**** BEGIN LOGGING AT Sat Jan 05 02:59:57 2008 Jan 05 03:06:02 mornin' Jan 05 05:52:24 agittins: are you around? Jan 05 09:37:27 hi Jan 05 14:14:41 hi all Jan 05 14:14:57 cyrilRomain: you said : * cyrilRomain hates such 'rm -rf tmp', especially when bugs then disappear :[ Jan 05 14:15:07 cyrilRomain: it happened..! Jan 05 14:15:55 cyrilRomain: bug disappeared building from a fresh /tmp Jan 05 14:15:59 ant: hi! you don't have the bug anymore, even without the insane.bbclass patch from yersterday ? Jan 05 14:16:12 cyrilRomain: exactly Jan 05 14:16:14 gees Jan 05 14:16:29 cyrilRomain: dunno what to think... Jan 05 14:17:12 ant: I already had really weird issue with OE that just disappear after a complete rebuild from scratch Jan 05 14:17:23 cyrilRomain: keep the patch somewhere ... if somebody else notices the issue...you know... Jan 05 14:17:35 ant: yes Jan 05 14:17:50 cyrilRomain: I feel so sorry ... Jan 05 14:18:01 cyrilRomain: I took your time for nthg Jan 05 14:18:06 ... Jan 05 14:18:10 when such a bug appears, it is likely we did something wrong or the environement state is wrong Jan 05 14:18:31 but it is often not straightforward to figure out why Jan 05 14:18:55 ant: don't be sorry, I better know insane.bbclass :) Jan 05 14:19:17 cyrilRomain: about the qa_logs, the reside in the respective /tmp directory... Jan 05 14:19:19 ant: and by helping you I'm helping myself to better understand OE internals ;) Jan 05 14:19:26 cyrilRomain: he he Jan 05 14:19:47 cyrilRomain: why not append all QA to one single file in let say /stuff Jan 05 14:19:49 ? Jan 05 14:19:51 ant: qa log file reside in respective /temp directories Jan 05 14:20:34 and a qa log file is created for a package only when at least one QA issue happend Jan 05 14:20:35 cyrilRomain: I have difficulties grepping the entire tree for QA_LOG Jan 05 14:20:51 ant: right Jan 05 14:21:13 cyrilRomain: basically like Gentoo does Jan 05 14:21:14 putting all the QA log in one file is not a good solution IMHO Jan 05 14:21:36 cyrilRomain: you can split them Jan 05 14:21:40 but qa logs could be send to tinderbox for instance Jan 05 14:21:57 from which one should be able to see all qa issues Jan 05 14:22:08 cyrilRomain: or ato checking the length Jan 05 14:22:12 *auto Jan 05 14:22:32 cyrilRomain: of the MAIN_QA.log file Jan 05 14:22:49 * cyrilRomain didn't used tinderbox for month but perhaps it already provide such a feature now Jan 05 14:23:13 ant: I don't think creating one big file and then split it is the right solution Jan 05 14:23:13 cyrilRomain: if it is the case we need to learn much more.... Jan 05 14:23:26 ant: do be consistent with other log files in OE Jan 05 14:23:33 cyrilRomain: I mean let the logs as they are Jan 05 14:23:42 cyrilRomain: plu adding one main lofile Jan 05 14:23:56 * jeez the keyboard... Jan 05 14:24:34 cyrilRomain: just echo >> Jan 05 14:25:03 cyrilRomain: or better cat >> logfile Jan 05 14:26:13 cyrilRomain: new question: how am I supposed to find cruft in /tmp ? Jan 05 14:26:56 ant: cruft ? Jan 05 14:27:14 cyrilRomain: extraneous - forgotten files Jan 05 14:27:40 ant: what do you mean by extraneous ? Jan 05 14:27:42 cyrilRomain: I'm looking dir by dir ... Jan 05 14:27:58 ant: find -name '*loq.qa*' ? Jan 05 14:28:07 cyrilRomain: and in angstrom-gpe-task-settings Jan 05 14:28:58 cyrilRomain: is angstrom-gpe.task-settings.bb Jan 05 14:29:08 cyrilRomain: .bb ? Jan 05 14:29:13 ant: but yes there are too much file in tmp dir to use find :[ Jan 05 14:30:19 ant: sorry I'm afraid I don't understand what you are looking for Jan 05 14:30:41 cyrilRomain: forget about it, I'll check Jan 05 14:32:57 cyrilRomain: find is slowwww Jan 05 14:33:27 ant: yes forget about finding things in tmp directory: too much files here Jan 05 14:33:43 cyrilRomain: but I found one before by hand..he he Jan 05 14:33:51 cyrilRomain: ...evil resides in ... Jan 05 14:33:55 cyrilRomain: he he Jan 05 14:33:59 :) Jan 05 14:34:35 cyrilRomain: find didn't actually find nothing... Jan 05 14:34:50 cyrilRomain: I insist with grep.... Jan 05 14:36:31 ant: with tinderbox, you can actually show the whole log of a build Jan 05 14:36:46 ant: and looks for 'QA Issue' in that log Jan 05 14:37:04 cyrilRomain: ok, thx Jan 05 14:37:40 ant: but raw log can be quite huge (up to 40Mo!!) Jan 05 14:40:13 ant: a seppuku is also available and is aimed to directly report bug to bugzilla Jan 05 14:40:56 ant: so that QA issue are reported on a per-package basis Jan 05 14:41:23 cyrilRomain: good, will fill the issue-list Jan 05 14:41:25 ant: but I guess they didn't included QA issue not to bloat the bugzilla Jan 05 14:41:56 cyrilRomain: fill<>fullfill Jan 05 14:42:05 * cyrilRomain is looking at seppuku to see if mail can be send to someone else Jan 05 14:43:29 ant: if you care about autobug http://www.openembedded.org/wiki/AutoBug Jan 05 14:45:01 ant: before toying with seppuku, you'd better asking to zecke about QA Jan 05 14:45:29 cyrilRomain: very interesting btw Jan 05 14:48:05 ant: maybe the easiest is as you said to create in addition a 'global' qa log file e.g. in tmp/qa.log Jan 05 14:48:28 ant: but that file would be erased on each build Jan 05 14:48:44 ant: so it would only contains all the QA issues of the last bitbake build command Jan 05 14:49:31 cyrilRomain: you could log-rotate it... Jan 05 14:49:58 cyrilRomain: let sey keep 3 files (actual,previous,old) Jan 05 14:50:32 cyrilRomain: or even one would be fine after all... Jan 05 14:51:13 cyrilRomain: or append the creation time Jan 05 14:51:34 cyrilRomain: or better: the build # Jan 05 14:51:36 ant: anyway they are two kind of QA issues: those that are fatal and that stop the build (i.e. you can't miss them) Jan 05 14:52:29 and those that are not fatal (and that does not prevent the package to be installed and used) <- those are less easy to find since they do not stop bitbake Jan 05 14:53:44 cyrilRomain: you could improve it further but I think the logging is ok now Jan 05 14:54:02 cyrilRomain: this has to be verified on field... Jan 05 14:54:28 ant: yes, and probably zecke has already further plan in mind :) Jan 05 14:54:57 cyrilRomain: if new major errors pass the qa filtering..then improve it Jan 05 15:44:46 anyone running Google android here ? Jan 05 15:47:06 no, we are waiting for Google to realise their mistake and use Angstrom Jan 05 15:47:35 XorA|gone: hehe ;) Jan 05 15:48:18 /trigger andriod /kickban %u Jan 05 15:52:51 heh Jan 05 15:53:18 whats that sneaky guy mickey doing around here :-) Jan 05 16:07:39 lurking Jan 05 16:07:40 what else :D Jan 05 16:14:49 :) Jan 05 16:15:03 one can run android as a chroot under angstrom Jan 05 17:37:01 summatusmentis: you pinged me several hours ago? Jan 05 18:33:38 RP: ping Jan 05 18:34:00 RP: I'm again in the gfortran mess... Jan 05 18:34:06 