**** BEGIN LOGGING AT Wed Aug 22 02:59:56 2007 Aug 22 03:28:33 how to make a fetch a certain rev from svn? Aug 22 03:28:49 my reicpe fetches current from svn now, but I need a certain rev (one that works) Aug 22 03:34:55 search for tag= in current .bb's and you should find an example Aug 22 03:44:33 thanks, it was rev= Aug 22 03:44:51 I kinda figured but... well we are all good now Aug 22 03:57:04 has anybody seen anything like this? makedevs: /home/henry/oe/dev/tmp/rootfs/bin/tinylogin: file can not be created! Aug 22 03:58:42 hvontres|home: lots of openmokians are reporting it with oe.dev head Aug 22 03:59:00 seems that makedevs has stopped doing mkdir -p like things. Aug 22 03:59:02 rwhitby: any ideas? Aug 22 03:59:27 or the device table has suddenly got subdirs where it didn't before. Aug 22 03:59:33 I haven't looked into the real cause yet. Aug 22 04:00:46 rwhitby: do you know if this could be related to the host os? I just switched from fedora to kubunu Aug 22 04:02:00 it works for me in Debian Etch Aug 22 04:02:36 and device_table-minimal.txt hasn't changed recently as far as I know Aug 22 04:06:36 Sounds like it might be fakeroot acting up again? Aug 22 04:09:26 does anyone here know the bb name for the kernel image for spitz? Aug 22 04:09:44 mwester-road: hmmmm, did we ever figure out what was causing fakeroot to act up in the first place? Aug 22 04:10:14 mwester-road: btw, are you going to send your kexec patches to the mailing list soon? Aug 22 04:10:48 Not really. HopsNBarley determined that it was something to do with the inode caching in fakeroot, but never sorted it out. It had been observed elsewhere at the time. Aug 22 04:10:49 wirelessdreamer: spitz uses linux-rp_2.6.21 Aug 22 04:11:02 hvontres|home: thanks Aug 22 04:11:07 I ran into it when I upgraded to Fedora 7, and solved it by upgrading fakeroot to 1.7 Aug 22 04:12:03 Re: kexec changes -- I can send those out. Probably when I return later this week... Aug 22 04:12:30 mwester-road: cool. I told RP about them and he sounded really interested Aug 22 04:12:58 Well, the sysfs changes are really bad, but the rest is probably ok to use. Aug 22 04:14:10 hvontres|home: can you pastebin the log file from the rootfs creation problem? Aug 22 04:14:27 mwester-road: well, at least we can get the other guys thnking about how to solve this better Aug 22 04:14:35 ~pastebin Aug 22 04:14:36 hmm... pastebin is a place to paste your stuff without flooding the channel - try http://pastebin.ca, or http://channels.debian.net/paste, or http://rafb.net/paste/, or http://pastebin.com is usually painfully too slow and unresponsive to use, use one of the other pastebin sites, or dpaste.com is a very nice pastebin as well Aug 22 04:18:07 mwester-road: http://pastebin.ca/666457 Aug 22 04:20:22 Well, not conclusive, but that's similar to the sort of thing one would see with the fakeroot problem. What version is fakeroot-native in your tmp tree? Aug 22 04:21:57 mwester-road: 1.7.1... but I just tried to rebuild it and this is what I got: http://pastebin.ca/666460 Aug 22 04:22:13 mwester-road: it looks like fakeroot is failing on configure.... Aug 22 04:23:09 There's a patch for that... Aug 22 04:24:08 mwester-road: ooops.. tried to rebuild fakeroot instead of fakeroot-native... one moment.... Aug 22 04:25:55 mwester-road: hmmm....fakeroot-native builds ok.... Aug 22 04:26:58 * * OE Bug 2304 has been RESOLVED (FIXED) by xjqian(AT)gmail.com Aug 22 04:27:00 * * texinfo fails to build Aug 22 04:27:02 * * http://bugs.openembedded.org/show_bug.cgi?id=2304 Aug 22 04:27:42 1.7.1 has many fixes, especially for 64 bit filesystem calls... perhaps something isn't quite sorted out for 64-bit systems? (I only run 32-bit at the moment). Aug 22 04:28:43 mwester-road: hmmmm...that could be it... I just got a nice new dual-core turion laptop.... and this is my first attempt at