**** BEGIN LOGGING AT Wed Feb 21 02:59:57 2007 Feb 21 03:59:47 <_TalonAWD> hey im looking for a coder to pull something off a eprom on a ECU willing to $$ Feb 21 05:52:33 hi * Feb 21 05:53:48 i want to setup new target (at91rm9200), where can i start w/ OE/BitBake now ? Feb 21 05:55:15 have a look at /conf/machines and find one that matches your cpu Feb 21 05:56:44 Ifaistos: i dont' have conf/machines directory, where can i get it ? Feb 21 05:57:08 Ifaistos: i checkedout source openembedded monotone Feb 21 06:01:26 its in the org.openembedded.dev dir Feb 21 06:02:38 Ifaistos: oh, i see Feb 21 06:03:46 Ifaistos: after found one matched w/ my CPU, copy it to build/mydistro , then build it ? Is that correct ? Feb 21 06:04:36 nop. you don't need to copy anything Feb 21 06:04:56 hold on a sec i'll try to find a page that explains the details a bit Feb 21 06:05:20 Ifaistos: thx guy, Feb 21 06:10:56 nm : its outdated in some parts but will give you a starting point -> http://www.openembedded.info/wiki/Main_Page Feb 21 06:11:18 nm : for a bit more advance options -> http://bec-systems.com/web/content/blogsection/7/9/ Feb 21 06:11:56 nm : An example how to set up the build enviroment -> http://www.hrw.one.pl/2006/11/22/my-openembedded-environment-ii/ Feb 21 06:13:18 Ifaistos: okie, i'll try it now Feb 21 06:27:37 good morning Feb 21 06:32:36 hi gremlin[it] Feb 21 06:35:27 hi Ifaistos Feb 21 07:21:35 good morning all Feb 21 07:26:59 morning b8jlene Feb 21 07:27:04 morning v8jlene Feb 21 07:27:24 likewise: Morning Feb 21 07:59:33 hi Feb 21 08:00:13 i have a question, i've been recompiling packages for the uclibc and i've find a problem Feb 21 08:00:38 packages like coreutils ask for a defined HAVE_MEMPCPY and if defined use __mempcpy Feb 21 08:00:58 uclibc exports mempcpy but not __mempcpy Feb 21 08:01:16 morning Feb 21 08:01:17 they hide __mempcpy Feb 21 08:01:24 morning hrw Feb 21 08:01:47 morning all Feb 21 08:01:51 hi Dirk Feb 21 08:02:01 hey Marcin Feb 21 08:02:05 so, what will be the correct action, hack uclibc? undefine HAVE_MEMPCPY? hack packages? or just do not include those packages? Feb 21 08:02:28 saladino: ade|desk is local uclibc mastah Feb 21 08:02:33 iirc Feb 21 08:03:46 humm Feb 21 08:06:07 saladino : which gcc ver and which uclibc ver ? Feb 21 08:06:27 uclibc svn Feb 21 08:06:35 saladino : it seems that gcc 4.1.x need uclibc-0.9.28-svn to work Feb 21 08:06:51 gcc 3.4.4 Feb 21 08:07:13 it is preferred in the ucslugc conf file Feb 21 08:07:24 i was thinking about changing it Feb 21 08:07:59 but first i wanted to achieve a compilation by default, i have managed to compile image and now i am with packages Feb 21 08:08:08 saladino : i am using 4.1.1 with uclibc-0.9.28-svn for x86,powerpc and works ok (minor glitches to the packages for things not defined in site files) Feb 21 08:09:00 saladino : check also the uclibc config as sometimes needs things to be added/excluded Feb 21 08:09:08 Ifaistos do you compile all packages? coreutils included? Feb 21 08:09:45 saladino : bootstrap-image builds Feb 21 08:10:08 yesterday i found that packages like libtcl used functions of libmath that werent included in uclibc although they were there, a patch for makefile.in solved that Feb 21 08:10:53 i don't know much about what do those builds include, would you tell me if libtcl and coreutils are compiled? Feb 21 08:11:13 saladino : don't remember will try and tell you Feb 21 08:11:35 ok Feb 21 08:11:40 thanks Feb 21 08:11:41 saladino : libtcl is need it by what package ? Feb 21 08:12:15 i don't know, it justs compiles when doing ucslugc-packages, i'm trying to compile all packages to keep ucslugc alive Feb 21 08:12:21 it was going to be dropped Feb 21 08:13:11 did you manage to build bootstrap-image ? Feb 21 08:13:30 bootstrap-image generates a .bin image to upload? Feb 21 08:14:05 depends on what you tell it in local.conf... Feb 21 08:14:19 but in general yes Feb 21 08:14:39 as far as i know, ucslugc does not provide a .bin image yet, it provides a kernel and a rootfs file to upload, i managed to do that Feb 21 08:14:40 it gives you a minimal image to boot you syste, Feb 21 08:15:06 ok. did you noticed any headers files been shipped in the rootfs ? Feb 21 08:15:22 no, i can't try it yet Feb 21 08:16:01 do a find rootfs/ -name *.h and see what comes out Feb 21 08:16:01 i have some kernel misconfiguration and my ucslugc is not booting yet Feb 21 08:16:06 im waiting for my serial port Feb 21 08:16:10 hmm Feb 21 08:16:12 from here Feb 21 08:16:13 wait Feb 21 08:16:35 nope Feb 21 08:16:52 saladino@korn:/mnt/almacen/Arkaitz/src/ucslugc/tmp/rootfs$ ls Feb 21 08:16:52 bin boot dev etc home initrd lib linuxrc media mnt proc root sbin sys tmp usr var Feb 21 08:16:52 saladino@korn:/mnt/almacen/Arkaitz/src/ucslugc/tmp/rootfs$ find -name "*.h" Feb 21 08:16:52 saladino@korn:/mnt/almacen/Arkaitz/src/ucslugc/tmp/rootfs$ Feb 21 08:17:20 you need to add ./ before the name Feb 21 08:17:34 find ./ -name *.h Feb 21 08:17:43 it is the same Feb 21 08:18:21 as far as i know the find command uses the actual path if no path provided Feb 21 08:18:30 but still i've tried and no .h found Feb 21 08:18:33 uclibc is 0.9.28 svn ? Feb 21 08:18:45 yep Feb 21 08:18:53 is gettext installed ? Feb 21 08:18:59 in the rootds Feb 21 08:19:01 uclibc-0.9.28+svn20070221-r4 Feb 21 08:19:09 it should Feb 21 08:19:25 wait Feb 21 08:20:46 Ifaistos i think not Feb 21 08:21:01 i cannot find the gettext executable Feb 21 08:21:07 but it is compiled Feb 21 08:21:17 i saw it compiling lots of times Feb 21 08:21:18 :) Feb 21 08:21:24 it should be there? Feb 21 08:21:34 gettext-native or gettext ? Feb 21 08:22:20 ahumm Feb 21 08:22:23 maybe native Feb 21 08:22:53 native Feb 21 08:22:54 yes Feb 21 08:23:14 i gess that it may compile after coreutils in ucslugc-packages Feb 21 08:23:58 once you get it compiles and in the rootfs check for header files again Feb 21 08:24:24 ok Ifaistos Feb 21 08:24:29 i take note Feb 21 08:24:58 did you check for the coreutils compilation? Feb 21 08:24:59 i am getting a strange dependansy from gettext. gettext depends on gettext-dev which depends on uclibc-dev and as a result i get all kind of header files in the rootfs Feb 21 08:25:14 yep, you told me yesterday, sounds strange Feb 21 08:25:16 not yet all machines are doing build Feb 21 08:25:57 will check and let you know Feb 21 08:26:12 ok Feb 21 08:26:20 morning all Feb 21 08:26:49 hi RP Feb 21 08:28:05 hi RP Feb 21 08:30:26 does somebody use distcc for compiling? is it possible? Feb 21 08:31:21 i do :) Feb 21 08:32:10 something special to do it? Feb 21 08:32:11 distcc or icecc? Feb 21 08:32:18 icecc sorry Feb 21 08:32:21 :-) Feb 21 08:32:26 are they similar? Feb 21 08:32:38 icecc supports cross-compiler a tad better Feb 21 08:32:40 distcc fro cross compiling is a no-no Feb 21 08:32:54 i'll try icecc Feb 21 08:33:05 do you use any special conf? Feb 21 08:33:15 or is just install and run? Feb 21 08:33:19 there is a wiki page on icecc Feb 21 08:33:29 oki Feb 21 08:33:35 hi RP Feb 21 08:34:23 saladino : http://www.openembedded.org/wiki/IceCC Feb 21 08:37:04 softfp in anything else thatn arm seems to be not working Feb 21 08:38:23 do13_: hi. for info, kexec is now in mainline :) Feb 21 08:39:35 RP: cool Feb 21 08:51:01 lrg: good morning! Feb 21 08:51:20 hi Liam, ade Feb 21 08:51:49 hi hrw et al. Feb 21 08:51:59 hey ade|desk Feb 21 08:52:25 morning Koen, Marcin, all Feb 21 08:56:14 hi ade|desk Feb 21 08:56:47 i've a question about uclibc Feb 21 08:57:15 i have a question, i've been recompiling packages for the uclibc and i've find a problem Feb 21 08:57:15 packages like coreutils ask for a defined HAVE_MEMPCPY and if defined use __mempcpy Feb 21 08:57:15 -:- SignOff bluelightning: #oe (Read error: 110 (Connection timed out)) Feb 21 08:57:15 uclibc exports mempcpy but not __mempcpy Feb 21 08:57:26 so, what will be the correct action, hack uclibc? undefine HAVE_MEMPCPY? hack packages? or just do not include those packages? Feb 21 09:01:59 03rwhitby 07org.oe.dev * r9148d2b8... 