**** BEGIN LOGGING AT Tue Jul 15 02:59:56 2008 Jul 15 07:34:54 good morning Jul 15 07:45:22 03utx 07org.oe.dev * rfd7e5dad... 10/ (4 files in 3 dirs): Jul 15 07:45:22 pidgin updated to version 2.4.3: Jul 15 07:45:22 * ICQ works again Jul 15 07:45:22 * pidgin-cross-python.patch: worked-around python cross include clash Jul 15 07:45:27 03mickeyl 07org.oe.dev * r93d087e4... 10/ (1 packages/linux/linux-openezx-devel_svn.bb): linux-openezx-devel bump to 2.6.25 Jul 15 07:45:32 03mickeyl 07org.oe.dev * rd39fcbb2... 10/ (8 files in 3 dirs): Jul 15 07:45:32 linux-openmoko-devel: clean up a bit, compute defconfig for om-gta01 and om-gta02 out of one defconfig, Jul 15 07:45:32 enable IKCONFIG in defconfig, disable APM_POWER in defconfig. Jul 15 07:45:37 03koen 07org.oe.dev * r7c888879... 10/ (4 files in 3 dirs): neuros-pkggen: update sources Jul 15 07:45:42 03koen 07org.oe.dev * ra2e94fc6... 10/ (1 conf/machine/include/neuros.inc): neuros.inc: don't use lzo compression for kernel jffs2 partition, the current uboot config doesn't like it Jul 15 07:45:47 03koen 07org.oe.dev * rfa252fa4... 10/ (3 files in 3 dirs): linux omap2 git: add patch that should fix hang on serial Jul 15 07:45:52 03koen 07org.oe.dev * r8dfd9755... 10/ (1 packages/linux/linux-neuros_git.bb): linux-neuros: bump SRCREV Jul 15 07:45:57 03koen 07org.oe.dev * r7297ef56... 10/ (1 packages/u-boot/u-boot_git.bb): u-boot: bump SRCREV for neuros-osd2 Jul 15 07:46:03 03koen 07org.oe.dev * rd4fd51a1... 10/ (1 packages/rt2x00/rt73-firmware_1.8.bb): rt73-firmware: add 1.8 Jul 15 07:46:08 03koen 07org.oe.dev * rc8d00699... 10/ (3 files in 2 dirs): zd1211-firmware: add 1.4 Jul 15 08:03:54 bonjour Jul 15 08:35:52 03tmbinc 07org.oe.dreambox * r9606f7d5... 10/ (3 files in 3 dirs): netbase: fix wpa_supplicant handling Jul 15 08:35:56 03tmbinc 07org.oe.dreambox * rf5db6a6b... 10/ (3 files in 3 dirs): busbox,opendreambox: add default_gw handling Jul 15 08:54:49 03koen 07org.oe.dev * r04195e19... 10/ (1 classes/qtopia4core.bbclass): qtopia4core bbclass: set QT_DIR_NAME Jul 15 09:17:50 03koen 07org.oe.dev * r79ac6734... 10/ (3 files in 3 dirs): neuros-qt-plugins: add qt input driver for the neuros IR remote Jul 15 09:21:40 morning all Jul 15 09:21:44 mickey|zzZZzz: I'm back ;-) Jul 15 09:23:14 RP: hey, welcome back Jul 15 09:38:57 morning Jul 15 09:42:32 hi methril Jul 15 10:02:05 03koen 07org.oe.dev * ra6503453... 10/ (1 packages/neuros-public/neuros-lib-widgets_git.bb): neuros-lib-widgets: add neuros Extra QT base widgets Jul 15 10:11:12 03koen 07org.oe.dev * rdea5a280... 10/ (1 packages/neuros-public/neuros-qt-plugins_git.bb): neuros-qt-plugin: it's an kbddriver, not an inputmethod Jul 15 10:11:38 RP: wb :) Jul 15 10:16:28 hi zecke Jul 15 10:16:44 we should make Tokyo OE capital: http://www.flickr.com/photos/16845425@N00/2671056278/ Jul 15 10:20:21 zecke: ask Laibsc* for the meaning ;-D Jul 15 10:28:39 hi zecke Jul 15 10:30:39 zecke: Hehe, Jul 15 10:30:45 Very funny Jul 15 10:30:48 Awesome Jul 15 10:30:53 Laibsch: any meaning? Jul 15 10:31:17 koen (actually, it should be kouen or k^oen (o with the thing on top)) means park Jul 15 10:31:22 pulic park Jul 15 10:31:22 Laibsch: I have a lot of "mickey's" too, pictures for someone with lolita goth girls Jul 15 10:31:40 mickey, like mickey mouse? Jul 15 10:31:54 Laibsch: no, like mickey's fish restaurant :) Jul 15 10:32:05 * Laibsch does not know that Jul 15 10:32:45 koen is actually pronounced like a German would pronounce koen Jul 15 10:32:49 ko-en Jul 15 10:33:08 unfortunately not koon or kuun like koen's name Jul 15 10:33:13 but very funny indeed Jul 15 10:33:47 公園 Jul 15 10:33:53 would be the Kanji for it Jul 15 10:34:04 first Kanji is public Jul 15 10:34:11 second is enclosure Jul 15 10:38:18 oe Jul 15 10:39:06 Laibsch: the sound is pretty straight forward most of the times Jul 15 10:39:11 oe Jul 15 10:39:29 yes, very much Jul 15 10:39:56 A german would usually be able to read the news off a piece of paper and be understandable Jul 15 10:40:01 oe Jul 15 10:40:09 There are a few kinks here and there, but nothing serious Jul 15 10:40:39 oe Jul 15 10:41:06 oh, too bad Jul 15 10:41:19 I was just about to put the swiss lunatic on ignore ;-) Jul 15 10:42:10 he's on the big ignore now :) Jul 15 10:42:33 kloeri: thanks, any idea what is wrong with him? Jul 15 10:42:42 03koen 07org.oe.dev * rf2bd2108... 