**** BEGIN LOGGING AT Fri Aug 14 02:59:57 2009 Aug 14 03:03:58 denix0: repo, as in the android tool? Aug 14 03:04:07 yep Aug 14 03:04:30 i monitor the mailing list, but unfortunately it seems to be moving away from being an independent tool. its getting more locked in to the use of gerritt and stuff Aug 14 03:04:36 which makes me sad, i quite like the tool Aug 14 03:05:24 kergoth: from what I undertood, it can do the task of managing slave repositories/overlays based on the setup of the master one... Aug 14 03:05:29 * kergoth is quite interested in tools like git-submodule, piston, repo, etc, but has yet to find anything he's really happy with Aug 14 03:05:51 I'm pretty sure that functionality is now deprecated Aug 14 03:06:07 * kergoth tries to dig up the email Aug 14 03:06:32 * kergoth grumbles, gmail hates him Aug 14 03:06:34 eh is the modules-xxx.tgz the initrd? do I have to untar and make a cpio out of it? Aug 14 03:06:48 kergoth: so, for this specific task you had to write your own fetcher for the compressed overlays, right? Aug 14 03:07:22 J13R: no, it is no initrd Aug 14 03:07:30 J13R: it's just a collection of modules, but you sure can make an initrd out of it :) Aug 14 03:07:55 how? i just want a initrd *sob* Aug 14 03:08:00 montavista Linux 6 provides a few tools. mvl-project does project area setup, generates the local.conf, calls out mvl-content to actually download stuff from the montavista content delivery system based on what the user has purchased, msd wise, and adds the downloaded collection tarballs to COLLECTIONS Aug 14 03:08:23 I thought the linux-rp would spit it out directly Aug 14 03:08:24 then we also inject a new fetcher into bb.fetch to grab actual sources from our content server for that side of things Aug 14 03:09:14 kergoth: I see, nice setup Aug 14 03:10:13 kergoth: I'm working on something similar, not so sophisticated, but a way to manage collections... was also thinking about some sort of a project setup tool Aug 14 03:11:28 what I'd really, really like to see long term is us move to something like what the vcs-pkg folks are pushing. a git repo containing the sources to the package, with branches for managing features and distro specific metadata and all, then a project setup tool to pull down those git repositories and setup bitbake to use the recipes they include, that fetching would be kind of like repo also Aug 14 03:11:31 I'm thinking like.. Aug 14 03:11:51 oe-project new -m mymachine foo; cd foo; oe-project add task-boot nano; . ./setup.sh; bitbake nano Aug 14 03:12:05 and all that would be in the project is the metadata + sources for *exactly what you wanted in the project* Aug 14 03:12:09 nothing else Aug 14 03:12:22 and since the recipes woudl be in the source trees, itd be much less painful to do active development Aug 14 03:12:38 thinking oe-project would be like.. apt-get Aug 14 03:12:42 cool! Aug 14 03:13:18 again, I'm thinking along those lines as well for our project, but on a much smaller scale :) Aug 14 03:13:22 :) Aug 14 03:14:33 * kergoth 's been trying to figure out how to move toward some of these long term goals of his without having to rewrite anything.. trying to be more incremental, but it's not particularly easy in some cases, especially since some of his ideas are still quite fuzzy Aug 14 03:16:08 I know that feeling :) Aug 14 03:18:30 i actually started coding something to take a recipe in our repo and convert it into a git repo of the source, with topic branches for our patches and everything, retaining our history of recipe changes and patch changes... it was twisted Aug 14 03:18:41 * kergoth needs to figure out what he did with that, so he can try to finish it someday Aug 14 03:22:24 hi Aug 14 03:22:33 sombody here? Aug 14 03:22:52 no Aug 14 03:23:05 everybody Aug 14 03:23:08 is Aug 14 03:23:09 gone Aug 14 03:23:18 :) Aug 14 03:23:37 haha Aug 14 03:23:46 I've have a simple question Aug 14 03:25:02 I'm new in oe Aug 14 03:25:04 sushisan: welcome to irc, just ask... maybe you get an answer maybe no Aug 14 03:25:25 I'm constructing a kaelios image...something generic Aug 14 03:25:42 but I want to modify the x11-image standard Aug 14 03:26:19 and put my nre bb somewere outside oe directoy (may be in kailos directory) Aug 14 03:26:40 what env variable can I modify to add my new image? Aug 14 03:27:30 BBFILES or COLLECTIONS. add the directory where your recipes are to whichever of those your'e currently setting. Aug 14 03:27:38 and it would be in local.conf, not the env, most likely Aug 14 03:29:17 I have in build/kaeilos/conf/site.conf Aug 14 03:29:19 BBFILES = "/home/koan/devel/openembedded/recipes/*/*.bb" Aug 14 03:29:46 put my directory here? Aug 14 03:29:51 somthing like: Aug 14 03:30:17 BBFILES = "/home/koan/devel/openembedded/recipes/*/*.bb /my_new/director/recipes/*/*.bb" Aug 14 03:30:39 ? Aug 14 03:31:21 assuming your recipe is in /my_new/directory/recipes/*/, yep Aug 14 03:31:30 TXS!! Aug 14 03:33:21 zecke: shouldn't you be outside rescueing some people? Aug 14 03:34:18 J13R: I don't have a helicopter Aug 14 03:36:21 zecke: awake godzilla or something Aug 14 03:57:09 connmand also has issue on mipsel when compiled