| superm1 | at some point did things switch around so that ubiquity doesn't run the whole process as root? | 03:25 |
|---|---|---|
| evand | yes | 03:26 |
| evand | sorry, I should've told you that | 03:27 |
| superm1 | whew that's not good, a lot of my scripts broke | 03:27 |
| evand | my apologies | 03:27 |
| superm1 | hopefully not a big deal.. | 03:27 |
| evand | well going forward you can control what gets run as root and what doesn't | 03:27 |
| evand | obviously some things need to run as root | 03:28 |
| superm1 | were these very recent commits (i'll look over the history to see what you adjusted for items) | 03:28 |
| evand | I'll get you the exact revno | 03:28 |
| evand | bzr diff -r2464..2465 | vim - | 03:29 |
| evand | should give you most of it | 03:29 |
| evand | there was one or two changes after that | 03:29 |
| superm1 | okay great thanks, i'll let you know if i have some difficulties switching things around | 03:30 |
| evand | yeah, please do | 03:30 |
| evand | best of luck | 03:30 |
| superm1 | thx | 03:30 |
| CIA-44 | ubiquity: superm1 * r2481 ubiquity/ (debian/changelog ubiquity/frontend/mythbuntu_ui.py): adjust permissions before/after launching external applications from ubiquity | 04:02 |
| superm1 | luckily very easy fix :) | 04:03 |
| evand | yay | 04:03 |
| evand | cjwatson_: if you have a free moment today, can you take a look at the current state of ubiquity-dm? I'm not entirely confident in the approach and I'd like the reassurance of a trained eye. | 05:50 |
| evand | and yay, CIA is down | 05:50 |
| * evand kicks CIA-44 | 05:50 | |
| CIA-44 | ow | 05:50 |
| superm1 | evand, you weren't really consistent with spacing in ubiquity-dm changes. i'm making a few other changes to that file right now, i'll clean those up for you | 06:02 |
| evand | oh? Thanks | 06:04 |
| CIA-44 | ubiquity: superm1 * r2483 ubiquity/ (4 files in 2 dirs): add mythbuntu to list of possible frontends to look for upon ubiquity startup | 06:11 |
| CIA-44 | ubiquity: superm1 * r2484 ubiquity/ (bin/ubiquity-wrapper debian/changelog): check for jockey-gtk instead of restricted-manager | 06:11 |
| xivulon | If anyone would like to help review the english text in wubi/umenu please see the following 2 files: | 10:39 |
| xivulon | http://bazaar.launchpad.net/~ubuntu-installer/wubi/hardy/annotate/agostino.russo%40gmail.com-20080225023058-shsf3oawibio0nxt?file_id=english.nsh-20070529205356-9mvslu0v3aodhtoy-1 | 10:39 |
| xivulon | http://bazaar.launchpad.net/~ubuntu-installer/umenu/devel/annotate/evand%40ubuntu.com-20080220141346-5prmrb6va03jdtw1?file_id=english.nsh-20080212233008-kvy2sm0usmkk5nf3-18 | 10:40 |
| xivulon | shouldn't take more than 5m | 10:40 |
| CIA-24 | installation-guide: cjwatson * r409 ubuntu/ (656 files in 174 dirs): merge from Debian 20080211 | 12:15 |
| CIA-24 | installation-guide: cjwatson * r410 ubuntu/en/appendix/chroot-install.xml: adjust chroot-install for Ubuntu sources.list | 12:18 |
| CIA-24 | installation-guide: cjwatson * r411 ubuntu/ (10 files in 5 dirs): | 12:40 |
| CIA-24 | installation-guide: * Bump kernelversion to 2.6.24. | 12:40 |
| CIA-24 | installation-guide: * Bump GNOME version to 2.22. | 12:40 |
| CIA-24 | installation-guide: * Bump release version and names for Hardy. | 12:40 |
| CIA-24 | installation-guide: cjwatson * r412 ubuntu/debian/changelog: releasing version 20080211ubuntu1 | 12:57 |
| xivulon | cjwatson, evand, is there any interest in having lupin-casper ported to the alternate ISO? | 13:11 |
| xivulon | That would be mostly to support loopinstallations for distros that do not come with a live CD. | 13:11 |
| xivulon | If that is something you would like to see, I'd be glad to work on an implementation. | 13:12 |
| cjwatson | xivulon: well, I did partman-auto-loop as a udeb for a reason; but there's no rush for hardy, I think there are probably more important things to work on for that | 13:32 |
| cjwatson | speaking of which | 13:33 |
| cjwatson | evand: I diagnosed bug 188492 - comment in the bug | 13:33 |
