**** BEGIN LOGGING AT Thu Dec 01 02:59:58 2011 Dec 01 06:43:01 hm, maintenance branch, oct 10 pull: Dec 01 06:43:03 ERROR: Nothing RPROVIDES 'perl-native' (but virtual:native:/home/hudson/jobs/ERC18_2/workspace/org.openembedded.dev/recipes/autoconf/autoconf_2.65.bb, virtual:native:/home/hudson/jobs/ERC18_2/workspace/org.openembedded.dev/recipes/automake/automake_1.11.1.bb RDEPENDS on or otherwise requires it) Dec 01 06:48:58 ah nevermind, user error Dec 01 08:36:17 good morning Dec 01 12:09:11 03Richard Purdie  07master * ref0888f83f 10bitbake.git/lib/bb/build.py: (log message trimmed) Dec 01 12:09:11 build.py: Be determistic about a function's cwd Dec 01 12:09:11 There is a subtle but nasty problem that a function's cwd can vary Dec 01 12:09:11 depending on whether ${B} (often ${S}) exists before the funciton is Dec 01 12:09:11 called or not. Most functions in the system can cope with this but Dec 01 12:09:11 its bad practise and I've just witnessed build failures resulting Dec 01 12:09:12 from this during image generation from bootimg.bbclass. I also Dec 01 12:34:07 What could the cause for kernel not being uncompressed ? (trying to test 3.2-rc2) Dec 01 12:34:43 missing compression support in defconfig? Dec 01 12:35:10 Which ones CONFIG_* ? Dec 01 12:35:30 I had some problems with my deconfing this morning Dec 01 12:36:41 Just grepped for 'COMPRESS' and all those are enabled in .config Dec 01 12:36:47 look at linux.inc there is some mangling of it Dec 01 12:37:05 well you didn't even say if you're using linux.inc or whatever.. Dec 01 12:41:01 Can't find anything related Dec 01 12:43:35 Reflashed the NAND to make sure, and after 'Starting kernel ...' no more output Dec 01 12:43:47 No uncompressing linux or something. Dec 01 12:47:50 Any debug options in u-boot which could help finding me cause? Dec 01 12:48:16 kernel_conf_variable KERNEL_LZMA y Dec 01 12:48:30 but maybe it's used only in linux-kexecboot, sorry for confusion Dec 01 13:21:37 do_install for my first "systemd-recipe" shows app.service as "installed but not shipped in any package". Dec 01 13:22:55 tasslehoff: you maen do_package right? Dec 01 13:23:11 tasslehoff: you have to include it in some FILES_PN-systemd Dec 01 13:23:30 JaMa: I mean do_package :) Dec 01 13:23:58 ah. thanks Dec 01 13:28:36 hi.. when sending a patch for a new recipe, Does it need PR = "r0"? Dec 01 13:28:46 zrafa: no Dec 01 13:28:55 I have read for a while mailing list archives and I am a bit confused if it should go or not Dec 01 13:29:06 JaMa: ah.. okey. Thanks Dec 01 14:30:20 hi ericben Dec 01 14:39:22 ericben: is this change http://build.shr-project.org/tests/jama/qt.diff also expected result of 2e635b4fe6be9d5bfcdd837a523833b7eb0819a0 ? shouldn't we replace those sysroot libdir paths with -L{libdir} ? Dec 01 14:39:44 ericben: ericben just wondering what caused this http://paste.pocoo.org/show/515331/ Dec 01 14:40:45 also there was some issues with already stripped binaries in qt4-x11 Dec 01 14:40:52 if I remember well Dec 01 14:42:16 in this case adding -lX11 manually to aurora-daemon was enough to fix it but I'm not sure what change triggered it Dec 01 14:49:57 I think that aurora now uses X11 tough Dec 01 14:51:26 JaMa: interesting I had a similar error under angstrom with mplayer Dec 01 14:56:29 JaMa: http://www.pastie.org/2949791 Dec 01 14:56:37 so I guess it may be a .pc or something like that Dec 01 14:56:57 because adding -lx11 to every recipe that fails isn't that scalable Dec 01 15:29:11 hi gnutoo hi JaMa Dec 01 15:29:45 JaMa: you mean yous til get -L/OE/shr-core/tmp/sysroots/nokia900/usr/lib -L/usr/lib -L/OE/shr-core/tmp/sysroots/nokia900/usr/lib in the .pc files of QT ? Dec 01 15:30:23 ericben: sorry the diff is vice-versa.. later (nokia900 is old version) Dec 01 15:30:40 ah ok Dec 01 15:30:43 I understand Dec 01 15:31:19 maybe -L{libdir} would be fine here. I looked at some other .pc