**** BEGIN LOGGING AT Thu Jun 21 02:59:57 2007 Jun 21 03:56:11 03mickeyl 07org.oe.dev * r2a2e9f06... 10/ (9 files in 5 dirs): esound 0.2.36 clean up, add no-docs patch Jun 21 03:56:19 03mickeyl 07org.oe.dev * ra1d9bb1b... 10/ (10 files in 3 dirs): Jun 21 03:56:19 directfb update: Jun 21 03:56:19 * bump ++dfb to 1.0.0 Jun 21 03:56:19 * bump examples to 1.0.0 Jun 21 03:56:19 * bump fusionsound to 1.0.0 Jun 21 03:56:20 * clean up a bit Jun 21 03:56:24 03mickeyl 07org.oe.dev * r3c408cdc... 10/ (1 classes/efl1.bbclass classes/e.bbclass): classes: update efl1 and e for E revamp Jun 21 03:56:30 03mickeyl 07org.oe.dev * r1104edc7... 10/ (40 files in 6 dirs): Jun 21 03:56:30 massive E revamp. all shiny and new. no more binconfig madness, no longer Jun 21 03:56:31 m4 patch madness, much less hackish and fragile. all hail pkgconfig Jun 21 03:56:36 03mickeyl 07org.oe.dev * r2cbc6468... 10/ (1 conf/distro/include/sane-srcdates.inc): sane-srcdates.inc: add new entries for EFL and E17 packages Jun 21 03:56:41 03mickeyl 07org.oe.dev * r7d8469a5... 10/ (36 files in 6 dirs): e17 applications: catch up with EFL overhaul Jun 21 04:00:34 03mickeyl 07org.oe.dev * rb86bcddf... 10/ (1 packages/esound/old): esound: remove unused directory Jun 21 04:00:39 03mickeyl 07org.oe.dev * r96a3fd53... 10/ (6 files in 5 dirs): move older efl stuff to nonworking Jun 21 04:00:44 03mickeyl 07org.oe.dev * r2e7bef94... 10/ (6 files in 4 dirs): elitaire is back in EFL directory Jun 21 04:00:48 03mickeyl 07org.oe.dev * r9ab74777... 10/ (3 files in 2 dirs): python-pygtk: explictly specify gtk version we depend on Jun 21 04:01:45 :D Jun 21 04:13:01 * * OE Bug 2487 has been RESOLVED (FIXED) by mickeyl(AT)handhelds.org Jun 21 04:13:03 * *  Problems with python-2.4.4 manifest Jun 21 04:13:05 * * http://bugs.openembedded.org/show_bug.cgi?id=2487 Jun 21 04:13:53 ~praise cdbot2 Jun 21 04:13:57 All hail cdbot2! Jun 21 04:29:31 03mickeyl 07org.oe.dev * r231950a6... 10/ (4 files in 3 dirs): python-2.4.4: update manifest, update manifest generator, close #2487 Jun 21 04:41:01 * * OE Bug 2459 has been RESOLVED (FIXED) by mickeyl(AT)handhelds.org Jun 21 04:41:03 * *  Bitbake build for python-setuptools Jun 21 04:41:05 * * http://bugs.openembedded.org/show_bug.cgi?id=2459 Jun 21 04:41:12 * * OE Bug 2460 has been RESOLVED (FIXED) by mickeyl(AT)handhelds.org Jun 21 04:41:14 * *  Bitbake build for python-simplejson Jun 21 04:41:16 * * http://bugs.openembedded.org/show_bug.cgi?id=2460 Jun 21 04:49:00 * * OE Bug 2282 has been RESOLVED (FIXED) by mickeyl(AT)handhelds.org Jun 21 04:49:02 * *  python_2.4.4 packaging QA issue Jun 21 04:49:04 * * http://bugs.openembedded.org/show_bug.cgi?id=2282 Jun 21 05:00:02 03mickeyl 07org.oe.dev * r53dc3050... 10/ (1 packages/e17/rage_0.2.0.003.bb): add rage, the Enlightenment Media Center Jun 21 05:00:08 03mickeyl 07org.oe.dev * r56187b1b... 10/ (4 files in 3 dirs): python 2.4.4 properly ship debug information in appropriate packages Jun 21 05:00:16 03mickeyl 07org.oe.dev * r21548369... 10/ (3 files in 3 dirs): add python-setuptools. closes #2459 Jun 21 05:00:24 03mickeyl 07org.oe.dev * re51ad54c... 10/ (3 files in 3 dirs): add python-simplejson. closes #2460 Jun 21 05:00:32 03mickeyl 07org.oe.dev * rb87ef6b9... 10/ (1 packages/python/python_2.4.4.bb): python 2.4.4 ship manpage. (finally) closes #2282 Jun 21 05:10:00 mickeyl : Good morning ! Started early i see :) Jun 21 05:40:01 * * OE Bug 2466 has been RESOLVED (FIXED) by mickeyl(AT)handhelds.org Jun 21 05:40:01 yeah Jun 21 05:40:03 * *  Bitbake build for python-fam Jun 21 05:40:05 * * http://bugs.openembedded.org/show_bug.cgi?id=2466 Jun 21 05:40:10 early bird catches the worm... Jun 21 05:42:00 * * OE Bug 1581 has been RESOLVED (FIXED) by mickeyl(AT)handhelds.org Jun 21 05:42:02 * *  distutils-base.bbclass include .debug files in main package Jun 21 05:42:04 * * http://bugs.openembedded.org/show_bug.cgi?id=1581 Jun 21 05:43:23 mickeyl|fic : it looks like they were bugs and not worms :) Jun 21 05:44:02 that too :D Jun 21 05:44:10 heh we made the news headlines (again)-> http://www.linuxdevices.com/news/NS9122739256.html Jun 21 05:45:23 awesome! Jun 21 05:45:25 good PR for us Jun 21 05:59:24 03mickeyl 07org.oe.dev * r9a039e37... 10/ (3 files in 3 dirs): add python-fam, bindings for the file alteration monitor. closes #2466 Jun 21 06:39:01 * * OE Bug 2541 has been created by jake(AT)cogini.com Jun 21 06:39:03 * * diet-x11 Jun 21 06:39:05 * * http://bugs.openembedded.org/show_bug.cgi?id=2541 Jun 21 06:54:01 * * OE Bug 2542 has been created by jake(AT)cogini.com Jun 21 06:54:03 * * Allow glibc to build with i686 Jun 21 06:54:05 * * http://bugs.openembedded.org/show_bug.cgi?id=2542 Jun 21 07:12:49 good morning all Jun 21 07:14:39 http://labs.vivi.eng.br/blog/wp-content/uploads/2007/06/building-mamona.pdf Jun 21 07:20:45 cool Jun 21 07:22:09 hey mickey|fic Jun 21 07:22:23 good morning koen Jun 21 07:22:59 * koen looks at the e updates Jun 21 07:23:08 maybe it actuall builds now... Jun 21 07:26:15 i worked on that efl revamp the last couple of days, dozens of hours Jun 21 07:26:36 now that all uses pkgconfig and has more sane autotools usage, i could remove a lot of insane stuff Jun 21 07:27:34 justin did some really strange things Jun 21 07:32:10 efl was in the same boat as slugos Jun 21 07:32:40 weird and seemingingly unfixable voodoo that makes me scream Jun 21 07:32:46 :D Jun 21 07:32:55 seemingly. Jun 21 07:33:29 20 lines of python voodoo instead of 2 variable assignments, etc Jun 21 07:33:43 "there must be a good reason for doing it this way" Jun 21 07:39:21 sometimes there is Jun 21 07:39:23 sometimes not Jun 21 07:40:14 most of the times it's a leftover from ancient OE time Jun 21 07:40:15 s Jun 21 07:46:29 mickey|fic: could you do rename python debug packages to -dbg? Jun 21 07:46:44 mickey|fic: and _please_ inherit insane.bbclass Jun 21 07:47:04 hmm, for now it's python-core-dbg Jun 21 07:47:11 since there is no python.ipk that matches Jun 21 07:47:23 ERROR: QA Issue: non debug package contains .debug directory: python-core-debug path /work/armv4t-angstrom-linux-gnueabi/python-2.4.4-ml2/install/python-core-debug/usr/bin/.debug/python2.4 Jun 21 07:47:25 but i can reconsider that Jun 21 07:47:45 uhm. that check is very conservative Jun 21 07:47:52 it only looks for ${PN}-dbg ? Jun 21 07:48:02 yes Jun 21 07:48:05 hmm Jun 21 07:48:15 i don't like that. it should look for *-dbg Jun 21 07:48:15 since package.bbclass adds some magic for those packages Jun 21 07:48:35 it says the package is called 'python-core-debug' Jun 21 07:48:41 not python-core-dbg Jun 21 07:48:41 ya, that's wrong anyways Jun 21 07:48:43 good morning Jun 21 07:48:48 should be named python-core-dbg Jun 21 07:48:57 it looks for *-dbg Jun 21 07:49:08 ok, good. i'll make the change now then. Jun 21 07:49:08 not PN-dbg as I said earlier Jun 21 07:49:21 and inherit insane.bbclass :) Jun 21 07:51:48 yes, when i'm back in FFM Jun 21 07:51:50 :D Jun 21 07:53:50 ok, python-core-dbg is it then. change done, will push later Jun 21 08:18:02 hello room Jun 21 08:19:05 does OE run on Pocket PC? Jun 21 08:20:15 not yet. Jun 21 08:20:35 he's gone, but some ipaqs are compatible, aren't they ? Jun 21 08:21:20 yes, but's that a totally different thing Jun 21 08:21:33 to me, PocketPC is not a hardware category Jun 21 08:21:36 i guess to him neither Jun 21 08:22:13 besides OE not running anywhere, except you really mean the buildsystem and not the images generated by it, but that's yet another story of impriceness Jun 21 08:22:16 heh Jun 21 08:24:11 sigh building on a laptop sucks Jun 21 08:24:15 sure Jun 21 08:24:54 * theBrave think that if building on a laptop sucks, you would't want to buil on you smartphone Jun 21 08:25:30 heh, yeah Jun 21 08:26:00 Mem: 2066748k total, 1947732k used, 119016k free, 170024k buffers Jun 21 08:30:58 peeps, I need help please Jun 21 08:31:13 I start my device, but I don't get login.... Jun 21 08:32:00 you'll have to take a look at linux traces Jun 21 08:32:48 most likely if kernel boot and you don't get login you have screwed something with the rootfs Jun 21 08:52:07 .me just fucked up his laptop Jun 21 08:52:10 * steliosk just fucked up his laptop Jun 21 08:52:59 i wonder what ice cold coffee does to motherboards Jun 21 08:54:01 except short circuiting them Jun 21 08:54:04 :( Jun 21 08:56:34 don't turn it back on. wash it with water with detergent. wash it with distilled water. let it dry for a week. try it then. Jun 21 08:57:04 never tried that myself, but it's been known to work for some people. Jun 21 08:59:41 03pfalcon 07org.oe.dev * r8dcbee44... 