**** BEGIN LOGGING AT Sat Jun 13 02:59:57 2009 Jun 13 04:58:50 Did anybody try installing open embedded on motorala devices? Jun 13 05:21:18 cbrake_away: who is the postmaster on melo? Jun 13 06:18:15 Laibsch, hi did you receive my mail about guile? Jun 13 06:18:35 yes, that it compiles for you Jun 13 06:18:41 Laibsch, yes Jun 13 06:18:44 Don't know anything else Jun 13 06:18:53 I have disabled the recipe locally, here Jun 13 06:19:01 Laibsch, could you post or send me your log files? Jun 13 06:19:06 Maybe I will revisit that stuff one of these days Jun 13 06:19:12 ok Jun 13 06:19:16 Gnutoo: I think I already did, no? Jun 13 06:19:23 ah I'll look again Jun 13 06:19:25 Oh, no Jun 13 06:19:28 There is no log Jun 13 06:19:35 The compilation just hangs Jun 13 06:19:44 I abort with Ctrl+C Jun 13 06:19:49 ok Jun 13 06:20:06 BTW, you added 1.8.6 because of an incompatibility with libtool? Jun 13 06:20:15 yes because it didn't compile for me Jun 13 06:20:21 Interestingly enough, that incompatibility does not exist here Jun 13 06:20:32 Newest libtool, but 1.8.5 compiles fine Jun 13 06:20:52 could you try with a fresh rootfs? or is it not an option for you? Jun 13 06:21:10 you mean TMPDIR? Jun 13 06:21:23 not an option at the moment Jun 13 06:21:32 ok Jun 13 06:21:37 yes I meant that Jun 13 07:15:55 hi **** ENDING LOGGING AT Sat Jun 13 07:19:27 2009 **** BEGIN LOGGING AT Sat Jun 13 07:20:00 2009 Jun 13 07:22:52 morning Jun 13 07:25:47 ircing from train Jun 13 07:26:54 cbrake_away: I packaged tinderbox and will install that on melo in a minute Jun 13 07:27:05 cbrake_away: We need more mysql users Jun 13 07:27:23 AFAICS, we currently don't have one for tinderbox Jun 13 07:48:10 lumdidum Jun 13 07:48:25 is there now postgres backend for tinderbox? Jun 13 07:48:41 no Jun 13 07:52:44 woglinde_ice: tinderbox runs off PG on amethyst **** ENDING LOGGING AT Sat Jun 13 08:24:55 2009 **** BEGIN LOGGING AT Sat Jun 13 08:26:00 2009 **** ENDING LOGGING AT Sat Jun 13 08:27:09 2009 **** BEGIN LOGGING AT Sat Jun 13 08:27:40 2009 Jun 13 08:49:54 03Phil Blundell  07org.openembedded.dev * rde8f4ecb08 10openembedded.git/ (conf/checksums.ini recipes/libidn/libidn_1.15.bb): libidn: add 1.15 Jun 13 08:49:57 03Phil Blundell  07org.openembedded.dev * r03f3d62589 10openembedded.git/recipes/enchant/enchant_1.3.0.bb: enchant: use more sensible FILES Jun 13 08:49:58 03Phil Blundell  07org.openembedded.dev * rae82ac27d6 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@new.openembedded.org:openembedded into org.openembedded.dev Jun 13 08:57:10 sigh, what's this util-linux-ng stuff all about? seems to be teh suck as far as I can tell. Jun 13 09:33:21 03Phil Blundell  07org.openembedded.dev * rf725a81c04 10openembedded.git/recipes/openssl/openssl.inc: openssl: avoid empty PREFIX since this causes Configure to do the wrong thing Jun 13 09:36:58 I am trying to follow the web instruction to bitbake helloworld but I am getting "IO Error: file 'conf/bitbake.conf' not found Jun 13 09:37:13 what should be in the bitbake.conf? how to fix this problem? Jun 13 09:37:21 I am just a newbie, would appreicate any help. Jun 13 09:40:10 whitefox: it's meant to be in your checked-out oe tree Jun 13 09:40:53 i can bitbake base-image though Jun 13 09:41:11 you mean "bitbase base-image" works, but "bitbake helloworld" fails? Jun 13 09:41:29 yes Jun 13 09:41:56 that is very weird. it's hard to imagine what could possibly cause that. Jun 13 09:42:18 each time I do bitbake do I need to do source profile.sh? Jun 13 09:42:27 no, you just need to do that once per session Jun 13 09:46:47 pb_: are you familiar with the .dot files and how to inspect them? I have two weird issues I fail to analyze. Can you take a look? Jun 13 09:49:09 hi, i managed to bitbake now, but I am getting strange error: Jun 13 09:49:10 NOTE: Resolving any missing task queue dependencies Jun 13 09:49:11 ERROR: Required build target 'helloworld' has no buildable providers. Jun 13 09:49:11 Missing or unbuildable dependency chain was: ['helloworld', 'stagemanager-native'] Jun 13 09:49:21 how to fix that? Jun 13 09:50:51 what machine and what distro are you building for? Jun 13 09:51:01 dev or stable? Jun 13 09:51:22 angstrom Jun 13 09:51:40 i am building with Ubuntu host for beagleboard Jun 13 09:52:14 dev or stable? Jun 13 09:52:29 in my local.conf, I have DISTRO = "angstrom-2008.1" Jun 13 09:52:36 how do I tell if it is dev or stable? Jun 13 09:52:50 "git branch" Jun 13 09:52:53 in your repo Jun 13 09:53:04 the one with the asterisk Jun 13 09:53:28 it is stable/2009 Jun 13 09:54:37 I'm on dev, sorry Jun 13 09:54:57 is it something wrong with my profile.sh? Jun 13 09:55:04 i am just compiling helloworld only Jun 13 09:55:16 sorry, I am not even using profile.sh Jun 13 09:55:30 I know nothing about it Jun 13 09:55:59 but a quick inspection looked like there was no difference between dev and stable for recipes/stage-manager/stagemanager-native_0.0.1.bb Jun 13 09:56:21 what is "md5sum recipes/stage-manager/stagemanager-native_0.0.1.bb" for you? Jun 13 09:57:33 it is 7ba55d461930568e3bc954d18fa5f6bc recipes/stage-manager/stagemanager-native_0.0.1.bb Jun 13 09:58:15 sorry, absolutely no idea, then Jun 13 09:58:55 somehow bitbake thinks that stagemanager-native is not buildable for you Jun 13 09:59:53 sill, thanks