[10:57] <jibel> hggdh, https://code.launchpad.net/~jibel/ubuntu-server-iso-testing/live_session/+merge/72168
[10:58] <jibel> it was more difficult to integrate to usit than expected, I'd appreciate a review
[12:12] <hggdh> jibel: looking
[12:14] <jibel> hggdh, good morning, no emergency :)
[12:14] <hggdh> jibel: good afternoon, no problem :-)
[12:29] <jibel> I need someone to reproduce bug 829370 on an oneiric box upgraded from natty
[12:29] <jibel> the test is to enable accessibility, reboot, login and check if the desktop crashes or not.
[12:47] <hggdh> jibel: must be upgraded from natty?
[12:50] <jibel> hggdh, I'm not sure that's a key component of the crash. You can try even on a non-upgraded system I think.
[12:50] <hggdh> k
[12:54] <hggdh> jibel: there seems to be another issue on accessability -- I requested flashing of the window title and whole window, then clicked on 'test'. Nothing happened
[12:54] <hggdh> meanwhile, rebooting
[12:57] <hggdh> jibel: first thing on the change: at the beginning you glob.glob a directory, and then 'cp -rf "the return from glob" newdir'
[12:58] <hggdh> you shoudl instead do a for entry in glob,glob cp -rf yadda
[12:59] <hggdh> if glob.glob returns more than one entry, it will do the wrong thing right now
[12:59] <jibel> hggdh, why couldn't I copy more than 1 entry to destdir at a time ?
[12:59] <hggdh> jibel: cp a b c d e f?
[13:00] <jibel> the goal *is* to copy more than one entry, and there's nothin in python to do that if target dir already exists
[13:00] <hggdh> of course, it will work, the last entry in the cp is a dir
[13:00] <hggdh> duh
[13:01] <hggdh> should not do that before drinking coffee
[13:01]  * jibel rereading his code
[13:03] <jibel> hggdh, you're right that would fail if target was a file, but it's the result from copytree above
[13:03] <jibel> which would fails if it is a file and not a directory.
[13:03] <hggdh> jibel: yes
[13:04] <jibel> I could add an extra test and exit if it is not a directory.
[13:04] <hggdh> but jibel just for sanity, it might be good to test
[13:04] <hggdh> heh
[13:04] <hggdh> yes :-)
[13:05] <jibel> this is really a corner case, like dpkg creating a file instead of a directory which is doubtful
[13:06] <jibel> of a I/O error or something bad like this.
[13:07] <jibel> lp hates me today and keeps crashing on my script.
[13:08] <jibel> hggdh, did you reboot with a11y enabled finally ?
[13:11] <hggdh> jibel: just did
[13:11] <hggdh> it is slow, the netbrick
[13:12] <jibel> hggdh, anything in /var/crash ?
[13:12] <hggdh> jibel: it did not crash, but I was not up-to-date; I am running a dist-upgrade now
[13:12] <hggdh> jibel: on the other hand...
[13:13] <hggdh> ally seems a bit hosed there
[13:13] <jibel> hggdh, no *at-spi2-core* or *unity-panel* crash file ?
[13:13] <hggdh> jibel: absolutely no crash file, of any name
[13:14] <hggdh> but wait for the dist-upgrade to end, and a new reboot (also remember, this is a fresh install of oneiric)
[13:16] <jibel> hggdh, I'll try on a fresh install on the same system and see if it doesn't crash,  that would be the upgrade then, some lib left behind or alike.
[13:16] <jibel> hggdh, many thanks for testing
[13:17] <hggdh> jibel: not done yet, now rebooting after dist-upgrade, with high contrast enabled
[13:18] <hggdh> I only have oneiric right now on a netslug
[13:23] <hggdh> jibel: seems to have crashed
[13:23] <jibel> \o/
[13:24] <hggdh> but it was compiz...
[13:24] <jibel> easy target ;-)
[13:26] <hggdh> trying again
[13:28] <hggdh> jibel: weird: compiz *is* running, unity is not, but no crash file
[13:29] <hggdh> ah, hell, the dist-upgrade removed unity
[13:29] <hggdh> from the /var/log/apt/history.log:
[13:29] <hggdh> Remove: unity:amd64 (4.8.0-0ubuntu3), unity-2d:amd64 (4.0.0-0ubuntu1), libunity-2d-private0:amd64 (4.0.0-0ubuntu1), unity-2d-spread:amd64 (4.0.0-0ubuntu1), unity-2d-panel:amd64 (4.0.0-0ubuntu1), unity-2d-places:amd64 (4.0.0-0ubuntu1), libgjs0b:amd64 (1.29.0-1ubuntu1), unity-2d-launcher:amd64 (4.0.0-0ubuntu1)
[13:30] <jibel> hggdh, simple fix to many problems
[13:30] <hggdh> yes, so true...