10/ (1 packages/dsplink/codec-engine_2.10.bb): codec-engine: package and stage cmem*.a Jul 15 10:42:47 03koen 07org.oe.dev * r84b25d66... 10/ (4 files in 2 dirs): dsplink: package and stage dsplink.lib Jul 15 10:44:19 zecke: he's 4 years old (mentally at least) and have serious trouble growing up Jul 15 10:45:02 he's been doing stuff like this at least for the past year and never seem to change Jul 15 10:50:16 thanks, kloeri Jul 15 10:50:19 appreciated Jul 15 10:50:22 np Jul 15 11:07:05 03koen 07org.oe.dev * r4a1fbbf2... 10/ (3 files in 3 dirs): u-boot env: update neuros-osd2 env Jul 15 11:53:04 http://rafb.net/p/TXMcWr25.html << Error while trying to bitbake qt-embedded (qmake2-native) Jul 15 11:55:17 piroko: interesting, where do you build? Jul 15 11:57:38 zecke: I'm not sure what you mean by "where" Jul 15 12:00:05 piroko: which machine, architecture and distribution Jul 15 12:01:02 there are two interesting things armv5te-angstrom-linux-gnueabi/qmake2-native-2.10a-r2, a native in that directory Jul 15 12:01:28 and that when compiling src/tools/moc it wants a mocced file... indicating that there is something really messed up with your environment Jul 15 12:01:34 zecke: x86 archlinux, compiling for an at91sam9263ek angstrom Jul 15 12:01:46 so messed up I have never seen it and have no straight forward answer Jul 15 12:03:37 ERROR: Multiple .bb files are due to be built which each provide qtopia-core (/home/jeremy/stuff/org.openembedded.dev/packages/qt4/qt-embedded_4.4.0.bb /home/jeremy/stuff/org.openembedded.dev/packages/qt4/qtopia-core_4.4.0.bb).' Jul 15 12:03:41 There's that too ^^ Jul 15 12:05:28 piroko: is that stock oe? or some gumstix stuff or similar? Jul 15 12:06:02 zecke: Stock oe dev branch, just updated yesterday Jul 15 12:07:38 piroko: hmm, might be the best thing to write a mail... I can start a build overnight to try to reproduce this... Jul 15 12:13:02 Well while we're at it, I'm trying to get native-compiling on my embedded machine. But for whatever reason cc1 is missing. What package provides it? Jul 15 12:14:15 i would assume gcc Jul 15 12:14:18 bbl Jul 15 12:18:05 zecke: I installed gcc. gcc is there, just not cc1 and the stdc library (printf etc) Jul 15 12:24:48 what do i need to get full i18n support on a oe image ? Jul 15 12:25:12 i ve mode a gtk app which checks for locale and it returns "C" Jul 15 12:26:52 just locale-base-LL-CC and its dependencies should be enough, I think Jul 15 12:29:08 morning Jul 15 12:29:37 hmm i am building uclibc .. and i dont see all those locale-LL ipkgs ..?? Jul 15 12:29:55 oh, uclibc Jul 15 12:30:00 no idea in that case, sorry :-} Jul 15 12:31:32 thanks anyway Jul 15 12:38:08 piroko, have you installed gcc-symlinks? Jul 15 12:38:13 binutils? Jul 15 12:38:21 diego__: Yes and yes. They are both in IMAGE_INSTALL Jul 15 12:38:53 binutils-dev too Jul 15 12:39:52 try to manually install it: ipkg install gcc-symlink Jul 15 12:41:42 Brb, gotta plug the ethernet into the embedded box Jul 15 12:44:21 can I do something like PR_machine = "5" ? Jul 15 12:48:16 diego__: I removed gcc-symlinks, then re-installed it. It installed fine, but running gcc on my hello world program failed with: gcc: error trying to exec 'cc1': execvp: No such file or directory Jul 15 12:53:06 * * OE Bug 4429 has been created by diegor(AT)tiscali.it Jul 15 12:53:09 * * apr compilation failed Jul 15 12:53:11 * * http://bugs.openembedded.net/show_bug.cgi?id=4429 Jul 15 12:54:11 piroko, use find to look for the cc1. Jul 15 12:54:33 find /usr -name \*cc1\* -print Jul 15 12:54:58 diego__: It's not there Jul 15 12:55:01 Did that already Jul 15 12:55:17 /usr/libexec/gcc/arm-angstrom-linux-gnueabi/4.2.4/cc1plus is there Jul 15 12:55:20 But no cc1 Jul 15 12:59:19 Here's the bitbake I'm building: http://rafb.net/p/s3oK0s38.html Jul 15 12:59:47 Offhand does anyone know how to get gcc to correctly align data when its using double instructions like strd and ldrd on arm? Jul 15 13:03:26 So is gcc-symlinks supposed to provide cc1? Jul 15 13:03:33 piroko, do you have memory constraints? Jul 15 13:04:08 diego__: We're trying to keep it small, but not really Jul 15 13:04:36 eventually you can install task-native-sdk Jul 15 13:04:47 eventually? Jul 15 13:04:51 it should install all you need Jul 15 13:05:21 (or was task-sdk? i don't remember O:-) ) Jul 15 13:05:31 task-native-sdk works Jul 15 13:08:18 RP: http://wiki.debian.org/ArmEabiFixes Jul 15 13:08:21 Cool. We get the real-deal with sed/grep/gawk and stuff too Jul 15 13:08:59 03utx 07org.oe.dev * rd8ffa6c8... 