**** BEGIN LOGGING AT Sun Jan 16 02:59:58 2011 Jan 16 05:54:14 thank khem` for pushing my 2 patches Jan 16 12:17:23 do we have image that contains development tools inside? e.g. gcc and gdb Jan 16 13:06:31 03Koen Kooi  07org.openembedded.dev * rb996f7e054 10openembedded.git/recipes/lm_sensors/ (sensors-applet_2.2.1.bb sensors-applet_2.2.4.bb): Jan 16 13:06:31 sensors-applet: update to 2.2.7 Jan 16 13:06:31 Signed-off-by: Koen Kooi Jan 16 13:06:48 03Koen Kooi  07org.openembedded.dev * r2d89c7274d 10openembedded.git/recipes/lm_sensors/lmsensors-apps_3.2.0.bb: Jan 16 13:06:48 lmsensors-apps: update to 3.2.0 Jan 16 13:06:48 Signed-off-by: Koen Kooi Jan 16 13:06:50 03Koen Kooi  07org.openembedded.dev * r15a7458d0a 10openembedded.git/recipes/lm_sensors/ (4 files in 2 dirs): lmsensors-apps: remove 2.10.1 and rename files/ dir Jan 16 13:14:44 well.. then I need good startpoint to create one :) Jan 16 13:22:03 03Michael 'Mickey' Lauer  07org.openembedded.dev * rbd892a60e9 10openembedded.git/recipes/vala/vala_0.11.4.1.bb: vala: add 0.11.4.1 Jan 16 14:41:17 nice, qemush4 built (minimal) for all of the images and stuff Jan 16 14:41:22 kicking uclibc, running away again Jan 16 14:49:39 http://tinderbox.openembedded.net/builds/119066/ Jan 16 14:49:49 is devimage-image supported this time? Jan 16 14:50:42 hehe.. newer changed from 2007 year Jan 16 14:50:53 s/newer/never/ Jan 16 15:25:12 03Koen Kooi  07org.openembedded.dev * r9a3e44e3d3 10openembedded.git/recipes/mplayer/ (4 files in 2 dirs): Jan 16 15:25:12 mplayer svn: bump SRCREV, refresh patches Jan 16 15:25:12 Signed-off-by: Koen Kooi Jan 16 15:25:14 03Koen Kooi  07org.openembedded.dev * r00abeb986a 10openembedded.git/recipes/ffmpeg/ffmpeg_svn.bb: Jan 16 15:25:14 ffmpeg svn: fix hardfp bug Jan 16 15:25:14 Signed-off-by: Koen Kooi Jan 16 16:38:49 * Jay7 wondered that we have no recipe for terminus font Jan 16 18:01:55 hm.. can't understand.. Jan 16 18:03:17 I've added kexec-tools to ANGSTROM_EXTRA_INSTALL and got same problem about can't satisfy kexec-tools Jan 16 18:04:56 opkg_install_cmd: Cannot install package kexec-tools. Jan 16 18:06:42 kexec-tools recipe provides only kexec and kdump.. Jan 16 18:06:44 Jay7: see if it is built Jan 16 18:07:15 Jay7: I had it once and it was because tmp crazyness that the package wasn't really built Jan 16 18:07:17 it is Jan 16 18:07:28 hm.. Jan 16 18:07:51 03Otavio Salvador  07master * r4b8d851deb 10openembedded.git/recipes/hal/ (hal_0.5.9.1.bb hal_git.bb): Jan 16 18:07:51 hal: drop git and 0.5.9.1 versions since no distro uses them Jan 16 18:07:51 Signed-off-by: Otavio Salvador Jan 16 18:07:57 hehe.. seems I understand Jan 16 18:08:05 there is no kexec-tools package Jan 16 18:08:10 because it is empty Jan 16 18:08:19 and provices only kexec and kdump packages Jan 16 18:08:28 heh see? Jan 16 18:08:41 those are horrible to figure out Jan 16 18:08:54 it's not build problem I'm sure Jan 16 18:08:57 03Otavio Salvador  07master * rd958139882 10openembedded.git/recipes/slang/ (slang/fix-uclibc.patch slang_2.2.2.bb): Jan 16 18:08:57 slang: fix 2.2.2 compilation with uclibc Jan 16 18:08:57 Signed-off-by: Otavio Salvador Jan 16 18:09:14 I suspect recipe should contain some variable about empty packages Jan 16 18:09:18 can't remember