**** BEGIN LOGGING AT Thu May 17 02:59:57 2007 May 17 03:38:47 03lenehan 07org.oe.dev * re5ed7f26... 10/ (3 files in 3 dirs): May 17 03:38:47 blueprobe 0.18: Fix for uclibc. Termios speed above 115k are not supported, May 17 03:38:47 so check for the defintion of the faster speeds before trying to use them. May 17 03:38:51 03lenehan 07org.oe.dev * rfb9c7c92... 10/ (3 files in 3 dirs): May 17 03:38:51 bluez-utils 3.9: Fix for uclibc. Termios speed above 115k are not supported, May 17 03:38:51 so check for the definition of the faster speeds before trying to use them. May 17 04:25:14 03lenehan 07org.oe.dev * r64fe11e0... 10/ (5 files in 3 dirs): May 17 04:25:14 grep 2.5.1: Fix for uclibc. Grep uses __mempcpy which glibc provides but May 17 04:25:14 uclibc doesn't. Since both glibc and uclibc provide the mempcpy version use May 17 04:25:14 that instead. May 17 06:44:32 morn :) May 17 06:53:34 morning May 17 06:55:00 morning gentlemen, does anyone know a way to make bitbake ignore a different md5? May 17 06:56:38 it's really killing me, the source file and the .md5 file both have the same md5 but bitbake still wants another md5 that its getting from *somewhere* May 17 06:56:52 check conf/checksums.ini May 17 06:57:19 I dont seem to have that file May 17 06:57:33 jus tconf/local.conf May 17 06:57:52 OE/conf/checksums.* then? May 17 06:58:02 nope, afraid not May 17 06:58:36 really? May 17 06:58:51 conf/checksums.ini is present in OE metadata May 17 06:59:51 wait, I found one in linux/conf/ May 17 07:00:09 and in org.openembedded/conf/ May 17 07:11:34 hrw: thank you, changing that file fixed it May 17 07:11:49 this has been with me since last night and all today May 17 07:12:06 post failed entry to bugtracker May 17 07:12:20 ah, it's a bug of my doing May 17 07:12:30 I was playing with the source file May 17 07:32:13 hi koen|away May 17 07:42:17 hey hrw May 17 07:43:06 ERROR: QA Issue: non -dev package contains symlink .so: qte-mt path '/work/tosa-angstrom-linux-gnueabi/qte-mt-2.3.10-r13/install/qte-mt/opt/QtPalmtop/lib/libqte-mt.so' May 17 07:43:19 yes, qte is broken May 17 07:43:26 there is a bug for that somewhere May 17 07:43:51 koen that was you morning wakup call! May 17 07:43:54 hi May 17 07:44:49 :) May 17 07:44:53 good morning all May 17 07:46:39 !oebug1785 May 17 07:47:01 should be easy to fix, shouldn't it May 17 07:49:13 !oebug 1785 May 17 07:49:30 * * Bug 1785, Status: NEW, Created: 2007-01-12 14:36 May 17 07:49:31 * * : ERROR: QA Issue: non dev package contains .so: qte-mt May 17 07:49:32 * * http://bugs.openembedded.org/show_bug.cgi?id=1785 May 17 08:02:16 http://blog.haerwu.biz/archives/ - I found new plugin May 17 08:04:43 hrw: if I want to tackle the tosa suspend/resume bug, where/which file should I start with May 17 08:05:02 hrw: i'm totally new to this, and the ipaq-sleep package May 17 08:05:13 I'm having a problem compiling module-init-tools, http://pastebin.ca/492537 any suggestions? May 17 08:05:28 xjqian: remove usb modules May 17 08:06:36 xjqian: some of the debugging output is found in ohci-hcd.c that's where I've been looking May 17 08:07:01 but bitbakes been giving me the runaraound for the past 12 hours! May 17 08:07:35 regarding ohci-hcd.c May 17 08:07:42 AsLan^: I've seen this quilt problem myself, can't remember which package I'm baking May 17 08:07:44 morning May 17 08:07:50 xjqian: most of problems came from usb host driver which lack suspend/resume May 17 08:08:01 I see May 17 08:08:30 hrw: looking at the code... the ohci_usb_reset function is called, which leads to more errors May 17 08:08:58 thanks. I will look into ohci-hcd.c May 17 08:09:04 what about power.c? May 17 08:09:08 03koen 07org.oe.dev * r1a21079a... 10/ (1 packages/gcc/gcc4-build-sdk.inc): gcc4-build-sdk: fix USE_NLS May 17 08:09:15 03koen 07org.oe.dev * re7f1b733... 10/ (1 classes/insane.bbclass): insane.bbclass: fix sdk builds May 17 08:09:17 this is the patch im working on http://pastebin.ca/492538 May 17 08:10:37 but can someone help me unscrew my OE environment! May 17 08:11:16 teh output says something about "package module-init-tools-3.2.2-r3" but I don't have that file May 17 08:12:35 AsLan^: looks like you have weird quilt version May 17 08:13:33 version 0.45 May 17 08:14:04 AsLan^: I don't remember what's the solution for mine, maybe rm -rf tmp May 17 08:14:18 and rebuild for another three hours! May 17 08:14:23 AsLan^: and starting from scratch May 17 08:14:26 I'm not ready for that just yet, already done it twice May 17 08:15:09 AsLan^: I've never had that problem for a while May 17 08:15:35 AsLan^: maybe -c rebuild quilt-native May 17 08:15:53 I'll give that a try May 17 08:16:31 AsLan^: try 'bitbake quilt-native' May 17 08:17:44 alrighty, tried both but it still doesn't work May 17 08:17:54 earlier.. I made a mistake May 17 08:18:11 I accidently did a bitbake in my OE directory rather than OE/build May 17 08:18:20 this seems to be part of the problem May 17 08:20:49 AsLan^: at least thats easy to undo May 17 08:22:12 