[13:32] <hggdh> ok, trying again
[13:35] <hggdh> jibel: I cannot reproduce -- unity is again installed, and I am running with high-contrast. Session is opened OK, no crashes
[13:35] <jibel> hggdh, and orca enabled and running ?
[13:36] <hggdh> no, orca is not running.
[13:36] <hggdh> will set it to speak, and reboot/login
[13:37] <hggdh> ls -l /var/crash
[13:38] <hggdh> _usr_lib_at-spi2-core_at-spi-bus-launcher.102.crash
[13:38] <jibel> hggdh, cool, that's the one
[13:38] <hggdh> jibel: there you go. Did not even need to reboot. As soon as I enabled orca, it barfed
[13:39] <jibel> hggdh, if you logout, login that should crash unity-panel
[13:39] <hggdh> will do it now
[13:39] <jibel> hggdh, it is a fresh installation, isn't it ?
[13:40] <jibel> hggdh, could you add a comment on bug 829370, please ?
[13:40] <jibel> I'll assign it to the desktop team
[13:41] <hggdh> will do it now. And yes, unit-panel now crashed
[13:42] <jibel> hggdh, so please confirm but 829367 as well. Thanks!
[13:47] <hggdh> jibel: done, and done
[13:48] <hggdh> jibel: and yes, it is a fresh install (fresh meaning it was an Oneiric install, no upgrade)
[14:08] <jibel> that new switcher is driving me mad. No one on DX has ever worked with spreadsheets I guess.
[15:43] <hggdh> jibel: there still?
[15:43] <jibel> hggdh, Yep on #u-meeting
[15:43] <hggdh> k, I will wait
[15:43] <jibel> hggdh, how can I help ?
[15:44] <hggdh> in templates.desktop/test_cases/livesession/custom-installation/hooks/casper-bottom.sh
[15:44] <hggdh> you start with "If you change this, please also change the copy in ubiquity-hooks/30accessibility."
[15:44] <jibel> hggdh, bah, copy/paste failure
[15:44] <hggdh> ah, OK :-)
[15:44] <hggdh> apart from that, go for it
[15:45] <jibel> hggdh, I saw that but forgot to remove it. I'll do, thanks for the reminder
[15:45] <hggdh> and I just approved it
[16:36] <jibel> charlie-tca, morning
[16:36] <jibel> charlie-tca, bug 829370
[16:36] <charlie-tca> looking
[16:37] <jibel> charlie-tca, do you know if it's known or in progress or any status ?
[16:37] <jibel> I reported it today but maybe that's expected at this stage
[16:38] <charlie-tca> I don't know what is going on with it. I will check with TheMuso tonight when he comes on
[16:39] <charlie-tca> I don't have at-spi2 on this system because dasher still requires the older version
[16:39] <jibel> charlie-tca, okay thanks, I'm usually offline when he comes on.
[16:40] <charlie-tca> No problem. Thanks for bringing it up.
[16:40] <jibel> charlie-tca, the problem being that on a fresh install orca doesn't work because of this crash
[16:40] <charlie-tca> Even in unity 2d?
[16:40] <charlie-tca> accessibility is going to be using 2d instead of 3d
[16:43] <jibel> charlie-tca, I haven't tried.
[16:43] <charlie-tca> To the best of my knowledge, we don't care if orca works in 3d now
[16:43] <jibel> charlie-tca, do you mean the if a user selects "a11y" on the  boot menu he will be logged to unity-2d and this settings will be kept after installation ?
[16:43] <charlie-tca> desktop images are broken again, so can't try it in live sessions
[16:43] <charlie-tca> yes
[16:44] <jibel> charlie-tca, okay, good to know. I'll test on 2d then
[16:44] <charlie-tca> a11y and accessibility installs will default to 2d
[16:44] <jibel> charlie-tca, that lower the importance a bit even if it's still high.
[16:44] <charlie-tca> 3d just doesn't work for it
[16:44] <charlie-tca> why?
[16:44] <charlie-tca> it still is broken if it won't work
[16:45] <charlie-tca> why are those needing that second-class users
[16:46] <charlie-tca> Well, I guess if it works in 2d but not 3d, yes, it is lower priority
[16:46] <jibel> charlie-tca, I mean if it works on 2d and the user is logged there by default, it is still a crash, but less critical since by default the user has a working environment
[16:46] <charlie-tca> right
[16:46] <charlie-tca> sorry
[16:46] <jibel> no worries :)
[16:46] <charlie-tca> fingers type, then brain engages
[16:46] <jibel> I'll test on 2d, to be sure it's working
[16:46] <charlie-tca> Thank you
[16:47] <charlie-tca> I will point the bug to TheMuso, anyway, tonight
[16:47] <jibel> yes please. Thanks