Jan 16 18:11:49 Jay7: yes; to force empty packages? Jan 16 18:11:57 ALLOW_EMPTY or something similar Jan 16 18:11:59 yes I mean this Jan 16 18:12:09 thanks, I'll grep :) Jan 16 18:14:00 Jay7: ALLOW_EMPTY_foo = "1" Jan 16 18:14:51 ALLOW_EMPTY_${PN} widely used Jan 16 18:21:12 * khem` yawns Jan 16 18:21:26 gm all Jan 16 18:21:39 Jay7: you need toolchain bits on target ? Jan 16 18:21:49 Jay7: you can use native-sdk-image Jan 16 18:22:25 otavio: did u get further into the sem_open case ? Jan 16 18:26:03 khem: I did not; still trying to figure why it fails Jan 16 18:26:13 khem: did you have been able to reproduce it? Jan 16 18:27:37 khem: is native for host-like system? Jan 16 18:30:01 03Khem Raj  07master * r004ceb0909 10openembedded.git/recipes/uclibc/ (uclibc-initial_git.bb uclibc_git.bb): Jan 16 18:30:01 uclibc-initial_git.bb: Use install_startfiles to generate startup files Jan 16 18:30:01 * There is no need to do manual make instead use available target Jan 16 18:30:01 install_startfiles Jan 16 18:30:01 Signed-off-by: Khem Raj Jan 16 18:31:04 well.. now it builds ok Jan 16 18:31:18 I'll push change to kexec-tools later Jan 16 18:34:15 How I can make an sdk for usage by our developpers? any doc where I can find information about it? Jan 16 18:34:23 Jay7: native in this case means "native on target" Jan 16 18:34:56 so you will get gcc which will be partof your image Jan 16 18:35:12 and when you flash or run that image on target you can use this gcc on target itself Jan 16 18:35:53 otavio: I did not try it on x86 Jan 16 18:36:05 but on arm I could not reproduce it Jan 16 18:36:12 khem: did you try it with eglibc? Jan 16 18:36:16 otavio: I dont have fast machines Jan 16 18:36:32 I have queued a x86/qemu build Jan 16 18:36:39 but god known when it will reach there Jan 16 18:36:55 right now its trying to build a different image for ar Jan 16 18:36:56 m Jan 16 18:37:49 I am quite lost on this Jan 16 18:38:00 this works on Debian but fails on OE Jan 16 18:38:10 otavio: should be easy if I can see it happening on my system Jan 16 18:38:11 khem: is something specific I would need on kernel for it? Jan 16 18:38:32 otavio: I dont think so Jan 16 18:38:43 let me gaze at the code once more Jan 16 18:41:32 otavio: so what I think it failing for you is that its unable to detect /dev/shm Jan 16 18:42:01 the error you are getting ENOSYS is only returned if it have mountpoint.dir == NULL Jan 16 18:43:07 otavio: what does /proc/mounts say on the target Jan 16 18:43:12 khem: I see but why it should fail? Jan 16 18:43:33 what is /etc/fstab too ? Jan 16 18:43:42 khem: I am not with that working today but I can get all that info from it tomorrow Jan 16 18:43:50 khem: we have a self-written init Jan 16 18:43:52 oh j Jan 16 18:43:57 khem: it does the mounting for us Jan 16 18:43:59 hmmm Jan 16 18:44:13 so that seems suspicios Jan 16 18:44:19 khem: BUT IIRC it was setting it exactly as our hosting system Jan 16 18:44:29 hm Jan 16 18:44:38 let me grab the code doing it Jan 16 18:44:39 hold Jan 16 18:46:43 tmpfs /dev/shm tmpfs rw,relatime,mode=777 0 0 Jan 16 18:46:59 is that something /proc/mounts Jan 16 18:47:03 has ? Jan 16 18:48:42 khem: http://paste.debian.net/104787/ Jan 16 18:49:21 mkdir("/dev/shm", 1777); Jan 16 18:49:26 