well... I basically just moved the new tmp dir into build May 17 08:22:31 * koen recommends setting TMPDIR in local.conf May 17 08:22:33 I'm guessing theres some temp configuration files somewhere that still have the old path May 17 08:22:36 AsLan^: that definitely won't work May 17 08:22:56 koen: I'll definitely take a look May 17 08:22:57 AsLan^: I've tried that and learned a hard lesson May 17 08:23:03 :D May 17 08:23:07 yay hard lessons May 17 08:23:23 was there an easy way to do it? May 17 08:23:50 AsLan^: seems you have to rm -rf anyway;) May 17 08:24:11 well, I"m going to first just try removing the offending packages from aarmvte dir May 17 08:24:21 ~lart gnu for using private symbols May 17 08:24:21 * ibot drops a baby grand on gnu for using private symbols May 17 08:29:50 people are really quite today, I've yet received any mail from any oe, opie, gpe lists... May 17 08:30:11 christian holiday May 17 08:32:35 koen: pushed something for you May 17 08:32:36 its because someone trademarked the world, we cant send email anymore May 17 08:33:32 03koen 07org.oe.dev * r57252914... 10/ (1 classes/sdk.bbclass): sdk.bbclass: also package debug files May 17 08:33:37 03koen 07org.oe.dev * rb9c37660... 10/ (1 packages/binutils/binutils-cross-sdk_2.17.50.0.5.bb): binutils-cross-sdk: add 2.17.50.0.5 May 17 08:34:09 2.17.50.0.5.. I forgot to merge it from poky.. May 17 08:34:24 the sdk stuff makes my eyes bleed May 17 08:34:47 not only yours May 17 08:36:03 I'd like to hand people a qemu image that includes and sdk, but that doesn't work for armv6 May 17 08:36:23 armv6 is not in qemu... May 17 08:39:37 koen: well if you already had the SDK pain, time to feel the real pain in mozilla code, when you finished that you will be a first class massochist May 17 08:40:08 My patches for mozilla are already upstream.... May 17 08:40:22 libtool: link: cannot find the library `/data/build/koen/OE/build/tmp/angstrom/work/armv5te-angstrom-linux-gnueabi/libxcb-0.9.93-r2/libxcb-0.9.93/src/libxcb.la' May 17 08:40:32 that's on the target May 17 08:41:17 hrw: has minimo recipe and patches been committed? May 17 08:41:36 hrw: I can't find packages/minimo May 17 08:41:58 xjqian: packages/mozilla/minimo May 17 08:42:08 ugh, thanks May 17 08:42:21 koen: does it work yet? May 17 08:42:38 XorA: the sdk? May 17 08:42:45 mozilla May 17 08:42:55 people claim minimo works May 17 08:43:03 koen: it almost does May 17 08:43:14 koen: so back to the wheel of woe for you :-D May 17 08:43:27 noooooooooooooooooooooooooooo May 17 08:43:38 but then the only browser we have that ssl works in is links :-) May 17 08:43:52 so its not working in pretty much the same way all the rest not work May 17 08:44:10 although I didnt try that new konqueror for opie May 17 08:47:18 how do I undo something with git? May 17 08:47:32 'git revert' is the most braindead command ever May 17 08:49:29 * koen restores the file manually from a tarball May 17 08:53:13 xjqian: gah an entire day wasted with this crap! I've gone and removed tmp/ and starting again :( - any luck poking around ohci-hcd.c? May 17 08:53:14 koen: git reset May 17 08:53:26 koen: oh actually git checkout May 17 08:53:39 koen: git was designed by cvs users remeber May 17 08:53:55 s/designed/flug together in 10 mins/ May 17 08:55:33 03hrw 07org.oe.dev * r8390c5b6... 10/ (3 files in 2 dirs): powertop: update to 1.2 May 17 09:26:20 Koen: did you see my mail to knutt ? May 17 09:49:08 gerwinin: the one from a few days ago? May 17 10:26:30 hi all ... i have a question about the poky image for qemux86 dod someone try it ? May 17 10:27:45 gremlin I played with the scripts to create the qemu image May 17 10:29:04 just a stupid question probably ... it boot and after a few it ask to calibrate the touchscreen ... but the mouse seem to not work as simulation of touch ...how i can do _ May 17 10:29:06 ? May 17 10:29:53 you need qemu from OH May 17 10:29:54 boot : i mean qemu boot ... May 17 10:30:09 yes ye sunder qemu :) May 17 10:31:17 basically the question is ... can qemu simulate a touchscreen with the mouse ? May 17 10:37:17 gremlin[it]: OH qemu do that May 17 10:38:01 seem no ... i try but i stay stucked at touchscreen calibration :( May 17 10:38:53 gm May 17 10:38:54 no no wait ... the qemu packeged by OH ??? ... i have the one provided by Fedora Core 6 ... maybe that is ... May 17 10:38:59 gm May 17 10:39:04 morning likewise May 17 10:41:19 hi gremlin[it], wookey_ May 17 10:44:20 gremlin[it]: Angstrom images built for a non-accel fb work well in qemu/integrator - both x11 and opie May 17 10:44:31 gremlin[it]: I run h4000 images in such way May 17 10:45:28 psokolovsky_, i woul'd just have a look at poky work ... and at now (for work) i'm working on x86 platform :) May 17 10:45:33 morning psokolovsky_ ! May 17 10:45:43 hi! May 17 10:46:52 well, again - angstrom has all needed setup to work with moust instead of touchscreen May 17 11:43:44 * wick720 is away: sl33p May 17 11:51:23 hi May 17 11:52:45 I'm getting this error the latest