building on it :) Aug 22 04:31:01 I also run only 32 bit and everything works. Aug 22 04:35:43 I wonder if this bug might be relevant? http://bugs.uclibc.org/view.php?id=371 Aug 22 04:41:24 mwester-road: hmmm,could be interesting Aug 22 05:10:58 * * OE Bug 2820 has been created by hvontres(AT)gmail.com Aug 22 05:11:00 * * do_make_rootfs fails on makedevs Aug 22 05:11:02 * * http://bugs.openembedded.org/show_bug.cgi?id=2820 Aug 22 05:54:23 koen: ping Aug 22 07:22:02 was there a fix for the tinylogin do_rootfs issue Crofton ? Aug 22 07:34:30 micropal : I stumbled into that myself Aug 22 07:35:10 not sure whats causing it though Aug 22 07:38:34 only the dev directory is created. I think it tries to copy tinylogin into bin which doesn't exists. But I'm still a newbie in OE, so I might be wrong Aug 22 07:39:31 micropal : yep, thats the problem Aug 22 07:39:50 micropal : what i can't figure out yet is why bin is not created Aug 22 07:48:48 It's makedevs that tries to create /bin/tinylogin. It's not that it tries to copy tinylogin into /bin Aug 22 07:49:43 why is /bin/tinylogin, /bin/mount and /bin/umount in files/device_table-minimal.txt? Aug 22 07:50:45 Is it just to get the permissions right or something? Aug 22 08:04:44 morning Aug 22 08:08:37 good morning all Aug 22 08:08:56 I thought you all were americans ;) good morning Aug 22 08:10:59 ~ugt Aug 22 08:11:00 from memory, ugt is Universal Greeting Time. Created in #mipslinux, it is a rule that states that whenever somebody enters an IRC channel it is always morning, and it is always late when the person leaves. The local time of any other people in the channel, including the greeter, is irrelevant. http://www.total-knowledge.com/~ilya/mips/ugt.html Aug 22 08:11:15 heh ;) Aug 22 08:12:14 anyways, steliosk if you comment out that /bin/tinylogin infiles/device_table-minimal.txt you can build your image. Aug 22 08:12:34 I'm going to try the efika image on the freescale m-itx board later this week. Aug 22 08:24:14 hi Aug 22 08:24:35 isn't there a package for bitbake for ubuntu feisty? Aug 22 08:31:59 Crofton: http://www.angstrom-distribution.org/unstable/images/gumstix/20070821/ Aug 22 08:37:04 koen : hi ! Package will leave today Aug 22 08:37:07 micropal : well, it will work, but that's not the general idea when trying to find bugs Aug 22 08:37:23 steliosk: thanks! Aug 22 08:42:53 * XorA should update mtn more often Aug 22 08:43:59 XorA: I'm doing a final test on abiword*2.5.2 now Aug 22 08:44:49 koen: cool Aug 22 08:45:16 and the ezx dudes now have a working asoc driver and can make calls with it Aug 22 08:45:18 will a PREFFERRED_VERSION_glibc="2.3.6" take affect in my local.conf or will i need to place it somewhere else ? i just notice that it still builds glibc-intermediate-2.5 Aug 22 08:46:58 * * OE Bug 2821 has been created by autobuild(AT)openembedded.org Aug 22 08:47:00 * * wv-1.2.0-r0-do_configure Aug 22 08:47:02 * * http://bugs.openembedded.org/show_bug.cgi?id=2821 Aug 22 08:49:57 * * OE Bug 2821 has been RESOLVED (FIXED) by Aug 22 08:50:00 * * wv-1.2.0-r0-do_configure Aug 22 08:50:02 * * http://bugs.openembedded.org/show_bug.cgi?id=2821 Aug 22 08:50:22 seppuku.bbclass rocks Aug 22 08:52:58 rob_w|laptop : You have to override that also then, beacuse the distro you are using defines it Aug 22 08:54:13 steliosk, how do i ? Aug 22 08:57:14 DISTRO=angstrom-2008.1 :) Aug 22 08:58:05 koen: so those moto phones are finally useful? Aug 22 08:58:36 XorA: looks like it Aug 22 09:00:25 good morning Aug 22 09:04:07 steliosk: yeah, I know. It's a quick hack. Aug 22 09:04:32 03koen 07org.oe.dev * r18a264d6... 10/ (3 files in 3 dirs): linux 2.6.21: remove localversion for gumstix to get module deps working properly Aug 22 09:04:40 03koen 07org.oe.dev * rad632a7e... 10/ (1 packages/wv/wv_1.2.0.bb): wv: add libgsf to depends Aug 22 09:04:46 03woglinde2 07org.oe.dev * r366af8b1... 