**** BEGIN LOGGING AT Fri Jun 14 02:59:58 2013 Jun 14 03:00:40 why SRCREV we should give.. i have a git repo where i have the soruce code directly.. if i make SRCREV="" then only it is fetching.. Jun 14 03:03:01 otherwise i made tag="" then it is working.. otherwise please set SRCREV a valid value.. error Jun 14 03:19:31 you could use a) a tag, b) a commit hash, or 3) ${AUTOREV} Jun 14 03:25:31 just don't set it blank... Jun 14 07:30:55 good morning Jun 14 07:57:55 good morning Jun 14 08:08:48 hi apelete, all Jun 14 08:24:13 hi bluelightning, hi all Jun 14 08:44:33 morning all Jun 14 09:08:53 moarning all Jun 14 09:09:03 *morning all Jun 14 09:09:31 morning silvio_l_ Jun 14 09:16:23 hi silviof Jun 14 09:58:58 what is the best way to replace a configuration file? file like meta/recipes-core/init-ifupdown/.../interfaces or so ... Jun 14 09:59:56 silviof: in a custom layer, bbappend the recipe that installs the file (init-ifupdown for /etc/network/interfaces yes) Jun 14 10:01:18 bluelightning: thx Jun 14 10:08:10 bluelightning: ahh and then change the "FILESEXTRAPATHS"?! Jun 14 10:09:44 silviof: yes Jun 14 10:48:40 hi Jun 14 10:48:55 how do I make oe skip the patch step for my local checkout Jun 14 10:49:17 I use S_pn-foo = "/home/..../mycheckout" Jun 14 10:49:38 and INHERIT_append_pn-foo = "externalsrc gitpkgv" Jun 14 10:49:48 in a bar-local.conf Jun 14 10:52:56 heeen: does INHERIT_append_pn-xyz actually work? Jun 14 10:53:18 shouldn't it? Jun 14 10:53:29 I got that from someone else Jun 14 10:53:44 bluelightning: how would you do it Jun 14 10:54:16 if it does work, no problem... I just never tried that and because INHERIT is kind of special I figured there was a chance it might not work as expected Jun 14 10:55:05 ugh Jun 14 10:55:07 https://github.com/webOS-ports/webos-ports-setup/blob/master/conf/local-builds.inc#L14 Jun 14 10:55:18 does NOT work Jun 14 10:55:21 fff Jun 14 10:56:07 so what is the proper way in the recipe Jun 14 10:56:17 INHERIT_append? Jun 14 10:56:25 INHERIT_append_pn? Jun 14 10:56:35 INHERIT_append_pn-qtdeclarative? Jun 14 10:57:50 or just inherit externalsrc Jun 14 10:59:37 JaMa: ^ Jun 14 11:00:06 heeen: I'm thinking the only way it can work is to have a bbappend for the recipe in an extra local layer that does "inherit externalsrc" Jun 14 11:02:50 hmpf Jun 14 11:03:02 why can I change S_pn but not inherit Jun 14 11:03:10 from the foobar-local.conf Jun 14 11:03:40 inherit operates at a different level; same with include/require Jun 14 11:09:50 hello Jun 14 11:09:58 any gvfs expert over here Jun 14 11:10:39 few months back samba is enabled in gvfs in meta-gnome Jun 14 11:11:22 I can't build GPLV3 as samba is GPLV3 so I am not able to build gvfs Jun 14 11:11:49 is it acceptable that that I disable samba support in gvfs? Jun 14 11:13:03 Noor: we probably ought to have a PACKAGECONFIG to disable that, I'm assuming it's possible with a configure option... Jun 14 11:13:47 bluelightning: i was thinking on that terms as well Jun 14 11:14:18 so by default we want to enable it or disable it Jun 14 11:15:12 Noor: I suspect default enable since it's currently enabled within meta-gnome Jun 14 11:16:20 heeen: we might be able to fix externalsrc to not require the inherit at the recipe level; I'll look into it Jun 14 11:29:10 I am looking into old commit of gvfs Jun 14 11:29:14 EXTRA_OECONF = "--disable-samba \ Jun 14 11:29:15 --with-samba-includes=${STAGING_INCDIR} Jun 14 11:29:58 isn't it wrong ... it is saying that samba is diabled and it is giving samba includes as well Jun 14 11:30:47 with sambs disbaled we should not provide its include folder and lib Jun 14 11:38:25 bluelightning: it would be nice to have all the customization for local builds in one file Jun 14 11:41:33 my bitbake broke after we upgraded to dylan Jun 14 11:41:37 I can only use make Jun 14 11:48:23 Is it possible to have per "package" ${prefix} for autotools? Jun 14 11:48:38 I have /usr as default, but want /opt for one package.. Jun 14 11:58:55 I have trouble building qtdeclarative from a local tree Jun 14 11:59:06 it complains about a missing moc file Jun 14 11:59:31 in fact there are no moc files at all Jun 14 12:25:52 gonzzor: that should work yes, just set prefix= in the recipe Jun 14 12:42:37 I've started to get ERROR: ld.so: object 'libpseudo.so' from LD_PRELOAD cannot be preloaded: ignored. when pulling latest oe-core + meta-openembedded... anyone else seeing this? Jun 14 12:49:45 /qtdeclarative/src/qml/util/qqmladaptormodel.cpp:977:32: fatal error: qqmladaptormodel.moc: No such file or directory Jun 14 12:50:01 whyyy Jun 14 12:52:33 bluelightning: Thanks Jun 14 12:53:28 roric: you're not using the external python tarball (buildtools-tarball) are you? Jun 14 13:04:29 bluelightning, my machine is sadly a ubuntu installation so it's all packaged by them, then nothing magic among my recipes Jun 14 13:04:51 bluelightning, same thing works with a clone of oe-core which is a couple of months old Jun 14 13:22:38 bluelightning, I will dig into this later, any pointers? Jun 