[01:47] hey guys can someone point me to a ubuntu how to for beaglebone. Seem I'm not using the right image [01:48] infinity: ogra_: bug 983555 [01:48] Launchpad bug 983555 in compiz-plugins-extra "compiz-plugins-extra FTBFS on ARM" [Medium,Confirmed] https://launchpad.net/bugs/983555 [01:48] and attached a debdiff containing the patch that disables the plugins that are gl-only compatible [01:49] can someone tell me if http://cdimages.ubuntu.com/releases/oneiric/release/ubuntu-11.10-preinstalled-server-armel+omap.img.gz is the right image for beaconboard? [01:49] I mean beaglebone [01:50] rsalveti: \o/ [01:50] electroglue: Other than the part where I'd strongly recommend using precise/armhf instead of oneiric/armel, yes. [01:51] infinity: armhf performance is much better than the soft one? [01:51] * twb perks up - ubuntu has a solid armhf now? [01:51] Better performance, plus we're dropping support for armel. [01:51] twb: Hrm? [01:51] twb: Where have you been for the last 6 months? [01:51] infinity: last time I looked, which was months- asleep [01:52] thanks infinity [01:52] wrt ubuntu arm I well and truly have my "user" hat on and my "dev" hat off :P [01:52] I have a few too many dev hats. [01:53] rsalveti: So, I'm guessing no Linaro folk got around to taking gnat off markos' plate? [01:55] rsalveti: Also, uploaded that compiz fix for you. [01:56] infinity: I thought SteveMcIntyre was looking at it, need to ping folks to check [01:56] let me write down an email about that [01:56] infinity: thanks [01:56] rsalveti: Unless someone magically produces a compiler in the next few days, we're pretty much officially "too late" anyway. [01:56] yeah =\ [01:56] Thankfully, no one actually uses ADA... Right? [01:57] well, there are always some weird folks that might be using it, but afaik not that many at least :-) [01:58] Well, the pet hobby language of the month seems to be haskell. [01:58] yup [01:58] I'd still love it if we got ghci working for those folks. [01:58] But at least ghc works. [02:00] cool, and do we have any issue with other haskell related packages? [02:01] I remember they were always a bit of a pain go get all in sync and building for arm [02:01] issues with builder and such [02:01] haskell-src-exts was having OOMing issues, that was fixed. [02:01] cool [02:01] So, it's only ghci that needs love. And I'm told that porting ghci to a new arch is a lot of Not Fun. [02:01] Which is why upstream only supports two arches. :/ [02:02] oh, ok, so will probably take time then [02:02] But I know, for instance, that CompSci classes teach interactive ghc usage, etc, so to many people, ghc without ghci is "useless". [02:02] infinity: GHC struggles on "second class" archs because it has to compile to C first, then run a C compiler [02:03] Something to ponder spending time on for Q anyway. Maybe. [02:03] infinity: last time I looked upstream GHC team didn't support arm at all, that was all Debian [02:03] infinity: if they want GHCI they can bloody well ssh into the student shell server :-/ [02:03] twb: Yeah, most of the porting effort with GHC has been Debian, and the Debian maintainer's awesome, but he's also a limited resource. :P [02:04] I used to maintain Darcs in Debian, from 2.0 up to about 2.4 or .5 [02:06] Hrm, Uploaders gets filtered out by dpkg-gencontrol, doesn't it? [02:06] So I can still fix this source without rebuilding the binaries. [02:06] Shiny. [02:07] you you guys saying alot is broken with armhf? [02:07] scientes: Other than gnat, no, it's just as broken (or slightly less) a armel. :P [02:07] s/a armel/than armel/ [02:08] cause i ran into the opposite with firefox [02:08] to run on armv5 you have to turn off the JIT or you get segfaults [02:08] Meh, ff is always a problem child [02:09] scientes: you'll want Debian's build if you're using armv5 [02:09] twb, ever try to build ff? [02:09] Yes [02:09] micahg, yeah that is what i was using [02:09] Although nowadays it's separate from xulrunner so the heavy lifting happens there [02:09] micahg, it works except it doesn't turn off the jit by default for some reason [02:09] twb: AIUI in Debian xulrunner and Firefox are built from the same source now [02:10] oh. [02:10] the 2GB for libxul linking sure caught be off gaurd [02:10] grumble grumble [02:10] twb: AIUI in Debian xulrunner and Firefox are built from the same source now [02:10] correct [02:10] i like that [02:10] scientes: hey man you can't even CHECK OUT emacs from bzr unless you have 800MB of RAM [02:10] Let alone compile it [02:10] well that a bzr problem [02:10] Yeah I know ;-) [02:10] which is not exactly something i care about [02:10] Checking out emacs seems more like a personal problem. [02:11] haha [02:11] infinity: it's a communicable disease [02:11] This is right up there with "I find it really difficult to masturbate to rotting corpses". [02:11] i just dont see why in bzr explicit renaming support is touted as a "feature" [02:11] The obvious response being "don't do that, then." [02:11] the content-centric approach of git seems much better [02:12] scientes: it's better enough that Darcs adopted it too [02:12] > its cancer enough that darcs adopted it too [02:12] FTFY [02:13] infinity, so are you guys going to maintain some sensible armel multiarch so that broken packages can still be installed as armel? [02:13] If you analyse the Θ of typical operations in a traditional delta-store backend vs. a content-oriented database, the latter is a clear win for VCS [02:13] scientes: We do support multiarch, yes. [02:14] so you will have a armel and armhf repo like debian, just not support the armel for primary arch? [02:15] We have both for now. We'll have both for the life of precise. I can't say if we'll have both forever. [02:15] (really, probably not) [02:16] well then they cant accuse your of breaking stuff that doesn't work on armhf, they can install the armel packages, and port it if someone cares [02:16] sounds perfectly reasonable to me [02:16] I thought not all of the multiarch ducks were lined up yet [02:17] Like, it's in dpkg now, but not in apt yet [02:17] I have a fine collection of ducks here. [02:17] It's been supported since oneiric... [02:17] haha, yeah there is alot of gui lack, and i've seen bugs in aptitude [02:17] We shipped with multiarch on by default for amd64 in oneiric. [02:17] aptitude is buggy, yes. [02:17] but ubuntu has had it for a while unlike debian [02:17] But who cares, it's aptitude. :P [02:17] infinity, I use it! [02:17] aptitude is better in precise, but still needs help [02:18] infinity: oooh, so you did biarch while waiting for Debian to solve the general case? [02:18] i use apt-get and aptitude [02:18] * micahg would like apt-getitude [02:18] i use to use synaptic, but now heavily dislike it, along with software-center [02:18] twb: [02:18] twb: Err, what? [02:18] twb: No, multiarch in oneiric was multiarch. [02:19] So I could add precise armhf to my existing oneiric/precise armel sources.list, and it should Just Work? [02:19] Well, to precise/armel, sure. [02:19] twb, if you have armv7 hardware, yes [02:19] Is there a howto wiki page somewhere? [02:19] Given that multiarch needs package versions to match, oneiric/precise won't exactly work. [02:19] And it's not added in sources.list at all. [02:20] * scientes uses [arch=amd64] in sources to make downloads faster [02:20] infinity: what I mean is I'm currently using oneiric and cherry-picking things I care about from precise [02:20] twb: echo "foreign-architecture armhf" > /etc/dpkg/dpkg.cfg.d/multiarch && apt-get update [02:20] twb: Congrats, you're multiarch. [02:20] ^^bingo [02:20] Thanks [02:20] and you can also change the primary arch to armhf [02:21] twb: I'd recommend having an armhf base, though, and cherry-picking in the other direction if absolutely necessary. [02:21] (Which, unless you really like ADA, it shouldn't be) [02:21] armhf can bring like 25% performance improvement [02:21] infinity: pascal as well [02:21] I'm currently running a random wacky android .36 kernel. Do I need that to be compiled with specific options for armhf to work, and if so, how do I check those options? [02:22] and even more on VFP-intensive stuff [02:22] micahg: I'm still pondering a 0-hour pascal transition. [02:22] scientes: yeah I was going to put Debian armhf on this box but my previous box died and I was in a rush :-( [02:22] twb, i think it even works with a armel kernel [02:22] twb: The kernel doesn't give a hoot about VFP. [02:22] That's what I thought, thanks [02:22] No such thing as an "armel" or "armhf" kernel. [02:22] but you can compile your kernel with VFP support, but there is very little FP in the kernel [02:23] Hm, I can't see apt-get update pulling down armhf entries yet... [02:23] the option is there [02:23] in make nconfig [02:23] http://paste.debian.net/163468/ [02:23] I guess because it hasn't got to the precise one yet, and there isn't one for oneiric? [02:24] There's no armhf in oneiric, no. [02:24] Yep, there's the armhf dl now [02:24] upgrade! [02:24] SHINY [02:25] The main thing is to avoid bricking it in any way, because this is my main / only workstation [02:25] You're as bad as ogra... [02:25] So-rry [02:25] ;) [02:25] If this was a normal server I wouldn't be worried about the userland bricking it because I can deal with that [02:26] twb, then use even more shiny like me and switch to btrfs :P [02:26] But this stupid thing can't exactly just boot off a USB live key atm :P [02:26] scientes: how do you think my last laptop died [02:27] * infinity slams his head on the desk. [02:27] I need to alias git/bzr/svn all to one wrapper that checks for .git, .svn, bzr and DTRT. [02:27] Not to worry tho, my scratch VM is still btrfs all the way [02:28] infinity: http://cyber.com.au/~twb/.bin/twb-get [02:28] So sick of running the wrong commands in the wrong projects. [02:28] infinity: or you could use Emacs which does exactly that ;-) [02:28] infinity: debcheckout/debcommit? [02:28] micahg: mm, that too [02:28] micahg: In this case, it was "bzr diff" in a subversion checkout. [02:29] Last night, it was git merge. Because I can't imagine merging without git. [02:29] But then I realised I had to. [02:29] Etc. [02:29] I need to alias git/bzr/svn all to one wrapper that checks for .git, .svn, bzr and DTRT. [02:29] i feel you [02:29] * scientes wishes it was all git [02:29] * scientes uses the git mirror of mozilla-central [02:30] I miss the fix-it-yourself appeal of RCS/CVS. [02:30] Of course, they also sucked. [02:30] hahahaha [02:30] A lot. [02:30] But, still. If I wedge a repository for any of the new shiny VCSs, it's over. [02:31] infinity: hey hey I still use RCS - for evil [02:31] http://paste.debian.net/163469/ [02:32] gpg + rcs = version-controlled, encrypted netrc files [02:33] http://paste.debian.net/163470/ is the manpage, since that sh code is pretty obtuse [02:37] why did ubuntu switch to statically linking libxul into firefox? [02:37] esp when they switched to shipping thunderbird by default [02:37] seems you could save a few MB of ram for nothing [02:37] You could, but the pain of keeping the world in sync was awful. [02:38] And when you're running firefox and thunderbird, "a few MB" is laughably nothing. [02:38] true.... [02:38] My Firefox is happily dining on 3G of RAM. [02:38] it use to crash at 2GB all the time [02:39] no that that is fixed there is no end in sight [02:41] * twb sighs [02:42] I know RAM is cheap, but "it's cheaper to add ram than code an O(logn) algorithm instead of an O(n) one" is like fingernails on a blackboard to me [02:42] well, xml is neither fast nor ram-sensitive by its very nature [02:42] The problem XML solves is not hard, and XML does not solve it well. [02:43] xml only creates problems [02:43] yes, it is context-free which is nice [02:43] but so is json, which is much better [02:43] So are sexprs, a better solution to the problem that existing in *1959* [02:43] are there any 100% xml-compatible binary representations [02:44] that we could support in browsers? [02:44] Who cares [02:44] The browsing public doesn't give a shit about anything except stupid shiny web apps [02:44] it would be easier to parse [02:45] Personally I avoid 90% of the problem by using w3m for everything except the odd broken site [02:45] thats pretty extreme [02:45] i mean even noscript and adblockplus gets you pretty far [02:45] adblockplus is a clusterfuck; that code belongs in DNS [02:45] afk food [02:46] thats what people say, but adblockplus works much better IMHO [02:46] cause it actually removes the elements from the DOM [02:46] and they are never rendered at all [02:46] or requested [02:46] I wouldn't know about that, I was pulling its data files and putting them into squid and polipo [02:46] and it works better than simple host forging [02:46] I haven't actually used ff since about 1.5 [02:47] Since webkit has slightly less bloat [02:47] well adblockplus is ported to chromium [02:47] but i use firefox cause its the only browser than you can coaxe into slightly keeping your privacy [02:52] twb: you should give Firefox 13 a go, much smaller memory footprint [02:52] and much faster javascript [02:52] and also video