10/ (1 packages/linux/linux-handhelds-2.6_2.6.21-hh9.bb): (log message trimmed) Jun 21 08:59:41 linux-handhelds-2.6: Add 2.6.21-hh9. Jun 21 08:59:41 * Changes since -hh8: Jun 21 08:59:41 1. aximx50: TS driver fixes. Jun 21 08:59:41 2. htc-egpio: New common driver for HTC device egpio chip. Jun 21 08:59:43 3. htcapache, htcmagician: Use htc-egpio Jun 21 08:59:45 4. htcuniversal: Add ts-adc-debounce support (need testing) Jun 21 08:59:47 03pfalcon 07org.oe.dev * r2b710e18... 10/ (1 conf/distro/angstrom-2007.1.conf): angstrom-2007.1.conf: Update linux-handhelds-2.6 to 2.6.21-hh9. Jun 21 09:04:52 03mickeyl 07org.oe.dev * r805f0a5b... 10/ (5 files in 2 dirs): libftdi[-native] 0.8 update to 0.9 (from OpenMoko) Jun 21 09:04:58 03mickeyl 07org.oe.dev * r8a4932c9... 10/ (3 files in 3 dirs): python 2.4.4 fix debug package name to python-core-dbg Jun 21 09:05:02 03mickeyl 07org.oe.dev * r3593b915... 10/ (1 packages/libftdi/files/doxygen-configure.patch): libftdi: update doxygen configure patch Jun 21 09:09:25 hmm Jun 21 09:09:27 | mv: cannot stat `/local/pkg/oe/fic-gta01/tmp/work/armv4t-angstrom-linux-gnueabi/perl-5.8.8-r22/image//usr/lib/perl/5.8.8/CORE/libperl.so': No such file or directory Jun 21 09:09:28 ? Jun 21 09:12:19 hi Jun 21 09:12:59 ~lart perl Jun 21 09:13:00 * ibot shoots perl in the head Jun 21 09:13:16 hey hrw Jun 21 09:18:50 03koen 07org.oe.dev * r0af25356... 10/ (3 files in 2 dirs): pidgin: update to 2.0.2 Jun 21 09:18:55 03koen 07org.oe.dev * ra481e13b... 10/ (1 packages/irssi/irssi_svn.bb): irssi: def pref -1 for svn Jun 21 09:22:36 mickey|fic: have you considered using telepathy in openmoko as the messaging/IM backend? Jun 21 09:23:01 * * OE Bug 2541 has been RESOLVED (FIXED) by Jun 21 09:23:03 * *  diet-x11 build fails, can't find patch Jun 21 09:23:05 * * http://bugs.openembedded.org/show_bug.cgi?id=2541 Jun 21 09:27:15 koen: no. then again, i have not considered any messaging backend yet ;) Jun 21 09:28:04 koen: my task in the next 2 weeks -- starting with when i'm back -- is writing up a report about our future software roadmap Jun 21 09:28:10 that will include messaging/IM Jun 21 09:28:57 mickey|fic: if you include avahi as well, you'll be GMAE compliant :) Jun 21 09:29:10 * koen mentions "EABI" again Jun 21 09:29:32 03polyonymous 07org.oe.dev * rd1eed5f4... 10/ (3 files in 3 dirs): packages/xorg-lib/: move ruutf8.patch and close bug #2541 Jun 21 09:30:05 well, it's from lennart, so i'll definitely take a look at that Jun 21 09:31:35 mickey|fic: I can do 'ssh neo1973.local' without having to check if I'm connected via bluetooth, usb, wifi, ethernet, etc Jun 21 09:31:49 coolness Jun 21 09:32:10 koen: BTW, we join you in the formula1. :) Jun 21 09:32:20 stefan_schmidt: :) Jun 21 09:33:04 koen: avahi is essential at wolfson so I can find the devboards quickly Jun 21 09:33:35 its also good for stealing peoples itunes libraries Jun 21 09:33:36 mickey|fic: telepathy + empathy works ok in poky Jun 21 09:33:49 in angstrom as well :) Jun 21 09:35:11 hrw: speaking of poky, I think python-zopeinterfaces is missing some bits, I need the full zope to get flumotion to work Jun 21 09:36:05 mickey|fic: should we schedule a formal OE-core meeting at guadec? Jun 21 09:37:51 koen: I am afraid that I cannot show you how I made flumotion working. it is too awful and will give you nightmares Jun 21 09:38:36 well, the one in .dev works, provides you have full zope installed Jun 21 09:39:08 provided* Jun 21 09:40:00 koen: I had to fit in <14M Jun 21 09:40:10 I stopped working on it because flumotion-admin + flumotion-manager + flumotion-worker == 64MB of ram Jun 21 09:40:27 and python is broken in ubuntu feisty, so.... Jun 21 09:40:40 koen: you do not need them to get flumotion working if you know what you want to do Jun 21 09:41:08 and I want to not talk about flumotion - it was nightmare project Jun 21 09:43:23 koen: btw, the AVR32 package looks indeed cute. :) Jun 21 09:44:45 I have the most important package: http://www.angstrom-distribution.org/unstable/feed/avr32/base/irssi_0.8.10-r2_avr32.ipk :D Jun 21 09:44:56 heh Jun 21 09:44:59 koen: without perl support? Jun 21 09:45:15 * stefan_schmidt hopes to play with the board at the we Jun 21 09:53:04 03mickeyl 07org.oe.dev * rfe6e384f... 10/ (1 packages/libid3 packages/libid3/libid3_1.1.bb): add libid3, (yet) another library for reading out id3 tags Jun 21 09:53:11 03mickeyl 07org.oe.dev * raa2d27a2... 10/ (5 files in 3 dirs): openocd[-native] update recipe, add link-static patch, put into deploy directory (from openmoko) Jun 21 09:53:18 03mickeyl 07org.oe.dev * r927140d7... 10/ (6 files in 3 dirs): s3c2410-utils: remove fixed date versions, keep svn versions Jun 21 09:53:24 03mickeyl 07org.oe.dev * rc49be424... 10/ (3 files in 3 dirs): add sidplay-base, a spartanic console sidplayer based on libsidplay 1.x (from OpenMoko) Jun 21 09:54:44 03mickeyl 07org.oe.dev * rd66ccd7e... 10/ (4 files in 4 dirs): initscripts: add seperate inittab for FIC GTA01 (ttySAC multiplexing...) Jun 21 09:59:47 03mickeyl 07org.oe.dev * re89c0961... 10/ (1 conf/machine/fic-gta02.conf conf/machine/fic-gta01.conf): fic-gta0x configurations: update and add note Jun 21 10:18:45 hmm, this u-boot directory is a mess. i'm not going to clean that up Jun 21 10:19:30 mickey|fic: uboot situation stinks Jun 21 10:19:48 gta01 uboot stinks even more - not buildable with eabi Jun 21 10:21:35 what error do you get? Jun 21 10:21:38 * koen points hrw and mickey|fic to their inbox Jun 21 10:23:01 * * OE Bug 2543 has been created by lukas.gorris(AT)gmx.de Jun 21 10:23:03 * * no icq in pidgin Jun 21 10:23:05 * * http://bugs.openembedded.org/show_bug.cgi?id=2543 Jun 21 10:23:35 morning Jun 21 10:23:57 koen: I pointed that to mallum. Jun 21 10:24:29 hi Jun 21 10:24:36 hola saladino Jun 21 10:24:47 hrw, is th e problem linking u-boot? Jun 21 10:24:58 hrw, may I try and test your hostap patch again? Jun 21 10:24:59 why wo we need this gta01-mkimage btw? Jun 21 10:25:26 aiva Jun 21 10:25:28 mickey|fic: because uboot is a steaming pile of fragile nonbuildable crap Jun 21 10:25:29 hola zecke Jun 21 10:25:42 Crofton: I do not remember what was a problem with it Jun 21 10:25:48 koen: ah. it's just that we don't need it in the OM overlay, so i wonder why Jun 21 10:25:49 mickey|fic: it only builds on full moon with the right compiler Jun 21 10:25:59 mickey|fic: fic uboot need OABI Jun 21 10:26:06 mickey|fic: fails with EABI Jun 21 10:26:27 "needless hardcoding of unneeded gcc options" comes to mind Jun 21 10:26:27 I have issues with the u-boot I am working with linking Jun 21 10:26:54 i.e -macps-32 Jun 21 10:27:00 hrw: I start to hate your government. Can you please impeach it? :} Jun 21 10:27:06 03koen 07org.oe.dev * r7c69d6f8... 