| ubotu | Launchpad bug 188492 in ubiquity "console-setup/layoutcode settings are ignored" [Medium,Confirmed] https://launchpad.net/bugs/188492 | 13:33 |
| xivulon | cjwatson, I was thinking of doing it using a udeb anyway, when you say that there is no rush for hardy, do you mean we should target hardy+1? | 13:48 |
| xivulon | for me too is a low priority, I have a few crashes to debug first... | 13:49 |
| xivulon | thanks for 188492 by the way | 13:49 |
| cjwatson | xivulon: that's what I meant, yes | 13:51 |
| cjwatson | I want the hardy installer to be really stable and don't want to rush many more features into it | 13:52 |
| xivulon | np | 13:52 |
| CIA-24 | ubiquity: jriddell * r2486 trunk/debian/changelog: set me in changelog | 13:59 |
| evand | cjwatson: oh wow, whoops. Thanks for catching that. | 14:02 |
| evand | cjwatson: did you happen to see my comments earlier in the day about ubiquity-dm? | 14:03 |
| cjwatson | evand: yeah, I wasn't sure exactly what was bothering you though? | 14:04 |
| evand | running X as root still. I'm not sure if there are any side-effects of that. | 14:05 |
| cjwatson | you pretty much have to, surely | 14:05 |
| cjwatson | I think that's fine | 14:05 |
| evand | oh? | 14:05 |
| evand | ok, works for me | 14:05 |
| cjwatson | direct hardware access ... | 14:05 |
| cjwatson | the one thing that jumps out at me is chmodding the dm log file to 666 | 14:06 |
| cjwatson | what's with that? | 14:06 |
| evand | ah, right | 14:06 |
| cjwatson | oh, and strictly we should fish HOME out of getpwnam too, although IIRC ubiquity has the same buglet in other places too | 14:06 |
| evand | and chmodding 666 so that "ubuntu" can write to it. | 14:07 |
| cjwatson | as a style point I would prefer 'import pwd' and 'pwd.getpwnam' just to make it clear where that function is coming from | 14:07 |
| cjwatson | chmod> shouldn't be necessary if you pass the fd around; it opens it with elevated privileges, doesn't it? | 14:08 |
| cjwatson | you can be handed an fd by root that you wouldn't be able to create yourself | 14:08 |
| cjwatson | the file mode just governs new opens | 14:08 |
| evand | I thought that was the case, but my testing seemed to prove otherwise | 14:08 |
| evand | I'll take another look | 14:08 |
| cjwatson | something else going wrong then, I guess | 14:09 |
| cjwatson | not a big deal under the circumstances, just jumped out | 14:09 |
| xivulon | evand re english^ | 14:11 |
| cjwatson | whoa, how did the code behind bug 194688 ever work? | 14:12 |
| ubotu | Bug 194688 on http://launchpad.net/bugs/194688 is private | 14:12 |
| evand | uhh odd. I know I've seen it work for me. | 14:13 |
| cjwatson | it's only the popup menu that's broken | 14:13 |
| cjwatson | it passes two extra user parameters that on_partition_list_edit_activate doesn't expect | 14:13 |
| cjwatson | ah | 14:13 |
| cjwatson | r2452 broke it | 14:13 |
| cjwatson | I think you forgot to update the popup menu code too? | 14:14 |
| evand | hrmm | 14:15 |
| cjwatson | though he says later on in the report that he actually meant to report a different bug he also encountered | 14:15 |
| evand | ah looks like it, but how on earth did I not run into that in UI testing. argh. | 14:15 |
| cjwatson | two different ways to do it ... | 14:17 |
| evand | ah | 14:18 |
| evand | yikes, so far not a great morning | 14:18 |
| cjwatson | plenty of time before the next alpha ;-) | 14:20 |
| evand | heh, that's one way of looking at it | 14:21 |
| evand | xivulon: can you elaborate? | 14:21 |
| evand | oh nevermind | 14:21 |
| xivulon | I mentioned the 2 english files that would need to be reviewed as we discussed earlier, no rush whatsoever | 14:21 |
| evand | indeed, I didn't see that at first | 14:22 |
| evand | I'll take a look after I fix the above mess. | 14:26 |
| cjwatson | evand: is bug 77966 handled by your fix for bug 48355? IIRC your solution was fairly general in the end and covered both | 14:29 |
| ubotu | Launchpad bug 77966 in ubiquity "MASTER: Should warn about insufficient disk space (manual partitioning)" [Medium,Confirmed] https://launchpad.net/bugs/77966 | 14:29 |
