**** BEGIN LOGGING AT Thu Apr 02 02:59:57 2009 Apr 02 03:17:12 * kergoth mulls over the possibility of recording the ABI of a lib and using it to invalidate the stamps of packages that link against it when the abi changes Apr 02 05:54:44 kergoth: ping/pong :) Apr 02 06:39:26 03Koen Kooi  07org.openembedded.dev * rb6f983eff8 10openembedded.git/recipes/e17/ (places/fix-api.diff places_svn.bb): places: add patch to fix API change introduced by SRCREV bump Apr 02 06:39:58 Hi guys, Is anyone familiar with openmoko? Apr 02 06:41:03 how to rebuild the image after I modified the source code in workdir? Apr 02 06:47:05 which XSERVER i should use for an embedded device, xserver-kdrive-fb or xorg xserver? Apr 02 06:50:54 how to rebuild the image after I modified the source code in workdir? Apr 02 06:52:18 i always notice when i load the image, i notice that the screen keeps flashing with gpe-login screen Apr 02 06:53:44 vsr: I argue you are using a x86 target, an't you? Apr 02 06:55:02 mckoan: i use a embedded omap based device Apr 02 07:02:28 03Roger Monk  07stable/2009 * r46f6f7e514 10openembedded.git/recipes/dsplink/ti-codec-combos_3.16.bb: Apr 02 07:02:28 ti-codec-combos: Use DVSDK Codec Combos to be aligned with DVSDK Apr 02 07:02:28 Signed-off-by: Koen Kooi Apr 02 07:02:28 Acked-by: Philip Balister (philip@balister.org) Apr 02 07:07:55 vsr: BTW your problem may be a wrong X setting Apr 02 07:33:32 good morning Apr 02 07:35:15 good morning Apr 02 07:39:17 morning Apr 02 07:40:42 good morning all Apr 02 07:52:16 03Jeremy Lainé  07org.openembedded.dev * r8f61b0343b 10openembedded.git/recipes/qt4/ (4 files in 2 dirs): qt4-tools-native: set LFLAGS to pick up zlib from staging Apr 02 08:21:18 03Koen Kooi  07org.openembedded.dev * rc03b0e6641 10openembedded.git/recipes/gnome/cheese_2.26.0.bb: cheese: add rrecommends to gconf-elements Apr 02 08:21:19 03Koen Kooi  07org.openembedded.dev * reff1b8de89 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Apr 02 08:51:16 hi, I've edited base-files and base-passwd in my local overlay, but even when I do a complete rebuild, it keeps taking these packages from the org.openembedded.dev repo. Is there any way to solve this? Apr 02 08:51:42 I copied task-base and task-boot to my overlay as well, but this makes no difference Apr 02 08:58:39 Aegiron give your overlay higher package version number and look if it builds them otherwise your overlay stuff isnt selected Apr 02 09:03:38 good morning Apr 02 09:03:48 'morning Apr 02 09:07:47 woglinde: that did the trick, thanks Apr 02 09:08:03 03Jeremy Lainé  07org.openembedded.dev * r86c175c766 10openembedded.git/recipes/makedevs/makedevs_1.0.0.bb: makedevs: use LDFLAGS (fixes QA error) Apr 02 09:08:13 03Jeremy Lainé  07org.openembedded.dev * rf77b7530d3 10openembedded.git/: Merge branch 'org.openembedded.dev' of ssh://git@git.openembedded.net/openembedded into org.openembedded.dev Apr 02 09:16:26 03Jeremy Lainé  07org.openembedded.dev * rf36e79c8e3 10openembedded.git/recipes/procps/procps_3.2.7.bb: procps: use LDFLAGS (fixes QA error) Apr 02 09:24:28 florian: good morning Apr 02 09:24:40 hi pb_ Apr 02 09:25:21 03Jeremy Lainé  07org.openembedded.dev * rb0f170f9ce 10openembedded.git/recipes/wireless-tools/ (3 files in 2 dirs): wireless-tools: use LDFLAGS (fixes QA error) Apr 02 09:26:43 03Frans Meulenbroeks  07org.openembedded.dev * rc96829e859 10openembedded.git/recipes/linux/ (linux-omap-2.6.29/ehci.patch linux-omap_2.6.29.bb): linux-omap_2.6.29: added ehci patch Apr 02 09:26:53 03Frans Meulenbroeks  07org.openembedded.dev * r5cba36b4a8 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Apr 02 09:30:41 03Jeremy Lainé  07org.openembedded.dev * r3a485444f0 10openembedded.git/recipes/zeroconf/ (files/zeroconf-ldflags.patch zeroconf_0.9.bb): zeroconf: use LDFLAGS (fixes QA error) Apr 02 09:30:42 03Jeremy Lainé  07org.openembedded.dev * r974ce4401a 10openembedded.git/: Merge branch 'org.openembedded.dev' of ssh://git@git.openembedded.net/openembedded into org.openembedded.dev Apr 02 09:32:20 woglinde: changing the version number of base-passwd did work, but now it keeps giving errors when it wants to applay a patch Apr 02 09:32:56 which is weird, as I did not edit anything exept for the version number Apr 02 09:34:33 it fails on the no_bash patch, but in fact I expect them all to fail as no_bash is just the first patch in the row Apr 02 09:34:35 RP: ping? Apr 02 09:36:02 Aegiron you forgot to copy the patch to the overlay Apr 02 09:36:25 no, I copied the whole base-passwd folder Apr 02 09:36:33 everything is there Apr 02 09:37:26 jeremy_laine: hey, QA log battle today ! Apr 02 09:38:39 ant_work: indeed Apr 02 09:38:53 ant_work: am trying to cut down http://tinderbox.openembedded.net/packages/?qalog=1 Apr 02 09:39:31 ant_work: the next culprit is ncurses, for some reason LDFLAGS is set to empty, I was wondering if anybody knows why? Apr 02 09:39:42 woglinde: this is what it returns: http://pastebin.com/d549bb6a6 Apr 02 09:40:13 3 times in a row even Apr 02 09:40:56 jeremy_laine: I was about asking you to fix ncurses :-) Apr 02 09:41:09 morning Apr 02 09:41:14 morning Apr 02 09:43:46 hrw: hi your ADSL is back ? Apr 02 09:46:49 jeremy_laine ncurses no, but I saw the qa issuses too Apr 02 09:46:54 but nobody cared Apr 02 09:47:39 what about fixing these? Apr 02 09:47:43 :-) Apr 02 09:49:01 it appears that changing the version of base-passwd from 3.5.20 to 3.5.21 makes the patches invalid? Apr 02 09:49:17 Aegiron didnt you say you modified this file? Apr 02 09:49:29 than it is clear the patch dont apllies Apr 02 09:49:46 it did'nt modify the bb file itself Apr 02 09:49:48 mckoan: not adsl but cable and yes Apr 02 09:50:44 I only edited a patch to set a default root password, just the same as they do with openslug Apr 02 09:51:30 if I edit the original patch, everything works, but I'd like to have my own edits in the overlay, rather than in the working copy of the git Apr 02 09:52:18 hrw: good to know :-) Apr 02 09:52:45 but it appeared that changing the version number does something weird Apr 02 09:53:26 therefore, I adjusted the version number of the original unmodified base-passwd, and that makes the patches invalid as well Apr 02 09:53:29 Aegiron: what for change version? Apr 02 09:53:55 just changing the filename frombase-passwd_3.5.20.bb to base-passwd_3.5.21.bb Apr 02 09:54:00 that's all Apr 02 09:56:10 Aegiron: no, I mean 'what for you updated to 3.5.21' Apr 02 09:58:05 hrw: if I copied the the base-passwd folder from the repo to my overlay, it still keeps choosing the repo's version Apr 02 09:58:22 tho my overlay works fine for every other package Apr 02 09:58:53 I have some problem with applying a patch. I have a package that fetches svn AND ftp thing, unpacks them into work dir where I have 1) linux-2.6.24 directory, 2) nxp directory. I'd like to apply the patch for nxp directory. I have a patch that starts filenames for patching with 'nxp/...' I specified the patch in my SRC_URI as "patch=1" but the patch fails. What's the problem? Apr 02 09:59:31 Aegiron: read about bitbake collections Apr 02 09:59:33 incrementing the version number solved the overlay problem, but now the patches do not apply anymore Apr 02 10:00:08 booxter: because S=linux-2.6.24 so it use quilt in that dir to apply patches Apr 02 10:00:41 hrw, how can I make 'nxp' the S dir? Apr 02 10:00:52 or just a dir for applying the patch? Apr 02 10:02:00 I would drop ';patch=1' and do patching in extra task Apr 02 10:03:30 hm, but why is linux-2.6.24 the default S thing? can I change the