**** BEGIN LOGGING AT Thu Feb 14 03:00:00 2013 Feb 14 08:20:44 good morning Feb 14 09:55:41 hi does the core package includes support for tun/tap interface or do we have to add it Feb 14 09:55:49 somewhere like local.conf Feb 14 10:09:02 RP: having "fun" with jabber today Feb 14 10:09:38 rburton: I noticed you were going on/offline a bit... Feb 14 10:10:02 RP: i wasn't even aware i was getting that far Feb 14 10:10:15 niteshnarayanlal: "core package?" You mean is the kernel module included by default? Feb 14 10:10:40 rburton: you did enter/leave a couple of times Feb 14 10:11:04 shocked Feb 14 10:12:51 RP: i'm also taking an early lunch to go swimming with the kids Feb 14 10:15:28 RP I was using bitbake to build the core package Feb 14 10:20:01 Q: who should run /etc/rpm-postinsts at boot/first boot? Feb 14 10:28:56 mcfrisk: there should be an initscript which does that Feb 14 10:37:09 RP: ah, the one created by meta/classes/rootfs_rpm.bbclass. But that is also missing any error handling, like set -e and set -u. Feb 14 10:37:46 mcfrisk: patches to improve it welcome... Feb 14 10:42:11 RP: Ok, I'll try to figure something out. Some packages try to do #!/bin/sh -e but that gets pre-pended by default #!/bin/sh, for example. Feb 14 10:42:24 Hi everybody. I'm building after a "repo sync" but my build dir is growing too much (>100GB). Is there a way to clean old builds of a target but not the latest? Feb 14 10:43:08 panda84kde: there are some scripts in contrib which might help Feb 14 10:47:11 * panda84kde searching Feb 14 10:50:25 RP: poky's contrib? Feb 14 10:51:47 panda84kde: sorry, I thought scripts/contrib but looks like they're in scripts/ Feb 14 10:52:40 panda84kde: "cleanup-workdir" sounds promising for example Feb 14 10:53:27 RP: thanks, will take a look! Feb 14 11:06:49 RP: thank you, it worked Feb 14 11:09:32 panda84kde: great :) Feb 14 13:05:11 what does ${PF} represent? Feb 14 13:05:16 :D Feb 14 13:10:07 I have encountered it in a .bbclass Feb 14 13:16:54 Corneliu: See bitbake.conf Feb 14 13:17:25 PN-PV off the top of my heard Feb 14 13:17:29 head Feb 14 13:18:55 RP: then it would mean package footer? Feb 14 13:19:33 Corneliu: more likely it stood for package file Feb 14 13:19:49 Corneliu: but the name isn't particularly helpful Feb 14 13:20:12 RP: indeed :) Feb 14 13:20:19 Corneliu: I suspect it was created as a lot of sources are called PN-PV.tar.gz Feb 14 13:30:36 otavio: Hi. I'm getting this warning http://pastebin.com/JtZ2xUTZ Feb 14 13:31:05 It appears with your all patches applied from patchwork Feb 14 13:31:16 panda84kde: yes; I got it here as well Feb 14 13:31:35 panda84kde: I am working on this right now Feb 14 13:31:45 otavio: fantastic, thanks! Feb 14 13:40:58 panda84kde: which image you've build? Feb 14 13:41:28 otavio: i've built fsl-image-gui Feb 14 13:43:20 panda84kde: ok Feb 14 13:59:26 hi, I have a 3rd paryt binary I need to add to my rootfs, but when I try my recepie I get 'Architecture did not match (62 to 3) on ' and that is tru its a 32bit bin I need on a 64bit arch. Can I tell my .bb file to ignore this? Feb 14 14:57:52 gm all Feb 14 15:13:35 rp: hi! did anything happen with my sftp patchset? i hope the ball wasn't in my court Feb 14 15:14:39 zibri: no, definitely in mine, sorry :/ Feb 14 15:15:06 ah, no problems. i just wanted to make sure that i hadn't missed anything :) Feb 14 15:16:11 did you see my patch for ConfHandler? any first impressions? (i think i accidently added anchoring for the regexp though...) Feb 14 17:03:11 RP: around? Feb 14 17:03:30 JaMa: kind of ;-) Feb 14 17:04:03 can you quickly scan through https://bugzilla.yoctoproject.org/show_bug.cgi?id=3890 and let me know if I'm missing something obvious? Feb 14 17:04:04 Bug 3890: enhancement, Medium, 1.4, richard.purdie, NEW , @contains does not create vardep automatically Feb 14 18:02:14 JaMa: no, not missing anything obvious. We need code for it. I think I know how to do it though Feb 14 18:02:30 JaMa: will look at it tomorrow Feb 14 18:06:53 RP: thanks! Feb 14 18:55:38 bug 2743 Feb 14 18:55:39 Bug https://bugzilla.yoctoproject.org/show_bug.cgi?id=2743 normal, Medium, 1.1.2, liezhi.yang, ACCEPTED , Useradd command fails after 10 retries with 1.1.2 20120630 build Feb 14 18:55:47 autobuilder is seeing these Feb 14 18:56:40 or maybe 3047 Feb 14 18:56:44 bug 3047 Feb 14 18:56:45 Bug https://bugzilla.yoctoproject.org/show_bug.cgi?id=3047 normal, Medium, 1.3 M4, liezhi.yang, VERIFIED FIXED, do_rootfs: ERROR: tried running useradd command 10 times Feb 14 20:28:38 msm: which autobuilder and which release? Feb 14 20:36:37 RP: nightly ppc has it Feb 14 20:36:49 nightly mips lsb Feb 14 20:37:01 nightly arm Feb 14 20:37:02 few others Feb 14 20:37:09 it