10/ (1 packages/kxml2/kxml2_2.3.0.bb): Aug 22 09:04:46 kxml2: more unification of recipe Aug 22 09:04:46 * change hardcoded download file to more general Aug 22 09:04:50 03ifaistos 07org.oe.dev * rf145dcb7... 10/ (1 packages/linux/linux-ml403-mvista-2.6.x_git.bb): Aug 22 09:04:50 packages/linux/linux-ml403-mvista-2.6.x_git.bb : Add DEPENDS="git-native" to use native git Aug 22 09:04:50 when doing the pulls Aug 22 09:04:54 steliosk, you mean to overrwrite glibc-intermediate="2.3.6" Aug 22 09:04:55 03ifaistos 07org.oe.dev * rba5f0bfd... 10/ (1 packages/linux/linux-ml403-slab-2.6.x_git.bb): Aug 22 09:04:55 packages/linux/linux-ml403-slab-2.6.x_git.bb :Add DEPENDS="git-native" to use native git Aug 22 09:04:56 when doing the pulls Aug 22 09:11:16 rob_w|laptop : yes Aug 22 09:13:03 ~fish lrg Aug 22 09:13:04 * ibot slaps lrg around with a large trout Aug 22 09:13:36 ~plague lrg Aug 22 09:14:09 * lrg can relax now as there is no ~plague Aug 22 09:14:31 hey lrg Aug 22 09:14:42 hey koen Aug 22 09:21:38 morning all Aug 22 09:23:47 hey RP Aug 22 09:24:01 hey RP Aug 22 09:24:12 good morning Aug 22 09:24:30 good morning, rp and florian Aug 22 09:31:47 morning Aug 22 09:32:00 morning mickey|taxes Aug 22 09:33:47 hey mickey|taxes & florian Aug 22 09:35:10 hi mickey|taxes Aug 22 09:35:27 mickey|taxes: uh, you spent the night working on taxes?! Aug 22 09:35:59 florian: it's much worse... i fell asleep over searching the necessary papers :/ Aug 22 09:36:23 the apartment move and the water incident didn't help my organization Aug 22 09:36:33 usually i have all papers in place Aug 22 09:36:36 not this time :/ Aug 22 09:37:14 mickey|taxes: What a well known situation... I have to search for papers all the time. But don't worry - taxes is harmless unless you have employees. Aug 22 09:37:35 mickey|taxes: Try to get a huge amount of money from the state - that's fun ;) Aug 22 09:38:31 RP: Hi ! I updated to bitbake 1.8.9 but if the build fails i still get *.lock files all over the place Aug 22 09:39:57 steliosk: They should only appear in DL_DIR Aug 22 09:40:34 steliosk: and they get left behind even when it succeeds intentionally to allow multiple bitbakes to use the same DL_DIR Aug 22 09:40:51 If it removed them, it has the potential for races :/ Aug 22 09:40:55 RP: its in the package dirs Aug 22 09:41:19 RP: but i'll try a clean build as these might be leftovers from earlier bitbake builds Aug 22 09:41:24 steliosk: Can you remove them and confirm that, it should have been fixed Aug 22 09:47:58 * * OE Bug 2822 has been created by autobuild(AT)openembedded.org Aug 22 09:48:00 * * abiword-embedded-2.5.2-r0-do_compile Aug 22 09:48:01 * * http://bugs.openembedded.org/show_bug.cgi?id=2822 Aug 22 09:51:41 Can somebody remind me how a 64bit machine can be forced to do a 32bit build Aug 22 09:51:44 florian: hehe. well, EÜR and Vorsteuer are keeping me busy enough... i can imagine employees are a nightmare (taxes-wise) Aug 22 09:52:09 then again, I have the motivation to go though this at least once without a tax consultant Aug 22 09:52:17 because i actually want to understand what's going on Aug 22 09:54:07 steliosk: 'linux32 ' Aug 22 09:54:34 mickey|taxes: This works pretty good without, but I'm quite sure that you won't stay with EÜR for a long time... Aug 22 09:54:39 morning Aug 22 09:54:53 | checking for headers required to compile python extensions... not found Aug 22 09:54:53 | configure: error: could not find Python headers Aug 22 09:56:15 mickey|taxes: python on x86-64 Aug 22 09:56:27 mickey|taxes: python-dbus to be exact but this can happen with any Aug 22 09:57:02 hmm Aug 22 09:57:18 task-python-everything compiles on my x86-64 buildmachine Aug 22 09:57:23 although i'm cheating a bit Aug 22 09:57:24 mickey|taxes: look at OE archives - I wrote why this happen Aug 22 09:57:27 CHEATER! Aug 22 09:57:41 mickey|taxes: you