**** BEGIN LOGGING AT Wed Dec 12 02:59:57 2007 Dec 12 02:59:57 03philippe 07org.oe.dev * r80514cdf... 10/ (4 files in 3 dirs): Dec 12 02:59:57 gpe-mini-browser: fix QA issue with desktop file Dec 12 02:59:59 * added patch to fix up desktop file Dec 12 03:00:01 * bumped PR Dec 12 03:00:03 * Fixes bug #3445 Dec 12 06:19:17 Argh.. http://rafb.net/p/7mfQRG70.html Dec 12 06:19:53 ~blame pfalcon for breaking matchbox-keyboard-inputmethod in openmoko Dec 12 06:19:54 * ibot blames mwester-laptop and perhaps pfalcon for breaking matchbox-keyboard-inputmethod in openmoko and the end of life as we know it Dec 12 06:59:10 03john_lee 07org.oe.dev * rc6ccdf57... 10/ (1 classes/sourcepkg.bbclass): Dec 12 06:59:10 fix typo to make sourcepkg.bbclass work when inherited. Dec 12 06:59:10 sourcepkg.bbclass:109:addtask dumpdata after archive_bb before do_create_diff_gz Dec 12 06:59:10 archive_bb should be "do_archive_bb". Dec 12 07:29:59 good morning Dec 12 07:31:25 hey mickeyl Dec 12 07:34:40 morning Dec 12 07:34:57 i managed to get some sane wake up time again Dec 12 07:35:26 mickeyl: get some children and that's fixed for ever ;) Dec 12 07:39:36 hehe Dec 12 07:39:38 i can imagine Dec 12 07:58:41 03xjqian 07org.oe.dev * r5006f70b... 10/ (1 conf/distro/include/angstrom-2007-preferred-versions.inc): Dec 12 07:58:41 angstrom-2007-preferred-versions.inc: prefer octave 2.1.73 (stable). Dec 12 07:58:41 * the old 2.9.14 never built Dec 12 07:58:41 * the new 2.9.x is still in testing Dec 12 07:58:48 03xjqian 07org.oe.dev * rc1c864a1... 10/ (3 files in 2 dirs): octave: bump 2.9.19, fix -dbg += bug#2317 Dec 12 07:58:53 03xjqian 07org.oe.dev * r16b35033... 10/ (8 files in 4 dirs): zroadmap: move zroadmap recipies to packages/roadmap for merging Dec 12 08:30:27 'morning all Dec 12 08:31:56 morning Dec 12 08:31:57 you're late Dec 12 08:32:01 :D Dec 12 08:38:36 yeah Dec 12 08:38:48 slept like 11 hours Dec 12 08:39:42 mickeyl: what are your thought on a moko-minimal-image that only has dialer and today? Dec 12 08:41:51 sure thing Dec 12 08:41:53 sounds good Dec 12 08:42:23 yes yes... more images! Dec 12 08:42:27 hehe Dec 12 08:43:07 I just bought a GSM eval kit - looks like its time to add support for it to gsmd :) Dec 12 08:45:02 cool Dec 12 08:46:47 yeah... I only need some free time. But it should be really easy... the documentation is great. Dec 12 08:55:10 bonjour Dec 12 08:57:31 mickeyl: in order to cut buildtime for moko-minimal-image: http://www.openembedded.org/~koen/moko-task-splitup.diff.txt Dec 12 08:57:42 ehm Dec 12 08:57:53 * koen spots a huge bug Dec 12 09:02:27 koen: feel free to commit that (minus your huge bug) ;) Dec 12 09:02:50 the huge bug is that the new files aren't *.bb files :) Dec 12 09:02:55 ah, heh Dec 12 09:03:19 I did touch Dec 12 09:03:27 and then 'open *moko*' Dec 12 09:03:31 hehe Dec 12 09:18:19 mickeyl: split committed Dec 12 09:20:11 thanks! Dec 12 09:21:55 mickeyl: btw, the distro specific patches to mb-kb-im don't apply anymore Dec 12 09:22:20 oh that's unfortunate Dec 12 09:22:31 then we need to setback the date Dec 12 09:22:41 the creators no longer work on that Dec 12 09:22:47 have been moved to other teams Dec 12 09:24:00 is that patch still needed? Dec 12 09:24:08 and why was it distro specific? Dec 12 09:25:28 somehow recent matchbox keyboard miscalulated the necessary font size Dec 12 09:25:37 the patch really is just Dec 12 09:25:38 - kb->font_pt_size = 5; Dec 12 09:25:39 + kb->font_pt_size = 3; Dec 12 09:25:55 so should be possible to adjust easily Dec 12 09:26:01 hi lrg Dec 12 09:26:35 good morning Dec 12 09:26:38 yo lrg Dec 12 09:26:39 hi ant Dec 12 09:26:50 hey flo_lap, mickeyl, ant Dec 12 09:27:49 mickeyl: don't paul's patches fix that? Dec 12 09:28:06 what are paul's patches? Dec 12 09:28:08 CoreDump: ping Dec 12 09:28:21 pong Dec 12 09:28:25 2-Add-new-modifier--layout--Used-to-cycle-thru-all-available-layouts.patch Dec 12 09:28:25 3-Changes-to-improve-layout-rendering--especially-after-adding-support-for.patch Dec 12 09:28:25 4-Add-rendering-debug-logging.patch Dec 12 09:28:25 5-Add-support-for-loading-multiple-independent-layouts.patch Dec 12 09:28:26 6-Add-layout-switch-key-to-all-layouts.patch Dec 12 09:28:33 oh all these Dec 12 09:28:41 CoreDump: referring http://bugs.openembedded.org/show_bug.cgi?id=3459 Dec 12 09:28:55 i don't know whether they fix it, i just know that the fic-gta01 patch was necessary to get the keyboard fit Dec 12 09:28:59 CoreDump: I mean at the very first boot Dec 12 09:29:08 otherwise the keys were much too large Dec 12 09:29:19 CoreDump: you cannot insert the altbot pwd Dec 12 09:29:37 ahh, the password Dec 12 09:29:47 CoreDump: so you're stuck Dec 12 09:30:04 does the keyboard work in the other image in the VT? Dec 12 09:30:31 CoreDump: yes in console mage Dec 12 09:30:55 CoreDump: my keyboard here is a disaster... Dec 12 09:32:26 CoreDump: I cannot check right now but don't