behavior somehow? Apr 02 10:04:20 built a gpe-image, when i click on the media player in gpe and add some mp3 files, when i try to play them it crashes Apr 02 10:04:31 does the gpe media player work oob Apr 02 10:04:37 you can do anything in recipe Apr 02 10:12:44 ok, back, someone pulled my eth cable from the switch Apr 02 10:13:15 hrw: I've read about collections, and I think my overlay is configure properly, as it chooses my modified qt4 recipe from the overlay instead of the one from the repo (which is desired) Apr 02 10:20:39 * Aegiron off to eat Apr 02 10:28:20 ant_work: ok, I've got it for ncurses.. it really uses an esoteric build system Apr 02 10:28:42 ant_work: it uses an awk script to generate the makefiles.. Apr 02 10:31:08 03Jeremy Lainé  07org.openembedded.dev * r9633d62015 10openembedded.git/recipes/ncurses/ (ncurses/use_ldflags.patch ncurses_5.4.bb): ncurses: add patch to use LDFLAGS (fixes QA error) Apr 02 11:04:17 thebohemian: hey! Apr 02 11:04:24 zecke: if you have questions shoot :) Apr 02 11:04:56 thebohemian: ^^^ wow talking to myself :} Apr 02 11:04:58 *healthy* Apr 02 11:05:53 zecke: hello, what about pushing buitbake patch for '-g' (attachment to bug 5002) Apr 02 11:11:31 hehe Apr 02 11:11:35 not now :) Apr 02 11:11:39 still busy with work Apr 02 11:12:14 np, at least you know :-) Apr 02 11:14:10 * ant_work hints again: teh patches from Roman I Khimov deserve attention Apr 02 11:14:30 * Roman_Khimov agrees ;) Apr 02 11:14:36 ;-) Apr 02 11:25:37 i have built a gpe-image from gpe-image.bb, when i click on the gpe-brower nothing happens, when i try to run it from the command prompt i get the message "GLib-ERROR **: The thread system is not yet initialized." Apr 02 11:25:43 what could be wrong? Apr 02 11:31:31 ant_work: is there a way to make package QA errors fatal (like staging QA errors) ? Apr 02 11:31:56 oh, yes Apr 02 11:32:11 iirc cyril_romain gave some thresold Apr 02 11:32:39 03Jeremy Lainé  07org.openembedded.dev * rc7a2bcb77e 10openembedded.git/recipes/libmikmod/ (libmikmod/ldflags.patch libmikmod_3.1.12.bb): libmikmod: add patch to use LDFLAGS (fixes QA error) Apr 02 11:34:30 jeremy_laine: def package_qa_make_fatal_error Apr 02 11:34:35 in insane.bbclass Apr 02 11:36:13 ant_work: ehm, what's class "9" ? Apr 02 11:38:13 jeremy_laine: "LDFLAGS ignored", Apr 02 11:38:29 pls fix comment Apr 02 11:39:25 ant_work: where can one find the definition of this error class? Apr 02 11:40:01 ant_work: ah, ok the list ERROR_NAMES Apr 02 11:40:16 :) Apr 02 11:41:30 03Jeremy Lainé  07org.openembedded.dev * r9b8323739f 10openembedded.git/classes/insane.bbclass: insane.bbclass: document error class "9" (LDFLAGS ignored) Apr 02 11:42:02 ant_work: in my local tree I have made error class "9" fatal Apr 02 11:42:30 you'll see some major brakages... Apr 02 11:42:45 there are several '9' Apr 02 11:42:57 in oestats builds Apr 02 11:43:23 but again, way to go! Apr 02 11:43:25 thx Apr 02 11:43:47 ant_work: as far as I am concerned, once openssl is fixed I should be able to build my usual set of packages OK Apr 02 11:45:02 denix: is the "cppstubs" a package that is local to you? Apr 02 11:45:02 jeremy_laine: my build would stop checking klibc... Apr 02 11:45:15 ant_work: hm, ok wil check that one Apr 02 11:45:32 jeremy_laine: wait, there is more Apr 02 11:45:48 possibly klibc will be updated by XorA / Khem Apr 02 11:45:52 soon Apr 02 11:50:31 ant_work: you don't seem to have any builds under your name in tinderbox for a while, still using it? Apr 02 11:53:22 03Jeremy Lainé  07org.openembedded.dev * r3696d4c13f 10openembedded.git/recipes/linux/linux-2.6.29/boc01/defconfig: linux-2.6.29: enable legacy wireless regulatory support for boc01 Apr 02 11:53:44 I'm not feeding 'cause I'm doing dirty things...I'd be ashamed :-) Apr 02 11:53:49 * XorA is being blamed somewhere Apr 02 11:54:06 jeremy_laine: bootloader work is almost done, I'll feed again asap Apr 02 11:54:33 * Jay7 can setup periodical builds for akita Apr 02 11:54:52 ademarez: I've just commited your patch for wireless on boc01 Apr 02 11:56:15 ant_work: ooh, klibc is a different QA error, I thought it was an LDFLAGS issue Apr 02 11:56:24 http://tinderbox.openembedded.net/public/logs/qa/452952.txt Apr 02 11:56:26 jeremy_laine: Thanks Jeremy ! It is currently working with in-kernel ath9k + AR9160 + hostapd-0.6.9. I am going to test now with ath5k Apr 02 11:58:10 jeremy_laine: http://lists.linuxtogo.org/pipermail/openembedded-devel/2009-April/009377.html Apr 02 12:01:45 ant_work: for the "non dev package contains .so", isn't this sufficient: Apr 02 12:01:46 -FILES_${PN} = "${base_libdir}/klibc*.so" Apr 02 12:01:48 +FILES_${PN} = "${base_libdir}/klibc-*.so" Apr 02 12:01:49 +FILES_${PN}-dev = "${base_libdir}/klibc.so" Apr 02 12:05:15 jeremy_laine: I'd say yes, but being this is a *libc... Apr 02 12:05:52 dunno if this was left purposedly Apr 02 12:06:22 ant_work: hm, actually we needed to add ${PN}-dev to PACKAGES Apr 02 12:06:35 ant_work: damn, the indentation of klibc.inc is really funky Apr 02 12:10:25 ant_work: ok, I'm not really competent for klibc, I just sent my patch to the OE mailing list Apr 02 12:25:17 mckoan: any luck with Qt 4.5.0? Apr 02 12:28:41 jeremy_laine: yes :-D Apr 02 12:29:13 mckoan: cool, which flavour, X11 or embedded? Apr 02 12:29:28 jeremy_laine: x11 Apr 02 12:29:46 BTW today I'm preparing a at91sam9g20ek patch for stable/2009 Apr 02 12:30:37 mckoan: There's a bunch of fixes to the audio stuff it might be nice to pick up. Apr 02 12:31:15 mckoan: OK. I'll keep an eye on tinderbox to see which versions you build Apr 02 12:31:17 broonie: audio stuff is too much work for this week, maybe next Apr 02 12:31:28 which recipe provides dmesg tool? Apr 02 12:31:56 jeremy_laine: actually I'm building all x11 related stuff without tinderbox because is experimental Apr 02 12:32:36 ah.. busybox Apr 02 12:32:52 * broonie has his at91sam9g20ek running just now :) Apr 02 12:32:55 mckoan: ah. I was hoping to collect enough statistic to see whether we can make 4.5.0 the default version Apr 02 12:34:21 jeremy_laine: IMHO it's worth it Apr 02 12:37:15 does gpe-mini-browser work for anyone here? Apr 02 12:37:31 when i start it, i see nothing happening Apr 02 12:38:47 vsr: try gpe-mini-browser2 Apr 02 12:39:40 florian: im using gpe-mini-browser2 Apr 02 12:39:52 florian: it does not work Apr 02 12:40:03 vsr: ah nasty. what hapens if you run it from a terminal? Apr 02 12:41:12 i set the display environment variable and i run, i get the message "GLib-ERROR **: The thread system is not yet initialized." Apr 02 12:42:26 florian: is there any package version mismatch? Apr 02 12:45:00 Sounds like a bug or change in WebKit - it used to work. Apr 02 12:45:08 philippe might know Apr 02 12:45:26 you need to init GThread in the main() Apr 02 12:45:30 (before gtk_init) Apr 02 12:45:36 iirc g_thread_init() Apr 02 12:45:59 I'll check it and check in a fix later today Apr 02 12:46:05 zecke thanks for the hint Apr 02 12:48:08 03Jeremy Lainé  07org.openembedded.dev * r1f3327dcd8 10openembedded.git/recipes/bzip2/ (bzip2-1.0.2/ldflags.patch bzip2_1.0.2.bb): bzip2: add patch to use LDFLAGS (fixes QA error) Apr 02 12:48:17 03Jeremy Lainé  07org.openembedded.dev * r792aafa604 10openembedded.git/: Merge branch 'org.openembedded.dev' of ssh://git@git.openembedded.net/openembedded into org.openembedded.dev Apr 02 12:51:12 /clear/clear Apr 02 12:52:31 i see similar issue with starling as well, i load the music file and click on the icon to play, it just crashes Apr 02 12:53:03 when i try on console