angstrom test release with ipkg -> ipkg: fork failed May 17 11:53:14 on update or install (remote or local one) May 17 11:55:42 aswell during flashing, my home directory has been deleted... :-( May 17 12:16:58 hello zecke May 17 12:17:52 ho May 17 12:18:21 koen|away: I'm now at pciutls May 17 12:22:55 * chouimat kicks perl-native for not compiling May 17 12:23:51 chouimat, I have that problem to! May 17 12:24:07 How do you "kick" it? May 17 12:24:48 step2back: don't know :) May 17 12:24:55 to! = too! May 17 12:26:05 Do you out perl-native from buildings packages? May 17 13:06:28 hi, all. I have a error while building opie-image May 17 13:06:37 ERROR: Build of org.openembedded.dev/packages/perl/perl-native_5.8.8.bb do_compile failed May 17 13:07:06 no bugs added about it... :( May 17 13:08:52 Is it possible to do manual #Making x2p stuff May 17 13:08:55 make[1]: Entering directory `tmp/work/i686-linux/perl-native-5.8.8-r7/perl-5.8.8/x2p' May 17 13:08:55 You haven't done a "make depend" yet! May 17 13:15:14 step2back: you should not need any manual actions. rebuild, then file bug if that doesn't help. May 17 13:16:23 , how to rebuild? May 17 13:16:51 step2back: -c rebuild May 17 13:17:04 for perl-native, of course May 17 13:17:53 bitbake perl-native -c rebuild - do you mean? May 17 13:19:17 Hi all, is there any known reason why compiling of freetype-native-2.2.1-r2 fails ? May 17 13:19:48 no no no... I executed "make depend" in tmp/work/i686-linux/perl-native-5.8.8-r7/perl-5.8.8/x2p. Building continue! May 17 13:26:27 I think it's needed voodoo for building OE May 17 13:29:06 step2back: no, it is not. all you need is just build it right, report issues right, and fix them right. May 17 13:29:21 Actually, ond more thing would be helpful - continuous integration. May 17 13:30:00 I read something about links in the OE devlopement path. Since I use different computers I set-up the OE environment on a mobile disk and set a link to /usr/local/OE. Can that create problems ? May 17 13:30:46 Zuse: yes, it will. May 17 13:31:33 psokolovsky_: O.K. ;) May 17 13:31:57 , I have not such experience for that. May 17 13:32:12 step2back: get it, it's free! ;-) May 17 13:32:15 cool Running task 98 of 3897 May 17 13:34:11 some people use a $TOPDIR approach to run their configs against. Can anyone explain how that would work? May 17 14:44:04 What i have to do if I like to reset the OE environment (it looks broken due to some problems with harddisk space). I like to avoid that I have to download every source again. May 17 14:44:53 Zuse: remove tmp/work tmp/staging tmp/cross tmp/stamps... so basicly everything but $DL_DIR May 17 14:45:08 Zuse: rm -rf tmp will remove all the semi-compiled stuff May 17 14:46:11 NAbyss_: But that's included to delete 3 G of data. Any way to rescue something ? May 17 14:49:41 zecke: Sorry just didn't saw your lines. Thanks i will try May 17 14:49:48 Zuse: What explicitly do you want to rescue? It should automagically(TM) rebuild what's not good.. May 17 14:50:02 tmp/deploy is probablyt the only thing worth keeping May 17 14:51:38 NAbyss_: Thanks, no I just like to keep the downloads since I'm at home now and my broadband connections is very slow. I guess I just have to keep $DL_DIR as zecke suggest May 17 14:51:55 Zuse: Sources aren't in tmp/ afaik May 17 14:52:03 At least they aren't here May 17 14:53:37 NAbyss_: just had to check build/conf/local.conf ;) May 17 14:53:53 ah :) May 17 14:55:44 NAbyss_: I just wonder, if the download folder is external, what does blow up the /tmp/work folder so much ? May 17 14:56:35 Zuse: look there and you will see May 17 14:57:02 psokolovsky_: O.k. i will look there ;) May 17 14:57:39 Zuse: tmp/work contains the compiled software and unpacked tarballs May 17 14:58:12 NAbyss_: Anyhow now it's gone ... :) May 17 14:58:40 Basic workflow is source -> tmp/work -> tmp/staging -> tmp/deploy May 17 15:02:07 NAbyss_: Yes thanks, I think I start to understand. Just start to "play" with OE more intensive and much stuff is still confusing me. May 17 15:02:29 Zuse: I haven't got into it that deeply myself.. but it's good once you get used to it :) May 17 15:04:15 03pfalcon 07org.oe.dev * r9caf4879... 10/ (8 files in 8 dirs): May 17 15:04:15 linux-handhelds-2.6: Update defconfigs, from defconfigman r230. May 17 15:04:15 * Also htcuniversal defconfig, contributed by tsdogs (#htc-linux). May 17 15:05:27 holtmann: hey, do you test bluez-utils with uclibc? May 17 15:18:20 03freyther 07org.oe.dev * re41a7519... 