remember errlogs Dec 12 09:32:41 I'll have a look Dec 12 09:32:46 CoreDump: only one mismatch with some TTY Dec 12 09:32:55 CoreDump: ... Dec 12 09:34:00 CoreDump: ahh, was complaining about missing blueprobe. Has been removed ? Dec 12 09:34:18 no clue Dec 12 09:34:47 CoreDump: I'll check tonight. Dec 12 09:35:32 CoreDump: now I'm stuck with "Preiserhoeungen wird nicht akzeptiert" Dec 12 09:35:42 CoreDump: bigger problem than OE !:-) Dec 12 09:36:16 =) Dec 12 09:37:31 have to move, bbl Dec 12 10:02:04 * * OE Bug 3461 has been created by  Dec 12 10:02:06 * * gcc-cross-4.1.2-r13-do_compile Dec 12 10:02:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3461 Dec 12 10:03:05 * * OE Bug 3461 has been RESOLVED (INVALID) by Dec 12 10:03:07 * *  gcc-cross-4.1.2-r13-do_compile Dec 12 10:03:09 * * http://bugs.openembedded.org/show_bug.cgi?id=3461 Dec 12 10:04:24 jesus, koen is closing the bugs faster than the poor autobuilder can open them Dec 12 10:09:49 hehe Dec 12 10:10:01 i think the oe autobuilder should improve the fetcher checks Dec 12 10:10:11 e.g. sf.net cvs Dec 12 10:10:16 is often overloaded Dec 12 10:10:20 so it returns an error Dec 12 10:10:26 but once you try again, it works Dec 12 10:10:38 i don't fancy dozens of false positive bug reports Dec 12 10:11:02 well, perhaps bitbake should be improved Dec 12 10:11:05 trying to fetch more than once Dec 12 10:11:09 i'll think about that Dec 12 10:26:50 a fetch error is a fetch error Dec 12 10:30:13 not on some sites Dec 12 10:30:40 perhaps cvs.sf.net returns something that says try again? Dec 12 10:30:46 if the build halts, it's an error Dec 12 10:30:47 then we should honor this Dec 12 10:30:53 this is simple thinking Dec 12 10:30:58 and it will clutter our bug tracker Dec 12 10:31:04 i would rather have more high level bugs Dec 12 10:31:09 *shrug* Dec 12 10:31:31 wanna explain that to users that have a 'failed' build? Dec 12 10:31:41 hey everyone Dec 12 10:31:42 angstrom 2007 builds can halt on fetch error as it means the source archives arent updated :-) Dec 12 10:31:44 they would not if we fix bitbake Dec 12 10:31:55 "Yes, your build failed, but it's not an error" Dec 12 10:32:09 or we forgot to lock down something for release Dec 12 10:32:36 XorA|gone: or someone replaces recipes which renders the lockdown useless Dec 12 10:32:50 * XorA|gone whistles innocently Dec 12 10:33:07 and the autobuilder updates the source archives :) Dec 12 10:33:14 so if the fetch fails... Dec 12 10:35:04 Could someone give me a hint about kdrive? I'm using it, it works mostly but I'm having trouble with keyboard. The symbols are completely crazy. Dec 12 10:35:14 I'm using kdrive 1.3 Dec 12 10:36:09 I have read this very interesting post: http://blog.haerwu.biz/2007/12/07/armv6-booted-in-qemu/ ... does it mean that I can soon change ASSUME_PROVIDED += "qemu_native" for my armv6 processor ? Dec 12 10:36:58 vlo: once hrw|gone fixes qemu and it gets enabled in OE Dec 12 10:37:18 for running qemu in systemmode we also need th nptl patches Dec 12 10:37:28 03koen 07org.oe.dev * re27d6ace... 10/ (1 conf/distro/include/angstrom-2007-preferred-versions.inc): angstrom 2007 preferred versions: fixes the autobuilder needed to get things building Dec 12 10:37:34 03koen 07org.oe.dev * rcdb61d34... 10/ (1 packages/images/minimal-openmoko-image.bb): minimal-openmoko-image: image with only dialer + terminal Dec 12 10:37:45 03koen 07org.oe.dev * r69b3efcf... 10/ (1 packages/tasks/task-openmoko.bb): task-openmoko: fix typo Dec 12 10:39:42 koen: hum OK ... thanks Dec 12 10:41:40 maybe am I mixing things, but is it correct that this qemu thing is also related to the GLIBC_GENERATE_LOCALES value, and the fact that if I set this variable, my build fails ? Dec 12 10:42:17 because I would like to avoid building all these locals Dec 12 10:42:19 angstrom disables that for armv6 cpus Dec 12 10:43:53 what do you mean by "that" ? Dec 12 10:45:07 the locale generation thing Dec 12 10:47:58 3.6M Angstrom-minimalist-image-glibc-ipk-2007.9-test-20071212-a780.rootfs.tar.gz Dec 12 10:47:58 7.3M Angstrom-console-image-glibc-ipk-2007.9-test-20071212-a780.rootfs.tar.gz Dec 12 10:47:59 11M Angstrom-minimalist-gpe-image-glibc-ipk-2007.9-test-20071212-a780.rootfs.tar.gz Dec 12 10:47:59 19M Angstrom-x11-image-glibc-ipk-2007.9-test-20071212-a780.rootfs.tar.gz Dec 12 10:47:59 21M Angstrom-minimal-openmoko-image-glibc-ipk-2007.9-test-20071212-a780.rootfs.tar.gz Dec 12 10:47:59 33M Angstrom-openmoko-image-glibc-ipk-2007.9-test-20071212-a780.rootfs.tar.gz Dec 12 10:49:50 an order of magnitude difference :) Dec 12 10:50:05 (note that a780 puts the kernel in /boot, so that's 1.4MB) Dec 12 10:55:06 koen: do you have any idea why I could be having problems with keyboard? Dec 12 10:55:19 nope, sorry Dec 12 10:55:41 koen: do you have any image working with regular keyboards? Dec 12 10:56:02 koen: if yes, does delete, page/up and like works? Dec 