**** BEGIN LOGGING AT Sun Mar 15 02:59:57 2009 Mar 15 04:58:13 looks like that last commit to OE did some funny stuff to the repo. Someone with some git foo needs to take a close look at it... Mar 15 04:58:50 Can no longer view the version tree with gitk -- the dev branch is ambiguous, apparently. Mar 15 04:58:59 (that's on a fresh copy) Mar 15 04:59:28 Existing copies will do a merge after doing a git pull -- as if suddenly they pulled a different branch. Mar 15 04:59:52 Oh. A fresh copy is also missing many weeks of commits. Mar 15 04:59:58 :( Mar 15 05:00:26 [Actually I expect they are there somewhere, just not on what it now thinks is the dev branch] Mar 15 08:46:58 hello Mar 15 08:47:08 how can I turn on the gps of the n810 using OE? Mar 15 08:47:22 we have this http://bugzilla.o-hand.com/show_bug.cgi?id=729 Mar 15 08:47:26 pathc for gypsy Mar 15 08:47:35 http://bugzilla.o-hand.com/attachment.cgi?id=447 Mar 15 08:47:41 but it requires a gpsctrl.h Mar 15 08:47:52 which is supposed to be in libgpsbt Mar 15 08:47:54 but isn't Mar 15 08:48:37 <_diego_> have you tried to look in gpsd Mar 15 08:48:38 <_diego_> ? Mar 15 08:58:20 _diego_, not yet Mar 15 08:58:29 but gpsd uses libgpsbt to turn on the gps Mar 15 08:58:39 yet in the libgpsbt that I have on OE Mar 15 08:58:46 I see no function that does it Mar 15 12:51:56 Wow, busy place on Sunday mornings... Mar 15 12:52:57 OE git repo still broken... Mar 15 12:53:51 * mwester goes to get a cup of coffee and his git phrasebook, to see what the phrasebook says to type when the master git repo is messed up... :-D Mar 15 13:38:01 hi Mar 15 13:38:07 anybody here ? Mar 15 14:06:09 mwester, git pull is working for me Mar 15 14:06:45 can you check your "git log" and see if it reflects that you have just done a merge locally? Mar 15 14:08:39 The history for a pre-existing OE repo after a git pull now looks like this: Mar 15 14:08:58 Merge: (by me) Merge branch 'org.openembedded.dev' of git://git.openembedded.net/openembedded into org.openembedded.dev Mar 15 14:09:10 java.bbclass: Make oe_makeclasspath more flexible (partly by Marcin Juszkiewicz ) Mar 15 14:09:20 preferred-om-2008-versions.inc: Update to version that are available Mar 15 14:09:23 etc. Mar 15 14:09:50 That all looks ok, except for some strange messages from git, and the fact that the most recent entry (the local merge) should not exist. Mar 15 14:10:25 And if you clone the master git repo, you get only the java.bbclass at the head, and it is missing several months worth of commits. Mar 15 14:12:51 hmmm Mar 15 14:13:00 it did rebuild some crap Mar 15 14:13:13 and I have a Merge at the top Mar 15 14:13:29 commi tbelow that is java.bbclass Mar 15 14:15:06 I have no idea what will happen if that merge gets pushed back... that might fix everything, or it could make it worse. My git foo is lacking. Mar 15 14:26:56 mmm could i ask how to manually check out an hh kernel for h5000 ? Mar 15 14:27:31 oe knows how/where to check it out, but ... i am with out any working oe at the moment Mar 15 14:37:53 any h2200 users around? Mar 15 14:39:35 h5500, sorry Mar 15 14:39:56 pocek are you able to manually check out hh kernel ? Mar 15 14:40:13 flameman: never tried Mar 15 14:40:55 