10/ (31 files in 3 dirs): Jul 15 13:08:59 Added linux-rp-2.6.26: Jul 15 13:08:59 * based on linux-rp-2.6.25+2.6.26-rc4 Jul 15 13:08:59 (b82d34d538132713ba10d59d0209f562e276f44b) Jul 15 13:08:59 * ported sharpsl-rc-r1.patch Jul 15 13:09:00 * set DEFAULT_PREFERENCE_spitz (suspend works again) Jul 15 13:12:02 ant|work: I know that, that doesn't tell me my gcc isn't getting this right though :( Jul 15 13:13:08 RP: once upon a time I started looking for the alignments traps Jul 15 13:13:16 and I collected a couple of links Jul 15 13:13:25 (for the novices) Jul 15 13:13:43 this other is old but informative: http://netwinder.osuosl.org/users/b/brianbr/public_html/alignment.html#q9 Jul 15 13:21:45 bbl Jul 15 13:22:09 03koen 07org.oe.dev * rbab230e1... 10/ (3 files in 2 dirs): ffmpeg, omapfbplay: bump SRCREV Jul 15 13:30:03 funny, i have no more cc1 on my image too Jul 15 13:30:32 diego__: Woo! Jul 15 13:30:34 Jul 15 13:30:35 Sorry Jul 15 13:43:04 a couple of week ago i was able to compile Jul 15 13:43:20 I wonder what happened ... Jul 15 13:43:32 Does this mean I found a bug? :D Jul 15 13:44:40 maybe. Which branch are you using Jul 15 13:45:19 dev Jul 15 13:47:36 me too Jul 15 13:48:25 may be is the price to live on the edge ^^ Jul 15 13:51:02 03koen 07org.oe.dev * rdb911b92... 10/ (1 packages/gstreamer/gst-ffmpeg_0.10.4.bb): gst-ffmpeg: prefer 0.10.4 for armv7a Jul 15 13:53:26 are most OE developers using the monotone server or the git server these days? Jul 15 13:53:38 s/server/repo Jul 15 14:08:02 Well adding mdev to my IMAGE_INSTALL caused both mdev and udev to get launched :P Jul 15 14:09:54 Is that supposed to happen? Jul 15 14:09:57 yes Jul 15 14:10:18 Anyone know if codesourcery puts their 'make check' results online somewhere? Jul 15 14:12:36 hrw: So how do I tell openembedded I only want to use mdev? Jul 15 14:12:48 install mdev not udev? Jul 15 14:13:19 hrw: Well udev is getting installed by default. How do I tell openembedded not to install it? Jul 15 14:13:24 1h eh? :) Jul 15 14:13:26 ga Jul 15 14:13:50 piroko: look which pacakge get it as dependency? Jul 15 14:15:09 base-image pulls in hundreds of packages... :( Jul 15 14:16:00 do a bitbake -g Jul 15 14:16:04 and follow the -> :) Jul 15 14:16:16 Tartarus: Thank you :) Jul 15 14:16:44 Tartarus: thats build time Jul 15 14:16:54 piroko: read tmp/rootfs/usr/lib/opkg/status Jul 15 14:17:48 hrw: opkg doesn't exist in usr/lib Jul 15 14:17:49 hrw, right and it's probably some task or another saying it wants udev, yes? Jul 15 14:18:12 piroko: then ipkg/status Jul 15 14:18:19 Tartarus: exactly Jul 15 14:20:31 Ah. hal requires udev Jul 15 14:20:59 probably libs from udev but we did not split them to separate Jul 15 14:22:40 Well crud. ftpd-topfield and vsftpd both failed to build Jul 15 14:22:49 Any other ftp daemons in oe I could use? Jul 15 14:24:34 Tartarus: you can check gcc mailing list (gcc-testresults) people configure their testers to automatically mail results to this ml and you will find many results for arch and toolchain combo you are lookin for Jul 15 14:24:38 http://gcc.gnu.org/ml/gcc-testresults/ Jul 15 14:24:59 ok Jul 15 14:32:37 piroko: cc1 should be in cpp package Jul 15 14:33:19 diego__: Doesn't that get pulled in by task-native-sdk? Jul 15 14:35:18 dunno Jul 15 14:35:39 and of course you need cpp-symlink as well Jul 15 14:36:04 Wow. That did it Jul 15 14:36:46 So why doesn't the task-native-sdk file pull in cpp???? Jul 15 14:38:05 RP: -mpreferred-stack-boundary could be used to force stack alignment but EABI compiler should align data to its natural alignment. Jul 15 14:38:07 Blah. It seems I am encountering more and more problems with every one I get past Jul 15 14:38:43 Still can't find a flippin ftp server for ti Jul 15 14:38:44 *it Jul 15 14:41:27 khem: Its not stack data :( Jul 15 14:41:55 khem: and its not aligning the 64 bit instructions to 8 byte boundaries, only 4 :( Jul 15 14:42:33 piroko: a bug i suppose. Open