and not all have ${libdir} so I assumed this was not always necessary Dec 01 15:31:21 ericben: but ie in http://git.openembedded.org/meta-openembedded/commit/?id=dcf924eb24aee888323727a64c33e57317f4d9c7 I was trying to keep at least what's left after prefix Dec 01 15:31:34 ericben: in case there is something still needed there Dec 01 15:31:36 hmm, can I somehow pass an alternative opkg.conf file to opkg when generating the rootfs? i.e. so that one verison of opkg.conf is installed, but a different version used during rootfs generation Dec 01 15:31:37 ? Dec 01 15:32:40 JaMa: are you sure http://paste.pocoo.org/show/515331/ has something to see with this ? Dec 01 15:33:53 JaMa: can you try adding -L{libdir} manually to qtdbus.pc and see if that solves the problem ? Dec 01 15:39:02 not really it can be anything changed in last 2 days Dec 01 15:54:41 gnutoo: hey.. you there? Dec 01 15:58:31 Koem? you there?. Guys : http://thread.gmane.org/gmane.comp.handhelds.openembedded/49240 . Koen says the patch has style issues. Which kind of problems you see? Dec 01 15:58:35 I would like to fix Dec 01 15:59:56 zrafa: hi yes Dec 01 16:00:28 gnutoo: Hey. could you check this? http://thread.gmane.org/gmane.comp.handhelds.openembedded/49240 Dec 01 16:00:48 gnutoo: I would like to know which are the style issues problems Dec 01 16:00:54 zrafa: DEPENDS_${PN} is wrong Dec 01 16:01:00 not only style of it :) Dec 01 16:02:12 JaMa: :S .. Is that the only error you see? Dec 01 16:02:21 or are there more? :) Dec 01 16:02:32 zrafa: for the rest try meta-openembedded/contrib/oe-stylize.py Dec 01 16:03:21 JaMa: ah.. I did not know that. Let me try Dec 01 16:08:06 JaMa: http://paste.debian.net/147711/ Dec 01 16:08:26 JaMa: would be that output okey to submit as patch? Dec 01 16:10:50 is nodejs-native building for anyone? Dec 01 16:12:38 zrafa: give me one second Dec 01 16:13:40 zrafa: is it for oe-dev or core based openembedded? Dec 01 16:13:55 because if it's for the later it at least lacks the LIC_FILES_CHKSUM Dec 01 16:14:08 and koen is known for not beeing very precise Dec 01 16:14:42 also the license has to be checked(for instance GPLv2, GPLv2+ etc...) Dec 01 16:16:32 zrafa: better for sure, please resend and you'll see Dec 01 16:16:49 gnutoo: the license in COPYING file in sources says GPL version 2 Dec 01 16:16:52 and yes add LIC_FILES_CHKSUM Dec 01 16:17:23 if it says "or newer" then it's 2+ (but I haven't checked COPYING Dec 01 16:17:45 but with this path it looks like OE-classic Dec 01 16:17:53 JaMa: gnutoo : the main.c says "GPL version 2, or in your option , later version Dec 01 16:18:10 I will add GPLv2+ Dec 01 16:18:41 gnutoo: JaMa : it is for OE-classic. Should I add that LIC_FILES_CHKSUM as well? Dec 01 16:20:35 zrafa: you can even when it's mandatory, but will make migration to oe-core oe meta-oe easier Dec 01 16:22:54 ok Dec 01 16:33:51 JaMa: if I add LIC_FILES_CHKSUM oe-stylize.py says at the beggining "## Warning: unknown variable/routine "LIC_FILES_CHKSUM = "file://COPYING;md5=cbbd794e2a0a289b9dfcc9f513d1996e"" Dec 01 16:34:11 add it just after LICENSE Dec 01 16:34:52 okey.. let me put that there Dec 01 16:41:40 JaMa: gnutoo : sorry.. last question... Dec 01 16:41:44 COPYING file says : Dec 01 16:41:46 " GNU GENERAL PUBLIC LICENSE Dec 01 16:41:46 Version 2, June 1991 Dec 01 16:41:47 " Dec 01 16:41:52 but main.c file says : Dec 01 16:42:03 * This program is free software; you can redistribute it and/or modify * it under the terms of the GNU General Public License as published by * the Free Software Foundation; either version 2 of the License, or * (at your option) any later version. Dec 01 16:42:03 main.c is the one you should listen to Dec 01 16:42:29 gnutoo: main.c says : "under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or * (at your option) any later version" Dec 01 16:42:47 gnutoo: with that ^ .. should I put GPLv2+ ? Dec 01 16:48:30 