10/ (3 files in 2 dirs): ixp4xx-kernel: Update ARM kexec patch for new official syscall number. Feb 21 09:02:03 03mwester 07org.oe.dev * r6b385aa9... 10/ (4 files in 4 dirs): slugos-init: turnup disk -s fixed, default delay set to 10 secs Feb 21 09:02:06 03rwhitby 07org.oe.dev * r1b0fc9ca... 10/ (4 files in 3 dirs): ftpd-topfield: Updated to 0.7.0, and fixed usb.h header location Feb 21 09:10:06 seems i cant find a pmap bb .. is there a reason or do i just need to create it ?= Feb 21 09:15:47 jonbaine uhhi Feb 21 09:15:59 saladino, uuuhhhiiii ^^ Feb 21 09:21:37 http://www.youtube.com/watch?v=SDjvGu19XJQ&eurl= for Treo owners Feb 21 09:29:08 RP: did you see http://bugs.openembedded.org/show_bug.cgi?id=1892 already? Feb 21 09:29:33 hrw: I hit the same packaging bug, but in gcc, not glibc Feb 21 09:32:05 koen: good to know Feb 21 09:42:43 hi greentux, lrg Feb 21 09:43:19 03hrw 07org.oe.dev * r429208b5... 10/ (1 packages/libxslt/libxslt_1.1.17.bb): libxslt: use destdir.patch to fix do_install Feb 21 09:44:42 koen: I started with aløl new fresh OE build envirement today, and still udev failes. Feb 21 09:45:10 hey Dirk Feb 21 09:45:45 here is the pastebin: http://pastebin.ca/365895 Feb 21 09:45:58 good morning Feb 21 09:58:41 hi florian Feb 21 10:03:14 hrw: Do you have any idea why udev failes? Feb 21 10:03:41 goxboxlive: ERROR: QA issue: libvolume_id.pc failed sanity test (reference to /usr/include or workdir) Feb 21 10:03:52 libvolume_id.pc is fscked Feb 21 10:04:55 yes, bu Feb 21 10:05:11 t are there anything that is wrong in my setup or is it a bug? Feb 21 10:05:42 I saw a guy at bugtracker he removed the libvloume.id in the udev.bb file. Feb 21 10:05:48 is that the way to do it? Feb 21 10:10:38 I suggest look deeper into problem and fix it properly Feb 21 10:10:39 http://bugs.openembedded.org/show_bug.cgi?id=1884 Feb 21 10:10:51 I lack time to fix it Feb 21 10:11:17 ok, thanks anyway. Feb 21 10:22:08 hey pb_ Feb 21 10:30:58 saladino: take a look at how it is done in buildroot, it maybe that there is a config option in uclibc to enable this Feb 21 10:35:12 from uclibc ml : mike :- otherwise i'll be tagging Feb 21 10:35:12 0.9.29 when i get back from FOSDEM next week Feb 21 10:35:19 yay :) Feb 21 10:36:23 then i think that {arm,mips,x86} nptl will be in after that Feb 21 10:41:26 yay! Feb 21 10:46:26 ade|desk the fact is that uclibc guys decided not to export the __symbols, i've readed through google Feb 21 10:46:42 so i don't know what can be the best solution Feb 21 10:50:26 well coreutils builds in buildroot, so take a look there Feb 21 10:51:49 ok ade|desk i'll look Feb 21 10:51:50 thanks Feb 21 10:53:41 ac_cv_func_memcmp_working=yes would that be it ? Feb 21 10:54:13 don't know, still searching. :) Feb 21 11:03:33 ade|desk nop, i think that ac_cv_func_memcmp_working will not work Feb 21 11:03:53 shame Feb 21 11:04:21 i supose that a patch for coreutils that #define __mempcpy mempcpy will work, but for sure that coreutils won't be the only package with that error Feb 21 11:04:38 so, a patch for every package... i don't know Feb 21 11:06:27 wow Feb 21 11:06:34 sed -i 's:__mempcpy:mempcpy:g' lib/*.c Feb 21 11:06:34 xD Feb 21 11:11:25 can i use OVERRIDES or something similar to use a do_unpack only when a specific distro is compiling? i can't find such thing in the bitbake manual Feb 21 11:13:23 you can, but 99% of the time a distro override is bogus and needs another solution Feb 21 11:14:02 ok koen, now i have 2 things to override when using the uclibc Feb 21 11:14:10 now im doing overriding distro when ucslugc Feb 21 11:14:18 is there a better way? Feb 21 11:14:58 what do you want to accomplish? Feb 21 11:15:42 in the first override i want to retire a package from task-slugos.bb because it can be compiled with uclibc Feb 21 11:15:44 it is done Feb 21 11:16:19 in the other override i need to substitute all __mempcpy from the code because uclibc does not export __mempcpy Feb 21 11:18:57 ~seen gerwin Feb 21 11:19:11 ~seen gerwinin Feb 21 11:19:27 gerwin was last seen on IRC in channel #oe, 413d 22h 1m 44s ago, saying: 'zecke: no I think it was some called sipphone or something it was with video'. Feb 21 11:19:30 gerwinin was last seen on IRC in channel #oe, 20h 47m 38s ago, saying: 'Koen: is the pressrelease good enough to send out ?'. Feb 21 11:19:46 hmm, sent him a booth schedule ages ago and got no answer Feb 21 11:20:45 XorA: edit wiki page directly? Feb 21 11:21:00 anyway koen, what is the correct syntax to do a do_unpack override? Feb 21 11:21:25 hrw: I was trying to follow instructions in said wiki :-) Feb 21 11:24:25 hi florian Feb 21 11:24:59 03florian 07org.oe.dev * rbeda0ffc... 10/ (3 files in 2 dirs): task-gpephone: Do not depend on gpe-bluetooth. Feb 21 11:25:15 ah, important topic... Feb 21 11:25:26 * florian starts to work on gpe-bluetooth Feb 21 11:54:16 Looks like the new quality (great idea!) just rooted out an issue with udev: just got the following error when compiling udev: QA issue: libvolume_id.pc failed sanity test (reference to /usr/include or workdir) Feb 21 11:54:52 s/quality/quality tests Feb 21 12:01:47 Gerrath: I suspect it can be fixed with do_stage_append() { sed -i -e s::$STAGING_SOMETHING:g } Feb 21 12:02:38 koen, ok, I was just starting to hunt it down, I was going to look at the croco patch since that was the example given on the wiki. Feb 21 12:02:59 Gerrath: ah, cool Feb 21 12:03:54 koen, what is the correct syntax to do a do_unpack override? Feb 21 12:06:54 pb_: Do you have a free minute for taking a look at a license problem in gpe? Feb 21 12:10:46 re Feb 21 12:12:53 koen: I just looked at the log and the install process is putting that file here: itmp/work/armv5te-angstrom-linux-gnueabi/udev-092-r15/image/usr/lib/pkgconfig/libvolume_id.pc Feb 21 12:13:09 koen, isn't that were it should go? Feb 21 12:13:17 What causes libtool to try to link against ${libdir}/.libs? Feb 21 12:17:18 Gerrath: iirc it complains about the .pc in QA staging, which means