| ubotu | Launchpad bug 48355 in ubiquity "MASTER: partitioner should warn about too small disk while autopartitioning (some_device)" [Medium,Fix released] https://launchpad.net/bugs/48355 | 14:29 |
| evand | that would cover resizing as well, right? | 14:30 |
| cjwatson | eventually, though the resize might well fail first | 14:33 |
| evand | I figured, though never looked, that commit.d is run in all cases, but I'm pretty sure I fed partman a partition size that was too small using the resize widget and it let me go through. | 14:35 |
| cjwatson | it isn't run at resizing time; it would be run later when you actually try to finish off the partitioner | 14:38 |
| cjwatson | CIA is slow today | 14:38 |
| evand | better than last night when it didn't work at all | 14:38 |
| evand | right, but I'm saying I got all the way through to an ENOSPC on that one, I think. | 14:38 |
| evand | I should probably test again to be sure. | 14:39 |
| CIA-24 | ubiquity: evand * r2489 ubiquity/ (debian/changelog ubiquity/frontend/gtk_ui.py): | 15:13 |
| CIA-24 | ubiquity: * Update partman_popup with respect to other recent changes (LP: | 15:13 |
| CIA-24 | ubiquity: #194688). | 15:13 |
| evand | welcome back, CIA-24 | 15:14 |
| CIA-24 | ubiquity: evand * r2490 ubiquity/debian/changelog: Whoops, make sure launchpad-bugs-fixed picks up the previous revno. | 15:16 |
| CIA-24 | ubiquity: evand * r2492 ubiquity/ (debian/changelog ubiquity/components/console_setup.py): | 15:27 |
| CIA-24 | ubiquity: * Regain root when moving /etc/default/console-setup out of the way | 15:27 |
| CIA-24 | ubiquity: (LP: #188492). | 15:27 |
| cjwatson | does CIA just not like me? | 15:28 |
| cjwatson | ah, it would help if I set cia-project | 15:29 |
| cjwatson | I was switching over to a bzr repository and it got lost along the way | 15:29 |
| evand | :) | 15:29 |
| CIA-24 | ubiquity: evand * r2493 ubiquity/debian/changelog: Credit where credit is due. | 15:29 |
| CIA-24 | ubiquity: cjwatson * r2487 ubiquity/ (debian/changelog partman/check.d/03partition_too_small): | 15:29 |
| CIA-24 | ubiquity: * Slight shell efficiency improvement in | 15:29 |
| CIA-24 | ubiquity: partman/check.d/03partition_too_small. | 15:29 |
| CIA-24 | ubiquity: cjwatson * r2488 ubiquity/ (configure configure.ac): bump to 1.7.12 | 15:29 |
| CIA-24 | ubiquity: cjwatson * r2491 ubiquity/ (37 files in 10 dirs): | 15:29 |
| CIA-24 | ubiquity: * Use 'autoreconf -fi' so that Autotools files reliably get upgraded. | 15:29 |
| CIA-24 | ubiquity: * Upgrade to gettext 0.17. | 15:29 |
| cjwatson | there you go | 15:29 |
| evand | thanks! | 15:30 |
| evand | xivulon: "Reboot into a new world..." sounds too evangelistic. Perhaps somewhat replicate the text from the button, "Try or install Ubuntu"? | 15:45 |
| evand | Of course that suffers from the same problem that a few of the other strings have, that is explicitly naming the distribution. | 15:45 |
| xivulon | ahah loved that sentence | 15:47 |
| xivulon | I can name the disto by the way using a variable, at that stage I'd know the distro name anyway | 15:51 |
| evand | Doesn't that get you in trouble with translations? | 15:51 |
| xivulon | "Reboot and try ?buntu..." | 15:51 |
| xivulon | Should be ok | 15:51 |
| xivulon | If you look into wubi english.nsh there are a few $xxx variables within translation strings | 15:52 |
| xivulon | I will put a note not to translate $stuff | 15:52 |
| evand | I'd also change "booter" to "boot helper" | 16:00 |
| xivulon | evan pull both and feel free to change the text directly | 16:01 |
| evand | Access doesn't seem right, but I face the same problem when you asked me that originally. I cannot think of a word for acessibility that fits in such a small space. How many characters is the maximum for a button again? | 16:01 |
| evand | ok | 16:01 |
| xivulon | I will see if I can make the button larger, not sure about that. | 16:03 |
| evand | assistive would probably be better if you can fit it. Though I wont make the change until we have confirmation of that. | 16:04 |