make it mkdir("/dev/shm", 0777); Jan 16 19:01:14 otavio: btw. Do you have CONFIG_SHMEM=y set in kernel .config ? Jan 16 19:01:53 * otavio checks Jan 16 19:02:11 khem: no Jan 16 19:02:37 and CONFIG_TMPFS=y Jan 16 19:02:37 BUT tmpfs wasn't support to provide it? Jan 16 19:02:53 khem: ah, thanks Jan 16 19:02:59 and CONFIG_DEVTMPFS Jan 16 19:03:00 seems that is what I need Jan 16 19:03:00 neither Jan 16 19:03:14 no Jan 16 19:03:28 khem: but I think tmpfs is "always in" today, no? Jan 16 19:03:38 otavio: try to enable them and make that change Jan 16 19:04:41 otavio: its configurable Jan 16 19:04:45 khem: both or just shmem? Jan 16 19:04:45 you need ot turn it on/off Jan 16 19:05:02 turn on all the options I mentioned Jan 16 19:05:13 then you can try to disable them Jan 16 19:05:16 khem: we use udev so we'd try to avoid using devtmpfs to avoid memory usage Jan 16 19:05:19 khem: ok Jan 16 19:05:49 k Jan 16 19:05:59 then you can disable CONFIG_DEVTMPFS Jan 16 19:06:03 to start Jan 16 19:08:28 ok Jan 16 19:08:58 khem: just to enlight me throught the learning process ... about sdk inside oe Jan 16 19:09:03 khem: any material for me to read? Jan 16 20:43:28 otavio: meta-toolchain is your sdk Jan 16 20:44:02 manual mentions it but is not very extentive Jan 16 21:39:08 Having some trouble building SHR for the iPhone3G. Pidgin won't compile properly. I tried cleaning and rebuilding separate, but no luck. I'm not really sure what to do. Anyone mind taking a look at my pastebin? http://pastebin.com/CSuqrMkb Jan 16 22:11:03 Lopi, try in #openmoko-cdevel Jan 16 22:11:10 there are more people there currently Jan 16 22:11:21 (at the time of writing) Jan 16 22:12:02 GNUtoo|laptop: I asked in there too Jan 16 22:14:42 ok Jan 16 22:22:46 khem, do you have a sheevaplug? Jan 16 22:22:54 mwester: hey Jan 16 22:22:57 mwester: no I dont Jan 16 22:23:03 Ah, ok. Jan 16 22:23:10 mwester: my device list is small Jan 16 22:23:20 I usually use qemus Jan 16 22:23:37 do u have some spare :) Jan 16 22:23:48 I intend to get slugos working properly on qemuarm as well... Jan 16 22:23:55 cool Jan 16 22:24:06 you gave me one nslu Jan 16 22:24:13 which I keep uptodate Jan 16 22:24:33 No spare - which is my problem. I've reflashed and changed everything on my Sheeva, and thought you had one and could tell me what the original partitioning of the flash was. Jan 16 22:24:44 and its courtesy of that device that slugos now is uptodate wrt toolchain bits Jan 16 22:25:10 mwester: I think eFfeM_work has one Jan 16 22:25:44 Your work is really appreciated, as is that of eFfeM_work -- it's great to be able to spend time on SlugOS itself rather than build problems. :) Jan 16 22:26:03 cool Jan 16 22:26:16 yeah now you should have less build issues Jan 16 22:26:22 Hardly any. Jan 16 22:26:43 khem: native-sdk-image is right what I neeed, thanks :) Jan 16 22:26:50 btw. we should revitalise slugosc Jan 16 22:26:58 Jay7: ok np Jan 16 22:27:00 only git, rsync and kexec-tools should be added Jan 16 22:27:08 a few troubles with less-common parts of teh package feeds, none with the core SlugOS. The only issue I've not bothered with is the external toolchain/SDK. Jan 16 22:27:57 mwester: btw. meta-toolchain-slugos builds fine Jan 16 22:28:32 Not sure about