**** BEGIN LOGGING AT Mon Dec 01 02:59:58 2008 Dec 01 14:47:45 hey, anyone around? Dec 01 14:48:15 yup Dec 01 14:48:19 have a question? Dec 01 14:48:34 right... I'm currently trying to compile angstrom for the h3600... the config files are there. Dec 01 14:48:44 yet, bitbake fails at some point, I'm wondering if you could help. Dec 01 14:48:50 (1 sec, gonna pastebin) Dec 01 14:49:00 no promises, but I'll see if I have an idea Dec 01 14:49:34 yeah, it's no real build error, just not finding some files and got some problems with git Dec 01 14:49:51 btw, is it normal that my temp dir is like 6gb big? Dec 01 14:50:17 yeah Dec 01 14:50:31 OE builds a whole cross compiling environment Dec 01 14:51:03 yeah ok :) Dec 01 14:51:15 I'll have that paste in a minute, bb needs to rebuild the cache Dec 01 14:52:37 oh and, will it pose a problem if I use the arm bootloader from handhelds.org? Dec 01 14:53:05 I'm afraid I don't know much specifically about the ipaqs Dec 01 14:53:14 ok, then nvm, I'll just try. Dec 01 14:53:28 rebuilding takes ages :S Dec 01 14:54:12 right ... there we go Dec 01 14:54:39 you can use it interactively to avoiding having to rebuild cache all the time Dec 01 14:54:49 http://rafb.net/p/OwUd6H14.html Dec 01 14:54:50 ok Dec 01 14:54:57 it was just some "truncated cache" problem anyway Dec 01 14:55:15 it's a 404 not found Dec 01 14:55:22 the file is not present anyway Dec 01 14:55:30 but I'm wondering how I could solve this problem Dec 01 14:57:14 nvm, just downloaded the page from some other page. Dec 01 14:57:15 :D Dec 01 14:57:23 Thank you for your attention though Dec 01 14:57:27 ok, sounds good :) Dec 01 14:57:43 if you get really stuck ask me and I'll try to be a little more helpful :D Dec 01 14:57:46 ok :D# Dec 01 14:58:06 might take some time though... task 575 of 800 and afterwards I think I'm still missing the other 2 images... Dec 01 14:58:28 oh, how big does the angstrom env dir normally get? Dec 01 14:58:34 more than 10gb? Dec 01 14:58:50 if you're only building for one machine it will probably stay under 10GB I would think Dec 01 14:58:53 ok Dec 01 14:58:57 I guess it depends on what image you're building though Dec 01 14:59:10 sure, though as the ipaq here is quite old I don't expect it to be that big. Dec 01 14:59:33 It really is a self contained cross-compile env, which then builds a linux distribution from source Dec 01 14:59:36 yeah Dec 01 14:59:59 and cache doesn't get cleared as you go Dec 01 15:00:19 ccache? Dec 01 15:00:48 I just mean, the unpacked source directories stay around until you delete them Dec 01 15:00:54 ah yeah. Dec 01 15:00:56 and if you're using ccache of course that builds up too Dec 01 15:01:06 sure. Dec 01 15:01:10 though I disabled ccache Dec 01 15:06:56 great... perl-native fails... Dec 01 15:07:27 will it pose a problem if when I run make depend myself? Dec 01 15:08:30 as a workaround or as a troubleshooting step followed by cleaning it up? Dec 01 15:08:40 workaround Dec 01 15:08:49 it might, it might not Dec 01 15:08:52 what's the error? Dec 01 15:09:04 1 sec Dec 01 15:10:25 http://rafb.net/p/Xl1GYR84.html Dec 01 15:10:36 :D Dec 01 15:10:58 huh Dec 01 15:11:03 I wonder how that happened Dec 01 15:11:22 i think because I deleted the dir... Dec 01 15:11:31 :P Dec 01 15:11:36 but anyway... Dec 01 15:11:39 well, yeah, uhm, that would do it Dec 01 15:11:53 I just put it back there but it'll fail anyway Dec 01 15:11:57 just give me