for all your help Jun 13 11:55:49 bluelightning: hi Jun 13 11:56:00 hi Laibsch Jun 13 11:56:00 are you familiar with the kexecboot thing for spitz? Jun 13 11:56:15 Laibsch: kind of... usage anyway Jun 13 11:56:19 I can't seem to get it to work properly Jun 13 11:56:32 my rootfs is not recognized Jun 13 11:56:37 Here is what I have Jun 13 11:56:53 spitz with kexecboot kernel installed Jun 13 11:57:05 SD card with vfat single partition Jun 13 11:57:38 I think your card needs to be ext2... Jun 13 11:57:41 no matter what I seem to put on that thing, nothing shows up in the kexecboot splash Jun 13 11:57:52 Hm, I can't do that right now Jun 13 11:58:08 do you not have a card reader? Jun 13 11:58:24 I think if you have the rootfs in an ext2 file, it's supposed to work Jun 13 11:58:42 possibly, but I've never tried or read about that Jun 13 11:59:00 no, but all the SD cards I have at my disposition have to stay vfat for now Jun 13 11:59:15 I also tried to flash to the internal flash Jun 13 11:59:18 nothing happens Jun 13 11:59:25 Have you ever accomplished that lately? Jun 13 12:01:48 Laibsch: yes, I use a kexecboot enabled tgz image and flash as normal Jun 13 12:01:49 bluelightning: http://linuxtogo.org/gowiki/AngstromManual says to rename an jffs2 filesystem to initrd.bin and flash Jun 13 12:01:59 tgz image? Jun 13 12:02:12 I thought you have to use an ext2 file? Jun 13 12:02:26 hdimage.tgz... isn't that the standard way of flashing? Jun 13 12:02:41 Hm, I kind of lost track Jun 13 12:03:04 Above site says to "rename your-image-rootfs.jffs2 to initrd.bin and copy it on the card whith updater.sh" Jun 13 12:03:16 I'll try the hdimage.tgz thing Jun 13 12:03:20 Maybe that will do it Jun 13 12:04:39 Laibsch: actually it's called hdimage1.tgz (sorry) Jun 13 12:04:55 yes, I remember that Jun 13 12:05:00 I'll try it in a minute Jun 13 12:05:05 Need to reboot first, though Jun 13 12:05:30 and in case you didn't know, you can leave the zimage.bin off (or renamed) to avoid reflashing the kernel every time Jun 13 12:05:40 (which is not needed with kexecboot) Jun 13 12:08:33 thanks Jun 13 12:11:59 hi, could someone help me with this problem? Jun 13 12:12:00 NOTE: Resolving any missing task queue dependencies Jun 13 12:12:00 ERROR: Required build target 'helloworld' has no buildable providers. Jun 13 12:12:00 Missing or unbuildable dependency chain was: ['helloworld', 'stagemanager-native'] Jun 13 12:13:18 that is not a problem, but just a note to inform you what bitbake is doing Jun 13 12:13:32 Oh, only saw the first line Jun 13 12:13:50 Laibsch how to fix it? Jun 13 12:14:59 I would have told you a few hours ago if I had an idea Jun 13 12:16:24 Laibsch ah i was just hoping... Jun 13 12:16:42 sorry to disappoint ;-) Jun 13 12:16:59 the first line is just for information Jun 13 12:23:41 03Phil Blundell  07org.openembedded.dev * r17f3b570cf 10openembedded.git/recipes/irda-utils/ (irda-utils/sbindir.patch irda-utils_0.9.16.bb): irda-utils: respect $(sbindir) Jun 13 12:27:03 03Phil Blundell  07org.openembedded.dev * rccbfea17e3 10openembedded.git/recipes/tasks/task-gpe.bb: task-gpe: remove xmonobut since SRC_URIs for both versions are unfetchable Jun 13 12:37:36 pb_: ping Jun 13 12:37:39 Laibsch: hello Jun 13 12:37:43 Hi Jun 13 12:38:10 hi Jun 13 12:38:21 you left the xmonobut recipes themselves untouched? Jun 13 12:38:25 yes Jun 13 12:38:39 hm, that is kind of lame IMHO Jun 13 12:39:30 I'm trying to build a console-base-image, it builds averything fine as I can see all the ipk in the deploy/glibc/ipk/* directories, but in the jffs2 image, several packages are not installed. Jun 13 12:39:32 well, there's not much that I can sensibly do to them. I don't know of another SRC_URI, so I can't fix them. on the other hand, for anybody who does have a copy of the tarball, the recipe will still work fine so it doesn't seem like it should be deleted. Jun 13 12:39:44 well, that is true Jun 13 12:39:56 there is always the angstrom source mirror Jun 13 12:40:02 do you know about witpa? Jun 13 12:40:12 no, what's that? Jun 13 12:40:15 somebody should integrate that into bitbake Jun 13 12:40:44 In the end I get only libc6 udev sysvinit tinylogin psplash-angstrom initscripts opkg-nogpg makedevs angstrom-feed-configs sysvinit-pidof psplash-zap psplash-support opkg-collateral & update-rc.d Jun 13 12:41:01 but busybox & sysvinit-inittab are missiing (among other packages) Jun 13 12:41:01 http://thread.gmane.org/gmane.comp.handhelds.openembedded/17776 Jun 13 12:41:16 eric___: I guess you need to check the log to find out why they weren't installed. Jun 13 12:41:35 Hi could someone help me with this error when bitbaking helloworld app? Missing or unbuildable dependency chain was: ['helloworld', 'stagemanager-native'] Jun 13 12:41:43 does anyone know what that mean? Jun 13 12:41:43 pb_: I think witpa is in contrib Jun 13 12:42:46 whitefox: Yes, it means that bitbake does not have a valid recipe for stagemanager-native in your case Jun 13 12:42:48 Laibsch: thanks, I'll hae a look Jun 13 12:42:54 and have a look, too Jun 13 12:43:12 whitefox: the real question is why you don't Jun 13 12:43:29 pb__ : thanks, are you talking of work/machine/console-base-image/temp/log* ? Jun 13 12:43:33 does anyone know why that is the case? Jun 13 12:43:37 eric___: right