**** BEGIN LOGGING AT Mon Feb 12 02:59:57 2007 Feb 12 10:45:51 <_law_> angstrom-bootstrap-image boots on my axim x50v :-) Feb 12 10:46:11 <_law_> is it normal that sshd isnt started automatic? Feb 12 10:53:58 no, dropbear should start automatically Feb 12 10:57:39 <_law_> hmm i think there is a problem with /dev/random /dev/urandom Feb 12 10:58:20 wrong kernel config? Feb 12 10:58:20 hmmm, fixed the SIGILL in firefox Feb 12 10:59:11 XorA: http://bash.org/?4424 Feb 12 11:00:30 koen: so firefox 2.0.0.1 with a couple of patches doesnt SIGILL anymore, but it does return to its never opening a windows Feb 12 11:00:48 behaviour we have seen before Feb 12 11:01:58 ugly patches? Feb 12 11:02:28 nah, simple ones I pulled from iceweasel, ones we already had, xptstubs.c and xptcinvoke_arm.c Feb 12 11:02:29 <_law_> koen, CONFIG_HW_RANDOM=y Feb 12 11:02:43 koen: but updated to iceweasel 2.0.0.1 Feb 12 11:05:05 koen: firefox is stuck on a wait4 after a clone, sound similar to dbus problem Feb 12 11:05:26 aha Feb 12 11:05:33 we know how to solve that :) Feb 12 11:15:04 All this linux BT is complete shit. but I finally got my gps working Feb 12 11:16:03 bluez2 is definitely a must. those assh%les who did 3 better have been working on documention Feb 12 11:18:47 i'm trying to port a Farsync WAN card driver onto my CM X270 and encountered this error when loading Unable to handle kernel paging request at virtual address 0000206e Feb 12 11:19:12 Anyone have a pointer? Feb 12 11:20:19 to be exact this happen when executing "outw ( 0x440f, card->pci_conf + CNTRL_9054 + 2 )" Feb 12 11:20:27 TinY: driver not in mainline kernel? Feb 12 11:22:09 hrw: I'm not sure.. coz they come with a patch when i bought the card Feb 12 11:22:45 hrw: I usually recompile the kernel when i use it on x86 machines Feb 12 11:23:18 TinY: so the driver probably was tested only on x86 hardware? Feb 12 11:24:47 yeah.. that's wat the ppl at farsync said Feb 12 11:25:37 hrw: I thought i could just port it over to CM X270 as it come with a PCI.. think it's not that easy Feb 12 11:26:53 good luck Feb 12 11:27:02 I never did pci work Feb 12 11:27:29 hrw: haha.. tks.. any idea where i can get help? Feb 12 11:30:15 lkml? Feb 12 11:34:12 #0 0x40046230 in mult () from /usr/lib/firefox-2.0.0.1/libmozjs.so Feb 12 11:34:12 #1 0x40046c88 in pow5mult () from /usr/lib/firefox-2.0.0.1/libmozjs.so Feb 12 11:34:23 hmm, debugging works nice these days Feb 12 11:46:24 * XorA needs to relearn gdb Feb 12 11:48:58 XorA: something like http://www.openembedded.org/repo/org.openembedded.dev/packages/dbus/dbus-1.0.2/fix-waitpid.patch before the clone? Feb 12 11:53:47 koen: do you have an arm board that can run mozilla configure? Feb 12 11:54:10 XorA: I think so Feb 12 11:54:24 koen: Id love to see the generated hsautocfg for arm Feb 12 11:54:42 jsautocfg.h Feb 12 11:55:23 http://www.openembedded.org/repo/org.openembedded.dev/packages/mozilla/firefox-2.0/arm/jsautocfg.h Feb 12 11:55:56 koen: thats not the generated one Feb 12 11:59:49 I'll build js, since that will generate the same jsautocfg.h Feb 12 12:07:49 XorA: http://www.openembedded.org/~koen/jsautocfg.h Feb 12 12:08:25 koen: cheers Feb 12 12:08:26 from spidermonkey Feb 12 12:09:54 armv5te/EABI Feb 12 12:10:46 koen: 2 notable differences in JS_ALIGN_OF_INT64 and JS_ALIGN_OF_DOUBLE Feb 12 12:15:00 hmmm Feb 12 12:15:46 XorA: could you push the patches and the new jsautocfg? Feb 12 12:16:07 koen: working on firefox now Feb 12 12:16:39 * koen will build ff natively Feb 12 13:30:15 hmmm Feb 12 13:30:26 FF needs >350MB unpacked Feb 12 13:30:33 my sd card isn't that big Feb 12 13:31:07 koen: heh heh Feb 12 13:31:43 the ipk is 8M in size Feb 12 13:32:54 koen: I currently suffer from this problem the BSD guys found http://www.openbsd.org/cgi-bin/cvsweb/ports/www/mozilla-firefox/patches/Attic/patch-nsprpub_pr_src_misc_prdtoa_c Feb 12 13:33:55 eek! Feb 12 13:34:26 man, I fell asleep on the couch an hour ago and I got woken up by a thundercrash Feb 12 13:34:42 that was a bit more scary than that bsd patch ;) Feb 12 13:34:47 :-D Feb 12 