**** BEGIN LOGGING AT Tue Nov 03 02:59:57 2009 Nov 03 07:20:10 03Steffen Sledz  07org.openembedded.dev * rab0b2217fe 10openembedded.git/recipes/busybox/ (busybox-1.11.3/iptunnel.patch busybox.inc busybox_1.11.3.bb): Nov 03 07:20:10 busybox-1.11.3: undefined reference __cpu_to_be16 fixed for busybox-static Nov 03 07:20:10 A patch fixing this problem inside the linux kernel was not accepted. Nov 03 07:20:10 For details see: Nov 03 07:20:10 - http://lkml.org/lkml/2009/6/18/223 Nov 03 07:20:14 - http://patchwork.kernel.org/patch/31164/ Nov 03 07:20:16 Signed-off-by: Steffen Sledz Nov 03 07:30:21 steffen? Nov 03 08:20:16 03David-John Willis  07org.openembedded.dev * r55331de730 10openembedded.git/ (conf/checksums.ini recipes/xfce-base/xfce4-panel_4.6.2.bb): xfce4-panel: Add new upstream 4.6.2 release and checksums. Nov 03 08:29:20 gm Nov 03 08:54:28 morning all Nov 03 08:54:46 hi RP Nov 03 09:15:24 hi, is there a clean way to include some kernel-module.ipk in an image without including kernel-image ? Nov 03 09:15:54 no Nov 03 09:16:28 maybee with a dummy package Nov 03 09:16:33 but there isnt one yet Nov 03 09:16:57 ok thanks I'm going to try this Nov 03 09:33:05 * * OE Bug 5327 has been RESOLVED (INVALID) by ospite(AT)studenti.unina.it Nov 03 09:33:07 * * gcc-cross-intermediate-4.4.1-r6.1: oe_runmake failed. Nov 03 09:33:09 * * http://bugs.openembedded.net/show_bug.cgi?id=5327 Nov 03 09:35:54 Good mornig, I am attempting to build the qte sdk, however it fails on the final task when it is trying to populate the sdk, any ideas what could be the main cause of this? Nov 03 09:36:42 what error do you get? Nov 03 09:37:36 hi xora Nov 03 09:38:59 zub: ERROR: function do_populate_sdk failed Nov 03 09:39:36 runing task meta-toolchain-qte.bb, do_populate_sdk Nov 03 09:39:40 I suspect it's an issue about stripping static libs Nov 03 09:39:47 check the detailed error log Nov 03 09:40:27 don't remember now, but bitbake/OE should tell you where to look for it Nov 03 09:40:28 *sigh* Nov 03 09:40:46 jeff__ you could use tinderbuild stuff Nov 03 09:41:02 which is a really easy effort to share the error Nov 03 09:41:20 INHERIT += "oestats-client" Nov 03 09:41:20 OESTATS_SERVER = "tinderbox.openembedded.net" Nov 03 09:41:34 OESTATS_BUILDER = "yournick_or_whatever" Nov 03 09:41:42 put this into your localconf Nov 03 09:41:47 and send us the url Nov 03 09:41:52 when its failing Nov 03 09:43:04 hi florian Nov 03 09:43:37 has anyone managed to build glibc with framepointers? Nov 03 09:44:25 hm no Nov 03 09:44:32 didnt try it Nov 03 09:44:42 I am only using eglibc and uclibc Nov 03 09:45:42 woglinde: could you try with eglibc? :) Nov 03 09:45:46 good morning Nov 03 09:46:04 woglinde: Ok, changed conf, one moment while I run build again. Nov 03 09:46:28 florian, good morning ! Nov 03 09:46:58 florian: hey :) Nov 03 09:47:10 florian: who is approving new projects on LTG? Nov 03 09:48:47 03Martin Jansa  07shr/merge * r2260d1ba38 10openembedded.git/recipes/shr/e-wm-theme-illume-gry_git.bb: Nov 03 09:48:47 e-wm-theme-illume-gry: missing recipe from shr/import Nov 03 09:48:47 Signed-off-by: Martin Jansa Nov 03 09:48:48 03Martin Jansa  07shr/merge * r9a13bb5dc2 10openembedded.git/recipes/sqlite/ (sqlite3-native_3.6.19.bb sqlite3_3.6.19.bb): Nov 03 09:48:51 sqlite3_3.6.19: add tcl-native to DEPENDS Nov 03 09:48:53 * Header files are generated by tcl in newer sqlite Nov 03 09:48:55 Signed-off-by: Martin Jansa Nov 03 09:49:07 I wonder when jama is sleeping Nov 03 09:49:18 * JaMa is wondering too Nov 03 09:50:05 stupid q, if I have multiple libs with -l do they need to be in order (so things like libc last) or doen't that matter ? Nov 03 09:50:31 good that employer isn't wondering when I'm working for him.. I should change work to something more linux related.. Nov 03 09:50:42 effem order normaly doesnt matter Nov 03 09:51:03 effem I didnt see any stuff yet where two libs provides the same symbols Nov 03 09:51:38 jama *g* Nov 03 09:51:43 woglinde: thought that Nov 03 09:52:14 why the heck is tcl generating header for sqllite? Nov 03 09:52:23 this looks like stupid idea Nov 03 09:52:29 from developers Nov 03 09:54:23 woglinde: haven't checked what they're doing but mrmoku complained that it cannot be build on his buildhost.. because it calls tcl and used tcl from buildhost here instead of tcl-native.. Nov 03 09:57:02 sleeping... Nov 03 09:57:14 it is too good to be real :) Nov 03 09:57:22 having trouble with the qt4 mysql driver, the .so complains about missing mysql_server_init, I added -lmysqlclient to it, still i do not see that lib end up in the mysql driver .so file Nov 03 09:59:19 effem hm Nov 03 09:59:30 I didnt build qt lately Nov 03 09:59:36 should again Nov 03 09:59:45 but first I will finish boost stuff Nov 03 09:59:53 build boost with cmake Nov 03 10:00:08 is much better then with bjam Nov 03 10:00:11 woglinde: it all builds without problem, but if I load the mysql driver it complains about an undefined symbol Nov 03 10:00:36 mythbackend: symbol lookup error: /usr/lib/qt4/plugins/sqldrivers/libqsqlmysql.so: undefined symbol: mysql_server_init Nov 03 10:00:57 this .so file is delivered from qt4-plugin-sqldriver-sqlmysql Nov 03 10:01:01 hm ldd Nov 03 10:01:03 shows Nov 03 10:01:05 libmysqlclient_r.so.16 => /usr/lib/libmysqlclient_r.so.1 Nov 03 10:01:08 on my host Nov 03 10:01:30 woglinde: what is the exact ldd cmd you give ? Nov 03 10:01:41 ldd /usr/lib/qt4/plugins/sqldrivers/libqsqlmysql.so Nov 03 10:01:49 debian unstable Nov 03 10:01:54 package shows two libs Nov 03 10:02:03 one with _r one without Nov 03 10:02:16 there are two Nov 03 10:02:28 i have no ldd on my beagle and ldd cross does not seem to work Nov 03 10:02:51 * eFfeM just did strings mysql.so | less Nov 03 10:03:11 is there a cross ldd ? Nov 03 10:03:26 eFfeM: qemu :) Nov 03 10:03:38 eFfeM: you can use objdump -x and look at the SO_NEEDED entry Nov 03 10:04:58 cross ldd does not work, objdump -x has NEEDED entry, not SO_NEEDED and no mysqlclient in it Nov 03 10:08:01 hm, can't find a thing right now, need to leave kbd for 15 minutes Nov 03 10:09:05 oops Nov 03 10:09:43 03Frans Meulenbroeks  07org.openembedded.dev * r7bd8c333e2 10openembedded.git/recipes/mythtv/ (2 files in 2 dirs): mythplugins: fixed remaining staging issues Nov 03 10:09:53 03Frans Meulenbroeks  07org.openembedded.dev * rf0792ef5d7 10openembedded.git/: Merge branch 'org.openembedded.dev' of git.openembedded.org:openembedded into org.openembedded.dev Nov 03 10:10:01 mythtv the never ending story Nov 03 10:10:03 *g* Nov 03 10:10:58 hm Nov 03 10:12:13 woglinde: and after all that work I wonder what monster embedded platform is going to run it all ;-) Nov 03 10:13:31 djwillis hm some omap3 Nov 03 10:13:43 I could test it with beagleboard and usb tv stick Nov 03 10:13:54 if the day would have 120 hours Nov 03 10:14:48 woglinde: I tried on the Pandora, I gave up as I just came to the view that OMAP3 was a little light for it. It would be cool to use the Pandora as a Myth frontend (or even both). Nov 03 10:15:17 * XorA shudders at myth Nov 03 10:15:23 DJWillis hm light? Nov 03 10:15:31 omap3 isnt enough? Nov 03 10:15:46 appros pandora are they now available? Nov 03 10:16:23 03Martin Jansa  07shr/merge * r3d9d81c87f 10openembedded.git/conf/distro/include/sane-srcrevs.inc: Nov 03 10:16:23 sane-srcrevs.inc: missing srcrev for calc Nov 03 10:16:23 Signed-off-by: Martin Jansa Nov 03 10:16:41 woglinde: regarding the Pandora, there are near final units out and about but I have been hacking on it on and off for a few years, just as a hobby dev with early hardware. Nov 03 10:18:25 florian: good morning Nov 03 10:18:27 mickeyl: good morning Nov 03 10:18:29 woglinde: good morning Nov 03 10:18:31 zecke: good afternoon Nov 03 10:19:08 he pb Nov 03 10:19:14 only 4 days left Nov 03 10:19:35 pb__: hey! Nov 03 10:20:02 pbuilder rockz Nov 03 10:20:21 re Nov 03 10:20:26 re effem Nov 03 10:20:44 djwillis hm mythtv should work on sheeva Nov 03 10:20:52 as server Nov 03 10:21:25 Ok, I think this is it http://tinderbox.openembedded.net/builds/37210/ Nov 03 10:22:25 jeff did you pull the latest stuff? Nov 03 10:24:15 I did Nov 03 10:24:22 jeff ah thats a nasty error Nov 03 10:24:38 hm could you try to set Nov 03 10:25:30 osrry, set? Nov 03 10:25:35 sorry Nov 03 10:26:36 PREFFERED_PROVIDER_gconf = "gconf" Nov 03 10:26:42 in your local.conf Nov 03 10:26:43 sorry Nov 03 10:26:58 needed some time to verify it Nov 03 10:28:26 Just to confirm, I don't need to build anything else before I build the sdk? Nov 03 10:28:48 woglinde: now that's a cool idea, had not considered that, ok, now to get a Sheeva to mess with. Nov 03 10:32:14 woglinde: back to my orlginal Q: if I have two libs A and B and they are linked with -lA -lB and B references a symbol of A, will that resolve or does that give an order problem Nov 03 10:32:42 that is fine Nov 03 10:33:05 ok, just wanted to verify, i know in the old days order was important Nov 03 10:34:03 effem 30 years ago? Nov 03 10:34:12 it's important with static libs, but shared libs are more forgiving. Nov 03 10:34:26 woglinde: only 25 :-) Nov 03 10:34:32 mostly because shared libs are an "all or nothing" proposition: you can't load just part of the library like you do with .a files Nov 03 10:35:12 pb hm intressting Nov 03 10:35:20 every day I learn something new Nov 03 10:35:24 pb__: the issue is that the lib does not show up in objdump -x Nov 03 10:35:43 although I did -lmysqlclient on the linker command line Nov 03 10:35:46 okay time to hang out the whasing Nov 03 10:36:01 effem try -lmysqlclient_r Nov 03 10:36:06 eFfeM: "show up" in what way? you mean you don't get a NEEDED record? Nov 03 10:36:17 there is no NEEDED record Nov 03 10:36:26 damn mysql mirror is so slow Nov 03 10:36:28 woglinde: what is the diff between the regular and the _r version? Nov 03 10:36:33 thats I cant look up it myself Nov 03 10:36:39 in our building Nov 03 10:37:19 eFfeM: can you put the output of "readelf -a" in a pastebin? Nov 03 10:37:27 will do Nov 03 10:38:56 pb__ it is kinda much so redirected to a file and loaded to http://filebin.ca/tfyscf/readelf Nov 03 10:39:12 righto Nov 03 10:39:42 03Jeremy Lainé  07org.openembedded.dev * rb8f658eb91 10openembedded.git/recipes/linux/ (linux-2.6.31/boc01/005-091008-isl12024.patch linux_2.6.31.bb): linux-2.6.31: update boc01 RTC driver for upstream submission Nov 03 10:41:15 yeah, that is a bit odd Nov 03 10:41:23 as you say, no NEEDED, but it clearly does have relocs against mysql Nov 03 10:41:34 what's the exact final link command you are using? Nov 03 10:42:33 03Martin Jansa  07shr/merge * r3262724e9c 10openembedded.git/recipes/calc/ (calc_0.0.1.bb calc_0.0.2.bb calc_git.bb): Nov 03 10:42:33 calc: add 0.0.2 tag and lower default preference for untagged version Nov 03 10:42:33 Signed-off-by: Martin Jansa Nov 03 10:42:51 $(LINK) $(LFLAGS) -o $(TARGET) $(OBJECTS) $(LIBS) $(OBJCOMP) Nov 03 10:42:51 and libs contains -lmysqlclient Nov 03 10:43:16 makefile is genreated with /home/frans/oe/tmp_angstrom/staging/i686-linux/usr/bin/qmake2 -spec ../../../../mkspecs/linux-g++ -unix -after INCPATH+=/home/frans/oe/tmp_angstrom/staging/armv7a-angstrom-linux-gnueabi/usr/include/freetype2 LIBS+=-L/home/frans/oe/tmp_angstrom/staging/armv7a-angstrom-linux-gnueabi/usr/lib -o Makefile mysql.pro Nov 03 10:43:52 hmmm Nov 03 10:44:19 what does that actually end up expanding to? Nov 03 10:44:24 morning Nov 03 10:44:27 can you run make and paste the actual command that it runs? Nov 03 10:44:43 in qt4 I also added LIBS+=-lmysqlclient, don't see that one in the qmake file Nov 03 10:44:48 hi hrw Nov 03 10:45:14 pb__ what is the best way to do that? is this cmd somewhere in a tmp/*log* file ? Nov 03 10:45:16 temp Nov 03 10:45:25 eFfeM: it should be in log.do_compile Nov 03 10:48:11 pb__ this is (what I think is) the relevant part of do_compile: Nov 03 10:48:12 http://www.pastebin.ca/1654649 Nov 03 10:49:43 ah Nov 03 10:49:49 "-L -lmysqlclient" is no good Nov 03 10:50:17 ~curse keeping linux-libc-headers in local repositories Nov 03 10:50:37 May the fleas of a thousand camels infest your most sensitive regions, keeping linux-libc-headers in local repositories ! Nov 03 10:51:40 you need to either remove that -L, or supply an argument to it. as it stands, the -L will swallow the following "-lmysqlclient" as its argument, which will not produce any useful effect unless you happen to have a directory named "-lmysqlclient" with libraries in :-} Nov 03 10:54:07 hm, the -L is not mine Nov 03 10:54:25 woglinde: Same error http://tinderbox.openembedded.net/public/logs/task/3417672.txt I going to try fresh build of everything by getting rid of the tmp dir and starting afresh Nov 03 10:54:49 I just added the LIBS+=-lmysqlclient line to qmake Nov 03 10:55:31 * eFfeM is pondering whether he should just do LIBS+="-lmysqlclient -lmysqlclient" to work around the -L Nov 03 10:55:50 or qmake LIBS does something different than I've been told Nov 03 10:57:54 jeff__ could try bitbake -g sdkstuff Nov 03 10:58:06 and look where the gconf-dbus is sneaking in? Nov 03 10:58:12 or which package its need Nov 03 10:58:38 I dont know why the sdk needs gconf anyway Nov 03 10:58:41 zecke do you know? Nov 03 10:59:22 woglinde: SDK == meta-toolchain? Nov 03 11:00:34 zecke jepp meta-toolchain-qte Nov 03 11:00:38 http://tinderbox.openembedded.net/builds/37223/ Nov 03 11:00:56 pb__: do you hav a better suggestion as the duplicatd -lmysqlclient Nov 03 11:02:33 woglinde: gconf... Qt -> Phonon -> GStreamer - > Gconf Nov 03 11:02:45 ah Nov 03 11:02:47 right Nov 03 11:02:58 gconf-dbus should die Nov 03 11:03:28 effem hm so the spec file is broken Nov 03 11:03:58 oooo k Nov 03 11:04:16 breaking territory that is still uncharted by me ... Nov 03 11:04:43 effem you could begging zecke Nov 03 11:04:45 to fix it Nov 03 11:05:07 zecke: please? very please ?? very very very please ? Nov 03 11:05:16 for glib it seems right Nov 03 11:05:25 -lglib-2.0 Nov 03 11:05:29 without -L Nov 03 11:05:39 but this linking is wrong anyway Nov 03 11:06:01 ah gconf-dbus died? Nov 03 11:06:17 dont know Nov 03 11:06:22 seems I should google Nov 03 11:06:29 what it is usefull for Nov 03 11:07:23 hm okay Nov 03 11:07:28 missusing of dbus Nov 03 11:07:29 as ipc Nov 03 11:07:31 zecke any idea what I need to do to build qt4 mysql connector with -lmysqlclient in a proper way ? Nov 03 11:08:15 eFfeM: look at test Qt is executing to figure out if it is installed Nov 03 11:08:30 eFfeM: maybe some variable is not properly initialized (config.