| xivulon | user32::SetWindowPos might do the trick | 16:06 |
| === cjwatson_ is now known as cjwatson | ||
| CIA-24 | ubiquity: evand * r2494 ubiquity/ (bin/ubiquity-dm debian/changelog): * Minor changes to ubiquity-dm from suggestions by Colin Watson. | 17:04 |
| xivulon | evand got a log from the amd64 user | 17:40 |
| xivulon | http://ubuntuforums.org/showpost.php?p=4402468&postcount=6 | 17:40 |
| xivulon | does not look to me like it is the same issue (freeze around console-setup), but the error is strange: | 17:41 |
| xivulon | 1) no swap (not sure why, does that cause a failure-command by the way?) | 17:42 |
| xivulon | 2) I/O error, dev fd0 (not sure if that is very relevant) | 17:42 |
| xivulon | 3) unsupported locale setting (LANG=en_US)? | 17:46 |
| * xivulon likes failure-commands | 17:47 | |
| xivulon | ...would be even better if we could split out non critical warnings though... | 17:48 |
| xivulon | "interactive" warnings that is | 17:50 |
| CIA-24 | ubiquity: evand * r2495 ubiquity/ (debian/changelog desktop/ubiquity-gtkui.desktop.in): * Remove suffix from icon path in .desktop file (LP: #188221). | 17:50 |
| evand | hooray for omni-completion. Thanks for hacking that up, cjwatson. | 17:50 |
| xivulon | evand I was thinking that ubiquity should have 3 modes: interactive, automatic, semi-automatic. | 17:52 |
| xivulon | the last if for wubi type installation where the user is not supposed to answer any question, but he might be shown a failure message. | 17:53 |
| xivulon | after the failure message we can still trigger a failure-command | 17:53 |
| evand | 1) because it's asking a question. As I explained in a post on ubuntuforums, we had intended to split failure_command into a complete failure and an "automation failure" (when ubiquity requires human interaction for something) | 17:58 |
| evand | You can avoid this question by preseeding it away. | 17:58 |
| evand | we had intended> I had intended to do it at the start of the Hardy cycle, but I completely forgot. It's fairly straightforward and I've added it to my todo list. | 17:59 |
| evand | 2) No, don't worry about that. | 18:00 |
| evand | oh, I missed that comma. | 18:00 |
| evand | So yes, it does cause a failure_command. | 18:00 |
| evand | back to 1) now that I've reread it | 18:02 |
| evand | debconf (developer): <-- GET partman-auto-loop/recipe | 18:02 |
| evand | debconf (developer): --> 0 /ubuntu/disks/root.disk 3000 30001 30001 ext3 method{ format } format{ } use_filesystem{ } filesystem{ ext3 } mountpoint{ / } . | 18:02 |
| evand | so it looks like you're not preseeding swap creation in that case. | 18:02 |
| evand | really not sure why, as I imagine that's always done in your code, not just sometimes. | 18:03 |
| evand | 3) that should be en_US.UTF-8 | 18:05 |
| evand | oh hrm | 18:05 |
| evand | you do set it as that | 18:05 |
| xivulon | hmm really strange, that should be in the preseed file | 18:08 |
| xivulon | Maybe it would help to copy preseed.cfg config.dat (other?) in the failure-command zip file. | 18:08 |
| xivulon | I mean swap recipe | 18:08 |
| evand | preseed.cfg> That would end up pulling in their password. | 18:09 |
| xivulon | 3) Isn't the preseeding supposeed to be "locale en_US"? | 18:09 |
| xivulon | might sed the password out | 18:09 |
| xivulon | would that be in config.dat too? | 18:10 |
| evand | no, it wouldn't | 18:10 |
| evand | but I don't really see the need | 18:10 |
| xivulon | of config.dat? | 18:10 |
| evand | indeed | 18:10 |
| xivulon | I'll add a sedded preseed then | 18:10 |
| evand | well, I don't see the need of either of them | 18:11 |
| evand | you already have debug enabled | 18:11 |
| xivulon | yes | 18:11 |
| xivulon | but in the case above how would I know the exact recipe string? | 18:11 |
| cjwatson | evand: you're welcome | 18:12 |
| evand | You'd look in debug, just as I did | 18:12 |
| cjwatson | preseeding should be en_US.UTF-8 not en_US | 18:12 |
| xivulon | evand does that contain the original "raw" string? | 18:13 |