says it's master Feb 14 20:39:40 msm: wasn't that a while ago and its built successfully since? Was pseudo upgrade related? Feb 14 20:40:16 RP: we never could the root cause of the useradd groupdadd retrying 10 times Feb 14 20:40:43 mark hatle had a patch that retried many times before failing Feb 14 20:40:49 which papered over the issue Feb 14 20:40:59 msm: in the recent builds it was triggered by pseudo Feb 14 20:41:30 ://gate.crashing.org/~fray/shadow-native-block-groupadd-and-useradd.patch Feb 14 20:41:36 patch: ://gate.crashing.org/~fray/shadow-native-block-groupadd-and-useradd.patch Feb 14 20:41:43 http://gate.crashing.org/~fray/shadow-native-block-groupadd-and-useradd.patch Feb 14 20:42:01 RP: not sure did not look too close Feb 14 20:42:10 RP: are you saying it's already solved? Feb 14 20:42:45 msm: I'm saying that the build failure in the last couple of days like that was "fixed" by the new verison of pseudo in that the builds then started working again Feb 14 20:43:17 msm: There are occsional problems like bug 2743 out there but we've not seen them on the AB for a long time Feb 14 20:43:18 Bug https://bugzilla.yoctoproject.org/show_bug.cgi?id=2743 normal, Medium, 1.1.2, liezhi.yang, ACCEPTED , Useradd command fails after 10 retries with 1.1.2 20120630 build Feb 14 20:43:26 (other than the pseudo induced failures) Feb 14 20:47:59 RP: k, i've not seen it either until just recently Feb 14 20:48:11 RP: i bet if I pull down the latest master w/ pseudo it will go away too Feb 14 20:49:43 msm: highly likely Feb 14 22:03:30 fray: I have interesting bug in rpm pakcage class Feb 14 22:06:16 fray: package_rpm.bbclass: for keyword_die in "unpacking of archive failed" "Cannot find package" "exit 1" ERR Fail Feb 14 22:06:34 now I have a postinst from meta-java which has this string Feb 14 22:06:49 if grep -q "\-server ERROR" $D/${JDK_HOME}/jre/lib/${JDK_ARCH}/jvm.cfg && ! grep -q "KNOWN" $D/${JDK_HOME}/jre/lib/${JDK_ARCH}/jvm.cfg; then Feb 14 22:07:09 and it catches the ERR from that ERRor Feb 14 22:07:12 above Feb 14 22:08:06 and I get | log_check: There were error messages in the logfile Feb 14 22:08:07 | log_check: Matched keyword: [ERR] Feb 14 22:45:52 kergoth: btw, the codeparser "contains" patch is a kind of rfc. Was there a reason we didn't do this originally? Feb 14 22:46:25 kergoth: am also interested in the parsing pool cleanup patch if that is working? Feb 14 22:46:59 working so far, but i'm the only one testing it, so i'm a tad nervous about it :) Feb 14 22:47:04 will take a look at the contains patch Feb 14 22:47:38 kergoth: ask sgw to put it in the next mut Feb 14 22:47:53 kergoth: that will at least throw the autobuilders at it Feb 14 22:48:04 ah, good idea. okay Feb 14 22:50:50 RP: re: contains, I don't think you need to do base_contains, since the dep tracking code can parse that and will see that it calls one of the other contains functions. the only reason we need to manually track the others is it doesn't know how to compile and parse functions imported from modules. the intent was always to either 1) learn to parse those things, or 2) add a way to control the hardcoded list from the metadata, but neither got implemented Feb 14 23:03:34 kergoth: ah, yes. For some reason I thought base_contains was a meta/lib/xxx.py function Feb 14 23:04:07 kergoth: I was also thinking we should phase out lib.oe.contains in favour of the bb.utils one... Feb 14 23:04:36 kergoth: I guess for now we can add this until we do one of the two other things you mention which I agree would be better long term Feb 15 00:23:57 kergoth: RP: cannot we put deprecated warning in 2 contains functions to cleanup metadata at some point? Feb 15 00:58:08 Just sort of an idle question: Am I right to believe that I can reasonably assume any oe-core build system will have sqlite 3.7.x these days? Feb 15 01:19:11 where is the recipe? Feb 15 01:19:56 I' m confused by the question :) Feb 15 01:31:37 Background is: I was looking at a possible performance enhancement for pseudo, and discovered that two of my host systems have an sqlite3 with the adorable quality that the performance enhancement is a catastrophic performance penalty. Feb 15 01:31:53 ah, you men the machines we run oe-core on Feb 15 01:32:18 Yeah. And the WR systems are using their own sqlite which is bundled in oe-core, I think. Feb 15 01:32:35 But I am not entirely sure that this is a reliable or safe assumption for other people who are using oe-core. Feb 15 01:37:44 hmm, there is some check for the host distro and you could check those Feb 15 01:38:37 If I do find this to be worth the hassle, I'll probably make pseudo aware of the known-bad sqlite versions. **** ENDING LOGGING AT Fri Feb 15 02:59:58 2013