10/ (1 packages/pciutils/pciutils-2.2.4/configure-uclibc.patch): packages/pciutils/pcituils-2.2.4: Fix configure for gnueabiuclibc builds May 17 15:22:39 holtmann: do you know if I can make AC_CHECK_FUNC include another header for me? May 17 15:22:43 psokolovsky_: do we need to merge? May 17 15:22:50 zecke: No. May 17 15:23:30 zecke: ??? mtn? i always merge before commit, sorry if it slipped May 17 15:23:37 holtmann: the NEED_PPOLL test fails on my uclibc build, ppoll is declared in sys/poll.h on uclibc May 17 15:23:44 psokolovsky_: well, I might be guilty ;) May 17 15:24:34 zecke: That can be fixed. May 17 15:24:58 holtmann: well, let me take a closer look at the config.log May 17 15:25:25 There is a way to make it include sys/poll.h, but I can't remember how. May 17 15:25:41 Post a patch to bluez-devel and I include it. May 17 15:25:44 holtmann: well, I assumed it failed to compile, it failed to link though May 17 15:25:53 holtmann: will do when it is not an OE error May 17 15:26:56 it looks like an OE error May 17 15:33:48 holtmann: well. uclibc has a ppoll declaration in sys/poll.h, ARM looks like it is not defining a ppoll syscall so it is missing from the libc.so and bluez-utils detects that it has to use the stub but then gcc complains about the different prototypes (static vs. non-static) May 17 15:34:12 holtmann: option a) remove static from the ppoll.h stub and hope it will link May 17 15:34:32 holtmann: b) move the implementation to ppoll.c May 17 15:34:35 To I added a --disable-ppoll by any chance. May 17 15:34:46 holtmann: c) make ppoll a define May 17 15:38:54 holtmann: sorry? May 17 15:42:45 anyone used buildbot? May 17 15:43:04 Crofton: RP does for poky May 17 15:43:28 I am trying to set it up for some stuff here and it confuses me :) May 17 15:44:04 Crofton: well it is using twised so this is normal... May 17 15:44:31 heh May 17 15:47:11 holtmann: http://rafb.net/p/0FKkka12.html the patch with the least effort May 17 15:48:08 Does this solve your issue? May 17 15:48:33 holtmann: it compiles and links May 17 15:49:36 holtmann: Short summary on the issue. NEED_POLL is 1 as libc.so has no ppoll symbol but sys/poll.h has a ppoll prototype. So when including common/ppoll.h and sys/poll.h gets included as well gcc will complain May 17 15:49:54 Does it actually work? May 17 15:50:25 holtmann: I can probably tell you in a week or two (depending on my time) :) May 17 15:50:31 holtmann: or work as in compiles? May 17 15:50:50 I think that I lost my system without ppoll support. So I can't test it. May 17 15:51:13 holtmann: use your N770? May 17 15:52:31 I have the hacker edition running on it. May 17 16:12:46 robtaylor: did you find the pxa-dev-manual? May 17 16:12:53 I have a copy on me now May 17 16:15:59 wookey_: no, i havn't found it yet May 17 16:17:01 wookey_: your're brilliantm=, thanks! May 17 16:17:16 uploading.... May 17 16:17:47 it's huge May 17 16:20:15 * robtaylor does a df May 17 16:21:25 oops - uploaded to mahcine you can't see. let me try that again May 17 16:22:02 doh =) May 17 16:22:59 morn koen May 17 16:28:02 hi koen May 17 16:28:16 hi HopsNBarley May 17 16:28:46 hi likewise. how's it going with uclibc? May 17 16:29:18 robtaylor: http://balloonboard.org/hardware/300/ds/PXA270-dev-manual.pdf May 17 16:30:10 wookey_: you rock May 17 16:32:35 anybody working with Marvell Orion? May 17 16:32:56 wookey_: i owe you a beer or somesuch :) May 17 16:33:27 HopsNBarley: all credit to koen and a lot of others. May 17 16:33:59 HopsNBarley: 0.9.29 seems to be fairly easy on us, no true problems yet. May 17 16:34:14 hey all May 17 16:34:43 vvv May 17 16:37:36 likewise, great news. May 17 16:38:06 he koen May 17 16:38:06 hi koen. May 17 16:38:11 does atmel have offices in sunny places? May 17 16:38:18 koen hehe May 17 16:38:25 hey woglinde & HopsNBarley May 17 16:38:28 koen: probably yes :-) May 17 16:38:37 i know sunny places with lots of offices (-; san jose has 300+ days of sun a year. May 17 16:39:33 hopsnbarley hm very hot there May 17 16:40:07 03likewise 07org.oe.dev * r1d3cbf1f... 10/ (5 files in 4 dirs): helloworld: package plus image for truly minimal user space, a statically linked hello world executable. May 17 16:40:08 HopsNBarley: we should really do a OE startup then :-) May 17 16:40:28 * woglinde wants uclibc for oabi "duck and hide" May 17 16:40:39 that's what i'm doing!!! May 17 16:41:01 woglinde: DISTRO=angstrom-2007.1-oabi ANGSTROM_MODE=uclibc MACHINE=simpad May 17 16:41:14 woglinde, only about 2 or 3 weeks each summer. really. i moved from florida. it's hot *there*. May 17 16:41:26 hopsnbarley hehe okay May 17 16:41:27 koen: libnss-mdns is currently failing :) May 17 16:41:47 http://people.openezx.org/pmarques/pictures/rokr_e6_booting.jpg May 17 16:42:56 Marvel orion anybody? May 17 16:43:08 HopsNBarley: nope, orion = ? May 17 16:44:07 Orion is Marvel SOC, Arm 9 based i think. Got USB2.0, Sata, cheap. need code though. May 17 16:44:41 I saw a patch for orion on LAK iirc May 17 16:46:16 i'll check archives. thanks koen. May 17 16:46:32 HopsNBarley: I really liked Marvell's SATA controller (working with it in our industrial PVR system). Very high performant, SMART and hotplug working earlier than libata/linux did. May 17 16:47:09 SMART for SATA? May 17 16:47:25 HopsNBarley: yup, I added SMART support for SATA in the marvell driver and the smartmontools package even May 17 16:47:28 hopsnbarley yes May 17 16:47:37 libata now also has it May 17 16:47:39 sata has smart support May 17 16:48:14 