* and conigure) Nov 03 11:09:16 zecke: wehll mysql.pro does a check for isEmpty(QT_LFLAGS_MYSQL) and apparently it is not Nov 03 11:11:14 sorry, I don't feel weel right now... poke me tomorrow Nov 03 11:11:44 zecke config.tests/unxi/mysql/mysql.pro does contain LIBS += -lmysqlclient Nov 03 11:11:44 not sure how that is handled Nov 03 11:12:21 zecke: np, get well soon, i'll dig into it myself, will get back to you tomorrow or so if still needed Nov 03 11:12:27 and thanks for your pointer Nov 03 11:12:47 eFfeM: feel free to send me a mail with the problem description (i didn't follow this conversation) :) Nov 03 11:14:48 argh Nov 03 11:14:59 jo hrw Nov 03 11:15:01 unpacking by hand works, by bitbake fails Nov 03 11:17:04 someone uses OE or Poky on ubuntu 9.10? Nov 03 11:17:18 hrw not yet Nov 03 11:18:09 hrw, problem with recent uduntu? Nov 03 11:18:10 hrw: yeah me Nov 03 11:19:04 Crofton|work: http://pastebin.ca/1654688 Nov 03 11:19:13 when I call same command by hand it works fine Nov 03 11:21:08 florian: what's happening to ltg? Nov 03 11:22:22 ant_work: what's wrong? Nov 03 11:22:35 can't load any web page Nov 03 11:24:52 fscking crap Nov 03 11:25:01 florian: I can ssh, though Nov 03 11:25:23 18364 list 20 0 591m 522m 3232 R 100 6.5 423:41.25 python Nov 03 11:25:25 ant_work: hmm indeed, looks like apache had some problem... just restarted it and it works again Nov 03 11:25:43 thx Nov 03 11:28:01 one sec, i'll restart it again Nov 03 11:28:06 RP: ping Nov 03 11:28:33 morning mickey|office Nov 03 11:28:37 morning florian Nov 03 11:28:46 florian: who can write the tax receipt for Eric? Nov 03 11:28:59 which form do it need to comply to? Nov 03 11:29:28 florian time to swtich ti lighttp? Nov 03 11:30:09 or to drop debian... Nov 03 11:30:50 *v* Nov 03 11:31:13 mickey|office: all board members at least. i guess it shoudl be easy to find out how it works. if i get the data and if it is not really urgent i can do this Nov 03 11:31:37 florian: it might be urgent, depending on fiscal quarters Nov 03 11:31:43 florian: (relatively) Nov 03 11:31:53 ant_work: LTG should move to angstrom? Nov 03 11:32:02 hrw :) Nov 03 11:32:12 one day, perghaps,...step by step :] Nov 03 11:32:28 ant why drop debian? Nov 03 11:32:51 mickey|office : this is not urgent, I need just it before the end of december Nov 03 11:32:52 well, apt upgrades... Nov 03 11:32:59 can be dangerous Nov 03 11:33:10 eric__: ah, you're here :) Nov 03 11:33:16 ant_work, Jay7: drop me mail if you notice somethign is wrong. the fusionforge installation misght still need some fixes Nov 03 11:33:19 ant_work: o, rly? Nov 03 11:33:30 hi lrg Nov 03 11:33:30 eric__: good, we should aim to do it within Nov. then Nov 03 11:33:34 yo lrg Nov 03 11:33:38 much more risky than emerge system e.g. Nov 03 11:33:42 ;) Nov 03 11:33:42 hey florian mickeyl Nov 03 11:33:49 hi Liam Nov 03 11:33:49 florian: I'm still see no project's wiki.. :( Nov 03 11:34:01 per-project's Nov 03 11:34:01 and Marcin ;) Nov 03 11:34:01 mickey|office : ok fine, thanks Nov 03 11:34:04 hrw: or even bsd for *our* servers Nov 03 11:34:05 lol Nov 03 11:34:43 Jay7: right... I haven't managed to check what's missing there. Nov 03 11:35:11 bye for now Nov 03 11:35:33 florian: ok, I'll wait :) Nov 03 11:35:46 mickey|office: pong Nov 03 11:36:00 about apache.. you can install nginx as frontend Nov 03 11:36:25 and serve static images by nginx too Nov 03 11:36:35 all static content even Nov 03 11:36:51 free time is the question :) Nov 03 11:37:20 that's it Nov 03 11:37:31 RP: trying to understand the exec problem. what about http://python.pastebin.com/m4642faea ? Nov 03 11:37:48 zecke, all, not sure if this is right or not but if I go into the qt dir all exe files in config.tests/unix are ARM binaries, shouldn't this be x68 binaries ?? Nov 03 11:38:31 mickey|office: Did my use case break but yours works? Nov 03 11:40:58 RP: injecting bb into the globals seem to work with http://python.pastebin.com/m4642faea Nov 03 11:41:06 RP: at least the function call works Nov 03 11:41:23 mickey|office: ok, let me try a test Nov 03 11:45:09 03Frans Meulenbroeks  07shr/merge * r117419bac6 10openembedded.git/recipes/mythplugins/ (8 files in 2 dirs): mythplugins: created recipe, not fully tested yet Nov 03 11:45:11 03Leon Woestenberg  07shr/merge * r91403056b1 10openembedded.git/recipes/linux/ (3 files in 2 dirs): Nov 03 11:45:11 linux-kirkwood-2.6.31: Add PCI Express support for openrd-base. Nov 03 11:45:12 Note there are upstream problems involving devices using PCI IOMEM. Nov 03 11:45:14 Signed-off-by: Leon Woestenberg Nov 03 11:45:16 03Frans Meulenbroeks  07shr/merge * r9fd25648fc 10openembedded.git/ (4 files in 4 dirs): linux-kirkwood: updated defconfigs (more modules, more filesystems in kernel) Nov 03 11:45:19 03Frans Meulenbroeks  07shr/merge * r1377125138 10openembedded.git/recipes/mythplugins/mythplugins_0.21+0.22rc1.bb: mythplugins: added a few lines of comment for clarification Nov 03 11:45:22 03Frans Meulenbroeks  07shr/merge * r056193c663 10openembedded.git/classes/package.bbclass: Nov 03 11:45:25 Revert "package.bbclass: Cleanup getVar syntax to use True, not 1 as third argument (from Poky)" Nov 03 11:45:27 This reverts commit fcee557e1b0551d1743bbb3ad61102b9ccc4cba8. Nov 03 11:45:29 03Frans Meulenbroeks  07shr/merge * r0c5891b4c9 10openembedded.git/conf/checksums.ini: checksums: added mythplugins checksum Nov 03 11:45:38 03Frans Meulenbroeks  07shr/merge * r138f792724 10openembedded.git/classes/package.bbclass: Nov 03 11:45:41 Revert "Revert "package.bbclass: Cleanup getVar syntax to use True, not 1 as third argument (from Poky)"" Nov 03 11:45:43 Oops, accidently reverted a patch, reverting the revert. Sorry for any inconvenience Nov 03 11:45:45 This reverts commit 056193c6638892ddde7bddbfc6f6f06a2288ba64. Nov 03 11:45:47 03Holger Hans Peter Freyther  07shr/merge * r9393ac7170 10openembedded.git/recipes/binutils/binutils-cross-sdk.inc: Nov 03 11:45:50 Revert "binutils-cross-sdk: Avoid QA error on building" Nov 03 11:45:52 This reverts commit b294033116a6a4f33d1b1ec9f8256adaf7374298. This Nov 03 11:45:54 patch is no longer needed as .a files are no longer stripped. Nov 03 11:45:56 03Frans Meulenbroeks  07shr/merge * rbdc1df7330 10openembedded.git/recipes/mythtv/mythtv_0.22rc1.bb: mythtv: added staging Nov 03 11:46:11 03Frans Meulenbroeks  07shr/merge * r8b32d20a07 10openembedded.git/: Merge branch 'org.openembedded.dev' of git.openembedded.org:openembedded into org.openembedded.dev Nov 03 11:46:14 03Frans Meulenbroeks  07shr/merge * r2d365ab53b 10openembedded.git/recipes/mythtv/ (mythtv-0.21+0.22rc1/configure.patch mythtv_0.22rc1.bb): mythtv: updated staging, fixed configure; renamed recipe Nov 03 11:46:18 03Koen Kooi  07shr/merge * r5ccd6effbe 10openembedded.git/recipes/mythplugins/ (8 files in 2 dirs): mythplugins: move into mythtv directory Nov 03 11:46:21 03Frans Meulenbroeks  07shr/merge * ra26da59582 10openembedded.git/: Merge branch 'org.openembedded.dev' of git.openembedded.org:openembedded into org.openembedded.dev Nov 03 11:46:24 03David-John Willis  07shr/merge * r34616c490d 10openembedded.git/recipes/libgtkstylus/ (files/gtkstylus.sh libgtkstylus_0.3.bb): libgtkstylus: Fix session script to support loading of more than one GTK_MODULES. Nov 03 11:46:34 03Koen Kooi  07shr/merge * rf1a20d6191 10openembedded.git/recipes/xorg-lib/ (pixman/neon-24bpp.patch pixman/prefetch.patch pixman_git.bb): pixman git: add more NEON patches, bump SRCREV Nov 03 11:46:37 03Frans Meulenbroeks  07shr/merge * r13355c1620 10openembedded.git/ (17 files in 5 dirs): mythtv, mythplugins: updated to 0.22rc2 Nov 03 11:46:44 03Frans Meulenbroeks  07shr/merge * rd88287ca8a 10openembedded.git/: Merge branch 'org.openembedded.dev' of git.openembedded.org:openembedded into org.openembedded.dev Nov 03 11:46:47 03David-John Willis  07shr/merge * r8ad7a9b528 10openembedded.git/ (conf/checksums.ini recipes/xchat/xchat_2.8.6.bb): xchat: Add 2.8.6 release (and fix package to include d-bus service files). Nov 03 11:46:50 03David-John Willis  07shr/merge * r995cb3cbff 10openembedded.git/conf/machine/omap3-pandora.conf: omap3-pandora.conf: Merge in latest machine file. Nov 03 11:46:55 03Frans Meulenbroeks  07shr/merge * r24c8f2b9e0 10openembedded.git/recipes/mythtv/ (5 files in 2 dirs): mythplugins: added more plugins, packaging is not complete yet Nov 03 11:46:58 03David-John Willis  07shr/merge * rfd778a7191 10openembedded.git/recipes/xorg-xserver/ (2 files in 2 dirs): xserver-xorg-conf: Update xorg.conf for the omap3-pandora. i.e. cut out a load of cruft. Nov 03 11:47:02 (53 lines omitted) Nov 03 11:48:16 need a reboot, biab Nov 03 11:55:56 mickey|office: I can't see to reproduce the problem but it wasn't my imagination - am trying to come up with the test case :) Nov 03 11:56:14 btw I'm seeing the same -L -lmysqlclient when trying to build mythplugins with zoneminder support Nov 03 11:56:30 RP: oki Nov 03 11:59:17 mickey|office: Ah, I think this is due to the way methodpool works Nov 03 11:59:36 mickey|office: Imagine functionB was injected into the methodpool Nov 03 12:00:05 mickey|office: That adds it to __builtins__ which does not have bb in its scope? Nov 03 12:00:38 RP: perhaps we should change the way we inject functions into the methodpool then Nov 03 12:00:47 RP: could i ask for a minimal example again? Nov 03 12:00:52 i'd go from there then Nov 03 12:00:59 mickey|office: yes, working on one :) Nov 03 12:01:02 great Nov 03 12:01:18 mickey|office: I think this means we need to look at what methodpool is doing more closely :) Nov 03 12:01:35 * mickey|office debugs the case of a pointer changing between a method invocation *sigh* Nov 03 12:01:39 RP: yep Nov 03 12:01:58 * hrw curses at slow network connection of one customer Nov 03 12:02:23 fetching speed reminds me maemo.org websites... Nov 03 12:02:28 and their 33600 modem Nov 03 12:03:00 RP: thx for answer on Poky ML Nov 03 12:03:02 :)) Nov 03 12:06:55 mickey|office: http://python.pastebin.com/m460eed29 Nov 03 12:07:43 mickey|office: and I think now I have the test case I understand - methodpool should operate on some context, not __builtins__ and we should be using that context as g when calling bb.utils.better_exec ? Nov 03 12:07:56 RP: patch will be sent Nov 03 12:08:17 mickey|office: or we decide to use __builtins__ ? Nov 03 12:09:10 hrw: np, sounds good Nov 03 12:10:31 RP: pull then ;) Nov 03 12:11:32 RP: good news is that I have poky trunk build in progress now Nov 03 12:12:01 hrw: cool Nov 03 12:13:19 zecke: Any good reason why methodpool operates on __builtins__ rather than creating an environment and passing it around? Nov 03 12:13:23 RP: which branch do i need to make this example work? Nov 03 12:14:00 mickey|office: Any bitbake will do - the point it is doesn't work without an "import bb" in functionB Nov 03 12:14:17 mickey|office: and we want that to work Nov 03 12:14:41 *nod* Nov 03 12:14:53 give me some time to play with that Nov 03 12:15:48 mickey|office: np :) Nov 03 12:30:49 RP: I have to rebuild all from scratch tonite...is in your branch any nerw commit fixing packaged staging? Nov 03 12:32:32 ant_work: Not in the sense you're after. I'm working towards that Nov 03 12:32:36 I mean, giving correct arch to staging* ipk's Nov 03 12:32:39 ah, ok Nov 03 12:34:03 pls don't forget the ugly package_stagefile_shell Nov 03 12:36:07 eFfeM: no, not really. I guess you need to figure out where the -L is actually coming from and suppress it Nov 03 12:36:45 morning everyone Nov 03 12:38:20 hi pb___ Nov 