**** BEGIN LOGGING AT Mon Apr 06 02:59:58 2009 Apr 06 03:06:01 03Mike Westerhof  07org.openembedded.dev * r26b1e96866 10openembedded.git/recipes/slugos-init/ (files/leds slugos-init_5.0.bb): SlugOS: leds - fix logic to attempt to beep even if udev hasn't yet run. Apr 06 04:36:49 03Mike Westerhof  07org.openembedded.dev * r45c1142d3e 10openembedded.git/recipes/tasks/task-slugos-toolchain-target.bb: Apr 06 04:36:49 SlugOS: toolchain/sdk - add a number of missing components Apr 06 04:36:49 provided by task-sdk-bare, as noted by Tom Rini. Apr 06 06:46:55 morning Apr 06 07:16:18 hi Apr 06 07:33:42 03Koen Kooi  07org.openembedded.dev * r389043beae 10openembedded.git/ (3 files in 2 dirs): Apr 06 07:33:42 kernel,module-base class, bitbake.conf: introduce MACHINE_KERNEL_PR Apr 06 07:33:42 * bandaid till we have a better way to track ABI changes Apr 06 07:33:52 03Koen Kooi  07org.openembedded.dev * r594a529544 10openembedded.git/ (10 files in 4 dirs): omap3 boards: introduce MACHINE_KERNEL_PR, put it in conf/machine/include/omap3.inc Apr 06 07:38:07 morning Apr 06 07:43:38 morning hrw Apr 06 07:44:08 hrw: about sam9g20 patches I sent you, I'm preparing a new set Apr 06 07:44:47 I read Koen NACK to my stable patches, so I'm preparing a dev patch set Apr 06 07:45:09 mckoan: I am merging Ulf versions already Apr 06 07:45:31 so I wasted my time, ok Apr 06 07:45:44 I'll wait them before continuing Apr 06 07:46:16 mckoan: I looked at your patches and they looked at Ulf versions so I took his work and checked how it looks Apr 06 07:47:08 mckoan: I think that will merge/push it before today evening Apr 06 07:47:46 I did 3 patches and only one pf them was Ulf work, so I yesterday I prepared a new set with the first taken from Ulf branch as-is with preoper Ulf's credit and two new made by me Apr 06 07:48:24 however I'll wait, and I home next time could be possible a better coordination among us ;-) Apr 06 07:48:32 s/home/hope Apr 06 08:28:39 when i run gxine, i get a segmentation fault, when i try to use gdb to debug all i see is Program received signal SIGSEGV, Segmentation fault. 0x40037354 in pthread_mutex_lock () from /lib/libpthread.so.0 Apr 06 08:29:51 how do i do source level debugging using gdb, cause i dont know which function in gxine is calling pthread_mutex, i see that gxine does not contain any debugging symbols Apr 06 08:43:32 mckoan: pushed Apr 06 08:43:34 03Ulf Samuelsson  07org.openembedded.dev * rb42626ddd4 10openembedded.git/conf/machine/at91sam9rlek.conf: Add support for AT91SAM9RL64 Apr 06 08:43:36 03Ulf Samuelsson  07org.openembedded.dev * r82f0e66ef8 10openembedded.git/recipes/at91bootstrap/ (8 files in 8 dirs): Add recipe for at91bootstrap-2.10 Apr 06 08:43:37 03Ulf Samuelsson  07org.openembedded.dev * r038dd6020c 10openembedded.git/conf/machine/at91cap9adk.conf: Add support for at91cap9adk development board with a at91cap9500 processor Apr 06 08:43:39 03Ulf Samuelsson  07org.openembedded.dev * r0289d87d39 10openembedded.git/conf/machine/at91sam9g20ek.conf: Add support for AT91SAM9G20EK Apr 06 08:43:42 03Ulf Samuelsson  07org.openembedded.dev * r7753dd2891 10openembedded.git/recipes/u-boot/ (5 files in 2 dirs): Add recipe for u-boot-2009.01 Apr 06 08:43:45 03Ulf Samuelsson  07org.openembedded.dev * r2b0bd3ae7f 10openembedded.git/conf/checksums.ini: Add checksums for new packages Apr 06 08:43:48 03Ulf Samuelsson  07org.openembedded.dev * r2604bdcde3 10openembedded.git/recipes/linux/linux-2.6.28/ (4 files in 4 dirs): Add linux-2.6.28 support for SAM9260/1/SAM9G20/ATNGW100 Apr 06 08:43:51 03Ulf Samuelsson  07org.openembedded.dev * rfba2f2496d 10openembedded.git/conf/machine/ (8 files): Update sam9260/1/3 and ap7000 (stk100,ngw100) to use linux-2.6.28, u-boot-2009.01 and at91bootstrap-2.10 Apr 06 08:44:45 has anyone been successful in running gxine or any mplayer frontend recently? Apr 06 08:45:00 hrw: ok Apr 06 08:48:28 hello all Apr 06 08:48:58 mckoan: sorry for that Apr 06 08:49:12 i've run into a problem with my build tree. Apr 06 08:49:35 everything i try to build is failing at do_qa_staging Apr 06 08:49:45 because of imlib.pc Apr 06 08:50:10 which has a -rpath-link optinion in it, which i don't think should be there. Apr 06 08:50:20 the thing is it was working fine this morning. Apr 06 08:50:57 something somewhere has changed, so every time i rebuild imlib, that -rpath-link optinion is re added, and breaks qa staging Apr 06 08:51:07 any ideas ? what has gone wrong ? Apr 06 08:51:48 from openembedded/build/tmp/staging/486sx-angstrom-linux/usr/lib/pkgconfig/imlib.pc Apr 06 08:51:54 Libs: -L${libdir} -Wl,-rpath-link,/mnt/data/openembedded/build/tmp/staging/486sx-angstrom-linux/usr/lib Apr 06 09:06:21 hrw: is it true is openmoko dead? Apr 06 09:06:47 I just read this at slashdot Apr 06 09:07:05 cityLights1: I was never related to Openmoko so cannot comment. Apr 06 09:07:40 hi hrw, I saw you at fosdem a year ago Apr 06 09:07:59 when you when with that other guy and sean Apr 06 09:08:42 cityLights: it was not me then Apr 06 09:08:42 that other guy (which I just 4got his name now) used my cellular battary to boot the first 1975 Apr 06 09:08:53 but you where there Apr 06 09:08:56 