hey, that's great news. i must have been working with a controller that didn't support it. May 17 16:48:17 smartmontools support for marvell is upstream (-d marvell). May 17 16:48:34 HopsNBarley: which controller? which controller is in the orion? (googling...) May 17 16:48:50 it's integrated in the SOC. dunno which one. May 17 16:51:55 koen: why do we need libnss_mdns? this won't ever work on uclibc ;) May 17 16:53:22 Can I disable ipkg ending up in my rootfs and image? May 17 16:53:30 03freyther 07org.oe.dev * r3072b3ae... 10/ (3 files in 3 dirs): May 17 16:53:30 packages/bluez/bluez-utils: angstrom uclibc fixlet. Make custom ppoll not clash with sys/poll.h May 17 16:53:30 uclibc on ARM has no ppoll symbol but a declaration in sys/poll.h. Give May 17 16:53:30 the bluez stub another name to avoid compile errors. May 17 16:53:34 zecke: dunno May 17 16:53:35 03freyther 07org.oe.dev * r386f0541... 10/ (3 files in 3 dirs): May 17 16:53:35 packages/strace: Apply patch from uclibc buildroot to fix compilation uclibc angstrom May 17 16:53:35 Without that patch the Angstrom uclibc EABI build fails. Patch was taken May 17 16:53:36 from the buildroot of uclibc. May 17 16:55:12 koen: do we have an override for uclibc? May 17 16:55:38 _linux-uclibc and _linux-uclibcgnueabi May 17 16:55:42 aka target_os May 17 16:55:54 likewise: could you restore angstrom-minimal-image? May 17 16:56:06 koen: restore? May 17 16:56:22 koen: need to override RRECOMMENDS of avahi to not depend on libnss-mdns May 17 16:56:58 likewise: there are 2 hello-world images now :) May 17 16:57:15 zecke: see the gdb recipe May 17 16:57:56 koen: thanks May 17 16:58:21 likewise: see the bottompart of http://lists.linuxtogo.org/pipermail/openembedded-commits/2007-May/005557.html May 17 16:58:45 koen: I'll reject my commit, probably hit save too soon in my editor... May 17 17:00:41 koen: I should review the mtn command string needed to view outgoing pushes... May 17 17:01:40 fwiw, the commit message show the affected files as well May 17 17:01:51 message-editor, that is May 17 17:02:43 koen: using -m... May 17 17:03:51 wow I need to buy the "magic opener" May 17 17:04:09 likewise: I like your do_fetch for helloworld.bb :) May 17 17:04:14 :-) May 17 17:09:34 koen hm dont work for simpad May 17 17:09:41 it will build glibc May 17 17:09:58 woglinde: well May 17 17:10:04 woglinde: user error ;) May 17 17:10:41 zecke if you think May 17 17:12:57 angstrom-console-image-1.0: completed yipee May 17 17:13:07 hm angstroem-oabi.conf has TARGET_OS = "linux" ;) May 17 17:13:16 woglinde: hehe May 17 17:13:23 zecke which architecture? May 17 17:13:35 woglinde: Neo uclibc May 17 17:13:39 hehe May 17 17:15:01 hm May 17 17:15:02 TARGET_OS = "linux-uclibcgnueabi" May 17 17:15:53 not EABI May 17 17:16:03 hehe May 17 17:16:08 thats what oe makes May 17 17:16:19 clever May 17 17:17:24 hm thats needs some love May 17 17:20:00 qemu over remote X feels like a real neo :} May 17 17:22:33 woglinde: hi May 17 17:23:01 woglinde: you try to build with uclibc instead glibc, right? May 17 17:25:52 woglinde: and uclibc build should give smaller binarys? May 17 17:41:24 mickey|moving: thanks for the reply May 17 17:46:34 hey all May 17 17:46:38 koen: what is the status on icu? May 17 17:46:54 is mtn 0.31 supported? May 17 17:47:05 mreimer: yes May 17 17:47:14 mreimer: supported by who/what? May 17 17:47:37 * mreimer sheepish grin May 17 17:48:02 Can I use the OE db with mtn 0.31? May 17 17:48:35 mreimer: http://www.openembedded.org/snapshots/ May 17 17:48:36 I get "mtn: misuse: layout of database /home/mreimer/oe OE.mtn doesn't match this version of monotone" May 17 17:48:50 likewise: so get the 0.31 snapshot from Nov and then update it? May 17 17:48:55 mreimer: corect May 17 17:48:57 +r May 17 17:49:00 thanks likewise May 17 17:49:15 and thanks for the bad question smackdown, zecke :-) May 17 17:49:22 * mreimer was lazy May 17 17:49:43 zecke: same as before (icu) May 17 17:49:46 mreimer: You can always run the statically compiled mtn 0.33 or even 0.34 from monotone.ca. There have been numerous speedups since 0.31 May 17 17:50:18 likewise: the machine I'm using for this is running Ubuntu feisty fawn, which only has 0.31 May 17 17:50:33 and git 1.4 :-( May 17 17:50:35 mreimer: me too May 17 17:50:56 likewise: do you know of some newer mtn packages for ff? May 17 17:51:02 mreimer: so, I uninstalled monotone 0.31, then installed the statically compiled binary (0.34) from monotone.ca and use that May 17 17:51:13 likewise: ah I see. thanks for the tip May 17 17:51:20 mreimer: no, I used to build it myself, these days I just run the static exe May 17 17:51:29 it saves me much hassles May 17 17:51:36 (building boost lib etc) May 17 17:51:51 likewise: good idea. Is 0.35 ok, or should I stick with 0.33/0.34? May 17 17:52:01 0.34 works for me (tm) May 17 17:52:12 mreimer: it is okay :) May 17 17:52:12 0.35 is nice, it changes the text in the commit-message editer to include the branch and less fluff May 17 17:52:19 easier