a bug ^^ Jul 15 14:42:51 diego__: Dude, I could open like 20 bugs right now Jul 15 14:43:31 RP: I have seen alignment problems on ARM with gcc 4.x typecasts from char* to other data types caused problems. using -Wcast-align captured many issues Jul 15 14:44:17 RP: you mean the instructions that access 64 bit data like strd/ldrd ? Jul 15 14:44:34 khem: In this case its a uint64_t being stored with a strd :/ Jul 15 14:44:47 The uint64_t is in a c++ class definition Jul 15 14:44:57 but I'm also seeing the problem in cairo Jul 15 14:45:10 all from webkit in standard poky Jul 15 14:45:19 on armv5te Jul 15 14:47:08 piroko: I already have fixed your problem in my local tree Jul 15 14:47:34 RP: is it gcc 4.x ? Jul 15 14:47:46 khem: How? Jul 15 14:47:50 And which one? haha Jul 15 14:47:51 khem: yes Jul 15 14:48:11 piroko: cc1 missing Jul 15 14:48:24 khem: Ah. I just installed cpp and cpp-symlinks and it seems to have fixed it Jul 15 14:48:33 khem: 4.2.3 Jul 15 14:49:21 RP: seems a bug are you compiling for Xscale ? Jul 15 14:50:51 khem: yes Jul 15 14:51:10 khem: Compile for armv4 and the problem goes away since it doesn't use the double instructions Jul 15 14:54:34 I see Jul 15 14:59:50 -mabi=aapcs Jul 15 14:59:57 does that help Jul 15 15:03:43 is not aapcs old abi Jul 15 15:03:44 ? Jul 15 15:07:43 hrw: ah old ABI hmmm Jul 15 15:08:21 OABI could have a bug here Jul 15 15:17:28 RP: what is your opinion about my proposal to toolchain build sequence change ? Jul 15 15:35:46 trying to create a recipe for nbench which does not use autoconf and builds a c app to determine size of pointer on target - how should I deal with this for a recipe that would work on multiple architectures? Jul 15 15:36:09 use autoconf? Jul 15 15:36:48 nbench doesn't use it and I'm no autoconf wizard... :( Jul 15 15:37:10 any suggestions on low level CPU/mem benchmarks that would perhaps be better suited anyway? Jul 15 15:44:52 tharvey: autotools is often the easiest even for small apps as cross compiling manually is tedious: http://bec-systems.com/web/content/view/95/9/ Jul 15 15:45:06 cliff? Jul 15 15:45:31 cbrake, thanks for the quick ref - I'll go that route then and add that as a file for my recipe Jul 15 15:45:51 frikker: yeah -- who are you? Jul 15 15:46:23 hah sorry - i am 'blaine' from ohio - i emailed you a few times Jul 15 15:46:32 thanks for the help again :) Jul 15 15:47:23 frikker: ahh, OK. You're welcome :-) Jul 15 15:48:58 is anyone familiar with i2c? I'm confused on what it is Jul 15 15:49:38 frikker: I'm using it to talk with a small A/D convertor Jul 15 15:50:01 what is it exactly? Jul 15 15:50:10 related to gpio? Jul 15 15:52:04 frikker: its a 2-wire bus Jul 15 15:52:08 frikker, it is a serial 2 wire communications bus designed by Philips./ Jul 15 15:52:21 ahh i see Jul 15 15:52:30 thank you Jul 15 15:52:32 frikker: http://en.wikipedia.org/wiki/I%C2%B2C Jul 15 15:52:45 it has clock/data and then uses off clock (edge) transistions of the data line to provide start/stop states. Jul 15 15:52:48 do you use a kernel module to communicate? Jul 15 15:52:56 frikker: the PXA270 has a I2C controller and there are drivers available Jul 15 15:52:59 gotcha Jul 15 15:53:12 ok that makes sense. is that what ssp would be using? Jul 15 15:53:21 frikker: here is a little test app that might be helpful: http://pastebin.ca/1072473 Jul 15 15:53:33 frikker: SSP is SPI, which is a little different Jul 15 15:53:38 frikker, there is also a generic bitbang (done all software no I2C hardware involved) i2c driver for linux. you just assign the io in your machine file. Jul 15 15:54:03 the bitbang is nice since you can use any IO. Jul 15 15:54:09 ok cool. makes sense. Jul 15 15:54:22 frikker, it is bad becuase it is all done in software so it is slower. Jul 15 15:54:32 frikker, trade-off :-) Jul 15 15:54:33 frikker: SSP ~= http://en.wikipedia.org/wiki/Serial_Peripheral_Interface_Bus Jul 15 15:54:36 would you use this to communicate to with an lcd screen? or just devices? Jul 15 15:54:43 svolpe: good call :) Jul 15 15:55:06 frikker, it is only 100KHz, 400KHz max, not very good for refresh rates ;-) Jul 15 15:55:36 frikker, it is really only good for slow speed eeproms, ADCs,DACs, RealTime clocks.. Jul 15 15:56:23 frikker, unless the LCD was a character one or one that had its own controller on board (for refreshing) . Jul 15 15:57:29 03utx 07org.oe.dev * r62a12617... 