with -fPIC. Aug 14 03:57:24 grg: ouch Aug 14 03:57:39 grg: is that normal? Aug 14 03:57:43 methinks libdbus is the common problem with connmand and dbus-daemon Aug 14 03:58:09 other applications are successful after being built with PIC/PIE Aug 14 04:02:59 maybe i'm wrong. connmand doesn't get built with PIC/PIE, just the plugins and libs. Aug 14 04:08:31 grg: which normally is okay :} Aug 14 04:08:47 i suppose Aug 14 04:09:10 all .so files are PIC Aug 14 05:16:42 I give up Aug 14 05:16:54 (porting linux to hx2750) Aug 14 05:18:33 even though little is missing to complete the boot, neither bluetooth drivers nor wireless drivers seem to be supported, so this: http://thera.be/my_public/my_projects/HX2790/2/ is still the best variant. Aug 14 05:49:39 03Sebastian Krzyszkowiak  07shr/import * r7ac03b351f 10openembedded.git/ (3 files in 3 dirs): e-wm-theme-illume-niebiee: add new recipe Aug 14 06:34:43 03Klaus Kurzmann  07shr/import * r457ac4d920 10openembedded.git/recipes/intone/intone_svn.bb: Aug 14 06:34:43 intone: update PV Aug 14 06:34:43 Signed-off-by: Klaus Kurzmann Aug 14 06:45:24 Goodmorning Aug 14 06:46:58 I created an console-image and I think there is a script in it wich sets the time (at first boot) according to a timestamp in the image. Does anyone know where this is done? Aug 14 06:48:22 (I want to disable this, because I set hwclock with ntp in the setup) Aug 14 06:54:58 tsjsieb, /etc/init.d/bootmisc.sh Aug 14 06:56:40 Thnx grg Aug 14 07:03:39 03Denys Dmytriyenko  07org.openembedded.dev * r9e2e5d525f 10openembedded.git/recipes/u-boot/u-boot_git.bb: u-boot: bump the SRCREV's for DaVinci boards Aug 14 07:41:47 03Denys Dmytriyenko  07org.openembedded.dev * r1708207a54 10openembedded.git/classes/kernel.bbclass: Aug 14 07:41:47 kernel.bbclass: introduce MODULES_IMAGE_BASE_NAME Aug 14 07:41:47 Similar to KERNEL_IMAGE_BASE_NAME, allow overwriting from a recipe Aug 14 07:41:47 Signed-off-by: Denys Dmytriyenko Aug 14 07:41:47 Acked-by: Henning Heinold Aug 14 07:50:36 03Steve Sakoman  07org.openembedded.dev * rc5d2c1b6b2 10openembedded.git/ (6 files in 2 dirs): libgnomeui: add 2.24.1 and restructure to use libgnomeui.inc Aug 14 07:50:37 03Steve Sakoman  07org.openembedded.dev * r9671737de1 10openembedded.git/ (conf/checksums.ini recipes/gnome/libgweather_2.26.2.1.bb): libgweather: add 2.26.2.1 Aug 14 07:50:37 03Steve Sakoman  07org.openembedded.dev * r28c0117b46 10openembedded.git/ (conf/checksums.ini recipes/gnome/libbonoboui_2.24.1.bb): libbonoboui: add 2.24.1 Aug 14 09:30:18 hrw|gone: awake yet? Aug 14 09:56:37 good morning Aug 14 09:57:24 good morning Aug 14 10:11:58 what's the proper way to override /etc/network/interfaces for a specific image/machine? Aug 14 10:15:17 hallo Aug 14 10:15:24 hi Aug 14 10:36:03 03Koen Kooi  07org.openembedded.dev * r0740cae859 10openembedded.git/recipes/gnome/libgweather_2.26.2.1.bb: libgweather: fix packaging regression Aug 14 10:41:09 the linux-omap for the gumstix build is interesting - the do_fetch fetches something an empty git repo Aug 14 11:29:12 how can I get the netbase package to generate a machine specific version with my interfaces file? I tried adding netbase/mymachine/interfaces but that doesn't seem to be enough... Aug 14 11:35:23 never mind seemed that a clean did it ;) Aug 14 11:35:54 03Graeme Gregory  07org.openembedded.dev * r5b63b14a98 10openembedded.git/recipes/linux/linux-omap_git.bb: Aug 14 11:35:54 linux-omap_git.bb : increase the revision and fix PV accordingly, this Aug 14 11:35:54 version boots on beagle once more Aug 14 11:36:00 03Graeme Gregory  07org.openembedded.dev * r3782caa265 10openembedded.git/recipes/linux/linux-omap-2.6.31/beagleboard/defconfig: Aug 14 11:36:00 linux-omap-2.6.31/beagleboard/defconfig : add defconfig for beagle to this Aug 14 11:36:00 kernel Aug 14 11:36:05 03Graeme Gregory  07org.openembedded.dev * rcc4408884f 10openembedded.git/: Merge branch 'org.openembedded.dev' of git+ssh://git@git.openembedded.org/openembedded into org.openembedded.dev Aug 14 11:37:09 Which are the places to look for the kernel PRs ?for gumstix, the do_fetch fetches nothing and do_patch obviously fails. Aug 14 11:43:51 mickey|dinner: http://pastebin.com/f5c08d9d9 Aug 14 11:45:04 hi all Aug 14 11:45:38 I have a question Aug 14 11:47:02 I need to build a console rootfs which contains qt3 qt4 support and qws server of qt. Which recipe I should use? Aug 14 11:48:15 sx: check your openembedded/recipes/images dir for an appropriate console image and then add the required packages to that image Aug 14 11:48:31 sx: http://bec-systems.com/web/content/view/79/9/ Aug 14 11:49:08 * rkirti quotes cbrake_away's articles for the n'th time Aug 14 11:50:38 sx: why Qt3 and Qt4? Aug 14 11:50:39 rkirti, thanks. Aug 14 11:51:37 zecke, actually I didn't decide yet which version I'll use Aug 14 11:52:13 zecke, may be only qt4 is enough Aug 14 11:52:59 does someone knows the name of the qws server's package on OE ? Aug 14 11:54:15 sx: there is no qws server package Aug 14 11:54:27 sx: with Qt Embedded Linux any app can play server by starting it with -qws Aug 14 11:54:37 sx: or using a special QApplication c'tor Aug 14 11:56:21 zecke, hmm. then