**** BEGIN LOGGING AT Wed Aug 24 02:59:57 2011 Aug 24 03:10:58 hi Aug 24 03:14:02 anyone here? Aug 24 03:40:58 anyone know the max length of a shebang line? Aug 24 06:53:52 hi all Aug 24 06:54:11 PaulePanter: r u there? Aug 24 06:56:26 Noor: Just got up. I am here. Aug 24 06:58:08 PaulePanter: just wanna know the reason of putting '"' after the last line of SRC_URI Aug 24 06:58:41 PaulePanter: where are you based? Aug 24 06:58:52 Noor: I already wrote my answer to the mailing list. Aug 24 06:59:02 Noor: Berlin, Germany Aug 24 06:59:22 Noor: Where do you live? Aug 24 06:59:48 PaulePanter: Lahore, Pakistan Aug 24 07:00:09 PaulePanter: ok ill check out the answer on the ML Aug 24 07:17:15 PaulePanter: is this correct http://cgit.openembedded.org/cgit.cgi/openembedded/tree/recipes/lmbench/lmbench_3.0-a9.bb?id=54aaa2d908254f0226f6c335d8e3f6a742e1d8c2 to put in commit message for revision purpose Aug 24 07:38:24 good morning Aug 24 07:45:12 Noor: »Imported from OE rev …« should be enough. In `meta-openembedded` just do `git log` and take a look how others have written int the commit log. Although not everyone is doing it “right”. Aug 24 07:45:37 Noor: There are also examples and suggestions in the Wiki . Aug 24 07:46:09 hello Aug 24 07:46:37 I want to compile libva libraries in oe.... can anyone help me howto? Aug 24 07:51:10 PaulePanter: ok .... I was following ur suggestions for commit message ..... you mentioned to write revision as well so that is why looking how people have written the revision and got one example of cgit Aug 24 07:53:48 PaulePanter: rev is the commit has? Aug 24 07:53:55 hash Aug 24 07:55:47 Noor: Yeah, I did not know the correct terminology either. Commit id, hash or revision. Revision is a left over from Subversion/SVN I believe. Aug 24 09:20:12 does machine specific changes in recipes will go in meta-oe? Aug 24 09:21:04 like # --disable-threads is needed on epia/x86 with uclibc # RIK: or probably not with recent uclibc with NPTL EXTRA_OECONF_append_epia = " --disable-threads" Aug 24 09:22:13 this is from iperf recipe from oe.dev Aug 24 09:22:28 Noor: that should be in .bbappend in meta-epia Aug 24 09:22:42 Yeah thats what I was thinking Aug 24 09:23:08 so if I am movinf recipe to meta-oe from oe.dev I should remove these lines Aug 24 09:23:14 moving Aug 24 09:27:41 yes Aug 24 11:04:38 hi all Aug 24 12:07:24 Ü/na Aug 24 12:07:27 ops Aug 24 12:07:29 sorry Aug 24 13:46:15 Hi everybody Aug 24 13:46:50 Is there a way in openembedded/yocto to compile GTK+ 2 with framebuffer support (not X11) ? Aug 24 13:58:58 mrAlmond: gtk does not support framebuffer, it supported DirectFB. But AFAIK DirectFB was broken since 2.16. Not sure if anyone attempted to fix it Aug 24 14:05:11 anarsoul : What about this http://www.gtk.org/api/2.6/gtk/gtk-framebuffer.html ... is this an old page? Aug 24 14:05:35 mrAlmond: it's something pretty old Aug 24 14:06:04 ok tnx Aug 24 14:07:20 morning Aug 24 14:07:25 esbenh: ping Aug 24 14:24:46 ericben: hi Aug 24 15:41:13 another article about OE Aug 24 15:41:16 http://www.eetimes.com/design/embedded/4218490/Open-Embedded--An-alternative-way-to-build-embedded-Linux-distributions Aug 24 15:42:41 denix: nice work :) Aug 24 15:43:39 seems people from the buildroot world quite like the menuconfig interface over there... Aug 24 15:45:47 good idea to make 'menuconfig' bitbake UI :) Aug 24 15:46:02 I think someone already started an ncurses one ages ago Aug 24 15:46:48 in Yocto we are working on a GTK+ based BitBake UI (hob) as you may be aware Aug 24 15:47:49 I remember only goggle UI Aug 24 15:47:51 iirc Aug 24 15:48:36 