Jun 13 12:43:40 i am a beginer in oe Jun 13 12:44:07 why you don't have it Jun 13 12:44:31 how do i check it is has or not? Jun 13 12:44:57 whitefox: enable automatix failure reports -> look for it under FAQ in the wiki Jun 13 12:45:03 ~oestats Jun 13 12:45:10 !oestats Jun 13 12:47:50 pb_: it would be awesome if witpa was a fallback in do_fetch Jun 13 12:48:01 maybe you can cook something up? Jun 13 12:50:59 pb__ : thanks it says * ERROR: Cannot satisfy the following dependencies for task-base: * update-modules * Jun 13 12:51:17 I've forced a build of update-modules and the image is now OK Jun 13 12:51:57 any idea why update-modules was not build as a dependency (Im using angstrom-2008.1 distro and synced to the git tree a few minutes ago) ? Jun 13 13:00:09 re Jun 13 13:00:16 pb_: If you have a minute, maybe you can help me understand why bitbake wants to build multiple instances of bluez-libs and virtual/server: http://paste.debian.net/38976/ The dot dependency files are at http://paste.debian.net/38975/ and http://paste.debian.net/38977/ Jun 13 13:00:27 so guys, how are kernel-module files generated ? I've been trying to track down a sample one to see, and for the hell of me, I can't find a trace. Say "kernel-module-usbserial" Jun 13 13:00:27 sorry, baby was screaming for a moment there Jun 13 13:00:34 Laibsch: yeah, good idea, I'll see what I can do Jun 13 13:00:56 in fact update-modules doesn't seems to be listed as a dependency in task-base nor in console-base-image Jun 13 13:01:00 eric___: hm, no, not offhand. you'd need to find out which package has it as a runtime dependency and then check whether it's also in DEPENDS there Jun 13 13:01:02 Laibsch: i added that, but still the same Jun 13 13:01:32 could someone please help a newbie here? Jun 13 13:04:12 very annoying kernel bug when your keyboard gets stuck and no matter what key you hit all you get is "11111111111111" Jun 13 13:04:19 Laibsch: heh, that must be annoying Jun 13 13:04:41 regarding your bluez thing, what DISTRO are you using? I know there is some special stuff in there for angstrom. Jun 13 13:04:46 pb__ : yes, that's stange because udpate-modules is not listed in any depends*dot Jun 13 13:04:47 minimal Jun 13 13:05:05 I've disabled the PROVIDES_append_angstrom Jun 13 13:05:11 s/_angstrom// Jun 13 13:05:14 okay Jun 13 13:05:17 and I intend to commit that later Jun 13 13:05:24 But, it's still the same Jun 13 13:05:26 yeah, I get a variation on that same problem with DISTRO=micro Jun 13 13:05:48 Yes, there are still a number of problems that Angstrom masks away Jun 13 13:05:53 what happens for me is that it builds bluez-libs 3.36, then bluez4 (which stomps on bluez-libs in staging) and then it tries to build bluez-utils 3.36, which fails because it won't build with the bluez4 headers. Jun 13 13:06:04 so it's good that people are starting to build other distros but angstrom Jun 13 13:06:15 I don't quite understand where the bluez4 thing is coming from but that seems like it must be a bug. Jun 13 13:06:34 how good are you with inspecting those dot files? Jun 13 13:06:51 Laibsch : grep update-modules *.dot Jun 13 13:07:03 eric___: different discussion here Jun 13 13:07:07 sorry Jun 13 13:07:11 Laibsch, only if they know why they changing distros Jun 13 13:07:14 And of course, I've grepped my heart out Jun 13 13:07:15 moderately, I guess. the dot files are fairly straightforward although I haven't looked at them in a while. Jun 13 13:07:31 far to many people start messing with the distro, when what they need to do is create a custom image Jun 13 13:07:33 Crofton: looks like we again have a different perspective on things Jun 13 13:07:36 which is much easier Jun 13 13:07:38 eric___: yah, that's to be expected. try inspecting tmp/deploy/ipk/*/Packages instead, look for 'Depends:.*update-modules' Jun 13 13:07:52 crofton: I am aware that I don't *need* a distro Jun 13 13:08:01 that is completely beside the point Jun 13 13:08:08 Crofton|work hi Jun 13 13:08:34 Crofton|work can you hep me out with bitbaking helloworld?ĦĦI am getting a strange problem Jun 13 13:08:39 And I know the difference between distro and image and that more often than notall you need is tweak an image. Jun 13 13:08:49 s/notall/not all/ Jun 13 13:08:57 Laibsch, I am not so mucjh woprried about you personally Jun 13 13:09:07 what is in helloworld? Jun 13 13:09:13 I'm not advising anybody to build anything but angstrom Jun 13 13:09:32 If you want success fast, go angstrom, I guess ;-) Jun 13 13:09:50 But that is about as much advertisement I am going to make for Angstrom ;-) Jun 13 13:10:18 pb_: Have you looked at my dot files? Jun 13 13:10:48 pb__ : thanks, I'm learning a lot ! in fact several kernel-module have update-modules in Depends: Jun 13 13:11:00 Laibsch: looking now Jun 13 13:11:17 Crofton|work just printf("hello") Jun 13 13:11:29 bitbake bluez-libs builds 3.36, I think Jun 13 13:11:29 yeah, checking on .dev Jun 13 13:11:55 Crofton|work and #include "studio.h" Jun 13 13:12:11 Laibsch: ah, here it is Jun 13 13:12:12 "openobex" -> "bluez4" Jun 13 13:12:36 Crofton|work it is just a simple program that I am trying to compile Jun 13 13:12:52 and, sure enough. bitbake -e -b openobex |grep ^DEPENDS yields DEPENDS="pkgconfig-native autoconf-native automake-native libtool-native