13:37:15 * koen spots new opensync releases Feb 12 13:38:20 hmm i am trying to buy a A780 on ebay.de, but noone want to sell it outside Germany. DOes any one know a german webstore where i can buy it? Feb 12 13:39:42 koen: 1.0? Feb 12 13:39:57 hrw: no, 0.22 Feb 12 13:39:59 koen: or 0.stilldoesnotworks? Feb 12 13:40:13 0.worksabitbetterasbefore Feb 12 13:40:45 I'm waiting for 0.willnotfsckaddressbookonyourphone Feb 12 13:41:52 - c = Balloc(k); Feb 12 13:41:52 + c = Balloc(k+2); Feb 12 13:41:57 * koen wtfs at mozilla Feb 12 13:46:58 http://lists.debian.org/debian-arm/2003/10/msg00029.html Feb 12 13:48:24 -#if defined(IS_LITTLE_ENDIAN) Feb 12 13:48:24 +#if defined(IS_LITTLE_ENDIAN) && (!defined(__arm__) || defined(__VFP_FP__)) Feb 12 13:50:53 only pb_'s suggestion makes sense Feb 12 13:51:09 the rest of the peope are babbling about FPA, while we care about VFP Feb 12 13:51:24 koen: patched that in, time to retry Feb 12 13:51:52 BSD guys did sensible thing, they have a working dtoa in their libc and they patched that in Feb 12 13:52:46 indeed Feb 12 13:52:58 glibc doesnt seem to have that function though :-( Feb 12 13:55:38 koen: what is FPA and VFP Feb 12 13:56:03 dspencek: 2 different ways to store floatingpoint on ARM cpus Feb 12 13:56:18 FPA is old and weird Feb 12 13:56:53 see http://wiki.debian.org/ArmEabiPort#head-292fb677a89b53821056f269bc5744e014cacfd7 Feb 12 13:57:30 thanks Feb 12 14:00:27 koen, hi ..ehmm ... I have to say OE seems a bit better now Feb 12 14:00:31 :P Feb 12 14:00:49 It compiles without some wicked problems Feb 12 14:08:06 <_law_> why are already files in /dev in my angstrom image? i think udev adds them automatic Feb 12 14:08:17 <_law_> if i extract my selfbuild angstrom image.tar.bz2 Feb 12 14:09:45 _law_: udev starts in userspace, so you need to have some stuff in /dev to get udev working Feb 12 14:09:57 _law_, you obviously don't want to rely on udev stuff for all Feb 12 14:10:45 48M of debug symbols :-( Feb 12 14:11:08 XorA: at least you can add 500MB swap on the microdrive Feb 12 14:11:18 <_law_> koen, ok and in which .bb file can i find the devices? Feb 12 14:11:42 <_law_> i?ll try to add random and urandom... Feb 12 14:11:54 random and urandom are added by udev Feb 12 14:12:01 and both need kernel support Feb 12 14:13:53 hi all Feb 12 14:13:59 hey mr_nice Feb 12 14:14:56 koen: can you help me out? I have created a new simpad kernel.bb file and it compiles well. but I don't know which .bb I have to change to get a /dev/console entry Feb 12 14:15:22 none Feb 12 14:15:31 OE adds it by default to the rootfs Feb 12 14:15:38 there is no one Feb 12 14:16:19 I untared the tar.bz2 and looked at the /dev/ then I created one myself and re mkfs.jffs2 it Feb 12 14:16:46 do I need to add something at the machine config file? Feb 12 14:18:27 all the images I build have /dev/console Feb 12 14:20:49 files/device_table-minimal.txt has /dev/console Feb 12 14:21:49 koen: http://rafb.net/p/mgvFLa61.html Feb 12 14:22:00 Angstrom-bootstrap-image-test-20070211-simpad Feb 12 14:22:35 XorA: looks ok Feb 12 14:22:54 koen: still no window appeared, but its churning away at 100% cpu load Feb 12 14:23:00 XorA: apart from the fact you built firefox after gnome*, so it has all kinds of spurious deps :( Feb 12 14:23:01 koen, did you generally read all swearing related to bluez 3.x? it's awful. Feb 12 14:23:27 psokolovsky: I read a lot about bluez3 Feb 12 14:23:45 but I have bluez-gnome installed on my devices :) Feb 12 14:24:07 koen, so what's solution? if big guys still didn't conquer it, then it's likely broken in OE for sure... Feb 12 14:24:17 koen, is it in Angstrom? Feb 12 14:24:22 yes Feb 12 14:24:23 http://www.pastebin.ca/351989 does the USE_DEVFS = "1" overwrites the BOOTSTRAP_EXTRA_RDEPENDS += "udev module-init-tools" in the include file? does oe builds a /dev/console if it uses devfs? Feb 12 14:24:25 ok Feb 12 14:24:43 devfs? Feb 12 14:24:44 eeew Feb 12 14:25:08 psokolovsky: http://www.angstrom-distribution.org/repo/?action=details&pnm=bluez-gnome Feb 12 14:25:25 *** glibc detected *** /usr/lib/firefox-2.0.0.1/firefox-bin: malloc(): memory corruption: 0x00075890 *** Feb 12 14:25:42 koen: I don't understand the config it is for to different kernels 2.4 and 2.6 - I only changed the 2.6 Kernel entry Feb 12 14:26:00 XorA: it detected glibc! switch to uclibc! Feb 12 14:26:37 what's wrong with glibc? Feb 12 14:26:55 neal: [15:58] XorA: *** glibc detected *** /usr/lib/firefox-2.0.0.1/firefox-bin: malloc(): memory corruption: 0x00075890 *** Feb 12 14:27:06 it errors out on detecting glibc Feb 12 14:27:29 I see. Feb 12 14:27:34 http://dominion.kabel.utwente.nl/koen/cms/error-messages-that-arent-very-helpfull Feb 12 14:28:04 helpful has a single l Feb 12 14:28:08 at the end Feb 12 14:28:29 right Feb 12 14:28:58 so, I have a problem Feb 12 14:29:10 what's the right way to propagate unlikely errors Feb 12 14:29:32 like sqlite and I/O errors (disk full) Feb 12 14:30:02 if I use GError, it basically means adding a GError to every single function Feb 12 14:30:16 and everyone will just ignore them by passing NULL Feb 12 14:31:45 how can I figure out why there is no /dev/console created? is there a logfile for files/device_table-minimal ? Feb 12 14:33:05 neal: adding GErrors would be a start Feb 12 14:33:19 but if they are ignored, what is the point? Feb 12 14:33:26 right Feb 12 14:33:31 basically, the errors I listed above are unrecoverable Feb 12 14:33:35 you have to bail Feb 12 14:34:28 it's like malloc failing Feb 12 14:34:34 if malloc fails, all you can do is die Feb 12 14:36:17 http://search.cpan.org/~rmcfarla/Gnome2-GConf-1.000/GConf.pm#ERROR_HANDLING Feb 12 14:36:21 this seems like a good approach Feb 12 14:37:34 mr_nice, what prob do you have with /dev/console?: Feb 12 14:37:56 it's not in the image or you just dont see any output on your device? Feb 12 14:38:15 <_law_> koen, did you know who i can enable urandom support im kernel? Feb 12 14:39:27 there is no file /dev/console in rootfs Feb 12 14:39:42 also not in the image Feb 12 14:40:21 is base-files-3.0.14 the right palce to search? Feb 12 14:40:30 dang Feb 12 14:40:32 update your OE installation Feb 12 14:41:44 it is from yesterday evening Feb 12 14:42:09 but I will upgrade it if you think it will solve the problem Feb 12 14:42:24 -r58 Feb 12 14:42:26 * Marex just generated working rootfs Feb 12 14:42:37 :) Feb 12 14:42:42 * mr_nice upgrading Feb 12 15:22:19 Marex: unfortunatly it didn't helped - do I have to do a clean or something first? Feb 12 15:22:35 mr_nice, dunno ... Feb 12 15:22:44 ask koen , it's koens' distro Feb 12 15:22:45 :P Feb 12 15:23:10 it's not my distro Feb 12 15:24:27 * XorA stole it from koen Feb 12 15:29:27 XorA, hehehe :) Feb 12 15:29:40 XorA, and that way the XorAngstrom was born :P Feb 12 16:47:28 hi all! ehh... XorA gone... it was so tuochly(?) when he comment bug that i post... XorA|gone - thanks! Feb 12 18:37:42 koen: I need a faster ARM :-) Feb 12 18:38:14 XorA|gone: faster than the thecus you have? Feb 12 18:38:19 my firefox does get a whole lot further before dumping now Feb 12 18:38:29 koen: I couldnt afford the thecus Feb 12 18:38:33 ah Feb 12 18:38:46 want to borrow wolfson's netbook? Feb 12 18:39:09 it's kind of useless for EABI work since it only has 2.6.9 Feb 12 18:39:22 koen: that wouldnt help me debug firefox then :-) Feb 12 18:39:33 D'oh Feb 12 18:39:56 koen: I shall if I get time tomorrow morning check in the changes I made, then take a rest Feb 12 18:40:20 cool Feb 12 18:40:42 its important as even minimo will need the same fixes, if dougt ever resurects it Feb 12 18:40:55 * koen wants webkit Feb 12 18:41:07 dont we all, but that seems to have stalled Feb 12 18:41:14 that's webcore Feb 12 18:41:24 webkit is the thing apple and kde are working on Feb 12 18:41:29 * XorA|gone does wonder if its the js problem that makes OH Web crash Feb 12 18:42:13 probably Feb 12 18:42:28 js also uses jsautocfg.h Feb 12 18:42:39 yeah, and the same dtoa function Feb 12 18:44:03 time to eat chinese Feb 12 18:44:17 enjoy Feb 12 20:47:04 ~lart firefox for being low quality junk **** ENDING LOGGING AT Tue Feb 13 03:00:00 2007