**** BEGIN LOGGING AT Tue Sep 06 02:59:56 2011 Sep 06 05:32:28 ka6sox: whats up Sep 06 05:33:45 did you setup cgit on garnet? Sep 06 05:37:42 yes Sep 06 05:39:14 can you add meta-opie and meta-handheld to cgit's config? Sep 06 05:41:59 ka6sox: sure Sep 06 05:42:13 thanks bluelightning will be happy. Sep 06 05:46:07 done Sep 06 05:46:13 check it out Sep 06 05:46:19 should be functional Sep 06 05:46:28 * khem -> sleep() Sep 06 05:52:58 thanks! Sep 06 05:52:59 gn Sep 06 06:07:25 good morning Sep 06 07:12:39 hi all Sep 06 07:12:56 need a little clarification Sep 06 07:13:58 where DISTUTILS_BUILD_ARGS and DISTUTILS_INSTALL_ARGS variable is being used ..... i get the environment of the recipe but could not find the place where these varaible is being used Sep 06 07:20:45 no worries .... got it it is in distutils.bbclass Sep 06 07:20:51 hi ka6sox Sep 06 07:21:05 morning GNUtoo|work Sep 06 07:21:23 I've an issue with patchwork: I tried to create an account on it and I got a traceball/traceback or something with a similar name Sep 06 07:21:34 then I cannot: Sep 06 07:21:44 *create a new account(mail already taken) Sep 06 07:21:52 *login(account inactive) Sep 06 07:22:33 the mail is denis@eukrea.com Sep 06 07:23:06 and I didn't receive a confirmation mail either... Sep 06 07:26:16 GNUtoo|work: new job ? ;-) Sep 06 07:28:10 yes Sep 06 07:29:36 migrating a company(eukrea) recipes from oe.dev to oe-core Sep 06 07:53:54 good morning .. Sep 06 07:59:01 hmm keymaps building failed when i try to compile for MACHINE=spitz Sep 06 08:00:04 missing keymap files. anyone an idea how to convince oe-core to use te keymap-spitz from oe-classic ? Sep 06 08:02:10 pwgen: create keymap_1.0.bbappend Sep 06 08:02:56 pwgen: something like http://git.shr-project.org/git/?p=meta-smartphone.git;a=commit;h=c4d06016870a163fc81eb8abf1340b7050f9fc5b Sep 06 08:04:58 ping PaulePanter Sep 06 08:05:23 Noor: What’s up? Sep 06 08:05:29 HI Sep 06 08:05:41 just need little help on python-lxml Sep 06 08:06:33 I will see what I can do. But you have seen that I do not know much as Michael pointed out. ;-) Sep 06 08:06:34 JaMa|Off: any reason to keep tzdata/codes 2011g in meta-oe? Is even unfetchable... Sep 06 08:07:20 PaulePanter: Ok ...... just wanna ask about that :) Sep 06 08:07:54 PaulePanter: how can I confirm the correct behaviour ..... any logs anything>? Sep 06 08:08:24 JaMa: *G* seems to work . please give me a tip how to summit this change to meta-smartphone Sep 06 08:09:25 PaulePanter: althoough xml2-config is present in STAGING_BINDIR_NATIVE path Sep 06 08:10:10 PaulePanter: and there is only few differences between Native and cross xml2-config scripe Sep 06 08:10:37 ant_work: iirc no if oe-core version was updated.. Sep 06 08:10:50 pwgen: pls send it to bluelightning, for meta-handeld Sep 06 08:11:12 JaMa|Off: this makes my build oe-core+meta-oeùmeta-handheld fail (core-image-sato) Sep 06 08:12:11 pwgen: meta-zaurus was dropped today from meta-smartphone (I haven't pushed it yet just because I have other changes on top of it I have to test more) Sep 06 08:12:33 pwgen: and tips how to submit changes are written in README files :) Sep 06 08:13:05 JaMa|Off: at first sight we did not forget any bit...all is in meta-hh now. Thx for your time&support Sep 06 08:14:57 ka6sox, who should I talk to for the patchwork issue? Sep 06 08:15:27 GNUtoo|work, which issue? Sep 06 08:15:44 I've an issue with patchwork: I tried to create an account on it