**** BEGIN LOGGING AT Thu Feb 16 02:59:57 2012 Feb 16 04:58:09 looks like upstream Chromium almost fixed the build on arm :) Feb 16 05:07:33 micahg: I and markos attached patches that fix chromium on armel/armhf to the debian bug Feb 16 05:08:35 suihkulokki: now I find out, I just uploaded a new version :) Feb 16 05:08:51 ok, I"ll grab them for the next updat Feb 16 06:03:43 suihkulokki: I realized that I was only checking for armel in debian/rules, I fixed that for the next upload Feb 16 10:27:59 hey kalikiana, good to see you around :) Feb 16 10:31:59 heyhey Feb 16 10:32:27 I checked out #arm before, seems the cool kids are all here :-P Feb 16 10:32:55 well, here is where the community is ... way more important than the other channel ;) Feb 16 10:56:29 hey micahg, whats that patch you talked about last night ? i think kalikiana would like to know about it Feb 16 10:56:43 ogra_: patch, what patch? Feb 16 10:56:50 chromium Feb 16 10:56:58 micahg: oh, you're there, was just typing a mail Feb 16 10:57:22 I think it might work now without any patches Feb 16 10:57:36 I'm waiting on the first armel build from yesterday Feb 16 10:57:48 micahg: so who builds it? where's the package source? Feb 16 10:57:56 * ogra_ only saw that riku said he attached the patches to a bug ... but didnt mention the bug number Feb 16 10:58:10 kalikiana, in the ubuntu archive i guess Feb 16 10:58:17 (precise sources) Feb 16 10:58:17 I couldn't find working steps to get a cross-compile setup with all packages Feb 16 10:58:20 kalikiana: chromium-browser? the same way anything else in the archive is built Feb 16 10:58:56 micahg: well, but a human must have tested it :-) Feb 16 10:59:03 kalikiana: I think Debian must have upstream their patches Feb 16 10:59:15 meh, so yet another possible team to poke... Feb 16 10:59:23 why ? Feb 16 10:59:30 they will just end up in precise Feb 16 10:59:43 I also found a bug in the packaging, which I'll fix for the next upload to get armhf Feb 16 10:59:44 just pull the precise source package for your project Feb 16 10:59:56 we don't merge from Debian for Chromium Feb 16 11:00:00 micahg, that would be helpful, given we will make the switch today :) Feb 16 11:00:16 micahg: so do you have a working build setup? Feb 16 11:00:23 I'm running into broken deps Feb 16 11:00:24 (armel will become unsupported from today on) Feb 16 11:00:35 kalikiana: no, these are archive builds, I upload 2 recent stable releases to precise Feb 16 11:01:09 hmm Feb 16 11:01:26 I really need to build it locally Feb 16 11:01:28 ogra_: I figured another chromium upload will come in a week or 2, if you need it sooner and armel works, I can upload with the armhf fix Feb 16 11:01:49 yes please, we will stop caring for armel from today on Feb 16 11:02:10 we wont stop building it, but nobody will look at issues Feb 16 11:02:15 ok, I'll keep an eye on it Feb 16 11:02:18 micahg: can you recommend a person who cross-compiles usually? Feb 16 11:02:31 kalikiana: only one I know is suihkulokki :) Feb 16 11:02:36 ogra_: "nobody" is a bit much, I always watch build issues on all ports. Feb 16 11:02:44 ogra_: I just prioritize. :P Feb 16 11:02:57 infinity, hey, you are not supposed to be up at that time ! Feb 16 11:03:07 * micahg does that too :) Feb 16 11:03:10 ogra_: I'm not, it's all in your head. Feb 16 11:03:15 * micahg is also not supposed to be up now Feb 16 11:03:19 micahg: k, I'll wait for him then Feb 16 11:03:22 * ogra_ was just trying to secretly spread FUD while infinity wasnt watching ... Feb 16 11:03:25 damned ! Feb 16 11:03:26 (or her) Feb 16 11:04:01 I'll be sure to call Riku a her the next time I see him. ;) Feb 16 11:04:06 kalikiana: did you try the instructions on the linaro wiki? Feb 16 11:04:07 lol Feb 16 11:04:40 micahg: yes. but gconf2 and bzip2 at least are broken Feb 16 11:05:06 file bugs? Feb 16 11:06:44 * infinity wonders why everyone gets so excited about cross-compiling. Feb 16 11:06:56 Especially to the point of wasting days/weeks trying to make it all work right. Feb 16 11:07:04 infinity: because it is several times faster