libtool-cross gnu-config-native shasum-native coreutils-native virtual/i686-oe-linux-gcc virtual/libc libusb-compat bluez4" Jun 13 13:12:53 * Laibsch goes to look into openobex recipe Jun 13 13:12:53 yeah Jun 13 13:12:55 I read the file Jun 13 13:12:59 checking on stabkle now Jun 13 13:13:19 pb_: aha, thank you Jun 13 13:13:26 let's dig into this Jun 13 13:13:40 Laibsch: older openobex just has "bluez-libs" there, which seems more reasonable Jun 13 13:14:08 so, need to find out whether openobex 1.5 really does need bluez4, or whether that dependency is spurious. Jun 13 13:14:10 I wonder why in that case, 3.x is just not skipped completely Jun 13 13:14:49 did you say you have your bluez4 PROVIDEing bluez-libs at the moment? Jun 13 13:15:11 if so, might just need to make that the PREFERRED_PROVIDER to discourage bitbake from building 3.36 Jun 13 13:16:11 well, since it is newer it would normally be picked up automatically, no? Jun 13 13:16:38 no, I don't think bitbake thinks like that. Jun 13 13:16:40 openobex 1.5 was committed by koen, one change afterwards by mwester, but that is unrelated to this DEPENDS Jun 13 13:17:22 well, there has to be a reason, bitbake is not satisfied with bluez_4.x by itself Jun 13 13:17:53 I think it's just a deficiency in bitbake's dependency resolution logic. Jun 13 13:18:30 well, it does not build the old versions in other cases, so there must be something different here Jun 13 13:19:49 * Laibsch needs to go buy some food or be starving the rest of the day Jun 13 13:20:18 heh, right. do what you need to do. Jun 13 13:22:46 BTW, I'm not sure openobex is the only culprit Jun 13 13:23:00 bluez-cups-backend is only available in version 3.x Jun 13 13:23:16 Not sure if building that in 3.x will work with bluez4 Jun 13 13:23:31 dunno. bluez4 might include the cups bits for itself. Jun 13 13:24:10 bluez4 is compiled with cups disabled Jun 13 13:24:27 ah Jun 13 13:24:35 well, I guess someone who wants cups can debug that :-} Jun 13 13:27:23 openobex in debian does not depend on bluez4 Jun 13 13:27:38 I'm guessing it's in there for conveniency reasons Jun 13 13:28:49 sounds like it Jun 13 13:29:01 I'll see if it does actually build against 3.36, and adjust the dependency if yes Jun 13 13:31:44 hm, task 2874 of 3838. surely I must be close to having a buildable gpe-image now :-} Jun 13 13:31:44 03Phil Blundell  07org.openembedded.dev * r2e4f1691b3 10openembedded.git/recipes/tzdata/tzdata_2007k.bb: tzdata: fix installation with prefix='' Jun 13 13:37:07 sigh, now console-tools fails. what's that doing in gpe-image anyway? Jun 13 14:39:14 g'day kergoth Jun 13 14:45:55 03Phil Blundell  07org.openembedded.dev * r40d0c33b4a 10openembedded.git/recipes/console-tools/console-tools_0.3.2.bb: console-tools: inherit gettext Jun 13 14:46:03 03Phil Blundell  07org.openembedded.dev * ree690abcef 10openembedded.git/classes/update-rc.d.bbclass: update-rc.d: don't add runtime dependency if not doing online package management; only add it to ${PN} in any case. Jun 13 14:48:12 03Phil Blundell  07org.openembedded.dev * r12e94253b3 10openembedded.git/recipes/images/gpe-image.bb: gpe-image: figment is unfetchable, punt it from the image Jun 13 14:48:31 Laibsch: it seems that openobex does build okay against the older bluez Jun 13 14:48:51 nice Jun 13 14:50:14 03Phil Blundell  07org.openembedded.dev * rf1dc5457a0 10openembedded.git/recipes/openobex/openobex_1.5.bb: openobex: relax dependency from bluez4 to generic bluez-libs Jun 13 14:50:16 bitbake wants to build three different versions of bluez-libs for me by now :-/ Jun 13 14:50:37 doh, three? Jun 13 14:50:44 I thought there only were two available Jun 13 14:52:25 bluez4 (I relaxed that angstrom override) and 3.33 and 3.36 Jun 13 14:52:53 funky Jun 13 14:55:16 I'm ahead of .dev by 15 commits now ;-) Jun 13 14:55:44 heh Jun 13 14:55:59 by the way, I don't think we want that angstrom override for bluez4 PROVIDES to apply to all distros Jun 13 14:56:08 why not? Jun 13 14:56:34 the monolithic bluez4 is a bit of a PITA. it'd be better to fix up the split libs/utils and let non-angstrom use those. Jun 13 14:56:35 If anything, I'd rather make it default and let distros that don't want it undefine it Jun 13 14:57:14 can the split be done in bluez3 only? Jun 13 14:57:24 the package is certainly a monster Jun 13 14:57:33 I'm not sure I understand that question. Jun 13 14:57:38 bluez3 already has libs separate from utils Jun 13 14:58:09 yes Jun 13 14:58:27 so, you want to transfer that split to bluez4? Jun 13 14:58:33 the trouble with the all-in-one package is that it means you can't build bluez-libs (or anything that depends on bluez-libs) without building gstreamer and a pile of other stuff. Jun 13 14:58:45 or you want to generally use bluez3 where the split is already done? Jun 13 14:58:54 so, yeah, I want to create separate packages for libs and utils for bluez 4.x Jun 13 14:59:09 OK, I'm happy with that Jun 13 14:59:09 in fact, there's already a 4.x bluez-libs package, it just needs a bit of polishing to make it work properly. Jun 13 14:59:22 OK Jun 13 14:59:49 03Phil Blundell  07org.openembedded.dev * ra729981b8a 10openembedded.git/recipes/gpe-session-scripts/gpe-session-scripts_0.68.bb: gpe-session-scripts: respect ${bindir} Jun 13 15:00:07 do you know what is