could you ask your bitbake to do that, then tgz the source tree ? Mar 15 14:44:01 flameman: `bitbake -b packages/linux/linux-handhelds-2.6_cvs.bb -c fetch` doesn't work for you? Mar 15 14:44:36 i have no bitbake working on this machine Mar 15 14:49:36 mwester, agreed so is mine Mar 15 14:51:09 luckily, I have time doing carpet cleaning today Mar 15 14:57:28 * mwester will go replace a CPU cooler in the quad-core, I guess... Mar 15 15:06:05 i need it just to speedup things, i am developing a special rootfs from gentoo-arm Mar 15 16:20:20 ok i have fixed the oe on my host Mar 15 16:20:30 i have launched a fetch from bitbake Mar 15 16:20:36 it has success Mar 15 16:20:49 where does it park the unpaked files ? Mar 15 16:22:45 tmp/work/... Mar 15 16:24:40 umm my tmp/work seems to be empty Mar 15 16:25:26 i mean i asked bitbake to "bitbake -b packages/linux/linux-handhelds-2.6_cvs.bb -c fetch" Mar 15 16:26:19 i see a the tgz in the oe.distfiles folder Mar 15 16:26:45 i saw a list a files unpacking ... when i launched the command Mar 15 16:27:04 but ... i can't locate them Mar 15 16:31:46 ok i fixed the problem Mar 15 16:32:00 flameman: hey, how is going with your h5500? Mar 15 16:34:42 bad :( Mar 15 16:34:59 the hh kernel is not compiling Mar 15 16:35:07 Is there some way to disable compiled in input module? Mar 15 16:35:23 yes, editing the relative makefile Mar 15 16:35:32 my problem is ... sources issues stuffed Mar 15 16:35:47 i get issues just out our the unpack + defconfig Mar 15 16:36:04 and vanillas are too bad supporting :( Mar 15 16:36:15 hh.org is a shit place for kernel tree Mar 15 16:36:54 Jay7 .... off course, i found out the issues about the bootloader Mar 15 16:37:12 googling a bit i found out other people with this issue Mar 15 16:37:37 both bootldr and haret are issued :( Mar 15 16:37:53 also everything in the h5000 angstrom web folder is faked Mar 15 16:38:20 (i tested, nothing is really working .... neither the live, nor the kernel.bin) Mar 15 16:40:02 - Mar 15 16:40:28 Jay7 ... umm yesterday you had success in compiling a kernel for h5000 Mar 15 16:40:51 could you tell me details ? Mar 15 16:41:26 flameman: I just have changed MACHINE from akita to h5000 and bitbake linux-kexecboot Mar 15 16:41:40 well bitbake -b packages/linux/linux-handhelds-2.6_cvs.bb Mar 15 16:41:46 fails great Mar 15 16:41:53 ah, and change DEFAULT_PREFERENCE of kernel 2.6.21-hh Mar 15 16:41:58 cause of issues in the cvs tree Mar 15 16:42:23 yeah, it may be DEFAULT_PREFERENCE of kernel 2.6.21-hh is the fixup Mar 15 16:42:35 linux-kexecboot have using 2.6.21 as kernel for rootfs Mar 15 16:42:41 i know Mar 15 16:42:44 the point is Mar 15 16:42:45 but tries to use 2.6.26 as own kernel Mar 15 16:42:56 but fails too Mar 15 16:43:02 no defconfig Mar 15 16:43:14 so I switched it to 2.6.21-hh too Mar 15 16:43:21 that's all :) Mar 15 16:43:26 the point is ..... 