10/ (1 packages/efl1/evas.inc): evas.inc: don't depend on directfb Jun 21 10:27:23 hrw, I was not able to install your patchfile with ipkg Jun 21 10:27:27 koen: uhm Jun 21 10:27:31 koen: what are you doing? Jun 21 10:27:38 zecke: why now? they are assholes from beginning Jun 21 10:27:42 mickey|fic: fixing the build Jun 21 10:27:43 koen: i want the directfb backend always to be built Jun 21 10:27:54 directfb doesn't compile for various archs Jun 21 10:28:06 e.g.? Jun 21 10:28:20 * Crofton is glad to see people in other countires hate their governments also :) Jun 21 10:29:20 powerpc and avr32 are the ones that failed so far Jun 21 10:29:25 hmm Jun 21 10:29:42 so you want me to add a evas-directfb, splitting it up again and making it messy again? Jun 21 10:30:46 koen, hi, OpenEmbedded/build/tmp/work/armv5te-angstrom-linux-gnueabi/glibc-2.6-r0/locale-tree//usr/share/i18n/locales/am_ET:584: LC_COLLATE: section `LATIN' not known ... is this a broken locale in glibc2.6 ? Jun 21 10:31:15 Marex: you alter the config, so you know Jun 21 10:31:36 mickey|fic: let me see if I can fix directfb to compile Jun 21 10:31:50 koen: replied Jun 21 10:31:51 ya, that sounds like the better option :D Jun 21 10:32:03 * koen inspects the error message Jun 21 10:32:11 alternatively, we can do DEPENDS_whereitdoesnt work Jun 21 10:32:26 and altering EXTRA_OECONF Jun 21 10:32:33 although that*s really only the 2nd best option Jun 21 10:32:35 ah, it seems like the old libpng screwage all over again Jun 21 10:32:41 * koen rebuilds libpng Jun 21 10:34:04 or python __anonymous() with list of working archs Jun 21 10:35:27 koen, pardon? I didnt really get what you said Jun 21 10:39:59 hi everybody Jun 21 10:40:06 hi thesing Jun 21 10:40:57 koen, ok, removing the locale helped ... it probably was really broken (am_ET and byn_ER) ... it'd be nice if you can check it Jun 21 10:41:09 file a bug Jun 21 10:41:21 hi thesing Jun 21 10:42:10 so let me get this straight Jun 21 10:42:19 uboot-EABI is a general problem or only in gta01 version? Jun 21 10:42:46 mickey|fic: depends on the baord you are building for Jun 21 10:42:52 ah. Jun 21 10:43:02 so it might be handy to inform upstream Jun 21 10:43:03 did anyone yet? Jun 21 10:43:06 mickey|fic: since each board has its own makefile duplication the wrong options Jun 21 10:43:43 ~lart OE for autoload only in kernel.bbclass Jun 21 10:43:44 * ibot says "boot to the head" and knocks OE over for autoload only in kernel.bbclass Jun 21 10:45:04 can someone pastebin the u-boot error with EABI? Jun 21 10:45:09 sure thing: Jun 21 10:45:10 here it comes Jun 21 10:45:11 I want to see if we share a problem Jun 21 10:46:19 http://pastebin.ca/581043 Jun 21 10:47:08 hmm Jun 21 10:47:16 I don't have that problem Jun 21 10:47:46 but I think you have built one .o with OABI and the other with EABI Jun 21 10:49:04 http://pastebin.ca/581046 Jun 21 10:49:08 is what I get Jun 21 10:49:18 when I start trying to add NAND support Jun 21 10:49:51 doesn't look all that different Jun 21 10:49:58 yeah Jun 21 10:50:06 similar but differeent Jun 21 10:50:30 in my case, one problem is it is getting symbols from libgcc (which I think is bad?) Jun 21 10:56:37 mickey|fic: http://ewi546.ewi.utwente.nl/tmp/ChangeLog-20070614-20070621.txt Jun 21 11:00:22 03koen 07org.oe.dev * rf5d5ad96... 10/ (1 packages/efl1/evas.inc): evas.inc: add directfb back to DEPENDS, it turned out to be a problem in libpng Jun 21 11:01:46 koen: nice. Jun 21 11:01:47 however Jun 21 11:02:05 it'd be even nicer, if you could group it by directory as well Jun 21 11:02:06 e.g. Jun 21 11:02:07 perl 5.8.8: Don't set LD_LIBRARY_PATH when running miniperl. This was being Jun 21 11:02:08 perl 5.8.8: Directly patch the files that need patching instead of having a Jun 21 11:02:08 perl 5.8.8: Fix the regexp for removing /usr/local paths - it was eating the Jun 21 11:02:15 merging that to Jun 21 11:02:16 yeah Jun 21 11:02:17 perl 5.8.8: Jun 21 11:02:18 * foo Jun 21 11:02:19 * bar Jun 21 11:02:26 would that be doable? Jun 21 11:02:36 haven't looked at that yet :) Jun 21 11:02:36 yo cedric Jun 21 11:02:38 hehe Jun 21 11:02:48 the script is in contrib/mtn2cl Jun 21 11:04:07 mickey|fic: I want OE to start doing (tarball) releases ASAP Jun 21 11:07:35 how can i know why bibtake is choosing one filepath before than another for a config file? for koen with angstrom uclibc chooses the correct config file at the archs subdir and for me with ucslugc it doesn't , and there is no FILESPATH explicitly declared at uclibcs bb files Jun 21 11:07:52 can be modified by distros or machines vars? Jun 21 11:08:26 saladino: does ucslugc has distro part of config done? Jun 21 11:08:30 saladino: Take a look at the manual and search for the override Jun 21 11:08:47 hm it does not need Jun 21 11:08:51 koen: me too Jun 21 11:08:51 zecke but i shouldn't need to override Jun 21 11:08:52 me too Jun 21 11:08:56 slapin: then take a look at the bitbake.conf OVERRIDES variable to see the order and then guess if it is used from left to right or right to left Jun 21 11:09:01 the conf is the same Jun 21 11:09:06 hrw yes Jun 21 11:09:07 it has Jun 21 11:09:33 if i set the correct filespath var in uclibc-0.9.29.bb it works well Jun 21 11:10:46 it needs to get from uclibc-0.9.29/ixp4xxbe/uClibc.machine but it gets from files/uClibc.machine Jun 21 11:10:50 and that is just a template Jun 21 11:11:41 if i put in the filespath ${FILE_DIRNAME}/uclibc-${UCLIBC_BASE}/${MACHINE_ARCH}' first it does well Jun 21 11:11:54 but for koen works fine without that Jun 21 11:12:50 so i supose that must be something from the distro or machine config that makes the change Jun 21 11:14:42 bbl Jun 21 11:14:53 saladino: "bitbake -e uclibc" output please (via pastebin) Jun 21 11:15:30 hi all Jun 21 11:16:01 * * OE Bug 2544 has been created by joaquinduran(AT)adtelecom.es Jun 21 11:16:03 * * Updating gtkmm, glibmm, cairomm version of libraries Jun 21 11:16:05 * * http://bugs.openembedded.org/show_bug.cgi?id=2544 Jun 21 11:16:37 hi florian Jun 21 11:22:25 hrw http://pastebin.com/933287 Jun 21 11:23:39 saladino: you want nslu2be not ixp4xxbe Jun 21 11:23:43 OVERRIDES="local:nslu2be:nslu2:ixp4xx:ucslugc:linux-uclibc:armeb:build-linux:fail-fast:pn-bitbake:thumb:thumb-interwork:slugos" Jun 21 11:24:34 saladino: you have MACHINE="nslu2be" and complain about "ixp4xxbe" - decide.... Jun 21 11:24:43 ok, i fixed the difference in EABI levels Jun 21 11:24:48 however still have undefined references Jun 21 11:24:57 | /local/pkg/oe/fic-gta01/tmp/cross/lib/gcc/arm-angstrom-linux-gnueabi/4.1.2/libgcc.a(_dvmd_lnx.o): In function `__div0': Jun 21 11:24:57 | /local/pkg/oe/fic-gta01/tmp/work/armv4t-angstrom-linux-gnueabi/gcc-cross-4.1.2-r4/gcc-4.1.2/gcc/config/arm/lib1funcs.asm:1001: undefined reference to `raise' Jun 21 11:24:58 | /local/pkg/oe/fic-gta01/tmp/cross/lib/gcc/arm-angstrom-linux-gnueabi/4.1.2/libgcc.a(_divdi3.o):(.ARM.exidx+0x0): undefined reference to `__aeabi_unwind_cpp_pr0' Jun 21 11:24:58 | /local/pkg/oe/fic-gta01/tmp/cross/lib/gcc/arm-angstrom-linux-gnueabi/4.1.2/libgcc.a(_udivdi3.o):(.ARM.exidx+0x0): undefined reference to `__aeabi_unwind_cpp_pr0' Jun 21 11:25:37 this looks so familiar Jun 21 11:25:49 I think it comes from the nand stuff Jun 21 11:25:53 hummm, hrw i'm very newbie with this, i'll check why is setted like that, i just get the nslu2-linux config files from ucslugc and with some modifications it compiled and worked Jun 21 11:26:43 mickey|fic: tried editing the linker script? Jun 21 11:27:05 koen: i sedded all -mabi=foo out Jun 21 11:27:33 however i*m kind of surprised that still there are problems Jun 21 11:27:43 * Crofton isn't :( Jun 21 11:28:43 hrw in the nslu2be machine config sets the machine_arch to ixp4xxbe Jun 21 11:29:14 so i suposed that