12 10:57:34 the efika images work with a regular keyboard Dec 12 10:57:46 efika ones? Dec 12 10:57:57 have you tried it with an usb one? Dec 12 10:59:13 I've installed console-tools and used showkeys and it doesn't look to be a kernel issue since it does send the right symbols Dec 12 10:59:40 So I've installed xev and checked that it is, really, sending wrong ones, while on X Dec 12 11:00:02 efika only has usb plug Dec 12 11:00:47 mickeyl: 12MB saved when using the minimal openmoko-image Dec 12 11:01:54 koen: good Dec 12 11:02:22 otavio: is your xmodmap global or local to the keyboard? Dec 12 11:02:42 koen: how I can check it? Dec 12 11:02:48 dunno Dec 12 11:02:52 never really used xmodmap Dec 12 11:03:03 pH5 is the keyboard guru here Dec 12 11:03:10 koen: I've built a xkeyboard-config .bb recipe and it does load the keyboard layout Dec 12 11:03:22 bbl Dec 12 11:03:28 koen: Does the minimal image include libicu? Dec 12 11:03:37 koen: if I query the X using setxkeymap it gives the right anwsers Dec 12 11:04:32 koen: i'll wait him and see if he can help me Dec 12 11:04:37 koen: This is from two weeks ago, but still: http://rafb.net/p/Fhjh4960.html Dec 12 11:04:58 koen: thanks by your help Dec 12 11:05:21 CM: icu is needed by webkit, minimal doesn't have the feedreader and browser (if I didn't mess up) Dec 12 11:07:09 Ok, should save a lot of space then :) Dec 12 11:10:06 CM: you don't gain the 20+MB you expect due to compression Dec 12 11:10:43 Ok. That libvorbisenc sure shouldn't be there either.. Dec 12 11:11:19 koen: still, nice work. The smaller "base" the better :) Dec 12 11:29:50 CM: http://rafb.net/p/owH8nJ32.html Dec 12 11:30:27 CM: that's from a tarball, so that probably doesn't have an iconcache Dec 12 11:34:02 koen: Oh, neat :) Dec 12 11:54:41 re Dec 12 11:54:54 flo_lap: wb Dec 12 12:22:17 CM: http://dominion.thruhere.net/koen/cms/slimming-down-oe-built-images Dec 12 12:27:22 koen: Thanks for doing more than just looking at things that should be done :) Dec 12 12:43:33 re Dec 12 12:43:51 mwester-road: http://dominion.thruhere.net/koen/cms/slimming-down-oe-built-images Dec 12 12:45:51 :) Nice... Dec 12 12:46:18 * mwester-road will have room on his jffs2 rootfs again! Dec 12 12:48:33 mickey|lunch: who needs to be poked to so "sed -i s:kabel.utwente.nl:thruhere.net:g config.ini" on the openmoko planet server? Dec 12 12:51:34 koen: maybe roh? Dec 12 12:51:58 What uses libicudata? Dec 12 12:52:12 webkit I think Dec 12 12:52:28 webkit uses ICU for unicode conversions Dec 12 12:52:30 sadly Dec 12 12:52:52 If so, then webkit is consuming 16M in libs that only webkit uses. Dec 12 12:52:55 Ouch. Dec 12 12:52:58 koen: can't zecke fix that? ;) Dec 12 12:53:47 flo_lap: zecke and I decided that other webkit problems have a higher priority :) Dec 12 12:54:23 mwester: icu compresses well, so it comes down to 8MB, still a lot, though Dec 12 13:39:54 I have a probelm with "arm-angstrom-linux-gnueabi-as: unrecognized option `-Qy'" -> http://pastebin.ca/813334 Dec 12 14:10:22 Hi, is there a printable version (PDF,onepage html) of the oe user manual? Dec 12 14:12:09 * florian_ is not sure Dec 12 14:12:28 Reach: worst case you can generate a pdf from the manual sources Dec 12 14:13:15 hrm Dec 12 14:13:29 s/Reach/reppel/ Dec 12 14:13:34 yep :) Dec 12 14:17:04 ok I'm installing xmlto Dec 12 14:19:37 reppel: http://www.angstrom-distribution.org/openembedded-user-manual-available-pdf Dec 12 14:19:58 morning Dec 12 14:20:07 hey chouimat|work Dec 12 14:20:11 koen: thanks Dec 12 14:20:54 huhu Dec 12 14:21:03 koen: is it re-generated often? Dec 12 14:21:38 hi Jin^eLD Dec 12 14:22:40 reppel: the one on the website is regerated every hour, but there haven't been any changes to the docbook file the past months Dec 12 14:22:44 hi florian Dec 12 14:25:35 cbrake: hello Dec 12 14:25:56 cbrake: do you remember having serial port issues on cm-x270 back in 2.6.22 ? Dec 12 14:26:34 nud: no, I don't recall any issues Dec 12 14:26:44 I'm struggling with a strange one Dec 12 14:26:55 the first time I start my program, it sometimes doesn't work Dec 12 14:27:05 but it always work the second time Dec 12 14:27:08 nud: we are using the 2nd and 3rd ports pretty heavy in an application Dec 12 14:27:17 any idea where it could come from ? Dec 12 14:27:26 * nud wonders where is ttyS3 Dec 12 14:27:29 ttyS2* Dec 12 14:27:35 nud: are you using handshaking? Dec 12 14:27:42 handshaking ? Dec 12 14:27:48 nud: RTS, CTS, etc Dec 12 14:27:50 there should be no hardware flow control Dec 12 14:28:18 the device I try to communicate with doesn't support flow control Dec 12 14:28:45 nud: that is what I'm doing Dec 12 14:29:12 nud: you might want to dump the GPIO registers when it fails and make sure they are set up correctly, but I'm not sure why running your program the second time would fix that. Dec 12 14:29:28 how can I dump those registers ? Dec 12 14:29:42 devmem2 or pxaregs