cityLights: I was at 2007 and 2009 fosdem Apr 06 09:09:03 year ago I was not present Apr 06 09:09:05 right 2007 Apr 06 09:09:33 anyway - I just want to ask - is this proves the concept is dead? Apr 06 09:09:43 no idea Apr 06 09:09:51 I do not track Openmoko Apr 06 09:09:53 does it mean we will never see truly open cellular? Apr 06 09:09:57 ask on #openmoko rather Apr 06 09:10:20 cityLights: freerunner and neo1973 were free and freerunner iirc is still available Apr 06 09:12:24 cityLights: there were some big problems with the choice of hardware with the openmoko, whether or not they were forced into those choices because of the opensource nature of the platfourm is unknown to me. Apr 06 09:13:16 does gpe environment support any spreadsheet programs, looking to add a few, any suggestions Apr 06 09:13:29 gr88: gnumeric? Apr 06 09:13:45 gr88: on VGA it was usable Apr 06 09:14:25 thanks ill check Apr 06 09:14:54 any other apps you recommend, planning to give a demo Apr 06 09:17:08 abiword, gimp? Apr 06 09:18:15 gimp? i use a 200MHZ vga lcd, it wont work Apr 06 09:25:55 hrw: email clients? which one works good on matchbox with vga Apr 06 09:26:05 sylpheed or claws-mail Apr 06 09:27:07 claws-mail is not supported on latest git Apr 06 09:27:27 so sylpheed-claws then Apr 06 09:30:40 hrw:thanks Apr 06 09:31:36 lovely2: so the reason google sucseeded with android while open moko didnt was a hardware choise?\ Apr 06 09:32:13 nope. Apr 06 09:32:25 openmoko would of had a better chance with a better hardware choice. Apr 06 09:32:34 the phone was 2.5G Apr 06 09:32:51 2.5? Where's the .5? Apr 06 09:32:58 for example the radio modem couldn't be connected to the internet and recieve a sms at the same time. Apr 06 09:33:07 broonie: edge they called 2.5g Apr 06 09:33:14 gprs Apr 06 09:33:42 ah.. forgot Apr 06 09:33:42 hrw: Quite; they don't have EDGE, only GPRS. Apr 06 09:33:52 openmoko phones use 2000/2001 gsm modem Apr 06 09:34:19 which probably was because they couldn't get a more recent radio modem, with out signing lots of NDA Apr 06 09:34:22 and is that such a big issue? you still have wifi Apr 06 09:34:31 huge issue. Apr 06 09:34:38 Isee Apr 06 09:34:38 i didn't buy one. Apr 06 09:34:50 would of it had 3g, or at least good edge Apr 06 09:34:51 hrw: you used any gps programs Apr 06 09:34:53 cityLights: my openmoko phones do not have wifi Apr 06 09:35:26 the wifi, was only added in the lastest version. Apr 06 09:35:50 i love the idea of opensource mobile phone. Apr 06 09:36:01 however, if a company is going to release such a device. Apr 06 09:36:09 WiFi was in all the production ones. Apr 06 09:36:15 they have to do bit of the heavy lifting, of making it work Apr 06 09:36:28 software wise. Apr 06 09:36:40 you can't let the oss community do all of it. Apr 06 09:37:20 last year when i looked into its software development, almost none of the applications where even near beta. Apr 06 09:37:37 big warnings about how it was more of a toy than a phone. Apr 06 09:39:29 just my 2cents Apr 06 09:39:37 03Koen Kooi  07org.openembedded.dev * r87ece9c2e0 10openembedded.git/recipes/linux/ (70 files in 9 dirs): Apr 06 09:39:37 linux-omap 2.6.28: introduce MACHINE_KERNEL_PR Apr 06 09:39:37 linux-omap 2.6.29: introduce MACHINE_KERNEL_PR, seperate out dss2 patches, add isp and musb patches Apr 06 09:39:42 freaking RPATH!!!!!! Apr 06 09:39:44 grrr Apr 06 09:41:36 right, this was also waht I UNDERSTOOD Apr 06 09:41:51 good morning Apr 06 09:42:15 hi mpoullet|work Apr 06 09:43:27 well off for now Apr 06 09:44:03 thanks lovely2 Apr 06 09:49:30 hrw: hello Apr 06 09:53:29 Hey guys. Apr 06 09:53:38 Would any of you remember how to fix "cannot map archive header: Invalid argument Apr 06 09:53:39 " Apr 06 09:53:41 (calling localedef) Apr 06 09:58:53 good morning Apr 06 10:00:19 gm Apr 06 10:00:51 Ainulindale: add second 64MB ram to device or generate locales during building (if it is arm) Apr 06 10:01:25 Well the real problem seems to be Apr 06 10:01:26 open("/usr/lib/locale/locale-archive", O_RDWR|O_LARGEFILE) = -1 ENOENT (No such file or directory) Apr 06 10:07:56 florian: good morning Apr 06 10:08:14 Ainulindale: I think that's harmless, localedef will create the archive if it doesn't exist. Apr 06 10:08:33 Well it doesn't create it it seems Apr 06 10:08:37 Ainulindale: "cannot map archive header: invalid argument" generally happens if your archive is stored on jffs2 or some other non-mappable filesystem. Apr 06 10:08:51 Well there I'm on the device. Apr 06 10:08:59 (GTA02) Apr 06 10:09:06 It seems to work for other people. Apr 06 10:09:37 The locale generation scripts generally arrange for the archive to be temporarily evacuated into ramfs for the duration. Apr 06 10:09:50 If you are running localedef by hand then you'll need to take the appropriate action for yourself. Apr 06 10:10:11 Or, if you're using the normal scripts but they aren't working, you'll need to figure out why. :-} Apr 06 10:10:28 Well I have no idea what may be wrong there Apr 06 10:10:39 and if it is gta02 then by default OE should provide already generated locales.. Apr 06 10:10:45 Which are the normal scripts? Apr 06 10:10:51 hrw: Well