some proper error ... Dec 01 15:11:58 1 sec. Dec 01 15:13:21 right... Dec 01 15:13:26 I shouldn't have deleted that file... Dec 01 15:14:25 how can I reset a bb task? Dec 01 15:15:00 bitbake -c clean perl-native Dec 01 15:15:04 ok Dec 01 15:16:48 there we go... i love errors ;) Dec 01 15:17:08 http://rafb.net/p/rYtXwE28.html Dec 01 15:17:12 that's the one :) Dec 01 15:17:13 because they give you information to fix problems> Dec 01 15:17:45 if I were a betting man I would say it has to do with the lack of clist and shlist installed on your host Dec 01 15:17:51 though I don't know what those are :) Dec 01 15:18:22 I'd say so too... but those aren't any "real" tools... Dec 01 15:18:30 hmm, unless it was some kind of failed substitution Dec 01 15:19:21 out of curiosity, are you on an ubuntu or debian machine with sh linked to dash? Dec 01 15:19:41 gentoo actually... but those shlist and clist thingies are created using the perl makefiles Dec 01 15:19:50 I just did that... let's see if whether it works now :) Dec 01 15:26:19 something here appears to be terribly fucked up :/ Dec 01 15:27:02 that's the feeling I got Dec 01 15:27:06 yup. Dec 01 15:27:10 what *is* sh linked to? Dec 01 15:27:13 but it must be the bitbuild file Dec 01 15:27:19 bash Dec 01 15:27:26 the clist and shlist files Dec 01 15:27:29 should / are generated by make. Dec 01 15:27:47 at least .clist and .shlist are Dec 01 15:27:53 might be that they are just missing a make there... Dec 01 15:27:57 weird... why does it compile now o.O Dec 01 15:28:02 ah no, fail... Dec 01 15:30:28 right... remind me that the $MAKE env var should be set all the time ;) Dec 01 15:30:39 eeh no, next error -.- Dec 01 15:36:10 it's compiling now :) Dec 01 15:40:22 johnx, should I run bitbake as root? Dec 01 15:40:27 no Dec 01 15:40:30 never Dec 01 15:40:33 ok Dec 01 15:40:49 well, perl compile worked now Dec 01 15:40:56 now it's having problems with do_populate_staging Dec 01 15:41:11 permissions related? Dec 01 15:41:14 yup Dec 01 15:41:22 execvp lib/re.pm: Permission denied. Dec 01 15:41:27 sth like that Dec 01 15:42:04 well, is it eXecutable? Dec 01 15:42:04 the command it's executing is... Dec 01 15:42:06 nope. Dec 01 15:42:13 but on purpose. Dec 01 15:42:16 ah Dec 01 15:42:24 http://rafb.net/p/2skwub26.html Dec 01 15:42:44 ah well... Dec 01 15:49:20 w00t! perl finally compiled & installed :D Dec 01 15:49:39 what was the trick with that? Dec 01 15:49:54 editing makefiles by hand, executing some commands byh and : Dec 01 15:51:02 why is it compiling gnome though? Dec 01 15:51:08 or is gnome-common != gnome? Dec 01 15:51:19 well, what image are you building? Dec 01 15:51:21 gpe? Dec 01 15:51:28 h3600 ... opie Dec 01 15:51:30 at least... Dec 01 15:51:36 huh, well that is interesting Dec 01 15:51:43 shouldn't even build gtk in that case Dec 01 15:51:46 http://rafb.net/p/0Xzj1l71.html Dec 01 15:51:50 that's my local.conf Dec 01 15:52:14 or is the qte:qte the perpetrator there? Dec 01 15:52:56 nope Dec 01 15:53:01 weird then :D Dec 01 15:53:05 so you ran bitbake opie-image Dec 01 15:53:06 ? Dec 01 15:53:11 could be a weird dependency somewhere Dec 01 15:53:21 if it's just gnome-common it might be for icons or something Dec 01 15:53:24 I'm more or less following the developer faq on the angstrom pgae Dec 01 15:53:25 maybe? Dec 01 15:53:31 bitbake base-image ; bitbake console-image ; bitbake x11-image Dec 01 15:53:36 currently running console-image