i get segmentation fault Apr 02 12:53:30 http://pastebin.com/d40cf4a8c Apr 02 12:56:51 philippe: where should i call g_thread_init in the webkit sources? Apr 02 12:58:45 main.c in gpe-mini-browser2 sources i guess Apr 02 13:00:26 before gpe_init according to zecke Apr 02 13:00:33 vsr, indeed Apr 02 13:01:25 philippe: before gpe_application_init, there is no gpe_init Apr 02 13:01:43 then it is before gpe_application_init Apr 02 13:03:04 * mckoan wonders why there are some email in OE ML I never sent Apr 02 13:03:14 http://lists.linuxtogo.org/pipermail/openembedded-devel/2009-April/009453.html Apr 02 13:03:25 stolen identity ? Apr 02 13:04:46 seems that my commits goes to ML, is it possible ? Apr 02 13:05:59 vsr, I updated it in svn Apr 02 13:06:41 * philippe should check with a newer webkit... Apr 02 13:09:21 mckoan: sounds like a misconfiguration... but of what? Apr 02 13:10:04 florian: at tha time I did only git commit Apr 02 13:10:27 I thought I commited my local copy Apr 02 13:11:39 florian: I also have my standard settings in .gitconfig Apr 02 13:12:19 * florian looks around for someone who should gnow git better than he does Apr 02 13:13:01 hrm... using an OABI only kernel with EABI userlad is a bad idea ;) Apr 02 13:14:12 florian yes Apr 02 13:15:02 * florian is cleaning up all sorts of dusty projects at kc Apr 02 13:16:40 philippe: i made the changes manually and it works Apr 02 13:18:09 vsr, great! Apr 02 13:18:22 vsr, thanks for reporting the bug and confirming the fix works Apr 02 13:18:53 * philippe is reminded he should work a bit more again on gpe-mini-browser2 Apr 02 13:25:44 philippe: that' always a good idea Apr 02 13:25:55 * florian likes this tiny browser Apr 02 13:25:58 03Koen Kooi  07org.openembedded.dev * r817fa24308 10openembedded.git/recipes/gnome/gnome-keyring_2.26.0.bb: gnome-keyring: add missing dependency on gconf-dbus Apr 02 13:26:32 florian, idd. I need to get the settings working again, add the bookmarking and do something for downloading Apr 02 13:27:31 I was wondering just yet: when I reboot my embedded system, the screen shows some penguin and the text 'please wait: booting' Apr 02 13:27:32 * florian promises to work on improving the gpe images in oe as soon as he has a little bit of time Apr 02 13:27:47 philippe: i havent checkout your fix yet, i will check it out and tell you Apr 02 13:28:13 is it easily possible to add the current ip address to that text? Apr 02 13:28:22 philippe, florian: there are a lots of things that dont work like media player, file browser Apr 02 13:28:59 media player segmentation faults: http://pastebin.com/d40cf4a8c Apr 02 13:29:07 vsr: which image is that? Apr 02 13:29:55 gpe-image.bb Apr 02 13:30:03 Aegiron, I'm not sure you have an IP address to give at that point. Apr 02 13:30:07 angstrom distro Apr 02 13:30:13 omap3evm machine type Apr 02 13:32:39 florian: in file browser, it does not respond to mouse clicks, when you point to a directory. you need to use the keyboard and click on enter Apr 02 13:52:13 03Jeremy Lainé  07org.openembedded.dev * r2b85d56482 10openembedded.git/recipes/openssl/ (openssl-0.9.8j/oe-ldflags.patch openssl_0.9.8j.bb): openssl-0.9.8j: add patch to use LDFLAGS (fixes QA error) Apr 02 13:52:16 03Jeremy Lainé  07org.openembedded.dev * r111de86113 10openembedded.git/: Merge branch 'org.openembedded.dev' of ssh://git@git.openembedded.net/openembedded into org.openembedded.dev Apr 02 13:57:34 03Jeremy Lainé  07org.openembedded.dev * refbc7fe76c 10openembedded.git/recipes/libpcap/ (libpcap-1.0.0/ldflags.patch libpcap_1.0.0.bb): libpcap: add patch to use LDFLAGS (fixes QA error) Apr 02 14:04:10 ant_work: OK, I think that does it, I can now build all the packages we use without any QA errors :) Apr 02 14:04:31 jeremy_laine: many thx. I hate QA logs ! Apr 02 14:05:13 jeremy yes nice job Apr 02 14:06:27 jeremy_laine: wrt QA, see what Gentoo emits: Apr 02 14:06:31 * QA Notice: Package has poor programming practices which may compile Apr 02 14:06:33 * fine but exhibit random runtime failures. Apr 02 14:06:37 http://www.datenex.de/blog/archives/10-Fun-with-the-Beagle-Board,-part-1.html Apr 02 14:12:10 does bitbake have support for CMake? Apr 02 14:12:37 yes Apr 02 14:12:39 florian: do you recommend me using gpe-2.8? Apr 02 14:12:45 broonie: are you running linux from Dataflash or NAND ? Apr 02 14:14:52 Dataflash IIRC, though when I get round to it I'm going to switch to netboot. Apr 02 14:22:21 vsr: yes.. or whatever got updated in oe Apr 02 14:23:16 Meh. How do you get bitbake to rebuild a pacakge again? Apr 02 14:23:48 bitbake -c rebuild package-name Apr 02 14:23:51 broonie: -c rebuild but that may currently be broken Apr 02 14:23:58 broonie: if you use svn bitbake Apr 02 14:25:27 broonie: in fact seems that NAND support is not tested Apr 02 14:26:00 broonie: -cclean;then build Apr 02 14:26:17 florian: require conf/distro/include/preferred-gpe-versions-2.8.inc is not included in angstrom-2008.1.conf, so does that mean it is compiling the old gpe? Apr 02 14:26:47 vsr: no, it builds 2.8+ versions Apr 02 14:26:48 Yeah, rebuild is broken but clean worked. Apr 02 14:26:58 Thanks. Apr 02 14:27:49 thanks Apr 02 14:30:28 broonie: could you send me your .config ? Apr 02 14:31:35 http://opensource.wolfsonmicro.com/~broonie/sam9g20 Apr 02 14:49:09 03Koen Kooi  07stable/2009 * r9a423d2238 10openembedded.git/ (4 files in 3 dirs): Apr 02 14:49:09 webkit-gtk: update to 1.1.4 release to get some more bugfixes Apr 02 14:49:09 Signed-off-by: Koen Kooi Apr 02 14:49:09 03Koen Kooi  07stable/2009 * r9f55682d70 10openembedded.git/ (conf/distro/angstrom-2008.1.conf recipes/tasks/task-base.bb): Apr 02 14:49:09 angstrom 2009.X: switch to bluez4 Apr 02 14:49:13 Signed-off-by: Koen Kooi Apr 02 14:52:34 * hrw wants --verbose-wget in opkg.. Apr 02 15:02:07 morning Apr 02 15:02:50 denix: pong Apr 02 15:10:14 03Jeremy Lainé  07org.openembedded.dev * r7596f86858 10openembedded.git/recipes/perl/ (perl-5.8.8/shared-ldflags.patch perl_5.8.8.bb): perl-5.8.8: use LDFLAGS for shared libraries (fixes QA error) Apr 02 15:17:26 what's the difference between DISTRO = "angstrom-2008.1" in my local.conf, and DISTRO_VERSION = "2009.X-test-20090402" which bitbake outputs? Apr 02 15:17:43 shouldn't be DISTRO_VERSION something like 2008.... ?? Apr 02 15:20:39 angstrom-2008.1 is a file Apr 02 15:20:49 2009.X-test-20090402 is a string generated in that file Apr 02 15:21:24 XorA: ok, thanks! Apr 02 15:40:00 I'd like to set up an HTTP server to fetch from, but I'm confused between SRC_TARBALL_STASH, PREMIRRORS, SOURCE_MIRROR_URL, etc. any pointers? Apr 02 15:49:25 morning all Apr 02 15:49:48 kergoth: how are you? Apr 02 15:50:17 pretty good, you? Apr 02 15:52:57 hi kergoth Apr 02 15:53:01 hey pb_ Apr 02 15:53:27 kergoth: ok in general, but tired debugging one nasty issue... Apr 02 15:53:30 Is the OE list moderated for non-subscribers? Apr 02 15:53:46 denix: fun Apr 02 15:53:53 denix: I'm in autotools hell at the moment Apr 02 15:53:59 kergoth: yeah. Apr 02 15:54:06 yeah Apr 02 15:54:38 broonie: non-subscribers cannot post to prevent spam - normal practice these days Apr 02 15:55:04 Bah, humbug. Apr 02 15:55:12 It doesn't bounce back which isn't helpful. Apr 02 15:55:39 yeah, that is a bit annoying. Apr 02 15:56:04 look on the bright side though: for every message that does get through, you get an extra bonus copy at no charge Apr 02 15:56:20 kergoth: initially I was suspecting one of your core commits 2 weeks ago... but now I'm not so sure Apr 02 15:57:20 denix: what problem are you hitting? Apr 