broken with it? Jun 13 15:00:19 not offhand, kergoth tried it the other day and had some issue Jun 13 15:00:22 it'd be in the ibot logs Jun 13 15:00:36 you mean tinderbox? Jun 13 15:00:39 or IRC logs Jun 13 15:00:42 irc logs Jun 13 15:01:05 I have to go out to the supermarket (again, doh) now, but I can look at it when I get back Jun 13 15:01:14 apparently I failed to buy everything that my wife asked for last time Jun 13 15:02:49 bluez-libs on tinderbox is all successes Jun 13 15:06:33 pb_: while you were at it you could have fixed gpe-session-scripts_0.67.bb as well ;-) Jun 13 15:12:50 ahead by 16 commits now ;-) Jun 13 15:12:54 but will push one soon Jun 13 15:14:09 03Rolf Leggewie  07org.openembedded.dev * r24e5233b89 10openembedded.git/conf/distro/include/preferred-xorg-versions-X11R7.4.inc: Jun 13 15:14:09 preferred-xorg-versions-X11R7.4.inc: pin xextproto-native and xtrans-native Jun 13 15:14:09 * xextproto-native: 7.0.3 Jun 13 15:14:09 * xtrans: 1.2.1 Jun 13 15:14:10 03Rolf Leggewie  07org.openembedded.dev * r328a71a287 10openembedded.git/conf/machine/include/zaurus-kernel.inc: Jun 13 15:14:14 zaurus-2.6.inc: drop machine specific MACHINE_KERNEL_PR. Jun 13 15:14:16 Will result in kernel rebuild for tosa and collie. Jun 13 15:14:18 03Rolf Leggewie  07org.openembedded.dev * r603c6fc530 10openembedded.git/conf/machine/include/zaurus-2.6.inc: WIP zaurus-2.6.inc: some changes to machine config Jun 13 15:14:21 03Rolf Leggewie  07org.openembedded.dev * r7bed7f4722 10openembedded.git/classes/siteinfo.bbclass: siteinfo.bbclass: add support for via C3 host Jun 13 15:14:24 03Rolf Leggewie  07org.openembedded.dev * r097a6b1089 10openembedded.git/classes/rootfs_ipk.bbclass: WIP rootfs_ipk.bbclass: use force-defaults Jun 13 15:14:27 03Rolf Leggewie  07org.openembedded.dev * rf2f0eaee13 10openembedded.git/recipes/cairo/cairo-1.3.4/configure.in_requires_pkg-config-0.15.diff: WIP cairo: mark patch as unapplicable upstream Jun 13 15:14:30 03Marcin Juszkiewicz  07org.openembedded.dev * r03ac5be531 10openembedded.git/ (classes/kernel.bbclass conf/bitbake.conf): (log message trimmed) Jun 13 15:14:37 fix for MACHINE_KERNEL_PR stuff Jun 13 15:14:38 oh no, all those weren't supposed to be pushed Jun 13 15:14:39 This patch unbreaks current behaviour which was introduced by Jun 13 15:14:41 MACHINE_KERNEL_PR variable. Jun 13 15:14:43 As most of target machines do not use it they have PR with broken value Jun 13 15:14:45 (set to "r0" instead of value in recipe). I took other way which makes Jun 13 15:14:47 both types of users happy -- those with MACHINE_KERNEL_PR in use and Jun 13 15:14:49 03Rolf Leggewie  07org.openembedded.dev * r72b363cc94 10openembedded.git/conf/distro/include/sane-srcrevs.inc: WIP tslib: update SRCREV to 76 (currently fails to build). (Closes: #582) Jun 13 15:14:54 03Rolf Leggewie  07org.openembedded.dev * rd9e49307ad 10openembedded.git/recipes/images/opie-image.bb: WIP opie-image: add tslib (this change should be dumped) Jun 13 15:14:57 03Rolf Leggewie  07org.openembedded.dev * r28c50febb8 10openembedded.git/classes/rl_rm_work.bbclass: Jun 13 15:15:02 WIP: make my own rm_work class that does not depend on do_build and keeps stamps intact Jun 13 15:15:06 maybe I should extend the existing class with a variable Jun 13 15:15:08 03Rolf Leggewie  07org.openembedded.dev * rccda7cbeba 10openembedded.git/recipes/ (3 files in 3 dirs): Jun 13 15:15:11 WIP recipes set to DEFAULT_PREF -1 Jun 13 15:15:13 * guile 1.8.6 (hangs in do_compile) Jun 13 15:15:15 * dri2proto 2.0 Jun 13 15:15:17 * bluez-cups-backend 3.36 (fails do_compile, 'HCI_SECMGR' undeclared) Jun 13 15:15:19 03Rolf Leggewie  07org.openembedded.dev * r4318e5b038 10openembedded.git/recipes/images/gpe-image.bb: Jun 13 15:15:26 WIP: images: "psplash-angstrom" with "virtual/psplash" Jun 13 15:15:28 * done for gpe-image, needs to be done for others Jun 13 15:15:30 03Rolf Leggewie  07org.openembedded.dev * r053d5a46bc 10openembedded.git/recipes/images/gpe-image.bb: WIP images: virtual/psplash does not build, inquire some more Jun 13 15:15:35 03Rolf Leggewie  07org.openembedded.dev * r854002a9d1 10openembedded.git/recipes/images/sdr-image.bb: Jun 13 15:15:40 (7 lines omitted) Jun 13 15:20:19 03Rolf Leggewie  07org.openembedded.dev * ra5a20cbef6 10openembedded.git/conf/distro/include/preferred-xorg-versions-X11R7.4.inc: Jun 13 15:20:19 Revert "preferred-xorg-versions-X11R7.4.inc: pin xextproto-native and xtrans-native" Jun 13 15:20:19 This reverts commit 24e5233b89825da9ba62050f1ea4fba607cae5c7. Jun 13 15:20:19 reverting a bunch of WIP commits that were not yet meant to be pushed upstream. Jun 13 15:20:21 Sorry for the noise. Jun 13 15:20:23 Signed-off-by: Rolf Leggewie Jun 13 15:24:50 hi how do I call autoconf? it seems that it's not called automatically on inherit autotools... Jun 13 15:46:03 Gnutoo: it should be automatic Jun 13 15:46:18 pb__, ok thanks...strange Jun 13 15:50:56 03Rolf Leggewie  07org.openembedded.dev * rae823d80c9 10openembedded.git/recipes/gpe-session-scripts/gpe-session-scripts_0.67.bb: Jun 13 15:50:56 gpe-session-scripts: replace /usr/bin with ${bindir} in 0.67 Jun 13 15:50:56 * same as commit a729981b8a3e27e800078ccfd05249eef42b7ae4 Jun 13 15:51:07 wow, git is SO awesome Jun 13 15:51:21 it was a breeze getting all those commits off the public