I should add only qt-embedded package right? Aug 14 11:56:43 rkirti: next time just point to the FAQ section in the wiki Aug 14 11:56:45 http://wiki.openembedded.net/index.php/I_want_an_image_with_package_XYZ_installed Aug 14 11:56:58 I added your link to the existing wiki page Aug 14 11:57:13 Laibsch: will do that, thanks :-) Aug 14 11:57:30 zecke: do we have a Qt4 image in OE? Aug 14 11:57:46 I was wondering if I can migrate from opie to Qt4 Aug 14 11:57:48 sx: well, we have no contractal relationship, so I don't tell you what to do :) Aug 14 11:58:33 sx: qt4-embedded will pull everything into the image, all examples and demos, plugins Aug 14 11:58:47 Laibsch: Opie is a full environment, Qt4 is a framework Aug 14 11:58:58 details, details ;-) Aug 14 11:59:10 uff, as always quite complicated Aug 14 11:59:16 * rkirti notices bitbake behaving oddly - do_fetch fails yet it continues to patch Aug 14 11:59:27 let me put forward my real question, then Aug 14 11:59:48 zecke: you are probably aware that I want OE derived images to do some Japanese stuff Aug 14 12:00:26 I am more and more afraid opie may be bitrotted beyond hope by now (font support comes to mind, touchscreen is another problem area) Aug 14 12:00:59 some apps I need and are only available as opie apps Aug 14 12:01:03 what are my options? Aug 14 12:01:19 which are the opie-only apps that you need? Aug 14 12:01:19 I'm afraid the answer is "complete rewrite" one way or another Aug 14 12:01:25 zten for example Aug 14 12:01:29 if they are fundamentally qt apps rather than opie apps, you might be able to run them against qt-x11 Aug 14 12:02:01 it's a dictionary app that can read EPWING dictionaries Aug 14 12:02:13 I'm not aware of another app, although there may be one out there. Aug 14 12:02:31 pb__: how would I do that? compile an x11-image and then? Aug 14 12:02:52 do I need any extra package for qt-x11 or are they already included? Aug 14 12:03:47 Laibsch: in theory it should be as simple as staging qt-x11 and then compiling your apps. in practice I think there are probably a few minor adjustments needed to the application code in moving from qt/eleet to qt/x11 but I'm not exactly sure what those would be. zecke might be your resource, he has the qt 5ki11z. Aug 14 12:04:05 OK Aug 14 12:04:12 I'll try to go that path Aug 14 12:04:25 First I have to get mouse support working in qemu, though Aug 14 12:04:32 whether that's practical depends on whether zten has any serious opie dependencies or not. if it wasn't developed specifically for opie then you'd probably be ok. Aug 14 12:04:38 Interestingly enough, poky qemu images work Aug 14 12:17:05 Can I create an hidden folder in a recipe? I tried 'install -m 0700 -d ${D}/home/root/.ssh' and It does show up in the work dir, but nog in the rootfs Aug 14 12:17:16 nog -> not Aug 14 12:18:19 * Laibsch remembers something about "never touch /home from inside OE" Aug 14 12:19:10 tsjsieb: you might need to mention it explicitly in FILES, I don't think the normal globs will catch it. Aug 14 12:20:05 pb__ thnx, I will look for that Aug 14 12:21:52 Laibsch: why would that be inappropriate? (I'm trying to add ssh authorized_keys file for a user, I don't think I can manage that in another way) Aug 14 12:23:22 (but I'm still quite new to Linux, so please correct me If I'm doing it the wrong way) Aug 14 12:23:45 I only remember vaguely Aug 14 12:23:56 koen said something about that some time in the past Aug 14 12:24:01 and it made sense to me, then Aug 14 12:24:44 Laibsch: port them to Qt4 Aug 14 12:25:01 zecke: I'm afraid that will likely be beyond my skill Aug 14 12:25:12 But I'll come back to you with questions ;-) Aug 14 12:25:21 I hope I can actually make it that far Aug 14 12:25:23 Laibsch: is a japanese input method there as well? Aug 14 12:25:34 Laibsch: no one else is maintaining these apps? Aug 14 12:25:43 Been fighting with problems much more elementary for quite some time now Aug 14 12:25:50 like getting that darn mouse working Aug 14 12:26:05 Laibsch: create a project on gitorious, checkin the latest code, add me to the project Aug 14 12:26:05 zecke: I don't think they are widely used Aug 14 12:26:18 Japanese input method is important, too, of course Aug 14 12:26:32 and it's not quite clear which one will be the best Aug 14 12:26:37 Laibsch: checkin the apps, porting should be mechanical and I can do that at airports... Aug 14 12:26:55 It's not even clear there is an appropriate one for touchscreen use (there was one for opie) Aug 14 12:27:10 OK, will do that Aug 14 12:27:12 Thank you Aug 14 12:27:28 * Laibsch visits gitorious for the first time Aug 14 12:50:18 why did patchwork not pick up the patch I sent to the mailing list about ftp.matrix.com.br no longer existing? It does have [oe][PATCH] in the subject line Aug 14 12:50:36 pb__: would you be so kind to give me an ack for that patch so that I can push it? Aug 14 12:52:13 Laibsch: it stopped working after the 12th Aug 14 12:54:10 aha Aug 14 12:54:15 Anybody looking into it? Aug 14 12:54:26 * Laibsch suspects mail stuck at ltg Aug 14 12:54:27 ;-) Aug 14 12:55:08 Let's see if anything changed on melo around the 12th Aug 14 12:57:22 