02 15:58:03 i don't think i pushed anything that should affect autotools much, my invasive changes to autotools.bbclass are on my autotools branch Apr 02 15:59:56 kergoth: I'm building an image and then a meta-toolchain. it fails to _package_ a wrong version of curl package! basically, curl is 7.19, while curl-native and curl-sdk are 7.18. when it builds curl-sdk, it builds 7.18, but it wants to package 7.19 for some reason... Apr 02 16:00:54 Ah, no. It does send a bounce back, it's just exrremely slow bout it. Apr 02 16:01:01 it used to work fine 2 weeks ago Apr 02 16:02:04 denix: odd. you could do a git bisect. Apr 02 16:02:17 kergoth: I traced it to package.bbclass and package_ipk.bbclass - there PV resolves to 7.19, while processing 7.18 recipe... Apr 02 16:02:43 git bisect start; git bisect bad; (since this version is broken), then go back to a working version a few weeks ago, git bisect good;, and it'll help you do a binary search to find the commit that broke it Apr 02 16:03:08 you can specify good and bad commit IDs on the command line too. Apr 02 16:03:20 (original command line, that is) Apr 02 16:03:25 true, but i always forget the syntax :) Apr 02 16:03:41 git bisect start bad good Apr 02 16:03:52 * kergoth nods Apr 02 16:04:09 git bisect run Apr 02 16:04:24 kergoth: yeah, I was doing it manually (still learning git) - I checked out a rev 2 weeks ago, which used to work. but I'm not sure it works 100% of the time now Apr 02 16:04:31 i hate the overhead in the build cycles when testing changes to core classes. the builds just take so damn long since you have to wipe and start from scratch each time Apr 02 16:04:33 git bisect dwim-and-fix-the-bug-too-kthxbye :) Apr 02 16:04:41 root Apr 02 16:04:44 bisect is awesome Apr 02 16:04:52 such a time saver Apr 02 16:12:17 i did a git pull and tried to build a gpe-image, i get the error "ImportError: No module named _md5" while packaging in do_rootfs Apr 02 16:12:22 is oe broken atm? Apr 02 16:13:12 please see: http://pastebin.ca/1380073 Apr 02 16:14:49 vsr: that's due to the openssl version change Apr 02 16:15:23 yeah it rebuilt openssl Apr 02 16:15:41 what version number should i set? Apr 02 16:15:41 vsr: kill your python-native and rebuild it Apr 02 16:15:56 vsr: after that it will work Apr 02 16:16:21 I am wondering whether we shouldn't bump the PR on python-native Apr 02 16:16:27 http://www.flickr.com/photos/hrwandil/3407272062/ Apr 02 16:16:43 jeremy_laine: python running on my ubuntu? Apr 02 16:16:58 vsr: no the python-native package in OE Apr 02 16:17:12 vsr: bitbake -c clean python-native Apr 02 16:18:08 hrw: My case is smaller and I even habe the keyboard inside ;) Apr 02 16:18:26 hrw: But yes this looks pretty good. Apr 02 16:18:53 * florian still searching for a small case for the board only - without paying USD 42 for shipping Apr 02 16:20:10 florian: gpe media player is broken, any idea what could be wrong? http://pastebin.com/d40cf4a8c Apr 02 16:21:13 hrw: what's your plan for this "box"? :) Apr 02 16:21:59 vsr: Might be a missing gstreamer plugin. Apr 02 16:22:15 denix: I do not want to have bb floating over the table with all cables hanging from it. Apr 02 16:22:28 vsr: If you find out which package is missing I can fix the bb Apr 02 16:22:34 florian: i used gpe-image, is there anything i need to add Apr 02 16:22:46 florian: sorry Apr 02 16:22:57 florian: is there a way to find out? Apr 02 16:23:04 denix: inside will be bb, usbhub, usbethernet, serial port, lot of space, one power supply for everything (I have 2 5V/3.5A ones) Apr 02 16:23:21 denix: who knows.. maybe one day I will add second BB into that just for fun Apr 02 16:23:32 hrw: :) Apr 02 16:24:02 vsr: just install some gstreamer module packages. I suspect that there were more generated than ending up in the image. Apr 02 16:24:15 has anyone experienced problems trying to generate an ext2 / ext3 rootfs when building within a chroot? Apr 02 16:24:16 it is a good idea to put everything in a box with only necessary ports out... Apr 02 16:24:59 denix: too bad that it is B7 - lack of normal usb port ;( Apr 02 16:25:11 I got the following error: ext2fs_check_if_mount: No such file or directory while determining whether .rootfs.ext3 is mounted. Apr 02 16:25:20 florian: I can't believe you can't find a cheap plastic case Apr 02 16:25:59 florian: i will debug and let you know, i use alsa, gstreamer must have a plugin for alsa Apr 02 16:25:59 hrw: you are not ordering C2 due to postal/customs expenses? Apr 02 16:26:05 * mckoan bought 2 SheevaPlug Apr 02 16:26:25 * kergoth wonders when the sheevaplugs will ship Apr 02 16:26:27 mckoan: I ordered 3 of them ;D Apr 02 16:26:48 denix: first need to buy lcd monitor to check B7 output ;D Apr 02 16:26:50 hrw: SheevaPLug++ Apr 02 16:27:01 * kergoth only ordered one Apr 02 16:28:09 I didn't get any :) Apr 02 16:28:22 denix: you have TI toys ;D Apr 02 16:28:29 mckoan: well... yes, I didn't spend much time in searching. There should be quite some standard industry cases which will fit pretty well. Apr 02 16:28:50 * denix has too much hardware for DIY projects, but no time to do anything useful with it Apr 02 16:29:12 i might ahave to order a second though. I need a device to test OpenEmbedded builds, but I'm also thinking about using such a device to run my torrent client so i dont need to leave a desktop running 24/7 Apr 02 16:29:14 * kergoth thinks Apr 02 16:29:26 vsr: if you haave it might make use of esd/pulseaudio/oss emulation Apr 02 16:29:50 kergoth: what about storage? NAS? Apr 02 16:30:02 yea, i use a netgear readynas nv+ Apr 02 16:30:15 tried running the torrents off of it, but it pushed it a bit too hard :) Apr 02 16:31:16 I looked for NAS few years ago - they were either very expensive or not scalable... Apr 02 16:31:22 hrw: any luck trying the QA fix patches? Apr 02 16:31:43 heh, my dad found a deal on this one on ebay, or i doubt I'd have one either Apr 02 16:31:49 the drobo looks nice, but not cheap either Apr 02 16:31:50 jeremy_laine: acked all Apr 02 16:31:52 readynas nv+ was/is one of the best, but $600 just for the enclosure? Apr 02 16:32:02 yea its crazy Apr 02 16:32:14 I wish the sheevaplug had a wireless variant. Ethernet is... limiting. Apr 02 16:32:28 drobo is usb only, you need an extra device for eth connection Apr 02 16:32:52 I ended up building my own fileserver... Apr 02 16:32:54 true, and the device is still $500 msrp Apr 02 16:34:07 15 HDD slots... Apr 02 16:34:28 robtow: a powerline varient of sheeva would rock Apr 02 16:34:58 not all of those 15 slots are populated yet though... :) Apr 02 16:35:53 XorA - i'd prefer wireless, for interoperability with things like my iPod Touch and iPhone. Apr 02 16:36:07 where can I find information on how to bitbake selfmade qt embedded projects? Apr 02 16:38:15 hrw: speaking of toys (boards/devices) - like you mentioned in your blog, I too haven't written a real big app in years... Apr 02 16:40:52 03Koen Kooi  07org.openembedded.dev * r320a5e3633 10openembedded.git/recipes/gimp/gimp_2.6.3.bb: gimp 2.6.3: explicitly depend on webkit-gtk since it gets dragged in if present Apr 02 16:40:52 03Koen Kooi  07org.openembedded.dev * rda07a7a0c3 10openembedded.git/ (3 files in 2 dirs): Apr 02 16:40:52 gnome-panel: update to 2.26 Apr 02 16:40:52 libgweather: update to 2.26 Apr 02 16:44:00 * hrw|gone -> off for ~1-1.5h Apr 02 16:44:08 florian: while i debug that, is there an alternative gui video/audio player i can use for gpe? Apr 02 16:44:58 I've set SOURCE_MIRROR_URL and INHERIT += "own-mirrors" in my local.conf