2.6.21-hh has a lot of shit in the hh.org cvs Mar 15 16:43:49 there are dead and broken branch in there Mar 15 16:46:27 - Mar 15 16:46:59 umm where to set DEFAULT_PREFERENCES of kernel ? what it the exact line in the oe .conf ? Mar 15 16:47:19 i want to force 2.6.21-hhxx (xx=known good) Mar 15 16:49:24 jay7 have you still got the oe/work files ? Mar 15 16:52:11 flameman: it was needed only for linux kexecboot Mar 15 16:52:31 one moment I will check Mar 15 16:52:34 yes but ... if you still have the hh tree Mar 15 16:52:43 i can ask if you could tarball it Mar 15 16:52:52 cause ... here i am in a very deep jam Mar 15 16:53:03 that is yesterdays oe.dev Mar 15 16:53:16 yesterday Mar 15 16:53:43 your kernel is the only 2.6 thing i found out really booting and working Mar 15 16:54:00 only file changed is linux-kexecboot_2.6.21-hh20.bb Mar 15 16:54:18 yes, but i am on a ppcG4 Mar 15 16:54:27 with less then 400Mb of free space Mar 15 16:54:40 i already have all i need to crosscompile a kernel Mar 15 16:55:00 invoking bitbake ... is amazing a lot Mar 15 16:55:09 i only need a good and pretty kernel tree Mar 15 16:55:15 just a fetch Mar 15 16:55:21 well, one moment Mar 15 16:55:51 no problem Mar 15 16:56:02 :) Mar 15 17:03:49 flameman: well, by default bitbake linux trying to build 2.6.25 Mar 15 17:03:53 is it broken? Mar 15 17:04:21 dunno, never tested Mar 15 17:04:28 i tested 21, 29 Mar 15 17:04:33 both are broken Mar 15 17:05:16 well, will search where is stored default preffered version for rootfs Mar 15 17:05:20 mmm how you forced the bitbake to compile 2.6.21 ? Mar 15 17:06:54 it was for kexecboot Mar 15 17:07:01 I'm looking for this Mar 15 17:07:17 yes, but how you forced the preferred version ? Mar 15 17:08:33 well, I have found: bitbake virtual/kernel Mar 15 17:08:48 this will build linux-handhelds-2.6-2.6.21-hh20-r25 Mar 15 17:08:55 for h5000 machine Mar 15 17:08:58 ok Mar 15 17:09:03 but my problem is Mar 15 17:09:16 how to fetch a kernel tree Mar 15 17:09:20 this should be same kernel that I was sent you Mar 15 17:09:28 hm.. Mar 15 17:09:35 if i invoke the bitbake Mar 15 17:09:51 bitbake -c fetch virtual/kernel Mar 15 17:09:59 is it? Mar 15 17:10:06 may be, dunno at all Mar 15 17:10:16 well, wait for some time Mar 15 17:10:20 also ... how to force the version ? Mar 15 17:10:37 I think that it should be default 2.6.21 Mar 15 17:10:50 ok, and what if it is not ? Mar 15 17:11:35 well, PREFERRED_PROVIDER_virtual/kernel = "linux-handhelds-2.6" will point to linux-handhelds-2.6 in packages/linux/ Mar 15 17:12:18 so ? is not possible to force a version ? Mar 15 17:13:01 flameman: set DEFAULT_PREFERENCE = "1" in appropriate linux-handhelds-2.6*.bb recipe Mar 15 17:13:19 by default most of them have is "-1" Mar 15 17:13:23 *it Mar 15 17:14:05 ok Mar 15 17:14:09 but the problem is Mar 15 17:14:09 but 2.6.21-hh20 is biggest available version in linux-handhelds family Mar 15 17:14:23 the problem is Mar 15 17:14:37 cvs -d:pserver:anoncvs:anoncvs@anoncvs.handhelds.org:/cvs co -D "20090315 UTC" linux/kernel26 Mar 15 17:14:51 is the manual procedure for 2.6.21-hh Mar 15 17:15:02 it is pretty broken Mar 15 17:15:11 bitbake -b packages/linux/linux-handhelds-2.6_cvs.bb -c fetch Mar 15 17:15:17 it pretty broken the same Mar 15 17:15:55 i could invoke bitbake again for bitbake -c fetch virtual/kernel Mar 15 17:16:11 hoping it will fetch the right hh branch Mar 15 17:16:18 21-hh20 Mar 15 17:17:16 -rw-r--r-- 1 jay users 58560113 δΕΛ 20 2007 linux.kernel26_anoncvs.handhelds.org_K2-6-21-hh20_.tar.gz Mar 15 