Last significant change was on the morning of the 11th Aug 14 12:57:32 But I can't spot anything obviously related Aug 14 12:58:01 Laibsch: I think zecke already acked it Aug 14 12:58:42 but it looks fine to me, too. I agree that ftp.matrix.com.br appears to be out of action. Aug 14 12:59:21 AFAIK, I need two ACKs Aug 14 12:59:37 oh, do you? I am a bit baffled by the rules. Aug 14 12:59:42 okay, I will send another Aug 14 13:00:21 there you go Aug 14 13:00:41 http://wiki.openembedded.net/index.php/Commit_Policy Aug 14 13:00:47 two or more core devs Aug 14 13:01:00 I read that as "proponent + 2 ACKs" Aug 14 13:01:06 Thanks, pb Aug 14 13:01:22 ah yes, you're right. Aug 14 13:01:46 I'm not sure whether I qualify as a "core developer" or not, but I think you might as well go ahead and check it in anyway. I can't imagine that the patch will be controversial. Aug 14 13:02:34 also, I'm pretty sure that lots of other people check their changes in with just one ack, or indeed none at all :-} Aug 14 13:02:43 Crofton|work: Can you look into the patchwork issue when you have time? It seems the patchwork queue doesn't list any patches that arrived after the 12th. Let me know if there is anything I can do to assist you. Aug 14 13:17:20 03Rolf Leggewie  07org.openembedded.dev * rb039332721 10openembedded.git/classes/base.bbclass: Aug 14 13:17:20 base.bbclass: ftp.matrix.com.br no longer exists Aug 14 13:17:20 * Acked-By: Holger Freyther Aug 14 13:17:20 * Acked-by: Phil Blundell Aug 14 13:36:14 what's wrong with cgit? Aug 14 13:40:43 booxter: why? Aug 14 13:40:58 hi Aug 14 13:41:17 someone know how to make osx terminal pass Alt as Alt during ssh session? Aug 14 13:41:19 Laibsch: could this be your mirror thing or my change to cgitrc to add a new repo? Aug 14 13:41:35 zecke: http://cgit.openembedded.org shows index of files, not cgit interface Aug 14 13:41:50 I'll take a look Aug 14 13:45:22 My apologies for the interruption Aug 14 13:45:30 Things should be back in working order now Aug 14 13:45:38 My vhost definition was buggy Aug 14 13:46:01 zecke, booxter: fixed Aug 14 13:58:53 Laibsch: thanks Aug 14 14:06:36 Laibsch: tnx Aug 14 14:18:05 Hey guys Aug 14 14:18:13 Is there any project for an image for La Fonera ? Aug 14 14:23:40 by default when my image boots the first time it runs some configuration of packages. can I stop it from doing this? I'm actually using an immutable image so next time it reboots, the changes will be gone and it has to do it every time it boots Aug 14 14:31:13 morning all Aug 14 14:31:21 hi RP Aug 14 14:57:06 hi Aug 14 14:58:16 i just diagnosed a problem in bitbake and oe-stable and paralel builds. Aug 14 14:58:19 my oe-stable is a few days old so let's see if this is known and fixed already Aug 14 14:59:14 what happens (once in a while) is that some build process uses commands from coreutils-native while coreutils-native is being staged Aug 14 14:59:37 let's say the command is 'rm' Aug 14 15:00:01 so coreutils-native stages 'rm' while another build process trys to delete some file Aug 14 15:00:28 this leads to "rm: text file busy" Aug 14 15:00:42 because rm is still open for writing Aug 14 15:00:55 has anyone seen this? and is there a known fix or workaround? Aug 14 15:04:14 morning Aug 14 15:04:19 is anyone listening? Aug 14 15:05:43 I'll post to the list. have a nice weekend guys Aug 14 15:11:01 hi kergoth Aug 14 15:12:48 hey pb__ Aug 14 15:29:12 hi Aug 14 15:29:28 what is the command to push only up to a certain git revision? Aug 14 15:29:40 Let's say I have a-b-c commits locally Aug 14 15:29:48 But I only want to push c right now Aug 14 15:30:00 git push c:? Aug 14 15:30:13 well, is c on top of b, or vice versa? Aug 14 15:30:15 I can't get the thing to the right of the : right Aug 14 15:30:28 a on top of b on top of c Aug 14 15:30:41 a-b-c-remoteHEAD Aug 14 15:31:04 ah, right, so no need to rebase it first Aug 14 15:31:09 just git push c:org.openembedded.dev Aug 14 15:31:10 or whatever Aug 14 15:31:16 nothing to it Aug 14 15:31:20 :) Aug 14 15:34:16 03Rolf Leggewie  07org.openembedded.dev * r986c562d42 10openembedded.git/conf/checksums.ini: checksums.ini: remove obsolete and duplicate entry for sazanami tarball Aug 14 15:34:34 Well, here I needed a "git push oe $REV:org.openembedded.dev" Aug 14 15:34:48 Thanks for confirming that I was generally marching in the right direction. Aug 14 15:36:31 ah, right Aug 14 15:36:32 yeah Aug 14 15:36:47 the key is to remember that its the remote ref not including the remote or refs/heads or anything Aug 14 15:36:49 just the ref Aug 14 15:37:04 i.e. git push origin foo:master, not git push origin foo:refs/heads/master or anything Aug 14 16:04:40 RP: zaurusd applies to collie as well, doesn't it? I'm wondering if zaurusd shouldn't be added to MACHINE_EXTRA_RDEPENDS just like the rest of the Zaurus family. Aug 14 16:06:42 hmm Aug 14 16:07:07 http://svn.o-hand.com/repos/misc/trunk/zaurusd/config/machines would support that conclusion Aug 14 16:22:47 03Rolf Leggewie  07org.openembedded.dev * r3e633b49e2 10openembedded.git/conf/machine/include/zaurus-2.6.inc: Aug 14 16:22:47 zaurus-2.6.inc: add zaurusd to collie RDEPENDS Aug 14 16:22:47 * there is no need to deviate from the other Zaurus machines here Aug 14 16:22:47 http://svn.o-hand.com/repos/misc/trunk/zaurusd/config/machines Aug 14 16:22:47 * regroup some fields while we're at it Aug 14 16:49:28 dumb question, what is gles? It breaks my build for qt4x11 Aug 14 16:49:56 subset of GL Aug 14 16:50:34 hm qt4x11 should use mesa-GL Aug 14 16:52:28 there is a qt4-x11-free-gles recipe Aug 14 16:52:40 and unfortunately in my case, that is the one that gets picked Aug 14 16:52:56 but apparently the necessary driver is available only for omap3 Aug 14 16:53:09 virtual/egl Aug 14 16:53:35 right Aug 14 16:53:43 only omap3 so far Aug 14 16:53:54 intel poulsbo would be a second Aug 14 16:54:06 but to get the driver is the same as for omap3 Aug 14 16:54:22 only with some effort Aug 14 16:57:07 woglinde: I wonder how to get around the issue Aug 14 16:57:19 what are you doing exactly? Aug 14 16:57:24 weakly define qt4-x11-free as the default somewhere Aug 14 16:57:43 and override that to qt4-x11-free-gles in omap3 machine config? Aug 14 16:58:21 woglinde: "bitbake anki" Aug 14 16:58:51 NOTE: multiple providers are available for qt4x11 (qt4-x11-free-gles, qt4-x11-free); Aug 14 16:59:01 hehe Aug 14 16:59:05 So, I guess the solution is to define qt4-x11-free as deault Aug 14 16:59:06 so it depends on you Aug 14 16:59:08 default Aug 14 16:59:26 and let the omap people override it somewhere Aug 14 17:00:06 well, it's broken OOTB for most people, so it does and doesn't depend on me Aug 14 17:00:20 I'd like to fix it so that it generally works as expected Aug 14 17:09:05 * * OE Bug 5047 has been RESOLVED (INVALID) by Aug 14 17:09:07 * * compiling bluez problem Aug 14 17:09:09 * * http://bugs.openembedded.net/show_bug.cgi?id=5047 Aug 14 17:15:39 * kergoth wonders if things would build fine if we just killed the .la files everywhere Aug 14 17:16:50 ping kergoth about Aug 14 17:16:54 !oebug 4452 Aug 14 17:16:55 * * Bug 4452, Status: NEEDSINFO, Created: 2008-07-23 22:25 Aug 14 17:16:56 * * kergoth(AT)gmail.com: Fetchers not run with a clean environment causes wget to obey user wgetrc. Aug 14 17:16:57 * * http://bugs.openembedded.org/show_bug.cgi?id=4452 Aug 14 17:17:11 might have time to look at it this weekend Aug 14 17:17:27 kergoth: yeah, I've wondered that as well. I've never been completely clear on what the .la files are good for. Aug 14 17:19:18 i guess it was to handle lame linkers that can't handle one shared lib linking against another, it would have the dependency_libs specified in the .la? Aug 14 17:19:19 * kergoth shrugs Aug 14 17:19:46 regardless, we shouldn't need it now. for stuff being in non-standard locations, you can be explicit or use pkg-config, don't see the point in keeping them around at all, and libtool can function without them, as far as i can tell.. Aug 14 17:19:54 if we remove them, we don't have to deal with mangling them all to hell Aug 14 17:20:44 * kergoth tries a build without them Aug 14 17:21:04 * * OE Bug 1814 has been RESOLVED (LATER) by Aug 14 17:21:06 * * rx3000 support in irda-utils Aug 14 17:21:08 * * http://bugs.openembedded.net/show_bug.cgi?id=1814 Aug 14 17:21:32 XorA: is bug 1784 still an issue? Aug 14 17:21:35 has anyone run into issues with udev not creating certain devices on boot? Aug 14 17:21:43 !oebug 1784 Aug 14 17:21:43 a udevadm trigger after that results in them showing up Aug 14 17:21:44 * * Bug 1784, Status: NEEDSINFO, Created: 2007-01-12 14:21 Aug 14 17:21:45 * * ThoughtMonster(AT)gmail.com: Abiword exits when "italic" text formatting is selected. Aug 14 17:21:46 * * http://bugs.openembedded.org/show_bug.cgi?id=1784 Aug 14 17:21:47 so it seems to be just on boot Aug 14 17:22:47 kergoth: Have you looked in the bug tracker? I remember seeing some tickets that at least sound like they may be familiar. Aug 14 17:22:55 http://bugs.openembedded.org/buglist.cgi?quicksearch=udev Aug 14 17:23:11 !oebug 1784 Aug 14 17:23:13 * * Bug 1784, Status: NEEDSINFO, Created: 2007-01-12 14:21 Aug 14 17:23:14 * * ThoughtMonster(AT)gmail.com: Abiword exits when "italic" text formatting is selected. Aug 14 17:23:15 * * http://bugs.openembedded.org/show_bug.cgi?id=1784 Aug 14 17:23:42 Laibsch: probably Aug 14 17:23:59 OK, thanks Aug 14 17:24:30 Maybe you can test if you have a spare minute Aug 14 17:24:40 Laibsch: good point, i'll look through those, thanks Aug 14 17:25:01 Laibsch: no hardware to test on Aug 14 17:25:18 bummer Aug 14 17:25:29 You have no embedded HW anymore? Aug 14 17:25:50 Shouldn't it be possible to confirm even in qemu? Aug 14 17:26:06 I'll add you to the cc of the bug if you don't mind? Aug 14 17:26:49 hmm Aug 14 17:36:37 meh Aug 14 17:55:55 woglinde: you look at mips stuff from time to time don't you? Aug 14 17:56:03 !oebug 5061 Aug 14 17:56:04 * * Bug 5061, Status: UNCONFIRMED, Created: 2009-03-10 03:26 Aug 14 17:56:05 * * yajinzhou(AT)vm-kernel.org: gpe-scap-1.3 mips-strdup patch Aug 14 17:56:06 * * http://bugs.openembedded.org/show_bug.cgi?id=5061 Aug 14 17:56:32 Laibsch hm Aug 14 17:56:39 !oebug 2453 Aug 14 17:56:40 not really Aug 14 17:56:40 * * Bug 2453, Status: CONFIRMED, Created: 2007-06-09 00:38 Aug 14 17:56:41 * * jake(AT)cogini.com: MIPS patches for glibc-2.3.3 from linux-mips.org Aug 14 17:56:42 * * http://bugs.openembedded.org/show_bug.cgi?id=2453 Aug 14 17:56:55 let me see Aug 14 17:57:02 but I think you close it Aug 14 17:57:06 OK, I'm mistaking you for somebody else, then Aug 14 17:57:11 much too old Aug 14 17:57:22 Could still be valid Aug 14 17:57:45 I don't usually close bugs only because they are old Aug 14 17:58:04 laibsch read the last commit Aug 14 17:58:08 aeh comment Aug 14 17:58:10 and close it Aug 14 17:58:34 I never support kernel 2.4 again Aug 14 18:00:52 the way I understand it, it would be applicable to 2.6 as well, no? Aug 14 18:01:04 Or is this a 2.4-only patch? Aug 14 18:01:16 I'm not even sure we still have 2.4 kernel recipes in OE Aug 14 18:01:28 Laibsch please close it Aug 14 18:02:15 or would like to provide a generic framework in oe to build a toolchain with selectable tls Aug 14 18:05:45 hmm Aug 14 18:07:53 * kergoth thinks about $ORIGIN Aug 14 18:17:16 hmm, wonder if our toolchains are built to pass --enable-new-dtags to ld Aug 14 18:25:04 * * OE Bug 2453 has been RESOLVED (WONTFIX) by Aug 14 18:25:06 * * MIPS patches for glibc-2.3.3 from linux-mips.org Aug 14 18:25:08 * * http://bugs.openembedded.net/show_bug.cgi?id=2453 Aug 14 18:41:18 * kergoth tries to figure out if the benefits of private staging are worthwhile Aug 14 18:42:58 khem: ping Aug 14 18:43:39 john3909: hi Aug 14 18:43:48 khem: Did you make any progress on resolving the problem with building gnome-games? Aug 14 18:44:08 john3909: I did not get enough info from you Aug 14 18:44:22 could you show me the run.do_compile* file Aug 14 18:44:35 khem: sure Aug 14 18:44:50 is there a way I can force bitbake to show me what conf files its including something like gcc -E Aug 14 18:44:53 :) Aug 14 18:45:23 khem: I'm not sure Aug 14 18:45:43 re Aug 14 18:45:55 * khem tried -D Aug 14 18:46:25 ah bingo Aug 14 18:46:32 khem: this is the run.do_compile I pasted yesterday http://pastebin.com/m46f774cd Aug 14 18:47:05 * * OE Bug 5296 has been created by ospite(AT)studenti.unina.it Aug 14 18:47:06 * * libfsobasics_git.bb fails to configure Aug 14 18:47:09 * * http://bugs.openembedded.org/show_bug.cgi?id=5296 Aug 14 18:47:20 john3909: try to run this file indvidually Aug 14 18:47:28 do you end up in same error ? Aug 14 18:47:40 khem: let me try Aug 14 18:50:26 khem: I have to put things back as I was doing some experimenting Aug 14 18:52:33 03Khem Raj  07org.openembedded.dev * r786619e5a2 10openembedded.git/recipes/libtool/ (20 files): Aug 14 18:52:33 libtool: Remove FILESDIR, use INC_PR and FILESPATHPKG Aug 14 18:52:33 Signed-off-by: Khem Raj Aug 14 18:52:33 Acked-by: Denys Dmytriyenko Aug 14 18:52:39 03Khem Raj  07org.openembedded.dev * rdc0604b297 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.org:openembedded into org.openembedded.dev Aug 14 18:52:54 john3909: ok no problem Aug 14 18:52:56 try it out Aug 14 18:55:42 03Rolf Leggewie  07org.openembedded.dev * rfd0274113e 10openembedded.git/recipes/bluez/bluez4.inc: Aug 14 18:55:42 bluez4: make it generally provide bluez-libs and bluez-utils Aug 14 18:55:42 * reorder fields according to Style Guide Aug 14 18:58:58 khem: Yes, I get the same error Aug 14 18:59:34 john3909: hmm ok Aug 14 19:00:36 khem: if you do bitbake -c build gnome-games, do you not get the same error? Aug 14 19:00:38 error was missing /SDL Aug 14 19:00:57 john3909: I do not have build env right now where I could do that Aug 14 19:01:06 OK Aug 14 19:02:33 khem: I think this has something to do with 43100bfa8 posted by Roger Monk Aug 14 19:03:39 khem: The sed change is causing the -/SDL error Aug 14 19:05:34 is that a pastebin id or what Aug 14 19:06:19 khem: no it's the SHA1 ID from the OE GIT Aug 14 19:06:29 ah Aug 14 19:06:32 lemme check Aug 14 19:17:07 john3909: Can you add - infront of the I/usr/include in that sed expression ? Aug 14 19:17:26 khem: yes Aug 14 19:18:37 and do bitbake -c clean gnome-games; bitbake gnome-games Aug 14 19:19:15 khem: still gives and error Aug 14 19:20:19 what is the error this time Aug 14 19:21:24 khem: arm-angstrom-linux-gnueabi-gcc: unrecognized option '-/SDL' Aug 14 19:21:52 khem: seems strange, as I would have expected the - to disappear Aug 14 19:22:01 yeah Aug 14 19:22:30 Did u clean it Aug 14 19:22:38 khem: I'm modifying in the run.do_compile and then running it Aug 14 19:22:47 khem: no Aug 14 19:23:28 khem: should I run it through bitbake? Aug 14 19:23:56 khem: modify the gnome-games recipe? Aug 14 19:25:48 ah Aug 14 19:25:54 no modify the ecipe Aug 14 19:26:06 khem: OK, I"ll do that Aug 14 19:26:10 and then bitbake -c clean gnome-games; bitbake gnome-games Aug 14 19:27:58 khem: doing that now Aug 14 19:29:45 khem: clean done; now building Aug 14 19:30:56 khem: shouldn't this be $SDL? Aug 14 19:32:46 khem: SDL points to oe/openembedded/tmp/work/armv7a-angstrom-linux-gnueabi/gnome-games-2.24.0-r5/temp Aug 14 19:35:22 khem: failed again Aug 14 19:36:47 khem: arm-angstrom-linux-gnueabi-gcc: /SDL: No such file or directory Aug 14 19:37:01 hmm ok Aug 14 19:41:50 meh Aug 14 19:42:41 john3909: try setting --with-sdl-prefix Aug 14 19:43:37 khem: where do I do that? Aug 14 