RP: http://rafb.net/p/QLEFYj90.html Jan 05 18:34:28 RP: NOTE: package gcc-cross-4.2.2: failed Jan 05 18:35:14 RP: configure: error: GNU Fortran is not working; Jan 05 18:35:44 RP: DISTRO_VERSION = "2008.1-test-20080105 Jan 05 18:46:26 ant: mmm, i worked around that by installing my linux distro's gfortran package... Jan 05 18:48:02 jkent: there was a bug about the issue but IMHO was fixed... Jan 05 18:48:13 ahha. Jan 05 18:48:30 jkent: gfortran is not needed on host (if anything changed ) Jan 05 18:48:59 jkent: I know .dev is not a breeze ;-) Jan 05 18:49:04 as it shouldn't be.. but yes Jan 05 18:49:36 jkent: angstrom-2007.12-stable built ok (gcc 4.1x) Jan 05 18:50:08 jkent: has to do with gcc-4.2x Jan 05 18:50:22 *probably Jan 05 18:51:31 jkeny: anyway thanks, shall try Jan 05 18:53:07 ant: this is recent pull from yesterday, i just build 2008.1 Jan 05 18:53:54 jkent: I have to find an old gfortran for ArkLinux then... Jan 05 18:54:40 gfortran 4.2.1-4ubuntu2 is what i threw on here Jan 05 18:54:58 jkent: no gcc-3 ? Jan 05 18:56:33 jkent: mmh...gcc-fortran-4.21 or gcc-fortran-static-4.21 Jan 05 18:56:49 lets see, i have gcc-3.4 and 4.2 available Jan 05 18:56:57 jkent: have to dig in obsoleted Jan 05 18:58:54 jkent: s*@it...no gcc-fortran-3.3.5 Jan 05 18:59:38 ehm...I cannot mix the two versions... Jan 05 19:00:29 package system does not allow it without breaking things? Jan 05 19:00:58 jkent: I feel dirty only thinking at that... Jan 05 19:01:07 haha Jan 05 19:01:23 i know nothing about ark Jan 05 19:01:40 jkent: rapid to install.stop. Jan 05 19:02:05 jkent: provides a working gcc-3.3.5 Jan 05 19:03:01 jkent: KDE 2.1 GB on disk Jan 05 19:03:16 jkent: +openembedded of course... Jan 05 19:03:42 jkent: perfect if you have a small partition Jan 05 19:04:24 jkent: I only use it for building OE... Jan 05 19:04:51 jkent: until I take some time and retry on Gentoo without gcc-33 Jan 05 19:05:07 jkent: should work, now Jan 05 19:05:24 jkent: with standard gcc-4x on host Jan 05 19:05:32 hmm, okay Jan 05 19:05:39 * FYI ant never tried Jan 05 19:05:58 * quemu is a beast! Jan 05 19:06:04 that it is Jan 05 19:07:18 jkent: bbl (food) Jan 05 19:28:35 where is located the src code of the nslu2 branch for building the minimal image? or do you have a link to that docs? Jan 05 20:06:09 agittins: we're playing ping tag Jan 05 20:38:29 anybody know if angstrom supports prism2_srec flashing? and not just sending in non volatile ram? Jan 05 20:38:50 IIRC that has been intentionally disabled Jan 05 20:38:57 doh Jan 05 20:39:09 which device do you need it for? Jan 05 20:39:14 ambicom Jan 05 20:39:21 the pda Jan 05 20:39:28 zaurus slc3200 Jan 05 20:39:33 * CoreDump|bbl hrms Jan 05 20:39:45 i wanna send the firmware to the non volatile ram (flash) Jan 05 20:39:54 sick of uploading to volatile ram Jan 05 20:40:52 ugh... now I've gotta find another ubuntu cd... Jan 05 20:40:58 I hate hard drive failure Jan 05 20:41:52 Pixelized: Angstrom does the upload automatically, and it takes so little time you wouldnt notice Jan 05 20:43:02 IIRC it needs enabling in the kernel .config Jan 05 20:43:26 not running angstrom right now thinking of switching back Jan 05 20:43:36 pdaxii13 has a nice huge install but tons of bugs Jan 05 20:43:53 and takes forever to boot lol Jan 05 20:44:46 but would've been nice to have it permenantly installed in flash Jan 05 20:45:29 