03 12:39:34 hi ant_work Nov 03 12:44:18 JaMa: apropos your fontconfig crash, it looks like libpthread.so might be the guilty party. could you upload that binary? Nov 03 12:44:34 ant_work: ugly in what sense? Nov 03 12:44:56 is there some historical reason that dev.oe.org pull a git version of connman rather than a release tarball? It is labeled 0.42 but the git version apparently refers to 0.44 or so Nov 03 12:44:59 well, the recipes using it have all disabled PARALLEL_BUILD Nov 03 12:45:19 and is not friendly if you deplace DEPLOYITR Nov 03 12:45:30 *DEPLOYDIR Nov 03 12:45:49 ...see, it is used in zaurus-updater Nov 03 12:46:00 in poky you call it in do_install iirc Nov 03 12:46:08 in oe is in do_stage (iirc) Nov 03 12:46:16 spaetz_: yes, historical, back in the days, the releases were very seldomly. we should flag the GIT version PREFERRED_VERSION=-1 and use releases as much as possible Nov 03 12:46:27 yousee ...I'm waiting your commits for both reasons... Nov 03 12:46:48 the bastard is the first package failing during rebuild from pstage Nov 03 12:46:59 mickey|office: ok, thanks. Nov 03 12:49:57 ant_work: I was totally unaware of that problem. Is it on the OE list somewhere? Nov 03 12:50:13 yes, old thread...le me see Nov 03 12:55:23 pb__ guess the -L is from qmake, what I think is happening is that somewhere there is a -L${STAGING_LIBDIR} or so which is not properly expanded Nov 03 12:56:28 this was also what happened with the taglib-config soluion that you proposed (using ${STAGING_LIBDIR} (or I did something wrong there) Nov 03 12:57:16 ah, qmake Nov 03 12:57:37 I tend to abandon all hope when those five letters appear on my screen. Nov 03 12:57:59 if you talk sweetly to zecke then he might be able to assist you in navigating through its crazy universe. Nov 03 12:58:58 RP: I don't remember Nov 03 12:59:12 ant_work: sorry, I'm mostly afk... need to practice chinese Nov 03 13:01:32 03Sebastian Spaeth  07org.openembedded.dev * re972a049e4 10openembedded.git/recipes/connman/connman_0.46.bb: Nov 03 13:01:32 connman: include connman_0.46 (mv'd from connman_0.19) Nov 03 13:01:32 Signed-off-by: Sebastian Spaeth Nov 03 13:03:01 hmm.. connman... Nov 03 13:03:29 does connman has any frontend now? not counting clutter based ones Nov 03 13:06:20 03Sebastian Spaeth  07shr/merge * r4b38fcb9d6 10openembedded.git/recipes/connman/connman_0.46.bb: Nov 03 13:06:20 connman: include connman_0.46 (mv'd from connman_0.19) Nov 03 13:06:20 Signed-off-by: Sebastian Spaeth Nov 03 13:06:20 03Sebastian Spaeth  07shr/merge * r4057779c87 10openembedded.git/recipes/connman/connman_git.bb: connman: disable git preference for shr distro again. use 0.46 tarball Nov 03 13:06:50 hrw: on the freerunner the is mokonnect which provides a simple frontend. But I guess that is of little use on other devices Nov 03 13:06:57 there is... Nov 03 13:07:27 it uses connman as backend Nov 03 13:08:25 I see Nov 03 13:08:57 hmm... maybe good moment to curse^Wlearn python-gtk Nov 03 13:09:38 hehe Nov 03 13:12:09 RP: http://bugs.openembedded.net/show_bug.cgi?id=4580 Nov 03 13:13:05 RP: "package_stagefile_shell is there to manually mark files that need to get Nov 03 13:13:07 revived to make a build work, which as it happens is with bootloaders Nov 03 13:13:08 and kernel images" - koen Nov 03 13:14:08 RP: iirc actuallly only 5 recipes (bootloaders / updaters) are using it Nov 03 13:15:25 mmh, should have reset PR to 0 when I updated connman version. too late... now it's PR=1 Nov 03 13:15:29 sorry Nov 03 13:15:50 heh, don't worry about it. that's a pretty trivial issue in the grand scheme of things. Nov 03 13:16:16 pb__: http://jama.homelinux.org/org.openembedded.shr/fontconfig/libpthread-2.10.1.so Nov 03 13:16:50 pb__: could it be cause when eglibc is compiled without thumb enabled and rest of system has thumb enabled? Nov 03 13:17:14 JaMa: no, that's okay. eglibc needs to be compiled with interworking but it doesn't matter whether it is thumb or not Nov 03 13:17:44 however, it does look like that libpthread.so is very far from being interworking safe. Nov 03 13:18:09 any call to read(), for example, will be disaster Nov 03 13:18:18 which is a bit of a shame, obviously Nov 03 13:19:00 so should I try eglibc_2.9 if its better? Nov 03 13:19:27 method names seen in lots of segfault suggests that they call read() somewhere inside.. Nov 03 13:19:39 I think it should be fairly easy to patch eglibc to make it work Nov 03 13:19:50 I doubt 2.9 is any better than 2.10 in that respect, anyway. Nov 03 13:20:02 have a look at ports/sysdeps/unix/sysv/linux/arm/sysdep.h Nov 03 13:20:26 find the bit where it defines SYSCALL_ERROR_HANDLER inside a NOT_IN_libc guard Nov 03 13:20:52 the last line of the second section (i.e. !RTLD_PRIVATE_ERRNO) is the offending one: Nov 03 13:20:57 ldr pc, [sp], #4; Nov 03 13:21:09 that instruction is not interworking safe in v4t (although it is in v5t) Nov 03 13:21:40 try replacing it with "ldr lr, [sp], #4; bx lr" and see if the situation improves Nov 03 13:22:04 if yes, it will need a slightly more sophisticated patch to do the right conditional bits, but I think that is the approximate idea that you need Nov 03 13:23:05 I think clone.S is also broken for much the same reasons and needs a similar fix Nov 03 13:24:19 ok, I'll try later.. now meeting.. Nov 03 13:24:30 enjoy Nov 03 13:24:33 thanks! Nov 03 13:25:22 mickeyl: the above is probably what was causing your python crash as well Nov 03 13:25:48 anything linked with libpthread.so will be almost certain to crash immediately. Nov 03 13:26:08 pb__: another nice catch! Nov 03 13:27:18 pb__: wah, good find! Nov 03 13:28:40 03Sebastian Spaeth  07org.openembedded.dev * rf5701cca05 10openembedded.git/conf/checksums.ini: Nov 03 13:28:40 connman: add checksums for 0.46 Nov 03 13:28:40 Signed-off-by: Sebastian Spaeth Nov 03 13:29:36 03Sebastian Spaeth  07shr/merge * r0027909318 10openembedded.git/conf/checksums.ini: Nov 03 13:29:36 connman: add checksums for 0.46 Nov 03 13:29:36 Signed-off-by: Sebastian Spaeth Nov 03 13:33:40 03Matthew Dombroski  07org.openembedded.dev * rc510866574 10openembedded.git/classes/qmake2.bbclass: Nov 03 13:33:40 qmake2.bbclass: Add lupdate and lrelease. Nov 03 13:33:40 http://patchwork.openembedded.org/patch/987 Nov 03 13:33:40 Signed-off-by: Matthew Dombroski Nov 03 13:33:40 Acked-by: Holger Hans Peter Freyther Nov 03 13:37:17 mickey|office: heh, I guess I should stick to debugging libraries rather than organising meetings. :-} Nov 03 13:37:52 pb__: you did good work so far in both Nov 03 13:39:43 pb__: not your fault at all.. we (as in board) should have been more explicit... in one way it was clear that GA == OEDEM2009 Nov 03 13:42:13 * florian agrees Nov 03 13:45:27 hello Nov 03 13:45:37 hi Longfield Nov 03 13:47:09 I know it's pretty overkill, but I have colleague using openMP in one of the programs he wants to use on the robot. However it does not compile with the SDK (built from OE) I gave him. Is there openMP support in gcc-cross in OE (and gcc-cross-sdk) ? Nov 03 13:48:11 because for instance gcc-cross-sdk has "--disable-libgomp" in its EXTRA_OECONF, but gcc-cross does not Nov 03 13:48:34 Longfield: I don't know Nov 03 13:48:51 ant_work: ok, I'd never seen that bug before. The fix sounds reasonable Nov 03 13:49:16 03Holger Hans Peter Freyther  07org.openembedded.dev * rc724ad44d6 10openembedded.git/classes/base.bbclass: Nov 03 13:49:16 base.bbclass: Better error message in case the Config Parser fails Nov 03 13:49:16 http://patchwork.openembedded.org/patch/1088/ Nov 03 13:49:16 Tested-by: Ed Nelson Nov 03 13:49:29 This might be just wrong... imo both the SDK and the OE compiler should use the same options. Nov 03 13:49:59 03Ricardo Salveti  07org.openembedded.dev * r14caede54c 10openembedded.git/ (conf/checksums.ini recipes/eject/eject_2.1.5.bb): Nov 03 13:49:59 eject-2.1.5: adding HOMEPAGE and fixing SRC_URI Nov 03 13:49:59 As geocities is now down, changing the SRC_URI to point to OE Nov 03 13:49:59 mirror. Nov 03 13:49:59 Signed-off-by: Ricardo Salveti Nov 03 13:50:01 Acked-by: Holger Hans Peter Freyther Nov 03 13:53:12 funny how we can use eject_2.1.5 when upstream announces 2.1.0 as latest released version :) Nov 03 13:53:21 zecke, but not clear to the larger membership Nov 03 13:53:30 but as debian also uses 2.1.5 it seems to exist Nov 03 13:56:22 spaetz_: well, who is upstream? geocities showed it has none Nov 03 13:59:49 zecke, did you send around the proxy form? Nov 03 13:59:56 last time? Nov 03 14:01:32 zecke: well, the author has a new homepage and refers to eject.sf.net and ibiblio as canonical sources. And both show 2.1.0 as latest Nov 03 14:03:26 spaetz_: hehe then it is the usual hijacking :) Nov 03 14:03:50 :) Nov 03 14:05:00 Crofton|work: I sent the proxy form for the extraordinary assembly, when resetting the process it was kind of clear that GA will happen at OEDEM Nov 03 14:07:03 zecke, but no formalities were follwoed Nov 03 14:07:09 to many people assuming things Nov 03 14:07:30 the current problem is I can't find the proxy form Nov 03 14:09:18 Crofton|work: the proxy form is just a convience... anything like "I, ask to be my proxy " on anything (paper,fax, tissue is good enough) Nov 03 14:10:16 the basic problem is you did a really good job on the EGA announcement Nov 03 14:10:21 Crofton|work: "2009-08-21 13:59" is the date of the mail i sent for the ega Nov 03 14:10:36 so the lack of something similar for the actual GA was strange Nov 03 14:10:37 (and personally totally forgot about GA because I would not be there) Nov 03 14:11:17 :) Nov 03 14:12:42 Crofton|work: technically the new members were added after the "invitation" to the GA :} Nov 03 14:13:25 well, not circulating the invite to all members at the time of the GA was poor form Nov 03 14:14:21 Crofton|work: Looking at Afghanistan I don't think the GA situation is that bad. We have some constraints here... we will need a GA this year for the "authorities that we run a proper club" Nov 03 14:14:36 rofl Nov 03 14:14:55 I am not certain that I want the eV compared with Afghanistan Nov 03 14:16:04 Crofton|work: in one way messing up with sending proper forms is bad as we don't run a proper club... in another we have two options in terms of german authorities: a) Invite for another GA were no one but me in Taipei is showing up b) not hold one and face issues (e.g. removal of tax exemption status) Nov 03 14:17:09 Basically, we need to complete the board election Nov 03 14:17:10 Crofton|work a) and b) have both issues.... a) allows people in germany/taipei to show up and e.g. not take proxy forms from other people Nov 03 14:17:32 b) well not really nice either Nov 03 14:17:37 but it is good that more of us are reading the articles and thinking about how we appear to function Nov 03 14:18:18 Crofton|work and technically speaking. If "the board" sends the protocol to the court "one" can also send a letter of complain that the invitation was not proper and didn't arrive in time Nov 03 14:18:38 yeah, that is the sort of thing we need to avoid Nov 03 14:19:36 * Crofton|work really needs to pack and stop reading irc Nov 03 14:20:10 crap Nov 03 14:20:12 /kill Crofton|work packtime Nov 03 14:20:20 I leave 2 hours b4 I thought I did :) Nov 03 14:20:26 Crofton|work: The question is more like: Is anyone not able to give input or not? Nov 03 14:21:14 zecke gotta run Nov 03 14:21:31 basically, we need to treat this as the GA, for the board and finance reasons Nov 03 14:21:50 but, we have to get the larger membership caught up on things Nov 03 14:21:53 l8r Nov 03 14:54:56 03Koen Kooi  07org.openembedded.dev * rc55f2f7273 10openembedded.git/ (conf/checksums.ini recipes/perl/libxml-parser-perl_2.36.bb): libxml-parser-perl: add 2.36 Nov 03 14:55:07 03Koen Kooi  07org.openembedded.dev * raf05f65b38 10openembedded.git/contrib/angstrom/build-feeds.sh: Nov 03 14:55:07 angstrom feed builder: also clean perl and perl-native during architecture change Nov 03 14:55:07 * If recipes like xml-parser-perl fail with linking to your host LIBC, clean perl and perl-native Nov 03 15:10:01 how many of you use virtualization? Nov 03 15:10:48 morning Nov 03 15:10:48 I am curious which one (kvm, xen, vbox) can use more then one core of quadcore cpu. so far I am using vbox but onecore operation slows my VMs Nov 03 15:11:24 we use xen on about eight machines (with between 2 and 8 cpus each) and it seems to work pretty well for us. Nov 03 15:12:10 pb__: does it allows to use more then one core for VM? Nov 03 15:12:30 yeah, it would be a bit of a waste of our eight-core machines otherwise :-} Nov 03 15:13:24 you can assign an arbitrary number of virtual cpus to each of the virtual hosts. Nov 03 15:14:10 so, on one of our machines for example, we have four physical cores reserved for one particular vhost, and then all the other vhosts share the remaining four cores. Nov 03 15:14:19 zecke, woglinde, pb__, all: found what the problem was with my -L (but haven't cured it yet) Nov 03 15:14:24 eFfeM: good-o Nov 03 