slugosc -- openwrt seems to fill that void. But it might be a good exercise to make it work anyway. But I'm working on Sheevaplug support in SlugOS right now, since someone has expressed interest in that. Jan 16 22:28:55 does openwrt use OE Jan 16 22:28:58 I think not Jan 16 22:28:58 * mwester makes a note to check meta-toolchain-slugos again to see if the libc conflict has mysteriously fixed itself. Jan 16 22:29:09 No, openwrt has its own build system. Jan 16 22:29:14 and feeds Jan 16 22:29:35 meta-toolchain-slugos built fine last week Jan 16 22:30:07 Ok. I disabled it on my autobuilder a few weeks ago after I was unable to figure out the conflict. I'll turn it back on again and see what happens. Jan 16 22:31:07 khem: hi, got a second? Jan 16 22:32:07 RP: hi Jan 16 22:32:20 hm.. but native-sdk-image doesn't understand ANGSTROM_EXTRA_INSTALL :( Jan 16 22:32:35 ANGSTROM_EXTRA_INSTALL is angstrom specific Jan 16 22:32:45 khem: We've been experimenting with libtool 2.4 in poky and saw a lot of RPATH issues - as far as we can tell they're in OE also? Jan 16 22:32:50 while native-sdk-image is not Jan 16 22:33:13 RP: hmmm I dont know if they are in OE as well Jan 16 22:33:34 hm.. but should according to require console-image.bb Jan 16 22:33:35 khem: zenlinux_ did some tests which suggest they are :( Jan 16 22:33:37 RP: I have QA enabled im my builds if they were there then it should have caught them Jan 16 22:33:49 khem: angstrom 2010? Jan 16 22:33:50 khem: what is right way to add some packages to image then? Jan 16 22:33:59 ok it would be nice to know which packages Jan 16 22:34:05 DISTRO_EXTRA_INSTALL? Jan 16 22:34:10 khem: udev or libpng show them Jan 16 22:34:31 khem: I mailed the libtool list with some questions about it Jan 16 22:34:32 RP: hmmm ok Jan 16 22:34:59 khem: Interestingly poky is no longer doing a relink of the libs at install time with 2.4 Jan 16 22:36:44 Jay7: just checked my dev syste m; I'm using DISTRO_EXTRA_RDEPENDS += "kexec-tools" to pull in kexec user utilities into SlugOS. Not sure if that works in your distro, though. Jan 16 22:36:49 RP: there were few other changes I did where feeding -L was removed Jan 16 22:37:08 mwester: I fear all this is distro-specific Jan 16 22:37:17 yeah, could be. Jan 16 22:37:20 well.. will have some experiments then Jan 16 22:37:30 I'm using minimal this time Jan 16 22:37:34 Jay7: IMAGE_INSTALL_append_pn-native-sdk-image Jan 16 22:37:36 is what u need Jan 16 22:37:40 khem: I think poky did that years ago ;-) Jan 16 22:37:48 could be Jan 16 22:38:42 khem: Anyhow, I just wondered whether you'd seen it or had any thoughts about it. As far as I know the issue is there if the relink step doesn't happen. What governs that I haven't worked out yet Jan 16 22:39:16 RP: udevadm and udevd does not show rpath problem on my rfs Jan 16 22:39:36 khem: is *_pn-* literally 'pn'? or should be replaced by something? Jan 16 22:39:40 yes there was some hackery in classes Jan 16 22:39:43 which I removed Jan 16 22:39:52 Jay7: literally Jan 16 22:39:57 ok, thanks Jan 16 22:40:06 khem: Are they being relinked at do_install? Jan 16 22:41:40 RP: I changed autotools_prepackage_lamangler Jan 16 22:42:44 RP: autotools_prepackage_lamangler is not needed with 2.4 Jan 16 22:42:54 that