**** BEGIN LOGGING AT Wed Nov 29 02:59:58 2006 **** BEGIN LOGGING AT Wed Nov 29 04:12:42 2006 Nov 29 08:05:07 yo Nov 29 08:07:35 koen: angstrom finally built. now I need to adjust the kernel so it boots. Nov 29 08:07:47 that's good news Nov 29 08:07:47 hey Genesis Nov 29 08:09:55 yo koen Nov 29 08:58:25 koen: what does init=... have to be for angstrom? Nov 29 08:59:35 /sbin/init Nov 29 09:00:01 thx Nov 29 09:00:33 * rwhitby loves how quickly Apex rebuilds ... Nov 29 09:00:42 good morning Nov 29 09:03:10 hey pH5 Nov 29 09:03:15 koen: have you booted an angstrom ixp4xx image from the kernel, or just chrooted into it? Nov 29 09:04:14 still waiting for the dd-wrt shop to send the POE inserter Nov 29 09:05:20 the angstrom jffs2 rootfs I've built just doesn't want to boot. It mounts as root fine, but init doesn't take over from the kernel for some reason. Nov 29 09:05:38 does apex like symlinks? Nov 29 09:05:53 the kernel boots fine, so it's not apex Nov 29 09:06:08 /sbin/init is a symlink to /sbin/init.sysvinit Nov 29 09:06:08 I get to "Freeing init memory: 100K Nov 29 09:06:08 " and then stop Nov 29 09:06:30 yeah, but the kernel is interpreting the init= cmdline argument, not apex. Nov 29 09:08:30 ah, right Nov 29 09:09:45 * rwhitby wonders how to debug this ... Nov 29 09:09:46 rwhitby: check you have glibc in your image Nov 29 09:10:13 rwhitby: that was the problem I had yesterday, turns out package.bbclass has an ancient bug and calls the wrong objdump Nov 29 09:10:23 I have /lib/libc-2.5.so Nov 29 09:10:36 and libc.so.6 symlinked to it Nov 29 09:10:44 rwhitby: so not that problem then Nov 29 09:11:00 yeah, I built this image from scratch in the last 12 hours. Nov 29 09:13:19 my cmdline is "root=/dev/mtdblock2 rootfstype=jffs2 init=/sbin/init.sysvin Nov 29 09:13:19 it console=ttyS0,115200 rtc-pcf8563.probe=0,0x51 noirqdebug" Nov 29 09:13:38 (this is on an NAS100d) Nov 29 09:15:56 ok, I better try with a slugos rootfs to make sure that boots fine. Nov 29 09:16:22 damnations, fontconfig still doesnt generate shlibs correctly Nov 29 09:28:38 XorA: grrr fontconfig ;) Nov 29 09:29:27 ok, slugos rootfs boots with the same apex and kernel Nov 29 09:30:07 I guess the next test is to chroot into the angstrom rootfs to see if that works ... Nov 29 09:30:10 rwhitby: can you use an nfsroot of the fs Nov 29 09:30:36 ade|desk: not easily, but yes that will be a last resort measure Nov 29 09:31:06 I can rsync tmp/rootfs over to the slugos session and chroot into it Nov 29 09:31:30 thats useful Nov 29 09:31:51 quite handy having rsync running Nov 29 09:33:59 slugos retains all the settings like network setting and ssh authorized_keys too, so it works as soon as I flash the slugos rootfs and ipkg install rsync Nov 29 09:34:42 * XorA wishes he could locate a nas100d Nov 29 09:38:15 hmm: Nov 29 09:38:19 root@nas100d:/$ chroot /media/sda3 /bin/sh Nov 29 09:38:20 Illegal instruction Nov 29 09:38:43 is that with the eabi kernel? Nov 29 09:39:06 kernel is fine - a "chroot / /bin/sh" works Nov 29 09:39:42 I'll reflash the kernel from the build today, just to be sure. Nov 29 09:41:44 koen: you're thinking it could be an OABI kernel trying to run EABI rootfs? Nov 29 09:42:16 yes, otherwise you shouldn't get a SIGILL Nov 29 09:44:26 Hmm - I guess an EABI kernel doesn't run an OABI rootfs, right? Nov 29 09:44:46 it does Nov 29 09:45:06 BUG: warning at lib/kref.c:32/kref_get() Nov 29 09:45:06 Unable to handle kernel NULL pointer dereference at virtual address 00000000 Nov 29 09:45:10 OE sets CONFIG_OABI_COMPAT=y Nov 29 09:45:11 not here :-) Nov 29 09:46:20 ok, it's going to take a while to upload the angstrom rootfs again - it's