[01:05] hey guys, while you're running those jaunty live cds, can you follow the directions on https://wiki.ubuntu.com/Audio/AlsaInfoOutput ? [03:04] hi everyone, i'm trying to install ubuntu netbook remix 9.04 but ubiquity does not see my hard drive [03:05] does anyone can help me? [03:14] ubuntu detect it and mount it correctly but the partitionner does not show anything [03:20] that could well just be an installer semi-crash [03:20] file a bug and post /var/log/syslog, /var/log/partman, and /var/log/installer/debug [03:23] ok i'll check that [03:24] thanks [03:28] I just uploaded ubiquity 1.11.13, including some fixes from a colleague that might address this [03:30] is there is a way to test it now? [03:31] if i can download it, i can tell you if it works for me [04:00] i installed ubiquity from the trunk and it did not work :< [07:31] good morning everybody! [08:37] ara: so how did the first testing day go other than unr and lvm not working ? [08:38] davmor2: hehehehe [08:38] davmor2: it went ok. kirkland was very responsive for his screen-profiles bugs [08:38] davmor2: he fixed (and released) some [08:39] davmor2: I was happy to see new people coming over [08:39] davmor2: the lvm bug was also fixed, wasn't it? [08:40] ara: I don't know if it was the big bug which mean that it won't boot but the bad numbering should be now [08:40] davmor2: cool [08:40] davmor2: thanks for testing that :) [08:41] ara: did you have an issue with the wording in screen for changing from light to dark ? [08:41] wording? [08:41] what do you mean? [08:46] let me get the screen up [08:52] ara: right run screen then run select-screen-profile change to another one and it says 'Run "screen" to activate' if you follow those instructions it does nothing. You need to shutdown the terminal your in then open a new one then run screen for it to take effect [08:53] davmor2: are you running the latest? today there was an update [08:53] I'll try it in tty [08:53] no that is from yesterday I'll try it in a second [09:03] ara: Yay that's better :) [09:03] see I knew it was just the wording :) [09:03] davmor2: :) [09:05] davmor21: hello evil twin === mvo__ is now known as mvo [09:08] davmor21: moo ping and other things [09:09] Bar humbug no notifyosd [09:36] So notifyosd only works with chat and not irc. So in order to test you just need 2 machines and bonjour setup [10:10] bdmurray: When your up working on vfat install now === mvo__ is now known as mvo [10:50] D'oh forgot windows fat can't handle drives over 32 GB [12:37] * ara -> lunch [13:30] davmor2: ping === Panattan is now known as Kilin === Kilin is now known as kyrin === kyrin is now known as kylin === kylin is now known as kilin [13:59] Ah, the updated ubiquity now gets past the whole partition setup and into the installation phase :) vast improvement over yesterday. [14:01] \o/ [14:06] ara: pong [14:19] eek! ubiquity blew up at ~90% install. getting apport stuffs now. [14:20] meh. It won't complete collection because apparently grub is out of date now. [14:24] cjwatson: ^ [14:25] rmcbride: could I get a syslog? [14:25] rmcbride: get the logs together for cjwatson [14:25] I just did dist-upgrade to put the newer grub into place. I'll get the logs together as well [14:27] cjwatson: I'm about to start a bunch of tests too [14:27] I'll see if I can confirm [14:29] cjwatson: syslog is at https://pastebin.canonical.com/14161 [14:30] ok, though I'd recommend using paste.ubuntu.com in future where it doesn't need to be private [14:30] ok, that's the bug evand just fixed in bzr [14:30] cjwatson: didn't know about that one. thanks. [14:30] OK cool [14:31] rmcbride: delete the seds of $adjtimefile in /usr/lib/ubiquity/clock-setup/finish-install if you want to get past it [14:31] cjwatson: will do, thanks [15:15] Argh. Installation completes but on reboot, grub throws error 22... if this is the same thing I ran into when I last built this machine I think it's because grub gets confused as to the partition numbers when being installed from USB stick in some cases. I should be able to boot off the live cd and re-edumacate the grub install... cjwatson there any useful info I could gather that would help fix this case? It seems to happen on [15:15] netbooks with a 20 (really 4 + 16) GiB SSDs like the eee 900 [15:16] I doubt that the size is relevant [15:17] 22 : No such partition [15:17] This error is returned if a partition is requested in the device [15:17] part of a device- or full file name which isn't on the selected [15:17] disk. [15:17] rmcbride: /etc/fstab /boot/grub/menu.lst and 'sudo vol_id -u' of the devices backing your / and /boot (if distinct) would be a start [15:17] rmcbride: however to some extent this is a known general issue [15:19] sc [15:20] bdmurray: you about yet dude? [15:20] schwuk, cr3, just pushed