[06:53] mwharris: you could maybe abuse partman/early_command instead [08:20] infinity: bug 1172002 relates to a higher priority bug 1066480 which depends on bug 291494 [08:20] Launchpad bug 1172002 in ubiquity (Ubuntu) "Install doesn't mount encrypted swap for reinstall" [High,Confirmed] https://launchpad.net/bugs/1172002 [08:20] Launchpad bug 1066480 in ubiquity (Ubuntu S-series) "Installer doesn't show encrypted partitions" [Critical,Triaged] https://launchpad.net/bugs/1066480 [08:20] Launchpad bug 291494 in partman-crypto (Ubuntu) "Partitions selected for encryption cannot be erased anymore" [High,Triaged] https://launchpad.net/bugs/291494 [08:22] xnox: Fun, fun. That sounds like a tangled mess that you don't have a hope in hell of fixing in the next few hours. :P [08:24] infinity: yeah, I mean partman need fixing preferrably in debian to allow deactivating device mapper devices.... [08:59] xnox: I'm here now if you still need a hand [09:03] ev: nah, it's ok. i think you should visit SpecSavers..... wubi-r279-signed in your home dir was actually r275. I rebuild wubi, got it resigned and it got respun for already. [09:10] oh wow, sorry about that [09:10] I'll have a look and figure out what happened [09:10] this was for raring, not precise, right? [09:11] we didn't check precise [09:14] we did check that IS had done the right thing - wubi-r279.exe was r275 too [09:14] what's strange is that I do this: scp build/wubi.exe people.canonical.com:public_html/wubi/raring/wubi-r$(bzr revno).exe [09:14] stale build, maybe [09:15] Yeah, could be [09:17] FWIW, we took the opportunity to have cdimage fetch from ~ubuntu-archive/wubi/ instead, and prepopulated that with a copy of your directory - but I want to make that a symlink or give you an ACL on it or something so that we don't have to have an extra copy step [09:18] woo! [09:18] thank you! [09:18] yesterday it was mostly just so that we didn't have to have IS abuse privs and write to your directory [09:22] it is indeed a stale build [09:22] (just ran it in wine) [09:22] awful, sorry again. I'll make sure that I pay close attention to the bottom left corner of the main screen after testing a build [09:33] ev: Or, you could stay on IRC 24/7, even when on VAC, like a properly dedicated Foundations developer. [09:33] ev: Sheesh. Some people. [09:33] infinity: I was on a motorcycle! [09:33] No excuse. [09:33] hahahaha [09:33] it was later joked that I answer the phone while on the test course and just spin the bike in circles [09:33] waving down the examiner [09:49] patch for kubuntu: http://paste.ubuntu.com/5597907/ [09:54] looks good to me [09:55] we shouldn't show the bad password icon when the password box is empty, imo [09:56] I mean, yes, it's not sufficient to pass the dialog, but it kind of means you start up the dialog with an error message, doesn't it?, which looks a bit odd [09:56] cjwatson: true. [09:57] cjwatson: just allow_go_forword(False) on empty password? [09:57] Yeah [09:57] ok, let me test. [09:58] cjwatson, then kubuntu will be the first password checker in the world that doesn't do that [10:00] cjwatson: http://paste.ubuntu.com/5597937/ [10:00] Noskcaj: Er, the user setup page in the Ubuntu installer is the same as what I'm proposing [10:00] ok, second ;) [10:00] And it's common practice on web pages too [10:00] They generally only start complaining after you try to use the entry box [10:02] xnox: Technically right but it's a bit twisty [10:02] I'd prefer the allow_go_forward to be in a different conditional [10:02] Or use a variable and then allow_go_forward(complete) or whatever [10:14] http://paste.ubuntu.com/5597952/ [10:14] but maybe just a paste is easier to read. [10:15] would http://paste.ubuntu.com/5597947/ work as a fix to bug 1171989? [10:15] Launchpad bug 1171989 in ubiquity (Ubuntu) "Duplicate filesystem types" [Undecided,Confirmed] https://launchpad.net/bugs/1171989 [10:16] Noskcaj: you can probably drop the if statement. If the combo is generated afterwards anyway, clearing it when it's empty won't hurt [10:17] http://paste.ubuntu.com/5597961/ [10:17] (also you seem to have tab damage in that diff) [10:17] stgraber, ok. [10:17] Noskcaj: and what cjwatson said, make sure not to mix tabs and spaces, it looks like you've been using tabs for your change and that may confuse python [10:18] stgraber, i just copied the link for sadandblue [10:18] he made the fix [10:23] stgraber, cjwatson, Noskcaj: Cool thanks, I'll attach a proper diff to the bug report. [10:40] https://bugs.launchpad.net/ubuntu/+source/ubiquity/+bug/1171989 I added a (hopefully proper!) patch for it. [10:40] Launchpad bug 1171989 in ubiquity (Ubuntu) "Duplicate filesystem types" [Undecided,Confirmed] [10:47] stgraber: cjwatson: here is the final fix for crypto/partman qt bug