support [03:01] are you serious? They're up to 13 now? [03:01] micahg: well I was running it over X so I didn't have to pay the footprint penalty on my local netbook [03:02] TBH the video stuff annoys me because now I can't opt out by not installing flash [03:02] noscript should help there [03:02] If I want a video I'll fucking well fun mplayer, I don't want it loading without asking inside a browser, any more than I want geocities "background music" in a web page [03:02] micahg: so instead of taking stuff out, I have to add more stuff to stop the first stuff from doing anything? Yuk [03:03] 13 is current in the aurora channel which is ~7 weeks from release (1 week till beta migration + 6 week migration cycle) [03:03] * micahg thinks noscript is sensible in any event for browsing [03:04] IMO if you have js in a browser it should be off by default [03:04] It's absurd to me that you need a third-party module to make that the case [03:04] twb: if you saw how many bugs we get because people disable stuff, you'd understand why :) (Mozilla probably gets more) [03:04] But then, that's why I use w3m and I'm not an arsebook weenie [03:04] twb, you can in firefox settings [03:05] micahg: yeah I grant you moco has to deal with a lot of fuckwits [03:05] noscript just lets you set it on a per-domain basis [03:05] whic is more useful [03:05] same with cslite [03:05] * micahg loves permablocking the ad sites [03:05] scientes: I think if I was building it, I'd just have a keychord that said "toggle js in this tab" [03:06] micahg: I do that in hosts(5) ;-) [03:06] twb, yeah i was looking for a way to only have ajvascript run in current [03:06] but firefox actually does that by default now [03:08] infinity: btw with both oneiric and precise enabled, apt-get whinges because it can't find oneiric/armhf [03:08] http://paste.debian.net/163476/ [03:09] twb: Yeah, cosmetic bug worth fixing at some point, but it's not exactly a common use-case. [03:09] twb, infinity already said, oneric doesnbt have armhf [03:09] (ie: how often do we add ports, and have people attempt cross-series multiarch?) [03:10] infinity: okey dokey [03:11] make[1]: *** [stamp-conf-lite] Error 1 make[1]: Leaving directory `/home/ty/LuanVan/icewm/icewm-1.3.7~pre2' make: *** [build-stamp] Error 2 dpkg-buildpackage: error: debian/rules build gave error exit status 2 [03:11] I am ninja enough to ignore an error from apt-get ::P [03:11] i can't fix this err [03:11] pnphi: pastebin full transcript please [03:11] make[1]: *** [stamp-conf-lite] Error 1 [03:11] make: *** [build-stamp] Error 2 [03:11] dpkg-buildpackage: error: debian/rules build gave error exit status 2 [03:11] 2 days , i can't fix this err [03:12] ubot2: pastebin [03:12] For posting multi-line texts into the channel, please use http://paste.ubuntu.com | To post !screenshots use http://imagebin.org/?page=add | !pastebinit to paste directly from command line | Make sure you give us the URL for your paste - see also the channel topic. [03:12] pnphi: put the whole transcript, all those 100s of lines, into a pastebin [03:12] ok [03:13] http://paste.ubuntu.com/933470/ [03:14] pnphi: what is the version (first line) reported by iconv --version? [03:15] must install iconv ? [03:15] If iconv is not installed, that is probably the problem. [03:15] http://paste.ubuntu.com/933473/ [03:15] iconv --version [03:17] have iconv [03:20] There is something wrong with your system, because an earlier version of iconv (2.11, from lucid) passes that test: [03:20] echo foo | iconv --from iso-8859-1 --to UCS4//translit ==> foo [03:21] yes [03:21] pnphi: Are you developing on maverick? [03:21] yes [03:21] You realise it's EOL, right? [03:21] yse [03:21] yes [03:22] Of course, maverick's iconv *does* work, so I'm sure your problem is local. But still. You might want to use a newer release. [03:23] newer...ok [03:23] yeah iconv works... [03:23] must upgrade [03:25] i'm very tried, 2 days with the err [03:26] build by "dpkg-buildpackage -a" and "xdeb" , the same err [03:31] i must upgrade into 11.04 [03:34] 11.04 is natty [03:44] how using dpkg-cross [04:13] infinity: wooot, jockey worked fine from first boot :-) [04:14] rsalveti: Crazy talk. You must have done it wrong. [04:14] ;-) [04:14] Also: \o/ [04:14] I mean, it would be more "\o/" if the drivers were just free, but whatever. :/ [04:14] yeah [04:15] hopefully some crazy dudes will reverse