Feb 16 11:07:08 When you could just build natively (sure, more slowly), but be done with it. Feb 16 11:07:16 kalikiana: Not when you factor in all the annoyance. :P Feb 16 11:08:00 well, the setup is actually simple. the problem is really that arm is relatively recent Feb 16 11:08:02 yeah, its such a waste Feb 16 11:08:12 "recent"? Feb 16 11:08:16 heh Feb 16 11:08:22 we even wronte instructions to cross-compile chromium: https://wiki.linaro.org/Platform/DevPlatform/CrossCompile/ChromiumCrossCompile Feb 16 11:08:26 infinity: as in packaging :-) Feb 16 11:08:35 yes Feb 16 11:08:39 and those don't do it Feb 16 11:08:41 we have it since jaunty ... Feb 16 11:08:43 hey suihkulokki Feb 16 11:08:56 and debian has it since ... hmm, forever Feb 16 11:09:01 suihkulokki: gconf2 and bzip2 are broken Feb 16 11:09:08 bug# ? Feb 16 11:09:36 suihkulokki: the packages don't install Feb 16 11:09:50 infinity, do you know if anyone talked to kate about armhf ? we should make her announce it in the FF announcement i think Feb 16 11:10:28 suihkulokki: is it expected to work as described there, as in used by somebody on a regular basis? Feb 16 11:10:34 ogra_: I brought it up when I spoke to her on the phone on Monday, but we didn't discuss a formal announcement. Feb 16 11:10:48 kalikiana: yes if you don't deviate from the steps Feb 16 11:10:53 ok, i'll ping here if she is up to make sure she doesnt forget Feb 16 11:11:14 ogra_: Also, I think we need to be careful how we word such an announcement. Feb 16 11:11:32 ogra_: People seem gung-ho to say things like "we only support armhf now!" which is patently untrue. Feb 16 11:11:45 ogra_: It's just that we've selected armhf as the LTS arch. Feb 16 11:11:51 k Feb 16 11:12:01 lets make sure that goes into the announcement then Feb 16 11:12:14 suihkulokki: I don't. so is there somebody who I could poke about those packages? I'd think bugs might not exactly receive attantion very quickly Feb 16 11:12:16 though i suspect we will stop building armel images for some flavours Feb 16 11:12:26 at least until your LP changes are ready Feb 16 11:12:28 ogra_: (From a Canonical standpoint, yes, we only support armhf now, but Canonical != Ubuntu, and we need to be clear about that, or just steer clear of such statements, I prefer the latter) Feb 16 11:12:51 kalikiana: pastebin your errors please (as well as what is in your sources.list) Feb 16 11:12:54 ogra_: I intend to drop armel images for !omap tomorrow. Feb 16 11:13:11 ogra_: But installer/image support != archive support. Again, distinction is key. :) Feb 16 11:13:16 if we can it would be nice to keep el and hf for ac100 Feb 16 11:13:26 simply for the binary driver Feb 16 11:13:43 ogra_: Yeah, that's fair. It's also a good platform for us to keep testing armel on as a community port. Feb 16 11:13:53 ogra_: Alright, I'll keep omap and ac100, and drop mx5 and omap4. Feb 16 11:14:18 ++ Feb 16 11:14:20 (The reason I'm keeping omap is simply because it's the only kernel that builds from mainline, so it's "free" from a maintenance perspective) Feb 16 11:14:44 yeaqh Feb 16 11:15:09 Plus, we might free up some beagles from the DC soon, and pass them around to people, thus making omap useful. ;) Feb 16 11:17:19 ogra_: Oh, we did get some indication at Connect that nvidia is actually planning armhf builds of their binary driver. Feb 16 11:17:50 ogra_: I was told they've already done builds in-house, they're just working on the tarball drop. Feb 16 11:17:57 yeah, but they dont exist yet and we have no promises they will be there by release Feb 16 11:17:59 So, we'll see. Feb 16 11:18:30 i'm carefule about them if it comes to timelines (unlike ndec whom i belive blindly if he says it will come) Feb 16 11:18:39 Heh. Feb 16 11:18:45 Well, this was engineers, not legal. Feb 16 11:18:48 suihkulokki: http://pastebin.ubuntu.com/844219/ Feb 16 11:18:50 Which is a double-edged sword, I guess. Feb 16 11:19:01 The engineers are doing the real work, and they know what they've done. Feb 16 11:19:10 Then again, they don't get to do the final release to the website. Feb 