built in 32bit chroot? Aug 22 09:57:43 hacking uname and overriding BUILD_SYS is a good thing :D Aug 22 09:57:57 no, it's 64bit userland, i just prevent it to be detected Aug 22 09:58:09 otherwise, a lot of packages don't build here Aug 22 09:58:16 mickey|taxes: so do not tell that it compiles Aug 22 09:58:18 and i never had the time to actually look into that Aug 22 09:58:32 mickey|taxes: I wrote everything needed few months ago Aug 22 09:58:39 and even patched few recipes Aug 22 09:58:45 hrw: ok, as you might know i was away then Aug 22 09:58:52 give me some pointers again Aug 22 09:58:56 mickey|taxes: grep -- --with-python-includes packages/python/* Aug 22 09:58:57 and i'll look at that Aug 22 09:59:12 Crofton: You were also running into the makedev and tinylogin problem during building a rootfs, right? Aug 22 09:59:31 Seems just the /bin is missing Aug 22 09:59:32 mickey|taxes: the problem is that python-* do_configure look into staging/HOST/include Aug 22 09:59:48 Here is a suggested fix: http://bugzilla.openmoko.org/cgi-bin/bugzilla/show_bug.cgi?id=740 Aug 22 09:59:59 mickey|taxes: so on 64bit host it get wrong results which gives error (as python guys have check for it) Aug 22 10:00:00 Would you guys be fine with this? Aug 22 10:00:27 It fixes the build for me. Aug 22 10:00:33 stefan_schmidt: or remove binaries from device_table Aug 22 10:00:43 stefan_schmidt: that's a workaround Aug 22 10:01:23 hmm Aug 22 10:01:43 hrw: Which binaries do you refer to? Aug 22 10:01:58 hrw: is that message in the archives or bugzilla? Aug 22 10:03:31 stefan_schmidt: /bin/tinylogin in device_table-minimal.txt, and the others following it in that file Aug 22 10:04:26 rwhitby, hrw: Ah, now I really understand the problem. It's ugly indeed. Aug 22 10:04:56 hmm there is no glibc-2.3.6.bb .. which one will be compatible to a glibc-2.3.6 compiled apps ? Aug 22 10:05:58 seems that makedevs on 32-bit debian etch at least creates whatever intermediate directories it requires (like /bin), but for some reason that's not happening for some people. Aug 22 10:07:14 rwhitby: debian unstable 64-bit here. Not sure on which systems are the others. Aug 22 10:08:10 But if makedevs creates the dir it's more or less the same then doing it with install. The big question is now how you guys like see this fixed? Aug 22 10:14:58 * * OE Bug 2823 has been created by autobuild(AT)openembedded.org Aug 22 10:15:00 * * abiword-plugins-2.5.2-r0-do_configure Aug 22 10:15:02 * * http://bugs.openembedded.org/show_bug.cgi?id=2823 Aug 22 10:18:43 stefan_schmidt rwhitby : If its a host thing/problem how about get a makedevs-native and use that ? Aug 22 10:19:18 steliosk: Still the problem persists that makedevs installs binaries into /bin Aug 22 10:19:31 steliosk: I'll update http://bugs.openembedded.org/show_bug.cgi?id=2820 Aug 22 10:19:51 stefan_schmidt rwhitby : BTW what trigered that ? it did not showed up a few days ago Aug 22 10:20:41 steliosk: Perhaps a distro update for makdevs? I use debian unstable with more or less daily updates. Aug 22 10:22:03 stefan_schmidt : weird. I don't think my build machine was updated the last few days Aug 22 10:22:18 I build on a ubuntu box, and I have not updated that one as far as I know Aug 22 10:22:26 hmm Aug 22 10:22:35 So that theory goes away Aug 22 10:22:51 Any changes on makedevs in OE the last days? Aug 22 10:22:59 But, I havent build efika before. Aug 22 10:23:22 my previous build was for epia. Not that it probably matters. Aug 22 10:23:26 efika was building fine till a few days ago. my last clean build was last week Aug 22 10:23:31 mickey|taxes: OE archive Aug 22 10:24:01 mickey|taxes: 23.03.2007 - Building Python extensions - mail by me Aug 22 10:24:49 btw, makedevs is in the oe repository isn't it? I don't have it installed on my ubuntu box Aug 22 10:24:55 I have updated #2820 Aug 22 10:25:01 Time for