**** BEGIN LOGGING AT Mon Aug 21 02:59:57 2006 Aug 21 04:21:24 Stupid question. How do you make bitbake package a lib to the package.ipk instead of the package-dev.ipk? Aug 21 04:23:05 The .so file shows up in the package-dev/usr/lib directory and gets packaged into the package-dev.ipk Aug 21 04:40:16 DoTheDog: change the FILES_ variables. it puts the .so into the -dev by deafult, because for nearly every package that does what it should, the .so is just a symlink Aug 21 04:41:03 Thanks kergoth, but I am a noob at this. Where does it go and what does it look like? Aug 21 04:42:10 I have posted a .bb file here: http://pastebin.ca/141913 Aug 21 04:48:57 DoTheDog: grep is your friend. i suggest you look at some of the 3000+ .bb files Aug 21 04:49:32 thanks. Aug 21 04:57:52 RP: pong. Go ahead ;-) Aug 21 05:16:23 kergoth: I tried this: Aug 21 05:16:26 PACKAGES = "serializeclass serializeclass-dev" Aug 21 05:16:26 FILES_serializeclass = "${S}/*.so" Aug 21 05:16:26 FILES_serializeclass-dev = "${FILES_serializeclass}" Aug 21 05:16:34 and it didn't work. Aug 21 05:16:50 any ideas? Aug 21 05:19:13 There is a "libSerializeClass.so" in the ${S} dir. Aug 21 07:07:06 morning all Aug 21 07:09:26 morning.. Aug 21 07:09:50 I just stumbled over a problem with image_ipk.bbclass Aug 21 07:10:06 it is supposed to DEPEND upon makedevs-native Aug 21 07:10:14 which is all good and well.. Aug 21 07:10:51 except for when the image package is build from a metapackage Aug 21 07:11:39 eg. I have a metapackage named 'stv-flashimage', which pulls together two images and bundles them into a flashable file (with the help of imgloader) Aug 21 07:12:33 when I run bitbake stv-flashimage, it gets all the way to stv-image (whic inherits from image_ipk), and fails at do rootfs Aug 21 07:12:52 with 'makedevs: command not found' Aug 21 07:13:52 if I bitbake stv-image on it's own, it works fine Aug 21 07:15:57 stv-flashimage looks like this: http://rafb.net/paste/results/px7BSv84.html Aug 21 07:16:16 can anyone see anything that may be causing this problem? Aug 21 07:19:36 hmm... well, I moved the inherit image_ipk line down to the bottom of stv-image, and it seemed to fix the problem... Aug 21 07:19:51 I had it right at the top before... above DEPENDS = Aug 21 07:19:57 I think that makes sence Aug 21 07:25:30 another really annoying problem... unionfs-modules... Aug 21 07:26:03 tkp: annoyance? they are unmaintained Aug 21 07:26:09 hehe Aug 21 07:26:34 I have it in RDEPENDS for a PACKAGES package... Aug 21 07:26:46 the package appears to have been build, yet there is no stamp for it Aug 21 07:26:47 hi zecke Aug 21 07:27:01 so when it comes to doing the rootfs, it rekons it can't find it Aug 21 07:27:18 RP: you are insane! Aug 21 07:27:25 and it gives me "Cannot satisfy the following dependencies for stv-base-depends" Aug 21 07:27:29 RP: I received an mail at 8h clock, so when did you wake up? Aug 21 07:27:39 where stv-base-depends is the PACKAGES package from task-stv Aug 21 07:28:08 zecke: I woke up about half an hour ago, why? Aug 21 07:28:23 only way I can sort it out, is to clean it, build it manually, and then build the rootfs image again Aug 21 07:28:35 (clean unionfs-modules) Aug 21 07:29:00 Hmm. According to hh.org, I imagine things... **** BEGIN LOGGING AT Mon Aug 21 07:30:42 2006 Aug 21 07:35:11 morning Aug 21 07:41:41 morning Aug 21 08:27:03 I am trying to build an external kernel module for the ppc and i am faced with the problem that the ppc kernel is moving towards the powerpc tree but my devices are still in the ppc, as a result a number of include files are missing and as such the nodule fails to compile Aug 21 08:27:27 Ifaistos: I have hacked on icecc.bbclass Aug 21 08:27:28 i manage to get the 'correct' set of hedader files to make the module compile Aug 21 08:27:33 Ifaistos: I will need to try it though Aug 21 08:28:06 zecke : i am planning on setting the first cluster boxes today, later in the eveing Aug 21 08:28:37 zecke : once ready i'll let you know Aug 21 08:30:46 and here is the question ... Aug 21 08:31:03 what is the 'best' way to include these headers Aug 21 08:31:39 Ifaistos: I'm not that familiar to compiling kernel modules Aug 21 08:32:02 Ifaistos: but IIRC we have a directory with Kernel Headers (in contrast to sanitized headers for libc) Aug 21 08:32:09 one option is to include ALL headers as a patch and then point the module to them i.e /workdir/kernel fro