**** BEGIN LOGGING AT Mon Mar 30 02:59:57 2009 Mar 30 03:03:28 Arcom development kits for sale? Mar 30 04:16:50 03Angus Ainslie  07fso/milestone5.5 * re844b8eb62 10openembedded.git/conf/distro/include/sane-srcrevs.inc: bump frameworkd and efl srcrev Mar 30 06:49:13 morning Mar 30 06:55:09 morning recalcati_oe Mar 30 07:01:06 is oe broken? Mar 30 07:01:33 a couple of days back, i was able to compile midori, now i cant Mar 30 07:04:00 i get a configure error in gnone-keyring package: it says configure: error: asn1Parser tool is not installed Mar 30 07:04:00 i dont understand Mar 30 07:04:00 or is my setup broken, Mar 30 07:04:00 i removed tmp directory twice thinking i had a broken setup. i get new errors each time Mar 30 07:07:21 gr_8: what's your last pull? Mar 30 07:07:29 yesterday: Mar 30 07:07:31 [20:01] Koen Kooi org.openembedded.dev * rf3cc636c6d openembedded.git/recipes/gnome/gnome-keyring_2.26.0.bb: gnome-keyring: depend in libtasn1-native to get access to needed binaries Mar 30 07:07:33 [20:01] Koen Kooi org.openembedded.dev * rdfd310fbb6 openembedded.git/recipes/gnutls/libtasn1-native_1.7.bb: libtasn1: add native version Mar 30 07:08:39 yesterday Mar 30 07:09:21 good morning Mar 30 07:11:23 likewise: i will pull again and try now Mar 30 07:11:33 gm Mar 30 07:11:40 seems somechanges have been added recently Mar 30 07:11:43 +DEPENDS = "gtk+ libtasn1 libtasn1-native libgcrypt" Mar 30 07:14:44 morning mckoan, I was back to libwebkit-1.0-1 - 0.1+svnr40814-r8.1 and midori and epiphany worked again Mar 30 07:20:36 likewise: i have similar errors with libsoup as well, it is looking for sqllite in configure, can i do the same as koen has done to gnome-keyring? Mar 30 07:22:10 so libsoup needs a sqlite on the host??? Mar 30 07:22:26 gr_8: can you show what happens in pastebin.ca? Mar 30 07:22:55 configure: error: Could not find sqlite3 devel files: Mar 30 07:23:14 gnome-keyring required some binaries on the build host (i.e. a -native package). Mar 30 07:23:33 likewise: how do you diffrentiate that the said library/package is needed on host or target? Mar 30 07:23:48 from the configure error, it is usually the host Mar 30 07:23:59 gr_8: that *sounds* like a build time dependency against a target development package. Mar 30 07:25:08 gr_8: careful inspection of what it is building. Most packages build target binaries only, but some build intermediate tools required to run on the build host, as helpers. Mar 30 07:26:32 gr_8: step 1 is usually checking if the dependency can be disabled Mar 30 07:27:01 gr_8: after all we are targetting embedded systems, any optional dependencies/features that are excessive, we try to get rid of. Mar 30 07:29:04 likewise: i did not understand step1 Mar 30 07:29:31 can you suggest what this careful inspection is Mar 30 07:29:58 gr_8: looking at /tmp/work///../temp/log* Mar 30 07:30:00 so way where i can find what exactly is the issue. is it a dependency issue/it requires something on the host Mar 30 07:30:40 gr_8: it requires experience with configure / building packages from source, there is not magic rule there Mar 30 07:31:03 gr_8: inspecting the temp/run and temp/log are the only advice I can give Mar 30 07:31:49 likewise: i do that, but i really dont know what could cause the issue Mar 30 07:32:06 good morning Mar 30 07:32:21 likewise: i am able to find out where it failed, what is the error, i dont know how to solve the error Mar 30 07:33:05 likewise: a few days ago i got an error saying gdk.pc sanity check failed. i dont know how to solve it Mar 30 07:33:35 gr_8: me neither Mar 30 07:33:41 i knew something to do with gdk.pc, Mar 30 07:33:53 the bug is listed in oe bugzilla Mar 30 08:46:06 morning Mar 30 08:55:14 hrw: morning Mar 30 08:55:34 hrw: when does the stable branch start :)? Mar 30 08:57:07 mpoullet|work: I have some test builds now in progress to check does everything works Mar 30 08:58:08 so far x11-image/console-image/console-base-image built fine for armv5/v7/i586/486sx Mar 30 08:58:39 hrw: cool, I'm trying to get a machine for night builds at work too Mar 30 08:58:55 cool Mar 30 09:00:23 03Koen Kooi  07org.openembedded.dev * rd7ce4de288 10openembedded.git/recipes/efl1/ (ecore.inc ecore_svn.bb): Mar 30 09:00:23 ecore: fix packaging after last update to SRCREV Mar 30 09:00:23 * please pay more attention when updating SRCREV next time Mar 30 09:01:15 hrw: is someone writting some automated test build scripts? maybe you or Crofton? Mar 30 09:01:46 mpoullet|work: I prefer to use buildbot then autoscripts but those needs to be written too Mar 30 09:02:22 hrw: are you referring to the stable branch of oe? Mar 30 09:02:34 yes Mar 30 09:02:35 i meant oe recipes git? Mar 30 09:02:41 ok, I don't know buildbot, I'll have a look Mar 30 09:02:51 for me some packages dont compile Mar 30 09:03:34 gr_8: even in stable branch there will be recipes which will not build Mar 30 09:03:44 when i compile gimp i get the message which means gdk.pc has lost its sanity Mar 30 09:04:27 hrw: arent all packages supposed to get compiled? Mar 30 09:04:33 no Mar 30 09:05:11 gr_8: we have over 6 thousands of recipes and not so many devs interested in fixing one by one Mar 30 09:06:13 i thought oe says here are 2000+ packages you can compile, build, package these packages Mar 30 09:06:42 hrw: you say when i hit a problem, i need to fix it and send a patch since oe is not complete Mar 30 09:07:42 gr_8: there are 2000+ recipes which you can build and use. there are also many recipes which will not build Mar 30 09:07:55 okay Mar 30 09:08:04 gr_8: we are just humans and many of us do OE in free time Mar 30 09:08:13 i understand Mar 30 09:08:41 the problem is i know there is problem, what the problem is, but i dont know howto fix it Mar 30 09:08:55 that is if something does not compile Mar 30 09:10:06 do you recommend any study material Mar 30 09:16:22 http://blip.tv/file/1658991/ Mar 30 09:21:41 hrw: i know that Mar 30 09:21:43 lol Mar 30 09:23:25 that seems to be good study material to me Mar 30 09:23:41 ;D Mar 30 09:26:26 good morning Mar 30 09:26:56 oh..again tinylogin and dropping privileges... Mar 30 09:28:19 ant_work: good morning Mar 30 09:28:26 hrw: very interesting patches from Roman I Khimov Mar 30 09:28:32 mpoullet|work: morning Mar 30 10:24:25 florian: good morning Mar 30 10:24:31 who is the patchwork admin? Mar 30 10:25:18 florian: are you the creator of beagle+pico_projector+laser_keyboard+.... ? Mar 30 10:26:14 methri|work: are you talking of beagle board demo.bb? Mar 30 10:26:18 good morning Mar 30 10:26:24 methril|work: is that working for you? Mar 30 10:26:28 methril|work: yes Mar 30 10:26:47 florian: nice work :) Mar 30 10:27:08 gr_8: No, a small "computer device" with beagle Mar 30 10:27:26 gr_8: in the florian blog Mar 30 10:28:51 methril|work: Well its a start - to make it really useful there is quite some research and development left to do. Mar 30 10:29:17 florian: nice start :) Mar 30 10:29:34 florian: what is the url of your blog? Mar 30 10:30:22 gr_8: http://fl0rian.wordpress.com/2009/03/29/a-smart-beaglebrick/ Mar 30 10:32:17 * florian should get a case for the Beagleboard only and try to find the best device layout. Mar 30 10:37:43 florian: i like that Mar 30 10:37:54 nice Mar 30 10:39:36 it would be nice have a oe plane, isn't it? Mar 30 10:39:48 for some geek doing hw/sw blog posting Mar 30 10:39:52 related to oe Mar 30 10:40:18 planet.linuxtogo.org partially handles that Mar 30 10:43:09 hum yes, we should try to 'collect' some more blogs there. Mar 30 10:43:22 hrw: Do you have some missing ones in mind? Mar 30 10:43:26 i missed this hrw :) Mar 30 10:44:10 florian: thats not that way Mar 30 10:44:35 florian: it is easy to miss planet.ltg as it is not linked on OE website etc Mar 30 10:44:55 03Koen Kooi  07org.openembedded.dev * rcc89fef9b7 10openembedded.git/recipes/linux/ (linux-omap-2.6.29/omapzoom/defconfig linux-omap_2.6.29.bb): linux-omap 2.6.29: add omapzoom support and fix PV so it can find the proper patches Mar 30 10:44:56 03Koen Kooi  07org.openembedded.dev * r1f1f8701de 10openembedded.git/recipes/linux/linux-omapzoom/omapzoom/defconfig: linux-omapzoom: fix defconfig Mar 30 10:45:04 hrw: well, that should be easy to fix :) Mar 30 10:51:21 florian: beaglebrick seems the future of portable pc Mar 30 11:15:38 im trying to compile ffmpeg, i get an message saying libtheora not found, i see