17:17:36 good :D Mar 15 17:17:44 I can share it for you but only with 40-45K per second bandwidth :) Mar 15 17:17:45 i need exactly this file :P Mar 15 17:17:52 no problem :P Mar 15 17:17:54 :D Mar 15 17:17:55 :D Mar 15 17:17:57 well, wait some time Mar 15 17:18:19 do you trust me if i say ... we are developing a pretty patch set for h3600 ? Mar 15 17:18:37 it means that in the future bitbake will have to fetch from vanilla Mar 15 17:18:51 then simply applying a specific patch set Mar 15 17:18:54 easy easy :D Mar 15 17:19:09 flameman: http://jay-home.ath.cx/linux.kernel26_anoncvs.handhelds.org_K2-6-21-hh20_.tar.gz Mar 15 17:19:11 (also on h3600 we are around 2.6.29 :P) Mar 15 17:19:16 flameman: grab it :) Mar 15 17:19:21 thank you so much :D Mar 15 17:19:56 np :) Mar 15 17:20:44 :)\ Mar 15 17:21:02 also i am trying to emerge a complete stage3 nptl - less Mar 15 17:21:14 cause on h5500 we have a pretty 2.4 kernel Mar 15 17:21:38 i wonder if a 2.6 stage3 could work on 2.4 kernel if i re-emerge it nptl - less Mar 15 17:21:52 main problems are around glibc-2 and bash Mar 15 17:22:11 btw, who knows what city in Germany have this phone code (or is it mobile?): +491766? Mar 15 17:22:26 * Jay7 have missed call from such number :) Mar 15 17:26:00 Jay7: http://en.wikipedia.org/wiki/+49 Mar 15 17:26:05 (yes, mobile) Mar 15 17:27:02 pH5: thanks :) may be someone just mistyped number.. Mar 15 17:28:50 i have several problem with ubuntu 8.10 amd64 and oe Mar 15 17:29:29 i don't know if there are know bug (Ubuntu adm64/openembedded) Mar 15 19:07:38 jay7 are you still there ? Mar 15 19:07:59 yes Mar 15 19:20:10 is someone try to build beagle demo image with the current git tree? Mar 15 19:23:24 when I try it it always failled in the FFmpeg build Mar 15 19:25:27 jay7 .... could you see if the h5000 needs a patch, too ? Mar 15 19:25:37 it seems the touch screen is not working Mar 15 19:25:43 not supported, i mean Mar 15 19:30:37 flameman: look here http://cgit.openembedded.net/cgit.cgi?url=openembedded/tree/packages/linux Mar 15 19:31:25 files linux-handhelds-2.6_2.6.21-hh20.bb and linux-handhelds-2.6.inc for SRC_URI with file:// entries Mar 15 19:31:28 this is patches Mar 15 19:31:35 *are Mar 15 19:33:03 yes, but .... is any of these patches applied to the kernel of h5000 ? Mar 15 19:33:38 all specified in SRC_URI on both files Mar 15 19:34:29 file://linux-2.6.git-9d20fdd58e74d4d26dc5216efaaa0f800c23dd3a.patch;patch=1 \ Mar 15 19:34:29 14 http://www.rpsys.net/openzaurus/patches/archive/export_atags-r0a.patch;patch=1 \ Mar 15 19:34:29 15 file://gcc4x-limits.patch;patch=1 \ Mar 15 19:34:29 16 file://defconfig" Mar 15 19:34:37 these ones ? Mar 15 19:34:42 yes Mar 15 19:34:52 with ';patch=X' Mar 15 19:35:09 sorry, forget about it Mar 15 19:35:16 grr.. Mar 15 19:35:46 applied only entries with ';patch=X' at end Mar 15 19:35:56 uuu pretty :D Mar 15 19:36:44 other is just files that dropped to working directory of recipe Mar 15 20:55:02 Hallo everyone. I've been having some problem with compiling using autoconf 2.63-native. It was missing a patch for "--exclude". I've written the patch. What do I do with it? Mar 15 20:56:38 re Mar 15 20:56:57 catsmuvva: send it to the mailinglist Mar 15 