via serial cause the nas100d redboot doesn't do tftp :-( Nov 29 09:47:17 yep, CONFIG_OABI_COMPAT is set on the kernel I built. Nov 29 10:04:37 ok, angstrom rootfs is struggling to boot now Nov 29 10:05:19 looks like it was an OABI kernel to blame for not booting from rootfs. Nov 29 10:09:12 http://pastebin.com/835216 (pastebin.ca isn't responding for some reason) Nov 29 10:10:13 koen: so it does boot to the shell prompt. no networking of course, but that's all to come. Nov 29 10:10:46 * rwhitby is too used to slugos sysconf script restoring all network settings after reflashing. Nov 29 10:16:58 right Nov 29 10:17:16 the images include a splash screen by default Nov 29 10:17:28 I think I sent a mail to OE devel about that :) Nov 29 10:18:06 it still being added for machines without screen feature? Nov 29 10:19:27 yes Nov 29 10:20:36 ipkg update and ipkg list work after networking is brought up Nov 29 10:21:05 is there a tty0 feature I can turn off :-) Nov 29 10:23:13 ok, so that's enough of a proof of concept for me to say that SlugOS 4 will be the last version as it's own separate distro, and SlugOS 5 will probably be based on Angstrom. Nov 29 10:23:42 based or part of? Nov 29 10:23:50 good question Nov 29 10:24:08 what does "part of" mean in practice? Nov 29 10:24:35 thats good question.. koen? Nov 29 10:24:46 ehm.... Nov 29 10:25:00 write an RFC :) Nov 29 10:25:04 (i.e. we don't want to reinvent any wheels, and we do want to stay close enough to directly benefit from, and contribute to, the main Angstrom development) Nov 29 10:26:00 I thiink that we can join work and produce one distro for all supported machines Nov 29 10:26:17 e.g. will the Angstrom introduction end "for embedded devices like handhelds and set top boxes and network-attached storage devices." ? Nov 29 10:26:19 I think it would mean that the former slugos people work on an images for slug like machines Nov 29 10:26:42 one Angstrom instead of Angstrom 2007.X & slugos Y & other Z Nov 29 10:27:10 who are the Angstrom core team? Nov 29 10:27:17 (at the moment) Nov 29 10:28:20 "offically" mickeyl and me, unofficially anyone who wants to help Nov 29 10:28:55 where does the "set top boxes" bit come in ? Nov 29 10:29:12 dreambox, neuros, etc Nov 29 10:29:32 and you have core developers supporting angstrom on them? Nov 29 10:29:44 nope Nov 29 10:30:06 ok, so the NAS devices would be the first real non-handhelds in Angstrom. Nov 29 10:30:39 yes Nov 29 10:32:04 and koen and mickeyl intend running angstrom on their NAS/networking devices (Loft for koen, NSLU2 for mickeyl)? Nov 29 10:32:32 at least I plan to Nov 29 10:32:48 I will use angstrom on my progear (webpad on transmeta ts3200 (x86)) Nov 29 10:32:51 and secretly looking for a thecus n2100 Nov 29 10:33:13 koen: easy to get in the uk Nov 29 10:33:23 the big question is about BE support for EABI Nov 29 10:33:46 (unless everyone is just going to take the networking speed hit and run everything LE) Nov 29 10:34:07 rwhitby: I suspect it will "just work", but you'll need to patch a few recipes to accept armeb-linuxgnueabi as a valid target Nov 29 10:34:35 jacques though that armeb eabi was nonexistent at the moment. Nov 29 10:34:51 uclibc/EABI will work with upcoming .29 release Nov 29 10:34:52 or is it just untried? Nov 29 10:35:34 I don't think anyone has tried to build BE EABI Nov 29 10:36:27 I think I can persuade lennert to fix bugs once it builds Nov 29 10:41:02 right - will he be running armeb eabi angstrom on a nas device? or just fixing the toolchain? Nov 29 10:49:31 fixing the toolchain at first Nov 29 10:50:56 is there an angstrom mailing list separate from the oe mailing list (which