the .pc in image/usr/ isn't what's tripping the error Feb 21 12:17:24 (I could be wrong, though) Feb 21 12:29:50 03hrw 07org.oe.dev * r3ab7a13c... 10/ (1 packages/xorg-lib/libxfont-native_X11R7.1-1.2.0.bb): libxfont-native: add 1.2.0 needed for recent freetype update Feb 21 12:32:09 koen, I think the staging looks good too: tmp/work/armv5te-angstrom-linux-gnueabi/udev-092-r15/temp-staging/usr/lib/pkgconfig/libvolume_id.pc Feb 21 12:34:02 koen, I guess I need to look at insane.bbclass and try to figure out what its looking at. Feb 21 12:42:11 I see something strange Feb 21 12:42:14 koen, well I'm not sure I understand what is going on, it looks like to me that insane reads the .pc files and sees if they have /usr/include in them and if they do then it sets a flag. I looked at the pc files in my last oe build and it looks like they all have /usr/include. Feb 21 12:42:41 xorg-font/xorg-font-common.inc got: Feb 21 12:42:42 DEPENDS = " encodings font-alias font-util-native mkfontdir-native mkfontscale-native" Feb 21 12:42:46 RDEPENDS = "encodings font-util font-alias mkfontscale mkfontdir" Feb 21 12:43:00 but 'bitbake font-misc-misc' does not build mkfontscale mkfontdir ;( Feb 21 12:43:56 bitbake -e -b font-misc-misc.bb|grep RDEPENDS show proper value Feb 21 12:44:10 hrw: related to #1892 ? Feb 21 12:44:31 koen: maybe Feb 21 12:56:59 koen, what name does Holger Freyther go by on the IRC? I would like to talk to him about the insane.bbclass Feb 21 12:57:27 Gerrath: zecke Feb 21 12:57:38 hrw, thanks. Feb 21 12:57:54 I have X starting again. Feb 21 13:03:35 is there any particular reason libtiff's bb file is call tiff ? Feb 21 13:03:45 and not libtiff ? Feb 21 13:03:50 Ifaistos: upstream naming? Feb 21 13:04:50 koen: libx11 lack locale support Feb 21 13:04:58 or I have something fscked Feb 21 13:05:04 hmmm upstream the call it libtiff -> http://www.remotesensing.org/libtiff/ Feb 21 13:05:11 fluxbox: symbol lookup error: fluxbox: undefined symbol: Xutf8TextListToTextProperty Feb 21 13:05:17 only the source tar is call tiff Feb 21 13:07:02 Warning: X server does not support locale Feb 21 13:07:02 Warning: cannot set locale modifiers Feb 21 13:07:04 hm Feb 21 13:07:10 hrw: that means fluxbux has been built against full x11 :( Feb 21 13:07:28 hrw: I think it makes sense to put the full x11 in the x86 feeds Feb 21 13:07:32 koen: as rest of things here Feb 21 13:07:49 koen: I did not build diet-x11 in this builddir Feb 21 13:12:32 hi all Feb 21 13:12:34 if I'd want to help testing the upcoming 3.5.5 on my collie, what should I do? Feb 21 13:12:39 so the following in a .pc file is wrong: includedir=/usr/include Feb 21 13:12:44 (sorry for xpost :P ) Feb 21 13:12:59 which means there will be a lot of .pc files failing in OE. Feb 21 13:14:18 lestat: wait until I find time to go back to this Feb 21 13:14:30 hrw: ok Feb 21 13:19:23 koen: diet-x11 is r4, full-x11 is r1 ;( Feb 21 13:19:37 bump PR, please Feb 21 13:20:08 hrw: does armstrong work on collie, too? Feb 21 13:20:44 lestat: collie/2.6 is broken Feb 21 13:21:16 armstrong uses only 2.6? Feb 21 13:21:36 anyway I have fluxbox started Feb 21 13:22:45 03hrw 07org.oe.dev * r34eba82d... 10/ (1 packages/xorg-lib/libx11_X11R7.1-1.0.1.bb): libx11: bump PR to the same as in diet-x11 Feb 21 13:23:16 mmm... if you should advice the best (actually) distribution for collie? I tried PdaXRom too, but I couldn't get it working :/ Feb 21 13:23:30 lestat: hentges Feb 21 13:23:45 pdaXrubbish? Feb 21 13:24:10 hentges isn't an OZ mod? Feb 21 13:24:33 is Feb 21 13:25:36 mmm... and what are the main differences? (If you're busy, feel free to tell me to keep quiet! :p ) Feb 21 13:27:03 lestat: updates, polishing etc. Feb 21 13:27:56 03pfalcon 07org.oe.dev * r5e73970d... 10/ (1 conf/distro/angstrom-2007.1-oabi.conf): Feb 21 13:27:56 angstrom-2007.1-oabi.conf: Bump PREFERRED_VERSION_linux-handhelds-2.6_h3600 to point Feb 21 13:27:56 to current release away of old removed. Feb 21 13:30:31 thanks Feb 21 13:34:17 koen: I'd not seen that bug, no Feb 21 13:36:57 brb, food Feb 21 13:37:42 RP: the same thing happened before when building stuff in bitbake shell, but it seems insane now triggers the same bug :) Feb 21 13:37:48 s/:)/:(/ Feb 21 13:41:45 cu Feb 21 13:50:51 koen: insane triggers the bug? Thats interesting. I should really try insane I guess... Feb 21 13:51:19 RP: it's the only change I made, inheriting the new insane Feb 21 13:51:59 koen: You should mention that on the bug... Feb 21 13:54:10 added Feb 21 13:56:58 ade|desk do you know why can be this error -> /mnt/almacen/Arkaitz/src/ucslugc/tmp/staging/armeb-linux-uclibc/include/sys/ucontext.h:92: error: field `uc_mcontext' has incomplete type Feb 21 13:57:30 almost other uclibc errors where solved changing the config file, but this... Feb 21 13:57:46 hm smells like chrackter wide stuff Feb 21 13:59:13 i look at ucontext.h and i can't find where is the problem Feb 21 14:07:36 are there any of the perl-cpan modules availiable in OE ? Feb 21 14:08:13 so instead of perl -MCPAN -e "install Net::Telnet" install a package ? Feb 21 14:10:20 03pfalcon 07org.oe.dev * r70b4e64a... 10/ (3 files in 3 dirs): Feb 21 14:10:20 aximx50.conf: New machine, Dell Axim X50/X51/X50v/X51v. Feb 21 14:10:20 * Also, add as compatible to linux-handhelds-2.6.inc. Feb 21 14:18:01 did anyone noticed users of amd lx 800 geode`s yet ? Feb 21 14:20:29 03koen 07org.oe.dev * rb1155dad... 10/ (1 packages/openmoko-pim/openmoko-dates_svn.bb): openmoko-dates: fix svn location Feb 21 14:21:27 <_law_> psokolovsky_, thanks for adding axim stuff to OE Feb 21 14:22:36 _law_, thank you! Feb 21 14:23:51 03law 07org.oe.dev * rcf10d341... 10/ (3 files in 3 dirs): Feb 21 14:23:51 linux-handhelds-2.6: Add defconfig for aximx50 Feb 21 14:23:51 * Closes #1888 Feb 21 14:25:00 <_law_> law ? i dont have a OE account... Feb 21 14:25:08 what controls the name of the .ipk package that bitbake generates? Feb 21 14:25:43 I have a simple bitbake recipe that installs one library into /usr/lib, but it's being named wrong Feb 21 14:25:49 _law_, OE has ability to specify actual author of changes commited. Sorry, I missed to use that 1st time (multitasking ;-)). Feb 21 14:26:05 it's getting the name of the library, not the name of the .bb file Feb 21 14:26:22 xumbi8: right, it's supposed to work like that Feb 21 14:26:22 xumbi8, there's such thing as "Debian package renaming" Feb 21 14:26:32 <_law_> psokolovsky_, so i have got a OE account now? Feb 21 14:26:52 ahhh, ok, I'll run some searches on that Feb 21 14:26:58 _law_, no, but commit marked as yours Feb 21 14:27:09 <_law_> psokolovsky_, ok have to go now bbl Feb 21 14:27:33 xumbi8: seriously, it is supposed to be named after the library instead of after the .bb Feb 21 14:29:27 koen: right, gotcha, it's just that my library ends in 1_33_1.so.1.33.1, which I don't think bitbake expects Feb 21 14:29:45 koen: so my ipk ends in 1-33-1-1.33.1_1.33.1-r1_armv5te.ipk Feb 21 14:34:08 re Feb 21 14:35:14 mornin Feb 21 14:35:15 g Feb 21 