21:03:36 Does anyone know if the "Peek" mobile device is capable of running oe? Mar 15 21:03:51 Its based on the TI Locosto chipset Mar 15 21:03:51 http://blog.makezine.com/archive/2008/09/peek_email_only_device.html Mar 15 21:04:18 more info: http://focus.ti.com/general/docs/wtbu/wtbuproductcontent.tsp?contentId=15408&navigationId=12656&templateId=6123 Mar 15 21:04:33 http://www.geekypeek.com/?p=13 Mar 15 21:19:24 xlogik: Is there a linux port to this platform already? Mar 15 21:20:35 it is arm7, but appararantly only 8mb of ram Mar 15 21:21:04 Florian: I'm unsure.... Mar 15 21:26:07 that's no fun... mirght be more something for Nut/OS or sinmilar Mar 15 21:47:46 how can I find out what happened to my gamecube commit? I cannot find it with git log | grep gamecube anymore. Mar 15 21:48:34 java.bbclass Mar 15 21:50:26 git pull Mar 15 21:50:26 classes/java.bbclass: needs update Mar 15 21:50:26 conf/bitbake.conf: needs update Mar 15 21:50:26 fatal: Entry 'classes/java.bbclass' not uptodate. Cannot merge. Mar 15 21:50:27 Merge with strategy recursive failed. Mar 15 21:53:06 emlinux_: you likely have local changes intefering with the merge Mar 15 21:54:12 Thanks Mar 15 21:57:57 kergoth: there is a suspicious commit it seems... Mar 15 21:58:07 ah Mar 15 21:58:29 I've read some complaints in the logs Mar 15 21:58:41 and don't dare tu pull --rebase Mar 15 22:01:10 is it possible that something vanishes completely from the git log? if the commit got reverted would I see that? Mar 15 22:03:01 that java.bbclass commit has an insane parent (too old) Mar 15 22:03:49 ant_home: hey :) Mar 15 22:04:12 hello Jay7 Mar 15 22:05:55 java.bbclass - still "needs update" - Looks I have a clean local config, any idea what is causing this? Mar 15 22:05:59 bitbake virtual/kernel -c devshell & make zImage: arm-angstrom-linux-gnueabi-ld: unrecognized option '-Wl,-rpath-link,/backup/stuff/build/tmp/staging/armv5te-angstrom-linux-gnueabi/usr/lib' :-( Mar 15 22:06:09 Any ideas? Mar 15 22:07:32 pocek: heh, the do_compile() method in kernel.bbclass is there for a reason Mar 15 22:07:44 if "make zImage" was all that was needed then it would be much shorter :-} Mar 15 22:08:19 ah... thanks Mar 15 22:09:12 that particular problem is due to the kernel using an unexpected setting for ${LD} which doesn't agree with oe's default LDFLAGS. Mar 15 22:09:34 but, in general, you're probably best off just doing the same thing that do_compile() does. Mar 15 22:12:16 is it possible to chroot a rootfs made for kernel 2.6, into kernel 2.4 ? Mar 15 22:12:26 i mean suppose the roofs is NOT nptl Mar 15 22:12:34 do you think it is possible ? Mar 15 22:13:10 what do you mean by "made for kernel 2.6"? Mar 15 22:13:45 modules aside, there is no inherent reason why the two should be incompatible, though there are plenty of things that 2.6 can do and 2.4 can't. Mar 15 22:14:01 obviously if your rootfs happens to rely on any of those new things then it won't work under the older kernel. Mar 15 22:15:56 for example: no sysfs in 2.4, so udev won't work, so you might not have a /dev if you are relying on udev to make it. Mar 15 22:16:21 no eabi in 2.4, so you might not even be able to start init if your userland is compiled