| evand | Did we stop writing passwords to the debug output at some point? | 18:13 |
| cjwatson | no | 18:13 |
| xivulon | cjwatson is the locale going to be *.UTF-8 for all languages? | 18:14 |
| cjwatson | xivulon: unfortunately not quite. You need to know the actual locale from /usr/share/i18n/SUPPORTED not just make it up | 18:14 |
| evand | curious, I don't see the user's password in this debug log, and I doubt he stripped it out | 18:14 |
| cjwatson | I'd suggest incorporating that file in your build process somehow | 18:14 |
| xivulon | will do | 18:15 |
| evand | and xivulon doesn't appear to be sed'ing the passwords out of the debug log before zipping, as far as I can tell | 18:15 |
| evand | ah duh | 18:15 |
| evand | because it didn't get that far | 18:15 |
| evand | ok yikes, so using debug is probably a bad idea | 18:15 |
| evand | we've talked about enabling this by default before | 18:16 |
| xivulon | evand I don't sed anything at the moment, just zip the whole /var/log | 18:16 |
| evand | but the situation is even worse for wubi | 18:16 |
| evand | because there's no bold text warning them that their password will be in the logs | 18:16 |
| cjwatson | right, debug mode definitely isn't appropriate for any kind of default | 18:16 |
| xivulon | let me know if I need to exclude / sed any file in failure-command | 18:17 |
| evand | I think you should not use debug mode by default and add a command line option to wubi that enables it. | 18:18 |
| evand | that still doesn't address the lack of a warning in the UI that their password is going to be in the logs | 18:18 |
| xivulon | cjwatson I am looking at SUPPORTED, for US I have en_US.UTF-8, en_US, and en_US.ISO-8859-15 | 18:20 |
| cjwatson | use the one with 'UTF-8' in the second column | 18:20 |
| xivulon | evand you mean for failure-command? | 18:20 |
| xivulon | or in general | 18:21 |
| xivulon | re pwd in the logs, isn't it possible to simply sed it out? | 18:21 |
| cjwatson | you could if you're careful. I did ask about including a patch to debconf that makes it never be included, but Joey noted (rightly) that it was fundamentally utterly awful | 18:22 |
| evand | I don't think so. The text could change and you'd end up missing the password. | 18:22 |
| cjwatson | we have had very high-profile installer vulnerabilities before regarding leaked passwords, and it deserves the highest possible level of care | 18:23 |
| cjwatson | you don't get away with this sort of thing a second time | 18:23 |
| xivulon | I remember that news | 18:24 |
| xivulon | do all language have a second column with UTF-8? I also noticed that some lanugages have multiple @ variants all with UTF-8 | 18:26 |
| xivulon | what shall I do with the log? is an (english) warning sufficient? | 18:27 |
| xivulon | I will copy the log only if the user selects debug mode at boot | 18:28 |
| xivulon | re (ubiquity) debug, that is not really enabled by default | 18:34 |
| xivulon | the user has to hit esc and select verbose mode | 18:34 |
| evand | oh? | 18:34 |
| evand | ah, I had no idea | 18:34 |
| xivulon | is that good enough (+pwd warning)? | 18:34 |
| evand | yeah, I think if you can fit it, a password warning would be great | 18:35 |
| evand | "(password included in logs)" perhaps | 18:35 |
| xivulon | hmm that should be localized too... bugger | 18:35 |
| xivulon | the password is only in /var/log/installer/debug, correct? | 18:36 |
| cjwatson | all languages should have a UTF-8-encoded locale; if they don't you can safely ignore them | 18:36 |
| cjwatson | for ones with multiple variants containing @ or . there should be one without those modifiers and you should use that | 18:36 |
| cjwatson | ca_ES.UTF-8@valencia is the only one I can think of offhand | 18:36 |
| xivulon | cjwatson thanks, will do that | 18:36 |
| cjwatson | oh, I see there are a few others | 18:36 |
| xivulon | yes | 18:37 |
| cjwatson | but yeah, looking through them, it's fine to just use the unmodified version; I suspect Windows does not have much in the way of Valencian support anyway and I don't think d-i does either | 18:37 |