and I got a traceball/traceback or something with a similar name Sep 06 08:16:13 then no way to recreate the same account(user+mail already taken) or to login(account inactive) Sep 06 08:16:24 and I didn't receive a confirmation mail Sep 06 08:16:48 okay, that baby is something khem worked on...I'll get with him about that. Sep 06 08:16:58 ok thanks Sep 06 08:17:06 got this from python-lxml log.do_compile "warning: library search path "/usr/lib" is unsafe for cross-compilation" ..... it seems Michael was correct ..... need to check with STAGING_BINDIR_CROSS Sep 06 08:17:07 the mail is denis@eukrea.com Sep 06 08:23:49 * JaMa|Off bisecting bitbake to see where were RDEPENDS broken :/ Sep 06 08:24:18 Noor: Good find. I could not spot that in the build with oe.dev. Sep 06 08:25:28 Noor: I would reply to Michael to the list though that you just imported it from oe.dev and that he should clarify if this was a mistake that he chose this variable. Sep 06 08:33:02 PaulePanter: OK sure ...... Sep 06 08:38:49 PaulePanter: BTW same message is comming after converting NATIVE to CROSS :( Sep 06 08:39:11 will update you on it Sep 06 08:39:16 hmm NOTABUG :/ http://bugzilla.pokylinux.org/show_bug.cgi?id=813 Sep 06 08:48:05 Noor: What are the expanded expressions added in `--with-xslt-config=` and so on? Sep 06 09:15:19 JaMa|Off: have you noticed some file/magic mismatch building on Gentoo? It looks like /meta/recipes-devtools/file is an outdated 5.04. The version of magic installed on buildhost is 8 while the one in oe-core is still 7. This causes errors (see log.do_package) Sep 06 09:16:31 error: magic_load(ms, /usr/share/misc/magic) failed: File 5.4 supports only version 7 magic files. `/usr/share/misc/magic.mgc' is version 8 Sep 06 09:18:04 question: is it ok for this native recipe to peek on host libs? Sep 06 09:26:47 haven't noticed it yet, but iirc there was long discussion and some patches for it in old oe.dev Sep 06 09:27:09 and even -native recipe should provide own magic.mgc and use only that Sep 06 09:46:26 ant_work: btw I have 3.1.0-rc5 recipes for meta-handheld.. but cannot test it on target now :/ Sep 06 09:56:10 PaulePanter: --with-xslt-config='/home/nahsan/SB/sb_oe_core/build-eglibc/sysroots/i686-linux/usr/bin/pkg-config libxslt --with-xml2-config='/home/nahsan/SB/sb_oe_core/build-eglibc/sysroots/i686-linux/usr/bin/xml2-config with NATIVE Sep 06 09:57:04 PaulePanter: xml2-config is coming from libxml2-native package in this case package Sep 06 10:02:31 Although python-lxml does not depend on it directly but it depends on it indirectly 'python-lxml'-> 'rpm-native'-> 'gettext-native'-> 'libxml2-native Sep 06 10:03:32 PaulePanter: xm2-config is also coming from libxml2 recipe one which python-lxml directl depends Sep 06 10:03:56 PaulePanter: and that is present at CROSS variable Sep 06 10:05:01 Thank you. And what does `/home/nahsan/SB/sb_oe_core/build-eglibc/sysroots/i686-linux/usr/bin/pkg-config libxslt` return? Sep 06 10:06:06 With the correct environment variables set of course. Sep 06 10:06:18 PaulePanter: let me check Sep 06 10:10:45 PaulePanter: I manually run this command but it did not show anything or did not create any file Sep 06 10:13:26 Noor: I see. But I now think this is expected. Later on in `setup.py`(?) additional options like `--libs` should be passed which returns the necessary things. Sep 06 10:22:13 PaulePanter: BTW the problem is with xml2-config not with pkg-config Sep 06 