engineer it soon ;-) [04:15] I'm still waiting for some clever young kid with too much free time to make nouvea work on Tegra. [04:15] Or you know, the vendor gets a fucking clue [04:15] would be awesome if robclark would be allowed to do that [04:15] rsalveti: That might represent a SLIGHT conflict of interest for him. [04:16] robclark: did you look at the tegra support before doing the work for the other driver? [04:17] - ARM Fast Models as builders [04:18] rsalveti: ^-- You hate me, right? [04:18] infinity: oh yeah :-) [04:18] start crying ;-) [04:18] On it. [04:19] On the one hand, it would be nice to have the port ready before there's any GA hardware. On the other hand, OH GOD NO. [04:19] lol [04:20] we're using it for big.little, and yeah, it's a pain [04:20] rsalveti, I didn't really look at tegra.. although that was more to do w/ hw I have access to.. :-P [04:20] infinity: ^ just need to get a tegra board to robclark ;-) [04:20] robclark: I'll mail you some Tegra kit. [04:20] heheh, well, I think I'll be busy for a while ;-) [04:20] Actually, in all seriousness, we might have some spare ac100s floating around the company, I could ask around. [04:20] although find some nouveau folks, send 'em tegras :-P [04:21] I just have a gut feeling that making nouveau do tegraish things would be dangerously close to "trivial" for someone who understands the architecture. [04:21] Since the Tegra GPU is, by all account, just a slightly crippled Geforce. [04:22] Sadly, once it's made to work "just like the desktop GPUs", then someone needs to make it do GLES instead. :/ [04:22] I'd heard somewhere that some nouveau folks thought it would need a different driver, but I'm not sure how much that has to do w/ command stream and shader ISA's, vs UMA vs vram.. [04:23] I'd have to guess there is at least some family resemblance ;-) [04:23] It's the nose. It's always the nose. [04:23] heheh [04:24] Honestly, we suffer along fine with ati/nvidia binary drivers on the desktop, and I think I've mostly come to terms with it. [04:24] It's the part where ARM SoC vendors are a bit less organised that makes it more painful. [04:24] I don't suffer [04:24] * robclark used nouveau on old laptop (now have intel gfx) [04:25] (So, the omap4 + jockey thing is great news) [04:25] I outright refuse to use anything but intel GPUs on the desktop [04:25] nouveau supported gles nicely thx to gallium [04:25] (gles also works well on new laptop) [04:25] twb: If Intel ever caught up in 3D performance, I may take that stand, but I still like my fancy video games. [04:26] I have never once had a proprietary video game work properly under wine [04:26] (I hear that this year's crop of Intel GPUs are at least "not crap", compared to one generation removed for ati/nvidia) [04:26] Even with non-intel (ATI) GPUs [04:26] twb: I used to play WoW in wine, worked fine. [04:27] On both ATI and nvidia. [04:27] cube and tremulous worked OK tho [04:27] infinity: the two I can remember are KOTOR and SMAC [04:38] infinity: and after reboot, unity 3d running :-) [04:38] janimo`: with today's image unity-3d is working after installing the pvr driver [04:39] janimo`: what issue did you get exactly? [04:39] can you also start just X and call unity by hand to see if you get any useful logs? [04:49] rsalveti: Huzzah! [04:49] I really need to test all of this on my Panda and see it in action. [04:49] The poor thing's been on fire with toolchain builds all week instead of playing with shiny things. [04:50] But, once this gcc-4.6/eglibc upload for Debian is done building, it's free again. Finally. [04:50] hm, seems a bunch of gnome-related changes just landed, running dist-upgrade now [04:50] yeah [04:50] twb, i've had games work [04:50] infinity, i couldn't get a framerate over about 3 with ati [04:51] i also got a nasty corruption issue (which i managed to fix) [04:51] scientes: Ow. [04:51] nvidia +wow always worked fine however [04:58] I know it seems to be JUST me [04:58] everyone else seems to have wine work fine for everything [04:58] Although I suspect that's mainly nvidia customers, not the ATI r250 I was using at the time [05:00] nvidia certainly is the strongest 3d driver for linux [05:00] but nouveau and radeon are certainly coming along [05:01] ..however nvidia just broke for me is some way on precise... [05:01] since last few kernels === doko_ is now known as doko [07:40] bug 