example Aug 21 08:33:03 Ifaistos: and one option is to add (by patch/install) the kernel headers into the right staging dir Aug 21 08:33:09 zecke : if i move them to the kernel dir its probable that they will break something... as they were not intented to be there in the first place Aug 21 08:33:47 Ifaistos: hmm Aug 21 08:33:51 anyway need to work Aug 21 08:40:35 anyone know how to find the library paths that a ld will search? Aug 21 08:45:09 mithro: cat /etc/ld.so.conf + /lib /usr/lib? Aug 21 08:45:56 Ifaistos: use STAGING_KERNEL_DIR - if module does not build with them then it is broken Aug 21 08:48:34 hrw|work : It won't build. there are some headers 'missing' Aug 21 08:50:02 mithro: + LD_LIBRARY_PATH Aug 21 08:50:09 mithro: env variable Aug 21 08:51:36 x/cr Aug 21 08:51:55 Ifaistos: so it need full kernel sources? Aug 21 08:53:58 hrw|work : the ppc tree is moving to the powerpc tree so you compile the kernel as ppc arch but need to build the external kernel modules as powerpc arch... kind of crazy if you ask me... Aug 21 08:55:22 hrw|work : the ppc tree is moving to the powerpc tree so you compile the kernel as ppc arch but need to build the external kernel modules as powerpc arch... kind of crazy if you ask me... Aug 21 08:56:20 hrw|work : i found which header are missing in order to compile the module but if i add them to the kernel dir (as a patch) the kernel does not compile Aug 21 08:57:03 hrw|work : what i am afraid is that if i add them (the headers) to /staging/kernel they will break something else... Aug 21 08:57:40 For some reason, my OE on one of my computers is again refusing to compile stuff: http://rafb.net/paste/results/DBk6tB82.html. Strange thing is that "/usr/bin/env python" which seems to be the shell called by /export/interim/openembedded/bitbake/bin/bitbake straight from the command line goes into that shell just fine. Aug 21 08:58:03 cr Aug 21 08:58:06 ops Aug 21 08:58:39 hrw|work : so i was thinking of 'shipping' the all the kernel headers with the module and then point the module there, in order to avoid 'contamination' of the real kernel headers.... Aug 21 08:58:57 Ifaistos: a bit of hardcore... Aug 21 09:00:28 good morning Aug 21 09:00:36 hrw|work : even then i have to take care for different kernel versions i.e the swifting started from 2.6.13 so id kernel is up to 2.6.12 the module compiles fine with the /saging/kernel... Aug 21 09:00:56 florian : hi Aug 21 09:01:40 hi florian Aug 21 09:03:01 hrw|work : i am thinking of spliting the module in 2 bb files.one for all arches and one called ppc-module that all the 'magic' need for ppc will take place Aug 21 09:04:00 can anyone give me a hand? doing a rebuild and coreutils isn't building :s Aug 21 09:04:18 In file included from lbracket.c:2: Aug 21 09:04:18 | test.c:129: error: static declaration of 'eaccess' follows non-static declaration Aug 21 09:04:19 | /usr/include/unistd.h:266: error: previous declaration of 'eaccess' was here Aug 21 09:04:19 | make[3]: *** [lbracket.o] Error 1 Aug 21 09:05:21 Ifaistos: if it will work... Aug 21 09:06:34 mikearthur: There is a patch for that in poky Aug 21 09:06:46 RP: whats poky and where? Aug 21 09:06:51 ~poky Aug 21 09:06:53 i heard poky is http://projects.o-hand.com/poky, or a subset of OpenEmbedded, created by o-hand.com to testdrive some of their technology Aug 21 09:07:00 mikearthur: grab the coreutils .bb Aug 21 09:09:17 KILL mc Aug 21 09:09:40 RP: coreutils_5.1.3.bb or coreutils-native_5.1.3.bb? I'm building coreutils-native Aug 21 09:09:57 5.1.3? did not we have 5.3.0? Aug 21 09:10:09 http://svn.o-hand.com/repos/poky/trunk/meta/packages/coreutils/ Aug 21 09:10:13 thats where I'm looking Aug 21 09:11:33 hey mikearthur Aug 21 09:11:42 XorA: hey, and congrats on the offer Aug 21 09:12:44 mikearthur: cheers, one day they may even post it Aug 21 09:13:30 heh Aug 21 09:20:28 Ok, I found the problem: The NFS mount had the noexec bit set ;-) Aug 21 09:26:48 XorA: are there any other possible locations ld could search? Aug 21 09:28:39 * koen returns Aug 21 09:28:52 koen: ominous Aug 21 09:30:01 mikearthur: There is a patch for the issue you're seeing in there Aug 21 09:30:08 wb koen Aug 21 09:30:17 RP: coolio Aug 21 09:34:20 hi koen Aug 21 09:35:45 RP: re syslog-ng. Just go ahead. Aug 21 09:35:55 * NAiL is in Sweden at the moment Aug 21 09:36:04 NAiL: ok, thanks. Just wanted to check first :) Aug 21 09:38:13 * XorA hides from koen