I thought that too Apr 06 10:10:55 But weirdly, it doesn't Apr 06 10:11:03 Ainulindale: the postinsts for locale-base-xx, I think. let me check. Apr 06 10:11:17 Hmmm. Apr 06 10:11:23 I may have done something wrong. Apr 06 10:11:32 But locale -a is insultant :-) Apr 06 10:11:48 yeah, see locale_base_postinst() in glibc-package.bbclass Apr 06 10:12:11 it copies the archive into /tmp, runs localedef, then copies the archive back again Apr 06 10:13:18 on the other hand, yes, you would normally expect to have the locales pre-generated at build time on gta02. Apr 06 10:24:34 03Jeremy Lainé  07org.openembedded.dev * r121da71778 10openembedded.git/recipes/libnl/ (4 files in 2 dirs): libnl: use LDFLAGS (fixes QA error) Apr 06 11:01:09 03Jeremy Lainé  07org.openembedded.dev * raf0a523736 10openembedded.git/recipes/tcp-wrappers/ (tcp-wrappers-7.6/ldflags tcp-wrappers_7.6.bb): tcp-wrappers: use LDFLAGS (fixes QA error) Apr 06 11:09:40 03Jeremy Lainé  07stable/2009 * r3530709b1d 10openembedded.git/recipes/sqlite/ (sqlite-2.8.17/mainmk_build_dynamic.patch sqlite_2.8.17.bb): Apr 06 11:09:40 sqlite: use LDFLAGS (fixes QA error) Apr 06 11:09:40 Signed-off-by: Jeremy Lainé Apr 06 11:09:40 Acked-by: Koen Kooi Apr 06 11:09:41 03Jeremy Lainé  07stable/2009 * rfeb216073f 10openembedded.git/recipes/libpcap/ (libpcap-1.0.0/ldflags.patch libpcap_1.0.0.bb): Apr 06 11:09:44 libpcap: add patch to use LDFLAGS (fixes QA error) Apr 06 11:09:46 Signed-off-by: Jeremy Lainé Apr 06 11:09:48 Acked-by: Koen Kooi Apr 06 11:45:03 03Jeremy Lainé  07org.openembedded.dev * r95cd55f0b7 10openembedded.git/recipes/cups/ (cups.inc cups_1.2.12.bb cups_1.3.8.bb): cups: use LDFLAGS for shared libs (fixes QA error) Apr 06 11:53:14 03Jeremy Lainé  07org.openembedded.dev * r5ad39eb0bc 10openembedded.git/recipes/libmng/ (libmng/ldflags.patch libmng_1.0.8.bb): libmng: use LDFLAGS (fixes QA error) Apr 06 12:01:09 03Jeremy Lainé  07org.openembedded.dev * r42b3663b07 10openembedded.git/recipes/dhclient/dhclient_2.0pl5.bb: dhclient: use LDFLAGS (fixes QA error) Apr 06 12:04:38 ant_work: hello Apr 06 12:04:50 hi jeremy_laine Apr 06 12:04:56 jeremy_laine: lot of QA work Apr 06 12:05:00 jeremy_laine: QA battle is over? Apr 06 12:05:04 hrw: hi! indeed Apr 06 12:05:35 jeremy_laine: thx again for fixing these Apr 06 12:05:47 ant_work: there are still a bunch of packages, I was wondering which one you personally need to be able to set LDFLAGS errors as fatal? Apr 06 12:06:21 I'll have to do as you did and set QA9 fatal. Then I'll do a list Apr 06 12:06:47 there were not many anymore, though Apr 06 12:07:13 hrw: same goes for you, if you want to try setting this: http://pastebin.com/m1c2b7df5 Apr 06 12:07:44 ant_work: I know I still need to fix hostap-daemon for one :) Apr 06 12:16:36 03Jeremy Lainé  07org.openembedded.dev * rfd2069004f 10openembedded.git/recipes/pth/ (files/ldflags.patch pth_2.0.2.bb): pth: use LDFLAGS (fixes QA error) Apr 06 12:27:50 jeremy_laine: it looks lite I'm the one hitting most QA... Apr 06 12:27:57 :/ Apr 06 12:32:52 jeremy_laine: apart OPIE/GPE I see pine, boost, netsurf, libmng, pth, postgressql, midori, js, ioquake3, gd, libsdl-gfx, libsdl-ttf, flite Apr 06 12:39:22 <_cpo__> is there yet any meta package which can build a GPL export of an openembedded build environment? Apr 06 12:39:41 <_cpo__> to comply with the licenses? Apr 06 12:42:38 ant_work: damn, we have too many versions of so many packages.. I only fix the default versions Apr 06 12:42:49 oh, sh@t Apr 06 12:44:14 jeremy_laine: these recipes obviously miss an .inc? Apr 06 12:45:02 ant_work: bah, I think we should kill some old versions, for instance "pth 2.0.0" doesn't even compile (2.0.2 OK), why so many gd versions?, .. Apr 06 12:45:22 only answer is grep... Apr 06 12:45:52 some old-unmaintainde machine? distro? Apr 06 12:47:12 03Jeremy Lainé  07stable/2009 * rfa827926ea 10openembedded.git/recipes/libnl/ (4 files in 2 dirs): Apr 06 12:47:12 libnl: use LDFLAGS (fixes QA error) Apr 06 12:47:12 Signed-off-by: Jeremy Lainé Apr 06 12:47:12 Acked-by: Koen Kooi Apr 06 12:47:19 03Jeremy Lainé  07stable/2009 * rf459205665 10openembedded.git/recipes/tcp-wrappers/ (tcp-wrappers-7.6/ldflags tcp-wrappers_7.6.bb): Apr 06 12:47:19 tcp-wrappers: use LDFLAGS (fixes QA error) Apr 06 12:47:21 Signed-off-by: Jeremy Lainé Apr 06 12:47:23 Acked-by: Koen Kooi Apr 06 12:50:28 03Jeremy Lainé  07org.openembedded.dev * r877213524a 10openembedded.git/recipes/gd/gd_2.0.33.bb: gd: use LDFLAGS (fixes QA error) Apr 06 12:58:01 ant_work: ehm.. pine looks pretty wrong: Apr 06 12:58:18 RDEPENDS_pine = "pico" Apr 06 12:58:28 shouldn't that be a "+=" ?? Apr 06 12:59:08 usually is += Apr 06 12:59:49 ant_work: also, I see pine embedded libc-client, which is available as a separate package Apr 06 13:00:22 oh..sorry..I've never used pine...just found on the QA list...can't say offhand Apr 06 13:00:39 pine... my first mua Apr 06 13:02:03 jeremy_laine: I think it's safe to RDEPENDS_${PN} += Apr 06 13:02:21 ant_work: I think I'm not going to touch pine, too many unrelated changes Apr 06 13:02:22 and sane too Apr 06 13:02:50 np, we should bother the maintainer Apr 06 13:06:35 pine suxx Apr 