to spot mistakes :) May 17 17:52:28 * mreimer wget mtn-0.35 May 17 17:52:29 :-) May 17 17:52:52 mreimer: the db migrate takes ~30 seconds on OE.mtn with 0.35, it just needs to add an index to a table May 17 17:52:54 I thought I disapproved my commit and committed that disapproval... still waiting for it to show... May 17 17:59:35 koen: this box of Todd's has 2 Xeon 5110 (so 4 x 1.6 GHz cores), which ought to make a decent OE builder May 17 18:00:53 mreimer: btw, the TI BTS patch doesn't apply to bluez-utils 3.10 May 17 18:01:43 koen: bummer. Do you know if it is wildly broken, or just a little? May 17 18:02:11 there is support for some more TI radios in there now, so I can't say May 17 18:02:34 mreimer: 3.10 is wildly different from 3.9 (in a good way) May 17 18:06:19 koen: that's good to know May 17 18:15:57 likewise: do you know where I could get newer git packages for ff? I see gutsy has 1.5+ May 17 18:16:13 mreimer: no, I don't use git much (yet) May 17 18:16:34 mreimer: I *do* have a git question, though May 17 18:18:20 likewise: fire away May 17 18:18:29 mreimer: though I'm no expert May 17 18:18:51 mreimer: I'm cloning upstream linux, then fetch another branch from another git tree. May 17 18:19:41 mreimer: I would like to specify a specific version of that branch when checking out. Is specifying one id enough to do so? May 17 18:20:38 likewise: git-checkout -b mynewbranch v2.6.16.47 May 17 18:20:43 likewise: something like that? May 17 18:21:41 likewise: that would create a new branch named mynewbranch based off the 2.6.16.47 tag on the stable branch May 17 18:25:39 likewise: I think if you don't need to create a new branch, you can just "git-checkout v2.6.16.47" May 17 18:28:38 koen: I'm following your "Building Angstrom" instructions, and I see that you use bitbake 1.6.0. Is that the recommended version? May 17 18:29:11 mreimer: no May 17 18:29:21 zecke: 1.8? May 17 18:29:24 mreimer: yes May 17 18:29:28 thanks May 17 18:30:47 mreimer: I thought I removed all 1.6.x references... May 17 18:31:02 koen: "tar zxf /path/to/bitbake-1.6.0.tar.gz" May 17 18:31:13 koen: "PATH=/OE/bitbake-1.6.0/bin:${ORG_PATH}" May 17 18:31:27 crap, I didn't press "submit" May 17 18:31:50 the thought counts :-) May 17 18:34:11 mreimer: is the rendering problem still there after an ipkg upgrade? May 17 18:34:32 koen: which ipk? gtk+? May 17 18:35:12 all May 17 18:35:20 'ipkg update ; ipkg upgrade' May 17 18:35:41 koen: I'll try it. (The image I'm using I built myself.) May 17 18:35:41 03likewise 07org.oe.dev * r2f7cdf8e... 10/ (6 files in 6 dirs): angstrom-minimal-image: revert broken commit. May 17 18:35:47 03likewise 07org.oe.dev * r161f25ff... 10/ (5 files in 4 dirs): disapproval of revision '1d3cbf1f69da57387c0d0eb4329fc5054a9ccec0' May 17 18:40:33 gremlin[it]: You need a patched qemu for touchscreen calibration May 17 18:49:53 likewise: ehm... eh... is revision 2f7cdf8eacd78e273744583cd918480d48f69dc8 what you intended to commit? May 17 19:01:53 is this anything to be concerned about: "NOTE: preferred version 2.5 of glibc not available (for item virtual/arm-angstrom-linux-gnueabi-libc-for-gcc)" May 17 19:02:35 thanks RP :) May 17 19:03:25 mreimer: that one is harmless May 17 19:03:31 thanks koen May 17 19:04:04 RP: got all the Intel pxa27x infos. Do you have any specific information about how the Z handle the suspend ? May 17 19:05:15 koen, why does it do that? May 17 19:05:24 inquiring minds want to know :) May 17 19:10:54 Crofton: it evaluates every glibc recipe to check what satisfies glibc-intermediate (aka virtual/arm-angstrom-linux-gnueabi-libc-for-gcc) May 17 19:11:09 Crofton: and glibc_2.5.bb doesn't provide virtual/arm-angstrom-linux-gnueabi-libc-for-gcc May 17 19:11:30 should we make it do that? May 17 19:11:33 glibc_2.ancient does May 17 19:11:38 I know it confuses a lot of people May 17 19:11:48 Crofton: no, we want glibc-intermediate :) May 17 19:11:59 * Crofton does not understand this intermediate stuff May 17 19:12:10 the solution is to delete everything except glibc*2.5.bb ;) May 17 19:12:36 heh May 17 19:12:54 anyway, I need to go back to working what set of revs to run to collect some data May 17 19:13:05 svn is in a bad state here right now May 17 19:13:13 HEAD isn't really working May 17 19:17:08 03koen 07org.oe.dev * r093b89e2... 