up an update to suspend_test for checkbox, proposed for merging: https://code.launchpad.net/~apw/checkbox/suspend-resume/+merge/3902 [15:20] cjwatson: OK I'll get those. WHat seems to be happening is that the device IDs are different when the USB drive is no longer present. I don't think it has so much to do with size as the fact that there are two SSDs on the system, [15:20] apw: cheers! [15:21] or at least the way grub sees things [15:21] rmcbride: right, at the moment there is unfortunately no good fix [15:21] apw: for which ubuntu releases might this script work? [15:22] apw: Excellent. [15:22] rmcbride: grub has to locate its stage2 from its 446-byte stage1, and there isn't enough room in there to do proper dynamic detection, so it depends on what the Linux device <-> BIOS device mapping looked like during installtion [15:22] rmcbride: I think we may, reluctantly, have to have some manual UI hack in ubiquity [15:22] cjwatson: ah. crap. Yea it's sounding like it. [15:23] rmcbride: futzing with /boot/grub/device.map and rerunning grub-install is probably a viable workaround (and is what the implementation will have to look like) [15:24] cjwatson: yea that's what I had to do with my intrepid/unr alt install last time I installed on this hardware. [15:24] apw: would you mind if I report a bug so that the script doesn't depend on being run with sudo, and that it could instead be run as root? [15:24] apw: I've added you to the checkbox-dev group, so you can commit your changes directly. [15:24] cr3, not at all ... that would preclude use of the dbus test [15:24] but in V6 that test is a separate 'category' so that would work [15:24] cr3, got a use case? [15:25] (purley curious) [15:26] apw: for interactive use within checkbox. if you have a command line script requiring to be run with sudo and which prompts for a password in the background, that breaks the flow of the application [15:27] cr3, i am going to have to hack the script about to interface it to checkbox anyhow [15:27] apw: patch approved and merged with trunk [15:27] and am planning on doing that bit next [15:27] cr3, ta [15:27] cr3, so explain to me, when running tests whome are we running as by default? [15:27] apw: if you interface it to checkbox, you could add "user: root" when defining your test [15:28] apw: by default, as the user running checkbox. however, this can be overridden with the "user:" option [15:28] and how does it become root there [15:28] apw: through dbus [15:28] ie can i tell who the orignal user is? [15:28] apw: good use case! no, you can't tell who the original user is [15:29] hmmm ... one of the tests at least needs to know that [15:29] do i have it as a variable in theere anywhere, as there is no trust issue [15:29] apw: I'll have to look at the script more closely but it would be nice to run it on server as well, as long as it's automated, which doesn't have dbus installed by default though [15:30] yep, that test can and is disabled for server [15:30] apw: honnestly, I don't have an answer for you at this time, I'll have to think about it [15:30] and we can simply ommit it for that one [15:30] but even if we only had $user as an option in the command strings to run that would [15:30] apw: I can appreciate your use case though, checkbox should be able to support knowing the user [15:30] be sufficient for this use case, right now i just use SUDO_USER cause i can [15:31] can i sub in environement variables? [15:31] apw: the problem is that when commands are run as root, a request is made to dbus telling it to run a test by name, not by command [15:31] shame [15:31] well ... i'll let you think about that one [15:31] thats one of a larger number of tests [15:32] apw: well, that's how it can remain secure, I wouldn't want to be able to send arbitrary commands to run as root by dbus :) [15:32] and i can simply make that a manual test [15:32] ie say "press suspend dummy" [15:32] apw: when do you intend to work on the integration, I'd like to have a workaround ready for you before then [15:32] cr3, soon ish [15:33] but it won't hold me up. it might make more sense to do that test [15:33] manually now i think about it [15:33] as the test is inherantly trying to test the end-to-end experience [15:33] I know I can pull some rabbit out of my... hat. I'd just like to make sure it materializes as a rabbit rather than a gerbil [15:33] asking the user to hit suspend isn't such a bad idea [15:33] in case I'm not making sense, I'd like to find an acceptable solution rather than a hack [15:33] where acceptable == rabbit, hack == gerbil :) [15:33] its actually a better test. so for now i'd not worry about it, at least in the sense that it isn't a blocker for the work i want to do on the thing [15:34] apw: ok, so instead of reporting a bug against the suspend_test