16 11:19:19 yep Feb 16 11:19:47 fi engeneering had a say, we would have everything in since a year or two, i know Feb 16 11:21:21 * micahg still would like an mx51 kernel for precise, but fears that's not happening Feb 16 11:22:00 micahg: For babbage, or efika? Feb 16 11:22:03 micahg, complain to linaro, we donmt maintain that arch :) Feb 16 11:22:11 efika Feb 16 11:22:14 micahg: For efika, talk to markos. Feb 16 11:22:21 micahg: He promised me a kernel. Feb 16 11:22:34 for what ? netbook or nettop ? Feb 16 11:22:39 micahg: I even promised him unofficial image enablement if he got me a working kernel. :P Feb 16 11:22:46 ogra_: Yes. Feb 16 11:22:57 there was an or in the question Feb 16 11:23:00 ogra_: (He seemed to think he could build one that would work on both) Feb 16 11:23:06 ogra_: I have both, would prefer the smartbook Feb 16 11:23:09 unless they very recently started having a unified kernel Feb 16 11:23:24 * ogra_ too, i have one here as well Feb 16 11:23:28 although, I guess I'd prefer an armhf kernel for mx51 :) Feb 16 11:23:33 * infinity has a smarttop. Feb 16 11:23:43 micahg: Kernels are FP agnostic. Feb 16 11:23:52 oh, ok, cool Feb 16 11:24:05 packages arent indeed :) Feb 16 11:24:08 micahg: But seriously, nag markos. If he can get me something that's sane, I'll get it into universe. Feb 16 11:24:43 infinity: I'm happy to be the second reviewer on that ;) Feb 16 11:24:50 micahg: I wouldn't mind a vaguely current kernel on my efika either. :P Feb 16 11:25:09 kalikiana: argh. looks like I have uploaded a precise targetted gconf to the ppa's oneiric side Feb 16 11:25:14 (Though it is the slowest ARM device in my house) Feb 16 11:25:40 you dont have beacgle A/B versions ? Feb 16 11:25:48 kalikiana: tho the bzip2 should install just fine Feb 16 11:25:56 * ogra_ has a beagle A1, thats definitely the slowest Feb 16 11:26:05 suihkulokki: is there something we should fix in precise for cross compiling? Feb 16 11:26:10 ogra_: I have no beagles. Feb 16 11:26:15 ah Feb 16 11:26:31 ogra_: My N900 used to be the slowest ARM device on my desk, but I sold it to Sledge. Feb 16 11:26:54 yeah, and thats already twice as fast as an A1 Feb 16 11:27:07 ogra_: So, now I have the efikaMX, an i.MX53, a Panda, an ac100, and an LGP999 (Tegra2 phone). Feb 16 11:27:18 suihkulokki: ah. yes, sorry, bzip seems okay at this point. I tried a different ppa before, but made sure to clean up before I pasted Feb 16 11:27:19 and has enormous amounts of ram compared... Feb 16 11:27:25 It's a bit sad that the phone is actually the fastest of the bunch. Feb 16 11:27:48 suihkulokki: any chance you could re-upload it correctly? anything I can bribe you with? :-P Feb 16 11:27:49 stop using it as a phone then and send it to lamont as a buildd :) Feb 16 11:28:03 I used it as a buildd during the armhf bootstrap. :P Feb 16 11:28:08 heh Feb 16 11:28:32 we should make tobin build a cluster of them in a rack case :) Feb 16 11:31:37 kalikiana: testing.. will ping back in a hour or so Feb 16 11:32:20 micahg: gconf2 is probably the most annoying thing missing multiarch conversion in precise. Feb 16 11:33:02 micahg: So, I harassed markos, and he said they're (A) working on 3.2 for efika/mx51, and (B) it will be a unified smarttop/smartbook kernel. Feb 16 11:33:23 micahg: And he's going to try to get me half-functional bits ASAP so I can work with them in my copious free time. Feb 16 11:33:39 suihkulokki: we might be able to do something about that, it's still got a lot of reverse dependencies though Feb 16 11:34:04 infinity: that's awesome, let me know if I can help with my copious amounts of free time Feb 16 11:34:15 can't we just make chromium not use gconf? =) Feb 16 11:35:15 not for another year :) Feb 16 11:35:53 * micahg wonders if they have an option to s/gconf/gsettings/ Feb 16 11:37:42 suihkulokki: wrt bzip2, I get this http://pastebin.ubuntu.com/844238/ Feb 16 11:39:01 removing gconf from chromium shouldn't be that hard, but more time probably than fixing gconf2 packages :-P Feb 16 11:39:32 well, it eventually needs to move to gsettings