06 13:06:47 we should have alpine instead - more sane license iirc Apr 06 13:07:07 jeremy_laine: that QA level change... cannot it be set in local.conf? Apr 06 13:07:17 Yes, you can't distribute modified pine binaries. Apr 06 13:08:00 hrw: well sure, theoretically but we need an extra config variable Apr 06 13:08:43 hrw: I don't really want to make overiding QA errors too easy Apr 06 13:09:17 hrw: once most of the current QA errors related to LDFLAGS are fixed, I'd quite like to see these errors made fatal to catch any regressions Apr 06 13:10:10 jeremy_laine: next ennemy the 'non X contains .so' QA's Apr 06 13:10:32 (not so many happily) Apr 06 13:11:07 ant_work: yeah well I hope someone will step forward for those, I'm started to get kind of bored/pissed off with all these QA errors :) Apr 06 13:11:25 he he...the LDFLAGS were the easy ones ;-) Apr 06 13:11:35 ant_work: not realy, quite the opposite! Apr 06 13:11:49 ant_work: the .so errors shouldn't require patching the source Apr 06 13:12:14 yep, just parsing a long list of FILES_xy Apr 06 13:13:03 I'll have a look at the 'broken desktop' oneday Apr 06 13:13:15 these just need time Apr 06 13:13:16 03Koen Kooi  07org.openembedded.dev * rc273263b14 10openembedded.git/recipes/u-boot/ (files/dss2.patch files/fw-env.patch u-boot_git.bb): u-boot git: update overo and beagleboard to 2009.03 Apr 06 13:15:33 jeremy_laine: please post an RFC to have most QA as fatal. +1 from me Apr 06 13:16:25 so people (like me) won't risk to copy from an insane recipe... Apr 06 13:28:13 jeremy_laine: I replied to pine mail Apr 06 13:28:21 jeremy_laine: http://www.washington.edu/alpine/ Apr 06 13:35:54 Hi, anyone here who has some knowledge of qt embedded and pkg-config in oe? Apr 06 13:38:48 03Jeremy Lainé  07org.openembedded.dev * r66fe0d8e7f 10openembedded.git/recipes/postgresql/ (4 files): postgresql: use LDFLAGS for shared libs (fixes QA error) Apr 06 13:39:46 03Koen Kooi  07org.openembedded.dev * r616a026e64 10openembedded.git/ (6 files in 5 dirs): Apr 06 13:39:46 linux-omap 2.6.29: fixes Apr 06 13:39:46 * add patch to fix leds Apr 06 13:39:46 * tweak ISP patches to handle new API Apr 06 13:39:46 * tweak beagleboard defconfig Apr 06 13:39:56 03Koen Kooi  07org.openembedded.dev * r230bf4be4a 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Apr 06 13:41:33 03Jeremy Lainé  07org.openembedded.dev * rb3d66c0976 10openembedded.git/ (conf/checksums.ini recipes/postgresql/postgresql_8.2.4.bb): postgresql-8.2.4: fix SRC_URI and checksums.ini Apr 06 13:41:33 03Jeremy Lainé  07org.openembedded.dev * rd102e25d54 10openembedded.git/: Merge branch 'org.openembedded.dev' of ssh://git@git.openembedded.net/openembedded into org.openembedded.dev Apr 06 13:42:20 virk: yes, what is your question? Apr 06 13:42:33 hrw: sure, go for it :) Apr 06 13:43:42 I have a qt project here in which I've set CONFIG += link_pkgconfig PKGCONFIG += gsoap++ Apr 06 13:44:05 which works fine when building it locally, no problems whatsoever Apr 06 13:44:35 but when I bitbake it, qmake2 seems to mess up the include dirs Apr 06 13:45:14 first, here is the output of pkgconfig http://pastebin.com/d11269407 Apr 06 13:45:31 checked and double checked, everything in place there Apr 06 13:45:55 but have a look at this: http://pastebin.com/d53fcbb5e Apr 06 13:46:22 lines 4 and 5, just behind -DWITH_DOM Apr 06 13:46:33 virk: ehm, do you have "inherit qt4e" in your recipe? Apr 06 13:46:39 jep Apr 06 13:47:29 I tried with "inherit qt4e pkgconfig" as well Apr 06 13:47:38 but that makes no difference Apr 06 13:48:18 virk: oh, wait, it's not the QT includes dirs you're concerned about, it's the gsoap++ ones right? Apr 06 13:48:35 indeed Apr 06 13:50:13 I here is the bb file used: http://pastebin.com/d24a830a8 Apr 06 13:51:21 virk: by the way you can drop qt4-embedded from DEPENDS, qt4e will set it for you (better in case the package changes name) Apr 06 13:51:45 hmm, yes, that is quite logical indeed :) Apr 06 13:52:40 did you have a look at the makefile snippet? Apr 06 13:56:01 this line: -I/home/larm/oe/build/tmp/staging/armv5te-angstrom-linux-gnueabi/home/larm/oe/build/tmp/staging/i686-linux/usr/include Apr 06 13:56:30 is litteraly copied from the Makefile generated by oe's qmake2 Apr 06 13:56:59 virk: do you have a copy of your .pro file, I am trying to reproduce Apr 06 13:57:35 just a sec, it are actually multiple Apr 06 13:57:43 *.pro files Apr 06 14:00:16 virk: just the minimum to get the bad flags Apr 06 14:00:38 http://pastebin.com/d2378f9de Apr 06 14:02:02 soapH.h includes soapStub.h, which on its turn includes stdsoap2.h. The latter is provided by gsoap++ Apr 06 14:02:23 (and soapstub.h is generated by it) Apr 06 14:03:20 so in fact, any project in which you include stdsoap2.h should do, I'll create such small project Apr 06 14:06:31 virk: it's going to be a while, it turns out I don't have a built copy of qt4-embedded :( Apr 06 14:07:18 haha, ok, meanwhile I'll certainly have a tetsproject ;) Apr 06 14:08:46 (I cannot simply hand over this project, it's a (stupid) commercial thing and they're quite paranoid about leaks Apr 06 14:26:56 03Jeremy Lainé  07org.openembedded.dev * r0fa410aa02 