**** BEGIN LOGGING AT Sun Aug 02 02:59:57 2009 Aug 02 09:22:59 03Koen Kooi  07org.openembedded.dev * r0131d2ec1d 10openembedded.git/recipes/libtool/libtool_2.2.6a.bb: libtool 2.2.6a: fix libltdl problem, add default pref version so distros don't accidentally pick this one up Aug 02 09:35:51 hey all, I have strange problems while building SHR images. A bunch of packages that should (and always got) installed into the image via RRECOMMENDS do not get installed Aug 02 09:36:16 I'm not 100% sure when this started, but it might have been with the bitbake move from svn to git Aug 02 09:36:37 anybody knows if there were some changes in bitbake which could cause something like that? Aug 02 09:58:03 is there a bitbake recipe for "update all the packages I already built in the repository" ? ie after I do a git merge... Aug 02 10:00:21 ???? Aug 02 10:00:43 depens on the recipe if a newer version is build Aug 02 10:02:41 well, some package I built "manualy" (ie bitbake blahblah; bitbake package-index) -- so it doesn't get rebuilt unless I specify it again ? Aug 02 10:04:18 there is no automatic rebuild Aug 02 10:05:28 is the only way to make a fake "image" and manualy add package I wanted to it, and bitbake that after a merge then ? Aug 02 10:07:12 hm seems I dont understand what you want achieve Aug 02 10:08:30 imagine I have bitbaked "screen" and a bunch of others. I git merge; what bitbake do I do to make sure the stuff I bitbaked gets recompiled... Aug 02 10:08:35 "bitbake update" ? Aug 02 10:08:57 what you merge? Aug 02 10:09:17 what has bitbake to do with merging? Aug 02 10:09:29 probably you mean Aug 02 10:09:37 you have build screen Aug 02 10:09:39 before Aug 02 10:09:40 I merged in "some changes" from the git tree, that could have changed any recipe Aug 02 10:09:46 now you changed the recipe Aug 02 10:09:48 of screen Aug 02 10:09:52 merge your tree Aug 02 10:10:13 and want now a new build of screen? Aug 02 10:10:53 NO I want a new build of ANY PACKAGE i BUILT THAT MIGHT HAVE BEEN UPDATED BY THE MERGE Aug 02 10:11:44 whats complicated here ? "opkg update && opkg upgrade" but for the source tree. "apt-get update && apt-get upgrade" equivalent Aug 02 10:11:59 you mix to many thinks Aug 02 10:12:14 building of changes to recipe is controled via PR Aug 02 10:12:24 if you change something bump PR Aug 02 10:12:35 if the editor forgets to bump the PR Aug 02 10:12:43 the package is not rebuild Aug 02 10:14:06 so, imagine a recipe X changes. If I bitbake /anything/ afterward it will rebuild X ? regardless of the fact that it's not a direct dependency ? Aug 02 10:14:55 again if the editor dont forget to bump PR it will not build Aug 02 10:15:01 args Aug 02 10:15:05 will build Aug 02 10:15:58 So, X changes... I bitbake Y that has /zero/ dependency on X, and X will get rebuilt anyway ? Aug 02 10:16:23 again if X has a bumped PR it will build Aug 02 10:16:37 ah no Aug 02 10:16:45 only when you build image Aug 02 10:16:58 oh build its directly again Aug 02 10:17:18 yeah thats what I feared Aug 02 10:19:12 I'm looking for a "check what is in the already built opkgs against the possibly new recipes bunch, and rebuilt anything that needs it"... Aug 02 10:21:18 hm Aug 02 10:26:39 hm I would check agains stamps Aug 02 10:26:56 but thats need some more modifications to bitbake Aug 02 10:27:25 hi bluelightning Aug 02 10:27:56 hi woglinde Aug 02 10:31:22 woglinde, do you see the advantage of having that tho ? I could