10:22:36 for pkg-config we should have STAGING_BINDIR_NATIVE Sep 06 10:23:22 with native the --with-xml2-config is expanded to --with-xml2-config='/home/nahsan/SB/sb_oe_core/build-eglibc/sysroots/i686-linux/usr/bin/xml2-config Sep 06 10:23:46 and with NATIVE it is expnaded to --with-xml2-config='/home/nahsan/SB/sb_oe_core/build-eglibc/sysroots/qemuppc/usr/bin/crossscripts/xml2-config Sep 06 10:35:59 PaulePanter: and the diff between these two xml2-config scripts is http://pastebin.com/gbZG30DN Sep 06 10:44:57 Noor: So the one in `sysroots/i686-linux` should be used, should not it? But why do you get still the same error? Sep 06 10:46:53 Maybe Michael knows more about this. Just ask on the list. Sep 06 10:46:56 Good luck! Sep 06 10:47:31 PaulePanter: yeah and that is currently being used Sep 06 10:47:49 with STAGING_BINDIR_NATIVE varaible Sep 06 10:58:45 hi Sep 06 11:00:14 PaulePanter: I was wrong on getting the same error message ...... It was on different configuration and that is why I was getting the warning message Sep 06 11:01:28 PaulePanter: I think I got the fix ..... with STAGING_BINDIR_NATIVE usage we get these in log.do_compile Sep 06 11:02:57 and -L/home/nahsan/SB/sb_oe_core/build-eglibc/sysroots/i686-linux/usr/lib -L/home/nahsan/SB/sb_oe_core/build-eglibc/sysroots/qemuppc/usr/lib is given at the compilation time Sep 06 11:03:28 so we get this message in the log /home/nahsan/SB/sb_oe_core/build-eglibc/sysroots/i686-linux/usr/libexec/ppc603e-oe-linux/gcc/powerpc-oe-linux/4.6.1/ld: skipping incompatible /home/nahsan/SB/sb_oe_core/build-eglibc/sysroots/i686-linux/usr/lib/libxml2.so when searching for -lxml2 Sep 06 11:04:55 when we use STAGING_BINDIR_CROSS only this option -L/home/nahsan/SB/sb_oe_core/build-eglibc/sysroots/qemuppc/usr/lib is given while compilation and we dont get any messages in log files Sep 06 11:38:07 Noor: That looks good. Good work for figuring that out. Make sure to note that in the commit message and that this warning did not show up in oe.dev. Sep 06 12:23:54 while I am compiling perl I am getting this error http://pastebin.com/AWpi7i5K Sep 06 12:26:17 I am doing bitbake canadian-sdk, I am build stopped at native-perl package, The log is in http://pastebin.com/AWpi7i5K Sep 06 12:36:30 few days back I saw a thread on PR = "r0" thing and last mail said that we will move this discussion to oe-core mailing list but I think it never started there Sep 06 12:36:38 any updates on that Sep 06 12:45:02 JaMa|Off: sounds good Sep 06 12:47:30 JaMa|Off: about linux-kexecboot, probably the output of savedefconfig is very similar if not equal comparing 2.6.39 and 3.1 Sep 06 12:49:13 ant_work: yes it similar but not equal Sep 06 12:49:47 ant_work: you would loose rtc if you do just make oldconfig and usb_pxa270 is no longer under gadget Sep 06 12:50:00 ok, remember you can just run savedefconfig w/out building the kernels for the other machines Sep 06 12:50:14 so, if you have time, pls do it for all Z Sep 06 12:50:16 ant_work: you can check diff if you apply my patches.. Sep 06 12:50:28 I did only spitz Sep 06 12:50:40 thus spitz and akita ;) Sep 06 12:51:08 you can even test images if you want ;) http://jama.dyndns-home.com/org.openembedded.shr-core.images/spitz/ Sep 06 12:51:33 ok, good to know Sep 06 12:58:25 JaMa|Off: time for testing klibc-1.5.25 ? I have a doubt about dependencies Sep 06 13:01:30 git log doesn't suggest changed dependencies, but I did only quick scan Sep 06 