14:36:09 ok looks like if I remove the version from my .bb file name I get a good looking ipk at the end Feb 21 14:38:31 xumbi8: that's wrong Feb 21 14:38:46 outstanding Feb 21 14:39:24 xumbi8: you shouldn't care how "good looking" a package name is Feb 21 14:39:34 if you remove the version you can't upgrade it later Feb 21 14:39:43 definitely bad Feb 21 14:40:15 florian: okay. what's the problem? Feb 21 14:40:36 well I was having a hard time getting the package picked up by DISTRO_EXTRA_RDEPENDS, and I thought it was something to do with the versioning Feb 21 14:41:39 pb_: neal moved some code you wrote to import vcard and vcal files from the pim applications to libgpevtype. Feb 21 14:41:41 hi, all! Feb 21 14:41:50 florian: right Feb 21 14:41:59 pb_: now we gave another library that is mixed gpl and lgpl Feb 21 14:42:03 eh have Feb 21 14:42:08 ah, I see Feb 21 14:42:19 I guess relicensing all that stuff as lgpl is okay Feb 21 14:42:22 strange - binutils-cross doesn't build when built from scratch... Feb 21 14:42:55 pb_: ok, do ypu want to do this yourself os should i just change it? Feb 21 14:43:11 florian: please go ahead and just change it Feb 21 14:43:18 oh well.. i can't type anymore :-/ Feb 21 14:43:25 pb_: ok, many thanks! Feb 21 14:53:44 Jin^eLD: Well, I guess the compat-34 stuff was too good to be true :( Feb 21 14:54:30 hvontres|poodle: why that, it worked for me! Feb 21 14:56:10 how did it fail? Feb 21 14:57:17 Jin^eLD: well, qemu builds, but when I tried to rebuild glibc, it segfaults during locale generation. At least on my setup :( Feb 21 14:58:21 hmm.. well, of course I did not do a clean build... Feb 21 14:58:55 I first build without locale generation; then later enabled it and tried to build with compat-34 installed Feb 21 14:59:06 so it built quemu, and then generated the locale stuff without rebuilding the whole glibc Feb 21 14:59:09 and that did work Feb 21 14:59:25 Jin^eLD: what arch are you using? mine was arm, so maybe there is an arm specific issue here Feb 21 14:59:45 I built it for arm too Feb 21 15:00:17 Jin^eLD: hmmmm....very interesting....but also very stoopid.. Feb 21 15:00:23 indeed.. Feb 21 15:00:24 hmm Feb 21 15:01:15 Jin^eLD: would you mind trying a bitbake -c rebuild glibc to see if it works? (I know that is probalby the most painfull thing to do in OE...) Feb 21 15:01:24 ok, let me try Feb 21 15:01:43 Jin^eLD: BTW, what kind of build machine are you running on ? Feb 21 15:01:44 mtn: multiple update candidates: Feb 21 15:01:44 mtn: b1155dadb6ffc45920e64fbaa89240ba5c7fa0d4 koen@openembedded.org 2007-02-21T14:54:22 Feb 21 15:01:44 mtn: f6878aab9639c583f635cdd6149b4aee8f86cb50 pfalcon@openembedded.org 2007-02-21T14:52:11 Feb 21 15:01:48 any ideas? Feb 21 15:02:07 hvontres|poodle: dual xeon, 3GHz each, 2GB RAM Feb 21 15:02:30 slapin_nb, choose any, I'm merging now Feb 21 15:02:35 * hvontres|poodle bows down to supirior hardware.... Feb 21 15:02:47 well :) its my machine at work Feb 21 15:02:58 so I was just luckily they gave me a decent one Feb 21 15:03:02 Jin^eLD: heh, that makes my dual P-II as home look even slower.. Feb 21 15:03:34 Jin^eLD: but the price was right :) (being in good with the IT guys at work helps) Feb 21 15:03:52 heh.. :) well I do not even have a PC at home Feb 21 15:04:03 use my noteobok as a terminal to get to the Xeon at work :> Feb 21 15:04:38 Jin^eLD: well, I kinda run a retirement home for old PC's it seems. Feb 21 15:05:07 :) Feb 21 15:06:38 I think I threw away my K6/2 333 end of last year Feb 21 15:06:46 and only because the hw died after many years of nonstup running :> Feb 21 15:08:06 Jin^eLD: heheh...well, the dual P-II replaced my trusty old K6/2-400. Now I just need to talk myslef into a newer laptop... that old Pentium-mmx is not going to cut it forever :) Feb 21 15:11:55 hehe, I have one too :) barely using it though Feb 21 15:13:28 hvontres|poodle: NOTE: package glibc-2.5-r4: task do_install: started Feb 21 15:14:21 ok locale generation failed Feb 21 15:17:49 Jin^eLD: so the plot thickens. Feb 21 15:18:26 yes.. interesting that I could generate it in the "2nd round" without rebuilding Feb 21 15:18:57 Jin^eLD: hmm, I'll have to try that at home tonight Feb 21 15:19:24 please could somebody look at bug #1900, that's about build of binutils-cross Feb 21 15:20:21 hvontres|poodle: basically it is telling me that there was an error in execution, and prints the line and environment Feb 21 15:20:34 I'll try to launch it manually and see if it segfaults or what it does Feb 21 15:22:23 yep, segfault here as well Feb 21 15:27:59 waa Feb 21 15:28:00 | ERROR: "gcc" looks like gcc 4.x Feb 21 15:28:00 | QEMU is known to have problems when compiled with gcc 4.x Feb 21 15:28:22 robtaylor: spank the qemu dudes and do 'apt-get install gcc-3.4' Feb 21 15:28:43 (if that's qemu-native, for qemu, spank the qemu dudes again) Feb 21 15:28:59 koen: ok. another one to add to openembeddded-essential :) Feb 21 15:29:43 spanking the qemu dudes? Feb 21 15:29:45 ;) Feb 21 15:29:48 heh Feb 21 15:30:02 koen: what distro do you use, ooi? Feb 21 15:30:30 koen: speaking of qemu, have you tried rebuilding glic lately? Jin^eLD and I are getting segfaults during locale generation :( Feb 21 15:32:51 the stacktrace is 44 lines long :) but no debugging symbols, so can't see anything Feb 21 15:34:52 THe git fetcher doesn't really grok alternate heads, does it? Feb 21 15:36:13 how can i set to use a local mirror for debian packages need during build ? Feb 21 15:41:34 Jin^eLD: hmm, well, thanks for looking into this for me...I guess I'll go bake to turning off locale generation for now :) Feb 21 15:42:27 hi, i have a problem bitbaking glibc-2.3.5+cvs20050627-r3 (oz354x branch) : Feb 21 15:42:33 cannot open locale definition file `/usr/share/i18n/locales/sr_CS': No such file or directory Feb 21 15:43:15 robtaylor: I use debian on one buildhost and ubuntu on another Feb 21 15:43:30 hvontres|poodle: qemu Just Works(TM) over here Feb 21 15:43:35 looks like a cross_compile badness: it tries to use my native i18n stuff instead of the files from the package to compile Feb 21 15:44:57 koen: cool, i was thinking of setting up an apt repo, with openembedded-essential,laest monotone and bitbake build for feisty, etch and maybe edgy Feb 21 15:45:04 koen: what do you reckon? Feb 21 15:45:23 robtaylor: sounds usefull Feb 21 15:46:29 also updated openembedded-essential: Feb 21 15:46:29 * Add openembedded-essential.postinst for setting git alternatives and Feb 21 15:46:29 symlinking /bin/sh to /bin/bash if needed. Feb 21 15:46:29 * rules: Convert to using cdbs. Feb 21 15:46:29 * control: Add a whole load more dependancies Feb 21 15:47:05 robtaylor: ping hrw|gone and/or add it to bugzilla Feb 21 15:47:14 koen: will do :) Feb 21 15:47:23 thanks Feb 21 15:49:47 koen: Is glibc supposed to be doing this: I18NPATH="/usr/share/i18n" ? Feb 21 15:51:35 hvontres|poodle: no idea Feb 21 15:51:58 there's a bug in bugzilla about it iirc Feb 21 15:52:59 http://bugs.openembedded.org/show_bug.cgi?id=1821 