script, I'll report one against checkbox itself :) [15:34] good enough :) [15:35] cr3, remind me, was there a reboot capable test yet? [15:37] apw: nope [15:38] cr3: ping [15:38] on your plan list add coping with machine crashing as part of the test [15:38] ie. continuing after a reboot and sayign "that didn't work did it" [15:39] apw: reported bug #333884 and subscribed you [15:39] Launchpad bug 333884 in checkbox "Some tests should be run as root but know the original user" [Undecided,New] https://launchpad.net/bugs/333884 [15:40] ara: pong [15:40] cr3 sounds good [15:40] cr3: is the checkbox ppa version newer than the one in jaunty? [15:40] apw: hm, other people have expressed interest in that usability issue where checkbox should continue from where it stopped in case of crash [15:41] ara: yep [15:41] given suspend-resume is pants and blows up a lot [15:41] and indeed there is only right now any point in 'did that work == YES' else they arn't there to hit it [15:41] cr3: then versioning is wrong [15:41] cr3: apt-get thinks jaunty one is newer [15:41] cr3: https://help.launchpad.net/Packaging/PPA?action=show&redirect=PPAQuickStart#Versioning [15:42] cr3: to install the ppa i would need to uninstall checkbox and then install the ppa manually [15:43] schwuk: ^^^ [15:45] cr3, when i run checkbox-gtk it just whines that i need to be an adminstrator [15:45] thats hopeless for a pleb to understand [15:46] apw: yeah, the version with policykit support has been in the core-dev queue since FF [15:46] davmor2: I am! [15:47] cr3, and it just crashed my X-server into the bargain. top marks [15:47] apw: woohoo, it failed testing! [15:47] apw: seriously though, what test were you running which crashed X? [15:48] none, i was skipping them all, i ran exactly none [15:48] i _think_ it was about to ask me about my wavelan prism2.5, but its gotten much further this time [15:48] apw: can you reproduce the problem consistently? [15:48] so i can't be sure [15:48] not so far [15:48] bdmurray: Yay. Right I've had a quick look at jaunty and umenu/wubi is still borked should be uploading shortly. So I'm going to rub wubi from intrepid on a fat32 xp install is that alright for you? [15:48] /s/rub/run [15:48] davmor2: sure, that'd be great. the bug is about kernel upgrading after the install [15:49] cr3, can i tell it which suite to run ? [15:49] apw: the only thing that comes to mind is that one of the tests needs to run xrandr in order to display the question [15:49] apw: which suite: yes. which test specifically: no. [15:50] yeah i want to say just my resume suite. whats the option? [15:50] this manual page sucks [15:50] apw: for i in `seq 1 20`; do xrandr -q; done # does that reproduce the problem? [15:51] apw: --config=checkbox/plugins/suites_info/whitelist=name_of_suite_file.txt [15:52] nope to the xrandr reproduce, and the test was that resolution test [15:52] apw: yeah, the resolution test simply runs xrandr to determine the resolution [15:53] well unless it happens again i'll put it down to an X update happening recentlly [15:53] ara: The version in the PPA (0.5~ppa5) is the same as the one proposed for Jaunty (0.6). [15:54] schwuk: yes, but in jaunty now it is version 0.5 [15:54] schwuk: so if you try to install the ppa over 0.5, it should be possible, but it is not [15:54] cr3, at the recent sprint we were introduced to checkbox and had some feedback on it [15:54] ara: hmm [15:54] like on how you get the question before you have run the test [15:54] schwuk: https://help.launchpad.net/Packaging/PPA?action=show&redirect=PPAQuickStart#Versioning [15:55] is there now a way i should be encoding things to get that to happen right? [15:55] or is that still in the works? [15:55] (t [15:55] schwuk: dpkg --compare-versions 0.5~ppa5 gt 0.5 && echo "true" || echo "false" [15:55] (this is for tests with a TEST button) [15:56] for sure, a ~NNN is before the thing its attached to [15:56] apw: man, did I make that introduction to checkbox? I remember making a presentation but only emmet was there :) [15:56] heh no it was someone else, but i assume yo heard about all my pitiful whining [15:57] cr3: No, I did it, and yes I passed along some of the pitiful whining :) [15:57] apw: nope, I wish I did though. I love to hear whining, I actually do wdd (whining driven development) [15:57] ara: We'll get that sorted out asap [15:57] schwuk: cool, my memory must be failing me once again [15:58] so one thing was the fact that you get the question and answer bits before you have hit test which makes no sense [15:58] schwuk: cool, thanks [15:58] apw: by default, you get the question before running the test. isn't that what you're observing? [15:58] and i wondered if that was jst cause i wasn't doing the definition of the test right [15:58] yeah it says like "Hit test to suspend" Did