13:06:55 the recipe has changed Sep 06 13:07:00 I'll commit in two step Sep 06 13:07:02 s Sep 06 13:09:56 I sent you the squashed one-patch Sep 06 13:11:02 does anyone have experience with gensi-usa ? Sep 06 13:11:27 they provide the efika MX arm driven devices Sep 06 13:15:11 con someone provide me with tzcode2011g.tar.gz ? Sep 06 13:15:43 ant_work: why not send it to meta-oe or even oe-core for more people to comment on it? Sep 06 13:16:09 will be sent in two steps this evening Sep 06 13:16:15 you're beta-tester ;) Sep 06 13:17:01 ant_work: I'm leaving in few mins so sorry won't test soon.. Sep 06 13:17:38 enjoy ! Sep 06 13:17:46 ant_work: and the patch is not very usefull to review anyways.. Sep 06 13:18:19 I suspect th eutils need a dependency on klibc Sep 06 13:18:24 ant_work: try something like "git format-patch -s -1 -M -M -C -C -C" to detect renames better Sep 06 13:18:30 but as it is, those can be built Sep 06 13:18:35 alone Sep 06 13:19:33 JaMa|Off: I'll change the 1.5.24 then I'll update to 1.5.25 in two steps Sep 06 13:20:30 it was too late to do more work on it, I just tested it can build Sep 06 13:20:49 (but the build was interrupted by tzdata, eh) Sep 06 13:25:26 Hi. What's the variable `T' for? Is it the new TMPDIR or something? Sep 06 13:27:57 Or maybe a per package location of the temporary directory? Sep 06 13:28:40 one ob the best hidden secrets :) Sep 06 13:30:00 I'd say the bitbake `d' variable is more hidden. :-) Sep 06 13:30:38 T = "${WORKDIR}/temp" Sep 06 13:31:50 ant_work: yeah, I found that. I'd like to know what it is used for. On my env it points to a directory that doesn't exist (when I run "bitbake -e"). Sep 06 13:32:19 we actually use for unortodox jffs2.summary images Sep 06 13:32:30 I'm sorry if I sound totally lost. I really am, I must admit. :-) Sep 06 13:32:41 http://cgit.openembedded.org/cgit.cgi/meta-handheld/tree/conf/machine/include/zaurus.inc Sep 06 13:34:38 Hmmm. Interesting. I got to T just because I want to programmatically find out the location of the rootfs file. Sep 06 13:39:14 rootfs log file, I mean. Sep 06 13:50:03 JaMa|Off: klibc-2.0 is behind the corner (kernel.org outage delay) Sep 06 14:35:30 hi, did someone succeed to make qtdemo work? we have an issue that is hard to debug: it doesn't show anything on the screen but the rest seem ok Sep 06 14:36:15 we overrided busybox(for /dev/touchscreen0) and tslib(added git version) in our local repo Sep 06 18:48:08 I have MACHINE = "ossystems-x86" on my config file. When I run "bitbake -e | grep ^MACHINE_ARCH" I get MACHINE_ARCH="ossystems_x86". It seems that openembedded-core/meta/conf/bitbake.conf is translating `-' to `_'. Why is it doing that? Sep 06 19:04:06 specifically, the line doing that is Sep 06 19:04:08 MACHINE_ARCH = "${@[bb.data.getVar('TUNE_PKGARCH', d, 1), bb.data.getVar('MACHINE', d, 1)][bool(bb.data.getVar('MACHINE', d, 1))].replace('-', '_')}" Sep 06 19:16:18 - in machine names causes problems with some packages Sep 06 19:16:50 Hmmm. Sep 06 19:16:58 Thanks, incandescant Sep 06 19:18:47 mario-goulart: a recent example - http://bugzilla.pokylinux.org/show_bug.cgi?id=946 Sep 06 19:20:43 incandescant: it means that machine names should follow C's naming conventions for identifiers? Sep 06 19:22:57 mario-goulart: probably a safe assumption Sep 06 19:23:02 s/conventions/rules/ Sep 06 19:26:29 * housel renames his local MACHINE config Sep 06 19:30:50 Wouldn't it be better to raise an error