Feb 21 15:59:47 DoKu: thanks Feb 21 16:00:02 ~fish hvontres|poodle for not searching bugreports Feb 21 16:00:14 * ibot slaps hvontres|poodle for not searching bugreports around with a large trout Feb 21 16:01:12 03mwester 07org.oe.dev * rdb725bcc... 10/ (5 files in 4 dirs): Feb 21 16:01:12 slugos: demoted netconsole driver to module in rootfs, removed dsmg600 cmdline Feb 21 16:01:12 patch (replaced by shim), added linux-2.6.20.1 patch, to bring in sync with svn Feb 21 16:13:19 mhm.. an mplayer related question.. I have a device that has an LCD display; the display can be accessed via framebuffer; I compiled mplayer with aalib support for the device - any ideas how to tell mplayer to render aalib stuff to fbdev? Feb 21 16:13:28 with aalib it will always try to render in current console Feb 21 16:14:31 one of my ideas was to somehow spawn a console with output to the LCD :) but I did not figure out how to do it yet Feb 21 16:14:39 so any ideas are welcome :) Feb 21 16:15:54 Doesn't mplayer have fbdev support? Feb 21 16:16:22 CosmicPenguin: it supports it twice even Feb 21 16:16:25 yes, but then it will try to render "normal" stuff to fbdev Feb 21 16:16:33 I want to tell it to render aalib to fbdev Feb 21 16:17:31 it's just a fun experiment anyway, but still would be interesting :) monochrome LCD, 128x60 or something Feb 21 16:22:51 the problem with the i18n issue is os.join.path in glibc-package.bbclass (here line 257) Feb 21 16:23:27 the second argument is "/usr/share" which discards the first argument which is $treedir Feb 21 16:23:43 the leading "/" is not good in this case Feb 21 16:28:36 DoKu: did you manage to figure out where "datadir" gets set in the first place? Feb 21 16:29:39 not yet, i'm searching.... Feb 21 16:30:16 DoKu: good find tho :) Feb 21 16:31:34 found in line 172 the call: datadir = bb.data.getVar('datadir', d, 1) Feb 21 16:32:06 cbrake: are you there Cliff? Feb 21 16:32:56 koen: any ideas on why datadir in glibc.bbclass gets set to '/usr/share' ? Feb 21 16:34:17 * koen dunnos Feb 21 16:34:22 glibc is deep dark magic Feb 21 16:34:44 ain't that the truth! Feb 21 16:35:37 koen: and even more so if run through the deeper and darker magic of bitbake :) Feb 21 16:36:08 nah, bitbake is high and bright compared to glibc Feb 21 16:36:43 03koen 07org.oe.dev * rc7087b82... 10/ (1 conf/distro/include/openmoko.inc): openmoko.inc: remove PARALLEL_INSTALL_MODULES Feb 21 16:37:02 hmmm.. it should be possible to tell getty to use framebuffer output somehow? Feb 21 16:38:08 koen: so glibc is kinda like the "magic smoke" that makes all electronics work :) Feb 21 16:38:49 yep Feb 21 16:47:18 xumbi8: Hi Jim Feb 21 16:50:45 xumbi8: be back in 30 minutes -- lunch ... Feb 21 16:57:02 DoKu: sounds like it's time for some debugging in glibc-package.bbclass. maybe add a few bb.note calls to see what's going on Feb 21 16:57:45 yep, that's what i'll do. but i have to stop for today.... bye Feb 21 16:58:13 DoKu: np. I might play with it when I get home tonight, if I can find some spare time :) Feb 21 16:59:28 cbrake_away: ah nuts I missed you, well if you see this when you get back I have a quick question about the compulab machine .conf file Feb 21 16:59:37 Jin^eLD: Or do you think you might want to take a stab at this? it seems like /usr/share is sneaking into the cmdline for locale generation... so maybe qemu is not at fault after all :) Feb 21 17:00:59 hvontres|poodle: I could take a look later, maybe in 2 hours or so Feb 21 17:02:37 Jin^eLD: np. I won't get home for another 10 or so...:) Feb 21 17:02:44 :) Feb 21 17:08:53 I'm trying to get together a set of libtool patches that'll allow 1.5.22 to be built. Nowhere in the maintainers file or in bugzilla have I been able to find any indication as to who put the initial recipe together. Feb 21 17:09:25 re Feb 21 17:09:25 I got new desk and it looks like it is too small ;( Feb 21 17:11:32 hrw: my mother would say "No, you have too much junk" Feb 21 17:11:54 ;D Feb 21 17:17:23 on desk I have: progear with usb keyboard, spitz, husky, tosa, 20" lcd wide, 4slot card reader, usb hub which act also as cup heater, speakers, wireless phone (+ cradle), mouse, keyboard, small lamp, and lollipop with 'kocham ciÄ™' text ('I love you' in Polish) which I got from Ania Feb 21 17:17:26 uclibc for x86 fails with a very strange error -> http://rafb.net/p/4S3NJ373.html Feb 21 17:17:38 koen: does it is too much? :) Feb 21 17:18:21 hrw: *I* don't think that's too much ;) Feb 21 17:18:25 koen : could this be from the insane class update Feb 21 17:19:02 it used to compile fine till i add the changes to the insane class Feb 21 17:19:16 hrw, usb hub that acts like a cup heater?..where can I get one of those? Feb 21 17:20:12 Kristoffer: on 'I want one of those' Feb 21 17:21:20 Kristoffer: but this is usb 1.1 hub :( Feb 21 17:26:50 xumbi8: I'm back Feb 21 17:27:29 cbrake: howdy Cliff Feb 21 17:28:18 I have another nice device i can bring with me to FOSDEM :-) Feb 21 17:29:38 cbrake: I'm just trying to add a new machine configuration based on compulab-pxa270.conf, when I set PREFERRED_PROVIDER_virtual/kernel = "compulab-pxa270" instead of "${MACHINE}", I run into problems Feb 21 17:30:07 cbrake: I see this when running bitbake bootstrap-image: Feb 21 17:30:08 NOTE: multiple providers are available (linux-geodelx, linux-xxs1500, shepherd-kernel-2.4-embedix, LAB-kernel, linux-h1940, ipod, linux-omap-2.6, poodle-kernel-2.4-embedix, a780-kernel, linux-bast, linux-geodegx, corgi-kernel-2.4-embedix, xanadux-ba-2.6, linux, xanadux-un-2.6, poodle255-kernel-2.4-embedix, linux-h6300-omap1, linux-mtx-3, husky-kernel-2.4-embedix, linux-dht-walnut); Feb 21 17:30:10 NOTE: consider defining a PREFERRED_PROVIDER to match runtime kernel Feb 21 17:30:49 cbrake: it's like "compulab-pxa270" isn't a valid kernel or something Feb 21 17:31:03 xumbi8: strange -- bitbake compulab-pxa270 works fine Feb 21 17:31:12 hrw, nice :D I want one of those Feb 21 17:32:01 cbpage: the compulab should have a COMPATIBLE_MACHINE set Feb 21 17:32:01 em Feb 21 17:32:01 cbrake: Feb 21 17:32:02 xumbi8: edit compulab-pxa270 kernel recipe and check COMPATIBLE_MACHINE Feb 21 17:32:41 hrw: a-ha! that's got to be it Feb 21 17:32:52 * cbrake fixes ... Feb 21 17:33:10 cbrake: that doesn't need fixing Feb 21 17:33:43 koen: ohh, so xumbi8 needs to set for his machine then if he is doing a custom machine. Feb 21 17:34:05 cbrake: right Feb 21 17:34:27 basically I'm just trying to create a new project, but basing it on the compulab config files from Cliff Feb 21 17:34:29 COMPATIBLE_MACHINE = '{compulab-pxa270|xumbimachine)' Feb 21 17:34:37 ah cool Feb 21 17:34:42 ehm Feb 21 17:34:44 () Feb 21 17:34:46 not {) Feb 21 17:34:49 heh Feb 21 17:35:06 cbrake: any news on a non-evil NAND driver? Feb 21 17:35:23 koen: yes, just got complete JFFS2 system running Feb 21 17:35:33 cbrake: nice! Feb 21 17:35:51 cbrake: so finally compulab-pxa270 can join EABI family Feb 21 17:35:56 koen: the trouble is the BL does not support it, so you still need to boot off something else to flash the JFFS2 partition. Feb 21 17:36:26 xumbi8: yes, I have an updater that will reflash kernel, nor flash, or nand flash from USB disk on boot :-). Feb 21 17:36:49 cbrake: very cool Feb 21 17:36:54 hrw: koen still trying to figure out the best way to support jffs2 on compulab for the new user Feb 21 17:37:00 03erik 07org.oe.dev * r6fba3dd9... 