you resume sucessully? [15:58] apw: sure that makes sense. what if the test powers off your machine, that's quite a startling user experience. [15:59] yep you need the first half the text, so in this case [15:59] "Click on the test button to initiate a suspend. Once the machine has gone to sleep, wake it up again" [16:00] i'd like that to be there, then i hit test and i'd like the queston to appear either additionally or only in the box after the script completes [16:00] kinda two phase [16:00] cr3, anyhow, the short answer is, i am doing it right at the moment i think [16:00] apw: how about this: Hit the test button to suspend your machine which should power on automatically. Did your machine power on after a few seconds? [16:01] well thats kinda what it says already [16:02] apw: I'm not sure I follow your flow here: i'd like the queston to appear either additionally or only in the box after the script completes [16:03] the "did the machine resume?" and the yes/no/skip options are not valid until the Test button was pressed, well skip and only skip is [16:04] there are two things in the text presented. the instructions "Do this/Hit TEST" and the question "Did you hear a beep?" [16:04] and they are kinda separate [16:04] one way to represent that might be to add the question and enable Y/N only after TEST has been hit say [16:04] apw: I'm reluctant to disable yes/no if the test button has not been pressed [16:05] well how can they answer if test has not been hit? [16:05] apw: because if I know the suspend works, I don't want to have to wait to test it again [16:05] they blatently can't knw the answer [16:05] hmmm ... [16:05] apw: I might be suspending every day in the first place [16:05] now you are letting the user guess what my test does [16:05] i may be suspending using some other mechanism, how would they know that [16:05] an indeed i do do that, i suspend using dbus and also using pm-suspend [16:06] "on suspend works, so this 100 suspend cycle test is clearly yes" [16:06] is not a very safe option to put in a users hands IMO [16:07] cr3 i am not at alll picky, honest [16:08] apw: this is an accuracy concern though, but duly noted [16:08] right, testing is 100% useless if you don't know its veracity [16:08] it is worse than no information at all [16:09] do we at least record that they didn't hit Test, that is the politic middle ground [16:13] cr3 what order do my tests in my suite run in? [16:15] bdmurray: Right wubi installed I'll start the updates now [16:19] davmor2: okay, if all goes well you shouldn't land at the lp bug filing page [16:19] :) [16:22] stgraber: ping [16:24] bdmurray: Meh Report A Problem Sorry, the package "linux-image-2.6.27-7-generic 2.6.27-7.16" failed to install or upgrade I'm guessing this is what was already reported or do you need me to upload the apport report? [16:26] cr3, got any suite fragments for non-manual tests? [16:26] which plugin should i be using [16:26] davmor2: try and upload it and see what happens [16:29] apw: plugin: shell [16:29] bdmurray: Hmm there are a few bugs listed for it [16:30] cr3, so there is no feedback at all on shell test [16:30] davmor2: so it didn't say this problem has already been reported and show you bug 252900? [16:30] Launchpad bug 252900 in wubi "vfat : wubi 8.04.1-rev506 kernel upgrade fails" [Medium,Confirmed] https://launchpad.net/bugs/252900 [16:30] is there a way to initiate a test which shows someone something, but requires no interaction [16:30] i would like to tell them how far through they are and the like [16:30] bdmurray: top of the list [16:31] apw: by default, the exit code of the given command determines whether the test passed (0) or failed (!0) [16:31] bdmurray: next is 332476 and 332837 which look identical [16:32] yeah but say there are two tests, a variable timing test and a repeated iteration test [16:32] apw: the shell tests are all run after the manual tests and a progress bar is displayed during that time [16:32] davmor2: okay, apport should have caught it before letting you get that far [16:32] i would like the box on the screen to show 'variable test pass 10' or something [16:33] apw: is the purpose to be able to identify which pass number failed, if that happens? [16:33] bdmurray: I got the login page, next I got the summary description, then I got the bug list [16:34] the purpose is that the test is going to take 30 minutes and 30 mins of cylon isn't going to help them [16:34] so i'd like it to have the normal box without the pass/fail on the bottom [16:34] with something helpful in it [16:34] even if i have to do each pass as a separate test, if i had a non-interactive normal test like window [16:35] apw: my god man! 30 minutes just for suspend/resume? [16:35] then it could 'flick' from test to test [16:35] cr3, heh, we do 24hour