15:14:34 pb__: ok, will have to work on it then Nov 03 15:14:49 the cross mysql_config returns this Nov 03 15:14:50 L/home/frans/oe/tmp_angstrom/staging/armv5te-angstrom-linux-gnueabi/usr/lib -Wl,-rpath-link,/home/frans/oe/tmp_angstrom/staging/armv5te-angstrom-linux-gnueabi/usr/lib -Wl,-O1 -Wl,--hash-style=gnu -L -lmysqld -lz -lpthread -lcrypt -lnsl -lm -lpthread Nov 03 15:15:02 sadly I think debian are planning to drop support for xen in their next release, which is a bit of a shame. Nov 03 15:15:06 note the isolated -L Nov 03 15:15:09 iirc, their policy is "use kvm" or some such Nov 03 15:15:21 pb__: If waldi can get it plausibly working it might stay. Nov 03 15:15:24 eFfeM: ah, that is no good Nov 03 15:15:36 no idea yet how it must look, but Ihave students waiting so need to leave to school Nov 03 15:16:11 pb__: Basically, Xen dom0 support is total fail requiring extensive out of tree patches with questionable support status. Nov 03 15:17:25 yeah, so I gather Nov 03 15:18:08 broonie: good to know Nov 03 15:18:08 it isn't the end of the world if we need to build our own dom0 kernels, though, or indeed install a non-debian os for dom0. Nov 03 15:18:39 The existing support fails for some massive classes of hardware. Nov 03 15:18:40 my Debian machine do not use Debian kernels Nov 03 15:18:46 the latter is fairly tempting for other reasons, notably that almost every time grub gets upgraded in debian at least one of our machines becomes unbootable. Nov 03 15:19:08 Are you sure that isn't just grub? Nov 03 15:19:48 well, it's not the core bit of grub that goes wrong, it's the initramfs generating scripts. Nov 03 15:20:12 Oh, that's nothing to do with grub. Nov 03 15:20:13 they always seem to forget to include some or other important piece of infrastructure in the initramfs which makes it impossible to mount the rootfs, load the kernel or suchlike. Nov 03 15:20:25 Are you tracking unstable or something? Nov 03 15:20:44 * broonie notes that he seems to avoid all these problems on unstable but anyway. Nov 03 15:21:40 oh, sorry, I was mixing two different things up there. Nov 03 15:21:47 we do have the initramfs problem, but we also have the grub equivalent Nov 03 15:22:12 so, half the time we don't have enough grub bits to even boot the first stage kernel. Nov 03 15:22:27 then, when that does work, another half of the time we don't seem to have enough initramfs bits to mount the rootfs inside linux. Nov 03 15:23:14 the former problem is common enough that I now keep a special grub image available on a netboot server to rescue broken machines. Nov 03 15:23:15 You're the first person I've heard really complain about this stuff - like I say, is it unstable? Nov 03 15:23:39 I don't think so, though I'm not absolutely certain. Nov 03 15:23:53 I'm fairly sure we aren't running unstable on most of the machines with those problems. Nov 03 15:23:59 It sounds like you're getting very frequent updates which doesn't sound like an actual release. Nov 03 15:24:03 (could be testing too) Nov 03 15:24:26 well, "very frequent" is probably an exaggeration. I guess it tends to happen about once a year on each machine. Nov 03 15:24:45 but, given that we only reboot each of those machines about once or twice a year anyway, that is a fairly significant percentage of failed reboots. Nov 03 15:25:12 Certainly, I don't recall ever seeing an update of grub in a release. Nov 03 15:25:33 the only machine where it actually happens frequently (as in, every couple of weeks) is my desktop, which is indeed running unstable. Nov 03 15:25:53 I've given up trying to fix it on there, I just boot a rescue grub from tftp and then load the kernel manually. Nov 03 15:26:34 If you're running unstable report the bug. Nov 03 15:27:24 yah, I guess I should do that. Nov 03 15:28:17 "I didn't tell anyone about this problem and it's still not been fixed!" Nov 03 15:37:13 pb__: this reminds me to some install-instructions on the web... livecd linux # cp arch/i386/boot/bzImage /boot/kernel ....then install bootloader (which defaults to vmlinuz or bzImage usually...) Nov 03 15:37:29 you're done :-} Nov 03 15:54:17 anyone familiar with arm Alignment Trap crashes? Nov 03 15:54:45 a little, yes Nov 03 15:55:21 running eg connman, any time it tries to open a nonexisting file, I get a segfault and an Alignment Trap warning in dmesg Nov 03 15:55:47 Program received signal SIGSEGV, Segmentation fault. Nov 03 15:55:47 0x0001b72c in __connman_rfkill_init () at src/rfkill.c:112 Nov 03 15:55:48 112 src/rfkill.c: No such file or directory. Nov 03 15:56:13 which is this code: http://git.kernel.org/?p=network/connman/connman.git;a=blob;f=src/rfkill.c;h=ad624ffc6bcfb43ad7a4980d5fbea214de1faad5;hb=HEAD#l112 Nov 03 15:56:27 oh yeah, this is the one you mentioned in #openmoko-cdevel earlier Nov 03 15:56:35 right Nov 03 15:56:42 and at the same time I get : [ 4558.775000] Alignment trap: connmand (7401) PC=0x0001b72c Instr=0x2800fbd1 Address=0xffffffff FSR 0x813 Nov 03 15:56:49 the "alignment trap" part is somewhat misleading, it's just a regular segfault Nov 03 15:56:58 glibc issue, or connman issue? Nov 03 15:57:14 connman, probably Nov 03 15:57:18 connman does check for an error and returns, so it shouldn't segfault... Nov 03 15:57:37 oh yes, so it does. that connman code looks reasonable enough. Nov 03 15:57:56 if you have a copy of the (preferably unstripped) connman binary on hand, I will have a look at it Nov 03 15:58:18 mmh. ipkg file with -dbk .ikg file? Nov 03 15:58:31 ideally just the binary from the WORKDIR Nov 03 15:59:50 ok, let me check Nov 03 16:00:01 ipkg is here BTW http://sspaeth.de/uploads/tmp/connman_0.46-r1.4_armv4t.ipk Nov 03 16:00:06 as well as a dbg pkg Nov 03 16:00:12 let me check for the binary Nov 03 16:01:52 oh, this is another pthread package Nov 03 16:02:00 you probably have the same bug as mickey does with his python Nov 03 16:02:18 see above Nov 03 16:02:23 mmh, I remember what you talked about earlier Nov 03 16:02:40 that's a good hint. Thanks will follow that one Nov 03 16:02:57 recompiling the binary (it was rm'd) Nov 03 16:03:08 righto Nov 03 16:03:28 I don't think the binary will yield any useful information until your libpthread is fixed, so you might as well look at that first. Nov 03 16:03:40 hmm.. West US coast should wake up Nov 03 16:04:05 pb__: ok, thanks anyway Nov 03 16:04:28 I'm reasonably optimistic that fixing the pthread thing will clear up your crash, but if it doesn't then feel free to ask again. Nov 03 16:06:16 connmand binary is here in any case: http://sspaeth.de/uploads/tmp/connmand Nov 03 16:06:21 hi kergoth, kergoth_ Nov 03 16:06:24 I'll look into pthread next though Nov 03 16:06:37 righto Nov 03 16:06:39 there is a compile option for --disable-threads Nov 03 16:06:51 might check whether that solves it. would point to pthreads Nov 03 16:06:57 yes, worth a try. Nov 03 16:07:09 that'd be quicker than recompiling glibc, anyway Nov 03 16:16:42 ok, need to leave for now, thanks for the pointers and help so far. Will report back Nov 03 16:22:52 03Jeremy Lainé  07org.openembedded.dev * rd62c9c5244 10openembedded.git/recipes/ntp/ (ntp-4.2.4p7/ntpdate ntp_4.2.4p7.bb): Nov 03 16:22:52 ntp-4.2.4p7: invoke ntpdate using if-up.d, not an init script Nov 03 16:22:52 * provide an if-up.d script for ntpdate (based on Debian's) Nov 03 16:22:52 * remove invocation of update-rc.d busybox-cron, this should Nov 03 16:22:52 definitely not be part of the ntpdate package Nov 03 16:41:02 jo cbrake Nov 03 16:41:37 03Henning Heinold  07org.openembedded.dev * rd450fec33a 10openembedded.git/classes/cmake.bbclass: cmake.bbclass: use staging from autotools too Nov 03 16:44:12 jo gnutoo Nov 03 16:44:18 ~seen crofton Nov 03 16:44:32 crofton is currently on #oe (1d 20h 58m 26s) #edev (1d 20h 58m 26s). Has said a total of 2 messages. Is idling for 18h 2m 38s, last said: 'they are a frakking nuisance :)'. Nov 03 16:44:35 hi Nov 03 16:44:47 wifi problems(that's why I was disconnected) Nov 03 16:44:59 maybe I should change the wifi card for an ath9k? Nov 03 16:45:09 s/for/of the router for Nov 03 16:45:17 hm or ath5k Nov 03 16:45:24 what you have now? Nov 03 16:45:33 b43+free firmware Nov 03 16:45:47 or maybe it's the heat Nov 03 16:45:50 could be Nov 03 16:46:09 maybe I should change for the following setup: Nov 03 16:46:20 sheevaplug->ip04 Nov 03 16:46:31 sheevaplug->wrt54gsv4 Nov 03 16:47:20 hm b43 suckz Nov 03 16:47:22 a power quick would be expensive no? Nov 03 16:47:24 ok Nov 03 16:47:42 I went from rt2500 to rt61 to b43 Nov 03 16:49:08 rt2500 seemed worse than b43 Nov 03 16:49:19 hm Nov 03 16:50:20 zd1211rw is working fine Nov 03 16:50:25 athk's too Nov 03 16:50:33 xd1211rw is an usb stick? Nov 03 16:50:46 but maybe it's not the wifi card Nov 03 16:50:58 or maybe it was my laptop's ath5k Nov 03 16:51:01 anyway... Nov 03 16:51:11 yeah zyxel usbsticks Nov 03 16:51:33 I've an htcdrem....what would be the most urgent thing to do with it between *full SHR support *finish to debug the alsa driver to make it work *create a pulseaudio plugin for the DSP Nov 03 16:51:37 ok Nov 03 16:51:39 I've one Nov 03 16:51:57 I didn't know it could do ap Nov 03 16:52:04 hm so you read all the patches from pavel machek? Nov 03 16:52:43 some patches...the refresh one Nov 03 16:52:48 if it was from pavel Nov 03 16:54:53 I've read an enormous thread with pavel Nov 03 16:55:16 what I wonder is if SHR support will brings more dev people Nov 03 16:55:57 nope Nov 03 16:56:05 htc dream is to old Nov 03 16:56:18 you have luck that pavel isnt more a suse Nov 03 16:56:20 at Nov 03 16:56:32 ok Nov 03 16:56:40 yes it's old but: Nov 03 16:56:45 *easily rootable Nov 03 16:56:57 *sdcard(unlike palm pre) Nov 03 16:57:11 *kernel driver are free software Nov 03 16:58:52 but 320x480 screen Nov 03 16:59:07 jo ph5 Nov 03 16:59:07 I bet it's enough for wesnoth (300x400 minimum) Nov 03 16:59:47 I'd like to finish the alsa support in order to be able to do oe things... Nov 03 17:03:53 btw you said the htcdream was too old...what better devices are there out there? Nov 03 17:04:07 the pandora is not shipping yet Nov 03 17:04:12 the palm pre has no sdcard Nov 03 17:04:24 google g1 Nov 03 17:04:24 evening woglinde, Gnutoo Nov 03 17:04:28 iphone Nov 03 17:04:32 Gnutoo: will pandora ever ship? Nov 03 17:04:37 ph5 yes Nov 03 17:04:39 lol don't know Nov 03 17:04:44 g1 is htcdream Nov 03 17:04:55 iphone port has stalled last time I looked Nov 03 17:05:05 gnutoo hms oh right I thought it was older one Nov 03 17:05:08 sorry Nov 03 17:05:17 it is Nov 03 17:05:23 it's the older android one Nov 03 17:05:32 g1 = htcdream Nov 03 17:05:34 yes yes Nov 03 17:05:36 sorry Nov 03 17:05:42 okay that isnt that old Nov 03 17:06:15 palm pre is very nice...but without sdcard how would I put my music on it? Nov 03 17:06:55 I bet nokia n900 is too difficult to free Nov 03 17:07:04 hm Nov 03 17:07:13 yes some parts are closed Nov 03 17:07:48 in the kenrel? Nov 03 17:07:52 or only in userland? Nov 03 17:08:15 because in the case of the htcdream it's only in userland Nov 03 17:08:25 but everything is non-standard,that's the drawback Nov 03 17:08:59 the good is that there are repacements for closed parts...like sound and modem... Nov 03 17:09:05 gnutoo mainly userland Nov 03 17:09:09 ok Nov 03 17:09:11 powermanagement Nov 03 17:09:12 interesting... Nov 03 17:09:15 ouch... Nov 03 17:09:42 so they still closed pm this time....:( Nov 03 17:10:03 if it was only telephony it could be ok...but...pm is an important part Nov 03 17:10:41 hm pm was closes previous versions too Nov 03 17:11:44 indeed Nov 03 17:18:30 btw what would be the performance cost of a pulseaudio dsp plugin in a 528Mhz arm11 without vfp and 192M of ram? Nov 03 17:20:04 sorry dont know Nov 03 17:20:07 ok Nov 03 17:20:08 np Nov 03 17:20:12 try it Nov 03 17:20:14 and tell us Nov 03 17:20:15 but do some machine use that? Nov 03 17:20:22 I've to write the plugin first Nov 03 17:20:30 hm there is pulseaudio for the freerunner Nov 03 17:20:38 there was at the beginning Nov 03 17:20:52 but the freerunner is an armv4 with 128M of ram Nov 03 17:21:04 maybe I could try on the bug device Nov 03 17:21:32 but bug device has vfp Nov 03 17:21:39 maybe I could disable it Nov 03 17:21:46 in the kernel Nov 03 17:21:47 and try Nov 03 17:21:55 because the alsa driver seem so buggy Nov 03 17:22:06 so I wondered if a pulseaudio plugin was better Nov 03 17:22:38 hm if I remeber correctly pulse will not work without alsa Nov 03 17:22:39 because even if I make the alsa driver work for other things than aplay I wonder if it will be usefull Nov 03 17:23:13 the g1 has a custom sound system Nov 03 17:23:15 it's a dsp Nov 03 17:23:29 omap3 has dsp too Nov 03 17:23:35 n900 is using gstreamer Nov 03 17:23:39 as n8xx too Nov 03 17:24:07 ok Nov 03 17:24:17 it's for that kind of devices that I was asking Nov 03 17:24:39 so gstreamer and no pulse Nov 03 17:24:39 mmm Nov 03 17:24:40 but maybee I am mixing something Nov 03 17:25:13 maybe I should ask in the beagleboard channel then Nov 03 