| eggonlea | hi, any tips to solve the "sudo: must be setuid root" issue when running "sudo" command in a chroot+qemu-arm-static environment? | 06:49 |
|---|---|---|
| * persia tests | 06:52 | |
| persia | Well, it doesn't happen in my lucid armel chroot (on amd64), so I wonder if there is some other factor involved. | 06:53 |
| persia | What does `ls -l /usr/bin/sudo` show? | 06:53 |
| eggonlea | it's of correct attribute | 06:53 |
| persia | -rswr-xr-x ? | 06:54 |
| eggonlea | -rwsr-xr-x 2 root root 98880 Jan 29 15:29 /usr/bin/sudo | 06:54 |
| eggonlea | the problem, I think, is in /usr/bin/qemu-arm-static | 06:54 |
| eggonlea | is it? | 06:54 |
| persia | Hrm. I saw this before when the rootfs data had been copied in and out of a FAT32 filesystem, but that's not your issue.] | 06:55 |
| persia | Well, I'd agree, except that I don't get that message :) | 06:55 |
| persia | How did you construct your chroot? | 06:55 |
| eggonlea | build-arm-chroot lucid | 06:55 |
| eggonlea | I chroot this arm rootfs in a x86 PC | 06:56 |
| eggonlea | so "sudo" after "chroot" is executed via qemu-arm-static actually | 06:56 |
| persia | Indeed. | 06:56 |
| persia | Looking at build-arm-chroot, I don't know of any reason that command wouldn't build the chroot you expect. | 06:57 |
| eggonlea | so, do you mean you could run sudo command correctly in such a chroot env? | 06:57 |
| persia | I'm on amd64. Are you on i386? I wonder if that might be a difference. | 06:57 |
| eggonlea | y | 06:58 |
| eggonlea | I'm running Karmic on i386 | 06:58 |
| persia | In that case, I'm inclined to agree with you that it's qemu-arm-static, as that seems to have been changed between karmic and lucid. | 06:59 |
| eggonlea | I tried chroot into an "installed arm rootfs from livecd" as well. same error. | 06:59 |
| persia | build-arm-chroot is the same. | 06:59 |
| eggonlea | so I don't think it's caused by build-arm-chroot. | 07:01 |
| persia | Neither I, as my chroot is also created by build-arm-chroot (albeit embedded from another script) | 07:01 |
| persia | Anyone else using build-arm-chroot with Karmic? Do you get the error eggonlea mentions? | 07:02 |
| eggonlea | persia, what's your Ubuntu version on amd64? | 07:06 |
| persia | lucid | 07:07 |
| persia | Upgraded just a few hours back. | 07:07 |
| persia | I don't recommend this for production use unless you're prepared to get two pieces when it breaks, but it's a good way to check to make sure one's preferred bugs get fixed. | 07:07 |
| * eggonlea testing in Lucid/vmware | 07:08 | |
| persia | Now there's an idea! | 07:08 |
| persia | If you see a difference, file a bug clearly indicating that it doesn't work in karmic and it does work in Lucid. I'm not sure we can get an update, but a backport is likely possible if a few people can test. | 07:09 |
| * eggonlea hates the slow network... | 07:19 | |
| persia | Indeed. | 07:19 |
| eggonlea | same error in lucid/vmware + build-arm-chroot/qemu-arm-static | 07:52 |
| persia | Very odd. I wonder why I don't get the error. | 07:56 |
| persia | Please file a bug about this (`ubuntu-bug qemu-kvm`) | 07:57 |
| eggonlea | do you have i386 host? | 07:57 |
| eggonlea | ok | 07:57 |
| persia | I don't. Just amd64. | 07:58 |
| eggonlea | bug filed. | 08:09 |
| eggonlea | thanks for your kindness. | 08:09 |
| lool | persia: Great to see qemu-arm-static support in ubuntu-dev-scripts! You might want to test for presence of build-arm-chroot instead of checking whether the arch is either i386 or amd64 (error would then be "you need the qemu-arm-static package..." and you could use a suggests instead of recommends); BTW the package just got renamed to qemu-kvm-extras-static | 14:16 |
| dmart | asac: hi | 16:12 |
| persia | lool: I did in mk-sbuild-lv. I'm not as sure how to do that with pbuilder-dist. I'd appreciate a hand if you have any ideas. | 18:05 |
| lool | persia: Sure | 18:10 |
| lool | persia: I thought you implemented that though? | 18:10 |
| persia | python != shell :) | 18:12 |