10/ (6 files in 6 dirs): disapproval of revision '2f7cdf8eacd78e273744583cd918480d48f69dc8' May 17 19:17:35 pH5, XorA|gone: http://rafb.net/p/IVX4tv70.html May 17 19:20:32 Zuse: Not much other than whats in the code May 17 19:21:08 Zuse: The most interesting bit is that when charging the battery the zaurus wakes up, does things and then sleeps again every 10 minutes May 17 19:21:29 Zuse: Other than that, its the standard pxa27x code May 17 19:23:16 RP: OE isn't a distribution :) May 17 19:23:34 RP: could you change the description on pokylinux.org? May 17 19:24:25 koen: Ah, at the bottom? I didn't write that :) May 17 19:25:08 RP: you might want to link to http://www.openembedded.org/user-manual as well May 17 19:26:56 koen: I can't change this atm, I'll try tomorrow May 17 19:30:47 * RP wonders if he dare look at the sync status between OE and poky May 17 19:32:05 RP: Can you help me and point a bit to the code ? Just wonder which of the idle modes are used and wether it will change with increasing idle time. May 17 19:32:33 RP: hrw looked at it a bit May 17 19:32:51 RP: every now and then he hints that we need to merge something from poky :) May 17 19:33:00 dunno about OE -> poky, though May 17 19:35:45 just to switch in, sorry i still try to learn and to understand :). What actually is poky ? May 17 19:36:59 Zuse: http://pokylinux.org/ May 17 19:37:56 Zuse: There is only one idle mode used and a choice of two suspend modes May 17 19:39:18 Zuse: arch/arm/mach-pxa/pm.c is the suspend code, or at least part of it, the rest links from there May 17 19:39:33 koen: I could give hints about new things to merge :} May 17 19:40:19 koen: Have you looked at nokia800 support? May 17 19:40:30 koen: Thanks for the link. So, OE is a platform for distribution devlopement of embedded systems and OZ, Angstrom and poky are distributions created with the help of OE ? May 17 19:40:32 koen: I can't use "ipkg" with the latest angstrom test image (12/05/2007), I'm getting "fork failed" May 17 19:41:08 RP: Thanks for the pointer I will have a look to it. Still like to solve this resume problem on akita. May 17 19:41:52 Zuse: yes May 17 19:42:08 cedricb: that means "Need more RAM" May 17 19:42:30 :) hm uclibc is now fixed May 17 19:42:36 RP: I haven't had time yet to try a new image on my n800 May 17 19:42:46 will post the patch to bugtracker May 17 19:42:51 koen: The xserver in poky works properly now May 17 19:43:16 ah, cool May 17 19:44:13 koen: any change between April image? was working before... and by flashing it has erased my home partition or /home/root May 17 19:44:53 cedricb: I suspect the feeds got bigger, so ipkg needs more ram May 17 19:46:49 any other alternative to reduce it? but if I try to install a local ipk, I'm getting the same thing May 17 19:51:27 koen: 0.75MB of diff with OE May 17 19:53:23 RP: partly because I tend to run poky recipes through oe-stylize.py May 17 19:54:16 cedricb: add a swap partition May 17 19:54:19 or is that with --ignore-whitespace? May 17 19:54:34 Its not with ignore whitespace May 17 19:55:13 cedricb: also, one other thing you could do is to modify ipkg.conf to store the ipkg-lists somewhere other than tmp, since that also uses up ram May 17 19:55:39 koen: Did you develop a pointercal type package for Angstrom? May 17 19:56:45 RP: not yet May 17 19:57:05 RP: I was waiting for people to put more pointercals in the angstrom wiki May 17 19:57:11 I'll try to look at it tomorrow May 17 19:58:10 koen: Theres an exmaple .bb in poky ;-) May 17 19:58:26 I know :) May 17 19:58:31 * RP -> afk, back in an hour May 17 21:02:31 03likewise 07org.oe.dev * r6e2c3679... 10/ (4 files in 3 dirs): helloworld: fix broken commit. May 17 21:02:36 03likewise 07org.oe.dev * r90a3055e... 10/ (3 files in 3 dirs): helloworld: helloworld is easy, fixing my own commits is not. May 17 21:03:05 ehm, are the upstream servers sync'ed? May 17 21:10:35 "Here, the project in question ran `sed 's/gaim/pidgin/g'`, moved their bugzilla to another machine and thought that those actions justified closing all the bugs rather than migrating them. This is hardly a spongeworthy accomplishment." May 17 21:10:59 From http://www.advogato.org/person/cinamod/diary.html?start=143 May 17 21:11:38 reminds me of how we moved from RT to bugzilla May 17 21:19:31 hello cworth May 17 21:19:42 koen: thanks for fixing my bad. May 17 21:19:47 hello likewise May 17 21:20:39 Can anyone who uses pango try the 1.16.4 release in OE which has pref -1? It should look much better in some cases: http://behdad.org/blog/mces/image/metricshinting-kerning.png and will close oebug #2155. May 17 21:21:41 likewise: poky-bleeding is using it May 17 21:22:01 RP: and does is bleed, is it edgy? May 17 21:22:36 likewise: I must admit to not using it much but others in the company do May 17 21:23:04 RP: I'm not using pango myself yet, so I cannot close my own "feature" bug :-) May 17 21:25:02 hi rp May 17 21:26:20 RP: got my bike fixed (the shaft sensors had little hairlines/cracks) (not sure about the right terminology here) May 17 21:26:55 likewise: hairline cracks sounds right (and nasty) :/ May 17 21:27:19 Good to have found and fixed it though May 17 21:27:58 RP: Engine stopped two times on the highways after a 15mile drive in busy traffic at 120km/h. May 17 21:28:27 RP: which, for a noob biker, is always a bit too exciting (especially in Dutch traffic)... May 17 21:28:36 likewise: Yes, sounds nasty May 17 21:28:46 likewise: Enjoying it other than that? :) May 17 21:29:29 RP: sure, I was happy as a little kid today that I could drive around a bit (you know, the drives without any destination) May 17 21:31:29 RP: do you know of a way to leave ipkg out of the image, despite