hi gnutoo Aug 24 15:51:50 Jay7: the author of Goggle sits two desks away from me right now :) he doesn't work on Poky/Bitbake/OE anymore though Aug 24 15:53:10 broadly, hob provides more of the frontend configuration in addition to being a shell for showing the during-build output that goggle provided Aug 24 15:53:32 bluelightning: you have interesting people around :) Aug 24 15:53:52 indeed :) Aug 24 16:44:44 hmm why isnt nano a valid target.. Aug 24 16:45:01 angstrom-2010.x not a valid distro ? Aug 24 16:50:28 http://pastebin.com/TSWEFuZE maybe i have a path issue.. Aug 24 16:50:34 no idea what else it might be Aug 24 17:08:59 nvm all my fault.. typo in recepies vs recipes.. Aug 24 17:09:20 wish there was a check on that dir throwing a sane message if dir does not exist.. Aug 24 17:10:09 unsolo: I think if it discovers there are no recipes in recent versions at least you get an error to that effect Aug 24 17:24:16 hmm i did a git clone like 20 minutes ago.. Aug 24 17:25:13 it warns if the bbfile pattern for the layer doesn't match anything, but not if you add a glob to bbfiles that doesn't match anything, afaik, but could be remembering incorrectly Aug 24 17:25:54 kergoth: actually this was not the glob BB path but rather the one for recepies in local.conf Aug 24 17:26:01 yes.. Aug 24 17:26:13 recipes are added to bbfiles, usually with a glob. /path/to/foo/recipes/*/*.bb Aug 24 17:26:19 sometimes people add the dir, but that's less common Aug 24 17:26:40 my point was that if /path/to/foo/recipes does not exist.. Aug 24 17:26:50 it would be nice if one could see that Aug 24 17:27:04 I know Aug 24 17:27:39 bluelighting said that there was a warning, I explained why I didn't think there was in this case Aug 24 17:27:47 is that confusing, somehow? Aug 24 17:28:14 now its fine.. Aug 24 17:28:37 im simply pointing out possible improvements as i discover pitfalls ;) Aug 24 17:31:29 * kergoth nods, good thing Aug 24 17:31:35 it's not a bad idea Aug 24 17:32:10 anyone else had trouble with oenjade-native recently Aug 24 17:39:26 unsolo: you may have the latest metadata but you're using a fairly old version of bitbake Aug 24 17:39:44 hmm 10.2 was the one marked as stable in gentoo.. Aug 24 17:40:14 13.3 is reachable. Aug 24 17:40:36 I've always found it easiest to just download and use my own copy of bitbake Aug 24 17:40:40 maybe that provides more sane error messages.. Aug 24 17:41:29 then again, these days because of my work I'm more or less (bar a few deltas) using the very latest version from git master Aug 24 17:41:47 :) Aug 24 17:42:10 btw is there a guide on how to do OE Offline ? Aug 24 17:42:26 Noor, also dbench needs those patch headers added :) Aug 24 17:42:45 1.12 is the current "stable", 1.13 if you want new features. current oe-core requires 1.13.2 or 1.13.3, i don't recall which Aug 24 17:42:56 unsolo: not sure, but it's pretty straightforward... AFAIK you can just do bitbake -c fetchall ; then you can set BB_NO_NETWORK = "1" in your config Aug 24 17:43:55 cool Aug 24 17:53:59 can I specify if a recipe needs pythong 2.6 or 2.4, etc? Aug 24 17:54:12 msm: no Aug 24 17:54:17 pyThong.. hmm sexy.. Aug 24 17:55:04 well Aug 24 17:55:12 im getting some weird errors on mesa-xlib Aug 24 17:55:21 http://pastebin.com/6Da9Uven Aug 24 17:56:34 ive even tried just disabling the recipe as well with EXCLUDE_FROM_WORLD_pn-mesa-xlib = "1" Aug 24 18:04:19 msm, and it already has DEPENDS = python-native? Aug 24 18:06:30 Tartarus, have you tried building openjade-native recently (in dev)? Aug 24 18:07:08 something weird is going on, because it used to work :( Aug 24 18:07:24 Nope, my new build box arrives tomorrow Aug 24 18:07:27 and the PS friday :( Aug 24 18:07:36 (and then I'm off for a week of vacation) Aug 24 18:07:46 Tartarus: let me look Aug 24 18:07:59 awesome Aug 24 18:08:05 have fun Aug 24 18:08:30 I am really tempted to try the version from oe-core Aug 24 18:08:46 jsut need to get a sane set of stuff in maintenance so I can focus on oe-core Aug 24 18:08:59 Noor_: dbench and tiobench also need patch headers added Aug 24 18:09:44 Tartarus: hi Aug 24 18:09:51 Tartarus: it appears to be missing this depends Aug 24 18:10:19 Tartarus: will retest Aug 24 18:10:45 yeah .... just came to irc to know what I am missing Upstream-Status tag from the patches? Aug 24 18:11:31 coz I am not getting that what exacty is missing from my commit message Aug 24 18:12:40 Tartarus: kindly help me what exactly is missing Aug 24 18:14:46 Noor_, headers in the patch files Aug 24 18:14:51 That match what the wiki says Aug 24 18:16:17 hey noor, you commited some sgml-common changes Aug 24 18:16:32 can you check you can still build openjade-native? Aug 24 18:16:42 I think this might be my problem Aug 24 18:17:00 although this is not explaing the failures in maintenance .... Aug 24 18:20:43 the maintenance faiures may be related to me being an idiot Aug 24 18:23:00 Crofton: sorry I dont have access to my build machine atm Aug 24 18:25:29 ok, I will try a build withthe sgml stuff out and see what happens and send an email to the list Aug 24 18:25:43 release branch looks ok Aug 24 18:25:55 ignore my earlier rambling on that subject Aug 24 18:36:48 Tartarus: I think the mesa-xlib build scripts are references "python2" and consequently pulling in the host python possible? Aug 24 18:36:52 ly* Aug 24 18:37:37 possibly Aug 24 18:37:41 check log.do_configure :) Aug 24 18:38:02 and see if there's a way to force it to use a specific python, after adding python-native to DEPENDS Aug 24 18:39:57 noor never mind Aug 24 18:40:08 the change that is bothering me is from 2010 Aug 24 18:40:14 just happens to be dated Aug 23 Aug 24 18:40:17 just happens to be dated Aug 24 Aug 24 18:40:59 :-) Two days ago we released our first system based on OpenEmbedded-Core + Meta-OpenEmbedded Aug 24 18:41:02 :-) Aug 24 18:41:40 otavio, congratulations Aug 24 18:41:49 any public info Aug 24 18:41:51 ? Aug 24 18:41:54 Crofton: not yet Aug 24 18:42:46 Crofton: btw, it uses combo-layer to merge all the needed repositories Aug 24 18:43:52 Tartarus: ok nothing for configure but an export PYTHON2=python2.6 works Aug 24 18:44:04 along with the DEPENDS Aug 24 18:44:11 there ya go :) Aug 24 18:44:12 otavio, are you planning on going to linuxcon brazil? Aug 24 18:44:18 hi all Aug 24 18:44:19 Tartarus: patch to oecore? Aug 24 18:44:20 Crofton: yes Aug 24 18:44:21 afk a while, errands Aug 24 18:44:27 msm, if that's where it is, yes Aug 24 18:44:29 Crofton: I ought to be theer Aug 24 18:44:35 its in meta Aug 24 18:44:35 Crofton: are you going as well? Aug 24 18:44:42 it would be fun to go, but too far for me Aug 24 18:44:47 Tartarus: catching up a bit on OE stuff. Just remembered the 2011.03 java patch Aug 24 18:45:14 ok openjade-native is ok in release, but not dev Aug 24 18:45:33 Tartarus: Difficult situation. The patch Henning and Robert prepared is needed to make sure that we get openjdk-6 building reliable for BugLabs here. Aug 24 18:46:02 Tartarus: Testing the patch gives me a broken build here again. Aug 24 18:46:26 Tartarus: Still silently sucking return codes and the borken classpath is a big problem for the other users it seems Aug 24 18:47:23 Tartarus: We