yes Dec 01 16:48:32 gplv2+ Dec 01 16:48:41 basically COPYING is just a copy if the gplv2 license Dec 01 16:48:52 as required by the GPLv2 Dec 01 16:57:26 gnutoo: JaMa : The PATCH will be v2 now. SHould I write on body mail which are the differences against PATCH v1? .. SOmething like : "Fix style issues" ? Dec 01 16:58:11 yes but the fix style issue should not go into the commit message but outside of it Dec 01 16:58:23 I don't remember how to do it tough Dec 01 16:59:57 gnutoo: then I will not add that in the commit message. Dec 01 17:02:11 gnutoo: JaMa : well. Please. If you can check the v2 of this patch : http://paste.ubuntu.com/756143/ Dec 01 17:02:19 then I would send after your okey :) Dec 01 17:03:48 I haven't build it but looks ok Dec 01 17:04:56 it builds fine in my local repo. I use directvnc for ep93xx MACHINE, and angstrom distro. The ARCH is arm Dec 01 17:05:32 JaMa: okey.. thanks. I will send :) Dec 01 17:05:39 gnutoo: any comments? Dec 01 17:06:03 bbs Dec 01 17:16:18 zrafa: where does it take the source from? Dec 01 17:16:28 ah sorry Dec 01 17:16:37 looks ok Dec 01 17:16:38 I must go Dec 01 17:16:50 gnutoo: JaMa : thanks.. I am sending then Dec 01 17:17:07 I guess it works and you already tested it Dec 01 17:18:13 gnutoo: Yes.. I am using the Encore entc100 thin client with my OE distro built. As thin client. ANd that directvnc is the vnc client it uses :) Dec 01 17:19:02 gnutoo: also latest Linux kernel :) (that Encore entc-1000 thin client comes from factory with an old linux 2.4 version. We hacked it to have an openembedded distro and latest linux kernel Dec 01 17:19:44 ok nice Dec 01 17:20:54 bye Dec 01 17:23:17 is there any interest in having a host prepare script in oe-core? Dec 01 17:24:04 https://gist.github.com/1418337 Dec 01 17:41:38 pixman-0.24 fails to build here Dec 01 17:42:04 how? Dec 01 17:43:06 JaMa: it seems it lacks a depends on zlib Dec 01 17:44:05 JaMa: building it again for test Dec 01 17:44:46 msm: seems more logical to have a package that does it Dec 01 17:45:00 JaMa: yes Dec 01 17:45:09 JaMa: with DEPENDS += "zlib" it works Dec 01 17:46:23 otavio: what do you mean by package here? Dec 01 17:46:34 some bbclass? Dec 01 17:46:42 msm: a binary package that you install on fedora Dec 01 17:46:54 msm: or on debian Dec 01 17:47:18 well, why? if these things are common available on the host distro? Dec 01 17:48:23 you can easily add new depends and like Dec 01 17:48:45 JaMa: will you send a patch to pixman? Dec 01 17:49:27 why me? ;) Dec 01 17:50:14 does it link to zlib for you or isn't it used from shlibs? Dec 01 17:50:22 here it isn't SHR root@gjama ~ $ opkg info libpixman-1-0 | grep Depends Dec 01 17:50:22 Depends: libgcc1 (>= 4.6.2+svnr181430), libc6 (>= 2.14) Dec 01 17:51:16 JaMa: it doesn't Dec 01 17:51:28 JaMa: but it fails if it can't find zlib.h Dec 01 17:52:04 ok Dec 01 17:53:49 JaMa: can you try to reproduce it? Dec 01 17:54:58 not now Dec 01 17:57:37 otavio: weird is that there is no zlib.h in pixman sources Dec 01 17:57:55 otavio: so the depends is maybe missing in some pixman dependency Dec 01 18:04:04 JaMa: might eb Dec 01 18:04:08 JaMa: might be Dec 01 18:04:09 hum Dec 01 18:07:23 ug, libtool: install: error: relink `libnetsnmpagent.la' with the above command before installing it Dec 01 18:07:57 because /usr/lib/libm.so: file not recognized: File format not recognized Dec 01 18:09:12 Should I pass LDFLAGS=${BUILD_LDFLAGS} on the make install to fix this? Dec 01 20:51:14 koen: just tested sgx/linux 3.0 on 37XX and works fine Dec 01 20:51:29 so now I need to figure out why it is broken on 35XX Dec 01 20:54:16 koen: I suspect something is broken in the revision based fixup logic in the init Dec 01 20:55:23 sakoman wrong channel Dec 01 20:55:28 koen wasnt here for years Dec 01 20:55:40 woglinde_: right you are! Dec 01 20:55:43 khem: ping Dec 01 20:55:45 sakoman but sounds easy to fix Dec 01 