anyway Feb 16 11:39:32 * micahg has plenty of other archive cleanup work to do Feb 16 11:39:40 kalikiana / suihkulokki: That looks like the bzip2 in the linaro PPA was built with a broken gzip. We've fixed that bug in precise... Though, if this is an oneiric PPA, it's a crap shoot if it'll break or not. :/ Feb 16 11:40:39 yes, it's oneriric I need to build for Feb 16 11:40:47 kalikiana: You can work around it by just doing "dpkg --force-overwrite -i /var/cache/apt/archives/libbz2-1.0_1.0.5-6ubuntu2linaro1_armel.deb" Feb 16 11:42:49 I'll throw backport of gzip to the ppa Feb 16 11:43:26 infinity: same error Feb 16 11:44:45 kalikiana: Oh, right. "rm /usr/share/doc/libbz2-1.0/changelog.Debian.gz && dpkg -i /var/cache/apt/archives/libbz2-1.0_1.0.5-6ubuntu2linaro1_armel.deb" Feb 16 11:44:56 kalikiana: That'll make it happy. :P Feb 16 11:45:06 evil :-D let's try Feb 16 11:45:35 you won't believe, same error Feb 16 11:46:09 Err, that's not possible. Feb 16 11:46:11 Or, shouldn't be. Feb 16 11:46:36 * ogra_ wonders how that cross build setup is set up ... might be caused by that Feb 16 11:46:53 ogra_: No, it's just straight up multiarch and gzip sadness, I'm sure. Feb 16 11:47:01 weird Feb 16 11:47:16 But deleting the target before installing the other arch's package should work... Feb 16 11:47:29 kalikiana: It's not complaining about another file now, perhaps? Feb 16 11:47:32 kalikiana, seriously, how many days did you spend on getting that working now ? Feb 16 11:47:37 presumably "different" here includes that it exists Feb 16 11:48:27 ogra_: I had a different setup before that, native arm, wither other problems. although I could get it to sort of work, it wasn't exactly working well Feb 16 11:48:50 it took in the hour range to test anything Feb 16 11:50:16 well, the build only takes 12h Feb 16 11:50:30 at least the lsat successfull one in the archive did Feb 16 11:50:52 given that you alkready spent several days to get your cross setup working ... Feb 16 11:51:40 it was not 12h, more like 2h, but still, that's no basis for working on code Feb 16 11:52:19 2h ? for a plain make you mean, surely not for the package build (natively) Feb 16 11:52:24 * infinity decides to refrain from "when I was your age" stories. Feb 16 11:52:30 haha Feb 16 11:53:07 ogra_: when I say native I mean chroot. which I guess is faster, but still a pain Feb 16 11:53:07 Besides, if you're "working on code", you don't rebuild from scratch every time, you have a built tree, and you let the magic of make only re-compile and re-link the bits you changed. Feb 16 11:53:23 right Feb 16 11:53:34 and that should just take minutes Feb 16 11:53:42 unless you make clean every build Feb 16 11:53:51 infinity: unfortunately it includes changing the build system Feb 16 11:54:22 Fun. Feb 16 11:54:32 yes. plenty. Feb 16 11:54:52 Speaking of build systems. Feb 16 11:55:03 I'm grumpy that I need to teach fpc about armhf after all. :/ Feb 16 11:55:14 give it to NCommander :P Feb 16 11:55:23 he did it several times before Feb 16 11:55:25 Because, despite not being VFP-aware, it DOES hardcode the linker path when linking with C code. Feb 16 11:55:56 and doesnt have libO anymore on his plate .... he needs a new challenge Feb 16 11:57:09 suihkulokki: Say, about LibreOffice... Feb 16 11:58:56 infinity: looks like janimo is ahead of me.. Feb 16 11:59:15 Oh, I didn't realise Jani was working on it too. Feb 16 11:59:16 Hrm. Feb 16 11:59:22 Duplication of effort, for the loss. Feb 16 12:00:17 ogra_: I could always tell people to use ebuilds when they want to cross-compile ;) Feb 16 12:00:34 infinity: any idea how to bootstrap gnat-4.6? Feb 16 12:00:54 micahg: Cross-compile from another arch, native compile and check for correctness. Feb 16 12:01:04 micahg: markos was working on it, and seemed to be stuck on the latter step. Feb 16 12:01:39 infinity: right, but how to do for the archive, same? Feb 16 12:01:49 suihkulokki, lol Feb 16 12:02:39 micahg: Cross, native, plunk native in my stage-2 repo of doom, build on buildds. Feb 16 12:03:09 infinity: ah, cool, ok, that's what