I still haven't subscribed at the new location yet) ? Nov 29 10:51:19 yes: http://www.angstrom-distribution.org/contact Nov 29 10:51:47 rwhitby: you will find http://lists.linuxtogo.org/pipermail/angstrom-distro-devel/2006-November/000017.html interesting Nov 29 10:51:56 do I need to subscribe to both, or just angstrom now? Nov 29 10:52:07 only -devel Nov 29 10:53:21 I mean both angstrom and OE Nov 29 10:54:23 being subscribed to the OE list is pretty much mandatory as OE dev/user Nov 29 10:54:52 koen: I assume ipkg is dieing, hence your wish to look at dpkg Nov 29 10:55:24 ipkg isn't actively maintained Nov 29 10:55:51 but ipkg is a much for devices with a small rootfs Nov 29 10:56:15 ok, I'm subscribed to both now Nov 29 10:57:11 * XorA really should see about rpm building in oe one day Nov 29 10:57:27 XorA: Really Poor Maintained thing? Nov 29 10:57:41 XorA: rpm support is in the bugtracker Nov 29 11:03:45 Is there an Angstrom wiki? Nov 29 11:04:02 hrw|work: better that Damnnearimpossibletomaintain pkg :-) Nov 29 11:04:28 ;D Nov 29 11:05:10 but OE makes both a doddle Nov 29 11:05:27 god I need a fatter pr0n pipe Nov 29 11:05:57 I need working monotone 0.35 Nov 29 11:06:11 which will give me informations instead of noise Nov 29 11:06:18 mtn: warning: orphaned node conflict on node 18,372, dead parent 18,331, name files Nov 29 11:06:21 mtn: warning: resolve non-content conflicts and then try again. Nov 29 11:06:45 lardman: check the frontpage of angstrom-distribution.org Nov 29 11:06:49 it could be in suahili for >90% users too Nov 29 11:11:48 koen: It says wiki, but I can't see a link. I'll look again Nov 29 11:16:09 koen: I was thinking that it might be useful to have some info for would-be 'user' testers about the status of the various images (after seeing XorA's post about Xfbdev locking machines at boot) Nov 29 11:17:34 lardman: http://www.angstrom-distribution.org/ Nov 29 11:17:48 lardman: the first three occurences of 'wiki' are a link to the wiki Nov 29 11:18:28 lol, my 'followed link' colour is about the same as the text, must fix that Nov 29 11:18:36 I must fix that that is Nov 29 11:18:49 koen: can you change my newly created 'RodWhitby' account on the wiki to 'rwhitby'? Nov 29 11:18:54 koen: you must fix the css so that bullet lists are inside the boxes, not outside Nov 29 11:19:09 s/the wiki/www.angstrom-distribution.org/ Nov 29 11:20:15 rwhitby: I don't have 'root' on the wiki Nov 29 11:20:34 * koen doesn't even know if moin supports user renaming Nov 29 11:20:51 oh, looks like I wasn't registering for the wiki then anyway. looks like I was registering for drupal. Nov 29 16:40:17 ~lart WinNT BSOD Nov 29 16:40:17 * cdbot duct-tapes WinNT BSOD to the floor and drools on him Nov 29 16:40:50 What I was going to say, was is it worth putting something like a status page in the wiki, assuming we want more testing of the various Zaurus images Nov 29 16:41:42 someone need to share info which platforms are built, which of them will be supported Nov 29 16:42:01 lardman: you like for tracking bugs Nov 29 16:42:04 request testers, machine mentors Nov 29 16:42:11 lardman: isn't a bugtracker better for tracking bugs? Nov 29 16:42:32 koen: ;) depends on who's testing of course Nov 29 16:43:19 not really Nov 29 16:43:26 check rule #5 Nov 29 16:43:34 It might just be nice to see a verbal description of status Nov 29 16:43:41 yes, I saw the rules Nov 29 16:44:40 This is all going back to the question of whether to do another OZ release, and if not, whether to point people at the Angstrom builds Nov 29 16:45:25 and because my mail about OZ future was like it was, no one answered to it Nov 29 16:46:22 I