10/ (1 packages/tangogps packages/tangogps/tangogps_0.9.0.3.bb): Jul 15 15:57:29 Added tangogps-0.9.0.3: Jul 15 15:57:29 * based on openmoko-projects/tangogps_0.7.bb Jul 15 15:57:29 * removed openmoko specific requirements Jul 15 15:57:29 * updated to version 0.9.0.3 Jul 15 15:57:59 gotcha Jul 15 15:58:05 svolpe: you're talking about i2c right? Jul 15 15:58:16 frikker, yes. Jul 15 15:58:28 svolpe: and yes, the lcd is a character device that does have an onboard controller (i think) Jul 15 15:58:56 frikker, does it have an I2C or SPI (SSP) interface? Jul 15 15:59:05 svolpe: spi Jul 15 15:59:26 svolpe: we're having a hell of a time figuring it out though. we don't have much (at all) experience with getting something like it to work Jul 15 15:59:43 svolpe: just reading through piles of documentation at this point to see if we even wired it right Jul 15 15:59:48 frikker, well spi can be run much faster than I2C. A spi interface is just a shift register. Jul 15 16:00:42 svolpe: ok makes sense Jul 15 16:00:56 svolpe: do you use the ssp.c kernel module to talk to it? Jul 15 16:00:56 frikker, your processor probably has a spi interface, I'm assuming that is what your using? Jul 15 16:01:05 svolpe: yeah we have a pxa270 Jul 15 16:01:15 svolpe: chipset at least. its an arm 5te i think Jul 15 16:01:53 frikker, the bad thing about spi is there is no standard on rising edge/falling edge for clock transactions so you need to make sure you set the master (pxa270) up to use the correct clock edge transaction required by your display. Jul 15 16:02:47 frikker, I always do the best based on the data sheet for setting up a spi bus then normally hook up an oscilloscope to verify the spi wave forms and edge transactions. Jul 15 16:03:47 svolpe: what do you drive it with? Jul 15 16:03:56 svolpe: once you have itall set up? Jul 15 16:05:04 frikker, we can go on a side channel so we do not flood OE with unrelated stuff :-) Jul 15 16:05:17 svolpe: sounds awesome, thanks :) Jul 15 16:21:32 khem: I'm using EABI Jul 15 16:21:32 svolpe: BTW, I've been using git for most of my OE work, set up my own git server, etc. Jul 15 16:21:42 khem: and sorry, I haven't seen the toolchain proposal in question Jul 15 16:22:21 Has anyone had success installing qt-embedded on angstrom? Jul 15 16:31:06 Bah Jul 15 16:31:19 Anyone around with commit privs wanna make KERNEL_OUTPUT in kernel.bbclass a ?= Jul 15 16:31:32 piroko: hehe, didn't try it recently Jul 15 16:31:53 zecke: I filed a bug about it: http://bugs.openembedded.net/show_bug.cgi?id=4430 Jul 15 16:33:06 * * OE Bug 4430 has been created by obrien654j(AT)gmail.com Jul 15 16:33:07 * * task-qtopia-core-console fails to build Jul 15 16:33:09 * * http://bugs.openembedded.net/show_bug.cgi?id=4430 Jul 15 16:33:16 Lol. There ya go Jul 15 17:04:40 Does anyone know why kernel.bbclass' do_deploy uses arch/${ARCH}/boot/${KERNEL_IMAGETYPE} not ${KERNEL_OUTPUT} ? Jul 15 17:05:30 Tartarus: probably due to not noticing KERNEL_OUTPUT Jul 15 17:06:20 k Jul 15 17:06:36 i'll file and fix sometime today maybe along with KERNEL_OUTPUT not using ?= Jul 15 17:06:46 ?= is needed Jul 15 17:06:55 so people can overwrite it in kernel recipes Jul 15 17:09:38 hey hrw :) make it back alright? Jul 15 17:10:49 kgilmer: yes - just had to wait 2h on Tegel for microbus Jul 15 17:11:02 kgilmer: next time I will call them to wait 15 minutes more ;) Jul 15 17:11:32 hrw: ugh! i had a similiar problem. plane after landing at jfk was on tarmac for 2 hours before a gate was available. Jul 15 17:11:42 it wasn't ?