983555 [07:40] Launchpad bug 983555 in compiz-plugins-extra "compiz-plugins-extra FTBFS on ARM" [Medium,Fix released] https://launchpad.net/bugs/983555 [08:14] has anyone tried running the ubuntu Oneric 11.10 rootfs over android kernel 3.01 in ICS? [08:29] nimesh_accenture, there is no such this as "android kernel 3.01" [08:29] AFAIK [08:29] each device has a differn't kernel, which is part of the way ARM currently works [08:32] or doesn't work, as the case may be :P [08:35] tsk, tsk, CONFIG_INIT_PASS_ALL_PARAMS is still not merged mainline? [08:37] scientes: Android has its own patches over the standard linux kernel , which had not recently been upmerged , until 3.3 [08:37] nimesh_accenture, indeed, however most embedded devices use more than just android patches [08:38] nimesh_accenture, and as 3.3 doesn't have wakelocks booting android on it will eat your battery alive [08:38] only a select few boards work with upstream kernels, even with non-android [08:39] nimesh_accenture: are they all/mostly in, in 3.3. then? [08:39] Last I heard was "RSN, honest" [08:40] twb: i'm not sure , but I read an article a few weeks back that android patches have been ypmerged into the std linux kernel , not sure if all had been merged [08:41] shiiiiiiny [08:43] scientes: what i'm basically trying to do is create a new rootfs using chrooot in android ICS and then include ubyntu packages and the glibc and the std libc in that root to run applications like g-streamer [08:43] nimesh_accenture, that should work just fine, however what does uname -a say [08:43] nimesh_accenture, also, see if you have /proc/config.gz [08:44] also, bind mount /dev [08:44] i'm currently on my desktop [08:44] panda hasn't been booted yet [08:45] nimesh_accenture, https://play.google.com/store/apps/details?id=com.galoula.LinuxInstall&hl=en [08:45] thats debian [08:45] oh if you have a pandaboard just use the ubuntu pandaboard images [08:46] should have said you have a more open device [08:48] ya but the ubuntu panda board images contain the kernel et. all , i'll have to separate that out from the image and create a new rootfs [08:49] i mean i've not done it before [08:49] that is waht i assume [08:51] nimesh_accenture, are you trying to dual boot? [08:51] whats wrong with the ubuntu kernel? [08:52] rsalveti, wat I tested was an x86 wit pvr not panda, so likely oter issues [08:52] * janimo` needs to fix te h key [08:52] nothing is wrong. I want the Android front end , but the ubuntu back end for apps like g-streamer , which dont run on android root [08:54] so both will coexist. not dual boot. [08:54] gstreamer isn't an app [08:54] an engine i ment [08:54] I wonder if anyone has filed an ITP for dalvik [08:55] That would be one of the obvious prerequisites [08:55] not bothered much abt ITP's at the moment ... [10:34] guys ... I have a question about rootstock.... does rootstock download packages from the web or does it include it from the local machine? [10:46] nimesh_accenture, from the archives. not the local machine [10:47] so what do I do if i want to crreate a root fs from my local machine root files? [10:48] i basically have the server image of oneric and i want all the packages that are included in that package [10:52] nimesh_accenture, I don't think you can reuse your serer image iso from rootstock [10:52] i am having package problems with enabling my pandaboard, in 11.10 and 12.04 [10:53] ok... [10:53] one more thing , what does "--seed linux-image-omap" actually do? [10:53] and do we actually need to give that option? [10:54] what is the best sdcard image to start with and is there a version of omap extras to use [10:54] nimesh_accenture, I think with --seed you specify extra packages to be added to the image which otherwise are not in the default rootstock image [10:54] kernel must be seeded as that is specific to the platform you are building for [10:55] if you add --seed ubuntu-server it may add all dependencies of that metapackage too. I am not sure though I rarely used rootstock and een then I did not understand exactly how it worked [10:55] isn't rootstock largely outdated? there was even talk about removing it. [10:56] i dont need the kernel, as i'm trying to chroot on top of Android kernel [10:56] LetoThe2nd, yes it is. I don't know if there's a similarly easy way to build custom images though [10:56] live-build is less straightforward [10:57]