Dec 01 15:53:42 ah Dec 01 15:53:48 well that's ... interesting then Dec 01 15:53:49 or do I only have to run bitbake opie-image? Dec 01 15:54:01 might as well try console-image first Dec 01 15:54:05 ok Dec 01 15:54:08 hmm, I didn't even know you could build 2008.1 out of .stable yet... Dec 01 15:54:23 i think I'm building unstable :D Dec 01 15:54:24 I guess I haven't been paying attention :) Dec 01 15:54:32 org.openembedded.stable is just the git trunk Dec 01 15:54:39 I was too lazy to rename the local.conf Dec 01 15:54:42 dev is the unstable trunk Dec 01 15:54:48 right? Dec 01 15:54:54 i got no idea actually :D Dec 01 15:55:03 i followed the building-angström page there Dec 01 15:55:09 git clone git://git.openembedded.net/openembedded.git org.openembedded.stable cd org.openembedded.stable Dec 01 15:55:29 nah, you're right Dec 01 15:55:33 I'm just out of date :) Dec 01 15:55:42 I've been being just an angstrom user lately Dec 01 15:56:02 ok. Dec 01 16:11:23 oh btw, how long does the whole image generation take normally? Dec 01 16:11:57 depends heavily on the machine Dec 01 16:12:09 mmh i got some c2d 2ghz here... any rough estimations possible? Dec 01 16:12:46 less than 24 hours? Dec 01 16:12:56 ok :D Dec 01 16:13:14 well, I finished half of console-image now... Dec 01 16:13:20 when I was building images I was doing it on my friend's phenom quad core, and I would just start a build and check in some time later Dec 01 16:13:42 just think back to building XFree86 on a 486...something like that ;) Dec 01 16:14:02 or gnome 0.54 on a k6 266 Dec 01 16:14:16 i wasn't around back then Dec 01 16:14:25 or ... rather ... when was it released? Dec 01 16:14:30 1997 or so? Dec 01 16:14:42 yeah, 1998 or 99 maybe? Dec 01 16:15:19 unless you mean xfree86, and then the answer is quite a while ago Dec 01 16:15:43 well, that'd have been a great thing to do... being 6 and compiling gnome.. or should I rather say "jerking off while my code is compiling?" Dec 01 16:15:51 nvm thoug Dec 01 16:15:52 h Dec 01 16:16:39 heh, I would start a build, go to sleep, start another, go to school, come home, surf the web on dialup for a while, start another build Dec 01 16:17:01 got it compiled over a week or so and it pretty much sucked :) Dec 01 16:17:06 :) Dec 01 16:45:55 right... there we go again. Dec 01 16:46:04 NOTE: package console-image-1.0-r0: task do_qa_staging: started Dec 01 16:46:04 NOTE: QA checking staging Dec 01 16:46:04 ERROR: QA Issue: libncurses.la failed sanity test (workdir) in path /programming/angstrom/angstrom-stable//staging/arm-oabi-angstrom-linux/usr/lib Dec 01 16:46:09 how could I possibly solve this problem? Dec 01 16:46:51 figure out why it failed the sanity test Dec 01 16:46:58 how could I possibly do that? Dec 01 16:47:31 it usually involves paths that point to the system itself being included Dec 01 16:47:50 check the logs in your temp dir for ideas Dec 01 16:47:59 right :) Dec 01 16:55:08 w00h00 :) console image is finished now. Dec 01 16:55:16 woo! Dec 01 16:55:20 still... Dec 01 16:55:26 x11-image wants to do 3899 tasks ... Dec 01 16:55:32 sounds right Dec 01 16:55:45 will it hurt if I just do bitbake opie-image? Dec 01 16:55:47 or at least within an order of magnitude of right Dec 01 16:55:50 :D Dec 01 16:55:54 no of course not Dec 01 16:56:00 ah well... xD Dec 01 16:56:00 bitbake as many images as you want Dec 01 16:56:04 ok. Dec 01 16:56:07 all the common pieces will be re-used