nslu2be is a ixp4xxbe machine Jun 21 11:30:10 * mickey|fic needs to go. please ping me if you have solutions... Jun 21 11:32:28 saladino: it is ixp4xxbe machine but for OE it is nslu2be Jun 21 11:33:52 saladino: contact nslu2 maintainer and ask him why nslu2be is not ixp4xxbe Jun 21 11:36:53 ucslugc is deprecated Jun 21 11:37:14 rwhitby-n800 i know Jun 21 11:37:26 but i am working with a friend to make it work Jun 21 11:37:33 are we doing something wrong? Jun 21 11:37:44 no Jun 21 11:38:05 are you going to delete ucslugc from openmebedded tree? Jun 21 11:38:17 just setting the expectation of what you are working with Jun 21 11:38:50 not if you want it to stay and want to maintain it Jun 21 11:39:26 yep, im just trying to see if i am able to do that Jun 21 11:39:49 im a newbie at the embedded world Jun 21 11:40:08 just learning and i liked ucslugc Jun 21 11:40:57 it's not the best idea to learn on it Jun 21 11:41:04 why? Jun 21 11:41:20 learn on angstrom indstead Jun 21 11:41:44 then you can get help when required Jun 21 11:42:07 so i'll not get help if i work with ucslugc? Jun 21 11:42:49 unless you're looking for an uphill challenge where you are the sole person in the world supporting it Jun 21 11:43:35 my friend an i have taken this as a quest :), but i'll like to know if it is _really_ a bad idea Jun 21 11:43:50 no one else uses it, so what incentive do they have to help you? Jun 21 11:44:27 * koen hugs gnu-configize Jun 21 11:44:48 well, i thought that as ucslugc is slugos with uclibc there were many parts in common Jun 21 11:44:58 so i supose it wouldn't be too hard Jun 21 11:45:15 somebody asked at the list recently about ucslugc Jun 21 11:45:20 there are, but uclibc has historically not been well supported in OE. That may be changing, or it may be a transient improvement that declines again later. Jun 21 11:46:08 take openwrt, for instance. they *only* use uclibc, so you know that ucliubc is well supported if you are using openwrt. In OE, glibc is used by the majority of distros, so you know that glibc is supported well. Jun 21 11:46:40 if you want to go against the majority, then you need to expect an proportionally harder job. Jun 21 11:46:50 hmmm, yes, i supose that it'll be the easy way Jun 21 11:47:03 i don't fear the work, im here to learn Jun 21 11:47:08 I just don't want to see uclugc revived for a couple of months while you are interested in it, and then you go away and do something else and it falls into disrepair again. Jun 21 11:47:27 i understand you Jun 21 11:47:48 that's the main reason why it has been deprecated by nslu2-linux (which is not the same as deprecated by OE, but is effectively so if no-one else is interested in it either) Jun 21 11:48:27 there is plenty of work to do and to learn in the mainstream of OE, rather than working on something which is not going to benefit anyone else in OE due to your efforts. Jun 21 11:48:33 angstrom-uclibc works great on my ipaq and avr32 board Jun 21 11:48:56 so no need for ucslugc if angstrom-uclibc works Jun 21 11:49:04 koen: yep, and if someone wants to use uclibc, then I suggest they use angstrom-uclibc, not slugos-uclibc. Jun 21 11:49:18 (as I did above, by suggesting saladino use angstrom) Jun 21 11:50:02 ok, thanks, i'll talk with my friend and decide it, i don't know angstrom and need to see the differences Jun 21 11:50:57 you can help by getting sysconf and altboot to do the same in angstrom-uclibc, that turnup does in slugos. Jun 21 11:51:10 in case we decide to continue with ucslugc, will that be a problem for anybody apart from us? Jun 21 11:51:27 that would help both the angstrom community (by getting it working better on nslu2) and also help in the transition of the rest of the slugos community to something based on angstrom Jun 21 11:51:59 is the slugos community doing a transition to angstrom? i didn't know that Jun 21 11:52:12 there is no problem at all with you working on ucslugc, other than the frustration you will get when no-one here is interested (or has the slugos/uclibc knowledge required) to help you Jun 21 11:53:04 the same principle applies. why would we continue to develop a separate slugos distro, when we can simply base slugos 5 on the same basic configuration as angstrom, and benefit from all the work that all the other angstrom maintainers do Jun 21 11:53:48 (we will keep a separate slugos.conf distro file, cause that is where policy differences are set, but almost everything else will just reuse from angstrom) Jun 21 11:54:14 ok rwhitby, thanks, i see your practical view, maybe i just not thought much about it when i started with this Jun 21 11:54:29 note that this will not happen overnight. it's a long term plan for at least slugos5+ Jun 21 11:55:31 the best way to help, is to work out how to do things that slugos currently does but angstrom doesn't (like turnup and sysconf for the nslu2) and reimplement them in generic ways that all angstrom-based distros can use in OE. Jun 21 11:55:50 does angstrom-uclibc work on nslu2 now? Jun 21 11:55:59 I have made a start on that with sysconf, and CoreDump|afk is working on it for altboot replacing turnup. Jun 21 11:56:05 dunno, I haven't tried it. Jun 21 11:56:37 ok, i'll think about it Jun 21 11:56:44 must go Jun 21 11:56:47 bye Jun 21 11:56:51 I've been concentrating on openwrt for the nslu2, as I'm not convinced that uclibc in OE isn't just a passing fad :-) Jun 21 11:57:12 well, avr32 doesn't support glibc :) Jun 21 11:57:14 (and more openwrt for the dsmg600, nas100d and fsg3, rather than the nslu2) Jun 21 11:57:58 (but talking about openwrt is offtopic here, so I will stop) Jun 21 11:58:34 i liked uclibc because of the amount of free space that left in the device Jun 21 11:58:57 i didn't like to have a separate hd for work with the nslu Jun 21 11:59:06 im planning to put it in my car Jun 21 11:59:17 sysconf? Jun 21 11:59:28 03koen 07org.oe.dev * ref8c7228... 10/ (1 packages/db/db_4.3.29.bb): db 4.3.29 : run gnu-confize Jun 21 11:59:44 hrw: cpio'ing CONFGILES into a unused mtdblock Jun 21 11:59:44 hrw: yep, store network settings and all CONFFILES in a separate MTD partition, which can be easily restored if you reflash the rootfs. Jun 21 11:59:58 my nslu2 is on 2.5" hdd case Jun 21 12:00:04 rwhitby: nice Jun 21 12:00:27 reflash, and it reloads all your modifications on the first boot. if the conffile has changed, then it provides the normal upgrade, keep, diff question that ipkg gives Jun 21 12:00:53 very handy for ssh host keys, for instance. Jun 21 12:01:07 and /home/root/.ssh/authorized_keys Jun 21 12:01:22 * rwhitby hates typing passwords after reflashing devices Jun 21 12:01:22 must go again :) Jun 21 12:01:24 bye Jun 21 12:08:20 hi ph5 Jun 21 12:09:06 hey woglinde Jun 21 12:34:50 sweet Jun 21 12:34:54 1.4M avr32/gtk+_2.10.12-r2_avr32.ipk Jun 21 12:35:05 huh? Jun 21 12:35:10 so tiny? Jun 21 12:35:15 * koen hugs autotools site files Jun 21 12:35:40 ~lart libapm for apm_exists return codes Jun 21 12:35:40 * ibot runs at libapm with an origami Swiss Army knife, and inflicts a nasty paper cut for apm_exists return codes Jun 21 12:35:41 yes Jun 21 12:37:16 woglinde: 1.7M ../armv5te/base/gtk+_2.10.12-r2_armv5te.ipk Jun 21 12:37:24 gtk isn't that big Jun 21 12:39:52 time for av32 pda :-) Jun 21 12:40:24 heh you never know Jun 21 12:40:55 most likely there will be a avr32 mediaplayer Jun 21 12:41:17 koen: got dsp stuff? Jun 21 12:41:27 yes Jun 21 12:41:33 koehn hm yes I forget you need atk pango and serval other libs for working Jun 21 13:56:22 hey people, how can I be sure my rootfs is ok? Jun 21 13:56:30 hm Jun 21 13:56:34 still nfsroo problem? Jun 21 13:56:39 woglinde: yes Jun 21 13:56:59 I checked it with my kernel and