bring the whole binary distro up to date Aug 02 10:31:56 buserror hm Aug 02 10:32:12 better rise the topic on the mailinglistr Aug 02 10:34:02 well I asked here because I assumed it was in there already. so much stuff, all hidden :> Aug 02 10:34:26 * BusError discovered the bitbake dev-shell only recently, by chance! Aug 02 10:38:45 BusError: would not a simple bitbake of your image have the side affect of building any bumped PR's for components (giving you fresh IPK's anyway)? Aug 02 10:38:47 ???? Aug 02 10:39:08 DJWillis thats what I said Aug 02 10:39:39 but this dont covers packages you will add to your disbu but not the image Aug 02 10:39:41 woglinde: sorry, scanning the logs, must have missed that. Aug 02 10:40:24 bitbake -k world would be a solution too Aug 02 10:40:36 okay I have to reboot Aug 02 10:40:39 till later Aug 02 11:10:11 hm, trying to do a git pull after 3 weeks of holiday and I get "fatal: cannot pread pack file: No such file or directory" Aug 02 11:10:14 how to resolve this? Aug 02 11:10:39 eFfeM: where is that coming from? Aug 02 11:12:28 zecke, hello Aug 02 11:12:40 zecke, what is the state of the reworked parser branch in bitbake Aug 02 11:13:22 zecke, it comes from git pull Aug 02 11:13:37 immediately after receiving ojbects Aug 02 11:13:40 eFfeM: remote, local "target/destination" Aug 02 11:14:12 + Aug 02 11:14:15 oops Aug 02 11:14:16 mrmoku: what kind of state report do you expect? Aug 02 11:14:22 mrmoku: it is public and GPL Aug 02 11:14:43 zecke, more like if you would recommend to use it instead of 1.8 Aug 02 11:15:26 zecke compelte error log is at http://www.pastebin.ca/1515603 Aug 02 11:15:38 mrmoku: well, the speed up is not massive Aug 02 11:16:11 zecke, but it builds our images correctly (SHR) Aug 02 11:16:31 with 1.8 lots of packages do not get installed into the image Aug 02 11:17:08 mrmoku: well, that comes from something else Aug 02 11:17:24 mrmoku: my parser change should have no functional impact(tm), it is just cleaning up code to allow future work Aug 02 11:17:26 eFfeM, looks like it is a local git problem - you could try to run 'git gc' Aug 02 11:17:35 zecke, hmm, ok Aug 02 11:17:47 mrmoku: will do Aug 02 11:17:50 eFfeM: I would assume your local git is too old? Aug 02 11:19:55 mrmoku: git gc did not cure the problem Aug 02 11:20:08 zecke the git archive or the git app ? Aug 02 11:20:52 how can i repull ? Aug 02 11:22:20 eFfeM: most likely the git app. do you use http/rsync to copy the plain objects? Aug 02 11:22:48 ehh, how can i find out? Aug 02 11:22:55 * eFfeM is really a git n00b Aug 02 11:23:22 eFfeM: what do you pull? git:// git+ssh:// rsync:// http://? Aug 02 11:24:15 no idea, i want to update my local oe repo Aug 02 11:24:26 i always just did git pull Aug 02 11:25:00 will pastebin my config Aug 02 11:25:28 http://pastebin.ca/1515605 Aug 02 11:27:15 eFfeM: ah pulling oe... hmmm Aug 02 11:27:29 eFfeM: you can do git repack -afd on your system and then pull again Aug 02 11:27:45 eFfeM: I have seen git fail because it couldn't mmap every pack Aug 02 11:28:44 will try that Aug 02 11:29:05 (and apologies, I assumed it was obvious that I wanted to pull oe) Aug 02 11:31:04 will report back later Aug 02 11:32:04 oh, repack just finished, git pull after that still gives the same error Aug 02 11:32:06 now really afk Aug 02 11:35:21 eFfeM-lunch: hmm Aug 02 12:03:58 re Aug 02 12:06:31 zecke: should I just try a git clone in the same dir or should I