but it still fetches from the angstrom mirror, any idea? Apr 02 16:45:08 vsr: vlc-gpe, but I havent tried this for some time Apr 02 16:46:25 if my gstreamer issues are solved gpe media player is good enough, do you think? which handheld device gives gpe media player? Apr 02 16:47:56 i want gpe, i can only use gpe-xxx packages, or i can use any package in oe like midori/mplayer/gmplayer? Apr 02 16:48:47 vsr: you can use any package of course. Apr 02 16:49:28 florian: thanks, i will get back tommorrow, i got to go now, its 11pm here Apr 02 16:54:29 gb Apr 02 17:13:36 morning all Apr 02 17:16:29 kergoth: http://pastebin.com/m36c05ba Apr 02 17:16:46 strange Apr 02 17:16:56 * kergoth hasn't really looked at the sdk bits Apr 02 17:17:36 kergoth: not sure it's sdk specific Apr 02 17:18:19 kergoth: it's coming from package_ipk.bbclass line 161 - it calls bb.utils.lockfile() Apr 02 17:18:36 i honestly wouldnt know where to begin in diagnosing that, particularly if I'm unable to reproduce. i'll try building curl-sdk here Apr 02 17:18:59 for some reason PV resolves to 7.19 there... Apr 02 17:19:19 the thing is building just curl-sdk works fine Apr 02 17:21:03 since 7.19 was built for target, it picks 7.19 either from deploy or from packaged-staging. if I clean tmp/deploy and build just curl-sdk, it works. Apr 02 17:21:36 its strange that there'd be any overlap at all, any metadata it could be pulling from should be separated Apr 02 17:23:17 it's hard to reproduce - it only happens when I build sdk after I've built an image... Apr 02 17:23:38 let me try "bitbake curl curl-sdk" and see if it works Apr 02 17:24:44 * kergoth nods Apr 02 17:32:00 I am building a bb file for a small test app, but what do I need to 'inherit' in order to het it to build? Apr 02 17:32:41 I tried inherit qmake, but that wants to build qt3.3.5 Apr 02 17:32:52 and I am using qt4.5 Apr 02 17:35:08 "a bb file for a small test app"? I'd suggest telling us what buildsystem it uses, unless you expect us to pick a class for you to inherit at random Apr 02 17:36:43 ehm, yes, that was a little too shy on information :) I just figured out i needed to inherit qt4e Apr 02 17:37:10 ah, right then :) Apr 02 17:38:36 Am I correct when I think there is not much documentation regarding making .bb files, other than the collection of ready-made bb files in the OE repo? Apr 02 17:38:59 or am I just searching the wrong places? :) Apr 02 17:39:19 that'd be correct. i think there's a small amount of info on the wiki, but its more style stuff than content Apr 02 17:40:18 ok, then there's no need to look for that anymore :) Apr 02 17:40:47 although, perhaps there's info in the OpenEmbedded manual? i haven't looked at that lately Apr 02 17:43:07 Aegiron: are you using Qt/X11 or Qt/embedded? Apr 02 17:43:32 Aegiron: ok you found your answer obviously Apr 02 17:43:37 any cpp experts I am looking for a stringification macro which do not lose the formatting of the string Apr 02 17:43:56 03Koen Kooi  07org.openembedded.dev * r4af6bdcda0 10openembedded.git/ (2 files in 2 dirs): gnome-python-desktop: update to 2.26.0 Apr 02 17:46:30 hi jeremy_laine, khem Apr 02 17:46:43 like2wise: hi Apr 02 17:46:47 "stringification"? Apr 02 17:47:06 jeremy_laine: 83xx product taking shape? Apr 02 17:47:24 like2wise: yep nicely thank you :) Apr 02 17:47:51 jeremy_laine: xx13 or xx23? Apr 02 17:48:07 like2wise: 8313 Apr 02 17:48:31 jeremy_laine: ah cool, our first proto (8313) should be there next week Apr 02 17:49:19 Maybe I will bother you with some questions on the clock configuration (33 / 66 MHz) etc... Apr 02 17:49:21 kergoth: ok, it fail a clean build of "bitbake curl curl-sdk" - that's a faster way to reproduce it... Apr 02 17:49:30 cool Apr 02 17:49:36 well, not cool, but helpful :P Apr 02 17:49:55 let me try disabling packaged staging... Apr 02 17:51:01 ~seen effem Apr 02 17:51:05 effem was last seen on IRC in channel #oe, 1d 2h 30m 45s ago, saying: 'works now, apparently took a long time to get going'. Apr 02 18:02:42 like2wise: stringification I mean is that I want to construct a string from the parameters Apr 02 18:05:38 kergoth: nope, it's not packaged staging - still fails w/o it... Apr 02 18:05:46 interesting Apr 02 18:06:48 indeed Apr 02 18:11:14 morning Apr 02 18:11:17 denix: what is the error Apr 02 18:11:38 khem: http://pastebin.com/m36c05ba Apr 02 18:12:56 03Jeremy Lainé  07stable/2009 * r593caa19b9 10openembedded.git/recipes/zeroconf/ (files/zeroconf-ldflags.patch zeroconf_0.9.bb): Apr 02 18:12:56 zeroconf: use LDFLAGS (fixes QA error) Apr 02 18:12:56 Signed-off-by: Koen Kooi Apr 02 18:12:56 Acked-by: Marcin Juszkiewicz Apr 02 18:12:59 03Jeremy Lainé  07stable/2009 * ra67cad751a 10openembedded.git/recipes/ncurses/ (ncurses/use_ldflags.patch ncurses_5.4.bb): Apr 02 18:13:00 ncurses: add patch to use LDFLAGS (fixes QA error) Apr 02 18:13:02 Signed-off-by: Koen Kooi Apr 02 18:13:04 Acked-by: Marcin Juszkiewicz Apr 02 18:13:06 03Jeremy Lainé  07stable/2009 * rb24794ae92 10openembedded.git/recipes/wireless-tools/ (3 files in 2 dirs): Apr 02 18:13:09 wireless-tools: use LDFLAGS (fixes QA error) Apr 02 18:13:11 Signed-off-by: Koen Kooi Apr 02 18:13:13 Acked-by: Marcin Juszkiewicz Apr 02 18:13:15 03Jeremy Lainé  07stable/2009 * ra67276be7c 10openembedded.git/recipes/procps/procps_3.2.7.bb: Apr 02 18:13:18 procps: use LDFLAGS (fixes QA error) Apr 02 18:13:20 Signed-off-by: Koen Kooi Apr 02 18:13:22 Acked-by: Marcin Juszkiewicz Apr 02 18:13:24 03Jeremy Lainé  07stable/2009 * r0097af4c8b 10openembedded.git/recipes/makedevs/makedevs_1.0.0.bb: Apr 02 18:13:32 makedevs: use LDFLAGS (fixes QA error) Apr 02 18:13:33 Signed-off-by: Koen Kooi Apr 02 18:13:36 denix: I ran into it a few days ago... Apr 02 18:13:38 Acked-by: Marcin Juszkiewicz Apr 02 18:13:40 03Koen Kooi  07stable/2009 * rdae0d6d32f 10openembedded.git/recipes/gnome/gnome-keyring_2.26.0.bb: Apr 02 18:13:43 mkdir curl-sdk curl-sdk-certs curl-sdk-dbg curl-sdk-dev curl-sdk-doc curl-sdk-locale libcurl libcurl-dev libcurl-doc Apr 02 18:13:45 gnome-keyring: add missing dependency on gconf-dbus Apr 02 18:13:47 Signed-off-by: Koen Kooi Apr 02 18:13:49 Acked-by: Marcin Juszkiewicz Apr 02 18:13:50 are all the stable patches already in dev? Apr 02 18:13:51 03Jeremy Lainé  07stable/2009 * r51568eed6f 10openembedded.git/recipes/bzip2/ (bzip2-1.0.2/ldflags.patch bzip2_1.0.2.bb): Apr 02 18:13:54 bzip2: add patch to use LDFLAGS (fixes QA error) Apr 02 18:13:58 Signed-off-by: Koen Kooi Apr 02 18:14:02 Acked-by: Marcin Juszkiewicz Apr 02 18:14:04 03Jeremy Lainé  07stable/2009 * r6716cb0fc0 10openembedded.git/recipes/openssl/ (openssl-0.9.8j/oe-ldflags.patch openssl_0.9.8j.bb): Apr 02 18:14:05 in tmp/work/x86_64-armv5te-sdk-angstrom-linux-gnueabi/curl-7.19.0-r1/install Apr 02 18:14:07 openssl-0.9.8j: add patch to use LDFLAGS (fixes QA error) Apr 02 18:14:09 Signed-off-by: Koen Kooi Apr 02 18:14:11 Acked-by: Marcin Juszkiewicz Apr 02 18:14:11 khem: do_package_ipk() gets the wrong version - 7.19 while processing 7.18 recipe Apr 02 18:14:12 seems to solve the problem Apr 02 18:14:15 03Jeremy Lainé  07stable/2009 * re43e900b9f 10openembedded.git/recipes/libmikmod/ (libmikmod/ldflags.patch libmikmod_3.1.12.bb): Apr 02 18:14:18 libmikmod: add patch to use LDFLAGS (fixes QA error) Apr 02 18:14:20 Signed-off-by: Koen Kooi Apr 02 18:14:22 Acked-by: Marcin Juszkiewicz Apr 02 18:14:24 03Jeremy Lainé  07stable/2009 * r63b1f4d416 10openembedded.git/classes/insane.bbclass: Apr 02 18:14:31 insane.bbclass: document error class "9" (LDFLAGS ignored) Apr 02 18:14:33 (2 lines