= Jul 15 17:11:45 at least not in my snap Jul 15 17:11:47 so 9.5 hrs became 11.5 hrs Jul 15 17:11:50 i'll check current before i file a bug Jul 15 17:11:56 but that means wget'ing the db again i bet Jul 15 17:11:59 kgilmer: ugh Jul 15 17:20:04 I think I may have found why my qt-embedded build failed Jul 15 17:20:46 Last time I fetched the sources, the fetch operation just stopped at 100%, so I control-c'd it. I refetched the sources again, and it's doing the same thing right now Jul 15 17:21:20 Which is, nothing at all. Just sitting here. ps and top don't show any bitbakes running. What should I do? Just wait forever? My hard drive and processor are not being used either Jul 15 17:22:59 try bitbake -c rebuild Jul 15 17:23:11 s/target/recipe/ Jul 15 17:24:43 or use pstree to understand which process as hunged Jul 15 17:25:48 ├─urxvt───bash───python───python───wget Jul 15 17:25:52 wget is hung Jul 15 17:26:51 But I don't want to control-c it because my build failed the last time I did that. Not sure if it's related, but it's the only oddity I can find Jul 15 17:27:58 check on what it hangs Jul 15 17:28:47 hrw: It hangs on qt-embedded-4.4.0, after the fetch hits 100% Jul 15 17:30:18 strange Jul 15 17:30:36 Finally! It timed out and retried, finishing the connection Jul 15 17:30:53 Maybe the build will succeed this time Jul 15 17:33:59 Wait. Why is oe downloading the 4.3.3 AND 4.4.0 versions of qtopia???? Jul 15 17:37:41 This makes no sense Jul 15 17:38:50 it does Jul 15 17:39:06 qmake, uicmoc, qtopia are 3 components in misc versions Jul 15 17:39:16 there was something about it on ML Jul 15 17:39:26 Wonderful. More convolutedness Jul 15 17:39:44 Openembedded is one huge hack job after another Jul 15 17:40:16 Ah well. It's saved me lots of time, taken up lots too Jul 15 17:40:26 I guess it all evens out in the end Jul 15 17:40:34 piroko, stick with it and it starts saving you time :) Jul 15 17:44:52 Argh. It failed again. Same error Jul 15 17:47:13 ~hug ~/devel/sources/dl_dir with over 12GB of archives Jul 15 17:47:15 * ibot jumps into ~/devel/sources/dl_dir with over 12GB of archives's lap and huggles and *hugs* ~/devel/sources/dl_dir with over 12GB of archives Jul 15 17:48:07 * piroko rolls his eyes Jul 15 17:48:30 ? Jul 15 17:48:43 piroko, we are using poky but we build a jvm that depends on qt3, uimoc, etc. and it builds fine. you can see the sources at http://cvs.buglabs.net/cgi-bin/viewvc.cgi/com.buglabs.build.oe/meta-oe/packages/?root=cvs Jul 15 17:48:47 * piroko shrugs Jul 15 17:49:21 kgilmer: The shrug was not to you Jul 15 17:49:46 piroko: no worries :) Jul 15 17:49:59 kgilmer: I don't know why, but qmake2-native errors out not being able to find moc_qbuffer.cpp Jul 15 17:50:01 for some reason qt3 depends on mysql with i think is strange Jul 15 17:50:04 but...it works. Jul 15 17:50:10 hi all Jul 15 17:50:12 Well I'm doing qt4 Jul 15 17:50:15 kgilmer: QtSql :) Jul 15 17:50:19 oh ic. Jul 15 17:51:19 piroko: yes, the breakage of .dev is annoying, but mostly unavoidable. it is a matter of locking things down when you start (like I do with my openmoko branch) Jul 15 17:51:51 But the stable tree is quite old, no? Jul 15 17:51:52 piroko: my build is still going, I will have a look at the qtopia4 stuff, I mostly reworked it in december of last year, I assume the addition of QtE4.4 for neuros broke things... Jul 15 17:52:04 zecke: Ah ok. Thanks man Jul 15 17:52:04 piroko: yes, you can not have both though :) Jul 15 17:52:15 Understandable Jul 15 18:15:59 what package would flash_eraseall be part of? would think there would be flashutils or something Jul 15 18:22:18 mtd-utils Jul 15 19:04:21 question Jul 15 19:04:30 if i go into the kernel source directory Jul 15 19:04:36 and type 'make menuconfig' and change options Jul 15 19:04:39 then rebuild the kernel with bitbake Jul 15 19:04:48 will those new config options take hold? or is that completely seperate?? Jul 15 19:06:14 frikker: it depends, e.g. if the configure task ran or will be run... Jul 15 19:06:34 frikker: but bitbake -cmenuconfig virtual/kernel (IIRC) should be a safe bet Jul 15 19:09:45 thanks zecke Jul 15 19:48:00 re Jul 15 19:51:10 ply Jul 15 19:56:36 03koen 07org.oe.dev * r2d823aa4... 10/ (3 files in 3 dirs): linux-omap2 git: add patch to fix a race in i2c code Jul 15 20:19:11 RP: around ? Jul 15 20:23:56 hey khem Jul 15 20:24:15 Any chance people in the OE community care about fixing up gcc 4.2.4 you think? Jul 15 20:24:30 Debian has a buncha patches, and MIPS 'make check' results are nicer looking Jul 15 20:30:58 hi all Jul 15 20:32:28 Tartarus: you there? Jul 15 20:32:32 hi ant_ Jul 15 20:32:40 hi flo_lap Jul 15 20:32:50 Tartarus: sure Jul 15 20:33:24 Tartarus: What patches in particular. You can send it over on ml so they can be reviewed Jul 15 20:34:16 khem, in particular? There's a prsomething/prsomething-mips that I bet fixes one of the failures Jul 15 