remove the openembedded dir first Aug 02 12:09:42 decided to rename the dir and do a new clone, i'll report the results Aug 02 12:16:14 seems to result in a sound system, now building ... Aug 02 12:16:27 zecke: mrmoku|away: thanks for your suggestions Aug 02 12:26:30 eFfeM: hmm Aug 02 12:46:02 has anyone recently used INHERIT += package_tar? Aug 02 13:52:38 anyone know how to fix this? Aug 02 13:52:39 ERROR: QA Issue with nscd: No GNU_HASH in the elf binary: '/home/frans/oe/tmp_angstrom/work/armv7a-angstrom-linux-gnueabi/glibc-2.9-r32.0/install/nscd/usr/sbin/nscd' Aug 02 13:56:48 i am a little bit hesitating to touch the glibc recipe; would expect this to work as it is very core Aug 02 14:05:54 03Angus Ainslie  07fso/milestone5.5 * ra2b50e2283 10openembedded.git/conf/distro/include/sane-srcrevs.inc: sane-srcrevs : bump efl Aug 02 15:01:01 03Holger Hans Peter Freyther  07org.openembedded.dev * rb9eea9a5d6 10openembedded.git/recipes/meta/meta-toolchain-qte.bb: Aug 02 15:01:01 meta-toolchain-qte: Reorder the tasks to match other recipes Aug 02 15:01:01 Cleanup on the meta-toolchain-qte. Set the variables before Aug 02 15:01:01 rquiring meta-toolchain. Aug 02 15:01:02 03Holger Hans Peter Freyther  07org.openembedded.dev * r6db30a40ce 10openembedded.git/recipes/qt4/qt4.inc: qt4.inc: Fix of shell for some shells Aug 02 15:42:15 morning Aug 02 15:43:30 wrt the glibc priblem I reported 2 hrs back: Aug 02 15:44:07 I was building angstrom for beagleboard and got the problem. after that I tried angstrom for sheevaplug and the problem is not there Aug 02 15:44:15 * eFfeM is flappergasted Aug 02 15:47:09 flabbergasted :-) Aug 02 16:10:04 * * OE Bug 5290 has been created by oe.5.khenry(AT)spamgourmet.com Aug 02 16:10:05 * * devmem2.c has wrong checksum Aug 02 16:10:07 * * http://bugs.openembedded.org/show_bug.cgi?id=5290 Aug 02 16:27:20 eFfeM: did you look at this thread : http://www.nabble.com/No-GNU_HASH-found-in-elf-binary-td23072960.html ? Aug 02 16:27:35 eFfeM: and this solution : http://cgit.openembedded.org/cgit.cgi?url=openembedded/commit/&id=ac6ff675b74d1fb5d21d51f0718dbc6159564b52 ? Aug 02 16:29:10 rkirti: no, thanks for the link Aug 02 16:29:57 but it is odd that it works for sheeva, and it is also odd that i am the first to bump into this as glibc is a very core package Aug 02 16:31:01 eFfeM: hmm..this is while glibc for angstrom right ? Aug 02 16:31:07 yes Aug 02 16:32:02 eFfeM: I dont know much in this case, but I suspect something wrong with your setup. Angstrom doesnt have such errors, esp with beagleboard. I just built one here Aug 02 16:32:12 will dig a little bit deeper after dinner, probably do a full rebuild to make sure it is not some oddity due to upgrades Aug 02 16:32:21 and inconsistent pacakges Aug 02 16:32:36 rkirti: that indeed suggests it is setup related Aug 02 17:52:45 hi im having a problem with bitbake Aug 02 17:52:59 after im doing bitbake nano ( just to test the process ) Aug 02 17:53:07 i have an error on shasum-native package Aug 02 17:53:31 after i google it i found that other people also having the same problem... Aug 02 17:53:39 any quick fix? Aug 02 17:54:08 ~pastebin Aug 02 17:54:09 [~pastebin] A "pastebin" is a web-based service where you can paste anything over 3 lines without flooding the channel. Here are links to a few : http://www.pastebin.com , http://pastebin.ca , http://channels.debian.net/paste , http://paste.lisp.org , http://www.rafb.net/paste Aug 02 