lunch now. Later Aug 22 10:25:45 Maybe the makedevs/makedevs-native packages in OE been updated Aug 22 10:26:33 is this an efika only problem ? what other machines are efected ? Aug 22 10:27:04 makedevs doesn't install things in /bin - it just changes the permissions of existing things I believe Aug 22 10:27:41 I don't know. I don't have enough disk space for epia and efika at the same time. My last build on epia was a few days ago. I then changed to efika, and got the problem. Aug 22 10:28:01 I noticed tinylogin is not installed in efika Aug 22 10:28:35 I think it was in epia, but I might recall incorrecly. Aug 22 10:29:00 (used the same distro just different arch) Aug 22 10:29:21 http://imgs.xkcd.com/comics/orphaned_projects.png Aug 22 10:31:24 so looking at openembedded/packages/makedevs/makedevs-1.0.0/makedevs.c, it is clear that it will not create /bin if /bin/tinylogin is referenced in the device table, and there is no directory reference to /bin earlier than that in the file. Aug 22 10:31:54 so perhaps device_table-minimal.txt should be updated with a directory entry for /bin in it before the entries for /bin/tinylogin and others. Aug 22 10:32:29 erm, is it me, or is fiddling with /bin an insane thing to do in device_table-minimal.txt Aug 22 10:32:39 if it had an entry for /dev in it too, then the mkdir of /dev would not be required in image.bb Aug 22 10:32:43 that sounds like a hack that should die yesterday Aug 22 10:32:57 XorA: I guess it's there to set /bin/tinylogin to setuid. Aug 22 10:33:12 the other way is for that to be done in a postinst, right? Aug 22 10:33:26 rwhitby: its totally unintuitive and would be next to impossible to locate for people trying to debug Aug 22 10:33:34 XorA: fully agree. Aug 22 10:33:40 last place I would look for would be the device creator :-) Aug 22 10:33:51 the question is: is there another way to do the same thing? Aug 22 10:35:03 tinylogin, mount and umount are the three files in question, all are setuid. Aug 22 10:35:07 postinst should work, as none of those 3 commands is called by a non-root user until after ipkg configure in first boot Aug 22 10:35:50 although Im not sure tinylogin should be SUID Aug 22 10:35:57 RP : It looks like the .lock files where leftovers from other builds Aug 22 10:36:21 that sounds like a security flaw waiting to happen Aug 22 10:36:33 steliosk: there was a period where .lock files were left over from files:// uri's, but it's been fixed for a couple of days now. Aug 22 10:38:08 those lines in device_table-minimal.txt haven't changed since bitkeeper, so it's a long-standing thing. Aug 22 10:41:01 steliosk: great :) Aug 22 10:41:11 RP : One more question. I have a kernel that uses git-native and git-native is added on the DEPENDS list but when enabling mutlithread in bitbake, git-native does not get build before the do_fetch of the kernel. any idea why ? Aug 22 10:41:30 * RP has wondered about the sanity of the /bin entries in that device creation file Aug 22 10:41:48 well /bin/tinylogin as SUID is not sane :-( Aug 22 10:41:56 traditionally mount/umount are always SUID Aug 22 10:42:31 * RP also votes for adding a /dev entry to remove the need for the mkdir Aug 22 10:42:39 03mickeyl 07org.oe.dev * re542d324... 10/ (1 conf/distro/include/sane-srcdates.inc): sane-srcdates: pinpoint OpenMoko U-Boot patches Aug 22 10:42:52 steliosk: Good question, it sounds like something is broken... Aug 22 10:43:36 steliosk: git-native should be added as a dependency of the fetch task though, not added to DEPENDS Aug 22 10:43:47 steliosk: Have a look at the output of bitbake kernel -g Aug 22 10:44:22 looks like makedevs.c was last modified by likewise on 2007-08-20 Aug 22 10:44:38 Revision: b396aa55286b5d00bbc2fba14ef9dd30b82a3af9 Aug 22 10:44:45 makedevs: Add support for creating directories. Aug 22 10:44:46 Hmm. What did he do? :) Aug 22 10:45:42 