that libtheora has been compiled but it is not installed Mar 30 11:15:55 i dont see a install function within libtheora .bb file Mar 30 11:28:36 its staging that needs to get called Mar 30 11:29:12 yes, do_stage, it does not install libtheora Mar 30 11:30:30 odd ffmpeg is normally well tested :-) Mar 30 11:31:14 yes, thats why im asking Mar 30 11:31:23 am i missing something Mar 30 11:31:54 Xora: have you tested ffmpeg? Mar 30 11:32:02 I am just now Mar 30 11:35:24 Hello all Mar 30 11:35:31 xora: oe setup uses ffmpeg_0.5.bb Mar 30 11:35:45 hello otavio Mar 30 11:35:47 bah libgsm is broken here Mar 30 11:35:53 How are you all? Mar 30 11:35:55 ant_work: :-) Mar 30 11:36:30 otavio: it seems psplash is needing deep rework... Mar 30 11:36:44 ant_work: really? Mar 30 11:36:52 ant_work: shoot; if I can help Mar 30 11:37:04 mom Mar 30 11:37:10 ant_work: np Mar 30 11:37:52 xora: do i really need libtheora? Mar 30 11:37:58 otavio_ http://lists.linuxtogo.org/pipermail/openembedded-devel/2009-March/009171.html Mar 30 11:38:09 some .bb file dont have --enable-libtheora in ffmpeg Mar 30 11:39:34 ant_work: I have 11 changes queued; I'll see if I can look into this and push it all together Mar 30 11:39:58 np, is a minor issue Mar 30 11:40:29 ant_work: I somewhat understand Koen point; we do require a psplash-oe, ok, but I also belive we need to set the preferred version to angstrom Mar 30 11:41:02 or remov ethe virtual... Mar 30 11:41:45 I have no strong opinion on this, it's 'distro' thing Mar 30 11:42:54 ant_work: we need the virtual Mar 30 11:43:12 then e need preferred-version :-) Mar 30 11:43:17 ant_work: for example at initramfs-module-psplash needs to be able to be used with any of them Mar 30 11:43:30 ant_work: I use it; that's why I added the virtual Mar 30 11:44:06 I see, we had same issues with kexecboot-kernel logo Mar 30 11:44:17 angstrom uses own Mar 30 11:44:52 we did it using overrides / folder hierarchy Mar 30 11:45:59 it will be the same thing Mar 30 11:48:01 ant_work: I replied to Koen, let's see what he say's Mar 30 11:48:28 fine for me, thx Mar 30 11:49:04 I just wanted to silence that note :-) Mar 30 11:50:09 gtk-builder-convert - what provides it? Mar 30 11:50:18 gnome-keyring fails here due to lack of it Mar 30 11:50:48 hrw: some new patches added to gnome-keyring some 9 hours ago Mar 30 11:51:16 hrw: update you git tree, i have the same problem today Mar 30 11:52:19 gr_8: the problem is that I have those Mar 30 11:52:45 we lack gtk+-native Mar 30 11:53:58 hrw: are you referring to gnome-keyring not compiling due to libasn1? Mar 30 11:54:31 hrw: http://cgit.openembedded.net/cgit.cgi?url=openembedded/commit/&id=f3cc636c6d63a60442255017bef502cd5622eb99 Mar 30 11:55:27 no Mar 30 11:55:59 gnome-keyring calls gtk-builder-convert which my system do not have installed and we do not have gtk+-native which would provide such Mar 30 11:56:53 oe building varies from system to system? depending on the distribution Mar 30 11:57:10 and what you have/you dont have on your system? Mar 30 11:57:18 hrw: fsck Mar 30 11:57:37 XorA: indeed Mar 30 11:58:21 what was the joke? i didnt follow Mar 30 11:58:36 hrw: newbee here Mar 30 12:00:04 it was not Mar 30 12:00:41 hrw: am i correct in my understanding? Mar 30 12:03:06 gr_8: OE is self contained as much as possible Mar 30 12:03:14 but some corners were cut ;( Mar 30 12:03:31 * XorA wonders what has staged a gsm.h file that is not libgsm Mar 30 12:04:12 XorA: find tmp/work -name gsm.h Mar 30 12:04:33 hrw: how did you debug this issue, im trying to learn. how did you figure out that gtk+-native is required? Mar 30 12:04:52 hrw: bad staging function in libgsm, if you bump PR a newer one cannot stage :-( Mar 30 12:05:30 gr_8: build failed due to missing tool. I checked which Debian package provides it (as I use Debian) and then mapped it to OE names Mar 30 12:06:53 is this mapping given somewhere Mar 30 12:07:30 uf... gtk-builder-convert is Python Mar 30 12:08:57 /bin/sh: ./vp3huff: cannot execute binary file Mar 30 12:09:07 so thats why libtheora doesnt build Mar 30 12:09:54 Xora: i did not understand, what are you trying to say Mar 30 12:10:22 xora: somebody removed the stage function