https://code.launchpad.net/~xnox/ubiquity/fix-crypt-qt-page/+merge/160602 [10:47] not sure if we will have any other ubiquity fixes or this just for kubuntu. [10:47] Riddell: ^ [10:51] xnox: well, I guess we want to also pull the fix for bug 1171989 in any case [10:51] Launchpad bug 1171989 in ubiquity (Ubuntu) "Duplicate filesystem types" [Undecided,Confirmed] https://launchpad.net/bugs/1171989 [10:52] now I'm not convinced that any of those two is critical enough to warrant an ubiquity upload (which would require respinning pretty much all images) [10:52] so I think they fit in the (will be fixed if we find something worse that we have to respin for) [10:57] stgraber: the crypto page, is purely UI / UX bug. The crypto install will finish, it's just that the first password will be used. The image is shown when passwords do not match. So it's not all lost. [10:57] stgraber: i think infinity and cjwatson were considering to upload and respin only kubuntu [11:00] stgraber: What xnox said. We might do a ubiquity to raring-updates and rebuild kubuntu against it. [11:01] ok [11:16] cjwatson or somebody else with excellent English linguistic skills: is bug 1171477 which in essence is """The slide that says the launcher autohides to "preserve screen estate" should say "preserve screen real estate".""" [11:16] Launchpad bug 1171477 in ubiquity-slideshow-ubuntu (Ubuntu) "Typo in install slideshow - missing "real" before "estate"" [Medium,Triaged] https://launchpad.net/bugs/1171477 [11:20] followed up to the bug [13:17] server preseed testcase says boot with "auto auto=true url=http://...../preseed.cfg priority=critical" [13:18] * xnox is sure it's auto=true, and the auto is redundant.... [13:23] No, auto is an alias for auto-install/enable [13:23] Er [13:23] Actually that would apply to auto=true wouldn't it [13:23] I think you're *probably* right that auto is redundant [13:47] preparing ubiquity release. [13:47] Riddell: ^ [13:58] xnox: thanks === kentb-out is now known as kentb [15:45] anyone know if image upgrade as described in http://iso.qa.ubuntu.com/qatracker/testcases/1498/info is a valid test for ubuntu-server installs? I'm thinking it's not, but want to make sure [15:46] if so, it doesn't seem to work, but iirc, upgrading from the cd image is not a supported install option except on desktop [15:46] xnox: ^? Do you know? [15:46] not valid for server [15:47] plars: the test is not valid, but such funcationality does exist. [15:48] (e.g. upgrade via apt-clone and preserve partitioning) [15:48] not sure if that's tested or not. [15:48] Err [15:48] I'm pretty sure it doesn't exist from the server CD [15:48] I fixed it to work in quantal images.... [15:48] In what package? [15:48] partman-auto [15:48] I have not seen that option on 12.04. [15:49] partman-auto doesn't have any apt-clone code [15:49] it offers reuse & replace recipes.... what happens next is a mistery, but I see your point. [15:49] xnox: I don't seem to have an option for it - just guided( resize, reuse, entire disk, lvm-*, etc), and manual [15:49] That isn't sufficient to do an "upgrade" (aka install over the top and put stuff back) [15:49] Trust me, it's not there [15:49] ack. [15:49] plars: ignore me. [15:49] =) [15:49] xnox: ok, so not supposed to be there either, right? [15:50] plars: Correct [15:50] good [15:50] It was only ever implemented in ubiquity [15:50] As xnox says, some limited bits of the associated partitioning code exist in d-i too [15:50] But that's only a fairly small part of it [15:51] (I'm not sure what would happen if you tried to use the reuse recipe in d-i - either it'll debootstrap over the top and pray, or it'll explode) [15:52] * antarus drools over re-use [15:52] One test case that should be added is post-install/use oem-config-prepare. I seem to be hitting a wall with it, at least on Precise. [15:53] antarus: Probably not as good as you're hoping for [15:53] It's reuse without removing the contents, which I don't think is what you want [15:53] (IIRC) [15:56] I know ;) [15:58] GrueMaster, what "wall" are you hitting? The oem test runs through the user install also [15:59] I need to be able to do post-install testing and configuring on the imaged system before setting up oem-config. oem-config-prepare is supposed to do this for me, but oem-config crashes on reboot. [16:00] And setting debug-oem-config on the kernel boot prompt doesn't give any detailed info. [16:01] Unfortunately, my day job has been very busy with other tasks, and I haven't had a chance to sit down and really debug this. [16:13] * GrueMaster speaks and the room goes silent. :| [16:14] Nothing personal, very busy [16:14] hey GrueMaster [16:14] ogra_: !!! [16:15] :) [16:15] didnt see you here ... [16:16] * ogra_ is more busy in #ubuntu-touch nowadays [16:16] Ive been lurking for a few weeks now. [16:16] Wish I had time to play with