10/ (6 files in 3 dirs): Feb 21 17:37:00 sqlite3 3.3.13: Point release Feb 21 17:37:00 * Sqlite has released a new point release for Sqlite3. It fixes one bug when Feb 21 17:37:00 using ORDER BY. Feb 21 17:37:00 * The patches for 3.3.12 have been rebased for 3.3.13 and are included in Feb 21 17:37:01 the 3.3.13 directory. Feb 21 17:37:03 * The 3.3.13 bb file is an almost exact copy of the 3.3.12 bb. Feb 21 17:37:36 Just a question, how many swedes in this channel? :) Feb 21 17:38:02 Most seem to be german, english or dutch Feb 21 17:38:31 AFK food Feb 21 17:42:56 * koen reads http://www.linux.com/article.pl?sid=07/02/21/1340237 Feb 21 17:43:12 interesting observation about dynamically linked package managers Feb 21 17:47:07 nuts, now I get this error "ERROR: Cannot satisfy the following dependencies for task-base: sysvinit-inittab" Feb 21 17:48:50 xumbi8: bitbake -c rebuild sysvinit Feb 21 17:50:01 koen: d'oh, that should have been obvious, thanks! Feb 21 17:59:30 koen: in your update of openmoko-dates, shouldnt the protocol be "hhtps" and not "http" ? It failes without "https" Feb 21 18:00:12 s/"hhtps"/"https" Feb 21 18:00:50 goxboxlive: it should be http, but thanks to SRCDATE it will only work starting tomorrow :( Feb 21 18:01:06 ok Feb 21 18:02:10 it was moved from OH to openmoko today Feb 21 18:02:26 i c Feb 21 18:02:43 hrw: that's contacts, dates is still at OH iirc Feb 21 18:02:53 ah. Feb 21 18:03:01 hrw: it did move from branches/private/omoko to branches/openmoko Feb 21 18:03:49 koen: do you know does someone started work on X11R7.2? Feb 21 18:04:26 you mean on X.org proper or bbs for X.org? Feb 21 18:04:52 ah, 7.2 got released Feb 21 18:05:04 hrw: AFAIK neither ph5 or me are working on it Feb 21 18:05:18 wasn't there an RFC for epochs on the list? Feb 21 18:05:25 was Feb 21 18:07:58 * koen wants to work on gettting uclibc .29 in OE Feb 21 18:07:58 koen: There is also something strange about openmoko-panel-gsm. When i start a fresh bitbake openmoko-devel-image openmoko-panel-gsm failes installing (it compiles). It seems like it's starts builds to soon, should maybe have been buildt after libgsm and gsmd. Feb 21 18:08:02 .29 + nptl Feb 21 18:21:21 koen: Is 0.29 the one that includes NPTL or is it .30? Feb 21 18:21:34 .30 Feb 21 18:21:44 but I want to patch .29 to have ntpl Feb 21 18:21:54 nice! Feb 21 18:22:06 I'd be interested in testing that. Feb 21 18:22:32 WIth the added bonus of freeing more flash space on my slug! Feb 21 18:30:05 xumbi8: CONFIG_JFFS2_SUMMARY cust NAND mount time about in half. Still pretty long at 10s, so I'll probably still reduce the partition size. Feb 21 18:30:10 *cuts Feb 21 18:30:51 cbrake: and summary images are a bit larger Feb 21 18:30:56 03florian 07org.oe.dev * rfdb255e3... 10/ (4 files in 3 dirs): soundserver: Work around libtool breakage. Feb 21 18:35:14 hrw: yeah, I noticed that -- tradeoffs. After putting up with corrupted filesystems with the old NAND solution, I think everyone is going to be pretty happy with jffs2 -- slow mount times and all ... Feb 21 18:35:54 logfs! Feb 21 18:36:35 cbrake: iirc OLPC guys worked on speedup mount time - they have 512MB o flash Feb 21 18:37:05 hrw: dwmw2 said they were ugly hacks Feb 21 18:37:19 argh Feb 21 18:37:29 cbrake: send a mail to Joern Engel asking if logfs has settled on one format yet Feb 21 18:37:43 cbrake: he said to me he'd decide on one in ~march Feb 21 18:39:55 logfs stores the tree, so you don't have to reconstruct it everytime at mount Feb 21 18:52:00 hey koen. logfs sounds interesting. is it usable yet? Feb 21 18:52:25 mreimer: right now it's like reiserfs a few years ago Feb 21 18:52:51 it works great, but changes the on-disk format faster than most people change socks Feb 21 18:52:59 ha Feb 21 18:53:16 koen: do you know where the code is? Feb 21 18:53:35 there was a patch on lkml a while ago Feb 21 18:53:44 * mreimer googles Feb 21 18:53:44 august 2006 Feb 21 18:53:45 dunno where joern keeps his code Feb 21 18:54:15 don't confuse it with logfs.sf.net Feb 21 18:54:23 yeah, I saw that Feb 21 18:57:52 OK, I am almost on the way to FOSDEM Feb 21 18:59:18 Crofton i start making the suitcase just now :) Feb 21 18:59:49 I need a day or two to sort out jet lag Feb 21 19:00:15 where are u ? Feb 21 19:00:39 Roanoke Virginia Feb 21 19:00:41 USA Feb 21 19:00:53 Dulles Airport in about 2 hours Feb 21 19:01:11 uhh yes ... good fly ! Feb 21 19:06:29 Crofton: Good luck, and don't fly JetBlue :) Feb 21 19:06:51 Crofton: cu @fosdem then Feb 21 19:07:10 hvontres|poodle, I own stock in them :( Feb 21 19:07:23 They don't flow from this cow town though Feb 21 19:07:32 but at least there is free wireless Feb 21 19:07:41 hrw, see you in a few days Feb 21 19:13:09 koen: logfs looks interesting Feb 21 19:19:11 koen : its "official" the changes in the ipk dir, break gettext when compiling with uclibc on x86... Feb 21 19:19:42 koen : the strange thing is that for efika it compiles fine... Feb 21 19:19:52 Ifaistos: I doubt that Feb 21 19:20:00 Ifaistos: deploy/ipk ha snothing to do with compiling Feb 21 19:20:56 koen : i first run i clean build on efika-uclibc ucing the changed classed and it compiled ok Feb 21 19:20:56 koen, about fosdem, the only thing i don't know is the route from hotel to univrersity ... we meet sat morning in hall ? Feb 21 19:21:14 gremlin[it]: probably at breakfast Feb 21 19:21:39 08am ? Feb 21 19:22:53 koen : tried x86, failed complaining that a file was missing during configure -> http://rafb.net/p/4S3NJ373.html Feb 21 19:23:00 03koen 07org.oe.dev * re741e9a1... 10/ (1 packages/linux/linux-ezx_2.6.16.13.bb): linux-ezx: update to ezx9 Feb 21 19:23:21 cu guys Feb 21 19:23:29 Ifaistos: that's not because ipkgs moved around Feb 21 19:23:56 koen : not sure what is causing the problem Feb 21 19:24:16 most likely libtool(-cross) not being built Feb 21 19:25:10 koen : it was building fine before, but gettext has also some wierd dependacies when building for uclibc.... Feb 21 19:25:35 koen : there is something wrong... can't figure out what at the moment Feb 21 19:47:18 * gremlin[it] start crying ... this evening >>>ERROR: Build of angstrom-gpe-image failed<<< Feb 21 19:51:37 maybe found ... >>> | /media/hd1/oe/ipaq/tmp/rootfs/usr/lib/ipkg/info/udev-utils.postinst: line 10: /etc/udev/mount.blacklist: Permission denied <<< is it touching file outside chroot environment ? Feb 21 20:01:15 mhh how is generated or located the udev-utils.postinst file (for ipkg package) ??? Feb 21 20:03:36 gremlin[it]: pkg_postinst_append() { Feb 21 20:03:43 from udev_100.bb Feb 21 20:04:06 I'll fix it Feb 21 20:04:16 udev 