can cause the relinking mess Jan 16 22:43:49 if zenlinux_ can post the problem details he saw on angstrom 2010 then I can take a deeper look Jan 16 22:44:02 khem: Right, and I'm pretty sure zenlinux_ brought those changes in. So you're confirming that a) You're not seeing a relink at do_install time and b) objdump isn't showing a bad RPATH in your binaries/libs? Jan 16 22:44:20 lemme check libs Jan 16 22:44:25 I only checked bins Jan 16 22:44:59 libs have rpaths, too? didn't know that... Jan 16 22:45:17 mwester: they do indeed Jan 16 22:46:00 so the lib's rpath is used to find the lib's dependencies? In addition to the exe's rpath, or exclusively? Jan 16 22:46:31 mwester: that I don't know Jan 16 22:46:44 I shall google that. Jan 16 22:46:50 I'm intruiged. Jan 16 22:46:53 RP: no libs seem ok too Jan 16 22:47:02 what options do u use for objdump Jan 16 22:47:07 I am usin readelf Jan 16 22:50:45 cool Jan 16 22:51:00 now I have image to do quick tests of kexecboot Jan 16 22:51:34 khem: I tend to be lazy and use objdump -x | grep RPATH ;-) Jan 16 22:52:37 03Khem Raj  07master * r59ff7a666c 10openembedded.git/conf/distro/ucslugc.conf: Jan 16 22:52:37 ucslugc.conf: Bring it forward to use sane-toolchain Jan 16 22:52:37 Signed-off-by: Khem Raj Jan 16 22:53:54 RP: OK used same options here there is no RPATH in bins or libs Jan 16 22:54:35 khem: Hmm, I wonder what the difference is :/ Jan 16 22:56:05 RP: dont know, oh I use DISTRO=minimal Jan 16 22:56:12 and the binaries are for arm Jan 16 22:56:16 JFYI Jan 16 22:56:41 this webkit is a hog Jan 16 22:57:41 1 G of objects thats insane Jan 16 22:58:41 khem: it is a bit crazy, the -dbg package is huge Jan 16 23:04:20 it shows that everything can be abused Jan 16 23:04:26 look at the templates Jan 16 23:04:34 in webkit its crazy Jan 16 23:04:42 they are merciless on compiler Jan 16 23:04:59 and inturn compiler is merciless on disk and user :) Jan 16 23:05:49 now people talk about compiling in an emulated env Jan 16 23:05:55 in the past i noted that webkit links 3 times. not sure if this is still the case Jan 16 23:06:04 webkit devs need to be issued Pentium-III systems to build on. That'll fix it. :D Jan 16 23:06:12 I would say compile webkit in qemu and come back andtell me how you feel about it Jan 16 23:07:11 grg: yeah it takes 25-40 mins on my T61 laptop to link Jan 16 23:07:18 qt Jan 16 23:07:22 err webkit Jan 16 23:07:26 ah.. I've remember this too Jan 16 23:07:45 I was wondered Jan 16 23:07:52 mwester: some folks know C++ and thats it Jan 16 23:08:00 they dont know aftermaths of it Jan 16 23:08:03 Java. Jan 16 23:08:15 Java -- this generation's COBOL. Jan 16 23:08:24 heh right Jan 16 23:08:34 btw, what is status of fennec? Jan 16 23:08:35 one reason why android is successful Jan 16 23:08:36 dead? Jan 16 23:08:52 fennec is still there I think Jan 16 23:09:08 it was always alpha-quality iirc Jan 16 23:09:22 Yeah, android is successful, but look at the type of hardware it needs -- my HTC Desire is a 1GHz ARM processor... for a cell phone! Jan 16 23:09:31 fennec requires 128mb ram. Presumably they are waiting for that to be the lowest common denominator for all handheld devices Jan 16 23:09:37 Think of what one could do with C or C++... Jan 16 23:09:40 here