vendor's provided rootfs - and it works Jun 21 13:57:04 hm do you see the nfs questions on he host? Jun 21 13:57:13 ah okay Jun 21 13:57:25 hm Jun 21 13:57:32 maybee wrong glibc format Jun 21 13:57:34 or something Jun 21 13:57:36 when I'm trying to load my rootfs, its like init never starts Jun 21 13:57:44 could you try file Jun 21 13:57:50 file? Jun 21 13:57:54 file somefile from vendor Jun 21 13:58:01 file somefile Jun 21 13:58:10 file is a commando Jun 21 13:58:27 to see which endianess and arch the file is for Jun 21 13:58:48 well, the arc is arm Jun 21 13:58:54 than file some_file_from_vendor_rootfs Jun 21 13:59:07 and file some_file_from_your_rootfs Jun 21 13:59:18 and look if it differs Jun 21 13:59:45 it sure does Jun 21 14:00:03 hm can ypu paste it? Jun 21 14:00:22 I've build mine with OE, they had Debian Jun 21 14:00:37 hm Jun 21 14:00:39 ah Jun 21 14:00:44 maybee eabi error Jun 21 14:01:14 well, it seems that kernel is alive, and connecting USB devices shows on console... Jun 21 14:01:23 just init doesn't start Jun 21 14:01:26 yeah Jun 21 14:01:44 hm can you look in the kernel config if eabi is enabled Jun 21 14:01:58 on which I compiled? Jun 21 14:02:01 thats nothing the kernel did startet some kerneld Jun 21 14:02:08 yes Jun 21 14:02:13 the kernel you now run Jun 21 14:03:27 ah, ok Jun 21 14:03:42 not the host kernel Jun 21 14:03:47 the kernel for device ;) Jun 21 14:05:22 woglinde: sec Jun 21 14:05:31 morning Jun 21 14:05:37 hm be quick I will heading home Jun 21 14:05:40 hi choiumat Jun 21 14:05:53 woglinde: nope, its not enabled Jun 21 14:06:00 hehe Jun 21 14:06:06 CONFIG_AEABI is not set Jun 21 14:06:13 than build a new one and enable it Jun 21 14:06:23 than init should work Jun 21 14:06:31 hm Jun 21 14:06:36 but then it means that the other rootfs which works is an OABI, right? Jun 21 14:06:42 can bee Jun 21 14:06:45 I only guess Jun 21 14:07:02 yea, it is Jun 21 14:07:08 ah Jun 21 14:07:10 * djay-il is slow Jun 21 14:07:15 okay fine we figure it out Jun 21 14:07:31 so I'll recompile kernel and try again... Jun 21 14:07:31 so I am now heading home Jun 21 14:07:34 thanks Jun 21 14:07:34 bye Jun 21 14:07:38 bye Jun 21 14:08:44 hmm, ERROR: No providers of runtime build target python-misc-dbg Jun 21 14:17:24 * cbrake removes that from his DISTRO_EXTRA_RDEPENDS Jun 21 14:18:04 looks like it is now python-core-dbg Jun 21 14:19:51 cbrake: mickeyl found time to look at python stuff Jun 21 14:21:42 cbrake: "bitbake python-misc-dbg -DDD" will tell you why there is no provider Jun 21 14:29:04 03koen 07org.oe.dev * r62812734... 10/ (1 packages/tasks/task-boot.bb): task-boot: task-boot is meant to provide the smallest possible set to get an image booting to a login prompt, so additional wishes should be in another task, or specified in the PACKAGE_INSTALL or the desired image. Jun 21 14:31:40 koen: thx. I will look and merge my changes Jun 21 14:32:29 * koen ponders about machines that need uImage in the rootfs Jun 21 14:34:31 koen: read.jffs2 exists for a reason Jun 21 14:35:31 koen , hrw : Removing the MACHINE_EXTRA_ from task-boot makes generic machines like i596 and i686 useless Jun 21 14:35:45 no Jun 21 14:35:49 * koen points to the commit message Jun 21 14:36:04 steliosk: add kernel-modules into ESSENTIAL for them Jun 21 14:37:46 koen : having a small image is the goal, but a small image that works Jun 21 14:38:14 steliosk: MACHINE_ESSENTIAL_* are for it Jun 21 14:38:22 koen, I found about four broken locales in glibc 2.6 now ... Jun 21 14:38:39 Marex: and I say again, file a bug Jun 21 14:39:04 koen, may I have just a question? Jun 21 14:39:19 I still dunno whether its problem on my side or not Jun 21 14:39:53 have you ever met this? Jun 21 14:39:54 /home/marex/OpenEmbedded/build/tmp/work/armv5te-angstrom-linux-gnueabi/glibc-2.6-r0/locale-tree//usr/share/i18n/locales/am_ET:584: LC_COLLATE: section `LATIN' not known Jun 21 14:39:54 NOTE: Task failed: localedef returned an error Jun 21 14:40:21 then there is some long unimportant stuff ... well I dunno if its on my side or OEbug Jun 21 14:41:00 lib/time.c:60: error: too few arguments to function 'gettimeofday' Jun 21 14:41:04 * koen stabs uclibc Jun 21 14:41:50 koen: http://blog.haerwu.biz/2007/06/21/extracting-diffs-from-vendor-kernels/ Jun 21 14:42:25 hmmm NX 3.0 is out ... Jun 21 14:42:30 hrw: iirc there's dont-diff file for the kernel Jun 21 14:42:34 oh well :-T ... Ill remove all the in my opinion broken locales then Jun 21 14:43:49 hrw: are you sure you want to look at their ugly hacks? Jun 21 14:44:00 chouimat: I have to sometimes Jun 21 14:44:16 chouimat: devboard on my desk use vendor kernel Jun 21 14:44:44 hrw: I know sometimes you have no choice ... Jun 21 14:45:49 zecke: ping Jun 21 14:45:55 pong Jun 21 14:46:37 ~kill u-boot Jun 21 14:46:37 * ibot shoots a super-inverse pseudophoton gun at u-boot Jun 21 14:47:25 is this gun better than the BFG-9000? Jun 21 14:47:57 zecke: oh.. you are fast :-) got a positive answer from Joey: Seems like they could help us if we want :) Jun 21 14:48:24 ~zap u-boot Jun 21 14:48:26 * ibot takes out a cattle prod and gives u-boot a good jolt. Jun 21 14:48:29 heh Jun 21 14:48:35 mickey|fic, I see we share a common enemy Jun 21 14:48:41 florian: I have two days off from work ;) Jun 21 14:48:45 mickey|fic: I see you're having fun Jun 21 14:48:49 the u-boot ml is very helpful either Jun 21 14:49:02 is? Jun 21 14:49:05 is'nt Jun 21 14:49:07 is not Jun 21 14:49:09 oops Jun 21 14:49:10 ya Jun 21 14:49:16 "use an older toolchain" Jun 21 14:49:18 big deal, thanks Jun 21 14:49:32 03hrw 07org.oe.dev * rb0652ec0... 10/ (1 packages/tasks/task-boot.bb): task-boot: bump PR after last changes Jun 21 14:49:48 zecke: We would have to prepare a short proposal and should send this to their steering committee. Jun 21 14:49:54 mickey|fic: they seems to be as helpful as the nice openwrt guys :) Jun 21 14:50:21 hmm we want to hire one of those Jun 21 14:50:26 aren't they any good? Jun 21 14:50:31 mickey|fic: well... port bootldr or redboot to a new device and understand ;) Jun 21 14:50:53 florian: sounds like what SPI requires. So they will be able to take donations? Have the "Gemein-Nuetzigkeit" and some/full control on the money :) Jun 21 14:50:54 mickey|fic: one of uboot or openwrt? Jun 21 14:50:57 no, that hiring comment referred to openwrt guys Jun 21 14:51:15 the u-boot guys are cool, just a bit sucky at communicating sometimes Jun 21 14:51:15 zecke: obviously... Jun 21 14:51:18 mickey|fic: they are good ... some of them have no people skills at all on the ml and the irc channel Jun 21 14:51:33 guess that's a habit with most devels Jun 21 14:51:37 florian: I think we should 'try' it? Jun 21 14:52:07 is there a u-boot irc channel? Jun 21 14:52:10 mickey|fic: and a few also have the primadonna syndrom ... Jun 21 14:52:18 mickey|fic: i guess most of the u-boot people are ppc instead of arm hackers. that makes these things a little bit complicated sometimes.. Jun 21 14:52:38 yeah, the arm guy seems overworked Jun 21 14:53:05 zecke: It would be worth a try, especially because it will save us some nasty setup and maintainance overhead. Jun 21 14:53:13 :) Jun 21 14:53:20 mickey|fic, where are the abi flags again? Jun 21 14:53:39 no idea, i just sed'ed all -mabi=acps out Jun 21 14:53:43 in all Makefile and config.mk Jun 21 14:53:59 no longer EABI version conflicts then Jun 21 14:54:10 only thing missing is the duplicated symbols now Jun 21 14:54:14 and/or missing ones Jun 21 14:54:36 can't find anything to cure that though :/ Jun 21 14:54:40 yeah, seems like some kind of linkage issue with libgcc? Jun 21 14:56:44 shrug I'm all but a toolchain guy. where's one when you need one Jun 21 14:56:47 ~seen pb_ Jun 21 14:56:50 pb_ was last seen on IRC in channel #handhelds.org, 70d 21h 21m 19s ago, saying: 'bd2: hi'. Jun 21 14:56:55 mmh Jun 21 14:56:57 oh Jun 21 14:57:26 ok, missing flsah stuff Jun 21 14:57:29 and this Jun 21 14:57:35 He seems to be too busy to answer emails since ~fosdem Jun 21 14:58:04 http://pastebin.ca/581438 Jun 21 14:58:11 i mean the thing that annoys me is that it just works with oabi Jun 21 14:58:11 why Jun 21 14:58:54 Crofton: yes, minus the flash i'm getting now exactly the same Jun 21 14:58:55 hmm... iirc i compiled it (for pxa) with an eabi toolchain Jun 21 14:59:00 ok Jun 21 14:59:03 but thats some time ago Jun 21 14:59:06 I am still working on flsh Jun 21 14:59:23 I use an eabi tool chain to compile it for OSK also Jun 21 15:00:35 hmm Jun 21 15:00:37 http://www.nabble.com/Re:-gcc-4.1.2-eabi-softfloat-toolchain,-undefined-reference-to-%22raise%22-p10253066.html Jun 21 15:00:42 claims it to be a gcc bug Jun 21 15:00:45 with a fix somewhere Jun 21 15:01:18 good find mickey|fic Jun 21 15:01:59 florian: should I mail the OE ml to ask for final comments? Jun 21 15:02:37 now that's fun, i always only find references to the patch, but not the patch itself Jun 21 15:02:37 heh Jun 21 15:02:40 internet is fun Jun 21 15:03:01 zecke: yes please Jun 21 15:03:17 * cbrake is liking the size of uclibc images :-) Jun 21 15:04:26 :) Jun 21 15:13:34 are there modules for using usb gamepads in the kernel? Jun 21 15:16:47 which kernel? Linux? Hurd? XNU? Jun 21 15:21:49 TRying to connect with the serial cable to my spitz, I now get "cu: /dev/ttyS0: Line in use". lsof /dev/ttyS0 returns nothing. How do I free ttyS0 without rebooting my laptop which would be inconvenient right now Jun 21 15:22:49 if nothings using it, it might free up after a few mins Jun 21 15:23:04 Why was it taken in the first place? Jun 21 15:23:09 I never use the serial port really Jun 21 15:23:11 no idea? :) Jun 21 15:23:19 So I doubt it will get freed Jun 21 15:23:52 mickey|fic, if you solve the eabi thing, can you let me know? Jun 21 15:24:08 I will be on vacation from June 23 - July 4 and likely hard to contact Jun 21 15:24:18 off to the Dolomites again Jun 21 15:27:53 Laibsch: dont do cu as root :-) Jun 21 15:32:50 cu Jun 21 15:32:56 XorA: I did not Jun 21 15:33:18 XorA: I just used another, more cooperative computer now and that did it ;-) Jun 21 15:34:55 zecke, err the one I have on my zaurus :) Jun 21 15:35:43 mickey|fic, ping Jun 21 15:37:45 Crofton: sure, i'll let ya know Jun 21 15:37:58 dolomites? Jun 21 15:37:59 fun Jun 21 15:38:08 I think I need to make a board specific nand file? Jun 21 15:38:09 * mickey|fic almost forgot the meaning of vacations Jun 21 15:38:10 yeah Jun 21 15:38:37 * Crofton isn't ready, to much infinished work Jun 21 15:38:40 hmm, getting there. __div gone, 3 errors to go Jun 21 15:38:44 Laibsch: throw bad machines out of window Jun 21 15:39:46 mickey|fic, would you have a sample nand board file? Jun 21 15:40:07 dcordes: dunno, don't have a zaurus here Jun 21 15:44:44 Crofton: try looking in svn.openmoko.org/src/target/ Jun 21 15:44:52 there are all our uboot patches Jun 21 15:45:07 ok Jun 21 15:45:10 thanks Jun 21 15:45:59 we should switch to uclibc for everything Jun 21 15:46:11 that compiles in like 10 seconds instead of 10 hours :) Jun 21 15:46:14 how good is c++ support? Jun 21 15:46:34 There is a uclibc++ project somewhere Jun 21 15:47:09 Crofton: c++ as in stl? Jun 21 15:47:39 zecke: yes iirc Jun 21 15:48:22 chouimat: I know that uclibc++ is a stl implementation... but the question is if he needs that or pure C++ Jun 21 15:48:36 zecke: sorry ;) Jun 21 15:50:08 yes, stl Jun 21 15:51:22 * mickey|fic more uboot hackyness Jun 21 15:52:32 * chouimat|busy is away: cleaning the mess before the guests arrive this weekend Jun 21 16:03:44 mickey|fic: test it in RAM first before flashing it Jun 21 16:03:52 (unless you have a JTAG fetish) Jun 21 16:03:59 :D Jun 21 16:04:00 you bet Jun 21 16:04:24 since harald refuses to look into that, i'm going to give him my boards to unbrick :D Jun 21 16:05:10 ~seen codyl Jun 21 16:05:15 codyl was last seen on IRC in channel #openmoko, 55d 10h 10m 48s ago, saying: 'your article is up there ;-)'. Jun 21 16:05:27 mickey|fic: what was the fix for the eabi version mismatch? Jun 21 16:05:41 I solved it a few months a go, but forgot to document it Jun 21 16:05:43 version mismatch is remove flags to do OABI Jun 21 16:06:05 hmmm Jun 21 16:06:13 ecore-native and tetex don't compile Jun 21 16:06:38 anyway, time for food Jun 21 16:17:05 Laibsch: How much success did you have yesterday? Jun 21 16:21:32 * Laibsch joins #linuxtogo Jun 21 16:21:44 florian: Quite a bit I'd say Jun 21 16:21:49 read Changelog Jun 21 16:22:12 koen, what about generating glibc locales in parallel? Jun 21 16:30:21 * florian shoots monotone Jun 21 16:31:44 Having a file in the local tree that gets added by an update messes up the whole tree :-/ Jun 21 16:32:01 florian, nothing surprising Jun 21 16:32:21 I think svn would be much better for oe purposes, but whatever ... Jun 21 16:33:24 Marex: yes, sad but true... everything a distributed scm can do for oe can be handled with multiple package locations in oe directly. Jun 21 16:35:26 no it can't Jun 21 16:35:43 try commiting or diffing over multiple revision without an internet connection Jun 21 16:36:27 koen, why not? Jun 21 16:36:35 svn -r1234 diff ? Jun 21 16:36:52 that needs to connect to the svn server Jun 21 16:37:21 * koen points to the "without an internet connection" bit Jun 21 16:38:51 koen: hum well... is that really important? most of us usually have internet connectivity almost everywhere anyway. Jun 21 16:39:04 yes, that is important Jun 21 16:39:15 * florian trades this feature for less breakage Jun 21 16:40:03 try doing development when you're on GPRS or UMTS Jun 21 16:40:11 suddenly 'diff' takes 20 minutes Jun 21 16:40:49 and svn can't do merges Jun 21 16:42:17 koen: i was able to live with hh.org cvs for years - always slow and no offline operations at all Jun 21 16:42:44 florian: wc operations with svn is dead slow, a non working grep suck :) Jun 21 16:43:33 * psokolovsky wouldn't trade distributed scm for non-distributed nowadays Jun 21 16:47:50 hi, all! Jun 21 16:48:03 * chouimat|busy gives svk to koen :) Jun 21 16:48:23 could anybody advice some cheap arm evolution board for amateur needs? Jun 21 16:48:37 chouimat|busy: that's doesn't help against the svn repo breaking at every commit Jun 21 16:48:55 slapin: www.olimex.com Jun 21 16:49:21 hello, for OZ, i thought that all packages available via the feeds would be in one of the oe branches checked out with monotone, but i see things are missing. Jun 21 16:49:22 question may sound stupid, but does uClinux use a special kernel or how does it work? can existing drivers (kernel modules) be used there too or is some porting effort involved? Jun 21 16:49:22 Is there a central source repository for packages not in the oe branches, I want the source and OE build files etc so I can do some dev on them. thx Jun 21 16:50:48 I know this question is probably somewhat offtopic in OE, so sorry if it does not fit ;) but you guys surely know the answer :> Jun 21 16:51:59 kalineon: what do you think is missing? Jun 21 16:52:51 um i noticed gaim-protocol-oscar is for one Jun 21 16:53:18 kalineon: grep a bitbake manual and read Jun 21 16:53:40 kalineon: one .bb file can emit multiple packages. This is why the package contains the name of the bbfile (it used to) Jun 21 16:54:23 slapin: the Colibri boards are nice too Jun 21 16:55:11 ahh okay thanks, last time i worked on famil i thought everything had it's on bb file maybe it didn't or things have changed, thanks Jun 21 16:55:22 Or Triceps... but these are quite expensive. Jun 21 17:07:12 mtn-0.32: bytes in | bytes out | certs in | revs in Jun 21 17:07:13 mtn-0.32: 591.4 k | 228.8 k | 310/354 | 77/87 Jun 21 17:07:45 * koen wonders what those revs are Jun 21 17:11:39 * koen wonders who pushed 80+ merges Jun 21 17:13:06 * koen bets on the automerger Jun 21 17:14:14 03Laibsch 07org.oe.dev * r2de119df... 10/ (3 files in 2 dirs): Jun 21 17:14:14 glibmm: update to 2.12.10 Jun 21 17:14:14 * verified successful compilation Jun 21 17:14:14 * verified that no declared dependencies on the old version 2.8.4 exist in OE Jun 21 17:14:26 03Laibsch 07org.oe.dev * raaafb999... 