Dec 12 14:30:32 hehe devmem2 is evil :) Dec 12 14:30:51 hell, I only have 'cat' Dec 12 14:31:04 * * OE Bug 3462 has been created by  Dec 12 14:31:06 * * librecord-0.1-r0-do_fetch Dec 12 14:31:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3462 Dec 12 14:31:33 nud: bitbake pxaregs Dec 12 14:31:42 yep that's what I'm doing right now Dec 12 14:31:43 florian_: the gpephone entries in the angstrom lockdown file are not up to date Dec 12 14:32:00 florian_: they still use the 0.1 files that the orange engineers messed up Dec 12 14:34:04 * * OE Bug 3463 has been created by  Dec 12 14:34:06 * * librecord-1.0-r0-do_configure Dec 12 14:34:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3463 Dec 12 14:35:08 koen: uh right, i might have an update at kc... Dec 12 14:45:01 hey zecke Dec 12 14:51:57 koen: i just told the OM planet admin Dec 12 14:52:04 he will change it soon Dec 12 14:54:12 cbrake: are you able to decrypt what pxaregs give ? especially, where should I look for serial port configuration Dec 12 14:54:33 mickey|lunch: thanks Dec 12 14:54:50 np Dec 12 14:56:39 03pfalcon 07org.oe.dev * r68bc31a5... 10/ (3 files in 2 dirs): linux-handhelds-2.6: Use deployed kernel filename consistent with main kernel recipe. Dec 12 14:56:43 03pfalcon 07org.oe.dev * rda5473b9... 10/ (1 contrib/angstrom/build-release.sh): build-release.sh: Add h3900. Dec 12 14:57:51 * koen ponders on victims to test http://www.angstrom-distribution.org/unstable/autobuild/toolchains/ Dec 12 14:58:13 what was that tool again to convert between different SCMs? Dec 12 14:58:25 tailor Dec 12 14:58:27 ah, right. thanks Dec 12 14:58:41 or 'zecke', if tailor doesn't work Dec 12 15:01:16 mickey|lunch: the gtk-iconcache is supposed to be a mmap'able file, right? Dec 12 15:02:08 * koen wonders how that works on jffs2, that can't mmap files Dec 12 15:03:03 koen: i never dug so deep into gtk, can't answer that Dec 12 15:03:11 jffs2 definitely has an mmap call Dec 12 15:03:18 i don't know how they're implemening it Dec 12 15:03:25 but lots of stuff would break without one Dec 12 15:04:22 AFAIK most apps have a fallback Dec 12 15:04:44 yaffs cites mmap as a feature over jffs2 Dec 12 15:05:02 but we all know not to believe yaffs propaganda Dec 12 15:05:19 yes please Dec 12 15:05:32 anyone who tried logfs? Dec 12 15:05:57 There was a talk at the meeting in Linz which was promising. Dec 12 15:06:41 I talked to joern about it last year Dec 12 15:06:47 I still have high jopes Dec 12 15:06:47 mmh Dec 12 15:06:50 hopes* Dec 12 15:06:54 I happen to have CRTSCTS Dec 12 15:06:57 set Dec 12 15:07:17 mickey|lunch: no need to use tailor for cvs->svn btw. Dec 12 15:07:22 cbrake: so yes I'm using handshaking Dec 12 15:07:56 "My goal is to push for inclusion into 2.6.25" - logfs.org Dec 12 15:08:16 nud: check the GPIO direction and the alternate function registers for the pins that are used for the serial port Dec 12 15:08:20 koen: yep, seems to be a smart guy and even if it does not seem to be finished it might do a good job already Dec 12 15:08:58 nud: see PXA270 dev manual for definitions of the GPIO registers Dec 12 15:09:24 florian_: jffs2 doesn't support >=4GB partitions, logfs does Dec 12 15:11:03 koen: well... i didn't run into this limitation yet Dec 12 15:11:45 Well.. it might becom an issue on the eepc which has up to 8gb of flash. Dec 12 15:12:06 I tried to buy one here, but the better ones are hard to get... Dec 12 15:12:18 florian_: eeepc uses a CF bridge Dec 12 15:12:48 koen: ah right Dec 12 15:15:30 volunteers for adding eeepc machine descriptions to OE? Dec 12 15:16:06 if i get one :o) Dec 12 15:16:26 Xorg is in a reasonable shape judging from my efika Dec 12 15:16:33 * Genesis has prefer bye a ibm X31 on ebay last sunday Dec 12 15:17:06 Genesis: good choice, i'm just using one :) Dec 12 15:17:30 hehe a friend of mine has one too Dec 12 15:17:39 so i just dd his disk for my install :) Dec 12 15:18:12 i paid it 300 € , i'm waiting now :) Dec 12 15:18:18 koen: yeah, but i guess that will have to wait until they are available in germany... we even migt get them for the kc shop Dec 12 15:19:07 Genesis: that's not really much, i have seen these much more expensive on ebay lately Dec 12 15:19:49 it's cheapeast than a eeepc for sure. Dec 12 15:20:00 and i participate to recyclage :) Dec 12 15:20:44 i'll put gentoo on it and develop some things for OE on it Dec 12 15:21:21 first thing : an diskdruid tools that works with image_tables Dec 12 15:21:29 on images files Dec 12 15:21:44 to make multipartitions images disks with OE Dec 12 15:22:19 i don't find any project like that Dec 12 15:23:05 Genesis: well, a ~300EUR device cant't compete with the x31. Dec 12 15:25:17 :) Dec 12 15:25:36 it's second hand but it's robust Dec 12 15:27:18 Genesis: I stopped to buy new PC stuff... even the company uses used stuff mostly. Dec 12 15:27:56 when i see the poor garanties with new material