repo again Jun 13 15:51:30 * Laibsch was sweating a bit Jun 13 15:51:32 :-D Jun 13 15:52:12 I also just discoverd that qgit replaces a commit ID with a link to that commit and the glyphs with the commit message Jun 13 15:52:18 very neat Jun 13 15:56:11 Laibsch: heh Jun 13 16:01:21 phew. gpe-image build completed! Jun 13 16:01:33 that only took, what, three days? Jun 13 16:01:34 really? Jun 13 16:01:38 Amazing! Jun 13 16:01:45 I'm still at it Jun 13 16:01:58 But I have for whatever reason about twice the tasks than you Jun 13 16:01:59 :-/ Jun 13 16:02:04 over 7.000 Jun 13 16:02:19 yeah, that is a bit odd Jun 13 16:02:37 it's tempting to put it down to distro bloat but it does seem like a startlingly large number Jun 13 16:04:57 also, for laughs, I started a build of micro-image for angstrom in another window and I don't think I'm getting significantly more tasks for that one than I do for micro. Jun 13 16:06:13 http://ftp.sh.cvut.cz/MIRRORS/rock/ROCK-2.0/gnome2/figment/ Jun 13 16:06:21 maybe you want to add figlet back in Jun 13 16:06:26 It sounds like a nice tool Jun 13 16:06:30 aha Jun 13 16:06:33 yeah, will do Jun 13 16:09:11 actually, I do get 50% more tasks with angstrom than micro for the same image, 631 vs 444 Jun 13 16:09:31 but your discrepancy is even greater, so that's still a bit weird Jun 13 16:14:03 03Graeme Gregory  07shr/import * rcf5e3ffe9b 10openembedded.git/recipes/unifdef/files/unifdef.c: Jun 13 16:14:03 recipes/unifdef/files/unifdef.c : more modern glibc expose getline from Jun 13 16:14:03 Signed-off-by: Koen Kooi Jun 13 16:26:38 I wonder if there is an easy way to collect number of tasks for various machine/distro/image combinations? Jun 13 16:27:20 whitefox, it should work Jun 13 16:32:29 can I take a look at your recipe? Jun 13 16:32:51 the recipe for gnuradio? I want to try it on beagle cuz I read about that being done. Jun 13 16:40:34 pb_: SUCCESS! Jun 13 16:40:53 Laibsch: hurrah! Jun 13 16:41:00 After several days, my little C3 machine has indeed successfully compiled gcc-cross now Jun 13 16:41:11 viac3-linux did the trick Jun 13 16:41:22 thanks to your support Jun 13 16:41:32 jolly good Jun 13 16:41:45 I can only be another half day before do_install and the rest are also done ;-) Jun 13 16:42:39 heh Jun 13 16:43:07 I'm just going to build a kernel, then I can try booting this image in qemu. Jun 13 16:43:15 I wonder what the chances are that it will actually work. Jun 13 16:45:20 slim Jun 13 16:45:24 at least not OOTB Jun 13 16:45:27 yeah, that's what I figured Jun 13 16:45:47 I'm spending most of my time just catching up with build-time issues Jun 13 16:45:51 I've never even tried gpe under qemu before, and I suspect this micro build probably has at least some files missing. Jun 13 16:46:00 I hardly ever get to fix on-device bugs Jun 13 16:46:08 still, only one way to find out... Jun 13 16:46:24 try poky-qemu for booting that image Jun 13 16:46:37 or if not that, then the qemu from OE itself Jun 13 16:47:01 qemu itself is a rather tricky beast Jun 13 16:47:30 regular qemu seemed to work ok for x86 last time I tried, without needing any special magic. Jun 13 16:47:38 * Laibsch hugs packaged staging Jun 13 16:47:40 if I run into trouble then I'll try poky-qemu though Jun 13 16:48:04 that makes nuking TMPDIR much less frustrating Jun 13 16:48:28 yeah, I should totally turn packaged staging on. I keep forgetting to do that. Jun 13 16:48:29 please look on the device if tslib is installed Jun 13 16:48:35 It wasn't on my opie-image Jun 13 16:48:39 Just tslib-conf Jun 13 16:48:51 which may be all that is needed, but I'm not sure Jun 13 16:49:29 don't worry, I've heard rumours, it will be default you can't turn off sooner than later Jun 13 16:49:44 it=packaged staging Jun 13 16:57:24 so, for interest's sake, micro-image built with angstrom currently contains 498 files and directories, whereas the same target built with micro-uclibc has 380 objects. Jun 13 16:57:48 the angstrom one is also rather bigger, at 4.8MB compared to 1.6MB, but I think this is mostly because angstrom uses glibc out of the box. Jun 13 16:59:46 Laibsch: I do have libts-1.0.so.0 in my gpe-image Jun 13 17:00:09 what does "opkg list_installed | tslib" say? Jun 13 17:00:17 dunno, I don't have opkg Jun 13 17:00:33 I can check the logs though, just a minute Jun 13 17:01:26 hm, seems to lose if the rootfs is mounted read-only. S37populate_volatile (ironically) is trying to write to / Jun 13 17:02:19 Laibsch: ah, I don't have the tslib package installed. just tslib-conf and libts-1.0-0 Jun 13 17:02:24 is there meant to be tslib itself as well? Jun 13 17:02:36 don't know Jun 13 17:02:49 doh, S99gpe-dm is pointing to the wrong place Jun 13 17:02:51 It was just something I suspected may be causing my calibration problems Jun 13 17:03:01 "/usr/bin/gpe-dm: not found" Jun 13 17:03:14 There is also another package called IIRC calibriation-data Jun 13 17:03:22 and that was not pulled in Jun 13 17:03:46 and that helped me overcome calibration issues in the past Jun 13 17:03:51 the bug tracker knows more about it Jun 13 17:04:20 it seemed natural to me that a missing tslib package may be the root cause Jun 13 17:04:29 And I still think that may be the case Jun 13 17:04:51 heh, that was dumb Jun 13 17:04:59 memo to self, don't try to press