in bay area if a restaurant loses out and gets bad reviews and loses business they just rename it Jan 16 23:09:40 instead of improving its mem usage Jan 16 23:09:46 staff is still same Jan 16 23:09:50 :) Jan 16 23:09:53 hehe Jan 16 23:10:11 grg: does webkit working with smaller RAM? Jan 16 23:10:23 Jay7, only for small pages Jan 16 23:11:29 hm.. good idea to render pages on some fast server then send it to device :) Jan 16 23:11:58 thats what google wants Jan 16 23:12:15 it should be easily doable Jan 16 23:12:20 and cloud computing Jan 16 23:12:25 heh! VNC to remote device from central webkit web browser. Jan 16 23:12:27 user input is question Jan 16 23:12:28 where's is google's "no i don't want your fucking ajax in my web search results" option anyway? Jan 16 23:12:49 just because a browser has js and its enabled doesnt mean i want to use it Jan 16 23:12:57 grg: you want! they know better! :) Jan 16 23:13:21 i dont really need a preview of the entire page in the results Jan 16 23:14:01 yeah.. time to start new search service ;) Jan 16 23:14:03 They're making room for a new search service. We just need to find it. Jan 16 23:14:20 Like opendns -- opensearch? Jan 16 23:15:16 grg: http://patches.openembedded.org/patch/116/ Jan 16 23:15:25 did u look further into it Jan 16 23:16:10 khem, busybox-mdev worked ok when i enabled it in an -image.bb Jan 16 23:16:29 but i don't know about if changing distro policies is desired Jan 16 23:23:03 grg: mdev makes sense for micro Jan 16 23:23:12 and even for slugos I think Jan 16 23:23:27 so you should propose a patch for micro atleast Jan 16 23:23:45 ok. i'll do some more testing and send a patch sometime this week. Jan 16 23:23:53 am busy with some non oe stuff right now Jan 16 23:24:13 sure np Jan 16 23:36:21 I'd like to be able to support mdev for SlugOS. Would be excellent if we could actually support "opkg install udev" (with required reboot, of course), and have that actually work. Jan 16 23:36:51 But failing that, just having an mdev-based SlugOS variant would be useful, esp. with uclibc. Jan 16 23:38:35 mwester: yeah I just revived uclugc Jan 16 23:38:59 FYI, I don't think it has been built or tested for years, and it has never been released... Jan 16 23:39:11 So I don' t know how well it ever worked. Jan 16 23:48:09 mwester: doesnt matter there always is a beginning : Jan 16 23:48:16 :) Jan 16 23:48:24 mwester: I am sure it should build something Jan 16 23:48:27 may be not Jan 16 23:48:43 I prolly have to add config for uclibc for armv5teb Jan 16 23:48:49 which is simple Jan 17 00:09:30 mwester: I suppose ./meta/slugos-packages.bb nd ./tasks/task-slugos-extd.bb could have an issue with kexecc-tools Jan 17 00:09:52 how so? Jan 17 00:09:55 the recipe splits in 2: kexec and kdump Jan 17 00:10:01 since 2 years... Jan 17 00:10:17 Ah. I don't use kdump, just kexec. Jan 17 00:10:25 well, anyhow you'd refer to kexec and kdump packages Jan 17 00:10:39 yes, that was the reason: skip beefy kdump Jan 17 00:10:57 I suppose we should put kdump in the package feeds, though - that would be useful. Jan 17 00:11:06 Jay7 noticed that in task-devimage Jan 17 00:11:14 andd suggested to grep ;) Jan 17 00:11:47 building kexec-tools packages both. you can install each or both Jan 