and the time lost trying to get it oki Dec 12 15:28:10 and the 2 months of dell for deliver us our material Dec 12 15:28:42 i don't see the real interest to get first hand hardware. Dec 12 15:29:06 cutting down on compile times might be a good motive ;) Dec 12 15:29:29 i've a quad core xeon serveur Dec 12 15:29:59 My quad box arrives tomorrow or the day after...finally Dec 12 15:30:34 CoreDump: well... just bought a used 3,6 GHz Xeon for almost nothing. Dec 12 15:31:15 single core? :\ Dec 12 15:31:30 dual/quad _does_ make quite a difference here Dec 12 15:32:25 anybody tested the tickless kernel stuff and high-res timer under ARM with the newer kernels? pass or fail? Dec 12 15:33:21 CoreDump: dual Dec 12 15:33:40 CoreDump: and a spare socket for an additional cpu Dec 12 15:33:47 sweet Dec 12 15:34:06 VoodooZ: imho it works on omap devices Dec 12 15:35:43 good. any major power saving observed or performance side-effects you've heard abou t? Dec 12 15:38:17 I'm also wondering about highres timer for robotics use.. Dec 12 15:38:25 VoodooZ: I think that the Nokia tablets use this for power saving - they run all the time si its not a common "sleep" mode. Dec 12 15:44:59 florian_: thanks. I'm developing on NSLU2 for now but might start looking in something more "potent" for opencv.. Dec 12 15:45:13 those little eee pcs look nice and small... Dec 12 15:58:04 crap, timedout Dec 12 15:59:44 so.. what is the correct way to limit a package to be compatible with only a set of machines? I tried adding more than one machine to COMPATIBLE_MACHINE variable but that did not work out then Dec 12 16:09:17 cbrake: one last question: does your serial port work *everytime* or does it fail sometimes ? Dec 12 16:09:24 nud: every time Dec 12 16:09:46 it's rather strange here the port fails something like once every ten boots Dec 12 16:10:03 nud: what is the failure mode? Dec 12 16:10:06 do you have some sample code you use to init the serial port ? Dec 12 16:10:22 nud: sure, just a min ... Dec 12 16:10:23 well in fact the port is opened without error but nothing is transmitted Dec 12 16:12:01 nud: http://pastebin.ca/813502 Dec 12 16:12:40 cbrake: thanks. The code I use for my serial port is this one: http://pastebin.ca/813503 (nothing clever beside spawning a thread to use blocking reads on the serial port) Dec 12 16:12:58 rs485 ? Dec 12 16:13:16 * nud uses rs232 Dec 12 16:13:30 nud: yes, just a difference in transcievers at the hardware level Dec 12 16:13:49 so they are interoperable? Dec 12 16:14:02 nud: no, different levels, etc Dec 12 16:14:22 the port I use is marked as TTL in the dev doc Dec 12 16:14:29 is it a rs485 one? Dec 12 16:14:48 nud: at the PXA270 level, everything is the same Dec 12 16:15:04 nud: TTL signals go from the PXA270 to an external transciever Dec 12 16:15:13 ah ok Dec 12 16:15:17 nud: and then get shifted to RS232, 485, etc Dec 12 16:15:47 ok Dec 12 16:16:30 * nud finds it rather strange it fails sometimes that way Dec 12 16:21:29 Genesis: btw. be careful with cheap replacement batteries at ebay... there is one at least selling faked "high capacity" ones Dec 12 16:22:38 thx for the advice Dec 12 16:22:55 i'll command at the same chinese vendor of my friend if i need it Dec 12 17:04:22 pH5: are you there? Dec 12 17:04:52 hi otavio Dec 12 17:05:00 pH5: I'm having trouble with keyboard and koan told me you're the OE expert on it and would like to ask some things to you if you're available Dec 12 17:05:22 zecke: had time to look at the mirror? I'd like to sync with lastest OE Dec 12 17:05:41 zecke: if you're busy, I could help to work on the script Dec 12 17:06:24 otavio: the script is in the public :) Dec 12 17:06:43 zecke: what error do you have? Dec 12 17:11:08 otavio: sure Dec 12 17:12:08 pH5: we're having problem with keyboard symbols Dec 12 17:12:21 pH5: we want to use xkb to allow user to set the keyboard layout Dec 12 17:12:34 pH5: so we've enabled it back on libxt, kdrive and libx11 Dec 12 17:13:01 pH5: all looks OK and setxkbmap looks to work too Dec 12 17:13:21 pH5: it does change the layout but some keys give completely crazy symbols Dec 12 17:13:37 pH5: as: delete, insert, page/up, page/down and like Dec 12 17:17:21 bye Dec 12 17:18:30 uh-oh, I'm 'expert' in xkb removal, not using it. Dec 12 17:20:17 pH5: but how we might identify what's the problem cause? Dec 12 17:20:28 pH5: we're lost since it looks to be on X itself Dec 12 17:21:03 pH5: and we looks to need xkb to be able to change the layout in an easy way, won't we? Dec 12 17:25:01 CM: ping Dec 12 17:25:55 koen: pong :) Dec 12 17:26:07 watch the next commmit Dec 12 17:26:18 * CM holds his breath Dec 12 17:26:39 * koen waits for CIA Dec 12 17:27:14 03koen 07org.oe.dev * ra5273f9b... 10/ (3 files in 3 dirs): Dec 12 17:27:14 webkit-gtk: fiddle with exceptions and rtti cflags Dec 12 17:27:14 before: 3.0M armv5te/libwebkitgtk1_0.1+svnr27222-r3_armv5te.ipk Dec 