot> davidjames: sheriffs: davemoore, gabeblack, kochi, penghuang, falken
[22:37:17] <davidjames> sheriffs: Time to open the tree?
[22:39:03] <penghuang> PFQ issue has been fixed. How about another build test problem?
[22:39:37] <jrbarnette> penghuang:  i can give you another build test problem :-)
[22:39:39] <davidjames> penghuang: The binary builders are all passed unit tests by now
[22:39:57] <jrbarnette> ... but maybe you meant "is there another build test problem?"...
[22:40:26] <penghuang> :)
[22:40:44] <penghuang> I think we can open it now
[22:41:56] <crosbot> tree became 'Tree is open (some bots already turned green)'
[22:52:06] *** BladeFreak has joined #chromium-os
[23:05:54] *** Styx has quit IRC
[23:14:53] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot" on "stumpy canary" from None: )'
[23:15:18] <redpig> jrbarnette: my bash didn't do that ;)
[23:15:36] <redpig> jrbarnette: ah you unlocal'd x
[23:15:49] <jrbarnette> that's why your bash didn't do that
[23:15:51] <redpig> jrbarnette: yeah mistakes and bad practices can caust that
[23:16:25] <crosbot> tree became 'Tree is closed (gabeblack is investigating)'
[23:16:26] <jrbarnette> agreed; the point was that discipline and defined coding conventions are mandatory
[23:16:38] <redpig> yup :/
[23:17:32] *** rharrison_chrome has quit IRC
[23:22:25] <gabeblack> I'm investigating the tree closure
[23:22:32] <gabeblack> it's a test failure, SessionManagerTest.SetOwnerKeyShouldFail
[23:22:52] <gabeblack> in chromeos-login
[23:23:04] *** mtl has joined #chromium-os
[23:23:19] <gabeblack> Is this a known issue?
[23:24:24] <srao>  "<davidjames> penghuang: Notice also stumpy-binary is about to fail due to chromeos-login unit test failure :("
[23:24:33] <srao> sounds like it could be the issue they just fixed
[23:27:30] *** cros_ has quit IRC
[23:29:51] <davidjames> srao: Yeah canaries take longer to fail
[23:30:10] <gabeblack> So do you think I should reopen the tree?
[23:30:23] <davidjames> gabeblack: Yes
[23:30:43] *** dkrahn has joined #chromium-os
[23:30:43] *** ChanServ sets mode: +v dkrahn
[23:30:58] <crosbot> tree became 'Tree is open'
[23:31:30] <davidjames> gabeblack: TOT preflight queue looks unhappy: http://chromeos-botmaster.mtv.corp.google.com:8026/builders/TOT%20Pre-Flight%20Queue/builds/4558/steps/BuildTarget/logs/stdio
[23:33:08] *** powderluv has joined #chromium-os
[23:33:08] *** ChanServ sets mode: +v powderluv
[23:33:41] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot_master" on "TOT Pre-Flight Queue" from 8b33ed9307f04ceda8ff058500c68e96efa4fc7a: gbillock at chromium dot org <gbillock at chromium dot org@0039d316-1c4b-4281-b951-d872f2087c98>)'
[23:33:46] * powderluv urges folks to start using cros_sdk instead of ./enter_chroot.sh
[23:33:52] <gabeblack> It looks like this has been failing for a while
[23:34:34] <davidjames> gabeblack: Yeah I saw that show up earlier and then disappear
[23:42:44] <gabeblack> davidjames: It's back again
[23:43:22] *** dkrahn has left #chromium-os
[23:44:29] *** piman_ has quit IRC
[23:45:24] *** dkrahn has joined #chromium-os
[23:45:25] *** ChanServ sets mode: +v dkrahn
[23:45:49] *** seanpaul has quit IRC
[23:49:52] <davidjames> gabeblack: That looks like it was introduced with satorux_ 's change
[23:50:37] <davidjames> gabeblack: the cromo ebuild needs to be rev'd so that its static library gets rebuilt... revving the ebuild should fix the problem
[23:51:35] <gabeblack> revving the chromo ebuild?
[23:51:48] <davidjames> gabeblack: cromo





top