Aug 21 09:38:31 mithro: only LD_PRELOAD that I know of Aug 21 09:38:58 XorA: wait till I have showered before trying to unhide :) Aug 21 09:39:03 i'm trying to write a quick shell script which searchs all the library paths for a special directory Aug 21 09:39:28 koen: em, now I truly am scared Aug 21 09:40:47 anyon e know the search order? Aug 21 09:42:06 XorA: working at lowlands.nl in the rain + mud + sewage in a night shift Aug 21 09:46:19 koen: you did this of free will? Aug 21 09:47:27 XorA: yes, and getting paid for it :) Aug 21 09:47:50 koen: I preffer being paid to stay out of the sewage Aug 21 09:48:30 XorA: luckily the sewage part only lasted a few minutes Aug 21 09:50:40 ANgstrom, sewage powered :-) Aug 21 09:56:02 morning all Aug 21 09:56:07 hi liam Aug 21 09:56:16 hi Liam Aug 21 09:57:41 koen: is there a way of limiting the glibc locale generation to certain locales only? Aug 21 09:58:00 hey liam Aug 21 09:58:47 mikearthur: iirc RP had a patch for that Aug 21 10:14:49 koen: wtf... Apple Mail sent an empty mail to you Aug 21 10:15:35 koen: http://www.oesf.org/forums/index.php?showtopic=21120 Aug 21 10:17:36 zecke: heh, and I was cursing thunderbird for noyt displaying your mail :) Aug 21 10:17:45 hrw|work: cute Aug 21 10:22:40 mikearthur: poky is patched for that Aug 21 10:26:41 RP: uh, that would be nice to have upstream Aug 21 10:27:17 florian: It breaks the other locales and I could be crucified if I checked it in - its a hack :-( Aug 21 10:27:41 s/could/would/ Aug 21 10:27:56 RP: ah, too bad Aug 21 10:28:42 florian: Its on my list of things to fix up at some point but its no straighforward and testing is a pain with glibc taking so long to package... Aug 21 10:28:58 RP: indeed Aug 21 10:34:18 lrg: got patch, will test when my image buld finishes Aug 21 10:35:41 lrg: RP: is it expected that a gcc 3.4.4 compiled alsa lib will fail calling ioctl's on a gcc 4.1.1 compiled EABI kernel? Aug 21 10:37:14 XorA: not sure, I would enable the backwards compat with the old ABI in the kernel though. Aug 21 10:37:32 XorA: It depends whether the compatibility layer was enabled in the kernel Aug 21 10:37:36 lrg: I can boot OZ on EABI kernel, only thing that fails is alsa Aug 21 10:37:55 XorA: buts its also possible someone broke compatiblity in ALSA Aug 21 10:39:00 I get innaproprate ioctl errors ( not the exact messaqge as machine is not booted) Aug 21 10:39:05 libxine: | dsputil_arm.c:54: error: PIC register `r10' clobbered in `asm' ? Aug 21 10:39:16 Has anyone seen that before? Aug 21 10:39:46 RP: as that file is probably same as one from mplayer, Id do a diff and see if mplayer guys bugfixed it Aug 21 10:40:22 XorA: I can guess the fix if not, I'm just wondering why its only shown up now... Aug 21 10:40:44 RP: I certanly havent seen it and compile gxine a couple of days ago Aug 21 10:42:52 RP: using gcc 4.1.1 Aug 21 10:43:05 * XorA had to check which buld tree the ipk was in :-) Aug 21 10:43:39 This is 3.4.3... Aug 21 10:44:21 its not in my OZ feed Aug 21 10:50:13 psokolovsky: nice work on the bootchart Aug 21 10:50:41 psokolovsky: did you consider timing a first, second and third boot to see how much they differ? Aug 21 10:51:20 koen: Hi! Aug 21 10:52:08 koen: Yes, I did several runs, but so far mostly craed for relative time, not absolue precise measurements. Aug 21 10:52:46 koen: To do that, first need to make bootchart stop automatically at the last stage of boot. Aug 21 10:53:15 koen: Will work on that, submit patches for bootstrap-image-bootchart, now that it's committed. Aug 21 10:55:33 ~hail dvbsnoop Aug 21 10:55:35 * ibot bows down to dvbsnoop and chants, "I'M NOT WORTHY!!" Aug 21 10:55:41 anyone remeber the $1 like thing for all arguments given? (in sh scripting) Aug 21 10:56:16 mithro: $@ or use shift Aug 21 10:56:32 $@ is all of them? Aug 21 10:56:44 mithro: yes Aug 21 10:56:53 thanks Aug 21 10:57:11 mithro: be aware that arguments with spaces will get split Aug 21 10:57:43 anyway to chance to fix that? Aug 21 10:57:57 "$@" in right places afaik Aug 21 10:59:15 :/ Aug 21 10:59:20 tim@ultraslim:~/oss/tp/tpclient-pywx/doc$ ./tp-pywx-installed test abc "testing abc" Aug 21 10:59:21 test abc testing abc Aug 21 11:12:00 suihkulokki: any idea? Aug 21 11:12:37 mithro: use shift then re-quote all args in your new commandline Aug 21 11:13:04 XorA: :/ Aug 21 11:13:10 mithro: or perl Aug 21 11:14:19 mithro: or use a shell that has array support Aug 21 11:22:27 Hi, trying to build tslib here. Something seems to be wrong with the fetch URL (CVS). Can somebody take a look? http://rafb.net/paste/results/FXEQDg69.html Aug 21 11:27:23 Laibsch: update bitbake to 1.4.4 or 1.6.0 Aug 21 11:27:43 seems this works Aug 21 11:27:45 ARGS="" Aug 21 11:27:46 for ARG in "$@"; do Aug 21 11:27:46 ARGS="$ARGS \"$ARG\"" Aug 21 11:27:46 done Aug 21 11:35:11 03hrw 07org.oe.dev * re5fc9068... 