12 17:27:14 after: 2.3M armv5te/libwebkitgtk1_0.1+svnr27222-r3_armv5te.ipk Dec 12 17:27:38 CM: there, you go :) Dec 12 17:27:46 Nice, what flags did you change? Dec 12 17:28:11 I added -fno-exceptions -fno-rtti Dec 12 17:28:22 see http://bugs.webkit.org/show_bug.cgi?id=16302 for details Dec 12 17:28:23 Heh, good thing for 700k Dec 12 17:28:31 * koen should have put that in the commit message Dec 12 17:29:10 cool Dec 12 17:29:30 otavio: I'm taking a look Dec 12 17:29:39 koen: Can you review my patch posted here http://bugs.openembedded.org/show_bug.cgi?id=3460 Dec 12 17:30:28 I have built a qemuarm x11-image Dec 12 17:31:05 is there any info we have put together on how to run qemu on OE Dec 12 17:31:37 I have run it outside OE but this one does not work with the image from OE Dec 12 17:33:40 zecke: any guestimate on when you'll have time to update the moko stuff to the new webview api? Dec 12 17:34:06 koen: what about tonight? Dec 12 17:34:12 that'd be cool Dec 12 17:34:35 should I update webkit now in OE? Dec 12 17:34:51 not now, alp's seems to merge patches Dec 12 17:35:22 with 'update' I mean applying http://lists.linuxtogo.org/pipermail/openembedded-devel/2007-December/003716.html Dec 12 17:35:45 koen: I'm building the OM libs now Dec 12 17:36:52 otavio: if it is in X itself, I fear you have to wade through that code or ask somebody knowledgeable. I have no experience with the xkb translation. Dec 12 17:41:39 pH5: good :( Dec 12 17:41:56 pH5: I have no idea where it's since it all looks right but symbols are still wrong Dec 12 17:48:51 Hello Dec 12 17:48:56 hey sirfred Dec 12 17:50:51 brb, booting to OSX.... Dec 12 18:03:10 koen: thank you for looking at the patch. Do i increase PR in eglibc_svn.bb too Dec 12 18:03:25 yes Dec 12 18:03:28 * Khem has been reading the oe manual Dec 12 18:03:50 koen: how about for intermediate and initial bbs Dec 12 18:04:12 those don't generate packages and don't need to get rebuilt afaik Dec 12 18:04:24 actually I need to change PV also Dec 12 18:04:39 now glibc is moved on to 2.7.90 Dec 12 18:04:51 on trunk Dec 12 18:04:59 and so has eglibc Dec 12 18:05:31 PV = "2.7+svnr${SRCREV}" Dec 12 18:05:36 is it ok ? Dec 12 18:05:45 yeah Dec 12 18:06:04 * * OE Bug 3399 has been RESOLVED (FIXED) by raj.khem(AT)gmail.com Dec 12 18:06:06 * *  Add --disable-libgomp to gcc 4.2.x recipes Dec 12 18:06:08 or PV = "2.7.90+svnr${SRCREV}" Dec 12 18:06:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3399 Dec 12 18:06:16 * * OE Bug 2894 has been RESOLVED (FIXED) by raj.khem(AT)gmail.com Dec 12 18:06:18 * *  eglibc should provide glibc packages when chosen as default libc Dec 12 18:06:20 * * http://bugs.openembedded.org/show_bug.cgi?id=2894 Dec 12 18:07:04 * * OE Bug 3464 has been created by raj.khem(AT)gmail.com Dec 12 18:07:06 * * Multiple files for bluez-utils-dbus Dec 12 18:07:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3464 Dec 12 18:14:16 binutils-cross-sdk is doing weird stuff Dec 12 18:14:33 this can't be right? in the install directory: ./binutils-cross-sdk-dbg/opt/bco/arm/bin/.debug/.debug/.debug/.debug/.debug Dec 12 18:22:33 and if someone could have a look at this one: http://bugs.openembedded.org/show_bug.cgi?id=2482 Dec 12 18:22:36 :> Dec 12 18:26:37 RP: I need your help on setting up qemu Dec 12 18:36:44 does building meta-toolchain work for anyone? Dec 12 18:40:46 ~seen woglinde Dec 12 18:40:50 woglinde was last seen on IRC in channel #oe, 2d 23h 5m 25s ago, saying: 'hi rp'. Dec 12 18:50:03 koen: I'm updating them now Dec 12 18:50:14 koen: need to spend some time on compiling them though Dec 12 18:50:16 hi zecke Dec 12 18:56:01 why is meta-toolchain building the external-toolchain .bb file? Dec 12 18:56:19 aem, not building the file but causing the recipe to be build? Dec 12 18:56:33 from what I understand external-toolchain.bb will only build if the toolchain is already installed on the system Dec 12 18:56:43 however to obtain one, building meta-toolchain is required Dec 12 18:56:56 looks like running in circles.. ideas anyone? Dec 12 18:57:14 external-toolchain.bb is a meta file to tell OE you already have a toolchain Dec 12 18:57:31 koen: well, I don't, I want to create one by building the meta-toolchain package Dec 12 18:57:51 but this seems to trigger the building of external toolchain which of course fails, because it looks for the toolchain in /opt/bla and can not find it Dec 12 18:58:39 so I wonder which package is setting the dependency that causes external-toolchain to be built Dec 12 19:01:48 Jin^eLD: do bitbake -v -D Dec 12 19:01:55 Jin^eLD: and save the output Dec 12 19:02:42 indeed, good idea Dec 12 19:05:29 ok that's funny, I can't find anyone who would be selecting it Dec 12 19:05:32 looking 2nd time.. Dec 12 19:06:51 Jin^eLD: Look for Selecting with big 'S" Dec 12 19:08:29 noone is Selecting external-toolchain Dec 12 19:08:33 that makes no sense Dec 12 19:08:35 