have a branch for more stuff here anyway based on 2011.03 Aug 24 18:47:59 Tartarus: If it helps the patch can get applied to plain 2011.03 and I keep the old setup over here. Aug 24 18:48:30 Tartarus: That may means trouble for other people using openjdk-6 on the target though :( Aug 24 19:04:00 Crofton: :-/ Aug 24 22:20:16 hmm... does OE have any support for A9 based systems? Aug 24 22:31:49 tzanger, people are using Angstrom Distro for Panda I think. Aug 24 22:32:12 panda? Aug 24 22:32:26 pandaboard Aug 24 22:32:37 oh I see Aug 24 22:33:06 thank you, I'll have to check it out Aug 24 22:33:52 http://en.szider.com/products_detail/&productId=9fc24d29-0481-4559-9a5a-17d6c3d70274&comp_stats=comp-FrontProducts_list01-004.html <-- I'm looking at that, it runs android but I'm looking at OE perhaps Aug 24 22:34:46 it's not OMAP but same ARM core Aug 25 00:41:36 03Philip Balister  07master * r9e9d8ef5ea 10openembedded.git/recipes/openjade/ (2 files in 2 dirs): (log message trimmed) Aug 25 00:41:36 openjade: fix build with GCC 4.6 Aug 25 00:41:36 Copied from oe-core 5e58a44ee5d5be814614cdcc147d1ce67a9a401d Aug 25 00:41:36 Thanks to Joshua Lock. Aug 25 00:41:36 Commit message from oe core: Aug 25 00:41:36 In GCC 4.6 the compiler no longer allows objects of const-qualified type to Aug 25 00:41:37 be default initialized unless the type has a user-declared default Aug 25 01:58:25 So is OE moving towards the OE-Core layout? Aug 25 01:58:29 vs OE classic? Aug 25 02:00:55 yes Aug 25 02:01:15 I must finish the last of my classic work Aug 25 02:04:32 I've only found one wiki page about it Aug 25 02:04:33 maybe two Aug 25 02:04:39 are there any mailing list threads about it? Aug 25 02:05:47 And/or a minimum bitbake version? Aug 25 02:33:01 03Joshua Lock  07master * r0ed281feb6 10bitbake.git/lib/bb/fetch2/git.py: Aug 25 02:33:01 bb/fetch2/git: add checkstatus command Aug 25 02:33:01 Use git ls-remote to implement checkstatus command for the git fetcher. Aug 25 02:33:01 Signed-off-by: Joshua Lock Aug 25 02:33:01 Signed-off-by: Richard Purdie Aug 25 02:37:06 03Joshua Lock  07master * r0423587db0 10bitbake.git/lib/bb/ui/hob.py: Aug 25 02:37:06 hob: disable some menu entries whilst build is in progress Aug 25 02:37:06 It doesn't make sense to be able to modify the preferences and add/remove Aug 25 02:37:06 layers whilst a build is in progress - disable the relevant menu items once Aug 25 02:37:06 the build has started and re-enable them once the user has returned to the Aug 25 02:37:06 creation view. Aug 25 02:37:07 Signed-off-by: Joshua Lock Aug 25 02:37:16 03Joshua Lock  07master * ra17334e858 10bitbake.git/lib/bb/ui/crumbs/tasklistmodel.py: Aug 25 02:37:16 bb/ui/crumbs/tasklistmodel: fix find_reverse_depends method Aug 25 02:37:16 This fixes an embarassing typo which meant the method actually returned a Aug 25 02:37:16 list of packages which didn't depend on the passed pn. Aug 25 02:37:16 Signed-off-by: Joshua Lock Aug 25 02:37:16 03Joshua Lock  07master * r27dcf245ab 10bitbake.git/lib/bb/ui/ (crumbs/hobprefs.py hob.py): (log message trimmed) Aug 25 02:37:17 hob: don't set PARALLEL_MAKE and BB_NUMBER_THREADS based on cpu count Aug 25 02:37:17 This was actually broken with recent changes as the values were never Aug 25 02:37:18 persisted to a file (meaning they were unset on the server at reparse Aug 25 02:37:18 despite the UI indicating they were set). Aug 25 02:37:19 However, I've chosen to remove the 'feature' as pegging a users CPU without Aug 25 02:37:19 them asking to use high thread counts seems a little offensive. **** ENDING LOGGING AT Thu Aug 25 02:59:58 2011