| persia | Plus, the logic in pbuilder-dist is particularly opaque to me. | 18:12 |
| persia | (plus I don't use pbuilder, which makes verification interesting) | 18:13 |
| lool | persia: Sorry, to clarify: you're more confortable writing shell than python and you would prefer me to improve the logic in pbuilder-dist; correct? | 18:22 |
| lool | dmart: asac is travelling in the west coast; might be easier to email him | 18:22 |
| persia | lool: please :) | 18:23 |
| dmart | lool: I did... not urgent anyway. Thanks | 18:23 |
| asac | dmart: hi | 18:24 |
| dmart | Oh, hi there | 18:24 |
| asac | what can i do ;)? | 18:24 |
| dmart | how's the sprint? | 18:24 |
| StevenK | lool: He isn't travelling, he's at the sprint, and has been for 4 days | 18:24 |
| asac | quite good ... some crisis here and there ;) ... but we seem to have a good new lead on dove | 18:24 |
| asac | will know more soon | 18:24 |
| lool | StevenK: He is not at home, so he is on travel for work, hence travelling; sorry if my English explanations suck, but the effect is the same :-) | 18:25 |
| * persia blames the awkwardnesses of English spelling and grammar | 18:26 | |
| dmart | at least it's not C++ grammar | 18:27 |
| dmart | :P | 18:27 |
| StevenK | Haha | 18:27 |
| persia | At least that's documented in a single way :) | 18:27 |
| * StevenK laments his lack of C++ to make a joke that builds on dmart's joke | 18:28 | |
| lool | C++? haha | 18:28 |
| asac | dmart: wonder ... could you add examples or give pointers how to properly do the "mov" fixes for thumb2 | 18:28 |
| asac | ? | 18:28 |
| asac | would like to get the thumb2 list down a bit more this week | 18:29 |
| dmart | I had some thoughts about that... did you get a chance to look at the stuff on https://wiki.ubuntu.com/ARM/Thumb2PortingHowto | 18:30 |
| dmart | It still rather incomplete, but progressing gradually. | 18:30 |
| dmart | I've tried to keep the amount of extra information down, but it may still be quite indigestible to a newcomer. | 18:31 |
| dmart | I was thinking maybe we should schedule a sort of mini-sprint with some mobile team guys sometime next week... the best way to get people up to speed may be for everyone to pick up a package and try to work out how to fix it... then watch the discussion on IRC. | 18:33 |
| dmart | Does that sound like a good idea? | 18:33 |
| asac | dmart: no ... didnt know that page exists | 18:35 |
| * asac checks | 18:35 | |
| asac | looks like a good start | 18:35 |
| asac | will check that | 18:35 |
| asac | thanks | 18:35 |
| asac | dmart: ^ | 18:35 |
| asac | dmart: yes. i like the idea of making porting sessions | 18:36 |
| asac | i think we fixed almost all swp ones now | 18:36 |
| asac | (boost upload pending) | 18:36 |
| asac | so mov is the most pressing one left i think | 18:36 |
| dmart | Yes... particularly since these don't generally show up as ftbfs, we need to be more careful. | 18:36 |
| dmart | OK; I'll probably try and flesh out the missing bits of that wiki page by early next week and we can discuss the way forward at the ubuntu-mobile IRC meeiting. | 18:38 |
| dmart | asac: ^ I was trying to look at the evolution-data-server update to help sanity-check, but looking in debian/patches I couldn't find the new patches described in the update. | 18:40 |
| dmart | Should I see the patches after apt-get source, or do I actually need to build the package (or look in bzr etc.) | 18:40 |
| asac | dmart: i also committed it to bzr, yes. | 18:43 |
| asac | dmart: you couldnt find the patches? | 18:44 |
| asac | hmm | 18:44 |
| asac | guess i messed that | 18:44 |
| asac | darn ;) | 18:44 |
| dmart | I think I have the right package version; debian/changelog mentions the patches | 18:44 |
| dmart | I tried a build anyway... it hasn't fallen over yet, but it's not finished... | 18:45 |
| asac | yes | 18:45 |
| asac | i think i failed to add it to bzr | 18:45 |
| asac | darn | 18:45 |
| asac | sigh ... i knew i should have uploaded without doing bzr first | 18:45 |
| asac | guess i have to redo it | 18:45 |