instead of automatically mangling the MACHINE value? Sep 06 19:34:45 would it? why would it matter? Sep 06 19:35:57 Those who use tools on top of OE (like me) can get easily screwed because of this. Sep 06 19:37:16 fair enough, I'd recommend you take this discussion to the list and/or submit a patch Sep 06 19:37:20 aside; care to share which tool? Sep 06 19:38:08 It's custom tool by ossystems to make the generation of images automatic for multiple machines/distros. Sep 06 19:39:02 I'll talk to otavio. We probably can work around this issue. I just mentioned just in case it can affect others. Sep 06 20:41:45 I see that use of PR in recipes is not accepted anymore Sep 06 20:41:49 what replaces its purpose? Sep 06 20:42:41 it's about PR="r0" which is default value (the same without PR) Sep 06 20:42:52 ah gotcha Sep 06 20:43:27 JaMa|Off, koen suggested to remove "PR=r*" from all new recipes Sep 06 20:43:46 I guess he wants it to default to r0 , yep. k k Sep 06 20:43:56 ty Sep 06 20:44:10 no Sep 06 20:44:33 he just said that 'PR = "r0"' shouldn't be in recipe because it's useless Sep 06 20:52:32 * incandescant disagrees with Koen on this Sep 06 20:58:47 resistence is futile Sep 06 20:59:11 Hi! Have a question about libtool behaviou. When I start process of building some target, it build libraries and creates *.la files in sysroot. This file contains string "installed=yes" and then, when it comes to build my target (in this case it is xbmc), it fails to link some libraries, because it tries to link libraries from my host system. It happens because in *.la file value for "installes" is set to "yes". When I change it to "no", my target builds like a ch Sep 06 20:59:11 arm... What is wrong and what shoyld be done about it? Thanks Sep 06 21:00:04 one method is Sep 06 21:00:11 do deinstall the libs from your host Sep 06 21:00:29 hi woglinde_ Sep 06 21:00:57 oh...thats bad Sep 06 21:01:19 because i need those libraries Sep 06 21:01:32 you reinstall them afterwards Sep 06 21:02:00 mike111: What are you using to build XBMC? oe-core/meta-oe or oe.dev? Sep 06 21:02:22 angstrom-setup-scripts Sep 06 21:02:25 hi paul Sep 06 21:02:40 git://github.com/openembedded/openembedded.git Sep 06 21:02:51 hi woglinde_ Sep 06 21:04:49 mike111: That is the classic oe or oe.dev. (Everybody calls it differently.) Sep 06 21:05:06 mike111: Do you have an updated recipe of XBMC? Sep 06 21:05:37 i think no Sep 06 21:05:48 my revision is "r1" Sep 06 21:06:26 woglinde_: I guess you are swampt in work since I have not seen a lot of patches from you lately? Or are you working on oe-core? Sep 06 21:06:44 mike111: Alright, so it is the shipped recipe. Sep 06 21:07:04 mike111: I have never had problems using it? What libraries are causing issues for you? Sep 06 21:07:32 libz.so and libexpat.so Sep 06 21:07:55 mike111: And just for my interest, what machine are you building for? Sep 06 21:09:23 for babbage machine (its freescale imx51) Sep 06 21:09:45 mike111 they are dev libs only Sep 06 21:10:00 so deinstall the devlibs and rerun compiling Sep 06 21:10:12 fastest way for now Sep 06 21:10:52 how can it be done? Sep 06 21:13:00 use your package mamanger? Sep 06 21:13:03 What distribution do you use? Sep 06 21:13:05 ups manager Sep 06 21:13:57 oh...you mean remove dev packages from my host machine? Sep 06 21:14:04 oh...ok...i've got it Sep 06 21:15:29 thanks you. will try this way Sep 06 21:16:37 Now I remember that I hit this issue too after