100? Feb 21 20:04:21 I thought that one wasn't used Feb 21 20:04:24 it seem udev-092 for me ... Feb 21 20:04:31 092 should be the default Feb 21 20:04:40 since 100 was too buggy Feb 21 20:04:55 ah, ok Feb 21 20:05:03 it's probably the same there Feb 21 20:06:04 yeah, it's the same in all of them Feb 21 20:06:19 the reouble is that /etc/udev don't exist :( ... Feb 21 20:06:23 s/reouble/trouble/ Feb 21 20:07:09 also easy to fix Feb 21 20:07:12 NAiL: iirc slugos only uses 100 to have a feature slugos stopped using months ago Feb 21 20:10:04 Well, I'm fixing all of them ;-) Feb 21 20:10:38 gremlin[it]: fixed now ;) Feb 21 20:11:31 thanks NAiL Feb 21 20:11:34 03nail 07org.oe.dev * r0a98356e... 10/ (4 files in 2 dirs): udev: Fixup etc->sysconfdir, make sure sysconfdir/udev exists before writing into it Feb 21 20:12:35 np Feb 21 20:19:48 03rpurdie * r762 10bitbake/lib/bb/runqueue.py: runqueue.py: Fix recrdeptask handling (from poky) Feb 21 20:22:52 NAiL sorry what is the correct sequence of operation to get the full rebuild of udev and final image .. i have the same error also after pulling :( Feb 21 20:23:10 dunno Feb 21 20:23:36 RP: it looks like you fixed #1892 Feb 21 20:25:44 gremlin[it]: http://rafb.net/p/oVKIfJ12.html Feb 21 20:26:06 Do you have a clue what's wrong? I was told ASIC2 should work for h3800 Feb 21 20:27:18 hvontres|poodle: I fear I won't make it today Feb 21 20:27:51 philippe no nothinmore the something in the code around adc is wrong ... Feb 21 20:28:51 gremlin[it]: what are you working on atm? h3600 or h3800? Feb 21 20:29:45 philipe at now about nothing ... i'm more involved in personal thing .. i'm close to change work :) Feb 21 20:29:53 sorry Feb 21 20:30:23 philippe, 2.6.16-hh9 is very, very old kernel. use CVS HEAD ;-) Feb 21 20:30:32 gremlin[it]: no prob :-) HOwever we should catch up at Fosdem about this :-) Feb 21 20:30:47 psokolovsky_: I'll try... It is just that OE kind of forced me into using this one... Feb 21 20:30:51 for sure ! Feb 21 20:31:21 philippe, well, you tell OE what you want, not the other way around (local OE copy, of course ;-) ) Feb 21 20:32:13 psokolovsky_: true :-) It is just that I had something working there... Feb 21 20:34:08 channel hi Feb 21 20:34:16 errr, hi Feb 21 20:34:24 XD Feb 21 20:41:34 Jin^eLD: np. I probably won't either. But now I am curious Feb 21 20:42:28 hvontres|poodle: I'll let you know if I figure something out, once I get at it Feb 21 20:42:55 NAiL don't know why but install -d seem to not work :( Feb 21 20:46:35 how does it "seem" not to work? Feb 21 20:46:45 night everonye Feb 21 20:47:08 hey all Feb 21 20:47:17 /etc/udev isn't make ... Feb 21 20:47:36 hey kergoth Feb 21 20:48:32 NAiL >>> cannot create directory `/etc/udev': File exists <<< trying doing a mkdir into the script :( ... Feb 21 20:49:14 i think it try to do in wrog directory ... maybe on mu desktop /etc/ not the tmp/rootfs ;( Feb 21 20:49:30 s/mu desktop/my host/ Feb 21 20:50:42 gremlin[it]: mtn pull ; mtn up Feb 21 20:51:32 the postint stript contain the 'install -d' ... mys say i'm aligned ... Feb 21 20:51:42 s/mys/mtn/ Feb 21 20:52:04 I'm moving the install-dir-part to install instead of postinst ;) Feb 21 20:54:46 03nail 07org.oe.dev * r4037452c... 10/ (4 files in 2 dirs): udev: Fixup slight mistake in previous commit, bump PR Feb 21 21:04:10 haha, I just wanted to update sqlite3... Feb 21 21:08:21 no way Nail :( Feb 21 21:10:08 koen: I did? Feb 21 21:11:04 koen: I can't see what in r762 would fix that... Feb 21 21:11:32 hi kergoth Feb 21 21:11:56 Nail ... looking at /rootfs/usr/lib/ipkg/info/ipkg.postinst i see "install -d /media/hd1/oe/ipaq/tmp/rootfs//etc/rcS.d" ... while in udev-utils.postinst "echo "$root_partition" >> /etc/udev/mount.blacklist" ... where is /rootfs/ on the latter ??? Feb 21 21:15:42 woglinde: hey Feb 21 21:15:44 gremlin[it]: have you updated again Feb 21 21:15:55 ? Feb 21 21:16:05 yes Feb 21 21:16:12 koen: I can't see how I would leave STAGING_DIR for checks, os.walk is not following symlinks Feb 21 21:18:09 RP: well, gcc packaged correctly Feb 21 21:18:23 RP: which is what kept failing Feb 21 21:20:28 NAiL only other clue i can say that yestarday evening all was ok ... Feb 21 21:20:53 koen: All the change would do is change the number of tasks bitbake would run for recrdeptask which is basically images Feb 21 21:20:57 g'night Feb 21 21:21:03 nite gremlin Feb 21 21:21:10 RP: weird Feb 21 21:21:51 It explains why do_deploy wasn't running in poky properly. OE could probably use some fixes from poky in that area... Feb 21 21:22:42 RP: well, the o-hand poky page says to you are pushing stuff upstream instead of OE pulling in stuff, so go ahead :) Feb 21 21:23:58 koen: We do that now and again and that wasn't a hint, more an observation ;-) Feb 21 21:25:09 as another OE devel working on Poky I want to say that thare are few nice things in poky which would be nice to have in OE. Feb 21 21:25:19 qemu-native Feb 21 21:25:53 RP: I was reading the o-hand project pages today, and that jumped out ;) Feb 21 21:25:53 03koen 07org.oe.dev * r7e917aac... 10/ (1 packages/eds/eds-dbus_svn.bb): eds-dbus: add libglade to DEPENDS Feb 21 21:27:11 hrw: Anything else? Feb 21 21:27:21 koen: You'd agree we do push things? Feb 21 21:27:25 RP: I have to find time to do poky build first ;D Feb 21 21:27:35 RP: I agree *you* push things Feb 21 21:28:12 RP: which you say is in your own time, so that the wording is a bit misleading Feb 21 21:28:25 koen: I am the only one in OH who will touch monotone and am the nominated coordinator of such pushing Feb 21 21:28:25 RP: btw - my last build finished and works Feb 21 21:28:27 (an excuse for me to keep being lazy) Feb 21 21:28:36 hrw: Great :) Feb 21 21:28:54 RP: qvga looks funny in qemu Feb 21 21:29:15 hrw: Yes ;-) Feb 21 21:29:15 RP: hehe, that sounds funny (touch monotone) Feb 21 21:29:41 zecke: I found a QA problem Feb 21 21:29:46 ze: Feb 21 21:29:48 zecke: Feb 21 21:29:49 koen@bitbake:/data/build/koen/OE/build/tmp/angstrom/staging/powerpc-angstrom-linux$ find . -name xproto.pc Feb 21 21:29:50 ./lib/pkgconfig/xproto.pc Feb 21 21:29:50 ./share/pkgconfig/xproto.pc Feb 21 21:30:10 the one in share has correct paths, but not the one in lib Feb 21 21:30:13 zecke: did you looked into my insane changes to insane? Feb 21 21:30:14 koen: so fakeroot and xproto install the same file twice Feb 21 21:30:28 not the same file Feb 21 21:30:30 one good and one broken Feb 21 21:30:38 hrw: no, have slept most of the last two days Feb 21 21:30:48 zecke: added 32/64bit handling Feb 21 21:30:49 koen: wording mistake, still like fakeroot Feb 21 21:30:57 greentux: morning Feb 21 21:31:02 someone removed commented code Feb 21 21:31:10 zecke: I did - bitbake failed with it Feb 21 21:31:17 RP: are you interested in a backlight.c/corgi-bl.c patch to add a sysfs attribute and hook for enabling automatic backlight control, for devices with light sensors? Feb 21 21:31:21 hrw: interesting Feb 21 21:31:40 morning hrw ;) Feb 21 21:31:45 zecke: kind