eabi. Mar 15 22:19:38 plus of course sundry new syscalls and minor abi features which your libc might be relying on if it was built to expect 2.6 Mar 15 22:21:58 mr_nice: something is broken in the git repo. As soon as a git god wakes up, they will surely know how to fix it, and then all the missing commits should re-appear. Mar 15 22:22:36 mwester: thx Mar 15 22:22:37 emlinux_: same thing - git repo is very confused. You'll have to wait. Mar 15 22:23:28 pb_ chroot /mnt/diske (which has been made NPTL with 2.6.19 in mind) Mar 15 22:23:41 pb_ the chroot says "kernel too old" Mar 15 22:23:46 who is our git god? :) Mar 15 22:23:55 uname -t tells you 2.4.19 Mar 15 22:24:18 pb_ so now i am rebuilding glibc and bash NPTL-less Mar 15 22:24:53 Jay7: I'm waiting to find out. Mar 15 22:24:54 flameman: right, that is the last thing I mentioned. your glibc was built with --enable-kernel and is expecting something newer than what you are using. Mar 15 22:25:16 pb_ so good news for me: it is possible :D Mar 15 22:25:20 "kernel too old" isn't, in itself, diagnostic of an NPTL problem although nptl will certainly not work with 2.4. Mar 15 22:25:59 pb_ cause the problem is ... h5500 kernel 2.6 suxs great: nothing working at all Mar 15 22:26:37 nothing at all? that certainly is rather sad. Mar 15 22:26:37 pb_ so ... i have to remove every nptl and eabi from my rootfs, no problem about that :) Mar 15 22:27:10 pb_ at all: 2.6 vanilla are not booting, 2.6-hh (hh.org) are freezing, panic Mar 15 22:27:42 pb_ also .... hh.org tree are broken ... issues everywhere, a very amount of garbage Mar 15 22:28:11 pb_ bitbake has compile a 2.6.19-hh20 but ... it suxs great: freezing Mar 15 22:28:37 hoe good is OE at building cmake packages? Mar 15 22:28:52 boost 1.38 adds "experimental" cmake support Mar 15 22:28:55 pb_ so i am considering the 2.6 series DEAD about h5000 Mar 15 22:29:59 pb_ thank you for your explanation: it is very appreciated here :) Mar 15 22:48:17 good night Mar 15 22:52:02 :D Mar 15 22:55:00 Question to h2200 users: does power button work for you? Could you `evtest /dev/input/event0` or event1 and check "where it is", please? I don't understand why it doesn't work with 'h2200 buttons' driver but works with "gpio-keys" driver (2.6.21-hh20 kernel) Mar 15 23:24:26 nvm, I just found an old IRC log about this issue Mar 15 23:24:47 after searching for GPIO_NR_H2200_POWER_ON_N ;D Mar 16 00:13:52 i run Debian Lenny and am about to start developing an oe project. i read that oe has a high rate of development. is the Debian Lenny version lagging very behind (queue jokes about Debian Stable here)? Mar 16 00:14:48 in other words, is it reccommended to use the latest version via git? Mar 16 00:22:01 read OE and your distro wiki page Mar 16 00:22:08 yes git Mar 16 01:29:20 * mwester thumps the OE channel soundly Mar 16 01:29:31 OE repo is *****STILL***** borken! Mar 16 01:30:03 I guess I'll try to do a merge, and push the result... it would have been better if we had a git guru to consult, though. :( Mar 16 01:55:25 Ok, I don't think that worked quite right... so I think I best not push that result. It'll remain borken for a while, I guess. :( **** ENDING LOGGING AT Mon Mar 16 02:59:57 2009