omitted) Apr 02 18:14:51 f3nix: ok, it's not just me! Apr 02 18:14:53 i686-armv5te-sdk-none-linux-gnueabi interesting. Do you have this dir Apr 02 18:15:11 khem: yes Apr 02 18:15:25 f3nix: that's a workaround... Apr 02 18:16:25 khem: yep. I have "none" TARGET_VENDOR, f3nix has "angstrom" Apr 02 18:17:30 khem: do_package_ipk() gets the wrong version - 7.19 while processing 7.18 recipe Apr 02 18:18:31 this one is wrong - .../i686-armv5te-sdk-none-linux-gnueabi/curl-7.19.0-r1/install Apr 02 18:19:00 should be .../i686-armv5te-sdk-none-linux-gnueabi/curl-7.18.2-r0/install Apr 02 18:19:39 another workaround would be to sync-up native, sdk and target versions of curl Apr 02 18:20:14 as currently target is 7.19, while native and sdk are still at 7.18 Apr 02 18:20:58 reproducible by clean (may not be required) build of "bitbake curl curl-sdk" Apr 02 18:21:38 denix: does it build when you bitbake them separately Apr 02 18:21:50 bitbake curl;bitbake curl-sdk ? Apr 02 18:22:33 fails. it only works if curl-sdk is built when target curl was not built Apr 02 18:23:03 hmmm Apr 02 18:23:13 otherwise if target curl was built before, curl-sdk fails Apr 02 18:25:18 and I also tried disabling packaged-staging w/o success Apr 02 18:29:36 denix: I can reproduce it Apr 02 18:36:14 khem: I'm trying the same on i686-generic now... Apr 02 18:40:03 re Apr 02 18:40:45 like2wise: are there any cheap powerpc devboards? Apr 02 18:41:27 hmm Apr 02 18:42:26 there's always 2nd-hand macs Apr 02 18:43:24 * khem have a old g4 at work for abusing Apr 02 18:43:57 Tartarus: which are big Apr 02 18:44:02 denix: I think I have a hint what could be going wrong Apr 02 18:44:09 There's laptops too :) Apr 02 18:44:14 but yeah Apr 02 18:44:29 iirc tho, most of the refboards are regular PC mobo sized Apr 02 18:44:39 khem: good! keep me posted :) Apr 02 18:44:41 or old and slow Apr 02 18:45:16 lemme go poke some people quick tho Apr 02 18:57:23 hrw, http://www.micetek.com/englishweb/product/new/PQIIproDevelopmentProducts/MPC8349EmITX.htm Apr 02 18:57:29 Those are suppoed to be around $500USD Apr 02 18:57:44 Even at single unit prices Apr 02 18:57:56 03Koen Kooi  07stable/2009 * r6fa4b963cf 10openembedded.git/recipes/e17/ (places/fix-api.diff places_svn.bb): Apr 02 18:57:56 places: add patch to fix API change introduced by SRCREV bump Apr 02 18:57:56 Signed-off-by: Koen Kooi Apr 02 18:57:56 Acked-by: Marcin Juszkiewicz Apr 02 18:58:34 Oh ya, and there's always the efika Apr 02 18:59:08 hrw, Tartarus: mpc8313e and mpc8315e ($300) Apr 02 18:59:11 Tartarus: old g4 you can buy for 200 on ebay :) Apr 02 18:59:47 like2wise, got any urls for that? Apr 02 19:00:17 Tartarus, hrw: hold on Apr 02 19:01:12 The 8315e is nicer, has PCI Express. This is the 8313: http://www.micetek.com/englishweb/product/new/PQIIproDevelopmentProducts/8313E-RDB.htm Apr 02 19:01:30 http://www.micetek.com/englishweb/product/new/PQIIproDevelopmentProducts/MPC8315E-RDB.htm Apr 02 19:01:32 ^^ 8315 Apr 02 19:16:15 RP: hello? Apr 02 19:16:25 jeremy_laine: catch him on Jabber Apr 02 19:16:43 hrw: ehm, don't have his JID Apr 02 19:17:06 richard@openedhand.com Apr 02 19:19:59 hrw: he's not responding for now Apr 02 19:20:06 normal Apr 02 19:23:09 khem: hmm, i686-generic worked, because it built curl-sdk first and then curl target... :) I wonder why. going to try building them separately to force the order... Apr 02 19:24:43 denix: its something in sdk.bbclass Apr 02 19:26:43 denix: probably PKG_CONFIG_SYSROOT_DIR Apr 02 19:31:58 03Jeremy Lainé  07org.openembedded.dev * r64f58bd748 10openembedded.git/recipes/linux/ (linux-rp-2.6.26/defconfig-qemux86 linux-rp_2.6.26.bb): Apr 02 19:31:58 linux-rp-2.6.26: update qemux86 defconfig Apr 02 19:31:58 * add support for RTC Apr 02 19:31:58 * add some netfilter modules which were enabled in 2.6.24 Apr 02 19:31:58 * refresh defconfig with "make menuconfig" Apr 02 19:32:03 03Jeremy Lainé  07org.openembedded.dev * rbc4bf696b5 10openembedded.git/: Merge branch 'org.openembedded.dev' of ssh://git@git.openembedded.net/openembedded into org.openembedded.dev Apr 02 19:33:35 jeremy_laine: I would like to move qemuarm/qemux86 to linux_2.6.29 Apr 02 19:34:05 hrw: sounds cool, in the meantime I wanted something fresher than 2.6.24 that still worked Apr 02 19:34:26 hrw: if you wouldn't mind basing the defconfig off the 2.6.26 one, that'll save me some time :) Apr 02 19:38:13 guys: simple working audio/video player in OE? not using gstreamer Apr 02 19:38:27 mplayer? Apr 02 19:38:39 and with gui ;D Apr 02 19:38:44 gmplayer? :) Apr 02 19:39:46 but I'm unsure that here is recipe for it Apr 02 19:41:13 vlc? Apr 02 19:51:55 * mwester adds koen to his spam filter's "REJECT" rules, and makes his life much more pleasant. Apr 02 19:53:07 * kergoth 's still of the school that .dev is like a trunk branch in an scm, and that breakage *happens*. nobody intends it, but it's going to happen, so people just have to deal with that if they intend to track that branch Apr 02 19:54:47 * like2wise seconds that opinion Apr 02 19:56:26 mwester: maybe add STABLE into that rule? Apr 02 19:57:01 mwester: hmm? becasue of stable patches? Apr 02 19:57:08 kergoth: +1 Apr 02 19:57:10 stable sounds like a Good Thing, but i don't think we should take its expectations and try to apply them to .dev Apr 02 19:57:37 kergoth: we cant even apply those to .dev Apr 02 19:58:12 well, from the emails, it sounds like people are expecting .dev to never break, and that's just ridiculous Apr 02 19:58:27 STABLE is already in mine! :) Apr 02 19:58:44 maybe this should go to a separate list... Apr 02 19:59:02 timtimred: we had ML for stable - it lacked devs Apr 02 19:59:41 we killed that ML due to that Apr 02 20:00:15 :) ahh well... Apr 02 20:00:38 * kergoth still thinks some sort of split between individually maintained stuff and group maintained stuff would do more for build stability than a stable branch, and would require less upkeep Apr 02 20:01:02 i still think people should fork off dev if they want to stabilise their stuff :) Apr 02 20:01:43 that too. that's why thats what most companies do, its mostly just the hobbiests that whine when dev breaks Apr 02 20:02:01 hobbyists? thats better Apr 02 20:02:03 * kergoth can't spell today Apr 02 20:02:38 bye all Apr 02 20:02:50 hrw|gone: bye Apr 02 20:03:01 well, hobbyists often can't maintain a fork Apr 02 20:03:30 kergoth: correct, it's normal to maintain a snapshot of .dev for stability Apr 02 20:04:19 valholl: well, there are distros that have done so in the past, nslu2 folks for example, but its generally community driven rather than individual hobbyists Apr 02 20:04:26 it's when you try to update your snapshot to the latest .dev due to some new features/fixes - the fun begins :) Apr 02 20:04:47 cherry-pick is your friend :) Apr 02 20:05:14 hrw|gone: (sorry, on a phone call) -- no, nothing to do with stable branch, although I do fear that working with Koen on the stable branch is just not going to be practical, at least for me. It has nothing to do with the branch, the quality of the work, or anything of the sort, it's just a long-time personality conflict that causes my blood pressure to rise exponentially with each silly email exhange. Apr 02 20:05:17 yes, communities probably can Apr 02 20:05:31 kergoth: yeah, but it maybe too time-consuming for a single person to go through every single commit... :-P Apr 02 20:05:54 And I agree - a fork is impractical to maintain. Apr 02 20:06:04 its true. plus git isn't as smart about cherry picks as I'd like Apr 02 20:06:05 My position on this is