looking for a working viewmtn now ... Aug 22 10:47:33 rwhitby: Its a strange patch. It looks like it might have just used to have done nothing and it errors now? :/ Aug 22 10:47:51 can you pastebin the patch? Aug 22 10:48:06 RP : but if i have package A that depends on another (B), shouldn't (B) be build first (before A gets fetched or whatever) Aug 22 10:48:54 steliosk: yes Aug 22 10:49:15 steliosk: That is what should be happening. It isn't done through DEPENDS though, its a different kind of dependency Aug 22 10:49:54 rwhitby: http://pastebin.ca/666678 Aug 22 10:50:29 RP: yep, that's what's causing the error. Aug 22 10:50:44 rwhitby: but did it used to do nothing? Aug 22 10:50:54 so either that patch should be reverted, or directory entries need to be added to the device_table Aug 22 10:51:04 RP: seems so Aug 22 10:51:30 in which case, those lines could just be removed from device_table-minimal.txt with no adverse effects. Aug 22 10:51:30 rwhitby: or if those permissions we're created, remove them from the file Aug 22 10:51:35 right :) Aug 22 10:51:44 Does anyone have an older image to see? Aug 22 10:52:05 RP: OZ images shouldbe available Aug 22 10:52:16 RP: I've got slugos releases Aug 22 10:53:12 yes, at the time of slugos 3.10 release, makedevs.c had an empty function for add_new_file Aug 22 10:53:28 Nov 11 2006 Aug 22 10:54:02 but somehow tinylogin ended up setuid in tmp/rootfs/bin Aug 22 10:54:07 mount and umount seem to be none suid, tinylogin is suid in poky Aug 22 10:54:24 and I think poky has the same setup as OE for these bits Aug 22 10:54:42 mount is an alternatives symlink to mount.util-linux and mount.util-linux is indeed setuid Aug 22 10:54:56 I think tinylogin has to be suid so it can log in users and then drop to the right uid Aug 22 10:55:20 rwhitby: I just have busybox mount Aug 22 10:55:39 sluos used util-linux mount so we could do nfs rootfs Aug 22 10:55:56 (at that time busybox couldn't do nfs) Aug 22 10:56:33 umount.util-linux is also suid Aug 22 10:57:34 RP: tinylogin should be run by root, then drop, if its SUID then any user can elevate back to root again, which is dangerous, /bin/login is not SUID on any distro I know of Aug 22 10:58:16 XorA: true, it is started by root. Hmm :/ Aug 22 10:58:42 RP: unless them is some special case for tinylogin, Im not familar with its architecture Aug 22 10:59:07 * RP neither Aug 22 11:00:57 * XorA gets his terrier out and sets tinylogin to not SUID :-) Aug 22 11:06:03 RP: tinylogin == su so needs SUID :-D Aug 22 11:07:09 XorA: right :) Aug 22 11:08:53 hi zecke Aug 22 11:09:29 ho Aug 22 11:10:19 RP: What kind of dependency is that ? any example ? Aug 22 11:15:48 steliosk: base.bbclass, grep for "git". It basically sets do_fetch[depends] = "git-native:do_populate_staging" Aug 22 11:24:10 koen: did you have another go at llvm? Aug 22 11:24:30 zecke: I did, it now has a compile error Aug 22 11:30:04 koen: oh the arm backend was (partially) by indt? Aug 22 11:38:28 koen: which compile error? Aug 22 11:41:50 hiya Aug 22 11:43:19 is there any use in me patching makedevs.c to make full paths when requested? i.e. is it better to do such things as /bin/tinylogin there than from image.bbclass? Aug 22 11:46:32 03koen 07org.oe.dev * ra9b77dce... 10/ (4 files in 2 dirs): abiword*: update to 2.5.2 Aug 22 11:46:42 XorA: there you go Aug 22 11:47:03 XorA: now we can take distributed notes at OEDEM on our neos :) Aug 22 11:48:43 koen: you think that neos will be as much usable then? Aug 22 11:49:08 hrw: as a phone, no, but we only need abiword + mb-keyboard :) Aug 22 11:50:14 then I will run abiword on my d400 rather ;) Aug 22 11:53:23 * XorA doesnt even get mb-keyboard on neo at the moment :-( Aug 22 12:07:01 torpor: I think someone did patch it to allow path creation recently, thats part of the problem Aug 22 12:16:32 hello Aug 22 12:20:32 so what about the