20:34:27 but in general, it's a pretty big diff on the testsuite end Jul 15 20:35:09 http://gcc.gnu.org/ml/gcc-testresults/2008-07/msg00558.html is Debian Jul 15 20:36:45 actually, http://pastebin.com/m3af88b19 Jul 15 20:36:52 * flo_lap thinks about fixing 4.3.1 Jul 15 20:37:01 developer woke up and reset the board while libstdc++ was going :) Jul 15 20:39:07 ant_, back now :) Had to setup a new cable modem / AP Jul 15 20:39:30 hey, you seem very busy atm Jul 15 20:39:55 when you'll find time, please feel free to test udev_124 and let me know Jul 15 20:40:22 I suspect on c7x0 the timeout is caused by 1)mmc 2)mtd Jul 15 20:40:30 Tartarus: OK results look normal Jul 15 20:41:45 Tartarus: I think it is better to take patches in small rather than a big ones. I have done it in past and the if there was a following regression because of this patch it became tough to find which part of patch really caused it. Jul 15 20:42:09 khem, my plan is to do a wholesale switch-over Jul 15 20:42:29 hence maybe not push it back, dunno about avr32 :) Jul 15 20:42:33 flo_lap: gcc 4.3.1 is working well except I had issues buiding target gcc on uclibc Jul 15 20:43:26 Tartarus: wholesale may not be a good idea and many of those patched get committed in next minor releases like 4.2.5 in future so porting forward will need hacking this monster always Jul 15 20:43:30 khem: try to build gcc-cross-sdk Jul 15 20:43:47 flo_lap: havent done that Jul 15 20:44:01 flo_lap: What is the error do u have it handu Jul 15 20:44:07 handy Jul 15 20:45:07 ~oebug 4228 Jul 15 20:45:33 khem, I don't mean as 1 big diff, I mean as bringing in all of the dpatches. So next update would be a pita until debian follows Jul 15 20:45:49 khem, another reason it might not be suitable to push, but fine for my needs Jul 15 20:45:52 Tartarus: ah that would be ok Jul 15 20:46:02 hrm Jul 15 20:46:03 http://bugs.openembedded.net/show_bug.cgi?id=4428 Jul 15 20:47:53 !oebug 4228 Jul 15 20:47:54 * * Bug 4228, Status: UNCONFIRMED, Created: 2008-05-07 19:37 Jul 15 20:47:55 * * : various likely targets for backporting Jul 15 20:47:56 * * http://bugs.openembedded.net/show_bug.cgi?id=4228 Jul 15 20:48:32 flo_lap: you forgot to attach full log. The snippet in summary is not the cause of problem Jul 15 20:48:44 ah... hrm Jul 15 20:50:00 khem: strange... I selected that file, but it doesn't seem to work creating a bug and attaching a file in one step. Jul 15 20:50:06 I have proposed some changes in build sequence. Jul 15 20:50:27 flo_lap: this has worked for me. Atleast for patches Jul 15 20:50:55 strange Jul 15 20:51:20 Once I found out on which machine I built that package I can provide the full log Jul 15 20:52:03 flo_lap: OK I have only one machine :) Jul 15 20:53:19 khem: I have oe on two of my workstations and two blades at work Jul 15 20:55:44 hi, can you please apply this one to the .dev branch? http://shell.studenti.unina.it/~ospite/tmp/rsync.inc.diff Jul 15 20:57:06 khem: done Jul 15 20:57:31 ao2: can you open a bug and attach your patch to it plz. Jul 15 20:58:06 khem, ok. I thought it was faster here, but I'll do as you request. Thanks. Jul 15 21:04:31 ao2: its faster with bugs in future to track changes thats why bugs are requested. Jul 15 21:04:54 flo_lap: its cross-gcc recipe ? Jul 15 21:05:14 flo_lap: I think we need to look into your config.log Jul 15 21:06:05 khem: sure, will attach it as well... in fact I was not able to see why exactly this test failed. one sec Jul 15 21:06:05 ~seen xjqian Jul 15 21:06:07 xjqian was last seen on IRC in channel #oe, 26d 10h 54m 10s ago, saying: 'hrw: my mistake, never mind. But still trying to figure out why the kernel built from stable fails to boot'. Jul 15 21:07:03 flo_lap: Most proably some headers Jul 15 21:07:32 khem: iirc there was something like "syntax error" Jul 15 21:07:39 ugh he somehow became the 'mantainer' of speech-dispatcher...which I hopefully fixed Jul 15 21:08:07 * * OE Bug 4431 has been created by ospite(AT)studenti.unina.it Jul 15 21:08:08 * * rsync package: do_fech failed Jul 15 21:08:10 * * http://bugs.openembedded.net/show_bug.cgi?id=4431 Jul 15 21:12:10 flo_lap: Do may be you can pastebin it Jul 15 21:13:08 khem: sure, one sec Jul 15 21:14:38 khem: http://pastebin.com/m3c36e9cc Jul 15 21:14:50 flo Jul 15 