U-T. [16:17] I have a friend that works here at Intel on the Surface systems with Windows 8. Trying to push a "real" os on her. :P [16:18] haha [16:19] this is windows 8 surface [16:19] not like the surface microsoft invented 10 years ago but never sold [16:19] like the table [16:20] which honestly was pretty neat..I still don't understand how they failed to capitalize on it so badly [16:20] It had less demand than WebTV. [16:21] MS has a hard time inventing new markets. They are better at taking over existing markets by brute force. [16:21] well, who wants a tablet that only runs 3-4h [16:22] who wants a laptop that only runs 3-4h [16:22] * antarus covets his pixel [16:22] I DO [16:22] But it has the latest in Poulsbo technology. :P [16:22] * GrueMaster means hardware, not drivers. [16:22] that reminds me, I need to go find the power supply for my laptop, it's almost out of battery after 9 hours of work [16:24] stgraber: which model gives you 9h? [16:24] * antarus got about 7 on his x220 [16:24] antarus: thinkpad x230 with the 9cell battery [16:24] ahhh nice [16:25] * antarus needs to buy a new desktop :/ [16:25] I usually get around 9-10h on wifi (pretty low brightness though) and around 12-13 without wifi [16:26] I have a pixel (4h tops) and a t420s..which basically can't even run linux [16:26] get arm HW ! [16:26] the t420s just sits on my desk looking pretty [16:26] * antarus lost his x220i on a plane, and traded his other x220 for a pixel [16:27] ogra_: Sadly, Intel frowns on me bringing arm HW to work. [16:27] well, they start doing SoCs now at least [16:27] ogra_: give me a quad-core armv8 with 16GB of RAM that can last 10h on battery and I'll consider it ;) [16:27] * antarus boggles [16:27] (not that I've tried though) [16:27] seems they hired quite a few former pandaboard guys from TI [16:27] * antarus somehow doubts they frown on it at all ;p [16:28] The group I'm in? Yes, they do. [16:28] stgraber, no problem ... [16:28] * ogra_ hands stgraber a time tunnel ticket [16:28] just buy it a few years in the future :) [16:29] GrueMaster: heh, I'm the lead for Goobuntu at Google, and all my stuff at work is ChromeOS (because lets face it, goobuntu sucks) [16:29] I get made fun of, but not frowned upon ;) [16:30] My co-workers don't even like it when I point out that they use Arm processors for the BMC in their server platforms. [16:31] do they..have phones? [16:31] Yea...in india. [16:31] Oh, you mean my co-workers. yes. But the company is pushing to change that as well. [16:33] At any rate, I didn't mean to start a distraction in this room. I'm here mainly to monitor installer issues. [16:33] I'm in here mainly to troll cjwatson [16:34] I have beer, trolling possibilities are limited [16:35] when will d-i be rewritten in go? [16:36] * GrueMaster chokes [16:43] hey man, its better than sh [16:43] ;p [16:43] I thought d-i was partially perl based. [16:44] There are so many components to it. [16:45] the parts I read were sh and C [16:46] d-i isn't perl, no [16:46] No perl interpreter in the installer environment [16:48] I almost would say I wish perl was in there [16:48] but I don't [16:48] python now.. [16:48] I'd take that [16:48] antarus: Just be thankful it isn't written in tcl. [16:48] I don't know enough about tcl to know if it sucks worse than perl [16:49] could be worse [16:50] (it could be fortran or cobol ... ) [16:50] * xnox thinks all of the above should be in #ubuntu-offtopic. [16:51] * GrueMaster gets back to work. [16:51] * xnox wishes -offtopic was as good as it used to be years ago [16:55] is there a way i can promopt the user while running a d-i early_command? [16:55] id like to pop up a box and wait for them to hit enter [16:55] mwharris: look at our...late command? [16:57] ahh you want a saner thing than that [16:57] mwharris: so there is like the 'isntall failed, press to continue' thing [16:57] mwharris: I'm not sure where in d-i that lives [16:57] guessing it is dialog based [16:57] mwharris: maybe decompress the initrd and poke around [17:08] that lives in our bootstrap package, and does some debconf stuff. i don't know if that'll work outside a package [17:09] d-i early_commands can use debconf, although you'll have to dance around a bit to cram the appropriate templates into the debconf db first [17:09] yeah, i saw we'd need templates. [17:09] i think i'll try that, thanks. [19:42] anyone have a little magic that would tell ubiquity to skip the installation of libre office? [20:20] johnsterdotcom: no. [20:21] johnsterdotcom: ubiquity installation takes the whole of squashfs and copies all contents to the target drive. You can presseed to uninstall it straight after at the end of installation, or you need to remaster a cd to exclude libreoffice. There is a wiki page about cd remastering. === kentb is now known as kentb-out