I thought which is why I can't do it :) Feb 16 12:03:32 micahg: If you can get the first two steps done, I won't mind. :P Feb 16 12:03:37 \o/ libreoffice passes the bridge tests on armhf, on to the rest of the build Feb 16 12:03:49 micahg: If you give me a self-hosting gnat, I'll happily turn it into something archivey. Feb 16 12:03:58 janimo`: ! Feb 16 12:04:13 infinity, ! Feb 16 12:04:29 * janimo` does not let himself be intimidated by excalmation marks Feb 16 12:04:45 janimo`, !!!! Feb 16 12:04:47 ! Feb 16 12:04:49 Hahaha. Feb 16 12:04:52 * janimo` collapses Feb 16 12:04:58 janimo`: !o! Feb 16 12:05:06 janimo`: That's awesome news. And I suspect awesome news for suihkulokki, who now doesn't have to care. :P Feb 16 12:05:39 infinity, yep. I just need top make sure it works on OABI - the code still has that support although I am not sure who still cares Feb 16 12:06:12 xranby, is that the 'happy drummer' sign? Feb 16 12:06:34 janimo`: Pardon my French, but fuck OABI. Feb 16 12:06:48 janimo`: If anyone's building Open/LibreOffice for OABI, they should be shot. Feb 16 12:06:54 infinity, my thoughts exactly only I express them more delicately Feb 16 12:07:38 I was not sure how much Debian cares about baxckporting TBH, I am not sure either how much of Debian users are on legacy hw/installs Feb 16 12:07:52 janimo`: More to the point, while an upstreamable fix might need to keep OABI working (though, god knows why), a Debian/Ubuntu patch clearly doesn't. Feb 16 12:08:05 I plan on pushing upstream today Feb 16 12:08:13 and deal with bug reports if OABI people come along Feb 16 12:08:15 janimo`: Debian doesn't support OABI. Feb 16 12:08:18 OABI is forgotten in debian. however we should check it doesn't break EABI softfloat Feb 16 12:08:28 I'll put ifdef around FP regs usage and that should be all I guess Feb 16 12:08:36 suihkulokki: EABI soft, or softfp? Feb 16 12:08:40 suihkulokki, test passes on armel as well Feb 16 12:08:46 but only hardfloat tested Feb 16 12:08:51 (God, that's confusing) Feb 16 12:09:03 janimo`: Our armel is softfp, which is still VFP-aware. Feb 16 12:09:05 I have no soft (no FPREGS usage at all) setup Feb 16 12:09:08 janimo`: Debian's armel isn't. Feb 16 12:09:29 And somewhere along the way, my head explodes. Feb 16 12:09:47 aha, that indeed needs to be checked for, I'll make use of the SOFTFP define then and hope it is not too much work Feb 16 12:11:25 infinity, suihkulokki can I test a Debian/softfloat setup in an ubuntu hosted chroot ? just create a sid-armel one? Feb 16 12:11:39 janimo`: yes Feb 16 12:11:39 schroot I mean if it matters Feb 16 12:11:46 oh great then Feb 16 12:12:54 janimo`: Yeah, kernels are completely VFP agnostic, so debian-armel chroots work fine. Feb 16 12:12:58 (And armhf, and, and) Feb 16 12:13:07 excellent Feb 16 12:14:11 janimo`: didn't you ifdef the hardfb abi with__ARM_PCS_VFP already? that should not be set on either debian or ubuntu armel Feb 16 12:14:23 suihkulokki, yes that one I used Feb 16 12:14:35 but I noticed there's another SOFTFP variable too Feb 16 12:14:47 so it may all just work indeed Feb 16 12:14:50 Which may or may not relate to softfp. Feb 16 12:14:56 hardfb ? hard floating bubble ? Feb 16 12:15:11 Because the soft/softfloat/softfp naming between toolchain and userspace applications is never consistent. :P Feb 16 12:15:35 (Heck, it gets even worse when Oracle refers to softfp as "hard-float") Feb 16 12:15:49 heh yes. Feb 16 12:15:51 haha Feb 16 12:15:57 And they're not wrong. It is hard float. Just not what WE call hard float. Feb 16 12:16:09 its kind of funny oracle compared their tuned armel builds against icedte built using debian squeeze Feb 16 12:16:32 could have been worse Feb 16 12:16:40 they could have built against woddy Feb 16 12:17:00 We didn't have armel in woody, did we? Feb 16 12:17:07 nah Feb 16 12:17:11 woody would have been OABI, I believe. Feb 16 12:17:13 only arm i think, if at all Feb 16 12:17:17 But it's been a while. Feb 16 12:18:56 i think we are doing good, at least our work runs on the armhf ABI Feb 16 12:19:07 I dunno. I