17 00:11:56 Thanks! I'll make a note to add to teh packages (I think we don't want kdump in the extended image, though). Jan 17 00:12:35 So I should RDEPEND on just "kexec"? Jan 17 00:12:48 yes Jan 17 00:12:52 Cool. Jan 17 00:13:06 e.g IMAGE_INSTALL = "kexecboot-klibc" (small version) Jan 17 00:13:21 err.. kexec-klibc Jan 17 00:14:09 What's in that package though? All I wish to do is to be able to use "kexec" to boot a different kernel -- do I need klibc? Jan 17 00:15:19 kexec-tools recipe PACKAGES =+ "kexec kdump" Jan 17 00:15:51 so you just need 'kexec' Jan 17 00:16:03 sorry for the confusion Jan 17 00:16:11 klibc is a variant of the same soup Jan 17 00:16:27 just a bit thick Jan 17 00:18:13 Actually, according to my notes here, I see that kexec doesn't work on the current preferred kernel for teh NSLU2... so I guess it's rather a moot point until I patch the kernel. Jan 17 00:18:57 * mwester notes that this kernel has been in use for two years, but no users have ever complained... so probably nobody uses kexec except himself. Jan 17 00:20:26 :/ Jan 17 00:22:56 Yeah, not enough developers with SlugOS. And the ones that show up seem to leave as soon as they get their specific software package built... Jan 17 00:23:52 I'd like to blame OE for being too steep a learning curve, but I don't think that's really a problem. Jan 17 00:29:50 mwester: general problem I think Jan 17 00:44:45 hmmm.. Jan 17 00:45:16 autom4te complains about wrong m4 version when running inside qemu + native-sdk-image Jan 17 00:45:28 path to m4 is build path Jan 17 00:45:38 sysroot even Jan 17 00:46:51 autom4te: need GNU m4 1.4 or later: /var/tmp/oe/minimal/arm/sysroots/x86_64-linux/usr/bin/m4 Jan 17 00:47:22 called from aclocal which is called by autoreconf Jan 17 00:47:31 khem: can you have a look? Jan 17 00:49:21 yeah.. Jan 17 00:49:50 eval 'case $# in 0) exec /var/tmp/oe/minimal/arm/sysroots/x86_64-linux/usr/bin/perl -S "$0";; *) exec /var/tmp/oe/minimal/arm/sysroots/x86_64-linux/usr/bin/perl -S "$0" "$@";; esac' Jan 17 00:50:13 this is from head of /usr/bin/autom4te Jan 17 00:50:19 on target Jan 17 00:50:54 my $m4 = $ENV{"M4"} || '/var/tmp/oe/minimal/arm/sysroots/x86_64-linux/usr/bin/m4'; Jan 17 00:50:56 from there too Jan 17 00:52:42 /usr/bin/autoheader is broken same way Jan 17 00:52:58 even worst, wrong path to perl in shebang Jan 17 00:58:22 and all fails on can't locate Autom4te/ChannelDefs.pm in @INC... Jan 17 00:58:53 gnu-configize line 40 Jan 17 00:59:17 Jay7: yeah some other people have seen it too Jan 17 00:59:24 I thought it was fixed by trini Jan 17 00:59:32 Tartarus I mean Jan 17 00:59:50 my $datadir = $ENV{'autom4te_perllibdir'} || '@autom4te_perllibdir@'; Jan 17 00:59:52 # '/home/kergoth/code/build-arm/tmp/staging/share/autoconf'; Jan 17 01:00:06 hehe.. seems kergoth was done something with that code ;) Jan 17 01:00:12 hah Jan 17 01:00:33 perl must be falling in love with kergoth Jan 17 01:00:37 noooo Jan 17 01:01:21 there are other alike comments Jan 17 01:01:38 well.. too late here Jan 17 01:03:22 I'll raise this up tomorrow Jan 17 01:03:26 good night all Jan 17 01:03:36 gn Jan 17 02:47:37 anyone here interested in running linux on Adtran and/or Cisco routers? **** ENDING LOGGING AT Mon Jan 17 02:59:57 2011