10/ (1 conf/distro/include/openzaurus.inc): openzaurus.conf: force BitBake to be atleast 1.4.4 Aug 21 11:36:00 hrw|work: Do I need to check it out again? "svn up" tells me I have the latest version which is 574. Aug 21 11:36:18 it = bitbake Aug 21 11:36:22 Laibsch: well... did you checkout a tag? Aug 21 11:37:09 Laibsch: e.g. the http://svn.berlios.de/viewcvs/bitbake/ Aug 21 11:37:21 Laibsch: depending on where you are on, svn up will do nothing Aug 21 11:37:28 Laibsch: svn info, tells you where you are Aug 21 11:37:29 zecke: Not that I know of. Unless that is what was written on GettingStarted, but I do not think it was. Aug 21 11:38:29 Laibsch: well, GettingStarted is outdated again then :) Aug 21 11:38:40 zecke: http://rafb.net/paste/results/2USz8p72.html Aug 21 11:38:43 as it uses 'tags' Aug 21 11:39:04 Laibsch: I'm getting less and less conviced that having a GettingStarted is a good idea Aug 21 11:39:26 zecke: What is the alternative? Aug 21 11:39:28 Laibsch: what we see now is people stopping to think and blindly following directions Aug 21 11:39:40 Laibsch: I have no clue Aug 21 11:40:03 and I'm about to get moody/hungry so I better stop Aug 21 11:40:12 hehe Aug 21 11:40:17 Laibsch: you should go from svn://svn.berlios.de/bitbake/tags/bitbake-1.4.2 Aug 21 11:40:26 Laibsch: to vn://svn.berlios.de/bitbake/branches/bitbake-1.4 Aug 21 11:40:32 Laibsch: or svn://svn.berlios.de/bitbake/tags/bitbake-1.6 Aug 21 11:40:33 you mean move away? Aug 21 11:40:41 Laibsch: svn switch Aug 21 11:40:43 I do have 1.4.2 Aug 21 11:41:03 OK, I see. Aug 21 11:41:07 Laibsch: I write some 'SVN' conventions Aug 21 11:41:15 Laibsch: we have three dirs, trunk, branches and tags Aug 21 11:41:25 Is it Ok to go straight to 1.6? Aug 21 11:41:26 Laibsch: trunk contains latest, best, most broken stuff Aug 21 11:41:43 Laibsch: tags contains 'releases', even if we can we will not update the tag Aug 21 11:41:58 Laibsch: and branches currently contain 'stable' branches we do bugfixing on Aug 21 11:42:12 Laibsch: e.g. bitbake-1.4 and bitbake-1.6 are considered stable branches Aug 21 11:42:23 Laibsch: both should work, but bitbake-1.6 should be a bit faster Aug 21 11:42:28 Thanks, zecke. I'll take 1.6, then. Aug 21 11:42:45 Laibsch: using svn switch you do not need to recheckout stuff Aug 21 11:42:46 good choice Aug 21 11:43:10 koen: can drupal do tables? Aug 21 11:43:26 why would diet-x11 be looking at my /usr/lib/libXau.so rather than the arm one? Aug 21 11:43:39 mikearthur: because it stinks ;) Aug 21 11:43:45 zecke: the wiki? Aug 21 11:43:49 zecke: no idea Aug 21 11:44:07 zecke: rephrase the question then ;), how would I stop it looking at my /usr/lib/libXau.so rather than the arm one? Aug 21 11:44:26 mikearthur: find the reason for it looking there Aug 21 11:44:37 mikearthur: this can be either a broken .pc or .la file pointing to /usr/lib Aug 21 11:44:51 mikearthur: or a stupid configure check hardcoding /usr/lib as path Aug 21 11:45:07 mikearthur: in most cases consulting config.log auf the autoconf/configure run is a good idea Aug 21 11:45:13 ok Aug 21 11:45:51 koen: does drupal support kind of 'recent changes'? Aug 21 11:46:00 hrw|work: it has an rss feed Aug 21 11:46:27 hmm Aug 21 11:47:41 Well, no matter what I do, I get "svn: Cannot replace a directory from within" Aug 21 11:48:19 try --relocate Aug 21 11:49:36 zecke: libdir='/usr/lib' Aug 21 11:49:39 zecke: sucks Aug 21 11:50:41 mikearthur: and installed=true? Aug 21 11:51:04 zecke: Now I get "$PATH is not a valid relocation" for which google also is not much help. I think I will just check out again. Aug 21 11:51:07 just doing a clean and rebuild Aug 21 11:51:19 Laibsch: then recheckout ;) Aug 21 11:51:32 mikearthur: which .la contained that? Aug 21 11:51:46 mikearthur: or was that coming from a config.log Aug 21 11:51:55 zecke: config.log Aug 21 11:52:09 zecke hm it alos can be libbla-config Aug 21 11:52:11 mikearthur: that is okay Aug 21 11:52:26 woglinde: they should use pkgconfig Aug 21 