10openembedded.git/recipes/gsoap/ (gsoap/rename_bogus_ldflags.patch gsoap_2.7.7.bb): gsoap: use LDFLAGS (fixes QA error) Apr 06 14:28:30 jeremy_laine: yep...indeed...I've seen some needs nasty hacking for LDFLAGS Apr 06 14:28:37 mucho work Apr 06 14:28:54 we owe you 'a' beer :-) Apr 06 14:29:18 ant_work: heh, next time you stop over in Paris, give me a shout :) Apr 06 14:29:29 possibly nov/dec Apr 06 14:29:47 ant_work: ok, I'll be really thirsty by then but OK :) Apr 06 14:31:16 ah, btw, except for the custom kernel that was originaly provided with it, I have machine conf for the olimex sam9-l9261. Where do I send it? Apr 06 14:34:17 03John Lee  07org.openembedded.dev * r188eaf3673 10openembedded.git/conf/distro/openmoko.conf: Apr 06 14:34:17 openmoko.conf: update sources mirror URL Apr 06 14:34:17 Someone changed the location of om sources mirror without modifying Apr 06 14:34:17 related settings, nor posting a proper notice on any mailing list. Apr 06 14:40:05 03John Lee  07fso/milestone5.5 * r8c46122ad4 10openembedded.git/conf/distro/openmoko.conf: Apr 06 14:40:05 openmoko.conf: update sources mirror URL Apr 06 14:40:05 Someone changed the location of om sources mirror without modifying Apr 06 14:40:05 related settings, nor posting a proper notice on any mailing list. Apr 06 14:43:37 jeremy_laine: can you make git branch on OE server with your stable fixes? pulling changes to build would be easier ;D Apr 06 14:44:36 hrw: ehm.. that means I have to read up on git and branching Apr 06 14:44:47 btw Apr 06 14:44:54 hrw: so far I've used Koen's scripts to send/apply patches to/from patchwork Apr 06 14:45:09 pplz, who has benchmarked ubifs against jffs2? Apr 06 14:45:15 how is it? :) Apr 06 14:45:51 jeremy_laine: tht reminds me that I had to improve his script Apr 06 14:46:48 hrw: it may be worth actually commiting those scripts to contrib, no? Apr 06 14:46:51 yes Apr 06 14:50:14 jeremy_laine: I wrote version which will accept multiple numbers Apr 06 14:53:41 jeremy_laine: I made an extremely simple (10 lines) test project to reproduce the error I got with gsoap Apr 06 14:54:25 is resolution of the lcd built into the x/gtk libraries. if i need a different resolution. should i recompile all the libraries? Apr 06 15:00:51 gr88: ? Apr 06 15:01:03 gr88: x11 can check or be told about dpi Apr 06 15:02:14 hrw: i built a x matchbox file system, my lcd supports lower resolutions, i want to checkout them since i feel they run slower Apr 06 15:07:38 i have a 500mhz cpu, i feel apps are slower to load, esp. apps like abiword take 15 or more seconds to load Apr 06 15:07:50 is that normal? Apr 06 15:07:59 i feel it could do something with resolution Apr 06 15:08:01 03Jeremy Lainé  07org.openembedded.dev * r5820b5762a 10openembedded.git/recipes/flite/ (flite-1.3/flite-1.3-Makefile.patch flite_1.3.bb): flite-1.3: use LDFLAGS (fixes QA error) Apr 06 15:08:02 03Jeremy Lainé  07org.openembedded.dev * r8ba3987485 10openembedded.git/: Merge branch 'org.openembedded.dev' of ssh://git@git.openembedded.net/openembedded into org.openembedded.dev Apr 06 15:12:09 gr88: what cpu? where is rootfs? Apr 06 15:12:16 morning Apr 06 15:13:17 hrw: ti omap3 cpu and rootfs resides in sd Apr 06 15:13:48 hrw: my sd speed is approx 13Mbps Apr 06 15:14:58 'morning Apr 06 15:15:58 gr88: in other words: beagleboard like Apr 06 15:16:11 03Marcin Juszkiewicz  07org.openembedded.dev * r442dcb5cfd 10openembedded.git/contrib/patchwork/pw-am.sh: pw-am.sh: script to import patches from patchwork into git repository Apr 06 15:16:27 hrw: yes Apr 06 15:17:55 17:17 hrw@home:oe$ git push Apr 06 15:17:55 fatal: The remote end hung up unexpectedly Apr 06 15:17:58 I hate it Apr 06 15:28:19 jeremy_laine: could you try to run this testproject whenever qt4 is compiled? Apr 06 15:28:22 http://users.edpnet.be/aegir/soaptest.tar.bz2 Apr 06 15:28:30 is it possible to build a debug built with sources, i want to debug gxine/xinelib on my target, is there a switch to bitbake to build debug libraries and applications Apr 06 15:28:57 autoconf needs a single env var to inject paths into the m4 search locations for *all* the autotools scripts Apr 06 15:29:00 hrmph Apr 06 15:35:35 hey Apr 06 15:35:46 how would an rpath find it's way into a .pc file ? Apr 06 15:36:00 rpath-link Apr 06 15:37:31 is there anyone here with qt4 embedded compiled that want's to try out building this simple project: http://users.edpnet.be/aegir/soaptest.tar.bz2 Apr 06 15:38:01 just to test wether it can fnd the gsoap headers Apr 06 15:38:55 fuck git Apr 06 15:40:58 What's it done to you? Apr 06 15:41:44 g'day kergoth Apr 06 15:43:34 hrw: i enabled qvga in the kernel, i see the images are split into 2. i get 2 startup screens. do i need to pass some argument to the xserver. with vga it works fine Apr 06 15:44:52 broonie: this time pushing refuses to connect Apr 06 15:45:27 fatal: The remote end hung up unexpectedly Apr 06 15:45:33 virk: ok, have tried running your example, and qmake does indeed seem to get totally confused in its substitutions.. Apr 06 15:45:35 sounds like you should be annoyed with the server, or you have your repo misconfigured.. neither of which