17:54:15 Use that to tell us your error Aug 02 17:54:42 ok wait to copy and paste it.. Aug 02 17:55:01 * mwester wagers a doillar he knows this one... Aug 02 17:55:12 http://pastebin.com/d5b12e5d Aug 02 17:55:15 this is it Aug 02 17:55:17 doillar? sigh... Aug 02 17:55:21 I can't spell. Aug 02 17:55:44 dollar? Aug 02 17:56:06 Yeah, that's what I meant. Aug 02 17:56:16 :P Aug 02 17:57:08 any idea about my error? Aug 02 17:57:38 Ok, well that's a start. but that's the output from bitbake -- it contains insufficient information for the general case; you'll note that it points out the compilation log file in the messages it outputs. That log file will tell you what went wrong when it tried to compile that pacakge. Aug 02 17:57:49 this is my problem... Aug 02 17:57:51 the logs are empty :( Aug 02 17:58:12 Ok, you didn't tell us that. Aug 02 17:58:21 all the log.do_compile.$$$$ are empty... Aug 02 17:58:34 its not the first time i setup an openembedded Aug 02 17:58:42 but its the first time i had this problem... Aug 02 17:58:44 That's the signature of the ccache problem. Disable ccache, or place CCACHE = "" in your local.conf file. Aug 02 17:58:52 i also tried that Aug 02 17:58:57 but wait to show your my problem then... Aug 02 17:59:30 CCACHE? Aug 02 17:59:40 what exactly is the problem? Aug 02 18:00:27 ./usr/bin/ccache you mean that binary? Aug 02 18:00:34 compile cache? Aug 02 18:00:40 Heh. Apparently nobody knows, and it's not been important enough for the folks that understand bitbake well enough to look into it. This has plagued me for months now, to the point where I just disable CCACHE on all my build systems, because it's too erratic to sort. Aug 02 18:00:42 the oe cant find that package? Aug 02 18:01:17 Nope, apparently in some situations the ccache program just fails, or does nothing, or some such thing. Nobody knows, there are no logs or messages, and it seems to come and go. Aug 02 18:01:21 Fun, huh? Aug 02 18:01:24 axmm Aug 02 18:01:36 i see Aug 02 18:01:46 so if i put CCACHE = "" i disable that feature? Aug 02 18:01:58 Fedora and Redhat-based systems seem to have this more often, but IIRC others have reported it as well. Aug 02 18:02:07 im using centos Aug 02 18:02:11 so its redhat problem... Aug 02 18:02:12 :p Aug 02 18:02:19 Yep, just CCACHE = "" in your local.conf will do it. And CentOS is a redhat clone, so that's almost certainly the issue. Aug 02 18:02:30 yeap.. Aug 02 18:02:46 hey mdp_ Aug 02 18:02:48 ga Aug 02 18:02:50 strange cause i never had that problem again with centos last time... Aug 02 18:02:54 Well people would like to blame it on RedHat, they sure seem to enjoy that a lot (the OE team seems Debian-centric, I think) -- but in this case it's not restricted to RedHat. ;D Aug 02 18:02:57 even with the same version ( 5.3 ) Aug 02 18:03:04 Yep. Aug 02 18:03:38 When it started happening to me, I couldn't find anything that changed to trigger it, and other systems with the same Fedora version didn't have the problem. Then suddenly they did... Aug 02 18:03:52 yep i can see that from the "fits in with your distribution section" Aug 02 18:04:04 so is completely random :S Aug 02 18:04:27 same problem occured :S Aug 02 18:04:35 Well, it's probably quite deterministic, once we understand what actually triggers it! :D Aug 02 18:05:22 CCACHE = "" Aug 02 18:05:27 yeap i did that... Aug 02 18:05:34 in your local.conf, or whatever other .conf file is parsed by bitbake? Aug 02 18:05:36 