10/ (1 packages/cairo/cairomm_1.2.4.bb): cairomm: initial release with version 1.2.4. Partly closes 2544. Jun 21 17:14:40 03rwhitby 07org.oe.dev * r00921883... 10/ (3 files in 3 dirs): arm-kernel-shim: Fixed dsmg600 cmdline Jun 21 17:21:14 florian, I'd change mtn for svn too Jun 21 17:21:25 I dont like mtn or git too much :S Jun 21 17:21:55 and I think distributed scm is not good for projects like OE because it takes just too long to update everything Jun 21 17:22:05 heh Jun 21 17:22:11 that's a load of nonsense Jun 21 17:22:11 and if you change something in your local tree, mtn merge often breaks Jun 21 17:22:22 koen, ok, you are the king Jun 21 17:22:29 koen, see eg. poky ... they use svn Jun 21 17:22:31 :) Jun 21 17:22:50 koen: ? Jun 21 17:23:15 * Marex openly laughs at zecke's ignorance ... Jun 21 17:23:35 zecke: it's nonsense that an update that talks to a server is faster than an update that is local Jun 21 17:24:10 * koen spends most of the time at work behind gprs Jun 21 17:24:17 koen, why does it take that long to update my org.oe.dev tree then? Jun 21 17:24:29 koen: eeks Jun 21 17:34:38 rwhitby: super merge storm :-) Jun 21 17:46:53 03koen 07org.oe.dev * rc3bf7d42... 10/ (1 packages/linux/linux.inc): linux.inc: s/arm/${ARCH}/g Jun 21 17:47:02 03oe 07org.oe.dev * r714f8404... 10/ (8 files in 7 dirs): Jun 21 17:47:02 merge of '7092ab793b40d9bf0352adda7cb0de9e9d1b53d9' Jun 21 17:47:02 and 'ed60cd7b6b6312ef6a4497d871d3c2f91ca6b369' Jun 21 17:48:43 03oe 07org.oe.dev * r6125010e... 10/ (8 files in 7 dirs): Jun 21 17:48:43 merge of '714f8404b02451929d6ca37a4f8a38beffae0322' Jun 21 17:48:43 and 'c3bf7d428ff163b2d84f615a9787148c42aefa89' Jun 21 17:50:59 * chouimat|busy is back. Jun 21 17:54:01 koen: let us start here, I will send a summary to the ml Jun 21 17:54:15 start what? Jun 21 17:54:29 koen: well, you should not underestimate the time required to run an e.V. Jun 21 17:54:50 bbl Jun 21 17:55:03 you should not underestimate the time when you want to leave FFIS or when FFIS goes down the drain Jun 21 17:56:39 koen: well, you can prevent that by getting a member and it requires less time Jun 21 17:57:00 koen: they are fine with the tax authorities, and if they don't blow up until 2010 Jun 21 18:00:23 zecke: can you give some hard numbers instead of handwaving? Jun 21 18:02:36 12 hours a week for ~10 members, +100 hours to create one, running to the tax authorities, notary, etc. Jun 21 18:03:16 what do those 12 hours a week consist of? Jun 21 18:04:28 stupid devs wanting money, book keeping, reading up on laws of the e.V. ;) Jun 21 18:04:36 (the last one is the idle task) Jun 21 18:06:07 what are all these merge messages that showed up in the commit email ? Jun 21 18:06:08 koen: for every money you take and spend you must check it is in compliance with the law (this takes some time to get used to) Jun 21 18:07:15 steliosk: probably a typo in rwithby's makefile Jun 21 18:09:51 koen: specially the guy creating the receipts has to take special care :} Jun 21 18:09:58 or typing 'mtn merge' 80 times by accident ;) Jun 21 19:04:52 what was the switch to build uclibc angstrom? Jun 21 19:05:29 thesing: http://www.openembedded.org/wiki/OeFaq Jun 21 19:05:43 thesing: ANGSTROM_MODE = "uclibc" Jun 21 19:05:55 thanks. Jun 21 19:06:10 thesing: make sure you have a machine config for uclibc -- otherwise it tries to build an Alpha version :-) Jun 21 19:06:49 cbrake: this faq-entry should be in angstrom-faq. Jun 21 19:08:57 cbrake: I added a 'generic' arm config to combat that problem Jun 21 19:12:15 * koen should have done that earlier Jun 21 19:14:09 thesing: I think it is probably useful information for both faqs as most OE users currently use Angstrom. At any rate, you can then look at the Angstrom config files to figure out how to do it for other distros. Jun 21 19:21:10 koen: AFAICT, all merges started to go unfiltered to channel and ML. anything changed just recently? Jun 21 19:24:34 yes, the ciabot now reports everything Jun 21 19:25:13 03koen 07org.oe.dev * rb1be59f7... 10/ (3 files in 3 dirs): Jun 21 19:25:13 uclibc: add generic armeb config Jun 21 19:25:13 * where 'generic' read 'armv4teb' Jun 21 19:27:18 anyone want to talk about kernel bb file naming? Jun 21 19:28:25 basically, the linux-omap1_2.6.18-omap1.bb could be called linux-omap1_2.6.18.bb Jun 21 19:28:43 why do we put omap1 in the package name and the version number? Jun 21 19:29:13 koen: can that be fixed, or is it on cia side? Jun 21 19:30:23 Crofton: because ipk name will be just kernel- Jun 21 19:30:52 so the ipkg name in not linux-omap1 then? Jun 21 19:31:13 this is due to the virtual prviders? Jun 21 19:31:24 Crofton: nope, there're standard naming for kernel/kernel module ipk's. Jun 21 19:31:50 ok Jun 21 19:32:06 Crofton: think users who want to instal kernel-module-foo without knowing what recipe itw as built from or that there're things like recipes at all ;-) Jun 21 19:32:32 So my next kernel is linux-davinci_2.6.x+git-davinci.bb Jun 21 19:34:39 well, wait! apparently my brains boil and I utter nonsense %) Jun 21 19:35:09 * Crofton waits Jun 21 19:35:57 it's not that necessary to include additional identifiers into version... Jun 21 19:36:01 psokolovsky__: my aim was to keep people from doing a zillion spurious merges ;) Jun 21 19:36:32 Crofton: well, forget what I said, I can't find ins and outs in that myself now ;-), time to break Jun 21 19:36:41 :) Jun 21 19:36:49 something looks funny to me Jun 21 19:36:57 koen: well, but it's not people who do merges, it's mtn ;-) Jun 21 19:37:01 * koen hints at making a generic linux_git.bb and using machine overrides on SRC_URI Jun 21 19:37:03 * mwester regularly needs a moment for brain to run a garbage-collect Jun 21 19:38:02 mwester: that's called 'sleep' :) Jun 21 19:39:47 hehe - naw, I'm old -- that's a nap. Sleep is when we run a full-out fsck on the brain to see what's been damaged or lost! Jun 21 19:40:32 * Crofton notes the hint Jun 21 19:40:34 fsck -f -y /dev/brain Jun 21 19:45:04 koen, I'm not sure if a generic git target will work well Jun 21 19:45:27 the omap1 git kernel uses the linux-omap1.inc file Jun 21 19:45:54 I wonder if we can require_machine-name ? Jun 21 19:46:24 woglinde: hey Jun 21 19:47:53 he zecke Jun 21 19:48:26 ~lastseen djay-il Jun 21 19:48:31 ~seen djay-il Jun 21 19:48:56 djay-il is currently on #oe. Has said a total of 22 messages. Is idling for 5h 41m 17s, last said: 'bye'. Jun 21 19:49:38 Crofton: sometimes it's not possible to cram it into a generic kernel, but sofar linux_2.6.x.bb supports arm,avr32,bfin and x86 :) Jun 21 19:50:21 does that exist? Jun 21 20:08:13 I'll compromise, I'll work of the linux-omap1.inc file for now, and when it works I'll see if it can be more general :) Jun 21 20:09:02 basically, I can't see finishing u-boot before I leave for vacation, so I am trying to make some progress on other tasks Jun 21 20:11:01 * * OE Bug 2545 has been created by cliff.brake(AT)gmail.com Jun 21 20:11:03 * * uclibc build uses host curses. h file instead of file in staging