simple. Apr 02 20:06:22 if a commit is changed when cherry picked, git-cherry no longer recognizes that that commit was cherry picked over Apr 02 20:06:26 I do not expect dev to always work; I expect a good-faith effort to keep it building. Apr 02 20:06:31 makes it harder to maintain a long term branch Apr 02 20:06:32 denix, exactly :) Apr 02 20:07:01 And I have perceived a lack of consistency among expectations in the OE community. Apr 02 20:07:03 SPecifically Apr 02 20:07:12 mwester: nicely put about good-faith! Apr 02 20:07:18 The core devs can commit anything anytime and if it breaks something, then "deal with it" Apr 02 20:07:27 The others of us, we must commit perfection. Apr 02 20:07:47 Otherwise, the commtis are in danger of being reverted in some cases, and often we are told to go work on a branch. Apr 02 20:08:04 in .dev one has to expect a certain degree of accidental breakage.... but i do agree :) Apr 02 20:08:15 :) Apr 02 20:08:16 And now I am told that I'm not to commit SlugOS-specific fixes in order to avoid angering the core devs, either. Apr 02 20:09:07 So, I can go off an work with buildroot or another tool. Apr 02 20:09:14 Or I can work with private overlays. Apr 02 20:09:15 I've always just done builds off of dev and start fixing any bugs i run into.. i don't go track down who introduced those breaks and yell at them Apr 02 20:09:22 Or I can put Koen on ignore. Apr 02 20:09:27 I choose the latter. Apr 02 20:09:39 If he says something worthwhile, maybe someone will relay it. :-D Apr 02 20:14:02 khem: ok, if I force curl-sdk to build _AFTER_ curl target, it also fails on i686-generic, so it's not arch specific... Apr 02 20:18:00 03Jeremy Lainé  07org.openembedded.dev * r1ed4e64c7f 10openembedded.git/recipes/sqlite/ (sqlite-2.8.17/mainmk_build_dynamic.patch sqlite_2.8.17.bb): sqlite: use LDFLAGS (fixes QA error) Apr 02 20:20:10 03Frans Meulenbroeks  07org.openembedded.dev * r86576770c4 10openembedded.git/recipes/linux/ (linux-omap-2.6.29/beagleboard/defconfig linux-omap_2.6.29.bb): linux-omap_2.6.29: updated defconfig (more compatible with .28); moved srcrev to git head Apr 02 20:20:11 03Frans Meulenbroeks  07org.openembedded.dev * re93bd21c7a 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Apr 02 20:21:50 mwester: I shouldn't get too worked up, personality clashes will happen :) Apr 02 20:22:52 mwester: I'm personally quite happy you changed the openssl version, and the fallout was pretty easy to contain Apr 02 20:27:49 denix: comment out export PKG_CONFIG_SYSROOT_DIR in sdk.bbclass and retry the build Apr 02 20:28:34 khem: will do Apr 02 20:42:00 khem: dirty rebuild (w/o removing tmp) of curl-sdk still fails, I'll try a clean one... Apr 02 20:44:10 * kergoth is in autotools hell Apr 02 20:46:33 * eFfeM does not envy kergoth Apr 02 20:53:29 it's pity to see developers conflict :( Apr 02 21:01:38 denix, shouldn't bitbake -v -DDD show all of the bb files and such that get parsed, including stuff that's a 'require' ? Apr 02 21:06:33 kergoth: about openssl upgrade, it's pity OE lacks a script like 'revdep-rebuild' (Gentoo) Apr 02 21:10:16 btw I've read the md5 error is fixed by updating python-native.....which beautifully fails :-( Apr 02 21:10:28 | Python/graminit.c:1741: error: expected '=', ',', ';', 'asm' or '__attribute__' at end of input Apr 02 21:13:24 kergoth: at least 'equery depends openssl' as appetizer ... Apr 02 21:15:20 http://rafb.net/p/XpG9kj26.html Apr 02 21:23:27 ok, rebuilding python does indeed the trick Apr 02 21:42:22 Anyone around that's got a glibc, not eglibc build handy? Apr 02 21:42:51 yes Apr 02 21:43:17 Can you build an image w/ gdb or gdbserver installed and tell me if libthread-db1 gets pulled in? Apr 02 21:43:25 (ie /lib/libthread_db.so.1 exists in the image) Apr 02 21:43:40 sure Apr 02 21:43:43 thanks Apr 02 21:45:19 ah..I'm hit by openssl bug atm...you'll have to wait... Apr 02 21:45:27 k Apr 02 21:45:35 khem, semi-related ping Apr 02 21:45:40 Anyone here with a Topas 910 board? Apr 02 21:52:23 Tartarus: building now ;) Apr 02 21:53:25 Aegiron, k Apr 02 21:53:30 thanks :) Apr 02 21:53:37 np Apr 02 21:53:41 * Tartarus switches back to kicking gcc-cross-sdk about Apr 02 21:53:57 03Stanislav Brabec  07org.openembedded.dev * r7de2f107cf 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Apr 02 21:54:07 it'll take a while tho, my system is not the fastes there's to find Apr 02 21:54:07 03Stanislav Brabec  07org.openembedded.dev * rf318163737 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Apr 02 21:54:09 03Stanislav Brabec  07org.openembedded.dev * r9c67c93114 10openembedded.git/recipes/ (18 files in 5 dirs): Apr 02 21:54:09 virtual/gail: Introduction of new virtual: Apr 02 21:54:09 * provided by gtk+ >= 2.13.0 or gail Apr 02 21:54:10 * gtk+ >= 2.13.0 now stages libgailutil Apr 02 21:54:12 * use appropriate PROVIDES and DEPENDS for gail Apr 02 21:54:14 * bump PR to fix packages dependencies Apr 02 21:55:03 k Apr 02 21:55:12 I've got a buncha annoying things to figure out atm, so no worries :) Apr 02 21:55:44 Tartarus: me too I fixed the breakage, now can test Apr 02 21:56:12 thanks, both of you :) Apr 02 21:56:34 in my alittle old custom setup, with eglibc i'm not getting that lib pulled in for gdb or gdbserver Apr 02 21:56:58 kicking a current .dev build now trying to make gcc-cross-sdk relocatible first tho Apr 02 22:00:39 kergoth: no way to speed-up 'NOTE: Resolving any missing task queue dependencies' phase? Seems slower than before, lately... Apr 02 22:01:37 Tartarus: adding package gdb-6.8-r3 Apr 02 22:03:00 ant__, yeah but is libthread-db1 pulled in? Apr 02 22:03:11 (ie /lib/libthread_db.so.1 and db-1.0.so in the rootfs0 Apr 02 22:03:48 heh, denix, I get that curl problem now too :) Apr 02 22:04:03 Tartarus: Apr 02 22:04:04 ls tmp/rootfs/lib/libthread_db Apr 02 22:04:04 libthread_db-1.0.so libthread_db.so.1 Apr 02 22:04:06 * Tartarus thinks he'll just hack around it for the moment and fry this fish later Apr 02 22:04:12 Aegiron, thanks Apr 02 22:04:21 khem, another glibc/eglibc compat problem found :( Apr 02 22:04:38 Aegiron, can you pastebin the control file for gdb and libthread-db1 please? Apr 02 22:04:54 erh, and what file is that? :p Apr 02 22:05:01 heh Apr 02 22:05:04 Tartarus: I have not reached to reproduce lzo probem yet Apr 02 22:05:19 ar -x tmp/deploy/ipk/.../file.ipk, tar zxf control.tar.gz Apr 02 22:05:26 khem, odd Apr 02 22:05:31 where might I find a download mirror that would contain web.py-0.31.tar.gz? (http://webpy.org is down) - I recall an angstrom mirror that may have it? Apr 02 22:05:37 Tartarus: *libthread_db-1.0.so │ 26308│Mar 28 17:51│ Apr 02 22:05:42 Aegiron, or email both ipks to trini@embeddedalley.com please Apr 02 22:05:45 Tartarus: I think I have this all ARM thing Apr 02 22:05:45 i can extract the rest Apr 02 22:05:59 so I am guessing it happens in thumb mode Apr 02 22:06:00 ok, i'll do that Apr 02 22:06:31 Tartarus: ywhat is the issue with glibc Apr 02 22:07:01 Tartarus: @libthread_db.so.1 │ 0│Apr 3 00:03│ Apr 02 22:07:16 khem, try an eglibc image w/ gdb installed, libthread-db isn't being pulled in, at least for me Apr 02 22:07:26 (in a build w/o those gcc patches, fwiw) Apr 02 22:07:46 I faintly remember that Apr 02 22:07:48 problem Apr 02 22:07:51 oh? Apr 02 22:07:56 i'm snapped a little bit back Apr 02 22:08:05 I'll go diff eglibc dirs quick Apr 02 22:09:03 nope Apr 02 22:10:04 Tartarus: mail sending Apr 02 22:10:26 sent Apr 02 22:11:19 thanks Apr 02 22:12:05 got 'em Apr 02 22:21:23 hi Apr 02 22:21:42 how can I expess a package