Pixelized: we disabled it by default as too many people think that you have to upgrade to latest versions, then they end up with buggered wlan cards and blame us Jan 05 20:45:37 lol Jan 05 20:45:49 there are fixes Jan 05 20:45:52 Pixelized: if you have an OE build its one flip in defconfig Jan 05 20:45:53 even if you screw it up Jan 05 20:46:15 from what I remeber Jan 05 20:46:50 resurrection/prismdos.txt <-- mind u needs dos Jan 05 20:46:57 http://linux.junsun.net/intersil-prism/dos-resurrection/prismdos.txt Jan 05 20:47:03 for those who scraped their cards Jan 05 20:47:29 cool I never knew about that Jan 05 20:47:33 anyway I is off Jan 05 20:48:04 Pixelized: I'll build you a kernel, hold on Jan 05 20:49:02 i'll brb Jan 05 20:49:05 gotta do something Jan 05 20:49:11 CoreDump|bbl that would be nice Jan 06 01:04:01 *yawn* any news on whether an unsupported image will be compiled for the iPaq rx3000 series? Jan 06 01:04:24 OasisGames: You could prolly build one yourself.. Jan 06 01:04:48 It'll take 6-24 hours, depending on your build machine and image target Jan 06 01:05:33 I probably could but I'll probably break something (like how I somehow managed to get my battery to not be used in Opie...) Jan 06 01:06:15 You'd be building off the same tree as the rest of us, and anyways, no better place to learn :) Jan 06 01:06:50 OasisGames: how did you manage that? Jan 06 01:07:48 my phone recognizes vt100 as a word!! weird Jan 06 01:08:12 summatusmentis: Well, it's a product made by geeks ;) Jan 06 01:12:20 OasisGames: fyi, someone posted a patch to get rx3000 support into Opie... he has not yet responded to my question as to whether he wanted it merged into Opie CVS Jan 06 02:20:19 summatusmentis: you're it :-) Jan 06 02:23:38 summatusmentis: feel free to just ask away, I usually catch up on what i've missed. Jan 06 02:29:26 lol, alright, here I am. my question was if you knew anything about collie kernel stuff that's happened recently Jan 06 02:29:32 summatusmentis: I have no idea. Kernel boots, OpenEmbedded boot screen shows, gets to the "Tap the screen to continue" as it always does when the screen configuration is messed up or missing, and then *pop* back light goes out and the entire thing shuts down. Jan 06 02:29:50 I had hard drive failure, so I have to re-setup OE and stuff Jan 06 02:29:57 Wrong power management modules possibly. I was working off of an old (June/July) build that only just barely worked. Jan 06 02:30:27 any idea what kernel version that is? Jan 06 02:30:38 OasisGames: you're on zaurus 5500? Jan 06 02:30:44 rx3700 Jan 06 02:30:58 ah, ok, collie == zaurus 5500 Jan 06 02:31:14 I was referring to the question you asked me earlier (how I managed to get my battery to not work) Jan 06 02:31:17 oh wait, you're replying to my earlier question, never mind Jan 06 02:31:22 yeah, I got confused Jan 06 02:31:54 I then went and watched "Shoot 'em up" and came back to my laptop and saw the question. Jan 06 02:34:54 I'd probably be able to build my own image for my iPaq, but I've got my hands full with my Compiz plugin (http://forum.compiz-fusion.org/showthread.php?t=6395) and various other things Jan 06 02:39:38 There is a config file for the rx3000, I just need to set up the build environment on my laptop... (http://www.openembedded.org/repo/org.openembedded.dev/conf/machine/rx3000.conf) **** ENDING LOGGING AT Sun Jan 06 02:59:57 2008