11:53:30 zecke yeah but libxine is a candidate Aug 21 11:53:38 which uses libpng-config Aug 21 11:55:07 zecke: After looking at the tags, I found out that 1.6 and 1.6.0 are indeed different ;-) Aug 21 11:58:37 zecke: pastebin http://rafb.net/paste/results/c9xHkq41.html Aug 21 11:59:12 mikearthur: yes, config.log please Aug 21 11:59:45 sure Aug 21 12:00:47 zecke: http://rafb.net/paste/results/gmRb4o13.html Aug 21 12:01:48 * zecke orders thai Aug 21 12:02:20 mikearthur: no, I need more stuff ;) Aug 21 12:02:33 zecke: I'm confused :s Aug 21 12:04:10 mikearthur: well, such is life Aug 21 12:04:16 :D Aug 21 12:04:44 mikearthur: Somehow you have /usr/lib/libXau.so Aug 21 12:04:58 mikearthur: in your compile/link line Aug 21 12:05:01 yeh Aug 21 12:05:06 mikearthur: now we need to find it why that happens Aug 21 12:05:17 because OE hates me? Aug 21 12:05:32 svn fetches fine now. Thanks guys. Aug 21 12:05:34 mikearthur: no, just use grep Aug 21 12:17:28 woglinde: hrw|work http://www.openembedded.org/oedem Aug 21 12:20:40 zecke arrivals? Aug 21 12:22:59 zecke: added Aug 21 12:25:11 done Aug 21 12:33:12 morning Aug 21 12:33:25 hi chouimat Aug 21 13:13:07 getting a "./mkg3states: cannot execute binary file: Aug 21 13:13:39 while building liblbxutil-X11R7-1.1.0.1-r0 Aug 21 13:18:52 hmm theres a patch for it, but its not applied Aug 21 13:28:27 Applying it didnt make any difference though Aug 21 13:40:59 Argh Aug 21 13:47:14 RP, do13: ping Aug 21 14:00:43 re Aug 21 14:01:07 zecke: wb Aug 21 14:02:18 oedem people: what do you think about my mail/ Aug 21 14:02:20 ? Aug 21 14:05:06 hrw|work: silent agreement from my side Aug 21 14:05:19 hrw|work: s/beer/ACME alcohol/ Aug 21 14:06:27 acme alcohol? Aug 21 14:06:36 zecke: selfmade too? Aug 21 14:07:07 any sort of alcohol (cocktails etc...) Aug 21 14:07:50 sure - beer is just a name Aug 21 14:09:32 rsync rulez Aug 21 14:10:03 s/beer/dinner + drinks/ Aug 21 14:13:18 frikandel+drinks Aug 21 14:14:19 nachos Aug 21 14:14:23 hehe Aug 21 14:14:30 zecke: Nynke would kill me if we'd go out for frikandels Aug 21 14:14:31 * zecke hands money to woglinde Aug 21 14:16:59 he, why? Aug 21 14:17:49 I don't think she'd travel to berlin to eat some frikandels with a bunch of geeks Aug 21 14:17:53 (but I could be wrong) Aug 21 14:18:13 oh she comes to berlin? Aug 21 14:19:05 was is frikandels? Aug 21 14:19:16 sausage? Aug 21 14:19:50 gracias Aug 21 14:20:05 a deepfried sausage Aug 21 14:22:29 http://en.wikipedia.org/wiki/Frikandel Aug 21 14:22:43 Does anybody know what is up with CoreDump? Aug 21 14:22:43 * koen is surprised about wikipedia knowing about that Aug 21 14:23:28 koen: eeek, horse... Aug 21 14:24:19 koen wyh not? Aug 21 14:24:52 Yum!!!! Fried Mystery meat :) Aug 21 14:25:15 zecke horsemeat is tasty Aug 21 14:25:27 zecke: I think I'll arrive on wednesday if my uni schedule permits Aug 21 14:25:31 re Aug 21 14:26:29 that way I can see a bit of berlin :) Aug 21 14:28:17 hehe, location is 98% secured (if we do not outgrow it) Aug 21 14:28:29 accomodation is another hot topic... Aug 21 14:29:01 need to work about though Aug 21 14:29:23 koen: BTW: I love the name 'Nynke' Aug 21 14:30:15 the name of my firstborn must contain a 'ij' and 'y' Aug 21 14:31:04 "Friesischen Upsilon" Aug 21 14:31:17 As we taught our german boss Aug 21 14:31:18 lol Aug 21 14:31:30 Uepsilon? Aug 21 14:31:52 upsilon, üpsilon, whatever ;) Aug 21 14:32:11 bleh Aug 21 14:32:20 I've been awake for 29 hours now Aug 21 14:32:23 * koen gets some food Aug 21 14:32:29 oh Aug 21 14:32:51 koen get some slepp Aug 21 14:33:00 hrw|work: pong Aug 21 14:33:22 RP: I will add do_rm_work() {} to linux-openzaurus.inc Aug 21 14:33:38 RP: this allow to use rm_work.bbclass and build wlan-ng stuff Aug 21 14:34:44 hrw|work: ok. I've yet to actually try rm_work :) Aug 21 14:36:30 RP: who know how it will fsck multitasking ;D Aug 21 14:36:34 cbrake: hi Cliff Aug 21 14:37:18 morning Aug 21 14:37:31 my biggest wish right now is for some way for uclibc and gcc to co-exist peacefully Aug 21 14:37:34 woglinde: I'm trying to straighten out my sleep schedule after working a nightshift Aug 21 14:37:39 so you could make an initrd and the rootfs in the same build Aug 21 14:37:54 CosmicPenguin: not worldpeace? Aug 21 14:38:07 nah - world peace doesn't really benefit me, you know? Aug 21 14:42:07 hrw|work: It shouldn't as it won't execute the rm_work task until the build is complete? Aug 21 14:42:33 