gnored all you ARM weirdos back then. I was working on promising new future technologies like m68k and parisc. Feb 16 12:19:35 mips ? Feb 16 12:19:44 I wasn't a mips porter, no. :) Feb 16 12:20:07 68k, parisc, powerpc, and alpha. I sure can pick 'em. Feb 16 12:20:14 1 out of 4 isn't bad, right? :P Feb 16 12:21:06 mmm, alpha ... i loved them Feb 16 12:21:10 * janimo` completely forgot how he created the precise-armhf schroot, now needs to figure out same for debian.sigh Feb 16 12:21:21 I do not recall running debootstrap explicitly Feb 16 12:21:38 and my bash history agrees Feb 16 12:21:56 it may have been mk-sbuild Feb 16 12:22:28 janimo`: debootstrap --variant=buildd --arch=armel sid sid-armel http://ftp.debian.org/debian Feb 16 12:22:32 janimo`: Ish? Feb 16 12:22:59 yes, it was not that I guess, probably mk-sbuild did it all behinf the curtains for me Feb 16 12:23:18 do debootstrap and schroot have a common location for chroots in which case I may just do that above Feb 16 12:23:59 mk-sbuild --eatmydata --arch=armel precise was it Feb 16 12:25:55 * janimo` remembers seeing the one and only alpha machine sometime in 94 at an expo. It was higher Mhz that the 486's (150 I think) and it had a video promo of Jean Luc Picard extolling its virtues Feb 16 12:26:24 probably they though it was meant for enterprise servers Feb 16 12:26:26 janimo`: debootstrap has no canonical location for anything, it just creates the chroot where you tell it to. Feb 16 12:27:08 infinity, ok I just started a mk-sbuild so it is stored along with the rest of schroots Feb 16 12:28:25 perhaps I can do the debian/armel build and testing Feb 16 12:29:59 suihkulokki, sure thanks. I am setting up a sid chroot now but my panda would be happier if left to build only a precise armhf and not something else in parallel :) Feb 16 12:32:42 janimo`: ok, on it Feb 16 12:36:02 suihkulokki, softfloat in debian mean no actual VFP instructions in asm right? Not that they are trapped and emulated by the kernel Feb 16 12:41:00 suihkulokki, you can stop the build after a couple of hours (not sure when) once the deps for bridges/ and testtools/ are built. Feb 16 12:41:15 then make bridges.deliver testtools alternating with changes to code is easy iteratin Feb 16 12:42:06 janimo`: do you have a line to grep in the buildlog when bridges have compiled? Feb 16 13:04:00 suihkulokki, hmm, not sure I know where the bridges' module number is Feb 16 13:04:20 what I did was ctrl-c and see if make brdiges works or says it has files missing Feb 16 13:05:02 but gcc3_linux_arm is a dir within bridges that needs to be entered so maybe good for grepping Feb 16 13:11:44 last build of libo on armel in the buildd: 5 days 19 hours. Ouch Feb 16 13:11:59 janimo`: On a babbage, I assume? Feb 16 13:12:04 assuming a panda is used instead of babbage it still likely be 1-2 days Feb 16 13:12:11 infinity, yes some old board Feb 16 13:12:20 I think with pandas it was over a day too last cycle Feb 16 13:12:29 or whenever we had pandas built libo Feb 16 13:12:33 Yeah, a dayish sounds about right. Feb 16 13:12:33 by accident Feb 16 13:12:45 so I am not holding my breath today for it to complete Feb 16 15:06:06 .c Feb 16 15:48:02 infinity: actually, the test suite failure I was fixing in upstart was unrelated... the one on arm*/ppc is still there, and I'm not sure what to make of it Feb 16 15:52:34 slangasek: Oh, it just happens to be in the same test. But, you're right, different error. Feb 16 16:12:32 infinity: yes, the "same", 4kloc test ;) Feb 16 16:16:36 slangasek: Ugh. **** ENDING LOGGING AT Thu Feb 16 16:19:53 2012 **** BEGIN LOGGING AT Thu Feb 16 16:20:19 2012 **** ENDING LOGGING AT Thu Feb 16 16:24:35 2012 **** BEGIN LOGGING AT Thu Feb 16 16:26:17 2012 Feb 16 16:27:51 suihkulokki: did you upload updates for libbz2 or gconf2 yet? Feb 16 16:36:12 kalikiana: yes but bzip2 is still building Feb 16 16:56:53 this morning's dump of updates seems to have another alsa bug Feb 16 16:56:54 ALSA lib conf.c:1220:(parse_def) show is not a compound Feb 16 16:56:55 ALSA lib