should result in blaming git itself, since the connection to the server for a push is just ssh. Apr 06 15:45:52 hrw: git pull --rebase cured twice Apr 06 15:46:24 kergoth: or 5 other reasons Apr 06 15:46:31 anyway I gave up for today with it Apr 06 15:46:40 then let's hear them, complaining about shit it isn't to blame for is idiotic Apr 06 15:47:07 will rather fire solder iron and finally solder usb hsot cable which parts are on a table since friday Apr 06 15:47:35 jeremy_laine: the strange thing is that it does build locally Apr 06 15:50:05 vikr: upon closer inspection, the gsoap package seems strange, unless I am mistaken it doesn't seem to stage anything Apr 06 15:51:23 here i can find stdsoap2.h in tmp/staging/i686-linux/usr/include/ Apr 06 15:52:14 or isn't that the correct place o search for it? Apr 06 15:52:25 virk: what's your target platform? Apr 06 15:53:06 angstrom (armv5te) Apr 06 15:53:52 virk: ok, then it's wrong, it's as I suspected, gsoap is not staging anything Apr 06 15:54:28 virk: the stuff you see in tmp/staging/i686-linux is staged by gsoap-native Apr 06 15:55:08 is it possible to build debug libraries which contain debug symbols with source code Apr 06 15:55:51 yes, you just need to set the appropriate CFLAGS. Apr 06 15:56:10 (well, you don't actually get source code in the libraries, but you do get the debugging symbols that I think you are really asking for.) Apr 06 15:57:07 is it normal that gsoap doesntt stage anything? Apr 06 15:57:09 or did you mean that you actually want to package up the source code for installation on the target device? that would certainly be possible, though I don't think there is any existing infrastructure in oe to do it. Apr 06 15:57:26 if i want to do source level debug Apr 06 15:57:30 virk: absolutely not, it means you can't build anything using gsoap :) I am fixing it right now Apr 06 15:58:03 virk: still, that doesn't explain the weird flags we are seeing Apr 06 15:58:18 I just wanted to return to that Apr 06 15:58:30 pb_: xine/gxine dont work they crash, so i want to see why they crash, all i see is Program received signal SIGSEGV, Segmentation fault. 0x40037354 in pthread_mutex_lock () from /lib/libpthread.so.0 Apr 06 15:59:22 isn't there a quick (and dirty) workaround for it? like pausing bitbake just after qmake2 created the Makefiles so I can edit them myself and let bitbake resume thereafter? Apr 06 16:00:12 virk: I shouldn't worry about the flags for now, just by fixing gsoap your soaptest compiles for me Apr 06 16:00:33 ah, ok, that's awesome Apr 06 16:00:43 I tought that was the main issue Apr 06 16:01:12 03Jeremy Lainé  07org.openembedded.dev * rdd286ca541 10openembedded.git/recipes/gsoap/gsoap_2.7.7.bb: gsoap: actually stage headers and libraries.. Apr 06 16:01:17 virk: do a git pull and try again Apr 06 16:01:49 Would anyone be able to tell me what I'm supposed to do to regenerate the glibc binary locales? bitbake glibc? Apr 06 16:02:26 ok, just a sec Apr 06 16:03:53 virk: by the way, remove "files/" from your SRC_URI Apr 06 16:04:26 hi, how to fix such a warning: base_set_filespath usage is deprecated, squashfs-lzma-tools-native-3.1r2 should be fixed ? Apr 06 16:04:44 is 'files' the default search path or something? Apr 06 16:05:22 virk: yes, "${PN}-${PV}", "${PN}" and "files" are searched Apr 06 16:05:59 (No idea for locales regeneration?) Apr 06 16:06:00 ok, that's easy Apr 06 16:06:41 virk: so, does it now build for you? Apr 06 16:07:06 virk: hey, actually I don't seem to get those bogus flags now that gsoap.pc is actually found Apr 06 16:08:11 it's still building, for some reason it rebuilded python-native Apr 06 16:08:37 virk: also, you can drop the QMAKE_PROFILES definition Apr 06 16:09:24 virk: and get into the habit of only defining "PR" if its not equal to "r0", which is the default value (otherwise you are sure to get a comment on oe-devel@ if you submit a recipe) Apr 06 16:09:26 it now builds without a problem Apr 06 16:09:57 virk: cool! Apr 06 16:10:05 haha, I just set PR to zero after reading that I should make a habit of it of indeed adding it Apr 06 16:10:43 virk: PR= 0 is removal-candidate ;-) Apr 06 16:12:00 Ainulindale: isn't that related to the ENABLE_BINARY_LOCALE_GENERATION config variable? Apr 06 16:12:03 heh, I never really understood why having 'PR = r0' was such a big deal. sure, it's the default, but adding and removing it every time the thing changes is just more useless typing. Apr 06 16:12:35 jeremy_laine: Yeah, just changed it back to what it was before, and I was wondering how to properly regenerate stuff. Apr 06 16:12:59 pb_: kergoth's 'carnivore' script cought a lot of Easter-cleaning to do Apr 06 16:13:01 just so you'd know: http://docs.openembedded.org/usermanual/usermanual.html#recipes_versioning , in the blue box Apr 06 16:13:50 well, quite. I agree with the user manual :-} Apr 06 16:14:54 anyway, thanks a lot jeremy_laine Apr 06 16:15:06 Ainulindale: you could try 'bitbake -f -c package_do_split_gconvs glibc' or some such Apr 06 16:15:17 if that fails then cleaning and rebuilding glibc is probably the easiest option Apr 06 16:15:35 btw, where should I send additional