hi CosmicPenguin Aug 21 14:42:56 CosmicPenguin: Could we extend multimachine to cope with that? Aug 21 14:42:57 RP: hey Aug 21 14:43:14 RP: rm_work is after do_package and do_deploy Aug 21 14:43:47 wtf? Aug 21 14:43:57 bitbake svn seems to be totally broken Aug 21 14:44:01 RP: possibly Aug 21 14:44:05 koen: trunk? Aug 21 14:44:07 koen: svn what? Aug 21 14:44:12 hrw|work: yes Aug 21 14:44:21 koen: Define totally broken? Aug 21 14:44:32 koen: its WIP to multitask bitbake Aug 21 14:44:35 * zecke gets a CD from his wishlist ;) Aug 21 14:44:44 I still think that we can pull it off with intellegent grubbing of the spec files Aug 21 14:44:48 but I haven't had any time to investigate Aug 21 14:44:50 * zecke pokes RP to buy a CD from my Canossa list Aug 21 14:45:00 koen: multithreading got merged last night and some of the metadata in OE is broken ;-) Aug 21 14:45:02 RP: it's trying to build 2.4 zaurus kernels for any machine I try to use Aug 21 14:45:17 and failing to find an eabi 2.95.3 Aug 21 14:45:42 koen: You'd better mail me your local.conf :-/ Aug 21 14:47:04 Does anyone have any thoughts on the circular dependency problem I mentioned on oe@ ? Aug 21 14:47:38 no Aug 21 14:48:15 zecke: Would you agree its a bug in debian.bbclass rather than bitbake? Aug 21 14:48:30 let me reread it Aug 21 14:48:43 RP: http://ewi546.ewi.utwente.nl/tmp/bb-b0rked/ Aug 21 14:50:50 koen: Those removal messages were not it trying to build it, it was it noting that it couldn't... Aug 21 14:51:20 I agree, we need to improve the reporting of it as they're not errors. Trying to tell an error case from a none error case is tricky Aug 21 14:51:33 "ERROR: Trying to re-add a failed file? Something is broken.." is a problem though Aug 21 14:52:47 koen: What were you trying to build btw? Aug 21 14:52:50 * koen ponders about mandating COMPATIBLE_MACHINE in kernel recipes Aug 21 14:52:58 RP: bitbake angstrom-gpe-image Aug 21 14:53:12 RP: 'bitbake minimix' does seem to do something Aug 21 15:02:57 koen: It breaks for me as well which means its reproducable - I'll investigate later tonight Aug 21 15:10:12 RP: what multithreading: http://svn.berlios.de/viewcvs/bitbake/trunk/bitbake/ChangeLog?rev=566&view=markup ? Aug 21 15:10:44 koen: trunk is 1.7.x now Aug 21 15:11:05 nothing in the changelog about that Aug 21 15:11:26 koen: Its under development ;-) Aug 21 15:11:50 but I'll make sure it gets added to the 1.7.0 section of the changelog :) Aug 21 15:11:56 :) Aug 21 15:12:25 * koen wonders how to enable multithreading Aug 21 15:12:56 BB_NUMBER_THREADS=5 in loac.conf atm but that is subject to change Aug 21 15:13:27 Its a really stupid/simple algorithm atm - I'll fix the dependency handling code, then improve the multithreading bit Aug 21 15:16:20 zecke: I'll implement your comments (or at least try to) after I get some sleep Aug 21 15:18:49 (and after I finish getting of the mud of everything) Aug 21 15:18:54 RP: 1 thread compiling Aug 21 15:18:57 RP: 4 monitoring? Aug 21 15:19:42 zecke : I'll start working on the cluster i a while. I'll install debian is it ok ? Aug 21 15:20:06 zecke: If we're running less that 5 tasks and have a task to run, we run it :) Aug 21 15:20:15 zecke sleeps with ubuntu *g* Aug 21 15:21:41 Ifaistos: I only support Ubuntu Aug 21 15:22:37 woglinde: lol Aug 21 15:23:03 Ubuntu 6.06 LTS if you can't configure debian Aug 21 15:23:45 woglinde: ubuntu is an anicent word and means: Zecke is sick of fighting broken packages in debian Aug 21 15:23:58 woglinde: because he can't configure debian Aug 21 15:24:00 show me broken packages in stable Aug 21 15:24:14 woglinde: I can show you only insecure ones Aug 21 15:24:30 woglinde: and stable has pppd? (but debian can't be blamed for that one) Aug 21 15:24:40 hm I will test xen today I think Aug 21 15:25:21 woglinde: have fun :) Aug 21 15:25:25 Ifaistos: debian is just fine Aug 21 15:25:30 03rpurdie * r575 10bitbake/lib/bb/taskdata.py: taskdata.py: Add handling to remove/ignore files which fail after after the checks earlier in add_[r]provider Aug 21 15:25:40 so lunch now Aug 21 15:26:01 koen: That issue you saw if fixed. It will still print info about the kernels its ignoring but should at least build Aug 21 15:26:20 RP: thanks Aug 21 15:26:22 hi all Aug 21 15:26:35 zecke : OK. will probably have something ready later tonight. (2-3 machines at least to start) Aug 21 15:27:12 Ifaistos: I will go home in about an hour Aug 21 