in libtheora recently Mar 30 12:10:28 gr_8: building libthero is broken in the same way as gnome-keyring, it needs native tools Mar 30 12:10:56 xora: it builds fine here Mar 30 12:11:16 xora: can you paste the configure.log/error message somewhere Mar 30 12:11:22 i would like to see Mar 30 12:17:54 http://pastebin.ca/1376812 is my fix Mar 30 12:18:02 XorA: can you look? Mar 30 12:21:39 hrw: I think thats the sensible path when its a python script Mar 30 12:29:33 ~curse libtheora Mar 30 12:29:34 May the fleas of a thousand camels infest your most sensitive regions, libtheora ! Mar 30 12:29:44 I was told you discussed ffmpeg+libtheora problems today. any chance somebody can paste/write me the result? (-: Mar 30 12:30:07 oh, maybe you are still right in discussion?! (-: Mar 30 12:31:22 I am testing ugly hack Mar 30 12:32:41 if we disable building of documentation then it builds Mar 30 12:35:05 ~lart libgd for moving to cmake Mar 30 12:35:05 * ibot dumps 42 tons of dirt, manure, and fish heads on libgd for moving to cmake Mar 30 12:58:22 I hate situations when I have to fix recipes.. Mar 30 13:04:59 hi aloisiojr Mar 30 13:06:01 mckoan: hey Mar 30 13:06:46 aloisiojr: qt-4.5 works great Mar 30 13:07:07 very nice Mar 30 13:07:25 mckoan: did you test it in a qemu env, right? Mar 30 13:08:50 likewise: ping Mar 30 13:11:50 hrw: http://cgit.openembedded.net/cgit.cgi?url=openembedded/commit/&id=9bfd0bd36ab2caa37f34554204ff8e830f848507 Mar 30 13:11:59 aloisiojr: no, I testet them with an ARM9 AT91SAM9263 based board Mar 30 13:12:03 hrw: unfortuneately Ulf put that on the branch he then asked to be deleted Mar 30 13:12:22 yep Mar 30 13:12:39 that reminds me one thing.. Mar 30 13:12:46 I have 2 branches to remove Mar 30 13:15:29 ant_work: pong Mar 30 13:15:38 mckoan: very nice Mar 30 13:15:43 ant_work: what did I break? :-) Mar 30 13:15:48 likewise: you're on x86_64 isn't? I see in tinderbox bluez-utils 3.36 failing Mar 30 13:16:00 ant_work: nope, I moved to x86_32.... Mar 30 13:16:18 he..you're the only human there...only autobuilder Mar 30 13:16:20 mckoan: and did you take a look in the bbclass that a put in pastebin? Mar 30 13:16:20 ant_work: you mean build host or target btw? Mar 30 13:16:29 http://tinderbox.openembedded.net/packages/?status=Failed&name=bluez-utils&page=1 Mar 30 13:17:00 (I just closed bug #3764 and now see it fails on x86_64) Mar 30 13:17:03 shit.. local git say that I pushed 3 revs.. cgit do not shows it Mar 30 13:17:13 XorA: git pull .dev and tell me what you got please Mar 30 13:18:23 ant_work: it's my colleague (who doesn't know much OE yet): /home/tom, he plain copied my bitbake config it seems Mar 30 13:19:24 :) Mar 30 13:19:52 git-pull Mar 30 13:19:53 Already up-to-date. Mar 30 13:20:43 XorA: git log? Mar 30 13:20:56 commit cc89fef9b784660843e6ed79911111b866db49e1 Mar 30 13:20:56 Author: Koen Kooi Mar 30 13:20:56 Date: Mon Mar 30 12:42:37 2009 +0200 Mar 30 13:20:56 linux-omap 2.6.29: add omapzoom support and fix PV so it can find the proper Mar 30 13:20:59 shit Mar 30 13:21:00 is the top entry Mar 30 13:21:46 commit 5fbd9b895fb92083ee41d107648af60831ecf543 Mar 30 13:21:46 Author: Marcin Juszkiewicz Mar 30 13:21:46 Date: Mon Mar 30 14:59:47 2009 +0200 Mar 30 13:21:46 libtheora: added staging to get ffmpeg build Mar 30 13:21:49 top here Mar 30 13:22:56 03Ulf Samuelsson  07org.openembedded.dev * r27759557d3 10openembedded.git/recipes/libogg/ (libtheora-1.0/libtheora-1.0-no-docs.patch libtheora_1.0.bb): Mar 30 13:22:56 libtheora_1.0.bb : Do not build libtheora docs/examples since this will try Mar 30 13:22:56 to run a crosscompiled utility on the host Mar 30 13:23:45 hrw: nothing wrong with the git server then :-) Mar 30 13:24:03 aloisiojr: I need to verify why qt-4.5 image is 40MB and wxWidgets image is 20MB Mar 30 13:25:23 hrw: the problem is libthero does not get installed, the header files and libtheora.so are missing from the staging directory, can i send a patch? Mar 30 13:25:31 03Marcin Juszkiewicz  07org.openembedded.dev * r49905a2449 10openembedded.git/recipes/gtk+/gtk+_2.16.0.bb: gtk+: stage gtk-builder-convert to get gnome-keyring built Mar 30 13:25:40 mplayer compilation fails because of libtheora