| xivulon | by the way I'd need a mapping of windows registry keys to keyboard variants | 18:38 |
| xivulon | I have the code to use that info, and dvorak in as an example | 18:38 |
| xivulon | but not sure where to grab such mappings | 18:38 |
| cjwatson | I think somebody might have to invent them | 18:38 |
| xivulon | There is an hkl value in the registry, for instance 0x00010409=dvorak | 18:40 |
| xivulon | but do not know how to get the others | 18:40 |
| xivulon | ah http://softserv.murdoch.edu.au/pub/mswin/Citrix/Citrix4.21.779/MODULE.SRC | 18:43 |
| xivulon | there is also the issue of preseed.cfg | 18:58 |
| xivulon | that countains the pwd | 18:58 |
| xivulon | now it gets deleted upon successful installation | 18:59 |
| xivulon | but not upon unsuccessful installation | 18:59 |
| xivulon | should probably comment that in failure command too? | 18:59 |
| CIA-24 | ubiquity: jriddell * r2496 trunk/ (debian/changelog ubiquity/frontend/kde_ui.py): Fix QApplication args in ubiquity/frontend/kde_ui.py | 19:02 |
| === andrea-bs is now known as andrea-bs_ | ||
| === andrea-bs_ is now known as andrea-bs | ||
| bdmurray | evand: I'm planning on adding some bugs to the DebuggingUbiquity Known issues section and possibly try to make the page look more like the DebuggingTemplate page. Does that sound alright to you? | 19:30 |
| evand | bdmurray: works for me, thanks for doing it | 19:32 |
| xivulon | evand what's your take on preseed.cfg password? ^ | 19:32 |
| evand | hrm, yeah, I'd do what you can to strip it out in that event | 19:36 |
| xivulon | hmm the issue is that in many case the user will try to reboot and select "Ubuntu" to redo the installation if something goes wrong | 19:39 |
| xivulon | if I strip out stuff from preseed.cfg, then the user will need to start windows and reinstall again to regenerate a new preseed.cfg | 19:39 |
| evand | hrmm, indeed. | 19:39 |
| xivulon | preseed.cfg will be deleted on successfull wubi installation and of curse whenever wubi is uninstalled | 19:40 |
| xivulon | the problem are is if wubi installation fails and the user just keeps it around | 19:40 |
| evand | I guess we don't have any other option but to leave it in on a failed install then. | 19:40 |
| xivulon | I could add a note in failure-command, but if the error happens before (grub4dos / initrd) the user will not see that | 19:41 |
| evand | unless I split failure_command into failure_command and automation_failure_command. Then stripping the password out and showing the user-setup page wouldn't be an issue. | 19:42 |
| evand | It'd be ugly though, as they wouldn't be expecting that. | 19:42 |
| xivulon | you mean asking for username/password in ubiquity as opposed in wubi? | 19:43 |
| evand | right | 19:43 |
| evand | I'm not too keen on the idea though | 19:43 |
| evand | don't get me wrong, I'd like to split failure_command | 19:43 |
| xivulon | pros and cons, might be a bit late for that though | 19:43 |
| evand | all it would take would be splitting failure_command. ubiquity already handles showing a page when you haven't adequately answered all of the questions. | 19:44 |
| evand | I think we're fine with our current approach though, at least for the time being. | 19:45 |
| xivulon | I asked cjwatson whether we should ask for username/pwd in ubiquity and he rejected the idea. | 19:45 |
| xivulon | I guess also because it would look akward to have to go through 2 installer interfaces | 19:45 |
| xivulon | But on the other side, that would make wubi a true 1 click installer, and I like that! | 19:46 |
| evand | indeed, mind you only in the case where the install failed once, but still like I said, I'm not keen on the idea. | 19:46 |
| xivulon | ah I see | 19:46 |
| xivulon | Hmm I'd say either we do it always or never | 19:46 |
| evand | I vote for never | 19:47 |
| evand | wait a minute | 19:48 |
| evand | don't you use MD5 hashes anyway? | 19:48 |
| xivulon | yes | 19:48 |
| evand | so then I'm confused | 19:49 |
| xivulon | by what? | 19:49 |