17:25:22 or look at beagle board web pages Nov 03 17:25:58 jo hrw Nov 03 17:26:02 hm Nov 03 17:26:04 and gone Nov 03 17:26:24 hey - is there some way to make the gettext .ipk not so huge? Nov 03 17:26:34 alsa-utils depend on it but it's mostly developer crap in /usr/share/gettext Nov 03 17:27:00 it would be good to split out libintl from the rest of gettext Nov 03 17:27:53 dhd yes Nov 03 17:28:04 look into my gettext-branch Nov 03 17:28:12 I started it there Nov 03 17:28:22 but didnt find the time to finish it Nov 03 17:28:32 thanks! actually i suppose i could just compile alsa utils without i18n, too... Nov 03 17:28:53 angstorem and micro is using proxylibintl anyway Nov 03 17:29:00 hmm Nov 03 17:29:11 i have a custom distribution based on stable/2009 ... Nov 03 17:29:21 hm oh Nov 03 17:29:28 dont know if it is there Nov 03 17:29:31 thanks! Nov 03 17:29:38 maybee you can switch to proxylibintl too Nov 03 17:29:54 yeah proxy-libintl seems like a very good idea Nov 03 17:30:15 i just have to switch teh virtual libintl provider i guess Nov 03 17:30:16 cool Nov 03 17:30:32 jupp Nov 03 17:31:14 03Koen Kooi  07rpurdie/work-in-progress * rb2010b603e 10openembedded.git/recipes/abiword/ (abiword-2.5.inc abiword_2.8.0.bb): abiword 2.8.0: band-aid packaging before overhauling it Nov 03 17:31:14 03Koen Kooi  07rpurdie/work-in-progress * r703b3f6dc4 10openembedded.git/conf/distro/angstrom-2008.1.conf: angstrom: bump DISTRO_PR and some cleanups Nov 03 17:31:16 03Koen Kooi  07rpurdie/work-in-progress * r5da5ee1a57 10openembedded.git/ (conf/checksums.ini recipes/abiword/abiword_2.8.0.bb): abiword: replace 2.8.0 with 2.8.1, some crashes fixed and libtool macros updated Nov 03 17:31:17 03Marcin Juszkiewicz  07rpurdie/work-in-progress * r04bf7d5d15 10openembedded.git/recipes/u-boot/ (u-boot.inc u-boot_git.bb): Nov 03 17:31:19 u-boot: added Sheevaplug support Nov 03 17:31:21 Marvell devices needs u-boot.kwb instead of u-boot.bin due to header for Nov 03 17:31:23 first stage bootloader. UBOOT_BINARY variable was introduced due of Nov 03 17:31:25 that. Nov 03 17:31:27 03OE Builder  07rpurdie/work-in-progress * r4ab9ea9469 10openembedded.git/recipes/linux/ (3 files in 2 dirs): linux-2.6.31: add driver for boc01 buttons Nov 03 17:31:30 03Sebastian Spaeth  07rpurdie/work-in-progress * r11aa8fff67 10openembedded.git/conf/distro/include/sane-srcrevs.inc: Nov 03 17:31:33 sane-srcrevs: add webkit-efl|gtk Nov 03 17:31:35 Signed-off-by: Sebastian Spaeth Nov 03 17:31:47 morning all Nov 03 17:31:48 03Frans Meulenbroeks  07rpurdie/work-in-progress * r056193c663 10openembedded.git/classes/package.bbclass: Nov 03 17:31:51 he thesing Nov 03 17:31:51 Revert "package.bbclass: Cleanup getVar syntax to use True, not 1 as third argument (from Poky)" Nov 03 17:31:53 This reverts commit fcee557e1b0551d1743bbb3ad61102b9ccc4cba8. Nov 03 17:31:55 03OE Builder  07rpurdie/work-in-progress * r050c77a7fe 10openembedded.git/recipes/linux/linux-2.6.31/boc01/012-091019-capsense.patch: linux-2.6.31: make capsense keys remapable for boc01 Nov 03 17:31:58 03OE Builder  07rpurdie/work-in-progress * r7cfed4fa37 10openembedded.git/recipes/linux/linux-2.6.31/boc01/014-091030-buttons.patch: linux-2.6.31: add support for multiple buttons on boc01 Nov 03 17:32:15 03OE Builder  07rpurdie/work-in-progress * r27009b0a42 10openembedded.git/recipes/linux/linux-2.6.31/boc01/011-091028-gpio.patch: linux-2.6.31: improve GPIO support for boc01 Nov 03 17:32:15 03Frans Meulenbroeks  07rpurdie/work-in-progress * r8b32d20a07 10openembedded.git/: Merge branch 'org.openembedded.dev' of git.openembedded.org:openembedded into org.openembedded.dev Nov 03 17:32:15 03Frans Meulenbroeks  07rpurdie/work-in-progress * r138f792724 10openembedded.git/classes/package.bbclass: Nov 03 17:32:18 Revert "Revert "package.bbclass: Cleanup getVar syntax to use True, not 1 as third argument (from Poky)"" Nov 03 17:32:20 Oops, accidently reverted a patch, reverting the revert. Sorry for any inconvenience Nov 03 17:32:22 This reverts commit 056193c6638892ddde7bddbfc6f6f06a2288ba64. Nov 03 17:32:26 03Khem Raj  07rpurdie/work-in-progress * r9458828947 10openembedded.git/recipes/eglibc/ (4 files in 2 dirs): Nov 03 17:32:37 eglibc: Backport IO-acquire-lock-fix.patch Nov 03 17:32:39 * Move the svn recipe to latest git revision. Nov 03 17:32:41 Signed-off-by: Khem Raj Nov 03 17:32:43 03Frans Meulenbroeks  07rpurdie/work-in-progress * r9fd25648fc 10openembedded.git/ (4 files in 4 dirs): linux-kirkwood: updated defconfigs (more modules, more filesystems in kernel) Nov 03 17:32:46 03Leon Woestenberg  07rpurdie/work-in-progress * r91403056b1 10openembedded.git/recipes/linux/ (3 files in 2 dirs): Nov 03 17:32:49 linux-kirkwood-2.6.31: Add PCI Express support for openrd-base. Nov 03 17:32:51 Note there are upstream problems involving devices using PCI IOMEM. Nov 03 17:32:55 Signed-off-by: Leon Woestenberg Nov 03 17:32:57 (84 lines omitted) Nov 03 17:32:59 03Thomas Kunze  07rpurdie/work-in-progress * r0e9b71b72a 10openembedded.git/conf/machine/include/tune-strongarm.inc: tune-strongarm.inc: add EXTRA_FEED_ARCHS Nov 03 17:33:10 03Matthew Dombroski  07rpurdie/work-in-progress * rc510866574 10openembedded.git/classes/qmake2.bbclass: Nov 03 17:33:13 qmake2.bbclass: Add lupdate and lrelease. Nov 03 17:33:15 http://patchwork.openembedded.org/patch/987 Nov 03 17:33:17 Signed-off-by: Matthew Dombroski Nov 03 17:33:19 Acked-by: Holger Hans Peter Freyther Nov 03 17:33:19 wahhaha massflood Nov 03 17:33:21 03Sebastian Spaeth  07rpurdie/work-in-progress * rf5701cca05 10openembedded.git/conf/checksums.ini: Nov 03 17:33:27 because of branches merging Nov 03 17:33:28 connman: add checksums for 0.46 Nov 03 17:33:29 crap.. another package need rework Nov 03 17:33:30 Signed-off-by: Sebastian Spaeth Nov 03 17:33:32 03Ricardo Salveti  07rpurdie/work-in-progress * r14caede54c 10openembedded.git/ (conf/checksums.ini recipes/eject/eject_2.1.5.bb): Nov 03 17:33:35 eject-2.1.5: adding HOMEPAGE and fixing SRC_URI Nov 03 17:33:45 As geocities is now down, changing the SRC_URI to point to OE Nov 03 17:33:47 mirror. Nov 03 17:33:49 Signed-off-by: Ricardo Salveti Nov 03 17:33:51 Acked-by: Holger Hans Peter Freyther Nov 03 17:33:53 03Holger Hans Peter Freyther  07rpurdie/work-in-progress * rc724ad44d6 10openembedded.git/classes/base.bbclass: Nov 03 17:33:56 base.bbclass: Better error message in case the Config Parser fails Nov 03 17:33:58 http://patchwork.openembedded.org/patch/1088/ Nov 03 17:34:02 Tested-by: Ed Nelson Nov 03 17:34:04 03Richard Purdie  07rpurdie/work-in-progress * r3d1a3ae927 10openembedded.git/classes/package.bbclass: Nov 03 17:34:07 package.bbclass: Move variable checks to one place making the code more readable (from Poky) Nov 03 17:34:09 Signed-off-by: Richard Purdie Nov 03 17:34:13 (66 lines omitted) Nov 03 17:36:59 RP: for mvl6, just did a variation on IMAGE_FEATURES which has packages listed in IMAGE_FEATURE_ and has a flag IMAGE_FEATURE__OPTIONAL (which puts them into _ATTEMPTONLY). that seem reasonable to you? I think it's cleaner than base_contains calls, and lets the user/distro easily define new package groups from the configuration metadata Nov 03 17:37:42 kergoth: so IMAGE_FEATURE_debug etc? Nov 03 17:37:47 yep Nov 03 17:38:11 IMAGE_FEATURES = "base dbg vmlinux" from local.conf, or IMAGE_FEATURES_append = " doc dev dbg" or whatever.. Nov 03 17:38:12 nice Nov 03 17:38:33 kergoth: It sounds reasonable Nov 03 17:38:45 k, i'll see about getting it pushed to oe Nov 03 17:38:50 * kergoth has a horrible backlog again Nov 03 17:40:26 kergoth: I know the feeling :/ Nov 03 17:40:48 kergoth: That pstage fetching code would be good ;-) Nov 03 17:41:07 * kergoth adds to his todo. Nov 03 17:42:02 Those CIA messages look very confused :/ Nov 03 17:42:24 yeah, it is a real pain that cia doesn't understand different branches. Nov 03 17:43:04 Its more our cia script on the git server I suspect Nov 03 17:43:16 effect is the same Nov 03 17:43:17 maybe one of our git-masters could just filter out pushes to anything other than org.openembedded.dev before sending them to cia. Nov 03 17:43:26 .dev and stable/2009? Nov 03 17:43:27 well, it knows the branch the commits are associated with, but sadly it can get awfully spammy when merges happen, or rebases even.. itd be nice to somehow change the notification to only notify on commit objects that were uploaded this time, and as such were never in the repository before then Nov 03 17:43:39 ah, yeah, filtering would be easy enough Nov 03 17:43:44 The above was a rebase Nov 03 17:43:47 kergoth: I dunno, I think what I care about more is when they land in the trunk. Nov 03 17:43:53 * kergoth nods Nov 03 17:44:07 rather than when they were first committed on kergoth/crazy-hacking-branch which I will never use. Nov 03 17:44:16 good point Nov 03 17:44:25 pb__: right Nov 03 17:44:33 * RP wonders who has admin access these days Nov 03 17:44:39 point for oedem? Nov 03 17:44:44 I think cbrake is the main man. Nov 03 17:44:46 hacking session? Nov 03 17:44:48 haha Nov 03 17:46:03 Yes, a topic for oedem. I did have admin access at some point Nov 03 17:46:21 HDD connected by serialata to desktop: 100-110MB/s (hdparm -t), same connected by usb to sheevaplug: 30MB/s Nov 03 17:46:39 hrw uh :( Nov 03 17:46:50 hrw hm is it the usb chip or the driver Nov 03 17:47:00 i could see enforcing a branch layout, make per user branches go in user// and filter all of those out, so we woudln't have to manually maintain the list of branches to notify on, someday Nov 03 17:47:01 woglinde: it is fast enough still ;) Nov 03 17:47:02 I wonder if I can make the new laptop usable by Friday... Nov 03 17:47:05 30MB/s is probably about as much as you could expect from usb. Nov 03 17:47:09 it just isn't a very fast bus. Nov 03 17:47:22 pb__: exactly Nov 03 17:47:26 hm right Nov 03 17:47:32 kergoth: That would be rather easy to arrange - topic for OEDEM? Nov 03 17:47:32 usb3 will be higher Nov 03 17:47:36 RP: rsync oldlaptop: newlaptop: Nov 03 17:47:41 sounds like a good idea Nov 03 17:47:46 rp install karmicle? Nov 03 17:47:47 hrw: and keep all that junk? :) Nov 03 17:47:50 that discussion would be a good time to consider renaming dev to master Nov 03 17:47:54 we aren't on monotone anymore.. Nov 03 17:47:58 RP: laptop will be usable for oedem ;d Nov 03 17:48:11 kergoth: get MV to send you to oedem Nov 03 17:48:16 hrw: I could just have transferred the hdd for that Nov 03 17:48:27 kergoth: heh, I'm not really a fan of renaming branches just to suit the fashion of today's scm. Nov 03 17:48:28 hrw: hopefully they'll be less annoying about travel next year Nov 03 17:48:48 RP: ah right, you use newer laptops ;D Nov 03 17:48:49 pb__: you mean like we did when we switched to monotone? Nov 03 17:48:58 if you're going to go with an scm agnostic branch policy, then stick to it Nov 03 17:49:03 kergoth: right, I wasn't a fan of that either Nov 03 17:49:58 "that", in this case, being both monotone and the change in naming policy :-} Nov 03 17:50:12 * kergoth nods Nov 03 17:50:38 03Marcin Juszkiewicz  07org.openembedded.dev * r2d81468a71 10openembedded.git/recipes/usbutils/usbutils_0.86.bb: usbutils: added 0.86 Nov 03 17:50:40 03Marcin Juszkiewicz  07org.openembedded.dev * rea6f6abebd 10openembedded.git/recipes/openvpn/openvpn_2.0.9+2.1rc20.bb: openvpn: added 2.1rc20 and enabled reading passwords from file Nov 03 17:50:50 JaMa: add mesa-xlib 7.4 please. I did build of x11-image for BUG with it on your branch Nov 03 17:50:50 kergoth: I added both items to the agenda so perhaps we'll remember to discuss them :) Nov 03 17:52:20 * woglinde wonders if we could setup videostuff for kergoth Nov 03 17:52:54 I don't think we have the technology to do actual videoconferencing, but we can certainly do audio if he wants to dial in. Nov 03 17:52:56 so kergoth could attend afternoon sessions Nov 03 17:53:22 RP: can we move last both entries on sunday? they are interesting but I will have to leave earlier ;( Nov 03 17:53:45 hrw when you think we should leave? Nov 03 17:53:51 2 p.m.? Nov 03 17:54:08 woglinde: 18:45 plane so 17:10 train I think Nov 03 17:54:24 hm oh isnt that a bit risky= Nov 03 17:54:25 hrw: yes, that was just where there was space on the agenda Nov 03 17:54:26 woglinde: this gives us 1h on airport Nov 03 17:54:43 stansted, right? yeah, 1h should be fine if you don't need to check any bags. Nov 03 17:55:05 we have only paid for one knapsack Nov 03 17:55:06 pb__: I do not need to checkin, do not know how Henning or Robert Nov 03 17:55:08 *g* Nov 03 17:55:14 if you do have hold luggage then you might do better to take the next earlier train for extra safety Nov 03 17:55:33 robert will be travel one week around in england Nov 03 17:55:36 ok Nov 03 17:55:43 so its only us Nov 03 17:55:44 though, on sunday evening, I expect you would be fine with only 1h anyway, stn isn't usually all that busy at that time Nov 03 17:55:57 woglinde: so it makes us both and Marco iirc Nov 03 17:55:59 hrw: I added "discuss agenda" at the start yesterday btw - I