| * asac hopes he still has the build tree somewhere | 18:46 | |
| persia | As far as I can tell, if one uploads whilst ignoring bzr, automated systems do the right thing. | 18:46 |
| asac | good still have it in a porter bux | 18:46 |
| asac | persia: i remembered seb wanting me to commit it | 18:47 |
| asac | so i wanted to be nice | 18:47 |
| * asac goes and redoes | 18:47 | |
| lool | asac: You might have it in your checkout still | 18:47 |
| lool | or in the package itself | 18:47 |
| lool | Ah no | 18:47 |
| lool | Your debdiff is definitely empty | 18:47 |
| asac | i did the checkout in /tmp/ ;) | 18:48 |
| asac | but i found it on the porter machine :-P | 18:48 |
| asac | \o/ | 18:48 |
| asac | dmart: you can check boost1.40 | 18:49 |
| dmart | ok... I'll try and take a look, probably tomorrow. | 18:50 |
| asac | yes | 18:51 |
| asac | make more sense | 18:51 |
| asac | dmart: check boost1.40 | 18:51 |
| asac | i think thats uploaded ;) | 18:51 |
| dmart | ok, will do | 18:51 |
| Marrs | hi all, I'm looking for some information about running openjdk on ARM, does anybody here have links to some info on this (are there binary packages available for it, etc..)? | 20:29 |
| lool | Marrs: https://launchpad.net/ubuntu/+source/openjdk-6 | 20:39 |
| lool | Marrs: The latest lucid version is still building on armel, but yes, there are usually binaries available for armel | 20:40 |
| persia | There's been a couple reports about some code paths causing VM issues, so heavy testing and reporting of any issues would be very welcome. | 20:40 |
| Marrs | ideally, I'm looking for a stable binary release, so that would be the karmik one? | 20:40 |
| Marrs | karmic that is :) | 20:41 |
| Marrs | persia: I'm definitely going to try to kick its tires :) | 20:41 |
| persia | karmic is a stable binary release, but if you find issues, those are easier to fix in lucid (and moreso if previously verified in lucid). | 20:42 |
| persia | I'd recommend starting with karmic, and maybe testing against lucid in a chroot if you encounter issues. | 20:42 |
| Marrs | persia: understood | 20:42 |
| Marrs | btw, I'm not so experienced in the ARM architecture, I'm more of a Java guy, could you take a look at this ARM based board and tell me if you think it will run openjdk? (looking up link now..) | 20:43 |
| persia | Better question is whether it can run Ubuntu, and which releases :) | 20:44 |
| persia | Running OpenJDK is mostly a side-effect of that. | 20:44 |
| Marrs | fair enough... supplier says it can run ubuntu, I have yet to find out which version, but this is the board: http://www.msc-ge.com/de/produkte/com/exm32/3640-www.html | 20:44 |
| Marrs | whoops, that was the german page, hit the english flag in the top right corner for an instant translation ;) | 20:45 |
| persia | Deutsh ist also gut. | 20:46 |
| * persia can't spell today | 20:46 | |
| Marrs | ;) ok | 20:46 |
| persia | I'd recommend against this one. I think it can run Jaunty with a custom kernel, but it wouldn't be able to run karmic or lucid. | 20:47 |
| Marrs | ouch :( | 20:48 |
| Marrs | what's the main problem? | 20:48 |
| Marrs | or, what type of board should I look for | 20:50 |
| persia | You want something that can handle the ARMv7 ISA. | 20:56 |
| persia | So, if looking at something using an i.MX processor, you'll want at least the i.MX51. | 20:57 |
| persia | There's also stuff from other vendors, of course :) | 20:57 |
| Marrs | ok | 20:59 |
| Marrs | if you have any links to a board you could recommend, that would be welcome | 21:03 |
| lool | Marrs: The boards which are officially supported in Ubuntu kernels aren't that cheap or common, but it's improving quickly; popular boards around here are OMAP 35xx or 34xx based boards e.g. beagleboard | 21:06 |
| lool | beagleboard is probably the one we get the most pinged about | 21:06 |
| Marrs | thanks | 21:06 |
| Marrs | I'll look for those and see if we can use them | 21:06 |
| armin76 | lool: +available | 21:16 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!