19:43:57 Add it to EXTRA_OECONF Aug 14 19:44:13 khem: OK Aug 14 19:44:19 and may be set it to $SDL Aug 14 19:44:29 hmm, i think we need to make it less painful to import a custom .py from OpenEmbedded metadata Aug 14 19:44:55 maybe just add /lib for each dir in bbpath to sys.path Aug 14 19:45:14 khem: I have changed it to ${SDL} and building now Aug 14 19:46:25 khem: it doesn't make sense, as SDL is referencing the temp folder Aug 14 19:47:01 john3909: I just want to make sure that it tried to find there we can set it to correct path later Aug 14 19:47:09 khem: the gnome_games temp folder Aug 14 19:47:38 OK Aug 14 19:53:15 khem: games-sound.c:30:19: error: SDL.h: No such file or directory Aug 14 19:53:40 khem: and other related errors Aug 14 20:00:49 ok so it is tryin to look for sdl.h Aug 14 20:00:59 do you have it in yout staging area ? Aug 14 20:01:14 khem: let me check Aug 14 20:05:00 khem: It is not in staging/armv7a-angstrom-linux-gnueabi/usr/include Aug 14 20:05:37 Is there a way to globally specify a preferred_provider of something and deviate from it only for a single machine? Aug 14 20:05:44 ... outside of distro configs? Aug 14 20:05:55 khem: but it is in staging/armv7a-angstrom-linux-gnueabi/usr/include/SDL Aug 14 20:06:28 khem: OK, now I know what to do. I'll work on finding a solution Aug 14 20:08:27 Laibsch: PREFERRED_PROVIDER__ = "" Aug 14 20:09:15 s/distro/machine Aug 14 20:10:52 yes, but in what file? Aug 14 20:11:28 I am aware of overrides, I'm not sure where to place the global preference Aug 14 20:11:43 local.conf Aug 14 20:11:52 should be the highest priority Aug 14 20:12:29 global = I want it in OE.dev ;-) Aug 14 20:12:44 hmm Aug 14 20:12:49 my apologies for being unclear Aug 14 20:13:46 machine conf file is probably the second in row Aug 14 20:25:57 Yeah, I was going to use that for the exception Aug 14 20:26:13 But defining the default would mean I'd have to touch all machine configs Aug 14 20:26:34 And it's really only machine dependent in the case of the exception Aug 14 20:26:59 I guess maybe there is no other way than distro config :-/ Aug 14 20:27:16 I would say machine config Aug 14 20:27:27 if it is machine specifiv Aug 14 20:35:55 anybody familiar with how AUTOREV works? Aug 14 20:37:35 a little Aug 14 20:39:27 Crofton: I guess you're also mildly familiar with virtual/egl? Aug 14 20:39:37 no Aug 14 20:39:40 is there any way to make it identify the REV in the final package? Aug 14 20:39:46 hmmm Aug 14 20:39:52 I don't know Aug 14 20:39:56 I thought it did? Aug 14 20:40:51 basically, the problem is if your PV contains reference to SRCREV, which is set to AUTOREV, everything breaks when resolving PV... Aug 14 20:40:59 urg Aug 14 20:42:44 something like PV = "1.2.3+gitr${SRCREV}" would break Aug 14 20:42:56 so, SRCREV cannot be referenced for AUTOREV to work, and resulting package has no indication which rev or commit# is contains... Aug 14 20:43:30 it contains Aug 14 20:45:23 has anybody looked into this? I know AUTOREV is not really popular among OE/distro devs, but it's a nice way for continued dev/build cycle Aug 14 20:50:18 I've used it, but I do not remember how it behaves Aug 14 20:50:44 AUTOREV is result of a python function Aug 14 20:51:24 wild guess may be SRCREV := "${AUTOREV}" Aug 14 20:51:26 will work Aug 14 20:55:12 I can try that... Aug 14 20:55:57 nope, fails even sooner :) Aug 14 20:56:16 same message - ERROR: SRCREV was used yet no valid SCM was found in SRC_URI Aug 14 20:56:53 XorA: you have an omap board? Aug 14 20:57:01 Laibsch: two Aug 14 20:57:55 nice Aug 14 20:58:04 khem: the message is not exactly correct, but the problem is that when PV=${SRCREV=${AUTOREV}} is evaluated, SRC_URI is not yet set properly from the recipe... Aug 14 20:58:35 denix: thats why := Aug 14 20:58:42 XorA: does something like http://paste.debian.net/44233/ make sense? It would solve a problem I have building qt4x11. Aug 14 20:59:03 XorA: currently, it tries to build qt4-x11-free-gles which apparently right now is only applicable to omap. Aug 14 20:59:20 the list of machines may be wrong, I did not yet look up their proper names Aug 14 20:59:29 It's just if this generally makes sense Aug 14 20:59:55 with the right machines it probably makes sense Aug 14 21:00:11 OK Aug 14 21:00:19 thats koens baby Aug 14 21:00:35 Well, then he should have fixed it already ;-) Aug 14 21:00:40 Do all omap boards support this gles thing? Aug 14 21:00:40 denix what if you use autorev direcly in url Aug 14 21:00:55 Laibsch: no Aug 14 21:01:03 Can you tell me which do? Aug 14 21:01:12 I'll see if koen responds to an IRC ping Aug 14 21:01:16 Laibsch: there are different omap3 silicons, and Ive no idea which do or dont Aug 14 21:01:23 OK Aug 14 21:01:29 I'll talk to koen about it Aug 14 21:01:40 Thanks for confirming it should generally be a sane thing to do Aug 14 21:01:54 khem: do you mean in SRC_URI? what to set SRCREV to? Aug 14 21:02:12 I thought the qt4-x11 thing was supposed to be fixed Aug 14 21:03:14 I think it's a different thing Aug 14 21:03:31 Try "bitbake qt4x11" for a machine that is not omap Aug 14 21:03:47 denix: just leave SRCREV as it is Aug 14 21:03:50 for a try