having `zlib1g-dev` (Debian) installed. Sep 06 21:17:52 Somebody reported that already to their ticket system but it was not committed. Sep 06 21:19:01 wow 5 months Sep 06 21:20:13 mike111: You should definitly report your experience with XBMC to the list. In a month or so I will hopefully have time to get my local OE/XBMC stuff pushed. Sep 06 21:20:44 or just integrate the patch Sep 06 21:21:19 woglinde_: Yeah. They now also switched to GitHub and a pull request model so that the patches in Trac are mostly ignored (I think). Also a lot has happened in between so it probably fails to apply (cleanly). Sep 06 21:21:47 paule the fix is so easy you can do it by hand Sep 06 21:22:12 anyway oe should run autoreconf it self Sep 06 21:22:37 PaulePanter: you mean to post my expirience on openembedded-devel mail list? Sep 06 21:23:01 mike111 nobody will care Sep 06 21:23:17 I would find it interesting. Sep 06 21:24:58 well...i've created thread on xbmc forum http://forum.xbmc.org/showthread.php?t=109361 Sep 06 21:25:17 mike111 read the bug Sep 06 21:25:23 At least the current XBMC version in OE should not too well on BeagleBoard. Sep 06 21:25:27 no need for a discussion Sep 06 21:26:39 well...no discussion...i've just posted my build logs there Sep 06 21:26:52 and? Sep 06 21:27:08 problem is outdated libtool in some libs Sep 06 21:27:12 as stated in the bug Sep 06 21:27:21 Please point them to the bug report so it gets attention. Sep 06 21:28:14 316 active tickets for linux Sep 06 21:28:15 not bad Sep 06 21:31:07 Since a week or so XBMC’s HEAD stopped building in OE, because they added support for Airplay which depends on libplist and they did not make that configurable. Sep 06 21:31:34 libplist is missing in OE and therefore configure fails. Sep 06 21:33:07 woglinde_: Do you know if Robert fixed that UNZIP issue with one of the Java packages? Sep 06 21:34:17 UNZIP was used as the environtment variable for the executable and by `unzip` it is used as the path to the archive. Sep 06 21:37:29 hm Sep 06 21:37:37 dont remember Sep 06 21:37:43 but I think he fixed it Sep 06 21:38:56 woglinde_: I did not see a patch on the list or in the commit log. But I could have missed it. Sep 06 21:39:14 I will look again in some weeks and report back. Sep 06 21:40:51 mike111: You can also send me your experiences with Babbage(?), OE and XBMC privately. Or just the URL to a forums post or so. Sep 06 21:44:04 i can send it privatly Sep 06 21:44:40 after i've removed zlib-dev and libexapt-dev from my host machine, xbmc compiled without any problem Sep 06 21:45:13 As woglinde_ told us. Genius! Sep 06 21:45:59 mike111: Do you want to just test XBMC or build a media center? Sep 06 21:46:20 this can be somehow connected with version of angstrom...now i'm using 2010 and build failed, but before it was 2008.1 and build was ok Sep 06 21:46:39 i want to build a media center Sep 06 21:48:14 Oh, and I wanted to tell you about qtmediahub . This uses Qt and looks quite promissing. I have not had time yet though to test it. Sep 06 21:48:52 mike111: Please send the log of the error to openembedded-devel. Sep 06 21:51:04 only log? Sep 06 21:51:25 well...its pretty large (7.1 M) Sep 06 21:52:48 please no 7 mb to ml Sep 06 21:53:01 only the failures Sep 06 21:53:05 good nite Sep 06 21:53:18 ok... Sep 06 21:54:13 i will do it tomorrow, its pretty late here... Sep 06 21:54:46 bye Sep 07 00:26:31 whee **** ENDING LOGGING AT Wed Sep 07 02:59:57 2011