now im comfused... Aug 02 18:05:43 in my local.conf.. Aug 02 18:06:00 i will try to reinstall the ccache Aug 02 18:06:23 installed the rpmforge version of ccache Aug 02 18:07:06 in your tmp/work/i686*/shasum*/temp directory (where the log file is), what is the output from: Aug 02 18:07:20 grep sha256.c run.do_compile* Aug 02 18:07:31 0 bytes Aug 02 18:07:33 all the log files.... Aug 02 18:07:51 this was my first check when i saw that error... Aug 02 18:07:52 run, not log Aug 02 18:08:09 and i also found that some recipes files have some errors in the URI Aug 02 18:08:40 i have changed them localy Aug 02 18:08:52 but how can i help OE in their repo to fix that also? Aug 02 18:09:05 i can remember the PHP package at the moment... Aug 02 18:09:19 emails to the mailing list. Aug 02 18:09:32 right... Aug 02 18:09:35 ill do it thanks Aug 02 18:09:39 so what is the output from that grep command? Aug 02 18:12:52 i told you all the log files are empty... Aug 02 18:12:57 0 bytes Aug 02 18:13:16 im doing the reinstall of ccache... Aug 02 18:13:26 i will try both version from rpmforge and centos repo... Aug 02 18:13:42 "maybe" it will solve my problem if ccache is the cause.. Aug 02 18:17:19 the rpmforge didnt solve it... Aug 02 18:17:27 i will try other version... :S Aug 02 18:34:31 max_, I asked about the run.* file, not the log files -- but nevermind, I must run. Aug 02 18:34:47 try running bitbake with the debug or verbose options to get more detail. Aug 02 18:57:24 hi, I've fatal: Couldn't find remote ref master Aug 02 18:57:29 but I used tag= Aug 02 18:57:32 and a good tag Aug 02 18:57:44 I've verified that the tag existed Aug 02 18:57:46 etc... Aug 02 18:58:00 it may be because I'm very tired today Aug 02 18:58:50 I'll try doing something like with svn tag Aug 02 18:59:31 mmm fatal: The remote end hung up unexpectedly Aug 02 19:01:28 does git use ping? Aug 02 19:01:38 because ping don't seem to pass trough the internet Aug 02 19:02:52 ah I had tag twice Aug 02 19:02:57 I seem very tired Aug 02 19:25:03 jo florian Aug 02 20:47:40 so my bud just bought a mybook and its formated FAT32.... so it can only hold files up to 4gb. Now he has a mac, and I got a pc and I want to put some large files on there for him Aug 02 20:47:51 If i format NTFS the mac will get pissed, and I dont think formatting HFS+ is an option on Windows 7 Aug 02 20:48:05 any suggestions Aug 03 01:39:12 qemu-i386: error while loading shared libraries: cannot restore segment prot after reloc: Permission denied Aug 03 01:39:22 any answer on that error?:S Aug 03 01:44:43 I think... Aug 03 01:45:05 think loud please :p Aug 03 01:45:06 12:06 sudo -s Aug 03 01:45:06 12:06 echo -e "vm.vdso_enabled = 0\nvm.mmap_min_addr = 0" > Aug 03 01:45:06 /etc/sysctl.d/60-oe-settings.conf Aug 03 01:45:06 12:06 exit Aug 03 01:45:06 12:06 sudo invoke-rc.d procps start Aug 03 01:45:12 had to /last it :) Aug 03 01:46:29 how can i change with git to stable version Aug 03 01:47:42 http://wiki.openembedded.net/index.php/Stable\ Aug 03 01:48:07 Still going to have to do the thing I said to make qemu work tho, I bet Aug 03 01:48:33 :) thanks Aug 03 01:48:52 :S Aug 03 01:49:24 i had to do all that you paste for my error? Aug 03 01:50:15 everytime i reinstall centos and OE i found a new error... Aug 03 01:54:08 6657 packages in stable version... Aug 03 01:54:14 in dev version are 7003 ... Aug 03 02:03:13 hi,a ll **** ENDING LOGGING AT Mon Aug 03 02:59:57 2009