of 'unable to parse' Feb 21 21:31:47 mreimer: Someone emailed me about that recently and no, it does not belong in the backlight class Feb 21 21:31:53 koen: The intent is not to have to do the work with poky in my own time. Practial realities are different and I am working on that... Feb 21 21:31:57 RP: where would you suggest? Feb 21 21:32:10 mreimer: It needs its own class Feb 21 21:32:20 mreimer: or is done adhoc Feb 21 21:32:31 zecke: bitbake.conf sets PKG_CONFIG_PATH to "${STAGING_DATADIR}/pkgconfig" Feb 21 21:32:33 RP: ok Feb 21 21:32:57 zecke: I propose to symlink the lib/pkgconfig to share/pkgconfig or to change bitbake.conf Feb 21 21:33:06 mreimer: You have no idea how many random feature emails I get. We'd have about 40 attributes by now if I let them all in :/ Feb 21 21:33:45 RP: I understand Feb 21 21:34:39 btw - what kind of clothes are 'fosdem official'? t-shirt? shirt? sweater? Feb 21 21:34:51 mreimer: isn't a thing like that better done in userspace? Feb 21 21:34:56 hrw: suit or leather clothing Feb 21 21:35:02 koen: userspace can't frob gpios Feb 21 21:35:04 beard and sandals Feb 21 21:35:07 zecke: no latex? Feb 21 21:35:21 zecke: I need to bring my leathers? :) Feb 21 21:35:24 koen: beard I have. sandals too but it will be too cold Feb 21 21:35:29 mreimer: it can frob sysfs, right? Feb 21 21:35:29 hrw: sorry, no. We try to be less geeky Feb 21 21:35:55 koen: that's the question: where to put the sysfs knob that enables/disables the automatic backlight control Feb 21 21:36:04 koen: where will pkgconfig-native search for .pc files? Feb 21 21:36:19 koen: the attribute needs to hang off some kobj, but I don't know which one Feb 21 21:36:35 zecke: in PKG_CONFIG_PATH Feb 21 21:36:47 koen: right + builtin Feb 21 21:37:11 RP: packages/tasks/tasks_svn.bb? looks funny with task-oh.bb task-base.bb ;D Feb 21 21:37:15 mreimer: I meant let userspace read out the lightsensor and adjust backlight accordingly Feb 21 21:37:21 hrw: I know :-/' Feb 21 21:37:53 koen: on hx4700 it's automatically controlled. I set a gpio to 1 and the sensor automatically modulates the brightness. pretty nice Feb 21 21:37:59 RP: maybe we should 'mkdir oh-pim;mv dates/* contacts/* oh-pim/' etc? Feb 21 21:38:07 mreimer: aaah, now I understand Feb 21 21:39:00 hrw: Ask about that tomorrow. Others might have other plans Feb 21 21:39:03 time to hibernate - cu tomorrow or @fosdem Feb 21 21:39:09 RP: sure Feb 21 21:39:21 cu hrw|gone Feb 21 21:41:24 hi Feb 21 21:41:53 Ifaistos are you there? Feb 21 21:54:29 zecke: http://lists.linuxtogo.org/pipermail/openembedded-devel/2007-February/001492.html Feb 21 21:58:28 koen: 2nd is not practical Feb 21 21:58:46 koen: it is hard to get right, the question is why where two files installed? Feb 21 21:59:10 zecke: "xproto doesn't honour PKG_CONFIG_PATH and installs it in ${libdir}/pkgconfig." Feb 21 21:59:27 and pkgconfig.bbclass installs one to PKG_CONFIG_PATH on it's own Feb 21 21:59:35 koen: who installs the 2nd one, ah okay Feb 21 21:59:55 cat $pc | sed ${@get_pkgconfig_mangle(d)} > ${PKG_CONFIG_PATH}/$pcname Feb 21 22:00:47 there's option 4): have pkgconfig mangle every .pc file in staging everytime Feb 21 22:15:26 koen: hmm Feb 21 22:15:44 koen: we never had any issue due this. So it is tough to know what we will fix Feb 21 22:15:55 I made the symlink manually and insane is happy now Feb 21 22:16:17 koen: do you know out of your head which of the two files ended up being used? (do we set PKG_CONFIG_PATH)? Feb 21 22:16:31 the right ones are being used Feb 21 22:16:34 03freyther 07org.oe.dev * r1bfe0aeb... 10/ (1 classes/insane.bbclass): (log message trimmed) Feb 21 22:16:34 classes/insane.bbclass: Do not follow links, fix the issue that made koen add bb.fatal Feb 21 22:16:34 udev has a link to host binaries, do not check the architecture Feb 21 22:16:34 of these binaries. We will skip all links for now for the two Feb 21 22:16:34 reasons: Feb 21 22:16:34 -We either point to a file in our package and then we will Feb 21 22:16:36 check it anyway Feb 21 22:16:36 we set PKG_CONFIG_PATH Feb 21 22:17:24 zecke: all the solutions I see are awkward in one way or another Feb 21 22:24:06 hmm, 1st is the easiest, I fail to see the consequences though Feb 21 22:24:35 developer churn Feb 21 22:25:05 the deploy/ipk sort was easy as well, yet some people got angry because their upload scripts broke Feb 21 22:25:27 hallo Feb 21 22:26:06 xfsprogs in dev doesnt work for me unless i change the line in do_configure to export LIBTOOL="${STAGING_BINDIR_NATIVE}/${TARGET_SYS}-libtool" Feb 21 22:26:31 im still a bit new, this a bug or my config broken? Feb 21 22:26:39 timtimred: has xfsprogs a custom copy of libtool.m4? Feb 21 22:26:45 timtimred: in that case, remove it Feb 21 22:27:10 hmm no idea... i will check! Feb 21 22:28:19 hmm wait (validating my generic answer) Feb 21 22:28:36 export LIBTOOL="${STAGING_BINDIR}/${HOST_SYS}-libtool" seems weird Feb 21 22:28:57 if i download the source it does not contain that file Feb 21 22:29:23 yes and there is nothing in that directory Feb 21 22:30:10 timtimred: not knowing xfsprogs, do you have a $TARGET_SYS-libtool in your STAGING_BINDIR at all? Feb 21 22:31:35 zecke: STAGING_BINDIR_NATIVE should have all binaries you want to run Feb 21 22:31:46 zecke: I suspect RP missed xfsprogs in his mass conversion Feb 21 22:32:28 in the staging bin dir, if you mean for example (my target is powerpc) staging/powerpc-linux/bin/ then there is nothing in that dir Feb 21 22:32:37 * koen goes to sleep before saying (more) nonsense Feb 21 22:32:57 timtimred: more nonsense from here then Feb 21 22:33:21 so i should file this one in the bugtracker? Feb 21 22:33:22 timtimred: I doubt the file you point to exists Feb 21 22:33:48 timtimred: so even if it might compile, it only hides the issue Feb 21 22:34:01 whats the issue? Feb 21 22:34:05 timtimred: the first part STAGING_BINDIR vs. STAGING_BINDIR_NATIVE might be right Feb 21 22:34:10 timtimred: I don't know :) Feb 21 22:34:48 timtimred: I'm about to sleep as well, so a bug tracker is good, kicking people to remember bigs as well Feb 21 22:36:53 okay off to bed Feb 21 22:39:57 hi Feb 21 22:52:54 Hey CoreDump|home :) Feb 21 22:53:54 CoreDump|home: I backed up my changes, but now I need to update my altboot to the latest wip and test again. My current script is against 1.0.8 Feb 21 22:54:11 no hurry Feb 21 22:54:34 it would probably better to merge your changes with altboot_0.0.0.bb Feb 21 22:55:17 CoreDump|home: np. But as an "official junior developer" on OZ 3.5.5 I figured I'd try to get a head start on some of the integraton stuff I told hrw I could help out with :) Feb 21 22:55:42 ;) Feb 21 23:07:53 * kergoth starts a build with a mingw32 target and crosses his fingers Feb 21 23:20:35 time to go home :) Feb 22 01:17:15 how comprehensive is the CPAN build class? Feb 22 01:18:04 has anyone tried to build all available cpan packages using stub recipies? What percentage succeed? **** ENDING LOGGING AT Thu Feb 22 02:59:58 2007