using image.bbclass? May 17 21:33:16 likewise: No, we should really add some way of disabling that May 17 21:44:19 RP: as a test image, I have added helloworld-image which only depends on helloworld (which depends on nothing, statically linked). May 17 21:44:43 likewise: I use pango a lot. I'll check as soon as I get a chance May 17 21:45:24 likewise: There is a post install function somewhere in poky that nukes the packaging files from the image iirc May 17 21:45:28 mreimer: Are there any *simple* pango-using programs (i.e. render Unicode to PNG)? May 17 21:45:34 likewise: a post image generation function even May 17 21:45:45 RP: yes, I could use the image pre or post commands. May 17 21:46:20 RP: but I wondered where the dependency on ipkg is hidden, so that it (and thereby the C lib gets installed) May 17 21:46:45 likewise: top of rootfs_ipk.bbclass (I think) May 17 21:47:40 RP: ok, so maybe rootfs.bbclass must be added? May 17 21:47:43 hrms May 17 21:47:53 RP: which does not use .deb or .ipkg's May 17 21:47:57 libiconv fails with uclibc May 17 21:48:11 likewise: No, you need to use some kind of package system May 17 21:48:46 psokolovsky: ping May 17 21:48:50 RP: oh yes, indeed. But the deployment of the package tool on the target rootfs should be configurable. May 17 21:49:09 likewise: Most images wouldn't actually work without it May 17 21:49:33 RP: Because they require post-install scripts to run? May 17 21:49:39 RP: pong May 17 21:49:43 likewise: yes May 17 21:50:00 psokolovsky: Did you ever finish your make-ipkg-index speedup? May 17 21:50:04 RP: then that's something I would like to get rid of as well. Many users are generating read-only images. May 17 21:50:17 wookey_: why is that? May 17 21:50:18 RP: yep, it's in q to be submitted ;-\ May 17 21:50:31 psokolovsky: Have you something rought I could play with? May 17 21:50:37 psokolovsky: and maybe merge ;-) May 17 21:50:52 * zecke heads to bed May 17 21:51:00 'night zecke May 17 21:51:11 likewise: A worthy aim, although its actually hard May 17 21:51:22 RP: yes, I'll try to update it over next few days (likely, to monday), and post it to bugtracker May 17 21:51:24 gnite ze May 17 21:51:26 cke May 17 21:51:38 wookey_: sorry, you are a victim of tab completion May 17 21:52:01 psokolovsky: ok, thanks. I'm running some builds that could really use that speedup and I'd hate to duplicate the work! May 17 21:52:05 zecke: sorry you were the victim of a trigger happy enter finger May 17 21:52:15 ;) May 17 21:52:24 :-) May 17 21:52:34 sleep tight (in your new place?) May 17 21:52:44 s/in/at/ May 17 21:52:48 actually I will create a spreahsheet first May 17 21:52:53 no new place yet May 17 21:53:03 sorry confused with mickeyl... ;-) May 17 21:53:25 yeah spreadsheets make you get very sleepy May 17 21:53:28 well I'm searching something small and crappy May 17 21:53:34 oh and really cheap :} May 17 21:53:34 as in cheap? May 17 21:53:39 :-) May 17 21:53:58 * RP should start house hunting. I keep putting it off May 17 21:54:15 houses, rich upper class boys :) May 17 21:54:40 zecke: There are shoe boxes here they call houses ;-) May 17 21:54:42 talk to you in 5 years May 17 21:54:45 hehe May 17 21:55:17 btw, still renting here, I'm only student-off for one year now. May 17 21:55:44 ljp: hey lorn, I will test my uclibc neo image inside the emulator tomorrow and then you have something to play with ;) May 17 21:56:02 hmmm. ok. May 17 21:58:57 hehe May 17 22:01:40 Zecke: did you follow the discussion with knutt ? May 17 22:02:17 Zecke: if you have some time I would like to have your point of view as well May 17 22:02:50 gerwinin: skip the second 't' May 17 22:03:06 zecke: okay sorry May 17 22:03:12 gerwinin: I'm summarizing our thoughts on the Foundation May 17 22:03:45 gerwinin: a small table with the options we had May 17 22:10:07 zecke: okay , I think we need to meet up with the core team and with knut and me May 17 22:12:16 cworth: I need a debugging hint May 17 22:12:28 zecke: what's up? May 17 22:13:15 cworth: Let me get a backtrace to ask a sane question May 17 22:15:25 cworth: http://rafb.net/p/cJp7zN53.html any idea how one could debug it? May 17 22:17:15 zecke: Is your X server crashing or something? May 17 22:17:33 cworth: it crashed once, but not this time May 17 22:18:20 I can't imagine lots of other reasons for having an IO error trying to write to X, (unless this is a networked X sesssion and the connection is going away or something). May 17 22:18:21 cworth: this is WebKit/Gdk and I wonder if it would be possible to record the parameters passed to cairo and then replay May 17 22:18:47 zecke: There's a "libcairowrap" thing that might help with that. May 17 22:19:11 cworth: the X is local, but it is apple's quartzX server. I think other people have seen it on a X.org X as well May 17 22:21:20 * cworth restarts X server May 17 22:24:57 good nite ppl May 17 22:25:40 'night all May 17 22:29:29 night May 17 23:58:04 morning **** ENDING LOGGING AT Fri May 18 02:59:56 2007