[01:28] Is ubiquity defaulting to America/New_York for an install in Japanese a bug? [01:32] yes [01:32] I think that only happens if you picked some other language first and then went back and redid it, though? [01:32] I'll file it then. Thanks. [01:32] do give precise information on what you pressed from the bootloader on [01:32] I didn't do that intentionally. Just booted the -moble image. double-clicked the icon. Selected Japanese. [01:33] OK. Do you want any logs? It's not in debug mode. [01:33] I'm sure it's reproducible from that [01:33] OK. [01:34] The city name "Tokyo" was also in English, but I suspect that's just a missing string. [01:34] not entirely sure those are translated at all in the location map [01:36] Right. Not likely to change this time, but post-release I'll do a summary of all the remaining English bits in a Japanese install, and file the appropriate bugs. The default timezone I'll file now. [01:37] thanks [07:54] evand: morning/evening hello. I've had a thought about the dvd install which completed in the end... Would it help if I ran strace the other side of the stoppage so from the last user page till it stops? [07:55] davmor2: we've since figured it out. [07:55] thanks though [07:55] oh cool :) [07:55] it was the file blacklist generation (http://wiki.ubuntu.com/DVDPerformanceHacks) [07:56] or rather, is. It's not fixed yet, and cjwatson does not think optimization is OK this late in the cycle, so we might back it out for 8.10. [07:58] might want to add it as a release bug though so people don't think it has stopped due to the inactivity [08:00] indeed [08:01] evand: is there a bug or would you like me to write one? [08:03] davmor2: I'll take care of making one, thanks though [08:03] evand: Okay np's [08:10] davmor2: any idea where the "Known Issues" are collecting? [08:11] I can't find the right wiki page [08:11] for the RC, that is [08:13] I think it is https://wiki.ubuntu.com/IntrepidReleaseNotes (for errata) [08:14] I think that's only for the final [08:14] slangasek: is there a bug number? if so, it should have a task opened on the 'ubuntu-release-notes' project [08:14] But I see that slangasek is still up, so I'll prod him [08:14] slangasek: the actual release notes drafting is at https://wiki.ubuntu.com/IntrepidReleaseNotes, but I prefer that we have a bug open first [08:48] evand: does the blaklist build in ram before being written to drive then? [08:51] yes [08:52] so that explains the lack of disk activity then thanks [09:06] no problem [09:37] evand: well, to clarify, I'm concerned about optimisations that change the overall logic - but if there are micro-optimisations that help significantly, those might be worth it [09:38] Hi. Any idea why system-config-kickstart doesn't show anything in the package selection panel in Kubuntu Hardy? [09:39] cilkay: bug 21214 [09:40] Aha. Thanks. [09:40] It looks like it doesn't support PXE installations either so it's of limited value at the moment for me. [09:40] *blink* certainly ought to support PXE [09:40] what do you mean? [09:41] Really? I don't see it as an option in the "Installation Method" panel. [09:41] My preseed installation was getting increasingly complicated so I was kicking the tires on kickstart again. [09:41] But some of that complexity has nothing to do with preseed. [09:41] "Installation Method" means how packages are retrieved, not how you boot the installer [09:41] most PXE users would want to select HTTP there [09:42] and read https://help.ubuntu.com/8.04/installation-guide/i386/automatic-install.html for how to deal with booting [09:42] I was reading about cfengine. Seems like it would help to automate many of the things I'm trying to do but it's seems quite complicated. [09:43] I have PXE preseed installations working, mostly. [09:43] ... thanks to you in no small measure. [09:43] booting PXE Kickstart is pretty similar really - just ks= (and maybe ksdevice=) rather than url= [09:44] Sure. I just didn't see the option in the system-config-kickstart tool. [09:45] I' [09:45] I'm trying to create something like "roaming profiles". [09:45] KDE has kiosk but it seems to be unmaintained and crashes every minute or so on Hardy. [09:47] Preseed PXE install with Kerberos, OpenLDAP, and KDE kiosk with configuration management by cfengine is what I'm shooting for so it's a tall order. [09:48] Off for now. Thanks again cjwatson. [09:48] no problem [10:25] davmor2 I cannot reproduce the reboot error, tried both vista+real hw and xp in vm and both reboot properly === ubuntu860 is now known as xivulon [10:26] superm1: I cannot reproduce your issue either [10:26] okay testing is nearly complete. So I'll just finish that off and then I'll look at it again. [10:31] evand: that said, after I expressed concern about optimisation, it occurs to me that http://paste.ubuntu.com/61441/ could work === tjaalton_ is now known as tjaalton [10:39] uh jono did an album? that's great stuff! [11:02] oem-config: cjwatson * r551 oem-config/ (configure configure.ac): bump to 1.53 [11:07] oem-config: cjwatson * r552 oem-config/ (debian/changelog lib/frontend/kde_ui.py): [11:07] oem-config: * KDE frontend: [11:07] oem-config: - Hardcode command line (including setting the Oxygen style) rather than [11:07] oem-config: passing sys.argv to KCmdLineArgs, which will include options that KDE [11:07] oem-config: doesn't understand (LP: #218144). [12:36] xivulon, are you doing it with a mythbuntu or ubuntu? I'm wondering if it's only reaching that codepath on mythbuntu iso's with wubi? [12:46] I did it with today's ubuntu iso [12:46] could you try that too? === davmor2 is now known as davmor2_lunch [13:05] xivulon, yeah I'll try to grab it [13:47] persia: don't suppose we could get logs for bug 287857? it's