needs a version >=x of some other package? Apr 02 22:21:56 something like DEPENDS="x", but x >= some version Apr 02 22:22:13 I tried DEPENDS="x_version", but that didn't work Apr 02 22:22:28 if I'm not mistaken DEPENDS="x-version" Apr 02 22:22:41 aha -, I tried _ for some reason Apr 02 22:23:29 ok, - seem sto work, thanks :) Apr 02 22:24:37 ah-ha, found the relocation thing, or at least one of 'em Apr 02 22:24:58 minimal/arm/arm-oe-linux-gnueabi/bin/as -> /usr/local/minimal/arm/bin/arm-oe-linux-gnueabi-as Apr 02 22:28:33 dent: if I'm correcy DEPENDS will only take x=version, not x >=version Apr 02 22:28:45 correcy=correct, typo Apr 02 22:29:22 Aegiron: I thought so; I can do with = Apr 02 22:29:46 * Aegiron thinks that those small logitech keyboards are a pain in the *ss Apr 02 22:30:17 btw. I suspect the boost recipe to be broken... and after hours I think I managed to find what's wrong Apr 02 22:30:21 at least on my box Apr 02 22:32:30 ant__: agreed, not having commands to show that information is annoying, it's on my list of OpenEmbedded issues that bug me.. Apr 02 22:32:32 heh Apr 02 22:33:05 kergoth: I would hepl..unfortunately my python skillz are very poor Apr 02 22:33:10 hehe Apr 02 22:33:14 mine are slowly improving Apr 02 22:33:32 * kergoth is battling autotools and losing at the moment Apr 02 22:33:46 if I have a block of python code (like in python __anonymous() { ) Apr 02 22:34:23 I found on my box, if the commands start with a real tam (\t), the code is not executed... if I susbtitute it with 4 spaces (I guess other numbers, like >1 would do, I tried 4) - it works Apr 02 22:34:28 is that expected behaviour? Apr 02 22:34:39 (with \t, it seems to be silently ignored) Apr 02 22:34:58 ehm, ^ real TAB Apr 02 23:03:04 After update (today) building of x11 image failed: http://pastebin.com/m36c67803 Apr 02 23:03:57 1 hour ago I have success of building this image, I just right now, after update - error. Apr 02 23:08:44 force a rebuild of python-native Apr 02 23:10:56 how? :) Apr 02 23:13:50 fpga: bitbake -c clean python-native;bitbake python-native Apr 02 23:13:59 So, whee, I think this should let the SDK stuff be moved around, within reason Apr 02 23:14:24 Just needed to fix binutils and gcc symlinks from /path/to/bin/foo to ../../bin/foo Apr 02 23:14:40 Tartarus: why is it difficult to do so Apr 02 23:15:08 khem, it wasn't :) Apr 02 23:15:16 first test passed Apr 02 23:15:21 hello.c compiled w/ moved dir :) Apr 02 23:15:40 gn Apr 02 23:15:46 Ah but c++ fails Apr 02 23:15:54 headers Apr 02 23:16:03 yeah Apr 02 23:16:33 do you have sysroot relative Apr 02 23:16:44 yeah Apr 02 23:16:59 There's probably another silly thing somewhere, poking around Apr 02 23:17:49 bah Apr 02 23:17:55 it's not sysroot'ing c++ headers Apr 02 23:18:30 --with-gxx-include-dir Apr 02 23:18:33 is your friend Apr 02 23:18:53 not so sure about that ;) Apr 02 23:19:15 ignoring nonexistent directory "/usr/local/minimal/arm/arm-oe-linux-gnueabi//usr/include/c++" Apr 02 23:19:29 (when moved) Apr 02 23:19:50 I wonder... Apr 02 23:21:02 what do you have for --with-gxx-include-dir in your configure Apr 02 23:22:12 So, looking at a CodeSourcery configure, I'm gonna try this: Apr 02 23:22:16 -EXTRA_OECONF_PATHS = "--with-gxx-include-dir=${prefix}/${TARGET_SYS}/${layout_i Apr 02 23:22:17 +EXTRA_OECONF_PATHS = "--enable-poison-system-directories \ Apr 02 23:22:37 (Dropping gxx-include is the big part, other just seems like a good idea) Apr 02 23:22:40 --enable-poison-system-directories wont work Apr 02 23:22:50 eh? Apr 02 23:22:53 this is not upstream AFAIK Apr 02 23:22:59 we have the zecke patch Apr 02 23:23:04 which does similar stuff Apr 02 23:23:07 not added or not working? Apr 02 23:23:15 not added I think Apr 02 23:23:18 it works though Apr 02 23:23:20 configure didn't blow Apr 02 23:23:25 but, noted will check in a min Apr 02 23:23:38 man, i like this machine :) Apr 02 23:23:42 configure never blows on -with/-without Apr 02 23:24:13 k Apr 02 23:24:17 you need --with-gxx-include-dir to make it relative Apr 02 23:24:38 anyone played with distributed bug tracking systems at all? Apr 02 23:24:52 kergoth: which one Apr 02 23:25:10 any, i haven't tried any yet, wondering if any are actually suitable for day to day use Apr 02 23:25:14 khem, we had that, relative to sysroot and it was only looking in that hardcoded path, not being relative to sysroot Apr 02 23:25:17 I only know gnats bugzilla and few others Apr 02 23:25:35 it should be relative Apr 02 23:25:53 mark mitchell fixed it in 2007 ( a montavista bug) Apr 02 23:28:44 Nope Apr 02 23:29:05 It works now, with a caveat i need to add a arm../include -> arm../usr/include symlink Apr 02 23:29:10 more clean fixing in a min or 2 :) Apr 02 23:30:02 * khem should try sdk sometimes Apr 02 23:31:14 anyone with a bit knowleg Apr 02 23:31:16 damn Apr 02 23:32:15 i'm having troubles with the most easy qt-embedded project you could imagine :) Apr 02 23:32:59 http://pastebin.com/d698011b9 -> look at line 55 Apr 02 23:33:56 it does'nt recognise the stuff it just created... Apr 02 23:35:46 Aegiron: run file on /home/larm/oe/build/tmp/work/armv5te-angstrom-linux-gnueabi/embedtest-1.0.0-r0/image/usr/bin/embedtest Apr 02 23:37:49 hmm, that runs , which it probably should'nt :) Apr 02 23:39:21 what does it say Apr 02 23:42:07 o, I just figured out what the problem was Apr 02 23:42:47 it seemd that the sources files I packaged still contained the .o files and the i686 binary Apr 02 23:43:18 and apparently, these are not overwritten by bitbake.. Apr 02 23:46:56 * Aegiron seems to be getting a bit tired Apr 02 23:48:08 khem: thanks anyway Apr 02 23:54:38 * Tartarus pokes the CS stuff a bit more Apr 02 23:54:55 I think we just need to handle c++ includes "special" like they do and have /include/c++ Apr 02 23:55:03 and /usr/include Apr 03 00:08:28 ok, meta-toolchain fixup done, poking some various versions next Apr 03 00:09:53 Oh damnit Apr 03 00:10:14 * mwester suspects it may not be as done as Tartarus thought ;-) Apr 03 00:10:27 No, it's a stupid gcc-4.3.3.inc bug :) Apr 03 00:11:15 Only 4.3.3 was using gcc.releasenotes.org as a URI not GNUMIRROR Apr 03 00:11:30 And that sites been down for a month it feels like to me Apr 03 00:11:38 bugger, just created an image with default root password and loaded it on the board: Permission denied, please try again. Apr 03 00:22:35 03Tom Rini  07org.openembedded.dev * re053ff45ac 10openembedded.git/ (11 files in 2 dirs): Apr 03 00:22:35 gcc: Fix 4.1.[012], 4.2.[1234] and 4.3.[123] SRC_URI to use ${GNU_MIRROR} and update checksums.ini Apr 03 00:22:35 This fixes a big problem with 4.3.3 as it wasn't using ftp.gnu.org but Apr 03 00:22:35 an alternate mirror that's gone away. Since I had to fix one I noticed Apr 03 00:22:37 others not calling ${GNU_MIRROR} but ftp.gnu.org. Also a few weren't Apr 03 00:22:39 using ${PV}, so use that too. Apr 03 00:35:56 ok, kicking off: for VERSION in 3.3.4 3.4.4 4.1.2 4.2.4 4.3.3;do ANGSTROM_GCC_VERSION=${VERSION} MACHINE=qemux86 bitbake meta-toolchain-gpe;done Apr 03 00:36:14 with TMPDIR=.../tmp.${ANGSTROM_GCC_VERSION} and we'll see which of these work Apr 03 00:36:27 I suspect some of the 3.*'s won't, but i wanna know anyhow Apr 03 00:51:23 yeah, there blew up 3.3.4 Apr 03 01:10:34 is there any way to generate a password so I can put it in /etc/passwd on the root filesystem? Apr 03 01:10:56 not really generate, rather encode Apr 03 01:15:47 * Aegiron is off to bed. sleeping on a keyboard isn't exactly comfy ;) **** ENDING LOGGING AT Fri Apr 03 02:59:56 2009