conf.c:1686:(snd_config_load1) _toplevel_:24:26:Unexpected char Feb 16 16:56:55 ALSA lib conf.c:3406:(config_file_open) /usr/share/alsa/pulse-alsa.conf may be old or corrupted: consider to remove or fix it Feb 16 16:57:24 show { Feb 16 16:57:24 @func refer Feb 16 16:57:24 name defaults.namehint.basic Feb 16 16:57:24 } Feb 16 17:34:00 infinity: ah, I understand jhunt has figured out the race in the test that was causing things to fail, so we should have a fixed package tomorrow-ish Feb 16 17:34:11 slangasek: \o/ Feb 16 18:17:09 ALLCON: I am attempting to setup ubuntu server 11.10 on a Pandaboard es with no luck. I attempted multiple times with the same pre-installed distro with different outcomes. They range to when the system unpacks and reboots, I either booting and then nothing, booting and then root access. With root access I try to adduser and get one of two options: The user is added but not in the sudoers list, or the user is not added and I Feb 16 18:18:25 I gave up for a while and tried installing 12.04 and had no issues at all. It went straight to the installation screen, gave me all my options, and then worked great. I just went back to 11.10 since it is supported Feb 16 18:18:34 Is there anyone who may know what is wrong and could help Feb 16 18:18:58 orion___: 11.10 doesn't have full support for the 4460 in u-boot or kernel. Feb 16 18:19:05 It can cause overheating. Feb 16 18:19:16 is there a better option to go with Feb 16 18:19:51 I am testing 12.04 daily. It is very stable, but since it is in development, constantly updating stuff. Feb 16 18:20:43 You should be fairly safe running Alpha 2, then upgrading during milestones (beta 1, 2, etc). Feb 16 18:20:49 yeah, this is going onto a robot that needs to be mostly stable. I will only need usb, ethernet, wireless (maybe bluetooth), and the ability to work with the gpio if needed Feb 16 18:21:24 Ah. Yea, stick with the 12.04 and just turn off auto-updating. Feb 16 18:21:50 How are the service packages with 12.04. For instance, I tried to download openjdk as a quick test with 12.04 and couldnt find it Feb 16 18:22:39 It is there. I use it daily to run a Jenkins slave. I use open-jdk-jre-headless, but I also just tried default-jdk. Feb 16 18:22:57 got it Feb 16 18:23:02 just must have missed it then Feb 16 18:23:13 default-jdk will give you a lot more packages, but only really usefull cor compiling java. Feb 16 18:23:23 /cor/for Feb 16 18:23:52 and would you say that 12.04 server is the fastest build for pandaboard es right now? Feb 16 18:24:40 Yes. Especially the armhf release. BTW, we will no longer be making armel images after today. Feb 16 18:25:34 orion___: i will voice for 12.04 armhf builds be wicked fast Feb 16 18:25:34 got it. I would rather hf anyways Feb 16 18:25:41 err confirm Feb 16 18:25:50 my english this morning is a little off Feb 16 18:26:12 thanks to both of you. I will go back to 12.04 then Feb 16 18:26:12 (though slap a hard drive on the thing.. running off the sd card is pure pain) Feb 16 18:26:23 yeah, that is another area I am noticing Feb 16 18:26:28 Linux panda 3.2.0-1406-omap4 #8-Ubuntu SMP PREEMPT Tue Feb 14 16:12:38 UTC 2012 armv7l armv7l armv7l GNU/Linux Feb 16 18:26:33 thinking of getting an ssd Feb 16 18:26:36 here is what i did Feb 16 18:26:51 LABEL=home /home ext4 defaults,noatime,errors=remount-ro 0 0 Feb 16 18:26:51 LABEL=usr /usr ext4 defaults,noatime,errors=remount-ro 0 0 Feb 16 18:26:52 LABEL=var /var ext2 defaults,noatime,errors=remount-ro 0 0 Feb 16 18:26:52 tmpfs /tmp tmpfs defaults,noexec,nosuid 0 0 Feb 16 18:26:52 tmpfs /var/log tmpfs defaults,noexec,nosuid 0 0 Feb 16 18:26:54 tmpfs /var/run tmpfs defaults,noexec,nosuid 0 0 Feb 16 18:26:56 tmpfs /var/lock tmpfs defaults,noexec,nosuid 0 0 Feb 16 18:27:14 though ill probably migrate /var/log to the hard drive Feb 16 18:27:23 i had it in a ramdisk before i added the hd Feb 16 18:27:33 got it Feb 16 18:28:05 GrueMaster: btw it looks like that sound patch made it in the recent kernel build \o/ Feb 16 18:28:09 thanks to both of you Feb 16 18:28:10 for your robot usage, SD should be