epxect we can rearrange things then as needed Nov 03 17:56:06 hrw ah okay Nov 03 17:56:12 RP: cool Nov 03 17:56:33 XorA|gone: get better soon - I want to discuss packaging of transmission Nov 03 18:13:15 anyone using mailx/liblockfile? Nov 03 18:15:09 03Marcin Juszkiewicz  07org.openembedded.dev * rfa73b45447 10openembedded.git/conf/checksums.ini: checksums.ini: added usbutils and openvpn Nov 03 18:27:18 re Nov 03 18:29:18 found what is wrong with mysql_config the recipe says in stage_append: Nov 03 18:29:29 sed -i -es,^pkglibdir=\'/usr/lib/mysql\',pkglibdir=\'\', ${STAGING_BINDIR_CROSS}/mysql_config Nov 03 18:29:47 if I understand it properly it replaces pkglibdir with an empty string Nov 03 18:29:53 that is not sound Nov 03 18:30:31 I would expect STAGING_LIBDIR or so Nov 03 18:31:46 RP: the issue with pstage fetching is that currently i have it fetching it whenever it doesn't have it.. but of course, that means a -c clean and rebuild will still result in it using the prebuilt package.. any suggestions on a good mechanism to force the rebuild? right now we just have a variable that inhibits pstage usage in general, and use that to force it when necessary, but i'm not sure if the clean not resulting in a fresh build beha Nov 03 18:31:47 intuitive Nov 03 18:38:44 bye Nov 03 18:39:07 kergoth: how about "bitbake -c buildlocally foo" ? Nov 03 18:40:03 heh, hmm Nov 03 18:41:20 most of the time people will want to fetch packages instead of building them Nov 03 18:43:31 true Nov 03 18:49:23 03Klaus Kurzmann  07shr/import * re596ff6505 10openembedded.git/recipes/freesmartphone/ (7 files): Nov 03 18:49:23 recipes/freesmartphone: sync with shr/merge Nov 03 18:49:23 Signed-off-by: Klaus Kurzmann Nov 03 18:49:33 03Klaus Kurzmann  07shr/import * r2a9665afc7 10openembedded.git/conf/distro/include/shr-autorev-unstable.inc: Nov 03 18:49:33 shr-autorev-unstable.inc: add libfsoresource and set to AUTOREV Nov 03 18:49:33 Signed-off-by: Klaus Kurzmann Nov 03 18:52:11 why does openembedded.org to go the linux to go homepage, rather than to www.openembedded.org? Nov 03 18:53:29 probably just an apache misconfiguration Nov 03 18:53:43 yep, that's something I also noticed yesterday heh Nov 03 18:53:45 quite weird Nov 03 18:53:52 ltg is the default vhost, and I guess it isn't set up to recognise openembedded.org without the www. prefix Nov 03 18:54:03 probably Nov 03 18:55:45 ah, actually, no, it's not that Nov 03 18:56:31 the dns record for openembedded.org points to ltg, but www.oe.org is actually on amethyst. Nov 03 18:57:27 ah Nov 03 18:59:43 so, the right answer to that probably depends on the outcome of the "get services off amethyst" session at oedem Nov 03 18:59:52 if www ends up moving to ltg then it's easy Nov 03 19:00:25 if www goes somewhere else then it is very slightly more complicated Nov 03 19:34:52 re Nov 03 20:01:54 he crofton Nov 03 20:02:03 hi Nov 03 20:02:15 waiting for the first plane ... Nov 03 20:02:28 crofton for boostversion 1.4.0 it is possible to build it with cmake Nov 03 20:02:45 this works around the pain of bjam Nov 03 20:02:46 interesting Nov 03 20:02:49 yes Nov 03 20:02:54 I have a working version Nov 03 20:02:59 does 1.40 have better arm support? Nov 03 20:03:05 cool Nov 03 20:03:16 args compiling Nov 03 20:03:24 didnt actual test it on device Nov 03 20:03:35 but some apps like povray needs some patching Nov 03 20:03:48 didnt look at the arm support yet Nov 03 20:03:56 yeah, boost api is not very stable Nov 03 20:04:07 ah Nov 03 20:04:07 hm its only header stuff Nov 03 20:04:12 looks like we have a plane Nov 03 20:04:29 and adding some std:: before exception Nov 03 20:04:33 Crofton: that'll save you some arm waving then Nov 03 20:04:34 at least for povray Nov 03 20:04:55 har har Nov 03 20:05:04 ok, I need to see how this impacts gnuradio Nov 03 20:05:09 crofton I will check in it with DEFAULT_PREF = "-1" Nov 03 20:05:11 thanks for the heads up woglinde Nov 03 20:05:12 than you can test Nov 03 20:05:13 awesome Nov 03 20:05:15 yeah Nov 03 20:06:34 packing needs some tweaking Nov 03 20:06:37 too Nov 03 20:06:48 but that could be done in later step Nov 03 20:07:23 ok Nov 03 20:07:26 gotta run Nov 03 20:07:29 have a nice flight Nov 03 20:07:30 yes Nov 03 20:07:34 I will see you this weekend? Nov 03 20:07:57 poof Nov 03 20:07:58 yes Nov 03 20:08:08 hms to late Nov 03 20:08:23 hehe Crofton sounds like he was insulting people :-) Nov 03 20:08:30 03Koen Kooi  07org.openembedded.dev * ra6a93a9acb 10openembedded.git/recipes/powervr-drivers/libgles-omap3.inc: libgles-omap3: adjust for future releases Nov 03 20:08:35 hm zshell is not bad with git Nov 03 20:09:04 XorA|gone: heh, better hope he doesn't try that up in scotland Nov 03 20:09:49 I guess he doesn't want to learn about the old glasgow kiss. Nov 03 20:09:58 hehe Nov 03 20:10:51 hm let me google what it is Nov 03 20:12:20 hms ppor thesing they dont care about his collie-patches Nov 03 20:12:37 poor collie Nov 03 20:18:12 So I think we can have a GM at OEDEM if we jump through some hoops. I've mailed the membership list with details Nov 03 20:18:40 very good Nov 03 20:19:28 rp fine Nov 03 20:19:56 GM? Nov 03 20:20:08 normaly we german has such pedantic behaoviur Nov 03 20:20:11 args Nov 03 20:20:13 my english Nov 03 20:20:19 kergoth: e.V. annual General Assembly Nov 03 20:20:21 ah Nov 03 20:20:21 03Jeremy Puhlman  07org.openembedded.dev * ra9bbc07ae8 10openembedded.git/recipes/neon/ (neon_0.25.5.bb neon_0.26.0.bb): Nov 03 20:20:21 neon: Add m4 dir to autoreconf. Nov 03 20:20:21 Signed-off-by: Jeremy Puhlman Nov 03 20:20:21 Signed-off-by: Chris Larson Nov 03 20:20:33 03Chris Larson  07org.openembedded.dev * rfbccdee191 10openembedded.git/recipes/prelink/ (prelink_20061027.bb prelink_20071009.bb): Nov 03 20:20:33 prelink: pull in /etc/cron.daily/prelink in the postinst. Nov 03 20:20:33 The cron script obeys /etc/default/prelink, so this ensures that the postinst Nov 03 20:20:33 also obeys it. Also changes postrm to use the same logic as the cron script Nov 03 20:20:33 for the prelink uninstallation. Nov 03 20:20:34 Signed-off-by: Chris Larson Nov 03 20:20:35 kergoth: Needs to be held before the end of the year with six weeks notice :/ Nov 03 20:20:36 seems I can become a member soon *g* Nov 03 20:20:41 fun fun Nov 03 20:20:57 kergoth: no notice was given so its "find the loophole time" :) Nov 03 20:21:02 hehe Nov 03 20:21:13 RP if no one of the memebers objecting there is no problem Nov 03 20:21:43 woglinde: It doesn't work like that - you have to follow the rules Nov 03 20:22:08 Solution: Call an EGM, pass a resolution to relax the AGM time limits for this year, hold the AGM Nov 03 20:22:21 hah Nov 03 20:22:22 rp I know of none e.V. that was shutdow of not following the 4 or 6 weeks rule of sending the invation Nov 03 20:22:23 nice Nov 03 20:22:39 Assuming we can pass the appropriate number of voters in various cases Nov 03 20:22:55 woglinde: We've had one complaint though :( Nov 03 20:23:38 what was his real problem? Nov 03 20:24:32 re florian Nov 03 20:24:46 hm Nov 03 20:25:21 I cannt believe his problem was not getting the invation at the right time Nov 03 20:25:44 re Nov 03 20:26:14 florian: I sent an idea about how we could still hold the GA at OEDEM Nov 03 20:26:41 woglinde: Consider it a wake up call that we need to play by the rules exactly :/ Nov 03 20:26:48 rp yes Nov 03 20:27:02 we will do it better at the next meeting Nov 03 20:27:15 but for nobody wins when there is no GA Nov 03 20:27:18 RP: yep, just saw the mail - will take a look once I have finished cooking Nov 03 20:27:19 args now Nov 03 20:27:24 (yes food, bit bits!) Nov 03 20:29:02 RP: ingenious plan. I guess we need a german to confirm that it actually conforms to the statutes though. Nov 03 20:30:15 hm for this weekend I only see one solution, every member confirms that he got an invation early enough Nov 03 20:31:07 normaly workaround is, to make an extra-odinary-meeting with a smaller invation deadline Nov 03 20:31:26 that are my experience running and working with e.V. Nov 03 20:31:34 yes, if all the members could affirm that they were happy with the notice period then that would be fine. Nov 03 20:31:51 how many members were there prior to the berlin meeting? not many, so maybe that would be feasible. Nov 03 20:32:12 re gnutoo Nov 03 20:38:22 03Chris Larson  07org.openembedded.dev * r5c4f5bec44 10openembedded.git/recipes/autofs/ (autofs.inc autofs_3.1.7.bb autofs_4.0.0.bb autofs_4.1.4.bb): Nov 03 20:38:22 autofs: switch to a .inc. Nov 03 20:38:22 Signed-off-by: Chris Larson Nov 03 20:38:24 pb_: I just sent what I think is a list of members Nov 03 20:38:32 03Jeremy Puhlman  07org.openembedded.dev * r8370077581 10openembedded.git/recipes/autofs/autofs.inc: Nov 03 20:38:32 autofs: pass RANLIB into the build. Nov 03 20:38:32 Signed-off-by: Jeremy Puhlman Nov 03 20:38:37 pb_: I've asked that someone from the board confirm it though Nov 03 20:43:35 righto Nov 03 20:43:54 I guess the list on http://wiki.openembedded.org/index.php/Organization is the roll of founding members, i.e. the membership prior to the berlin meeting. Nov 03 20:44:32 I could ask robert tomorrow Nov 03 20:45:07 pb_: Right, that matches what I lifted from the minutes Nov 03 20:45:18 pb_: which I comibed with the list of members who were added Nov 03 20:45:29 right, very good Nov 03 20:45:44 03Frans Meulenbroeks  07org.openembedded.dev * rc206dabdc8 10openembedded.git/: Merge branch 'org.openembedded.dev' of git.openembedded.org:openembedded into org.openembedded.dev Nov 03 20:45:45 03Frans Meulenbroeks  07org.openembedded.dev * r1632c44d94 10openembedded.git/: Merge branch 'org.openembedded.dev' of git.openembedded.org:openembedded into org.openembedded.dev Nov 03 20:45:46 03Frans Meulenbroeks  07org.openembedded.dev * refa3f5fd40 10openembedded.git/recipes/mysql/mysql_4.1.22.bb: mysql: fixed mysql_config in staging Nov 03 20:47:10 hi mrmoku Nov 03 20:47:28 so boost now looks nice enough o check in Nov 03 20:47:54 hm okay I should look at the arm atomic support Nov 03 20:48:38 woglinde: hi Nov 03 20:48:39 hm Nov 03 20:48:41 in that case, per woglinde's plan, it doesn't seem completely hopeless to get those eleven people to waive the notice period for the GA. might be worth canvassing them, anyway. failing that, we can use the egm plan. Nov 03 20:48:54 gm Nov 03 20:49:11 latest gcc should have __sync_add_and_fetch for arm now Nov 03 20:49:48 pb_: We already have an email from Leon saying he didn't realise that was the plan? Nov 03 20:50:03 hi likewise Nov 03 20:50:10 hi fooks. Nov 03 20:50:30 hi likewise Nov 03 20:50:56 RP: hi, I see you just mentioned me, wassup? Nov 03 20:51:07 ah it's regarding the ev/OEDEM topic? Nov 03 20:51:51 likewise: yes, just talking about whether OEDEM can be a GA or not Nov 03 20:52:44 likewise: pb_ was thinking if the original members agreed that it had been public knowledge then there would not be a problem. Your email earlier says otherwise though Nov 03 20:52:51 but there is a plan b Nov 03 20:52:52 ok, I do not want to prevent anything from happening, it's just I didn't learn about the plans (my bad I presume). Nov 03 20:52:59 RP: no, that's not quite what I was thinking. Nov 03 20:53:45 pb_: No, rereading what you said its not :/ Nov 03 20:54:05 pb_: likewise has it in scrollback anyway Nov 03 20:54:43 clearly it is not the case that all the original members received adequate notice, so I don't think we can pursue that line of enquiry. but, if all the original members were to agree that, despite lack of notice, they were still happy for the GA to go ahead, then that would effectively remove the problem. Nov 03 20:55:38 Even if notice had been served at the last EGM, we'd still fall foul of the six week rule :/ Nov 03 20:55:52 there are always people who don't read their emails, snailmail, etc. It doesn't mean anything strictly speaking. Nov 03 20:55:54 obviously this is not in conformance with the letter of the statutes, but I imagine that german law follows the basic principle that the statutes exist to serve the members and not vice versa. Nov 03 20:56:26 pb_: its another idea certainly Nov 03 20:56:40 if all of the founding members (i.e. all the members who were entitled to receive six weeks notice) are prepared to state that they don't mind not having received that notice, then it seems vanishingly unlikely that the german authorities would have a problem with such a GA. Nov 03 20:57:08 after all, the real purpose of the notice period is to ensure that all members are alerted to it and have an opportunity to attend. Nov 03 20:58:55 pb_: Right. Equally if you can get >50% of the voting rights (some via proxy or otherwise) in one place I suspect there is a lot you can do... Nov 03 20:59:04 right, indeed Nov 03 20:59:34 *g* Nov 03 20:59:46 anyway, I guess florian or someone else from the board will have to sort this out in the end. Nov 03 20:59:52 I will stop being an armchair lawyer now :-) Nov 03 21:00:00 I never saw a german authorities at my e.V. GA's Nov 03 