| evand | how is the password exposed in preseed.cfg? | 19:49 |
| xivulon | in hashed form | 19:49 |
| xivulon | so the fact that preseed stays around on failure is half damage | 19:50 |
| evand | half damage? | 19:51 |
| evand | I don't think there's anything to worry about if you're using user-password-crypted. | 19:51 |
| cjwatson | what are the permissions on preseed.cfg? | 20:01 |
| CIA-24 | debian-installer-utils: evand * r650 ubuntu/ (debian/changelog update-dev): | 20:02 |
| CIA-24 | debian-installer-utils: * Avoid sound events as they will trigger audio to be played on every | 20:02 |
| CIA-24 | debian-installer-utils: call to udevtrigger; udevsettle (LP: #178057). | 20:02 |
| xivulon | cjwatson do you mean on the windows side? I think anyone can read it | 20:05 |
| cjwatson | can it be given stricter NTFS permissions somehow? | 20:05 |
| xivulon | can try that | 20:06 |
| xivulon | not that I think that will limit an attacker ability to access it anyway... | 20:06 |
| cjwatson | right, but if they have sufficient access on the Windows side they can do whatever they want anyway | 20:08 |
| cjwatson | what I want to avoid is the file being left around such that unprivileged users created on the Ubuntu side can read it | 20:08 |
| cjwatson | but if it's only on failed installs anyway, it's not a big deal | 20:09 |
| xivulon | yeah I'd skip that if possible it would involve adding an ACL plugin just for that | 20:09 |
| CIA-24 | debian-installer-utils: evand * r651 ubuntu/debian/changelog: releasing version 1.50ubuntu2 | 20:17 |
| bdmurray | evand: are there test cases documented for testing Wubi? Actually is it testable yet? | 20:26 |
| evand | bdmurray: It is testable, there are no documented test cases yet. | 20:27 |
| evand | umenu should also replace WinFOSS testing | 20:28 |
| bdmurray | In the test lists at iso.qa.ubuntu.com? | 20:29 |
| evand | yes, WinFOSS is no longer used. | 20:29 |
| evand | I'll add Testing/Cases/umenu and Testing/Cases/Wubi pages to the wiki if you'd like. | 20:30 |
| bdmurray | evand: that'd be great and I'll get them added to the iso testing tracker | 20:31 |
| xivulon | how do those tests work? can I help on my side? | 20:32 |
| bdmurray | Is that arch specific at all? | 20:32 |
| evand | sure | 20:32 |
| evand | amd64 and i386 | 20:32 |
| evand | works on both kubuntu and ubuntu (edubuntu is now an add on CD) | 20:33 |
| evand | xivulon: The test cases are linked from http://iso.qa.ubuntu.com . bdmurray can talk to stgraber to get them added. The tests themselves are on the wiki | 20:34 |
| evand | https://wiki.ubuntu.com/Testing/Cases | 20:34 |
| * xivulon reading | 20:34 | |
| stgraber | evand: I'll need to rework the testcases list anyway, so just put a list of what's wanted on the wiki Testing/TrackerUpdate (or something similar) or e-mail me | 20:35 |
| xivulon | ah those are human tests | 20:35 |
| stgraber | (following the Edubuntu changes and LTSP being now part of Ubuntu Alternate) | 20:35 |
| xivulon | re Edubuntu I do have it in the wubi list of desktops | 20:35 |
| evand | stgraber: ah, I had no idea you were in here :). Will do. | 20:35 |
| xivulon | is it available as a Live ISO so that it can be targeted by wubi? | 20:36 |
| xivulon | I mean I can see: http://cdimage.ubuntu.com/edubuntu/daily-live/current/ | 20:37 |
| xivulon | and http://cdimage.ubuntu.com/edubuntu/releases/8.04/alpha-5/ | 20:38 |
| xivulon | first link seems to indicate a full ISO, second link an addon | 20:39 |
| evand | I'm not sure how long cdimage will continue to generate full live CD images. | 20:40 |
| xivulon | ah pls let me know, in case we have to take it out of wubi | 20:40 |
| xivulon | I hope they'll keep doing full CDs of course | 20:41 |
| stgraber | well, the live CD for edubuntu isn't that usefull as we don't have enough place on it to put the educational softwares | 20:46 |
| stgraber | and it ends up being just an ubuntu with a different artwork | 20:47 |
| stgraber | the add-on CD contains all the educational stuff and is then installed on a standard Ubuntu | 20:47 |