Dec 01 16:56:07 where to find them afterwards? Dec 01 16:56:25 ./tmp/deploy/ Dec 01 16:56:38 nvm, found them :) yeah Dec 01 16:56:38 or possibly build/tmp/deploy depending on your setup Dec 01 16:56:50 yeah... Dec 01 16:57:10 so... I could have just run bitbake opie-image the whole time? Dec 01 16:57:29 yeah, but all the work you did making console-image run won't need to be rebuilt Dec 01 16:57:39 s/run/build Dec 01 16:57:42 <- tired Dec 01 16:58:32 still... I might have some problem in the end :/ Dec 01 16:58:48 the created jffs2 images are ~40mb (base-image) and ~80mb(console-image) big Dec 01 16:59:00 my h3600 only got umm... 32mb flash though Dec 01 16:59:23 that... might be a problem :D Dec 01 16:59:36 huh Dec 01 16:59:40 that's quite odd Dec 01 16:59:42 yeah... Dec 01 16:59:52 I'm used to them being <10MB Dec 01 16:59:57 o.O Dec 01 17:00:02 that _is_ weird then. Dec 01 17:00:16 80mb for a console image? Dec 01 17:00:20 did you build a jffs2 image as well as a .tar.gz? Dec 01 17:00:24 yup. Dec 01 17:00:30 my base console images weight in at 2mb to 3mb depending on file format Dec 01 17:00:35 o.O :S Dec 01 17:00:37 why?!? Dec 01 17:00:39 jff2s cramfs tar.gz cpio all in the 2.4 to 3mb Dec 01 17:00:45 yeah I know... Dec 01 17:00:47 still... Dec 01 17:00:53 but he also saw gnome-common being built... Dec 01 17:02:07 oh sorry :S Dec 01 17:02:15 my mistake Dec 01 17:02:17 they are quite small Dec 01 17:02:25 :P Dec 01 17:02:26 misread the byte count. Dec 01 17:02:30 du -h Dec 01 17:02:32 or du -hac Dec 01 17:02:33 yeah Dec 01 17:02:37 I'm not using -h normally though :D Dec 01 17:02:52 you're on a GNU system. might as well take advantage of it Dec 01 17:04:54 indeed. Dec 01 17:05:12 well, I'm loading the console image into flash now... let's see whether it works :) Dec 01 17:05:22 serial port is so slo~w Dec 01 17:23:36 Error: unrecognized/unsupported machine ID (r1 = 0x00000089). Dec 01 17:23:38 that is ... bad Dec 01 17:25:09 bad as in "shiny brick" or bad as in "annoying" Dec 01 17:25:10 ? Dec 01 17:25:16 well.. bad as in FAIL. Dec 01 17:25:19 fail at boot :D Dec 01 17:25:24 now I got to hack this ... Dec 01 17:25:26 -.- Dec 01 17:25:39 well, just a kernel rebuild Dec 01 17:25:43 could be worse :) Dec 01 17:26:01 sure. Dec 01 17:26:08 but I first have to find out how to hack it properly :D Dec 01 17:28:24 which... might pose some problems :( Dec 01 17:29:06 (mind that I'm trying to run the h3600 version on a h3800 machine, just as the handhelds project did it, too) Dec 01 17:29:51 right... Dec 01 17:29:53 great... Dec 01 17:30:32 i think i'm gonna delete my whole temp dir and start anew. Dec 01 17:31:46 that sounds somewhat rash Dec 01 17:32:00 well.. Dec 01 17:32:07 it's the easiest way for me :) Dec 01 17:34:43 deleting my temp dir is taking ages though -.- Dec 01 17:35:10 well yeah Dec 01 17:35:16 which fs? Dec 01 17:35:21 ext3 :/ Dec 01 17:35:28 better than xfs Dec 01 17:35:29 + just 5400rpms Dec 01 17:35:32 probably :) Dec 01 17:35:38 or what that jfs? Dec 01 17:35:42 probably jfs Dec 01 17:35:47 as xfs is ... better than ext3 iirc Dec 01 17:35:55 though I'll eventually switch to ext4 with 2.6.28 :D Dec 01 17:36:18 I let other people beta test fs'es for me Dec 01 17:36:21 xD Dec 01 17:36:25 quite efficient. Dec 01 17:39:08 works out great Dec 01 17:39:17 but OE would be a really great stress test for an FS Dec 01 17:39:45 probably :D Dec 01 17:39:53 I'm just scared for my hdd... Dec 01 17:40:02 it ... made some nasty noises earlier :S Dec 01 17:40:08 (while compiling) Dec 01 17:40:20 eh, magnetic disks don't suffer too much from that kind of thing Dec 01 17:40:41 yup. Dec 01 17:40:58 Well, if it's on its last legs then making it work is likely to stress it a bit. Dec 01 17:42:54 right... Dec 01 17:43:01 why didn't I do this h3800 stuff earlier... Dec 01 17:43:08 or rather, why was I having problems back then -.