Mar 30 13:25:41 gr_8: pull Mar 30 13:25:42 03Marcin Juszkiewicz  07org.openembedded.dev * r43427019cb 10openembedded.git/recipes/libogg/libtheora_1.0.bb: libtheora: added staging to get ffmpeg build Mar 30 13:26:36 hrw: i tried that here, it does not put the libraries and header files into staging Mar 30 13:26:41 hrw: PEBKAC ? Mar 30 13:27:55 NOTE: package ffmpeg-1_0.5-r1: task do_compile: started Mar 30 13:28:23 mckoan: there is a bug that i need to fix there... Mar 30 13:28:36 XorA: hard to tell. I resetted my repo, reapplied patches and pushed this time good Mar 30 13:28:39 mckoan: beyond this problem that you said Mar 30 13:28:51 mckoan: regarding qvariant + float Mar 30 13:29:34 mckoan: moc is compiled to assume that qreal is double inside qvariant, and the qt assume that it is float Mar 30 13:29:53 mckoan: this is a qt build system bug Mar 30 13:30:07 mckoan: we already talked to qt guys Mar 30 13:33:23 hrw: does that work?, i tried that yesterday, it didnt, this seems http://pastebin.ca/1376849 to work Mar 30 13:33:42 gr_8: ffmpeg is now fixed Mar 30 13:34:07 xora: are you seeing theora.h and libtheora in the staging directory? Mar 30 13:34:16 ffmpeg is building Mar 30 13:35:33 gr_8: I built ffmpeg already Mar 30 13:36:09 i had trouble building mplayer, so i made those changes and it worked Mar 30 13:37:13 if i add autotools_stage_all to do_stage it does not work, that is it does not install libtheora.so and theora/theora.h to staging directory as a result of which mplayer build fails Mar 30 13:37:19 aloisiojr: ok, thx Mar 30 13:37:33 gr_8: pull again. Mar 30 13:38:03 hrw: am I remembering right you did import qemu patches from poky? (I'd close 2589 and 2548) Mar 30 13:38:20 I thought i had the patches already a few minutes ago, but there was another commit Mar 30 13:39:39 03Stefan Schmidt  07fso/milestone5.5 * r8fe876d687 10openembedded.git/packages/cellhunter/cellhunter_0.4.2.bb: cellhunter: Add frameworkd RDEPENDS Mar 30 13:39:50 03Stefan Schmidt  07fso/milestone5.5 * r796f7c2065 10openembedded.git/packages/openbmap/ (openbmap-logger_0.1.0.bb openbmap-logger_0.2.0.bb): openbmap-logger: Add frameworkd RDEPENDS Mar 30 13:54:25 ant_work: long time ago Mar 30 13:55:17 hrw: one bug was closed, the other no. I don't build on x86_64 so can't really say Mar 30 13:55:33 We'll have to ask Stanislav Mar 30 13:56:09 yes Mar 30 14:01:57 hrw: are you building on x86_64? as I said it seems bluez-utils 3.36 could be broken if built on x86_64 Mar 30 14:02:13 03Koen Kooi  07org.openembedded.dev * r92ddcd0051 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Mar 30 14:02:14 03Koen Kooi  07org.openembedded.dev * r5defd177d4 10openembedded.git/conf/machine/omapzoom.conf: omapzoom: add xf86-input-tslib Mar 30 14:02:15 03Koen Kooi  07org.openembedded.dev * re8f0a65d79 10openembedded.git/conf/machine/omapzoom.conf: omapzoom: linux-omap work better on the device than linux-omapzoom, so switch to that Mar 30 14:02:16 03Koen Kooi  07org.openembedded.dev * r0546e921e1 10openembedded.git/recipes/linux/ (3 files in 3 dirs): linux-omap: add patch to fix mmc support and update defconfig Mar 30 14:04:46 hrw: before opening a bug I hope to hear it fails just for the autobuilder... Mar 30 14:43:40 03Robert Schuster  07org.openembedded.dev * r2263e973b5 10openembedded.git/recipes/cacao/ (3 files in 2 dirs): Mar 30 14:43:40 cacao 0.99.3: Added patch to fix PR99 (this time for real). Mar 30 14:43:40 cacao 0.99.4: Dito. Mar 30 14:51:17 03Graeme Gregory  07org.openembedded.dev * r80570a8a80 10openembedded.git/recipes/mozilla/ (13 files in 2 dirs): Mar 30 14:51:17 firefox_3.0.8.bb : add new version, seems to compile with modern gcc Mar 30 14:51:17 so made it the default. Mar 30 14:51:18 03Graeme Gregory  07org.openembedded.dev * r24931658f9 10openembedded.git/conf/checksums.ini: checksums.ini : add firefox 3.0.8 Mar 30 14:57:50 ant_work: my machine builds other stuff now ;( Mar 30 15:02:02 mplayer seems to be broken now Mar 30 15:02:10 i get the error: nellymoserenc.c:205: error: 'INFINITY' undeclared (first use in this function) Mar 30 15:05:10 hrw: when you'll find time have a look at packaged-staging Mar 30 15:05:31 (ot just ask mckoan :) Mar 30 15:08:29 gr_8: NOTE: package mplayer-0.0+1.0rc2+svnr28346-r9: task do_build: completed Mar 30 15:08:36 gr_8: for beagleboard Mar 30 15:08:39 ant_work: had some fun on Friday trying to check for a few things in a sanity class.. checked for recipes that can have S, FILESDIR, or FILESPATH overrides removed Mar 30 15:09:04 very nice... Mar 30 15:09:20 ant_work: found 1,168 recipes that can have their S= removed.. of course, a number of those came in via a .inc, so it looks like multiple recipes can have it removed, when its just one file.. no way to check for that from a task Mar 30 15:10:53 Ah, about this 'multiple recipes' see http://bugs.openembedded.net/show_bug.cgi?id=1688 Mar 30 15:11:31 nice, that's useful Mar 30 15:11:45 was Laibsch's first-try Mar 30 15:12:05 hrw: what could be wrong? my setup Mar 30 15:12:24 hrw: im trying to compile the same version of mplayer Mar 30 15:12:31 for omap3evm Mar 30 15:14:38 kergoth: new devs will copy from existing recipes...a clean base is future-proof Mar 30 15:19:19 03rmonk  07org.openembedded.dev * r4346939289 10openembedded.git/recipes/dsplink/ti-codec-engine_2.23.bb: Mar 30 15:19:19 codec-engine: Add development support for Codec Engine 2.23 Mar 30 15:19:19 Signed-off-by: Koen Kooi Mar 30 15:21:14 ant_work: NOTE: package bluez-utils-3.36-r4: task do_build: completed Mar 30 15:21:24 ant_work: angstrom/beagleboard on amd64 Mar 30 15:23:28 hrw: nice to hear Mar 30 15:23:56 how can i make oe use the mplayer_svn.bb and not mplayer_0.0+1.0rc2.bb Mar 30 15:28:22 I'm trying to use a python cmd in a var thats doing something a little funny: DISTRO_VERSION ?= "${@os.popen('svnversion $OEROOT | tr : _').read().strip()}" - the bitbake 'banner' displayed from base.bbclass shows correctly DISTRO_VERSION as the output of svnversion in my root dir, but when meta-toolchain.bb is processed DISTRO_VERSION seems to be 'exported' which means the var is getting evaluated multiple times, and perhaps Mar 30 15:28:22 in some strange environment Mar 30 15:29:35 03Koen Kooi  07org.openembedded.dev * r9749a93633 10openembedded.git/recipes/mozilla/firefox_3.0.8.bb: firefox 3.0.8: lower default pref Mar 30 15:35:57 i dont know whats happening in my setup, i tried first time mplayer takes mplayer_0.0+1.0rc2.bb and now after doing bitbake -c clean mplayer, when i build mplayer it takes mplayer_svn.bb Mar 30 15:36:35 can i specify the version number and date (incase it checks out of svn/cvs/git) of the mplayer Mar 30 15:36:38 which distro/machine combo? Mar 30 15:37:22 angstrom, omap3evm Mar 30 15:39:54 kergoth, you had pointed me to the use of python in variable assignments the other day - do you know why a 'svnversion $OEROOT' would work in some recipes but not others? I don't understand why the variable assignment would not take place only once Mar 30 15:40:12 is there any way i can print out the all .bb/version number that my build takes Mar 30 15:40:50 shouldn't that be ${OEROOT}? i really doubt you want the shell to be expanding that Mar 30 15:41:32 kergoth, perhaps but when base.bbclass spits out the banner info, the DISTRO_VERSION is correct yet when meta-toolchain uses it in its output filename its not? Mar 30 15:41:57 no idea Mar 30 15:42:12 but its probably what i just said Mar 30 15:42:29 i have no other ideas offhand, would have to do further investigation Mar 30 15:43:00 OEROOT is an env var thats whitelisted Mar 30 15:43:25 the addition of the curly braces around it in the conf does what? Mar 30 15:44:43 I'm not totally clear how bitbake parses its files Mar 30 15:44:55 it means its expanded by OE, whenever its used, anywhere. Mar 30 15:45:06 $OEROOT would only be expanded if it was emitted into shell somewhere Mar 30 15:45:36 ah... ok I see - and yes you are right (just tested it) Mar 30 15:46:47 so evidently the variable expansion is happening multiple times which is a bit strange - the first time (base.bbclass) the env was as I thought it would be but when meta-toolchain.bb was being processed it was not Mar 30 15:48:10 if you want it to be expanded once, you can use := to evaluate it *right there*. but i doubt thats helpful to you. the issue is its expanded whenever its used for