15:27:25 Ifaistos: and I don't know if I go to the gym Aug 21 15:27:33 Ifaistos: in the worst case we will meet tomorrow Aug 21 15:27:52 what kind of cluster do you build? Aug 21 15:28:12 zecke : p4 celerons @ 2.5 Ghz Aug 21 15:28:28 p4? and how many? Aug 21 15:28:38 how do you get rid of the heat? Aug 21 15:29:17 zecke : 10. aircondition :) In the entrace of the local electricity company my name is engraved on marble as one of their best customers Aug 21 15:31:06 Ifaistos: if they don't produce enough heat, you can do some more builds Aug 21 15:34:40 my first cluster (home made) -> http://www.stelioscellar.com/Linux/Helios/index.html Aug 21 15:35:04 ~lart monotone Aug 21 15:35:04 * ibot resizes monotone's terminal to 40x24 Aug 21 15:35:23 my first cluster (home made) -> http://www.stelioscellar.com/Linux/Helios/index.htm Aug 21 15:35:55 can i get a info text of a bitbake target? Aug 21 15:37:35 Ifaistos: nice Aug 21 15:39:15 koen: feedback on the new code is welcome - I know we need to do something with the logging :) Aug 21 15:40:16 RP: 'NOTE: Waiting for 1 active tasks to finish' isn't really helpfull Aug 21 15:40:49 koen: how much is netbook pro worth? Aug 21 15:41:07 hrw|work: E1500 when it came out Aug 21 15:41:24 ridiculously expensive Aug 21 15:41:34 koen: now how much is it worth Aug 21 15:41:38 NOW() Aug 21 15:41:44 hrw|work: I have no idea Aug 21 15:42:03 a few pints at fosdem? Aug 21 15:42:08 a+ Aug 21 15:43:23 koen: someone sells one on Polish auction system Aug 21 15:44:02 there's one on ebay for $100 with 6 days left Aug 21 15:44:06 ~lart monotone badly for broking my working copy Aug 21 15:44:06 * ibot cats /dev/urandom into monotone badly's ear for broking my working copy Aug 21 15:44:14 koen: that one is ~60 eur now Aug 21 15:45:35 mtn update == 'you have missing files'; mtn revert --missing;mtn update == 'there are _MTN/tmp/ already'; rm -r _MTN/tmp; mtn update == 'you have missing files' and story starts again ;( Aug 21 15:46:36 shit.. have to go ;( Aug 21 15:46:37 cu all Aug 21 15:47:01 RP: is it possible to list the tasks that are pending? Aug 21 15:47:30 koen: Other than looking at ps and making educated guesses, no Aug 21 15:47:40 koen: We ideally need a proper UI Aug 21 15:50:38 RP: http://rafb.net/paste/results/FUZk1i50.html Aug 21 15:53:18 koen: and then you get the output from other tasks corrupting the console :) Aug 21 15:53:23 koen: but yes, I agree Aug 21 15:59:04 RP: commit your bitbaked as well please Aug 21 15:59:12 How can I read out the SECTION and DESCRIPTION Variables for a target with bitbake? Aug 21 15:59:32 data.getVar('SECTION', d, True) Aug 21 15:59:32 zecke: bitbaked? Aug 21 15:59:55 The BitBake Daemon? Aug 21 16:00:04 The one that can distribute tasks over the network Aug 21 16:00:48 zecke: I'd have to write it first :-( Aug 21 16:01:27 zecke: how to use data.getVar('SECTION', d, True) Aug 21 16:05:10 mr_nice_slacker: tough question, I do not know what you want to do Aug 21 16:05:29 mr_nice_slacker: e.g. you could look at bittest to see how to load BBFiles and query attributes Aug 21 16:06:44 zecke: i want to sort out http://opensimpad.org/test/test33.php and kill the targets without sense - the list i got over a bitbake -nk world. how to get bittest? Aug 21 16:07:17 from SVN and it is in bitbake's trunk (bitbake_qa) Aug 21 16:07:53 mr_nice_slacker: BTW: http://www.angstrom-distribution.org/repo/repo.php Aug 21 16:08:05 http://www.angstrom-distribution.org/repo/repo.phps Aug 21 16:08:24 zecke: well yes Aug 21 16:08:35 zecke: how did they get the list? Aug 21 16:08:47 good question ;) Aug 21 16:09:02 mr_nice_slacker: take a look at BitTest it is using BBFILES Aug 21 16:09:10 loads every BBFile and then checks the data... Aug 21 16:09:47 zecke: ok - i will download and try it Aug 21 16:18:50 03rpurdie * r576 10bitbake/lib/bb/runqueue.py: runqueue.py: Print summary of active tasks when aborting a build run Aug 21 16:22:02 zecke: do i have to install it in a special way? if i start trunk/bitbake_qa/bin/bittest i get an error File "./bittest", line 42, in ? from bittest import * ImportError: No module named bittest Aug 21 16:22:41 how would I get bitbake to just perform the do_compile step onwards? Aug 21 16:22:49 if I'm modifying the source, say Aug 21 16:22:50 zecke: ah ok found the config :) Aug 21 16:25:17 mikearthur: bitbake xyz -c compile -f Aug 21 16:26:52 * lrg has his headphones on and cant hear