Jun 21 20:11:05 * * http://bugs.openembedded.org/show_bug.cgi?id=2545 Jun 21 20:15:24 * hrw for a moment Jun 21 20:15:46 Crofton: linux-omap_2.6.18-omap1 is because this kernel version is 2.6.18-omap1 Jun 21 20:16:10 hrw, I am just confused why the double mention of omap Jun 21 20:16:13 Crofton: linux_2.6.21-mm2.bb is because it is recipe for 2.6.21-mm2. Jun 21 20:16:29 Crofton: then make it linux_2.6.18-omap1.bb Jun 21 20:17:00 ok, so the preference is to have the architecture in the version string? Jun 21 20:18:02 Crofton: no. Jun 21 20:18:32 Crofton: kernel recipe can have any name. we used linux-MACHINE_version.bb usually Jun 21 20:18:57 hi hrw ... would you mind helping me a bit? Jun 21 20:19:00 ok Jun 21 20:19:02 Crofton: there are some exceptions like linux-openzaurus^Wrp which handle few different machines Jun 21 20:19:06 Marex: but fast Jun 21 20:19:14 I'll worry about the name when I am done with vaction :) Jun 21 20:19:19 thanks hrw Jun 21 20:19:34 Crofton: take a look at linux_2.6.21.bb - it is perfect example how kernel recipe can look Jun 21 20:19:43 ok Jun 21 20:19:47 hrw, is there any way how to generate glibc locales in parallel under OE? Jun 21 20:19:57 eg. running more qemu instances Jun 21 20:20:00 no idea Jun 21 20:20:19 have you tested glibc-2.6 yet? Jun 21 20:20:30 no Jun 21 20:20:39 oh ... ok :( Jun 21 20:20:55 Marex: I build poky usually not OE distros Jun 21 20:21:32 koen, have you tried building all the locales with glibc-2.6? Jun 21 20:21:56 * hrw goes Jun 21 20:43:07 what is the preferred way to handle situations when configure.in tries to run some C code during the configure process? Jun 21 20:43:55 cbrake: fix configure.in to have a cross compile default for AC_TRY_RUN Jun 21 20:46:43 cbrake: add the variable to the site files and extract the configure test and place it in contrib Jun 21 20:48:59 thanks! Jun 21 20:49:28 damn ... is there any way how to generate glibc locales in OE in parallel? Jun 21 20:58:30 feel free to ask more autofoo questions :} Jun 21 20:59:01 * cbrake is fairly impressed how easy mono builds in OE -- I guess its a good example of using autotools right :-) Jun 21 20:59:41 zecke: working on the site stuff now :-) Jun 21 20:59:52 that is black magic :} Jun 21 21:00:33 my aim is to have each test that we have a result for in our test set Jun 21 21:00:52 and for each new ABI, architecture we just need to run the test set once and have all the values we need Jun 21 21:00:55 *dream* Jun 21 21:02:40 zockt hier jemand alientrap? Jun 21 21:02:59 not me Jun 21 21:03:07 please speak english Jun 21 21:03:23 sorry. Wrong channel Jun 21 21:09:50 zecke: sounds good! Jun 21 21:22:58 cbrake: did you look at the mono-1.2.4 stuff that is already in oe? Jun 21 21:23:43 thesing: I'm currently fixing it Jun 21 21:24:00 thesing: there is a minor cross compile issue Jun 21 21:24:52 cbrake: which arch? Jun 21 21:25:46 cbrake: I made the current mono-bbs. Jun 21 21:27:52 thesing: cool! xscale eabi Jun 21 21:29:45 cbrake: look at site/mipsel-linux Jun 21 21:29:57 thesing: will do Jun 21 21:30:02 thesing: how is mono working out for you? Jun 21 21:31:11 works fine. but you need to copy the dlls manually. (compile the same mono version on host and copy them to your device.) Jun 21 21:32:08 thesing: so mono-dll does not provide them? Jun 21 21:32:41 it provide only the most basic dlls. (and not at current version afaik) Jun 21 21:33:27 thesing: ohh Jun 21 21:33:42 one could take them from the mono-native package, but afaik one can't create packages from native bbs easily Jun 21 21:35:04 but atm mono-native doesn't build the dlls. (if you enable it and it doesn't compile thy disabling parallel make) Jun 21 21:35:28 thesing: IC Jun 21 21:35:55 * thesing fought the whole day getting mono to compile with openwrt buildsystem. Jun 21 21:37:53 thesing: was doing it in openwrt a log more difficult than OE? Jun 21 21:37:58 *lot Jun 21 21:39:07 cbrake: a *lot*. this buildsystem doesn't do anything by itself. I ran into massive libtool and pkgconfig issues. Jun 21 21:39:42 cbrake: but I learned a lot about autotools ;) Jun 21 21:40:37 thesing: :-) so openwrt is based on buildroot Jun 21 21:41:21 cbrake: yes I think so. buildroot-ng or how the called it. Jun 21 21:41:57 thesing: could mono-native actually be mono-cross? gcc-cross packages stuff it build (libgcc.so) Jun 21 21:41:58 cbrake: but it has a nice gui to build images. and the images are much smaller than oe-build ones Jun 21 21:42:14 thesing: yeah, I look at other build systems and I strugle to find anything they offer over OE -- at least for the type of work I do. Jun 21 21:42:24 thesing: and OE uclibc support is coming along Jun 21 21:43:52 XorA: currently mono-native is used to stage monoburg which is used in mono. I don't know exactly what a foo-cross package does. Jun 21 21:47:20 thesing: thanks for the help! Jun 21 21:47:58 cbrake: have a look at http://www2.informatik.hu-berlin.de/~tkunze/mono if you want to rework the mono-bb Jun 21 21:49:04 you can drop the dependency to mono-native if you do something like the configure patch. Jun 21 21:49:04 thesing: OK, will do Jun 21 21:49:10 thesing: ok Jun 21 21:49:30 cbrake_away: I mean the part about BUILD_GLIB_* Jun 21 21:51:30 XorA: btw. is the multi-codec-support for asoc ready? Jun 21 21:52:56 thesing: its currently being reviewed I think Jun 21 21:53:12 thesing: its part a quite a large structural change Jun 21 21:56:09 XorA: is the a dedicated mailinglist for asoc or is this done on alsa-ml? Jun 21 21:56:55 thesing: should be on alsa-devel Jun 21 21:57:11 ok. thanks. Jun 21 21:57:20 thesing: opensource.wolfsonmicro.com/cgi-bin/gitweb/gitweb.cgi Jun 21 22:02:33 hi+ Jun 21 22:02:45 thesing: you probably want to pull the ppc-dev branch from asoc git as it has dual codec support Jun 21 22:15:50 koen, XorA: the merge was just my second development machine on which I build almost every day (and therefore when I update it sometimes does trivial merges), but almost never commit anything on. Yesterday I had to commit something there cause that's where the hardware under test was connected. Jun 21 22:16:09 so no mistakes, just normal monotone operation. Jun 21 22:18:05 rwhitby: I figured that, it just amused me Jun 21 22:18:47 rwhitby: it happened at exactly the same time I tested new mail client on my phone and I was getting new message beeeps every second for ages Jun 21 22:20:09 well, it worried me when I saw it happen (i.e. that many revs going out), but then I realised that it was the first push from that machine for a long time. Jun 21 22:20:55 rwhitby: could be worse, you could dissalow the first ever commit and push that :-D Jun 21 22:21:35 that would be classed as a D.o.S Jun 21 22:22:13 hmm, makes me think the servers should have a disallow window Jun 21 22:22:26 anything older than that cant happen Jun 22 00:21:17 is there a 'bitbake clean' type command, to clean up a failed build? Jun 22 00:22:26 summatusmentis: close dude, damn close Jun 22 00:22:37 bitbake -c clean foo Jun 22 00:22:54 oh cool, thanks Jun 22 00:24:01 * * OE Bug 2544 has been RESOLVED (FIXED) by Jun 22 00:24:03 * *  Updating gtkmm, glibmm, cairomm version of libraries Jun 22 00:24:05 * * http://bugs.openembedded.org/show_bug.cgi?id=2544 Jun 22 00:24:08 CoreDump|home: should there be output? Jun 22 00:25:46 well, yes Jun 22 00:26:05 my machines just slow, nvm Jun 22 00:26:53 CoreDump|home: oh... I forgot I had emerge world running :-P Jun 22 00:27:21 lol Jun 22 01:35:18 What's the package the generates updater.sh, and that stuff Jun 22 01:36:49 that would be zaurus-updater Jun 22 01:37:18 Thanks Jun 22 02:30:51 Crofton: ping **** ENDING LOGGING AT Fri Jun 22 02:59:57 2007