ctrl-alt-backspace in qemu Jun 13 17:05:35 so, issues so far: (1) the gpe-dm wrong path thing; (2) /etc/X11/Xserver loses if you don't have an /etc/profile; (3) it's trying to run Xfbdev and my image has Xvesafb. Jun 13 17:06:06 if you can't fix them right away, please report them in the tracker and make them block the qemu meta-bug Jun 13 17:06:28 the first two are specific to micro, I don't expect you'll get them with angstrom or minimal Jun 13 17:06:34 I believe mail from the bts is still not working. cbrake_away, mickey|DepecheMo? Jun 13 17:07:02 that does not matter, the bts is open for any OE-related issue Jun 13 17:07:31 and it's also a source to consult for how to fix problems Jun 13 17:09:55 pb_: is there any chance the kernel could be made to identify the CPU correctly as viac3 instead of i686? Jun 13 17:30:57 Laibsch: unlikely. you could fix it in oe, though. Jun 13 17:31:56 I'm just reading up on it Jun 13 17:32:22 It seems the reason that gcc behaves the way it does is that turning off cmov would hurt everybody Jun 13 17:32:33 how much of that would be a real issue for OE? Jun 13 17:33:28 I don't think you should turn it off globally, that would indeed be a nuisance. But you can patch base.bbclass to detect your situation and use "viac3" not "i686". Jun 13 17:35:44 how can I detect that? Jun 13 17:36:01 And if I can, why is it not done globally, ie outside OE? Jun 13 17:37:21 mostly because there's no obvious place where it should be done. Part of the problem is that OE is taking the string that it gets from "uname" and just using that as the cpu part of the configure triplet; I don't think there's any guarantee that this will always be a sensible thing to do. Jun 13 17:37:47 one thing you could try would be to have OE invoke "config.guess" to determine the GNU-ish system name for the build host. Jun 13 17:38:00 that's what GCC would do if it was left to its own devices, i.e. being built outside of OE. Jun 13 17:38:30 I think there's at least some chance that this will give you back the "viac3-linux" string that you want, and if it does then that's an easy win. Jun 13 17:38:49 if not, you can either fix config.guess, or add extra logic to OE to figure it out directly. Jun 13 17:39:10 in the latter case, you'd need to write a bit of python to inspect /proc/cpuinfo and establish whether the host is a via c3 or not. Jun 13 17:39:36 (obviously keyed on having detected i686-linux to start with; you wouldn't want to start trying to poke around in /proc/cpuinfo on OS X or the like.) Jun 13 17:42:20 invoking config.guess is a good idea, indeed. the gnu-config recipe gets built quite early on in the process, so it could use it safely after that Jun 13 17:43:01 course, it wouldnt be available until the full parse at task execution time, but i dont think the first parse would need it Jun 13 17:43:19 right Jun 13 17:43:43 alright, I'll think about this a bit tomorrow Jun 13 17:49:29 unfortunately, config.guess always returns i686-pc-linux-gnu Jun 13 17:52:03 it's still probably a good idea, just you'll have to do more than that, i guess :) Jun 13 17:52:25 I looked into config.guess briefly Jun 13 17:52:34 There is nothing about cmov in there Jun 13 17:52:55 and apparently in gcc terms i586 + cmov = i686 Jun 13 17:53:18 so, it's just not tested Jun 13 18:12:39 does openembedded/angstrom have a package manager for an installed system, or just for managing images and distributions? Jun 13 18:12:53 hrw: when are we going to see your MACHINE_KERNEL_PR patch? Jun 13 18:12:55 works fine here Jun 13 18:15:08 opkg? Is that it? Jun 13 18:15:36 mdavey2: yes Jun 13 18:16:03 awesome! Little by little I am learning :D Jun 13 18:16:03 mdavey2: angstrom even has http feeds like remote repositories in your lovely desktop distro Jun 13 18:17:10 mdavey2: but you can just push ipk file to opkg and pray for resolved dependencies Jun 13 18:26:44 Laibsch: I think that I will just push it Jun 13 18:26:50 Laibsch: during week Jun 13 18:33:25 03Tom Rini  07org.openembedded.dev * r42b8607cc4 10openembedded.git/classes/autotools.bbclass: autotools.bbclass: sdk packages should not depend on libtool-cross Jun 13 18:43:54 * Laibsch stopped this bluez madness by dropping bluez from gpe-task-base :-D Jun 13 18:44:07 heh Jun 13 18:44:28 don't all the PIM bits depend on it? Jun 13 18:45:17 we'll see Jun 13 18:45:19 03Phil Blundell  07org.openembedded.dev * ra1e13e08fc 10openembedded.git/recipes/ (figment/figment_0.3.5.bb images/gpe-image.bb): figment: update SRC_URI and reinstate in gpe-image Jun 13 18:45:20 03Phil Blundell  07org.openembedded.dev * r430f94204e 10openembedded.git/ (3 files in 3 dirs): Jun 13 18:45:20 gpe-dm: update to 0.52 Jun 13 18:45:20 xserver-common: update to 1.25 Jun 13 18:45:22 03Phil Blundell  07org.openembedded.dev * r6ea4aca3fb 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@new.openembedded.org:openembedded into org.openembedded.dev Jun 13 18:45:37 so, with those changes my qemu image gets as far as starting X Jun 13 18:45:50 sadly, I still don't get calibration or the login screen, but that's progress of a sort Jun 13 18:47:50 hm. well, xcalibrate not installed, no wonder I don't get calibration Jun 13 18:48:33 but x-window-manager looks ok Jun 13 18:50:11 oh, I take that back, it's got xtscal. just looking