the most part.. there's *sort of* a expansion cache, but it doens't work particularlyw ell, so things often get expanded numerous times in a single run Mar 30 15:48:14 unfortunately Mar 30 15:48:15 hi mickeyl Mar 30 15:50:17 kergoth, ok - I do recall the := syntax thx. Why do you think (in my case of trying to get at the svn revision of my project) that I wouldn't want to eval it 'right there'? Mar 30 15:51:26 oh, i suppose it would be okay for your case. was just thinking for the case where the command needs something from the recipe, you couldn't define it in a .conf and use := because the recipe hasn't yet been parsed at that point Mar 30 15:51:35 but for you, it looks global, so that probably is best indeed Mar 30 15:51:38 kergoth: about this expansions...I fear I misunderstand ...pls see the [PATCH] kernel.bbclass: Automate uboot entry/load addresses detection (c/o Ihar Hrachyshka) Mar 30 15:51:49 when are the UBOOT_x vars evaluated? Mar 30 15:52:00 I have mixed results...fails..works... Mar 30 15:52:07 can't understand why Mar 30 15:52:55 kergoth, right... ok - thx again Mar 30 15:53:01 np Mar 30 15:53:17 ant_work: i'll take a look in a bit, catching up on work email at the moment Mar 30 15:53:24 kergoth: np, it's there http://lists.linuxtogo.org/pipermail/openembedded-devel/2009-March/009245.html Mar 30 15:53:38 (lot of emails today :-) Mar 30 15:56:55 the bitbake manual does not mention the 'soft assignment' ?= variable assignment syntax? - I assume := is a hard assignment right? Mar 30 15:57:12 khem: ping Mar 30 15:57:30 http://bitbake.berlios.de/manual/ch02.html#id868911 Mar 30 15:58:06 := is immediate expansion syntax in gnu make too, there was a bit of influence by it Mar 30 15:58:51 Damnit Mar 30 15:58:52 strange that ?= isn't mentioned in the manual there though Mar 30 15:58:58 Does bitbake git support not do file URIs? Mar 30 15:59:03 agreed, that is strange Mar 30 15:59:12 g'day kergoth Mar 30 15:59:16 hey pb_ Mar 30 15:59:19 yo florian Mar 30 15:59:25 can i specify a date for my oe. i want to make sure that oe always downloads the same sources. this is in case of git/svn/cvs downloads Mar 30 15:59:33 the patch to update bitbake docs was submitted long time ago... Mar 30 15:59:58 gr_8: look for SRCREV Mar 30 16:01:05 or SRCDATE... Mar 30 16:01:28 hail mickeyl Mar 30 16:02:40 tharvey, kergoth: https://lists.berlios.de/pipermail/bitbake-dev/2009-March/000423.html Mar 30 16:05:11 denix, thx Mar 30 16:05:30 SRCDATE is defined in bitbake.conf so everytime (different day i build) it downloads the latest sources in case it is in git/cvs/svn? Mar 30 16:06:40 Ah yay I think, user error :) Mar 30 16:06:52 no more abiword Mar 30 16:06:54 ever Mar 30 16:09:44 hi, was just building a package that required fs and noticed the build failed (fs_cvs).. is this my fault or..? http://www.pastebin.ca/1376976 Mar 30 16:10:01 gr_8: set the SRCDATE in your recipe, or otherwise SRCDATE_pkgname elsewhere Mar 30 16:10:59 can this SRCDATE set globally for all packages? Mar 30 16:10:59 there should be sane-srcrevs and sane-srcdates for all svn/cvs recipes already Mar 30 16:11:16 if not please patch to add them for missing recipes Mar 30 16:12:04 during compilation i sometimes see the same package getting compiled over and over again Mar 30 16:13:14 i got to go now, thank you all Mar 30 16:13:45 <_boto> hello Mar 30 16:14:02 <_boto> i have a problem building boost for omap (beagle) Mar 30 16:14:20 Hmm Mar 30 16:14:22 <_boto> it seems that the host gcc is taken for building instead of the cross compiler Mar 30 16:14:25 Where do bitbake patches go? Mar 30 16:14:44 <_boto> well, what i see is that the patches are evaluated Mar 30 16:15:16 <_boto> building boost on neighboring machines makes no problems though Mar 30 16:15:24 <_boto> we use ubuntu as host Mar 30 16:15:45 <_boto> i don't know if something is installed on my machine which makes trouble for boost Mar 30 16:16:05 <_boto> i have boost also installed on my host before, but in the meantime i have removed it Mar 30 16:16:18 <_boto> the problem still exists though Mar 30 16:16:18 yo pb_ Mar 30 16:16:37 _boto: another dev reported issues with boost. Read the channel logs of last