mhm Dec 12 19:10:37 well, let me move away the external-toolchain.bb file and see who complains Dec 12 19:13:18 lol Dec 12 19:13:19 :) Dec 12 19:13:22 that makes no sense at all Dec 12 19:13:31 I remove the external-toolchain.bb and then meta-toolchain did build Dec 12 19:14:07 Jin^eLD: http://monotone.openembedded.org/revision/info/8c692ec4078a7d2e345213289da925bfca3cac63 Dec 12 19:15:44 uh Dec 12 19:15:49 I did a fresh pull/update today Dec 12 19:16:03 zecke: how's the compiling going? Dec 12 19:16:13 koen: feedreader2 should work Dec 12 19:16:40 is that in svn yet? Dec 12 19:18:27 koen: ok I am somewhat confused now, what you pasted - is that what causes my problem or is that what solves it? I did an update today so I should have this Dec 12 19:18:45 that show how to solve such problems Dec 12 19:19:24 koen: feedreader is in Dec 12 19:19:58 koen: well, but I should be able to see that someone is pulling in teh deb of external-toolchain when running with -D ? Dec 12 19:20:04 and I do not see it, that's what really confuses me Dec 12 19:20:15 Jin^eLD: no idea on that, sorry Dec 12 19:20:37 when I remove the .bb physically it's ok - then meta-toolchain builds, really weird Dec 12 19:20:57 well, I guess I'll have to talk to RP and florian Dec 12 19:21:17 zecke: is the kinetic scrolling bug fixed so feedreader2 can use that again? Dec 12 19:21:28 x11-image with eglibc comes up on wemu wow Dec 12 19:21:37 s/wemu/qemu Dec 12 19:21:38 koen: feedreader2 always used it?! Dec 12 19:22:02 keyboard is not working :* Dec 12 19:22:33 zecke: ah, then it was the browser that disabled it :) Dec 12 19:24:10 koen: don't ask for too much now :) Dec 12 19:26:11 * Crofton|home still hates painting Dec 12 19:30:19 koen: browser is updated as well, no kinetic scrolling Dec 12 19:30:55 otavio: hmm, the state is pretty bad and I might need to start over... Dec 12 19:31:18 zecke: well, many thanks for updating it to the new api Dec 12 19:36:03 koen: I need to update libmrss as well Dec 12 19:36:09 koen: but that is not critical Dec 12 19:37:04 * * OE Bug 3465 has been created by andrea.adami(AT)gmail.com Dec 12 19:37:06 * * zaurus-hinge cannot find bl (on c7x0) Dec 12 19:37:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3465 Dec 12 19:38:15 03koen 07org.oe.dev * r676667f8... 10/ (7 files in 5 dirs): webkit update: update webkit srcrevs, add midori 0.13, bump openmoko-feedreader2 and openmoko-browser2 Dec 12 19:38:18 zecke: http://lists.linuxtogo.org/pipermail/openembedded-devel/2007-December/003755.html Dec 12 19:39:04 koen: Thanks for the cluebat with -fno-exceptions and -fno-rtti Dec 12 19:39:22 koen: Does is ask for a new user initially when angstrom boots in X11 ? Dec 12 19:39:30 * Khem is booting in X11 for first time Dec 12 19:39:39 zecke: straight from webkit bugzilla :) Dec 12 19:39:57 Khem: is you built x11-image it should ask to create a new user Dec 12 19:40:51 s/is/if/ Dec 12 19:42:00 koen: http://www.flickr.com/photos/15140063@N04/1785562109/in/photostream/ Dec 12 19:42:24 "Oops! You don't have permission to view this photo." Dec 12 19:42:32 heh Dec 12 19:45:01 could you unmark it as 'private'? Dec 12 19:45:21 http://www.flickr.com/photos/15140063@N04/2105554761/ Dec 12 19:45:24 there we go :) Dec 12 19:45:41 its public Dec 12 19:46:36 koen, can you see it. somehow my keyboard and mouse is not working on that so I can not type in anything Dec 12 19:46:58 I haven't really tried qemu machine emu yet Dec 12 19:47:40 koen: ok Dec 12 19:47:52 Khem: http://www.rpsys.net/openzaurus/qemu/ Dec 12 19:48:26 Khem, is that a poky? Dec 12 19:48:31 dcordes: no Dec 12 19:48:34 qemu arm ox x86 ? Dec 12 19:48:35 I am using angstrom Dec 12 19:48:40 qemu arm Dec 12 19:48:55 Khem, which angstrom verison? Dec 12 19:48:57 ext2 x11-image Dec 12 19:48:59 2008.1 Dec 12 19:49:07 * Khem likes bleeding edge Dec 12 19:49:13 I should also try Dec 12 19:49:19 does it use matchbox2 stuff? Dec 12 19:49:29 not yet Dec 12 19:49:37 from what i know only a version of qemu user (and available) by poky emulate the touchscreen with the mouse so is usable for us ... Dec 12 19:49:49 qemu user -> qemu used Dec 12 19:50:26 I just need normal mouse and keyboard input on qemu Dec 12 19:50:35 I mimed the poky interface on my zaurus installing matchbox-desktop-2 matchbox-panel-2 but it's a pain because I run an ugly script for it Dec 12 19:51:32 mhh i used only qemux86 ... Dec 12 19:51:58 input ImExPS/2 Generic Explorer Mouse as /class/input/input1 Dec 12 19:52:28 AT raw set 2 keyboard as /class/input/input0 Dec 12 19:52:31 yes ... but 90% of angstrom stipt don't use it and instead use tslib ... Dec 12 19:52:46 are u pasting what ? Dec 12 19:53:00 gremlin[it]: the messages from my boot log Dec 12 19:53:28 zecke: http://webkit.org/blog/151/new-webkit-reviewer-2/ Dec 12 19:53:32 you're famous :) Dec 12 19:53:45 check if in /etc/profiles.d exist a tslib.sh or something like Dec 12 