21:00:12 pb_: yes, I've put the idea forward, lets see whats said Nov 03 21:00:27 pb_: Presumably you'll mention yours too... Nov 03 21:00:41 03Chris Larson  07org.openembedded.dev * rf1d0fab3ac 10openembedded.git/recipes/opkg/ (opkg.inc opkg/isatty.patch): Nov 03 21:00:41 opkg: Don't prompt for user input from stdin if it's not a tty. Nov 03 21:00:41 Also fixes a minor memory leak while we're at it, as the code was Nov 03 21:00:41 strdup'ing the malloc'd string read from stdin without ever Nov 03 21:00:41 freeing the original. Nov 03 21:00:43 Signed-off-by: Chris Larson Nov 03 21:00:51 yeah, I will send a note to the list Nov 03 21:00:51 * florian is back Nov 03 21:01:06 * XorA|gone has already organised his proxy for OEDEM Nov 03 21:01:54 hm __ARM_ARCH_7A__ is beagle board right? Nov 03 21:02:31 woglinde: yes Nov 03 21:02:54 ah, no need, florian is here :-) Nov 03 21:03:01 pb_, woglinde: I didn't have much time lately, but what I managed to find out is that even if we violate a rule in our statutes the GA itself is valid. Nov 03 21:03:06 okay will test boost 1.4.0 for it now Nov 03 21:03:13 if course another couse of action is to hold the meeting then call for an EGM at FOSDEM and first article is to ratify decisions made at OEDEM Nov 03 21:03:48 XorA|gone: if you're going to do that then you might as well ratify the decisions by internet vote. Nov 03 21:03:55 pb_: or that Nov 03 21:04:15 pb_: but we do need a new board as members have resigned Nov 03 21:04:45 florian: right, that's good to know. Nov 03 21:05:07 Someone *could* decide to go to court and fight a decision but its very unlikely to succeed. Nov 03 21:05:16 florian yupp Nov 03 21:05:55 yeah, that is what I figured. Nov 03 21:05:56 only think that he would achieve is that the GA's wasnt valid Nov 03 21:06:21 obviously I don't have any experience of german law, but in the UK the aggrieved party would need to demonstrate something beyond a technical breach of the statutes. Nov 03 21:06:51 pb but as I said, I never heard of such a case Nov 03 21:06:54 We could have changed the statutes... there does not seem to be a law that says how much time we need to announce the GA an advance. I have seen two weeks several times e.g. Nov 03 21:07:21 florian I mostly saw 4 weeks Nov 03 21:07:38 woglinde: right, the nature of e.V.'s seems to make this unlikely. I doubt many people will take them seriously enough to go to court. Nov 03 21:07:49 right Nov 03 21:07:58 pb hm depends Nov 03 21:07:58 unless there is a large amount of money at stake, I would expect most folks would just resign from the e.V. and leave it at that. Nov 03 21:08:06 when big money is in the play Nov 03 21:08:18 * kergoth tries to figure out how to make his granular dbg packaging cooperate with do_split_packages Nov 03 21:08:25 we have an animal rights e.V. which hosted some million of funds Nov 03 21:08:34 sure, but that clearly is not the case here. oe e.v. has, what, about 1500 EUR in the whole world? Nov 03 21:08:36 and the president took some of the millions Nov 03 21:08:44 but thats not our case Nov 03 21:08:45 *g* Nov 03 21:08:57 That's somethign I can say for sure: No big money involved currently :) Nov 03 21:10:19 The bank decided to notify us by a written note for every transaction - including drawing a fee for the notification. Nov 03 21:10:27 florian: doh Nov 03 21:10:51 Need to fix this somehow :) Nov 03 21:11:00 kergoth: ah, tricky. I guess the whole do_split_packages thing could probably do with being revisited. Nov 03 21:11:07 florian hm I hope that we get enough for travel cost at some point Nov 03 21:11:37 woglinde: if you are planning to spend more than 1500 eur on travel then I think you need to rethink your plans Nov 03 21:11:48 i'd kind of like a more generic mechanism of capturing information from a build and using it in subsequent builds, really.. but i'm not sure how such a thing would work Nov 03 21:11:49 heh Nov 03 21:11:56 florian hm depends on the bank and bundesland if online bankauszug is a valid document Nov 03 21:12:48 kergoth: yah, right. sort of like a cache of previous build information. Nov 03 21:12:53 yeah Nov 03 21:12:56 pb nope I am trying got away cheap as possible Nov 03 21:13:13 yeah, that would obviously be handy for a future oecommander (!) type thing. Nov 03 21:13:21 without sleeping under the bridge Nov 03 21:13:52 not sure that I would want bitbake itself to use that kind of cached data though. Nov 03 21:13:55 woglinde: We should have printed ones, but sending one via snailmail for every transaction is useless. Nov 03 21:14:17 yes Nov 03 21:14:21 for every suckz Nov 03 21:14:39 for my basketball e.V. we get one at each month Nov 03 21:14:41 Apart from the fact that it is less than 500m to grab them... Nov 03 21:15:15 hm florian they are send centralized Nov 03 21:15:45 03Koen Kooi  07org.openembedded.dev * r1aa6d29a4e 10openembedded.git/recipes/gthumb/gthumb_2.10.11.bb: gthumb: fix packaging Nov 03 21:20:24 woglinde: quite likely yes Nov 03 21:28:15 03Martin Jansa  07shr/merge * r8012bf3ced 10openembedded.git/conf/distro/include/preferred-shr-versions.inc: Nov 03 21:28:15 preferred-shr-versions.inc: minimal version Nov 03 21:28:15 * Use latest versions available Nov 03 21:28:15 * Few packages are preferred with higher version (which wouldn't be Nov 03 21:28:15 selected by bitbake because of DEFAULT_PREFERENCE = "-1" Nov 03 21:28:17 * One package is hold back, because of compile errors here Nov 03 21:28:21 Signed-off-by: Martin Jansa Nov 03 21:30:10 good evening. i'm new to oe. i[ve setup a dev enviroment yesterday and i'm trying to build linux-rp kernel. nevertheless i face an error on qemu compilation on file x86_64-linux/qemu-native-0.10.3-r1/qemu-0.10.3/target-i386/translate.c is this a known problem, gcc returns internal error Nov 03 21:30:12 RP: Did you work for lawyers befor joining O-Hand? ;) Nov 03 21:35:14 hm oh have we still not 0.11 of qemu in oe? Nov 03 21:35:34 qsp which linux distribution you are using? Nov 03 21:36:01 opensuse 11.1, gcc 4.3.2 Nov 03 21:36:09 qsp as workaround put ASSUME_PROVIDED += "qemu-native" into your local.conf Nov 03 21:36:28 mm, thank you. its a but in my gcc? Nov 03 21:36:39 dont know Nov 03 21:36:49 maybee gcc or qemu self Nov 03 21:37:00 maybee it is fixed in 0.11 Nov 03 21:37:13 because 0.10.3 is fairly old Nov 03 21:37:15 heh, no matter what qemu's source doing, gcc shouldn't be ICE'ing :) Nov 03 21:37:23 ok, thank you, if the config.local setting does the trick then its geat Nov 03 21:37:44 well, i've had some issues with gcc before, so i also habe gcc 4.4 Nov 03 21:37:46 qsp that aussems you install the qemu stuff from suse Nov 03 21:37:57 args assumes Nov 03 21:38:44 but i think i saw somewhere that gcc 4.x wasnt ok this case.. anyway, its perfect. i'll install qemu and rerun then. thank you Nov 03 21:39:15 i'm really new. and the aagstrom kernel doesnt boot on c700. so i'm trying to build linux-rp... Nov 03 21:39:26 :-) Nov 03 21:42:59 hm as I thought armv7 has __sync_add_and_fetch with gcc-4.4 Nov 03 21:43:50 florian: no, I decided a carear in law would be boring ;-) Nov 03 21:45:37 hm I am think about study law at remote university Nov 03 21:48:19 woglinde: ignore me, I'm sure it can be interesting :) Nov 03 21:49:32 rp I could fight against gpl volations *g* Nov 03 21:50:00 woglinde: indeed :) Nov 03 21:50:23 * RP thinks about his companies legal department Nov 03 21:52:22 jo ant Nov 03 21:52:45 abend woglinde Nov 03 21:53:04 jojo Nov 03 21:53:10 Hmm, first time I've ever seen a DISTRO_PR changed. An interesting concept :/ Nov 03 21:53:10 re jay7 Nov 03 21:53:24 hm? Nov 03 21:53:32 angstroem does it all the time Nov 03 21:53:34 *g* Nov 03 21:53:51 woglinde: :/ Nov 03 21:54:01 every time I'm updating and want to test new kexecboot DISTRO_PR is bumped :) Nov 03 21:54:05 Ah well, there goes that testing Nov 03 21:55:25 need a coffee.. Nov 03 21:57:11 qsp: qemu 0.10.5+ can be compiled with gcc 4.x Nov 03 21:58:21 woglinde: __sync_add_and_fetch is a builtin isnt it Nov 03 21:58:26 mickeyl: Did you have any thoughts on that import bb issue in the end? Nov 03 21:58:37 khem not in earlier gcc's Nov 03 21:59:12 woglinde: is it in in libgcc then on older ones ? Nov 03 21:59:27 khem and not for all archs Nov 03 21:59:38 but now it seems for all in with 4.4 Nov 03 21:59:42 boost builds Nov 03 22:00:20 boost 1.4.0 Nov 03 22:00:27 will commit now Nov 03 22:01:17 woglinde: are you worried about gplness of the intrinsics ? Nov 03 22:01:35 RP: the concept of DISTRO_PR and OELAYOUT_ABI in PSTAGE_PKGPATH would need more attention... Nov 03 22:01:52 khem for boost 1.3.36 there is a patch for armv7 in Nov 03 22:03:21 ant__: Not really, if its been added in there it should be fine Nov 03 22:03:34 well, here poky and Angstrom diverge Nov 03 22:03:57 ant__: Its not really a problem Nov 03 22:04:09 it smells somehow duplicate Nov 03 22:04:14 ant__: ? Nov 03 22:04:23 I know these are two concepts Nov 03 22:04:29 woglinde: yes older arms didnt really have atomic op equivalent for this operation Nov 03 22:07:56 03Henning Heinold  07org.openembedded.dev * rc0089b4e81 10openembedded.git/ (4 files in 3 dirs): Nov 03 22:07:56 boost: add new version 1.4.0 Nov 03 22:07:56 * switch to cmake as buildsystem Nov 03 22:07:56 * enable serialization and python Nov 03 22:08:00 * make own .inc for 1.4.x versions Nov 03 22:08:02 * don't make default for now, needs more testing with Nov 03 22:08:04 apps that are using boost Nov 03 22:15:49 03Richard Purdie  07master * r40856c0b1d 10bitbake.git/lib/bb/ (build.py parse/parse_py/BBHandler.py): Nov 03 22:15:49 Improve anonymous python function handling sp functions use the methodpool and are no longer concatenated Nov 03 22:15:49 Signed-off-by: Richard Purdie Nov 03 22:16:16 re crofton Nov 03 22:16:25 flight canceld? Nov 03 22:16:25 i'm still having troubles with gcc segfaulting with internal compiler errors. should there be issues to switch to gcc4.4? or i should clean the things it already built? Nov 03 22:16:29 * XorA|gone goes insane trying to pay his local council money Nov 03 22:16:51 hm what is local council money? Nov 03 22:16:53 qsp: this is the gcc on your build machine which is ICEing Nov 03 22:16:59 if it is so hard, do pay them Nov 03 22:17:13 no, four hour layover Nov 03 22:17:16 woglinde: I owe local government money, it seems they havent actually thought about how to receive the cash Nov 03 22:17:28 kdem: yep. my gcc... i have also 4.4 so i'll give it a try.. Nov 03 22:17:49 I guess I must be a rare case in that they didnt screw up and I do actually owe the money Nov 03 22:17:50 qsp: You can create a testcase by preprocessing the .c file where it is ICEing and let SuSE know about it may be they are interested to fix it Nov 03 22:18:06 khem they will say use newer qemu Nov 03 22:18:35 oh, its now on qemu now, its in gcc-cross-intermediate Nov 03 22:18:40 *not Nov 03 22:18:43 woglinde: I dont think so if you have a testcase which shows gcc is ICEing its a different story Nov 03 22:19:22 well, i can extract the gcc command that was ran and see if i can make one (test case) Nov 03 22:19:26 xora, check my twitter Nov 03 22:19:31 unless gcc version is EOL it will be of interest to fix gcc Nov 03 22:19:58 well, yes, thats also a problem, i dont think they have gcc.3.4 in the opensuse to come.. Nov 03 22:20:26 anyone: I have a reason to branch off in GIT, to be merged later. Any simple workflow for this in OE. (I'm overwhelmed by the git-branch options). I want my branch to keep in-sync with .dev during development. Nov 03 22:20:39 03Richard Purdie  07master * rf75ac81a55 10bitbake.git/lib/bb/parse/parse_py/BBHandler.py: Nov 03 22:20:39 BBHandler.py: Use exec_func, not exec_func_python Nov 03 22:20:39 Signed-off-by: Richard Purdie Nov 03 22:20:47 so i'll just test with gcc4 and see what happens. though its annoying sometimes, these opensuse modifications, it isnt the first issue with them :-/ Nov 03 22:20:57 likewise: create a remote user branch Nov 03 22:21:03 likewise, yeah but tohard to explain b Nov 03 22:21:03 Via crackberry Nov 03 22:21:18 and maintain it as long as you develop and then merge back into .dev Nov 03 22:21:26 git reabse origin Nov 03 22:21:29 Crofton|bold: good luck being non-bored during the layover. :-) Nov 03 22:22:00 qsp: so its the gcc-native that is used here right ? Nov 03 22:22:07 khem: thanks, googling for the teminology you splashed at me. Nov 03 22:22:58 khem: yes, it segfaulted on armv5te-angstrom-linux-gnueabi/gcc-cross-intermediate-4.3.3-r7.1/gcc-4.3.3/gcc/simplify-rtx.c:3391 Nov 03 22:23:22 khem, but its while it compiles gcc, using the machine's own gcc, so its the machines gcc fault i'd guess Nov 03 22:24:35 * Crofton|bold is rsorting to drinking Nov 03 22:24:50 qsp: yes Nov 03 22:25:35 khem: there isnt a way to define the use of a different gcc exec in local.conf? i'll need to change the symlink? Nov 03 22:25:52 khem: remote user branch means: git branch --track myfeature origin/org.openembedded.dev ? Nov 03 22:26:44 dont' need --track with recent git, its default for branches based on remote branches Nov 03 22:27:19 yeah, that is my experience also Nov 03 22:27:35 Crofton|bold: awesome Nov 03 22:28:38 if I get back thereI'll get a real photo Nov 03 22:30:35 where do the .debug directories come from in the install phase and how can i prevent them from being created? Nov 03 22:30:46 why would you want to? Nov 03 22:30:51 to save space... Nov 03 22:30:56 uh, no. Nov 03 22:30:58 oh Nov 03 22:30:59 .debug files go into -dbg packages Nov 03 22:31:06 if you don't want them, don't install those packages Nov 03 22:31:22 oh, this is my own .bb file - clearly i missed something :) Nov 03 22:31:37 sounds like you overrode one of the FILES vars and did it wrong. Nov 03 22:31:39 so if i add a -dbg package to the list of packages, then bitbake will put them in there... Nov 03 22:31:42 ah Nov 03 22:31:43 no. Nov 03 22:31:47 -dbg is in the package list by default Nov 03 22:31:50 unless, again, you change it Nov 03 22:31:57 bitbake -e | grep \^PACKAGES= Nov 03 22:32:01 right, I did override PACKAGES Nov 03 22:32:09 why are you doing that? Nov 03 22:32:20 the makefiles for this package are kind of screwed up - so it is a hack job Nov 03 22:32:30 not sure anymore... okay i'll check it out Nov 03 22:32:30 what does the makefiel have to do with packaging? Nov 03 22:32:37 it's a long story Nov 03 22:32:40 PACKAGES is just how you split up what gets installed in do_install Nov 03 22:32:45 absolutely nothing to do with the makefiles Nov 03 22:32:55 likewise: git branch user/branch origin/org.openembedded.dev Nov 03 22:33:05 kergoth: thanks - okay, that makes sense then Nov 03 22:33:10 qsp: there are several ways Nov 03 22:33:30 the do_package task takes what you installed in do_install and splits it up into the packages in PACKAGES, obeying the globs in FILES_ Nov 03 22:33:34 ah okay Nov 03 22:33:38 qsp when gcc is compiled it used CC env variable to know what compiler is used to compile it Nov 03 22:34:17 likewise: then you can do git push origin user/branch to create the remote branch Nov 03 22:34:44 03Richard Purdie  071.8 * r47dceffdfc 10bitbake.git/lib/bb/build.py: Nov 03 22:34:44 build.py: Add bb and os to __builtins__ backporting from trunk since we'd like to be able to depend on this behaviour in OE soon Nov 03 22:34:44 Signed-off-by: Richard Purdie Nov 03 22:34:56 kergoth: thanks - it makes sense now Nov 03 22:34:56 03Richard Purdie  071.8 * ra865dd4263 10bitbake.git/lib/bb/fetch/svk.py: Nov 03 22:34:56 fetch/svk: Fix up breakage. Evidently nobody uses svk Nov 03 22:34:56 Signed-off-by: Richard Purdie Nov 03 22:35:00 np Nov 03 22:35:11 khem, so it'll recompile? or i need to clean it on my own? i modified the symlink and i'm running it lets see if i manage to build something this time.. :-) tough work just to get it running Nov 03 22:35:13 kergoth: Any objection to rolling a bitbake 1.8 release? Nov 03 22:35:13 where are the default definitions for FILES? Nov 03 22:35:14 then you can create a local branch which tracks this remote branch and do your development and push it to the remote branch regularly Nov 03 22:35:18 rofl on svk usage Nov 03 22:35:26 dhd: bitbake.conf, same place where nearly every other default value is :) Nov 03 22:35:30 :) Nov 03 22:35:35 RP: nope, sounds fine to me Nov 03 22:35:52 qsp: hmm I would suggest to clean the recipe and rebuilt it Nov 03 22:35:58 kergoth: I think I'm going to tag, release and bump OE's dependency Nov 03 22:36:03 then set to on 1.10 Nov 03 22:36:05 dhd: conf/bitbake.conf is the only configuration file bitbake parses directly. it includes local.conf, machine .conf, distro .conf, etc. Nov 03 22:36:10 RP: sounds good Nov 03 22:36:25 usually it works but mixing objects generated by two different gcc versions is like wearning two different color shoes Nov 03 22:37:47 * Crofton|bold must remember the different colour shoes line Nov 03 22:37:48 khem, sure, i just delete the tmp dir? or there is a more "clean" way to do it? Nov 03 22:38:04 likewise: if you dont have to share your changes then you can use a local branch carved out of .dev Nov 03 22:38:21 qsp: bitbake -c clean gcc-cross-intermediate Nov 03 22:38:33 oh, thank you :-) Nov 03 22:38:35 will clean just gcc-cross-intermediate Nov 03 22:39:26 03Richard Purdie  07refs/tags/1.8.14 * r91a84dd349 10bitbake.git/ (bin/bitbake lib/bb/__init__.py): Nov 03 22:39:26 Release 1.8.14 Nov 03 22:39:26 Signed-off-by: Richard Purdie Nov 03 22:40:09 wow Nov 03 22:40:18 03Richard Purdie  071.8 * r91a84dd349 10bitbake.git/ (bin/bitbake lib/bb/__init__.py): Nov 03 22:40:18 Release 1.8.14 Nov 03 22:40:18 Signed-off-by: Richard Purdie Nov 03 22:40:26 03Richard Purdie  071.8 * r70f124715c 10bitbake.git/ (bin/bitbake lib/bb/__init__.py): Nov 03 22:40:26 Post release version increment Nov 03 22:40:26 Signed-off-by: Richard Purdie Nov 03 22:52:59 bother, I forgot my power adapter Nov 03 22:55:06 Crofton|bold: lower your screen brightness and stop generating interrupts on the keyboard :-) Nov 03 22:56:45 Well, I hope I got that right :) Nov 03 22:57:39 Crofton|bold: What exactly? I'm sure someone might be able to find something similar to give you some power at least Nov 03 22:58:03 khem: I would like to share my branch, i.e. make it remote. How would I keep the remote branch in-sync with .dev. though? Can I make it track .dev? Nov 03 22:59:31 o/ Nov 03 22:59:37 cbrake_away, around ? Nov 03 23:00:58 likewise I belive you will manually have to rebase against the origin Nov 03 23:01:40 rp, I'll likely pick up yet another adapter, or borrow one from my uncle Nov 03 23:03:02 Crofton|bold: ok Nov 03 23:10:25 rp, bring what you have, I am sure the guys from the continent may need some also Nov 03 23:15:59 Crofton|bold: This is UK mains to US mains, right? Nov 03 23:16:24 Crofton|bold: "power adpator" can mean a variety of things :/ Nov 03 23:23:23 RP: while you're there: any idea why 2.6.31 wouldn't boot from nand on c7x0? Boots from media if kexeced, though Nov 03 23:23:56 ant__: none Nov 03 23:24:07 (sane mtd config, of course) Nov 03 23:24:17 some obscure gcc 4.x bug? Nov 03 23:24:33 relocation? Nov 03 23:28:29 rp, correct Nov 03 23:36:46 RP: should I update bitbake to work with current .dev? Nov 03 23:36:59 Jay7: You don't need to yet Nov 03 23:37:09 ok :) Nov 03 23:37:23 Jay7: It will tell you when the time comes Nov 03 23:37:38 better is to tell to ML ;) Nov 03 23:37:54 I think I broke the 1.8.14 release anyway :( Nov 03 23:38:01 hehe.. Nov 03 23:38:14 I did, doesn't even run Nov 03 23:38:29 * RP sighs, too many different copies of bitbake :/ Nov 03 23:41:23 So many versions, so little time Nov 03 23:41:51 Actually, I've totally screwed this up :( Nov 03 23:46:13 it's ok Nov 03 23:46:44 I've semi-broken xpm parsing routine in kexecboot :) Nov 03 23:46:59 I noticed one of kergoths changes, thought it was a regression and removed it :/ Nov 03 23:53:32 03Richard Purdie  071.10 * rf2bc4198f2 10bitbake.git/lib/bb/ (11 files in 2 dirs): Nov 03 23:53:32 Sync module import cleanups back from bitbake trunk Nov 03 23:53:32 Signed-off-by: Richard Purdie Nov 03 23:53:33 03Richard Purdie  071.10 * r4d24b57c03 10bitbake.git/ (bin/bitbake lib/bb/cooker.py lib/bb/fetch/__init__.py): Nov 03 23:53:33 Add a compare-versions command which returns whether the SRCREV versions have changed since bitbake was last run (from Poky) Nov 03 23:53:37 Signed-off-by: Richard Purdie Nov 03 23:53:39 03Richard Purdie  071.10 * rfdd7407534 10bitbake.git/lib/bb/fetch/ (svk.py wget.py): Nov 03 23:53:41 wget.py: Don't check files exist when checking urls for validity Nov 03 23:53:43 Signed-off-by: Richard Purdie Nov 03 23:53:45 03Richard Purdie  071.10 * ra18e62fc26 10bitbake.git/lib/bb/fetch/hg.py: Nov 03 23:53:47 Add hg fetcher auto revisioning, patch from Carl Simonson Nov 03 23:53:49 Signed-off-by: Richard Purdie Nov 03 23:53:51 03Richard Purdie  071.10 * rc9b82d7336 10bitbake.git/lib/bb/ (event.py parse/parse_py/BBHandler.py): Nov 03 23:53:54 event.py: Add recipe parsed event backported from master Nov 03 23:53:56 Signed-off-by: Richard Purdie Nov 03 23:53:58 03Richard Purdie  071.10 * r6838d9eeae 10bitbake.git/lib/bb/build.py: Nov 03 23:54:04 build.py: Remove zero length logfiles (backport from master) Nov 03 23:54:06 Signed-off-by: Richard Purdie Nov 03 23:54:08 03Richard Purdie  071.10 * r361dc190c6 10bitbake.git/lib/bb/persist_data.py: Nov 03 23:54:10 persist_data.py: Add missing getKeyValues function for SCM version comparision code Nov 03 23:54:12 Signed-off-by: Richard Purdie Nov 03 23:54:14 03Richard Purdie  071.10 * r15f8e4b549 10bitbake.git/lib/bb/providers.py: Nov 03 23:54:18 providers.py: Add regexp cache speeding up certain operations Nov 03 23:54:20 Signed-off-by: Richard Purdie Nov 03 23:54:22 03Richard Purdie  071.10 * r0cbc684e94 10bitbake.git/lib/bb/fetch/__init__.py: Nov 03 23:54:24 fetch/__init__.py: Drop pointless pickle import Nov 03 23:54:26 Signed-off-by: Richard Purdie Nov 03 23:54:28 03Richard Purdie  071.10 * rd4bc9c01d2 10bitbake.git/lib/bb/parse/parse_py/BBHandler.py: Nov 03 23:54:35 BBHandler: Drop unused __functions__ variable (backport from master) Nov 03 23:54:37 Signed-off-by: Richard Purdie Nov 03 23:54:39 03Richard Purdie  071.10 * r52127bcebe 10bitbake.git/lib/bb/runqueue.py: Nov 03 23:54:41 runqueue.py: Update to the faster algorithm from master Nov 03 23:54:43 Signed-off-by: Richard Purdie Nov 03 23:54:45 03Richard Purdie  071.10 * rd8777b04e6 10bitbake.git/lib/bb/build.py: Nov 03 23:54:51 build.py: Drop unneeded directory changes (from trunk) and remove code duplication Nov 03 23:54:53 Signed-off-by: Richard Purdie Nov 03 23:54:55 03Richard Purdie  071.10 * r506ded268f 10bitbake.git/lib/bb/utils.py: Nov 03 23:54:57 (5 lines omitted) Nov 03 23:57:24 03Richard Purdie  071.8 * r72cc4aabaa 10bitbake.git/bin/bitbake: Nov 03 23:57:24 bin/bitbake: Fix typo during release process Nov 03 23:57:24 Signed-off-by: Richard Purdie Nov 03 23:57:25 03Richard Purdie  071.8 * r1c06edb545 10bitbake.git/ (bin/bitbake lib/bb/__init__.py): Release 1.8.16 Nov 03 23:57:35 03Richard Purdie  071.8 * r53f097ed9f 10bitbake.git/ (bin/bitbake lib/bb/__init__.py): Increment version post release Nov 04 00:00:37 kergoth: You and I are conflicting on a commit :/ Nov 04 00:00:49 what commit? Nov 04 00:01:06 http://cgit.openembedded.net/cgit.cgi/bitbake/commit/?id=1863e9954571e6fd6ee0a0fd307ca226bff86bad vs. http://cgit.openembedded.net/cgit.cgi/bitbake/commit/?id=60e36de1b7c72139933bc1760094f28a64dc6b43 Nov 04 00:01:29 kergoth: I changed that originally so we could remove the hardcoded assumption on "SRC_URI" in the fetcher Nov 04 00:01:43 kergoth: Obviously it has issues with your mirror changes though Nov 04 00:01:58 kergoth: Need to find something that fixes this for both issues :) Nov 04 00:02:14 ah, indeed Nov 04 00:02:16 hmm Nov 04 00:02:37 I only noticed when merging stuff around :/ Nov 04 00:02:38 * kergoth would like to just rethink the whole bb.fetch api at some point Nov 04 00:02:52 kergoth: no argument here Nov 04 00:03:20 bb.fetch.init probably needs to save further state Nov 04 00:04:46 on the plus side, we hopefully have a sane bitbake 1.8 release on the second try, a 1.10 branch that starts to look more like master and is in sync with poky Nov 04 00:05:21 and if we bump OE.dev to require bitbake 1.8.16 we can start to use BBCLASSEXTEND and droppping "import bb" everywhere Nov 04 00:07:19 that'll be very nice Nov 04 00:09:28 hmm, bb.fetch.Fetcher(url) factory -> fetcher object -> run download method, passing local path as an argument, or similar, would probably be a cleaner way of doing it Nov 04 00:21:07 pb_, do you know anything about the opkg 'flag' command? There is a comment in the sources saying you asked for some feature (not sure if it refers to a type of flag, or the flags themselves) Nov 04 00:21:39 The openwrt mob have removed the flags support in some of their patches. Nov 04 00:22:22 Are there real world use cases for these flags? Nov 04 00:43:54 03Richard Purdie  071.10 * rb2e24e190c 10bitbake.git/lib/bb/fetch/ (__init__.py git.py): Nov 04 00:43:54 fetch/__init__.py: Try and make the sortable_revision code mode readable Nov 04 00:43:54 Signed-off-by: Richard Purdie Nov 04 00:44:05 03Richard Purdie  07master * rb3af6f278a 10bitbake.git/lib/bb/fetch/ (__init__.py git.py): Nov 04 00:44:05 fetch/__init__.py: Try and make the sortable_revision code mode readable Nov 04 00:44:05 Signed-off-by: Richard Purdie Nov 04 00:44:30 gogogo RP Nov 04 00:47:57 03Richard Purdie  07master * r054758a54a 10bitbake.git/lib/bb/fetch/git.py: Nov 04 00:47:57 git.py: Search mirrors for tarballs before fetching, not just local directories Nov 04 00:47:57 Signed-off-by: Richard Purdie Nov 04 00:47:57 03Richard Purdie  071.10 * r0ed37ca237 10bitbake.git/lib/bb/fetch/git.py: Nov 04 00:47:59 git.py: Search mirrors for tarballs before fetching, not just local directories Nov 04 00:48:01 Signed-off-by: Richard Purdie Nov 04 00:48:38 * RP -> Zzzz **** ENDING LOGGING AT Wed Nov 04 02:59:57 2009