machine conf files? Apr 06 16:15:36 Well I tried that, I had a bad "configure: error: C preprocessor "arm-angstrom-linux-gnueabi-gcc -E" fails sanity check Apr 06 16:15:40 " Apr 06 16:15:40 So I'm not that happy :-) Apr 06 16:15:50 heh, oh dear. Apr 06 16:15:58 did you check config.log to find out what it thought was wrong with your preprocessor? Apr 06 16:16:09 Well, not yet Apr 06 16:16:11 Just saw that Apr 06 16:17:58 Hmmpf Apr 06 16:18:02 limits.h, no such f ile or directory Apr 06 16:18:04 sounds nice Apr 06 16:18:52 pb_: http://kergoth.pastey.net/111232 .. heh, was just a proof of concept, to see what sorts of things could be easily checked for, when it comes to recipe best practices.. didn't intend it to spark a lot of cleanup commits persay :) Apr 06 16:19:17 kergoth: doh, careless talk costs lives and all that Apr 06 16:19:22 * kergoth chuckles Apr 06 16:19:25 oops Apr 06 16:19:30 looks cute though Apr 06 16:19:37 http://kergoth.pastey.net/111234 Apr 06 16:20:02 heh. 'kergoth_sanity'? Apr 06 16:20:28 heh, I've never been very good at naming. just figured it was an extension of sanity/insane.bbclass, but for recipe metadata checks Apr 06 16:21:05 yah, just makes it sound like it's a class for checking your own personal sanity. :-} Apr 06 16:21:14 hehe Apr 06 16:21:38 Well I'm wondering why there is no limits.h Apr 06 16:22:38 Ainulindale: 5 bugs http://bugs.openembedded.net/buglist.cgi?quicksearch=limits.h Apr 06 16:23:00 Damn, google failed me. Apr 06 16:23:32 iirc was OpenSuse issue Apr 06 16:24:31 So, now, I need to find by which lib I should replace that one :-) Apr 06 16:24:49 I'll go with linux/limits.h Apr 06 16:29:40 pb_: if i want to change the cflags, where do i change it, the .bb file does not contain any cflag variable Apr 06 16:32:01 gr88: it's set in bitbake.conf but you can override it from anywhere you like. Apr 06 16:32:05 local.conf, bb file, command line, ... Apr 06 16:54:20 howto fix Apr 06 16:54:23 QA Issue: imlib.pc failed sanity test (tmpdir) Apr 06 16:54:28 ? Apr 06 16:55:04 03Jeremy Lainé  07org.openembedded.dev * rfe6822f3e8 10openembedded.git/recipes/gnome/libsoup-2.4_2.26.0.bb: libsoup-2.4: add missing DEPENDS on sqlite3 Apr 06 17:03:53 bye for now Apr 06 17:11:04 only 207 oe-dev emails to delete .... Apr 06 17:11:42 heh Apr 06 17:11:51 and review my sdk ones? ;) Apr 06 17:11:57 i think i need 2 acks Apr 06 17:12:44 depends how scary they look :) Apr 06 17:13:06 not at all :) Apr 06 17:14:00 those changes look good to me Apr 06 17:14:15 * kergoth battles autotools some more Apr 06 17:15:49 * kergoth chuckles at http://software.ericsink.com/entries/hg_denzel.html Apr 06 18:20:23 good morning Apr 06 18:22:58 hey khem Apr 06 18:23:30 I have now gcc svn recipe working Apr 06 18:23:44 but kernel does not compile due to what I think is a kernel bug Apr 06 18:23:48 heh Apr 06 18:24:02 Any chance you can look over my sdk changes today please? :) Apr 06 18:24:51 fs/ioctl.c: In function 'file_ioctl': Apr 06 18:24:51 fs/ioctl.c:358: error: case label does not reduce to an integer constant Apr 06 18:24:57 Tartarus: yes sure Apr 06 18:25:01 thanks Apr 06 18:57:48 pb_, pb__, did you test ONLINE_PACKAGE_MANAGEMENT stuff w/ an image that removes packages via ROOTFS_POSTPROCESS_COMMAND ? Apr 06 19:18:18 03Khem Raj  07org.openembedded.dev * r877fed8078 10openembedded.git/recipes/gcc/ (153 files in 5 dirs): gcc-svn: Add recipe to build bleeding edge gcc. Apr 06 20:30:49 hmmm Apr 06 20:31:35 * kergoth thinks Apr 06 20:33:55 don't overdo it Apr 06 20:34:07 :P Apr 06 20:37:47 who's maintaining git.openembedded.net? Apr 06 20:38:21 goddamnit, i have hiccups again.. hiccups + heartburn = ow Apr 06 20:42:35 khem, do you know who / if anyone maintains the gcc csl recipes? Apr 06 20:47:03 http://www.thinkgeek.com/geektoys/rc/b527/ = awesome Apr 06 21:13:46 hmm Apr 06 21:23:40 Tartarus: koen sort ok Apr 06 21:23:59 Tartarus: I also did some stuff but I dont use it. Apr 06 21:26:30 03Khem Raj  07org.openembedded.dev * rbe2795a6bf 10openembedded.git/conf/distro/include/sane-srcrevs.inc: sane-srcrevs.inc : Rev up eglibc and add revision for gcc-svn. Apr 06 21:33:59 ok Apr 06 21:34:17 Once I get a handful of images built (and thus I know should build) I'm gonna kick them into shape a wee bit Apr 06 21:34:54 khem, note that my task-slugos-toolchain-target patch isn't needed now that mwester switched to task-sdk-bare Apr 06 21:34:59 I should go update my bundle, oops :) Apr 06 21:38:34 yeah read that Apr 06 21:38:43 I read my mail from oldest :) Apr 06 21:51:06 khem, florian, OK to put Acked-by you guys in the commit msgs for the toolchain stuff? Apr 06 21:51:43 re Apr 06 21:51:49 Tartarus: yep Apr 06 21:59:44 re Apr 06 22:08:42 03Tom Rini  07org.openembedded.dev * r176aebf4b2 10openembedded.git/recipes/binutils/ (12 files): Apr 06 22:08:42 binutils-cross-sdk: Make relative, not absolute symlinks for 'as', etc, bump PR Apr 06 22:08:42 While I'm in here, make the cvs recipe use FILESPATHPKG not mung FILESDIR. Apr 06 22:08:42 Acked-by: Florian Boor Apr 06 22:08:42 Acked-by: Khem Raj Apr 06 22:08:45 Signed-off-by: Tom Rini Apr 06 22:08:47 03Tom Rini  07org.openembedded.dev * r439e60b246 10openembedded.git/recipes/gcc/ (15 files): (log message trimmed) Apr 06 22:08:50 gcc-cross-sdk: Fix relocation of the toolchain and bump PR. Apr 06 22:08:52 There's two parts to this. The first is to make relative, not absolute Apr 06 22:08:54 symlinks for 'cpp', etc. The second is that we need to configure without Apr 06 22:08:56 --with-gxx-include-dir and instead install the base C++ headers into the Apr 06 22:08:58 expected location. The path passed to --with-gxx-include-dir will not be Apr 06 22:09:00 relocated and is an absolute. Apr 06 22:09:02 03Tom Rini  07org.openembedded.dev * rda82d156d7 10openembedded.git/recipes/meta/ (9 files): (log message trimmed) Apr 06 22:09:05 meta-toolchain, canadian-sdk: Move C++ headers into the correct location, bump PR Apr 06 22:09:07 A gcc that has sysroot support expects that the default C++ headers Apr 06 22:09:13 (iostream, etc) are in /include/c++ while regular C headers Apr 06 22:09:15 are still in /usr/include. Apr 06 22:09:17 Acked-by: Florian Boor Apr 06 22:09:19 Acked-by: Khem Raj Apr 06 22:09:21 03Tom Rini  07org.openembedded.dev * r4c45e60d52 10openembedded.git/recipes/meta/meta-toolchain-gpe-sbox.bb: Apr 06 22:09:24 meta-toolchain-gpe-sbox: Drop task-sdk-gpephone (to match m-t-gpe), bump PR Apr 06 22:09:28 Acked-by: Florian Boor Apr 06 22:09:30 Acked-by: Khem Raj Apr 06 22:09:32 Signed-off-by: Tom Rini Apr 06 22:09:40 wonder what the status of richard's BBCLASSEXTENDS stuff is Apr 06 22:09:41 03Tom Rini  07org.openembedded.dev * r28d0f54e30 10openembedded.git/recipes/wireless-tools/ (wireless-tools.inc wireless-tools_29.bb): (log message trimmed) Apr 06 22:09:43 wireless-tools: Fix libiw/ifrename packages, bump PR. Apr 06 22:09:45 We weren't creating the above packages as we needed to use '=+' not '+=' here Apr 06 22:09:47 to take things from the default package. For these additional packages, add Apr 06 22:09:51 files with '=' not '+=' and we only need to add ${sysconfdir}/network to the Apr 06 22:09:53 base package glob. Also we need to run install-static to get libiw.a to be Apr 06 22:09:53 err, BBCLASSEXTEND Apr 06 22:09:55 installed. Apr 06 22:11:59 Somoene has fixed the message you get when running bitbake w/ metadata not in a git repo, yeah? Apr 06 22:18:23 03Andrea Adami  07org.openembedded.dev * r747d8eec1f 10openembedded.git/ (3 files in 2 dirs): initramfs-kexecboot-image: fix EXTRA_IMAGEDEPENDS for all Zaurus Apr 06 22:18:59 hehe Apr 06 22:19:12 ant__: final fix? :) Apr 06 22:19:34 hmmmm Apr 06 22:19:41 seems so..but personally I would get rid of those headers Apr 06 22:20:04 hey ant__ Apr 06 22:20:15 well.. one function left.. Apr 06 22:20:16 Have you done any thinking about initramfs stuff that's not in a cpio? Apr 06 22:20:18 hello Tom Apr 06 22:20:28 And I have to admit I haven't poked the initramfs stuff too hard Apr 06 22:21:52 not yet, we are concentrated with 2.6.28/9 kernels and kexec problems... Apr 06 22:22:16 k Apr 06 22:22:30 * Jay7 is concentrated on merging cfgfiles branch into master kexecboot branch :) Apr 06 22:23:00 have you had to worry about embedding modules into it yet or no? Apr 06 22:23:12 I just finished doing a one off where we had to, so this is still in my mind Apr 06 22:24:59 03Andrea Adami  07org.openembedded.dev * r55dc760645 10openembedded.git/recipes/linux/linux-rp.inc: Apr 06 22:24:59 linux-rp.inc: respect do_rm_work task Apr 06 22:24:59 - wlan-ng headers were needed before 2.6.28 Apr 06 22:24:59 - tosa was needing these but now uses 2.6.29 Apr 06 22:24:59 - not needed anymore it seems Apr 06 22:32:02 Tartarus: I've read interesting bits here...http://www.itp.tuwien.ac.at/~mattems/blog/kernel Apr 06 22:32:22 "The MODULES=dep mode walks /sys/ for the needed drivers of the specific box." Apr 06 22:32:46 is not yet clear to me if you need mdev/udev for this Apr 06 22:33:11 I guess that gets around doing depmod Apr 06 22:33:23 That was the biggest hurdle I had Apr 06 22:33:36 ant__: don't forget that we doing cross-compilation Apr 06 22:33:38 Need a kernel image for making modules.* Apr 06 22:33:56 you can't be sure what devices will be on target Apr 06 23:04:23 hi woglinde Apr 06 23:04:38 I've now understood the code with the help of the mailing list... Apr 06 23:05:14 again I didn't pay enough attention to what was written(in the lwn article) despite of having read it several times Apr 06 23:05:33 hm? Apr 06 23:05:39 so what should I do? Apr 06 23:09:17 by the way is http://www.ethernut.de/en/documents/arm-inline-asm.html outdated? Apr 06 23:09:34 because the ldrex-strex part were written with only one ':' Apr 06 23:11:10 dont know Apr 06 23:11:44 ok thanks anyway and thanks a lot for the link Apr 06 23:12:56 I'll use the syntax of the howto Apr 06 23:15:17 jo crofton Apr 06 23:36:42 * Jay7 -> sleep Apr 07 00:13:36 hey denix, did you ever figure out that curl thing? I think I'm seeing it w/ tcp-wrappers too Apr 07 00:15:08 hmmmmm Apr 07 00:15:24 * kergoth plays around with a git tree containing a package's source and a bitbake recipe Apr 07 00:32:19 denix, how big is your BB_NUMBER_THREAD, btw? Apr 07 00:35:01 denix, nm **** ENDING LOGGING AT Tue Apr 07 02:59:58 2009