makedev issue ? Should the commit creating the problem be disapproved until we found a way around it ? Aug 22 12:27:12 steliosk: No, we need to fix this properly. To do that we need some facts like whether the /bin entries do enter the image from the source somehow or not Aug 22 12:52:25 koen: do you have any issues with linking to the pdf copy of the OE manual from: http://www.openembedded.org/wiki/MainPage? Are the pdf and html copies kept in sync? Aug 22 12:54:11 no problem, but the pdf isn't updated automagically Aug 22 12:54:37 koen: is the html updated automatically, or how does that work? Aug 22 12:54:53 the html is update automagically Aug 22 12:56:06 koen, do you know if the gumstix bad magic number was resolved? Aug 22 12:56:33 nope Aug 22 13:00:10 03cbrake 07org.oe.dev * r1adf7728... 10/ (1 conf/machine/compulab-pxa270.conf): compulab-pxa270.conf: remove MACHINE_ESSENTIAL_EXTRA_RDEPENDS and replace with documentation Aug 22 13:01:06 are you talking about the bad magic error immediately after flashing the kernel in uboot? Aug 22 13:01:18 yeah Aug 22 13:01:44 i dont think it's fixed but i will reflash right now and let you know Aug 22 13:02:22 asmola_, you see bad magic number? Aug 22 13:02:46 im checking right now Aug 22 13:03:21 I think we need more gumstix machines Aug 22 13:03:28 one for each model Aug 22 13:03:38 * Crofton|home agreed Aug 22 13:03:40 i saw that people on the listserv were having issues, my unit booted fine after the changes as of 6pm est yesterday Aug 22 13:03:46 hmmm Aug 22 13:03:53 (since you can't build a mixed pxa255/pxa270 kernel) Aug 22 13:04:04 the problem we are having is the dev's do not have any gusmtix Aug 22 13:04:13 so we are working blind Aug 22 13:04:22 http://www.angstrom-distribution.org/unstable/images/gumstix/20070821/ Aug 22 13:04:26 steve is on low end gumstix Aug 22 13:04:36 asmola_, is on high end .... Aug 22 13:05:29 i also noticed that the kernel is detecting iwmmx coprocessor but it gives the 'kernel support is missing' Aug 22 13:05:52 there were some odd filesystem mount issues too, ill post a dmesg output asap Aug 22 13:06:03 (assuming that's something you would be interested in) Aug 22 13:06:16 koen: there was proposal for more gumstix configs Aug 22 13:07:37 hrw we may need more configs to support different kernels by machine? Aug 22 13:08:26 yeah, i was thinking of suggesting that, using buildroot they hve you select machine prior to doing any builds Aug 22 13:08:35 yeah Aug 22 13:08:54 the machine name is used to select the kernel source Aug 22 13:09:14 so maybe they have different patch sets by model? Aug 22 13:09:46 I still have tinylogin issues building an image Aug 22 13:10:11 by the looks of it only a new defconfig is needed for each machine Aug 22 13:10:22 ok Aug 22 13:10:34 * koen discovers buspowered usb scanners Aug 22 13:10:35 does that mean we need to use different machine names? Aug 22 13:10:47 yes Aug 22 13:10:53 e.g. gumstix-verdex Aug 22 13:11:00 i just reflashed, the magic number is being set properly Aug 22 13:11:02 ok Aug 22 13:11:07 for verdex :) Aug 22 13:11:11 yes :) Aug 22 13:12:04 if people weren't some damn nice, I would never hang out here ,.,,, Aug 22 13:13:44 asmola_, so you can build working images? Aug 22 13:13:52 with some hiccups? Aug 22 13:14:10 can you post the problems to the bug tracker Aug 22 13:14:31 bugs.openembedded.org, be sure to list gumstix-verdex in the report Aug 22 13:14:48 yeah i can do that Aug 22 13:15:04 thanks, that way we won't lose it, we will just be able to ignore it :) Aug 22 13:15:12 rofl Aug 22 13:15:44 hrw, you are teh celinux-test distro guy? Aug 22 13:19:14 morning Aug 22 13:19:19 gm Aug 22 13:19:21 just as I make te pxa255/pxa270 comment eric miao sends a patch to LAK to solve that :) Aug 22 13:20:00 hey chouimat Aug 22 13:20:27 Crofton|home: nope now Aug 22 13:20:37 