19:54:58 zecke: are your foss.in slides online? Dec 12 19:55:33 on the foss.in side Dec 12 19:55:50 yeah, I found them: http://foss.in/2007/register/speakers/talkdetailspub.php?talkid=343 Dec 12 19:55:51 koen: I'm sorry, I said you are a monster :) Dec 12 19:56:07 koen: you are famous too :) Dec 12 19:57:18 :D Dec 12 19:57:44 ~hail OE_devs Dec 12 19:57:45 * ibot bows down to OE_devs and chants, "I'M NOT WORTHY!!" Dec 12 19:58:19 zecke: your presentation looks familiar Dec 12 19:58:29 ;) Dec 12 19:58:39 koen: it can't though Dec 12 19:58:49 koen: the History part is 'new' and the Qtopia part as well Dec 12 20:00:01 koen: and I have used images Dec 12 20:01:09 gremlin[it]: etc/profile.d/tslib.sh is there Dec 12 20:01:21 gremlin[it]: it says TSLIB_TSDEVICE=/dev/input/touchscreen0 Dec 12 20:01:27 export TSLIB_TSDEVICE Dec 12 20:02:21 mhhh and /etc/X11/Xserver ... try to figure out if it use /dev/input/input[01] or /dev/input/touchscreen0 Dec 12 20:02:51 it should use /dev/input/touchscreen0 Dec 12 20:02:59 since /dev/input/touchscreen0 is under udev control Dec 12 20:03:07 ah ok ... Dec 12 20:05:42 yeah it checks -z "$TSLIB_TSDEVICE" Dec 12 20:06:15 how should touchscreen function on qemu Dec 12 20:06:24 through mouse ? Dec 12 20:06:47 only a version of qemu available on inhand site ... Dec 12 20:06:55 but i never try it Dec 12 20:07:01 the OE built qemu should have the touchscreen patches Dec 12 20:07:34 thats the one I am using Dec 12 20:07:50 the OE one Dec 12 20:41:25 morning all Dec 12 20:42:07 zecke: ok .. will you start it now? Dec 12 20:42:44 otavio: the machine is currently doing something... no idea if the trick helped or not Dec 12 20:50:36 otavio: seemed to have worked Dec 12 20:50:55 zecke: wow, nice! :-) Dec 12 20:51:22 also the cron job seems to work Dec 12 20:53:11 * cbrake gets mono 1.2.6pre4 to build and notes it fixes several critical issues on ARM ... Dec 12 21:26:33 cbrake: Ah, interesting. Did you see the the AOT stuff in after 1.2.7? http://www.mono-project.com/AOT Dec 12 21:32:27 CM: yes, that looks very interesting Dec 12 21:32:53 CM: mono is definitely a bit slow to start on ARM Dec 12 21:35:36 night guys Dec 12 21:36:39 night Dec 12 21:36:53 cbrake: I Dec 12 21:36:57 Sorry Dec 12 21:37:22 I'll try to get it installed on my neo soon using one of koens minimal images where there's enough space left :) Dec 12 21:37:33 CM: heh, sounds good Dec 12 21:37:58 CM: I'm pushing 1.2.6pre4 recipies shortly after I finish a bit of testing Dec 12 21:38:15 CM: recommend starting with that version Dec 12 21:42:22 Ok, thanks :) Dec 12 21:58:38 something is broken in X11 it does not recognise wacom usbdevice Dec 12 21:59:03 thats why I dont get touchscreen working on OE Dec 12 22:19:32 03khem 07org.oe.dev * rb777d230... 10/ (3 files in 2 dirs): Add glibc-* to equivalent eglibc-* packages in RPROVIDES. Dec 12 22:49:39 khem: do you recommend eglibc as a stable enough alternative for non-angstrom builds? Dec 12 22:51:13 hello Dec 12 22:59:30 03cbrake 07org.oe.dev * r0d12a42d... 10/ (8 files in 5 dirs): Dec 12 22:59:30 mono 1.2.6pre4: add new version to OE Dec 12 22:59:30 DEFAULT_PREFERENCE is set to -1 for this version Dec 12 22:59:30 until 1.2.6 is released. Dec 12 23:02:06 when using bb threads > 1 and DEPENDS Dec 12 23:02:41 it seems that each thread does note communicate with the others Dec 12 23:03:30 I mean I have package A that depends on B. one thread starts to build B while the other builds A like B was already compiled and staged Dec 12 23:04:12 an thus it fails. After running the same build-package-A command a few times finally the stage phase of B was done and A compiled Dec 12 23:04:47 is this a known problem? I'm using mokomakefile and building OE packages for the Neo1973 Dec 12 23:10:39 Fique: Right now it will be same for angstrom and non angstrom I think Dec 12 23:10:58 Fique: I encourage you to do so and find issues if any Dec 12 23:11:08 and report here we will fix them Dec 12 23:21:28 Ghiottone: what version of bitbake are you using? Dec 12 23:22:50 hvontres|work, 1.8.8 Dec 12 23:26:16 Ghiottone: hmm, not sure then. any particular packages that cause this? Dec 12 23:27:22 hvontres|work, no. I have seen this behavior with other packages Dec 12 23:28:06 Ghiottone: Next time you hit this, you might want to file a bug Dec 12 23:29:00 I think it is easy to reproduce with a clean build and a package that depens on another and uses a .h from it Dec 12 23:33:29 Ghiottone: I do clean build almost every day never ran into such an issue Dec 12 23:33:50 I use bitbake 1.8.8-svn branch Dec 12 23:34:30 hfsplusutils is an example of this, which fails to finish off rpm2cpio-native first Dec 12 23:35:43 kinda Dec 12 23:36:01 well a similar issue Dec 12 23:36:20 with bitbake 1.8.8 too Dec 12 23:36:45 does the clean operation removes header files installed in the stage phase? **** ENDING LOGGING AT Thu Dec 13 02:59:56 2007