| stgraber | so if we wanted to install everything out of the box we would need ~900MB on the LiveCD (as the add-on contains ~200MB of usefull stuff at the moment) | 20:47 |
| xivulon | stgraber: 900MB would not be an issue for wubi since no CD is burnt ;) | 20:48 |
| xivulon | but of course that would be too much of a special case I guess | 20:48 |
| evand | 15:45:28 < slangasek> my understanding was that edubuntu desktop would remain a discrete image, only server was being rearranged into an ubuntu-server adjunct? | 20:49 |
| xivulon | so my understanding is that it will be around but not too usueful because of lack of packages | 20:49 |
| stgraber | yes, we'll keep edubuntu desktop, it'll just be as usefull as it was in Gutsy | 20:49 |
| xivulon | my point is that someone uses wubi to demo edubuntu might get the wrong impression since it won't be fully loaded | 20:50 |
| evand | 15:50:06 < slangasek> evand: that's my understanding; plans may have changed when I wasn't looking | 20:50 |
| evand | gotcha | 20:51 |
| xivulon | for me it's fairly easy to add/remove distros, and I am not binded by ISO size restrictions, just let me know what you want to have in! | 20:51 |
| evand | I'll let stgraber and the rest of the edubuntu people speak for their distribution and needs. | 20:53 |
| evand | and its* | 20:53 |
| xivulon | absolutely, stgraber, let me know what you decide | 20:53 |
| evand | and then I can flip the necessary bits on cdimage, if any changes are needed | 20:53 |
| stgraber | xivulon: we have a meeting this wednesday at 12:00 UTC. I'll ask here (you can also attend it depending on your timezone) | 20:54 |
| xivulon | I'll try to. | 20:55 |
| xivulon | as mentioned, 900MB ISO, is still an option as far as wubi go | 20:55 |
| bdmurray | evand: was bug 48355 fixed in dapper updates or just Hardy? | 21:32 |
| ubotu | Launchpad bug 48355 in ubiquity "MASTER: partitioner should warn about too small disk while autopartitioning (some_device)" [Medium,Fix released] https://launchpad.net/bugs/48355 | 21:32 |
| evand | hardy | 21:32 |
| bdmurray | That's weird the comment on the wiki pages says it was fixed in 6.06.1 | 21:32 |
| evand | perhaps a different bug? Or does it explicitly reference that bug number? | 21:33 |
| bdmurray | That number explicity, I'll change it though. | 21:34 |
| cjwatson | evand: do you know what happened to the notes from the USB stick production thing we discussed at UDS? | 22:22 |
| cjwatson | xivulon: we've already stopped doing full Edubuntu CDs; the desktop is just an accidental hangover that nobody has got round to disabling yet, AFAIK | 22:23 |
| cjwatson | bdmurray: we probably ameliorated it in 6.06.1 | 22:23 |
| xivulon | cjwatson, I cannot do multiple ISO installations at this stage, if edubuntu is to be added the only option is to have a 900MB ISO | 22:25 |
| evand | cjwatson: I'll look through my laptop for it. | 22:25 |
| evand | xivulon: you wouldn't have to. They install Ubuntu using wubi, then use the edubuntu addon CD. | 22:25 |
| xivulon | yeah that would be good enough | 22:26 |
| cjwatson | bdmurray: I seem to have added that text deliberately but it was a long time ago | 22:26 |
| xivulon | I will remove edubuntu then | 22:26 |
| cjwatson | evand: ta | 22:26 |
| cjwatson | the gobby notes may be available from IS if need be | 22:26 |
| xivulon | re USB stick, I have vague memories and remember that Wubi was involved somehow | 22:26 |
| xivulon | can you remind me briefly what was the objective? | 22:26 |
| * cjwatson -> bed, sorry, wife will be annoyed otherwise :) | 22:27 | |
| cjwatson | another day | 22:27 |
| cjwatson | (it's not urgent) | 22:27 |
| xivulon | sure | 22:27 |
| xivulon | I'll be glad to hack something together anyway | 22:27 |
| xivulon | I'll wait wednesday meeting before removing edubuntu from wubi, I guess there is no rush | 22:29 |
| xivulon | cjwatson quick q | 22:34 |
| xivulon | might be good to use low swappiness as well? | 22:34 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!