21:15:06 are you the libtool expert? Jul 15 21:15:55 khem: line 533 Jul 15 21:16:09 ant_: no... luckily not ;) Jul 15 21:18:47 flo_lap: suggestions? ;-& Jul 15 21:19:39 ant_: wbout what? Jul 15 21:19:46 libtool Jul 15 21:19:49 eh about Jul 15 21:19:58 RP? koen? Jul 15 21:20:23 I guess so.. RP most likely. Jul 15 21:20:25 I'm unsure about how I fixed oebug 4391 Jul 15 21:20:36 worksforme (tm) Jul 15 21:21:15 unsure what caused it...libtools2? Jul 15 21:21:27 2.2.4 I mean Jul 15 21:24:29 * ant_ cannot believe nobody's using speech-dispatcher Jul 15 21:26:55 flo_lap: I think its the search path Jul 15 21:27:10 flo_lap: Do you have headers installed properly Jul 15 21:34:38 flo_lap: limits.h should be installed by glibc-intermediate in your case I guess Jul 15 21:35:06 Let me try the same think with my new build sequence and see if that helps Jul 15 21:35:32 oops Jul 15 21:35:34 re Jul 15 21:36:06 khem: Not sure yet... I have noticed that quite a lot of the gcc build system has changed. Jul 15 21:36:31 khem: It works for gcc-cross, so there must be a some small issue. Jul 15 21:44:51 khem: I see the c++ includes are set, but not the standard C includes. If the xgcc binary does not have the correct path built in this might be a reason why it fails. Jul 15 21:45:15 * ant_ wonders how gpsdrive_2.10pre4 now depends on mysql... Jul 15 21:45:57 maybe some 'clever' person activated mysql support in it Jul 15 21:47:00 flo_lap: the nasty bit is mysql fails to compile :-D Jul 15 21:47:41 but even, I cannot imagine it on Zaurus ;-] Jul 15 21:48:20 ant_: we should change that be to disable and not to require mysql Jul 15 21:48:55 we'll have to wait for xjqian's comeback Jul 15 21:52:14 XorA|gone might know Jul 15 21:55:26 flo_lap: in debian-stable libmysqlclient is 'suggested' but in unstable 'mysql-server' is suggested Jul 15 21:55:32 ? Jul 15 21:56:21 ant_: maybe it opens libraries on start instead of linking against them Jul 15 21:57:09 is there an easy to read tutorial on how exactly to create a bitbake kernel module package from scratch? Jul 15 21:57:37 specifically, linux/arch/arm/mach-pxa/ssp.c needs to be a kernel module for me. I can't figure out how to set that up... Jul 15 21:58:41 frikker: Just make it a module in your kernel configuration you provide in oe and it does what you want. Jul 15 21:58:52 its not a standard module though Jul 15 21:58:59 i *think* its supposed to be included in pxa3xx_spi Jul 15 21:59:06 but no such luck. we still can't use its functionality Jul 15 21:59:17 and i've seen it used as a stand along module with a custom compilation Jul 15 21:59:21 alone* Jul 15 21:59:45 frikker: hum well... that's a different thing. More kernel related than oe related. Jul 15 21:59:55 true actually, good point Jul 15 22:00:03 well Jul 15 22:00:19 besides that - is there a tutorial on how to make a kernel module bb package? it would be useful to me I believe. Jul 15 22:02:11 frikker: no need to do so - just make a kernel bb and OE will packae all built modules. Jul 15 22:05:18 frikker: You need to know how to build kernel modules and I recommend the Linux Kernel Development Book by Robert Love Chapter 16 Jul 15 22:07:12 flo_lap: what do you mean by making a kernel bb? a kernel module package? any references on that? Jul 15 22:07:26 khem: you're probably right - thank you for the reference Jul 15 22:08:23 flo_lap: --enable-cheaders=c_std Jul 15 22:12:59 khem: what exactly does this option set? Jul 15 22:22:07 * * OE Bug 4432 has been created by andrea.adami(AT)gmail.com Jul 15 22:22:09 * * gpsdrive_2.10pre4 r2 fails do_configure Jul 15 22:22:11 * * http://bugs.openembedded.net/show_bug.cgi?id=4432 Jul 15 22:34:57 flo_lap: This is already there in the recipe Jul 15 22:37:50 flo_lap: flo_lap docs say this Headers intended to include standard C headers, and put select Jul 15 22:37:54 ! names into the std:: namespace. Jul 15 22:43:29 khem: ah okay, so this is not our problem here Jul 15 22:47:19 flo_lap: it infact could be Jul 15 22:49:50 hmm... /me checks the context Jul 15 22:56:55 khem: I'll take a closer look at the configure options tomorrow. Jul 15 22:57:15 * flo_lap needs a little bit of sleep in the meantime Jul 15 22:59:46 nite all Jul 15 23:08:18 good night **** ENDING LOGGING AT Wed Jul 16 02:59:56 2008