20:56:06 woglinde_: hopefully so :-) Dec 01 20:56:16 I'm surprised no one noticed for this long! Dec 01 20:56:33 nobody uses older beagles anymore Dec 01 20:56:39 shows how much people use sgx on omap3 I guess :-) Dec 01 20:56:41 or at least with 3.x kernel Dec 01 20:57:30 right Dec 01 20:58:02 03Rafael Ignacio Zurita  07master * ra4f35de0f6 10openembedded.git/recipes/directvnc/directvnc_0.7.7.bb: Dec 01 20:58:03 directvnc: Add initial recipe with version 0.7.7 Dec 01 20:58:03 DirectVNC provides a very thin VNC client for unix framebuffer systems. Dec 01 20:58:03 Signed-off-by: Rafael Ignacio Zurita Dec 01 20:58:03 Build tested with `angstrom-2010` and `minimal{,-uclibc}` for `MACHINE = "beagleboard"`. Dec 01 20:58:03 Signed-off-by: Paul Menzel Dec 01 21:19:29 JaMa: yes Dec 01 21:20:36 khem: latest binutils in oe-core (2.22) have " --copy-dt-needed-entries is no longer enabled by default" can it cause errors like this? http://paste.pocoo.org/show/515511/ Dec 01 21:20:58 khem: I guess if answer is yes then correct fix is to add -lX11 to corresponding .pc Dec 01 21:26:31 JaMa: yes Dec 01 21:26:50 JaMa: btw. this is a basic behavior change in binutils Dec 01 21:27:11 we should read such changes and foreward people before such upgrades Dec 01 21:27:29 s/foreward/fore warn/ Dec 01 21:28:09 whoever provided the binutils 2.22 upgrade patch should have noticed it IMO Dec 01 21:28:18 khem: for now I know about nmon, mplayer, aurora-daemon failing Dec 01 21:29:08 khem: but the harder task is to find where it should be added, because work arrounding it by adding ie -lX11 in "target" build like in aurora-daemon is probably not right way to fix it Dec 01 21:29:31 actually thats the right way Dec 01 21:29:41 if you need it then you specify it on cmdline Dec 01 21:30:00 khem: I mean if something in Qt4 is depending on X11 then it should be stated in Qt4 .pc file no? Dec 01 21:30:01 but its a feature change that we have to document with bold letters Dec 01 21:30:21 since it will cause question for folks moving from older version of ld to newer 2.22+ Dec 01 21:30:39 could you please write e-mail to oe-core about it as you're for sure right person to report it? Dec 01 21:30:43 JaMa: yes as I said whoever needs it that package should specify it Dec 01 21:32:30 JaMa: ok. I think we should let such patches sit around for a while before cosuming them Dec 01 21:32:38 especially with toolchain Dec 01 21:32:40 so to be clear "first package needing it should specify it", because in aurora case x11events.o is not part of aurora but some dependency of aurora Dec 01 21:32:54 yes then that dep should ask for it Dec 01 21:33:02 and provide it Dec 01 21:34:37 ok, thanks Dec 01 22:30:00 hmm, I think we may be about to be slashdotted Dec 01 22:30:48 people still read slashdot? huh Dec 01 22:30:50 :) Dec 01 22:31:02 (article about gta04) Dec 01 22:31:15 well I don't know about people, I do... Dec 01 22:32:54 bluelightning, you're involved in the gta04? Dec 01 22:33:07 no, not at all... Dec 01 22:33:15 ok Dec 01 22:33:32 if it's any good I might get one to go in my old neo1971 case assuming it will fit Dec 01 22:34:12 ok Dec 01 22:34:24 I don't know exactly when mine will arrive btw Dec 01 22:35:00 btw I tried opie image on bug20 Dec 01 22:35:18 and it was nearly unusable(not related to opie but to the image itself) : Dec 01 22:35:33 * couldn't boot since no kenrel was present in /boot/ Dec 01 22:36:06 * couldn't log in with no password( but I guess it's set by opie right after) Dec 01 22:36:31 I guess I'll have to look into it Dec 01 22:36:53 I guess it's MACHINE_EXTRA_RRECOMMANDS that didn't made it into the image Dec 01 22:36:59 and similar stuff Dec 01 22:47:15 Hi there, I'm trying to compile the core-image-minimal of oe-core using an external toolchain (csl arm-2011.03). Therefore I've created external-csl *inc and *bb files based on the existent ones. Currently the