suspend resume tests on some h/w [16:36] apw: actually, a 30 minute test would be fine if it's interactive and the user can optionally skip it [16:36] bdmurray: http://www.davmor2.co.uk/lp.png is the page I'm on if you want I can continue [16:36] cr3, well right [16:36] davmor2: is this something you could test later after I sort out my apport bit [16:36] bdmurray: Yes no probs [16:36] so i guess i have to do it as an interactive test, even thought its not [16:37] and even then i don't have any feedback [16:37] thats pretty poor [16:37] apw: well, what does the test do exactly? repeatedly suspend/resume the machine? [16:38] bdmurray: I'll just wipe the wubi install and drop a base install back on so it's ready [16:38] davmor2: if you were to perform the upgrade again right now you'd get the same error [16:38] at least I think so [16:39] cr3, yep in one it does it a lot of times, the other it does it with varying times awake [16:39] So leaving as it is should be fine [16:39] as that tickles bugs often [16:39] apw: if your concern is the poor user experience of showing a progress bar for 30 minutes, that could be solved [16:39] so i guess i would be happy making the whole thing a single interactive test which said [16:39] "this takes a damn long time" [16:39] but some feedback would be nice [16:39] apw: it could be non-interactive, because it's not really interactive as you said [16:39] bdmurray: I want to make a back up of the windows install anyway so no issue there [16:40] cr3, though as you say that is one people might want to skip [16:40] apw: I'm reluctant to have each test provide their own custom mechanism for providing feedback. so, how about this... [16:40] so at least offering a skip is pretty important [16:40] and I want the broken wubi install off before I do that :) [16:40] cr3, right i don't want to do feedback in my code thanks :) [16:41] apw: what would you think if tests define a timeout and a reasonably high default. then, instead of showing a progress bar, I could show the maximum amount of time remaining which would slowly decrease as automated tests are completed. [16:41] cr3, thats just bound to get it wrong and annoy people [16:41] how about a standard reporting form [16:41] apw: it's an upper bound though, which would never go over [16:41] command_with_output: [16:42] apw: it's annoying to people when the upper bound is not respected :) [16:42] and that would display the stdout of the command in like an xterm style box [16:42] then we could display our current position in there [16:42] apw: ooo! that reminds me of the xterm box in gdebbi or somesuch [16:43] apw: very cool idea, I like it [16:43] well i'd hope you would put the panel in the same spot the test queston goes [16:43] so that would just scroll up, but dunno how hard that is [16:43] i presume i can fake it up for now with just an xterm [16:43] apw: the same spot would show the output of each of the automated tests being run, right? [16:44] yeah [16:44] while you think on that, i am going to cheat and use an xterm :) [16:44] apw: thanks, I'll report another bug for that because I really don't want to forget this one [16:44] enjoy :) [16:45] you are going to come to hate working with me :) [16:47] apw: as long as you have reasonable requests, I might react quite the opposite and enjoy working with you very much [16:49] cr3, heres hoping [16:49] cr3, so is there a way to not run a test cause one previous didn't work or was skipped [16:50] like it makes sence to do a single automated sleep/resume cycle. if that suceeds then its worth doing lots of repeats of it, else its madness === fader_ is now known as fader [17:07] apw: absolutely, depends: test_name [17:08] apw: are there specific requirements for running the tests by any chance, such as some kernel flag or some swap space or somesuch? [17:09] bdmurray: Right off for tea but I'll be back latter and I'll give it a retry for you then [17:10] apw: in case you need more swap space than memory, you could also add, requires: swaptotal > memtotal === fader is now known as fader|lunch [18:45] bdmurray: Right I'm back [18:45] davmor2: thanks, it'll be a while before my bit is squared away [19:15] sbeattie: ping [19:20] davmor2: what's up? [19:20] are you the controler for dl-ubuntu-test-iso? [19:21] I prefer to think of it as Chief Bug Producer, but yes. [19:23] jeos and gobuntu aren't in anymore and is there away to get the mobile .imgs tied in? [19:26] davmor2: jeos and gobuntu (and kubuntu-kde4) should be disabled by tip now (unless, ahem, I've produced another bug), though if you still have them in your config, it'll still try to pull them, [19:27] I'll peek at mobile imgs [19:27] sbeattie: Sorry I was looking at base and not main === fader|lunch is now known as fader [22:07] Right I'm off till tomorrow :)