pretty bizarre [13:47] Launchpad bug 287857 in ubiquity "Prompting for step 4 in --automatic mode only when not in --debug mode" [Undecided,New] https://launchpad.net/bugs/287857 [13:47] --debug really oughtn't to change a whole lot ... [13:48] cjwatson. Sure. I'll collect some. That was a placeholder for me to investigate as much as anything else, although I definitely appreciate the help. In about 20 minutes work for you? [13:49] oh, sure, I just wasn't sure whether you were stuck [13:49] and saw it mentioned in the meeting [13:49] not like I have lots of time ... [13:49] I'm mostly stuck because I've been up for 33 hours, rather than because I'm not understanding the code path. [13:49] I'd rather look at it tomorrow if that isn't worse for you. [13:50] we're running out of time is all [13:50] OK. Tonight then. [13:50] but 33 hours is not good for thinking [13:51] xivulon, just tested with an ubuntu disk too [13:51] it happens there as well with the same test procedure [13:52] er ISO at least [13:52] oem-config: cjwatson * r553 oem-config/debian/ (changelog po/gl.po po/hr.po po/ja.po po/mk.po po/pt_BR.po): Update translations from Launchpad. [13:56] oem-config: cjwatson * r554 oem-config/debian/changelog: releasing version 1.53 === davmor2_lunch is now known as davmor2 [14:03] cjwatson, The bug description is wrong. I think I can collect useful logs, but it will take a few more runs to ensure I have the right model. [14:04] ah [14:25] cjwatson: Just tested your patch. FWIW, it takes 9 minutes to run through generate_blacklist with it applied on a VM. [14:27] how long does it take without? [14:27] checking that now [14:48] 11 minutes [14:51] I'd love to put a progress message for the blacklist generation in there, but obviously that's not possible given where we are in the release. [15:00] a progress message seems entirely possible [15:00] it can't be translated, but it's better than none at all [15:01] ah, I thought string changes were expressly forbidden. [15:01] well, certainly you shouldn't be changing existing strings [15:01] and new strings should be kept to a minimum - but for this case I think it's fine [15:01] ok [15:02] so I assume then that you think we should keep the blacklist in place for the RC and the release? [15:02] phone [15:02] ok [15:09] cjwatson, 287857 isn't the installer at all. Sorry for the confusion. [15:11] oh, what was it? [15:11] evand, kourou not honoring arguments in Exec in .desktop files. [15:12] ah [15:12] So ubiquity was never being passed --automatic. The reason it appeared to work with --debug was that I ran that from the command line. [15:12] When I went back to retest, rather than waiting until tomorrow, I discovered that adding --debug in the .desktop file didn't do anything either, but it did for Desktop. At that point, it was just a matter of testing kourou with other applications to verify the cause. [15:13] ah [15:20] evand, maybe would adding a preseed option to turn it on and off be worthwhile? particularly in cases that all the language packs remain post install? [15:20] evand, then you could make a flip of the switch decision too on final whether it would stay [15:22] superm1 I'd need some windbg output [15:22] superm1: I agree [15:22] xivulon, how do i obtain windbg output? [15:23] you need to install windbg or equivalent tool [15:23] xivulon, is that the stuff that i get the option to submit to MS? [15:23] I do not find that memory dump very useful in most cases [15:25] but yes, post it anyway [15:26] well i couldn't copy and paste it when i tried [15:26] in fact I can open that using my windbg anyway [15:26] I do not know windows that well, but I guess they save it in some temp file [15:27] they love temp files... [15:27] is there a particular windbg then that you are preferring? or just follow http://www.microsoft.com/whdc/DevTools/Debugging/default.mspx ? [15:28] install debugging tools for windows [15:31] okay i'll install them tonight when I get home and see what I can get you [15:31] are there symbols for wubi that need to be installed too? [15:39] not really [15:51] ubiquity: evand * r2918 ubiquity/ (configure configure.ac): Bump to 1.10.8 [15:53] evand: can you have a look at bug 287747? [15:53] Launchpad bug 287747 in ubiquity "Installing into an existing / partition will result in a broken kernel" [Undecided,New] https://launchpad.net/bugs/287747 [15:53] cjwatson: will do [16:53] ubiquity: evand * r2919 ubiquity/ (3 files in 2 dirs): [16:53] ubiquity: * Optimize the blacklist generation. Thanks Colin Watson. [16:53] ubiquity: * Add ubiquity/install/generate-blacklist preseed key for file [16:53] ubiquity: blacklist generation. [16:53] ^ Just in case [16:54] did you make sure it was still producing correct results? [16:54] my DVD jigdo template rsync is still only at 51% :-( [16:56] it appears to be: http://evalicious.com/tmp/blacklist [16:56] yikes [16:58] * evand runs fibre from the DC to cjwatson's house. [17:05] evand: was that an English install? [17:09] cjwatson: indeed [17:09] I'll give it another go in other languages [17:09] just to be on the safe side [17:11] s/other/another/ [17:13] ok, thanks [17:13] I don't suppose the problem is just building that enormous list in memory? [17:14] perhaps [20:36] cjwatson: re 287747> initial testing looks ok, I am unable to reproduce the bug thusfar [20:37] I'll keep at it though [20:48] for some reason, when netinstalling the alternate image of ubuntu i386, I now see the main menu instead of having everything preseeded. not sure what might be wrong: http://people.ubuntu.com/~cr3/syslog [20:52] aha! I already see one problem because the wget in the installer environment doesn't seem to support http proxying === michael_e_brown_ is now known as mebrown