do_package task is failing with: Dec 01 22:47:15 | ERROR: QA run found fatal errors. Please consider fixing them. Dec 01 22:47:15 | ERROR: Function 'do_package_qa' failed Dec 01 22:47:15 NOTE: package external-csl-toolchain-arm-2011.03-r0: task do_package: Failed Dec 01 22:47:23 This is the corresponding "do_package"-log: http://paste.ubuntu.com/756511/ Dec 01 22:48:01 I guess I'm missing some "FILES_"-directives or maybe some "include" directive. Any ideas? Dec 01 23:16:55 03Richard Purdie  07master * rbd5a727c84 10bitbake.git/lib/bb/cooker.py: Dec 01 23:16:55 cooker.py: Allow the -e option to work with virtual classes and -b Dec 01 23:16:55 Using bitbake -e -b virtual:xxxx:/path/to/the.bb would result in Dec 01 23:16:55 zero matches since the virtual:xxxx piece wasn't being processed. Dec 01 23:16:55 This adds in the necessary functionality to handle it correctly. Dec 01 23:16:56 [YOCTO #1793] Dec 01 23:16:57 Signed-off-by: Richard Purdie Dec 01 23:23:04 kenws: the problem is the .debug directory in the wrong package Dec 01 23:23:37 kenws: although there are some other worrying messages in there that is the one that is giving you the failure Dec 01 23:35:26 khem: thanks for approving the account Dec 01 23:51:28 * wmat catches up on the Documentation thread on oe-devel Dec 01 23:57:27 wmat: if people could only spend as much time and energy on documentation :) Dec 01 23:58:43 this is an interesting thread and touches on many things I've encountered while working on elinux.org, i.e. editing wiki pages in vim/emacs/whatever, storing wiki changes in a DVCS, etc. etc. Dec 02 00:00:04 yeah, funny how everyone shares some of the same problems Dec 02 00:05:22 Crofton|irssi: I offerred to fix the wiki many, many months ago so that we could re-enable the ability for new users to edit Dec 02 00:07:11 personally I'd still prefer it if we could do that Dec 02 00:07:21 it's the way wikis are meant to be Dec 02 00:07:55 i presume that it is the way it is because the wiki got hammered by spammers at some point? Dec 02 00:08:22 I dealt with the same issue for elinux.org and it seems to be working presently Dec 02 00:08:42 it's just a matter of installing a few mediawiki plugins Dec 02 00:10:35 wmat: I believe so, according to tom there was spamming even with the captcha Dec 02 00:16:12 yeah, it definitely happens Dec 02 00:17:00 we've been pretty successful keeping them off of elinux.org with Captchas, as well as a few other extensions Dec 02 00:18:11 wmat: can you talk with ka6sox about this Dec 02 00:18:27 bluelightning: I owe you some qt + camke stuff also Dec 02 00:18:40 at a conference this week, so no time get stuff in order Dec 02 00:18:50 Crofton|irssi: will do Dec 02 00:18:56 thanks Dec 02 00:18:58 wmat, pong Dec 02 00:19:17 Crofton|irssi: I thought I owed you some stuff on that... sorry I really wanted to help you more directly with this stuff, I'm glad you've been able to make some progress though Dec 02 00:19:17 * Crofton|irssi is going for beer to decompress Dec 02 00:19:33 I think I have a handle in the issues Dec 02 00:19:42 with some fixes and some stuff we need to thinkn about Dec 02 00:19:49 ka6sox: was just reading the Documentation/wiki thread and noted that I can help with the MW config so that people can edit again Dec 02 00:21:37 wmat, kewl...let me know how I can enable you to do that or you can tell me how :D Dec 02 00:25:21 ka6sox: i'd need ssh access to the box the MW install is on, as well as being added to the Administrator group in MW Dec 02 00:26:12 wmat, username on the wiki? Dec 02 00:26:22 ka6sox: wmat Dec 02 00:27:14 done Dec 02 00:27:18 try now? Dec 02 00:27:48 great, thx Dec 02 00:31:50 wmat, see pm Dec 02 02:35:58 wmat, you still about? Dec 02 02:42:03 ka6sox: yep, for a bit longer Dec 02 02:45:38 wmat, working your login now. Dec 02 02:47:10 wmat, did you get pm? **** ENDING LOGGING AT Fri Dec 02 02:59:57 2011