http://bugs.openembedded.org/show_bug.cgi?id=2078 Aug 22 13:20:47 NOTE: package glibc-2.6.1-r0: task do_fetch: started Aug 22 13:20:52 lists it as distro, but the problem may be "generic" Aug 22 13:21:09 I am trying to eliminate all bugs with "omap" in them Aug 22 13:21:34 * XorA nukes Crofton|home and declares OMAP not supported, there problem solved Aug 22 13:21:54 hrw: did you get my mail about ELC in linz? Aug 22 13:21:54 * Crofton|home kicks XorA Aug 22 13:22:10 Crofton|home: patch looks ok Aug 22 13:22:13 * koen hides his 2 omap devices Aug 22 13:22:40 hyyeeee war ... diaspora !!! :P Aug 22 13:22:56 dos anyone know what runltp is ? Aug 22 13:23:05 gremlin[it]: I saw lacuna coil this weekend Aug 22 13:23:11 Crofton|home: its part of LTP package Aug 22 13:23:11 Crofton: linux test project Aug 22 13:23:23 koen: I did, but no idea about funds Aug 22 13:23:59 should we apply the patch and close the bug? Aug 22 13:24:02 uhhh nice ... i like them ... theyr cover of "enjoy the silence" is great :) Aug 22 13:24:17 Crofton|home: yes Aug 22 13:24:18 gremlin[it]: they played that as well :) Aug 22 13:24:52 koen, what omap devices? Aug 22 13:24:56 mhh they played also the cover of "big in japan" ... ? Aug 22 13:26:04 koen: going all teeny goth on us :-D Aug 22 13:28:31 hi Aug 22 13:30:09 Does anyone here use the ltp? Aug 22 13:33:10 Crofton|home: I am currently considering using it on a project. Aug 22 13:33:21 Crofton|home: but don't have any experience with it yet Aug 22 13:33:25 I am goign to apply the patch Aug 22 13:34:14 let us know if you have any problems? Aug 22 13:34:37 I will update the bug as patch applied, needs testing to close Aug 22 13:34:58 XorA: If I saw NiN and chris cornell as well, does that make me a teeny emy industrial goth? Aug 22 13:35:12 koen: heh heh yeah Aug 22 13:35:28 koen: NIN played Edin last night, but supporting the bloody Fu Fighters Aug 22 13:35:33 koen: so I didnt bother going Aug 22 13:35:45 they were LOUD Aug 22 13:36:14 koen: thats the way we like them Aug 22 13:36:15 that is why I where ear plugs at live music events these days :) Aug 22 13:36:32 I started wearing earplug 5 years ago :) Aug 22 13:37:09 back in the "good old days", when I frequented bars with bands Aug 22 13:37:18 LOUD as is "wear a belt or lose you trousers" Aug 22 13:37:24 there was "an old guy" who wore over the head ear muffs while dancing Aug 22 13:38:25 although loudest I ever heard was the Mission who didnt seem to get how to work volume controls, yanked everything to 11 and succeeded into turning all sound into square waves Aug 22 13:43:41 screen -rd Aug 22 13:46:07 03mickeyl 07org.oe.dev * rc3529caf... 10/ (1 classes/distutils.bbclass): distutils.bbclass: always install staging directory before calling python Aug 22 13:46:12 03mickeyl 07org.oe.dev * r9aa7ad52... 10/ (1 packages/python/python-evas_cvs.bb): python-evas: bump PR as per distutils change Aug 22 13:46:24 grrr, now QA issues with ltp Aug 22 13:46:50 ERROR: QA Issue: non debug package contains .debug directory: ltp path /work/armv5te-angstrom-linux-gnueabi/ltp-20070228-r0/install/ltp/usr/libexec/ltp/testcases/misc/math/float/.debug/float_iperb Aug 22 13:46:56 and millions like it Aug 22 13:46:58 * * OE Bug 2824 has been created by autobuild(AT)openembedded.org Aug 22 13:47:00 * * abiword-embedded-2.5.2-r0-do_install Aug 22 13:47:02 * * http://bugs.openembedded.org/show_bug.cgi?id=2824 Aug 22 13:47:10 * * OE Bug 2825 has been created by autobuild(AT)openembedded.org Aug 22 13:47:12 * * abiword-plugins-2.4.6-r0-do_compile Aug 22 13:47:15 * * http://bugs.openembedded.org/show_bug.cgi?id=2825 Aug 22 13:49:51 * XorA hates the way the press works sometimes, people narrowly escape fines by not actually doing anything illegal Aug 22 13:53:57 * * OE Bug 2825 has been RESOLVED (INVALID) by Aug 22 13:53:59 * * abiword-plugins-2.4.6-r0-do_compile Aug 22 13:54:01