fine. It really only sucks when you have a lot of io load. Feb 16 18:28:32 (by a lot he means more then 2 MB/s Feb 16 18:28:32 pbuckley: Yes, I was told today. Also fixed the ucm configs last week. Feb 16 18:28:33 ) Feb 16 18:28:37 the robot will be communicating with the google cloud and with other robots Feb 16 18:28:49 indeed.. there is a new alsa bug i mentioned earlier Feb 16 18:28:54 but it doesnt break functionality Feb 16 18:28:55 sounds cool. Feb 16 18:28:56 just throws a warning Feb 16 18:29:06 orion___: sweet.. graduate project? Feb 16 18:29:08 also, may be using opencv, so there is a bit of IO Feb 16 18:29:12 pbuckley: I'll look into it. Feb 16 18:29:21 pbuckley: yes Feb 16 18:29:29 reminds me of Feb 16 18:29:45 http://www.google.com/url?sa=t&rct=j&q=ryan%20walker%20robot&source=web&cd=1&ved=0CCYQFjAA&url=http%3A%2F%2Fweb.cs.swarthmore.edu%2F~meeden%2Fcs81%2Fs09%2Ffinals%2FRyan.pdf&ei=Eks9T4H4AuTUiALj-dzDAQ&usg=AFQjCNG3THVik7wByH61j67PnDqGDWn2Ww&cad=rja Feb 16 18:30:47 pbuckley: pioneer robots? Feb 16 18:32:11 pbuckley: same ideas. Going to be multiple ground and air vehicles cooperating togehter Feb 16 18:32:21 *together Feb 16 18:32:41 if you want me to hook you up with that author of that paper hes usually interested in talking about robots ;) Feb 16 18:33:14 that's be great, however this wont be my common irc name. Using webirc now, working on setting up a more permanent version right now Feb 16 18:37:10 strange Feb 16 18:37:57 pbuckley: think i got it setup Feb 16 18:39:36 pbuckley: yeah, looking at the paper, it seems pretty good with the adaptive controls Feb 16 18:40:09 orion__: When you start diving into this on the panda, I recomment you keep a complete copy of your workspace on your desktop that you can just dump into the SD after flashing a new image. It is much faster to flash a fresh image than it is to update on SD. Feb 16 18:41:40 GrueMaster: do you mean a complete copy of code, etc. I haven't moved anything to the distro yet. I am currently getting the board setup first to move code from a chumby hacker board Feb 16 18:43:16 Yes. What I do with my test projects is store them in a self-contained tarball, When I run them, I assume a clean image and install the necessary packages, change the environment settings, etc accordingly. Feb 16 18:44:14 Too often someone on here complains that they lose a lot of work when they decide to upgrade and they are on SD. Feb 16 18:44:59 I prefer to think of the SD as a good demo image/embedded project storage that is volatile. Feb 16 18:45:12 I see what you are saying. I will be working on setting up a cross compiler environment on my laptop soon. I will then be holding my projects both on the laptop and pandaboard so that shouldnt be a problem Feb 16 18:46:34 Can't wait to see your project on youtube. :P Feb 16 18:47:12 orion__: distcc is also fun Feb 16 18:47:28 me too. We already have 6 ground vehicles working well. Working on building 2 quadcopters at the moment, but those won't be done until the beginning of summer Feb 16 18:49:32 Anways, I don't want to tie up the board. I appreciate both of your help. Feb 16 18:50:36 come back soon ;) Feb 16 18:50:49 I'll be hanging out. Still have to see if this goes smoothly Feb 16 18:50:56 :D Feb 16 18:51:12 and the download for 12.04 is going very slow today Feb 16 18:55:18 pbuckley: This will be my main name if I am on Feb 16 18:58:02 alright.. ill let him know.. he's in the zone tippity tapping Feb 16 18:58:48 GrueMaster: Actually one question. Should I download the current daily build or the alpha 2 version the channel provides a link for Feb 16 18:59:26 Either or. The daily-preinstalled has a newer kernel with some fixes, mainly for audio. Feb 16 19:00:03 ok, then I will stick with that one. Thanks again Feb 16 21:25:06 Ooo, new shiny. http://makeplaylive.com/ Feb 16 21:27:13 Price and specs are comparable to a Nook Tablet. More usb ports and a camera as well. **** ENDING LOGGING AT Thu Feb 16 22:21:34 2012 **** BEGIN LOGGING AT Thu Feb 16 22:22:15 2012 **** ENDING LOGGING AT Fri Feb 17 02:59:57 2012