for the wrong thing. Jun 13 18:50:31 but these scripts are all infested with wrong paths, that's a bit sad. Jun 13 19:27:39 03Klaus Kurzmann  07shr/import * r16e0cbb5e7 10openembedded.git/conf/distro/include/shr-om-gta01.conf: Jun 13 19:27:39 shr-om-gta01.conf: keep OABI Jun 13 19:27:39 For gta01 we need to keep OABI in the kernel. Otherwise the Jun 13 19:27:39 proprietary GPS stuff won't work. Jun 13 19:27:39 Signed-off-by: Klaus Kurzmann Jun 13 19:39:32 03Phil Blundell  07org.openembedded.dev * r30a18b60ef 10openembedded.git/classes/rootfs_ipk.bbclass: rootfs_ipk: opkg (at least some versions) doesn't respect ${libdir}, so look in /usr/lib specifically for its bits Jun 13 19:41:18 03Phil Blundell  07org.openembedded.dev * re0f0ba11e6 10openembedded.git/recipes/tasks/task-gpe-settings.bb: task-gpe-settings: remove xmonobut, unbuildable and nothing to do with settings anyway Jun 13 19:42:44 03Phil Blundell  07org.openembedded.dev * rfb8cbd53a9 10openembedded.git/conf/distro/micro.conf: micro: disable ONLINE_PACKAGE_MANAGEMENT globally Jun 13 19:46:11 03Phil Blundell  07org.openembedded.dev * reb783af8f0 10openembedded.git/recipes/gcc/ (gcc-4.4.0.inc gcc-4.4.0/200-uclibc-locale.patch): gcc-4.4.0: add some uclibc patches Jun 13 20:27:53 so I'm trying to modify how uboot sets the pin mux on my device but I cant find the configuration file to do this. where is this file located? Jun 13 20:30:29 taelot: which device? Jun 13 20:30:45 taelot: typically, it's a C source file that does that. Jun 13 20:31:35 likewise: beagleboard. so its not configured in a recipe? i guess i should start looking in the work directory for source Jun 13 20:47:47 03Graeme Gregory  07shr/import * r4ef6774fbc 10openembedded.git/recipes/linux-libc-headers/ (files/unifdef.patch linux-libc-headers_2.6.23.bb): Jun 13 20:47:47 linux-libc-headers_2.6.23.bb : more modern glibc expose getline from Jun 13 20:47:47 Signed-off-by: Koen Kooi Jun 13 20:47:57 03Klaus Kurzmann  07shr/import * rfab6727c86 10openembedded.git/conf/distro/include/shr-autorev.inc: Jun 13 20:47:57 shr-autorev.inc: set preferred version for linux-libc-headers to 2.6.23 Jun 13 20:47:57 Signed-off-by: Klaus Kurzmann Jun 13 21:05:57 taelot: sorry I don't have a beagleboard build here, so I cannot help you quickly. Jun 13 21:06:01 hi all Jun 13 21:06:54 how can i remove a once compiled and installed recipe? is there any do_remove operation or something? Jun 13 21:06:58 likewise: no worries! i found the files i needed. now i just have to devise a way to make my changes persist across new versions from git Jun 13 21:15:42 Goddchen: bitbake -c clean Jun 13 21:18:36 ah, just found it in the documentation Jun 13 21:18:40 but thx anyway :) Jun 13 21:19:10 btw, if you're insterested, we decided to switch from our avr32 cpu to an ARM one... Jun 13 21:26:50 florian: good evening Jun 13 21:28:05 hopefully we get java running on that arm platform Jun 13 21:29:12 yeah, arm is rather better supported than avr32 Jun 13 21:29:43 the only bad thing is that I don't think there is any open source support for the arm java acceleration bits. Jun 13 21:30:02 our professor insisted on the avr32... we wanted an arm from the first second... Jun 13 21:30:24 so, if your cpu has those and you want to use them, you probably have to use a closed jvm. Jun 13 21:30:53 the cpu has it, but will first have a look on the performance without the hw acc Jun 13 21:31:17 if it's acceptable for the first prototype, it's ok... Jun 13 21:32:02 wow, the jamvm-native doesn't seem to be able to build classpath (do_configure fails because jamvm-native can't execute the Test-class) Jun 13 21:32:08 with cacao-native everything is fine... Jun 13 21:49:00 i'm off for today, gn8 @ll Jun 14 01:05:03 Anybody alive tonight? Jun 14 01:44:59 SDuensin: Ever heard about "don't ask to ask?" Jun 14 01:45:29 kergoth: ping Jun 14 01:45:37 rwhitby: ping Jun 14 01:45:47 Yep. Didn't have a real specific question. Just hangin' out. :-) Jun 14 01:45:53 Been bitbaking Angstrom all day. Jun 14 01:46:12 Funny, so indeed your question was only if anybody was awake ;-) Jun 14 01:46:30 Yep! :-P Jun 14 01:47:07 Unless you happen to know how to get Java going under Angstrom with OE. Jun 14 01:47:21 Laibsch: syn|ack Jun 14 01:47:38 SDuensin: sorry, no idea Jun 14 01:48:06 Laibsch - Me either. Was planning on worrying about it for real once the first bake is done. Jun 14 01:49:11 I use packaged staging. Yesterday I cleared TMPDIR with the exception of DEPLOYDIR and restarted building stuff. It went well at first, but there were always errors about "ERROR: no cross compiler built yet?" Was I doing something wrong? Jun 14 01:49:46 rwhitby: maybe you have an answer to above question? Second, I was wondering about mysql on melo. Jun 14 01:49:59 we recently moved bugzilla and that went really well Jun 14 01:50:47 I'm wondering if the users that have been set up are restricted in any way or if any user can currently basically access any database Jun 14 01:51:02 I was wondering who to ask for additional database users Jun 14 01:51:19 We don't have a dedicated one for tinderbox/oestats for example Jun 14 01:51:35 Laibsch: let's talk about this in another channel Jun 14 01:51:43 where? Jun 14 01:51:56 check your invites **** ENDING LOGGING AT Sun Jun 14 02:59:57 2009