Aug 14 21:04:57 * ant__ feels guilty for the noise of yesterday: Gentoo-ccache is not the culprit of failing rebuilds from pstage... Aug 14 21:05:14 ..configure: error: C compiler cannot create executables Aug 14 21:05:17 khem, your patch for libtool seems to fix the build issue I was having with meta-toolchain for the ppc603 machines Aug 14 21:05:38 what's strange this time, is about klcc! Aug 14 21:05:50 tharvey: cool its committed Aug 14 21:06:12 ant__: I added zaurusd to MACHINE_EXTRA_DEPENDS for collie today Aug 14 21:06:16 you may have seen it Aug 14 21:06:22 basically when /cross get repopulated the compilers disappear Aug 14 21:06:31 tharvey: btw I built a toolchain for lite5200 with gcc 4.4.1 eglibc 2.10 and binutils 2.19 yesterday Aug 14 21:06:42 Laibsch: yes, I really have no idea about collie Aug 14 21:07:22 btw, have you had the time to send that serial cable to Pavel? Aug 14 21:09:11 khem, nice - what did you need to tweak for that configuration? distro config? Aug 14 21:10:39 ah I created a new distro :) micro-eglibc and minimal-eglibc Aug 14 21:11:22 makes sense... will you be submitting those as patches? Aug 14 21:11:32 so DISTRO=micro-uclibc MACHINE=lite5200 bitbake meta-toolchain Aug 14 21:11:40 s/uclibc/eglibc/ Aug 14 21:11:46 yes Aug 14 21:12:23 thanks... I'll be on the lookout for that patch - will make my testing a lot easier :) Aug 14 21:19:52 kergoth: in run.do_configure I have export PATH="...:/oe/build/tmp/cross/armv5te/bin:..." but packaged staging recreates /oe/build/tmp/cross/bin and /oe/build/tmp/cross/arm-angstrom-linux-gnueabi/bin Aug 14 21:20:09 can it be an angstrom-only problem ? Aug 14 21:22:11 03Khem Raj  07org.openembedded.dev * rbc159cc29e 10openembedded.git/recipes/gnome/ (files/libproxy-move-define-__USE_BSD.patch libproxy_0.2.3.bb): (log message trimmed) Aug 14 21:22:11 libproxy: Define __USE_BSD before its needed. Aug 14 21:22:11 This fixed a build break on uclibc where Aug 14 21:22:11 it complained Aug 14 21:22:11 armv5te-oe-linux-uclibcgnueabi/usr/include/stdlib.h:215: Aug 14 21:22:15 error: expected '=', ',', ';', 'asm' or '__attribute__' before 'strtouq' Aug 14 21:22:17 The reason is Aug 14 21:22:20 03Khem Raj  07org.openembedded.dev * r068abe84a7 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.org:openembedded into org.openembedded.dev Aug 14 21:22:51 hm..khem... Aug 14 21:23:06 armv5te-oe-linux-uclibcgnueabi rings a bell... Aug 14 21:25:42 how can i change the ext2 total image (default seems to be 64mbyte) Aug 14 21:52:32 set IMAGE_ROOTFS_SIZE_ext2 Aug 14 21:52:51 in which conf file? Aug 14 21:53:29 probably in machine conf file Aug 14 21:53:58 size in bytes? Aug 14 21:54:10 yes Aug 14 21:54:23 thanks :) Aug 14 21:54:24 65536 is for 64k Aug 14 21:54:40 yes i got it ;) Aug 14 21:54:49 np Aug 14 22:45:13 03Khem Raj  07org.openembedded.dev * rad9bb3dc5a 10openembedded.git/site/common-uclibc: (log message trimmed) Aug 14 22:45:13 common-uclibc: Unset guile_cv_use_csqrt for guile Aug 14 22:45:13 guile's configure is not smart to guess if csqrt works Aug 14 22:45:13 in a cross compile environment so it ends up assuming Aug 14 22:45:13 that it works and fails to link. Anyways for uclibc Aug 14 22:45:15 complex functions are unimplemented so we help configure Aug 14 22:45:19 by unsetting here in site file for uclibc. Aug 14 23:58:03 * * OE Bug 5297 has been created by  Aug 14 23:58:05 * * python-pyqt fails do_compile Aug 14 23:58:07 * * http://bugs.openembedded.org/show_bug.cgi?id=5297 Aug 15 00:21:26 khem: ping Aug 15 00:21:48 a poky qemuarm image works fine with the host mouse. An angstrom or minimal x11 image does not. What areas should I have a closer look and inspect the differences between poky and .dev? Aug 15 00:22:33 machine config is an obvious one, anything else? Aug 15 00:23:06 khem: I replaced that line in gnome-games recipe with sed -i -e s:'/usr/include/SDL':/oe/openembedded/tmp/staging/armv7a-angstrom-linux-gnueabi/usr/include/SDL:g -e s:'-I /usr/include -I /usr/local/include'::g $i Aug 15 00:23:26 khem: now gnome-games compiles successfully. Aug 15 01:45:21 hi Aug 15 01:45:32 I have a (simple) question Aug 15 01:45:45 I have a bb with dependences: Aug 15 01:46:36 in may case said in install: Aug 15 01:46:37 IMAGE_INSTALL = "\ ${XSERVER} \ task-base-extended \ task-xterminal \ angstrom-zeroconf-audio \ Aug 15 01:47:14 and XSERVER hacve all modules that I turn on or off with a comentary Aug 15 01:47:28 XSERVER = " \ xserver-xorg \ xf86-input-evdev \ xf86-input-tslib \ # xf86-video-apm \ # xf86-video-ark \ # xf86-video-ast \ # xf86-video-ati \ # xf86-video-chips \ # xf86-video-cirrus \ # xf86-video-cyrix \ xf86-video-dummy \ xf86-video-fbdev \ # xf86-video-geode \ # xf86-video-gl Aug 15 01:47:31 ok? Aug 15 01:47:46 when I bitbake the image all ok Aug 15 01:48:22 but if I turn off some driver in XSERVR variable and remake de image compile and install the package again! Aug 15 02:46:39 john3909: It compiled fine for me Aug 15 02:46:44 with out any change Aug 15 02:47:07 so I dont know why you have problem What host do you use Aug 15 02:47:16 and do you have sdl installed on host **** ENDING LOGGING AT Sat Aug 15 02:59:59 2009