see no problem with having verbal descriptions, but I do have big problems with people using said wiki pages as a bugtracker Nov 29 16:46:26 hrw|work: Sorry, been busy, didn't actually see your post Nov 29 16:46:53 koen: Yes, agreed, it would just be nice to see a list of showstoppers, etc. Nov 29 16:46:58 http://article.gmane.org/gmane.comp.handhelds.openzaurus.devel/1320 Nov 29 16:47:21 showstoppers: http://bugs.openembedded.org/showdependencytree.cgi?id=1573 Nov 29 16:47:51 * lardman thinks it was a mistake not to ask for a digest of OpenMoko & Maemo, too much traffic Nov 29 16:52:46 you can turn them into a digest while being subscriber Nov 29 16:52:52 subscribed* Nov 29 16:53:07 koen: Yep, just about to do that Nov 29 16:53:10 koen: thanks Nov 29 16:53:37 * lardman trying to read through anything interesting in inbox since he left office this morning Nov 29 16:53:54 mailinglists don't hit my inbox Nov 29 16:54:01 procmail filters them to subdirs Nov 29 16:54:18 and I read most mailinglists via gmane anyway Nov 29 16:54:24 The wonders of limited mailbox space and using Windows Nov 29 16:54:48 koen: Your deptree link timed out for me Nov 29 16:55:37 Ah, I now see why I didn't get hrw's post - I've falled between the cracks with the ml move Nov 29 16:55:43 fallen even Nov 29 16:59:40 mencoder opie/multimedia Open Source multimedia player. Nov 29 16:59:41 ;D Nov 29 16:59:49 angstrom supports opie? :D Nov 29 18:22:56 morning Nov 29 18:25:04 hey rwhitby Nov 29 18:27:55 morning rwhitby Nov 29 18:28:11 hey koen, hrw|work Nov 29 18:33:19 hi koen. The plsplash interupts opie when opie start. The blining cursor (form the shell) is blinking through opie screen, and when it writes some messages it totaly flush the screen. I have to restart Opie to get rid of it. What i have done til now is to disable psplash, but now it has arivved again. Nov 29 18:33:33 I have also moved away "screen" form my machine file Nov 29 18:33:45 Is this i known bug? Nov 29 18:35:48 angstrom doesn't support opie Nov 29 18:36:29 koen: Not even if opie is maintained in OE? Nov 29 18:36:47 only if people want to maintain it in angstrom Nov 29 18:37:25 ok, i thought is was supported since i saw you had moved Opie over to task-base, wich works great BTW Nov 29 18:37:48 ? Nov 29 18:38:08 goxboxlive: I moved task-opie to 'task-base' format Nov 29 18:38:32 Ok, so it was you, sorry. thx, hrw, it works even bether now. Nov 29 18:38:53 It doesnt install hardware apps i dont have anymore. Like pcmcia Nov 29 18:39:51 Theonly hack i have to do now to get Opie work allright is to remove the psplash from installing to image. Nov 29 18:40:18 goxboxlive: or fix this Nov 29 18:42:14 I saw todays changes in the anstrom.conf that psplash where become dependent of the content in the machine file. So i removed "screen" from my machine file and start building, but it made the psplash anyway. Nov 29 18:42:24 thats workaround Nov 29 18:43:15 you need to check when psplash end his work and adapt opie image for it Nov 29 18:43:29 Actually i dont want psplash, beacuse i miss the bootup messages ( i know dmesg) if something goes wrong during bootup. Nov 29 18:44:05 goxboxlive: my home machine does not show any boot messages and does not use splash Nov 29 18:45:16 Well your homemachine has working drivers, we are still developing drivers for it. :-) Thats why i want it. When the Universal hits the 99% score we dont need it either. Nov 29 18:46:23 my home machine oops each time I plug bt dongle Nov 29 18:47:17 I see your point hrw. But i think you se mine to Nov 29 18:47:38 sure Nov 29 19:59:12 #LFD Nov 29 19:59:18 ops **** ENDING LOGGING AT Thu Nov 30 02:59:57 2006