14 13:23:46 roric: pseudo allows us to run things pretending to be root (but not really) so that when we create an image for example files have all the right permissions Jun 14 13:24:09 it operates as an LD_PRELOAD library and intercepts various library function calls Jun 14 13:25:03 there could be a number of reasons why it wouldn't be able to be preloaded - architecture mismatch, missing libraries it depends upon are two things I can think of Jun 14 13:35:53 bluelightning, ah interesting. I have an x86_64 which is multilib, I have some i386 on there as well.. But see no reason why bitbake should touch the i386 stuff... Jun 14 13:37:32 bluelightning, maybe it tries do load some arm lib from the sysroot or something.. Jun 14 13:37:58 bluelightning, is this interceptable by strace or is pseudo to magic? Jun 14 13:38:01 too Jun 14 13:39:38 roric: I don't know I'm afraid; you may want to try pinging seebs (he's in #yocto, but probably isn't awake yet) Jun 14 13:39:45 he's the author of pseudo Jun 14 16:25:19 http://www.mail-archive.com/openembedded-core@lists.openembedded.org/msg39645.html Jun 14 16:25:38 I have an error trying to build SHR Jun 14 16:25:47 ERROR: Nothing RPROVIDES 'dosfsprogs' Jun 14 16:25:53 as described in that email Jun 14 16:26:17 Christopher Larson says "Ignore this for the time being." Jun 14 16:26:43 how can I ignore a non-building dependency? Jun 14 16:36:09 rah: a patch has already been sent and merged for this Jun 14 16:36:46 rah: http://cgit.openembedded.org/openembedded-core/commit/?id=15e3c72bdcabf0a21efff60333323ac04908efb0 Jun 14 16:39:21 $ git remote -v Jun 14 16:39:21 origin git://github.com/shr-distribution/oe-core.git (fetch) Jun 14 16:40:36 * rah glares at JaMa Jun 14 16:42:28 rah: heh, at the time i sent the ignore this email, i didn't realize the patch series had already been merged. sent a patch to the list about 10 seconds after realizing it had been Jun 14 16:43:56 I'm wondering why there's a different openembedded-core repository for SHR Jun 14 16:45:16 probably forked it to apply bugfixes that aren't yet upstream? *shrug* Jun 14 16:46:55 sounds like a good idea Jun 14 16:47:21 pity the fork is unbuildable Jun 14 16:47:31 usually you dont need to fork oe-core, you can just use bbappends, but when you have fixes for classes things can get ugly. i've forked it to apply extra bugfixes to the denzil/danny release branches before Jun 14 17:29:47 what is sstate Jun 14 17:40:23 heeen: sstate, aka shared state, is a mechanism for skipping parts of the build by caching the output of tasks. essentially, it lets you have a binary cache of previous build output to speed up future builds. Jun 14 17:44:06 RP, RP_ I think there is a change done in bitbake/lib/bb/utils.py that breaks support for python 2.7.4 at least.. Jun 14 17:44:59 rah: updated, now you can submit patches to fix util-linux and nspr .bbappends Jun 14 17:45:46 RP, RP_: I see this: Exception: AttributeError: 'function' object has no attribute 'Pool', changing return multiprocessing.Pool.Pool(*args, **kwargs), to return multiprocessing.Pool(*args, **kwargs) works for me.. but don't know what happens to other python versions.. Jun 14 17:47:17 rah: that's what you get with unbuildable upstream revision.... Jun 14 17:55:16 roric_: sounds like a problematic search/replace Jun 14 17:57:28 kergoth, you mean a patch is simpler ;) Jun 14 17:58:09 I don't understand what you mean. I'ms aying he probably did a s/Pool/multiprocessing.Pool/ and it replaced an existing multiprocessing.Pool with multiprocessing.Pool.Pool Jun 14 17:58:45 kergoth, Ah now I see what you mean, I thought you meant my message wasn't so helpful Jun 14 19:09:22 are there instructuions for using buildhistry Jun 14 19:29:36 * mr_science enjoys packaging GNU projects Jun 14 19:30:35 nothing to do, mostly, which is nice to have once in a while when you just want to use the damn thing instead of debug a new recipe... Jun 14 19:31:39 even with a homepage defined it's a 7 line recipe... Jun 14 19:46:11 true, those bearded guys know how to write software :) Jun 14 19:48:16 yeah, just don't think about the stuff Stallman picks from his feet... Jun 14 19:51:44 does someone remember seeing libc sometimes loosing executable bit on /lib/ld-VER.so? it's lost in do_package (sometimes) Jun 14 19:52:38 iirc at some time we even had postinst script fixing it in do_rootfs (but maybe I messed that with work around for missing libc symlink) Jun 14 20:21:29 we humans are so forgetful... Jun 14 20:21:50 * mr_science makes a mental note to hassle the quality guy on Monday Jun 14 21:01:42 roric_: Fix pushed Jun 14 21:21:14 g 2 Jun 14 21:24:58 bingo! Jun 14 22:46:28 RP, great, thanks Jun 15 00:00:15 ah I started to hit multiprocessing.Pool.Pool issue too good I checked IRC Jun 15 00:00:39 one by one the penguins stole my sanity Jun 15 00:03:26 :) Jun 15 00:05:14 JaMa: Do we have new report on world builds Jun 15 00:05:23 after the latest round of fixe Jun 15 00:05:24 s **** ENDING LOGGING AT Sat Jun 15 02:59:58 2013