- Dec 01 19:27:57 right... apparently booting the kernel for my h3800 worked fine. still, after "Uncompressing linux..... done, booting the kernel." I get no further messages on the serial connection... is that good or bad? :D how could I eventually get some messages? Dec 01 19:29:57 nvm, found everything. Dec 01 19:50:36 how can I change the root partition that is passed on to angstrom as kernel arg? Dec 01 19:57:48 right... apparently, /dev/mtdblock1 is no valid root... I got no idea why though. How could I possibly find this out? Dec 01 19:59:39 well first Dec 01 19:59:46 the kernel you built for your device Dec 01 19:59:59 did you specify a start and end point for all your mtd partitions? Dec 01 20:00:24 then did you put an image of either jffs2 or cramfs on /dev/mtdblock1 (usualy using a initrd and nandwrite) Dec 01 20:02:06 umm... Dec 01 20:02:28 I'm using the arm bootloader rev 2.21.12 Dec 01 20:02:49 so it's doing basically everything automatically Dec 01 20:03:04 I just have to send it a jffs2 file via ymodem Dec 01 20:03:26 though I think that the problem is that it's a vfat system but the kernel has only got vfat compiled in as a module Dec 01 20:05:03 how can I recompile the kerneLß Dec 01 20:05:19 or what's its package name? Dec 01 20:06:48 nvm. Dec 01 20:08:48 hum Dec 01 20:08:49 afaik Dec 01 20:08:56 the kernel unless using redboot or something similar Dec 01 20:09:25 might not be able to read your partitions and might need the offsets built in kernel (depends on bootloader) Dec 01 20:09:29 never used yours so I dunno =) Dec 01 20:09:34 mmh... ok. Dec 01 20:09:41 well, I'm just recompiling the kernel with built in vfat support Dec 01 20:09:47 let's see whether it works then Dec 01 20:11:50 *if it works then Dec 01 20:20:28 vfat? Dec 01 20:20:37 shouldn't your flash be running like jffs2 or cramfs lol Dec 01 20:21:49 i know Dec 01 20:22:01 i just got no idea either... :D Dec 01 20:22:13 it's flashing using a jffs2 file Dec 01 20:22:19 so I think that it's using jffs2, too... Dec 01 20:22:25 but somehow... things don't work as they should Dec 01 20:22:36 [ 0.750000] No filesystem could mount root, tried: jffs2 Dec 01 20:22:37 [ 0.750000] Kernel panic - not syncing: VFS: Unable to mount root fs on unknown-block(0,0) Dec 01 20:22:43 nice, isn't it ? :D Dec 01 20:23:48 did you specify Dec 01 20:23:50 in your bootloader Dec 01 20:23:53 for the linux cmd line Dec 01 20:23:58 rootfstype=jffs2 Dec 01 20:23:59 ? Dec 01 20:24:03 mmh... no I didn't. Dec 01 20:24:07 I'll do that. Dec 01 20:24:09 mmmmmm you should! Dec 01 20:24:17 ok, I will do so then Dec 01 20:24:35 if your bootload can't pass cmd lines to kernel then you have to put it in the kernel itself Dec 01 20:24:58 mind you your bootloader should support it Dec 01 20:28:22 it can actually. Dec 01 20:28:50 still the same error Dec 01 20:29:06 even with the rootfstype one Dec 01 20:29:48 jffs2 support is built in Dec 01 20:29:49 weird ... **** ENDING LOGGING AT Tue Dec 02 02:59:57 2008