mikearthur asking questions Aug 21 16:27:48 RP: wierdly its complaining, for trying to compile linux-openzaurus, about "InvalidTask event exception" in package quilt-native Aug 21 16:28:14 mikearthur: Use -b /somepath/linux-openzaurus.bb Aug 21 16:28:47 ok Aug 21 16:30:20 mikearthur: The new dependency code doesn't have that problem :) Aug 21 16:30:27 :D Aug 21 16:56:45 x29a:ping Aug 21 16:57:02 poing Aug 21 16:57:23 x29a: did you managed to build the other cable? Aug 21 16:57:28 not yet Aug 21 16:57:43 i come home from work at 1700, then eat, now configuring my webspace Aug 21 16:57:50 im not very thrilled to do any jtag Aug 21 16:58:10 x29a: yes Aug 21 17:01:04 x29a: i will try to improve my crosscompiler compiling skills then - just drop me a line that i konw if i have to install the light linux on the sl4. Aug 21 17:01:52 k Aug 21 17:19:33 hm, did anybody know if this book(http://www.oreilly.com/catalog/belinuxsys/) is worth the money (it's from 2003)? Aug 21 17:19:48 -did + does Aug 21 19:26:56 mr_nice_slacker: WAAAH Aug 21 19:26:58 ARGH Aug 21 19:27:07 damn shit crap cable wont work Aug 21 19:27:26 just spent almost 2h soldering and now it wont detect Aug 21 19:31:19 hey something wierd with bitbake svn repo - try this link svn://svn.berlios.de/bitbake/branches/bitbake-1.6 Aug 21 19:31:30 sorry Aug 21 19:31:39 something wrong with my clipboard Aug 21 19:31:51 link = http://svn.berlios.de/viewcvs/bitbake/branches/bitbake-1.6/?rev=576 Aug 21 19:32:26 it says current rev is 570 Aug 21 19:32:51 isnt it supposed to be 576? Aug 21 19:34:02 because when i run svn info on the same url (except svn:// instead of http://) it returns rev as 576 Aug 21 19:37:09 well its almost the same url: svn info svn://svn.berlios.de/bitbake/branches/bitbake-1.6 Aug 21 19:37:40 hi andersee Aug 21 19:38:34 woglinde: howdy Aug 21 19:38:44 * andersee gets off to a late start today Aug 21 19:49:57 ok i think i see whats going on now - svn info is reporting 'revision' as the latest for the whole bitbake repo - whereas 'last changed rev' is the revision that last changed the specified directory (or anything under it) Aug 21 19:52:50 the svn-book documentation doesn't seem to mention this at all - just leaves it to the user to deduce - i think it would be ok if they would just change 'revision' to 'repository rev' - because when i ask for info on a directory i would expect something called 'revision' to be the rev of the directory (not the whole repo) Aug 21 20:00:21 anyone have any idea why the X libraries would be looking in my /usr/lib rather than my oe stuff? Aug 21 20:00:55 mikearthur: because their building is broken somehow perhaps Aug 21 20:03:11 /bin/sh ./arm-linux-libtool --tag=CC --mode=link arm-linux-gcc -I/home/mike/oe/tosa/tmp/openzaurus/staging/arm-linux/include -fexpensive-optimizations -fomit-frame-pointer -frename-registers -O2 -L/home/mike/oe/tosa/tmp/openzaurus/staging/arm-linux/lib -Wl,-rpath-link,/home/mike/oe/tosa/tmp/openzaurus/staging/arm-linux/lib -Wl,-O1 -o libXext.la -rpath /usr/lib -version-number 6:4:1 -no-undefined libXext_la-DPMS.lo libXext_l Aug 21 20:03:11 a-XLbx.lo libXext_la-XMultibuf.lo libXext_la-XSecurity.lo libXext_la-XShape.lo libXext_la-XShm.lo libXext_la-XSync.lo libXext_la-XTestExt1.lo libXext_la-Xcup.lo libXext_la-Xdbe.lo libXext_la-extutil.lo libXext_la-globals.lo -L/home/mike/oe/tosa/tmp/openzaurus/staging/arm-linux/lib -lX11 -ldl Aug 21 20:05:03 any ideas? Aug 21 20:14:50 hey... Aug 21 20:15:11 is the DISTRO_TYPE used for anything specific Aug 21 20:15:28 I see most distro.conf files setting it to either release or debug Aug 21 20:17:05 I mean, does setting it to 'release' actually have any effect on the way things get built? Aug 21 20:35:45 tkp: it is used by some distros and recipes to switch behaviour Aug 21 20:36:08 tkp: switching off kernel messages, forcing to have root password set etc Aug 21 20:39:00 hi hrw - trying to figure out what is in bitbake-data-local branch - do you know? Aug 21 20:39:13 (in teh svn repo) Aug 21 20:39:18 nope Aug 21 20:39:40 use svn log to check Aug 21 20:40:09 k thx Aug 21 20:42:35 hi TheMasterMind1 Aug 21 20:54:24 mr_nice_slacker: worked Aug 21 20:54:27 verify done Aug 21 20:54:33 ill get it over to the serial Aug 21 21:45:52 03rpurdie * r577 10bitbake/lib/bb/taskdata.py: taskdata.py: Fix so only warn about failed providers once. Aug 22 02:00:03 night **** ENDING LOGGING AT Tue Aug 22 02:59:56 2006