days. Mar 30 16:16:57 <_boto> ah, thanks for the hint ant_work Mar 30 16:18:53 <_boto> hmm, i cannot find the channel log :-/ Mar 30 16:19:21 http://www.hentges.net/irclogs/%23oe/ Mar 30 16:19:34 <_boto> thanks Mar 30 16:19:47 iirc yesterday Gnutoo and woglinde were talking about boost Mar 30 16:21:39 _boto, we were talking about implementing synd_add_and_fetch for arm < v6 (so no ldrex and strex support) but I have not done it yet...but I have others fixes for boost_1.36.0 Mar 30 16:21:57 <_boto> ah Mar 30 16:22:42 <_boto> my problem is building it with proper compiler: bitbake seems to take the host gcc instead of the cross compiler Mar 30 16:22:54 by the way...where can I find the documentation about the subs instruction...I looked with the search function in ARM Architecture Reference Manual but I didn't find the real doc about it Mar 30 16:22:56 ant_work: did you see koen answer? Mar 30 16:23:05 ant_work: update your patch with the comment he has suggested Mar 30 16:23:14 ant_work: and then we review it again Mar 30 16:23:52 _boto, mmm...and for the others packages it doesn't Mar 30 16:23:53 ? Mar 30 16:24:15 <_boto> no, all other stuff is built fine Mar 30 16:24:20 <_boto> only boost makes trouble Mar 30 16:24:44 <_boto> curiously, using the same build process on another machine here works Mar 30 16:24:47 ah ok...do you use the standard recipes? and which version of boost do you use? Mar 30 16:24:56 <_boto> it must have something to to with my host Mar 30 16:25:08 <_boto> i use standard recipes from oe Mar 30 16:25:14 I had this but it was for things like the autotools Mar 30 16:25:16 <_boto> i look for version Mar 30 16:25:22 <_boto> aha Mar 30 16:25:27 <_boto> how did you solve it? Mar 30 16:25:30 yes Mar 30 16:25:39 <_boto> boost_1.36.0.bb Mar 30 16:25:42 ok Mar 30 16:25:45 <_boto> i use 1.36 Mar 30 16:26:03 _boto, try my version from bugzilla Mar 30 16:26:19 but I bet it still won't work Mar 30 16:26:22 <_boto> i see arm-intrinsics.patch is applied Mar 30 16:26:28 yes Mar 30 16:26:39 <_boto> Gnutoo, you have a recipe? cool Mar 30 16:27:03 <_boto> would you mind to post the link to your bugzilla entry? Mar 30 16:27:14 http://bugs.openembedded.net/show_bug.cgi?id=4977 Mar 30 16:27:21 <_boto> thx alot Mar 30 16:28:12 _boto, but it may not work...I think the problem is with the state of your staging dir Mar 30 16:28:29 _boto, which gcc version do you have Mar 30 16:28:45 _boto, from oe Mar 30 16:28:47 <_boto> 4.3.2 Mar 30 16:28:52 <_boto> oh, i look Mar 30 16:29:04 yes there is using gcc 4.1.1 somewhere in the recipe Mar 30 16:29:11 it was from before so I didn't change it Mar 30 16:29:33 <_boto> so i would need to update it then? Mar 30 16:29:48 ah ok 4.3.1 Mar 30 16:29:53 try it Mar 30 16:30:04 I don't know boost's jam enough Mar 30 16:30:25 jam = a build system Mar 30 16:31:30 <_boto> i will try it Mar 30 16:32:41 <_boto> i see 4.3.1 is used as cross compiler Mar 30 16:33:21 <_boto> thanks a lot guys, i will be on tomorrow again Mar 30 16:33:32 <_boto> now i gotta go to do a bit sports :-) Mar 30 16:33:35 <_boto> cu Mar 30 16:34:54 ah ok it was in sub instruction Mar 30 16:35:04 ok cu Mar 30 16:39:24 sed, can you push the normal upstream tags over please? Mar 30 16:39:26 ga Mar 30 16:47:20 see you later Mar 30 17:10:37 hi jay7 Mar 30 17:10:49 pwgen: hey Mar 30 17:11:36 Jay7:i try to build the initramfs-kexeboot image for the openmoko and have problems getting it to work Mar 30 17:12:04 pwgen: that's not new ;) Mar 30 17:12:08 because the openmoko reads the 2 buttons from diffrent eventinterfaces . Mar 30 17:12:22 pwgen: yes, I know this already and have solution :) Mar 30 17:12:28 . i got it running useing a littel patched qi Mar 30 17:12:31 but not implemented yet, sorry Mar 30 17:12:58 Jay7: we are still waiting for voice-commands :-) Mar 30 17:13:01 i am think to add support for the touchscreen Mar 30 17:13:03 one key should be assigned as down, one - as select :) Mar 30 17:13:11 and menu should wrap Mar 30 17:13:23 Jay7: same thoughts I had Mar 30 17:13:27 pwgen: is your solution the same? :) Mar 30 17:14:01 XorA: koen notes about this sometime ago Mar 30 17:14:09 so that is not my idea :) Mar 30 17:14:23 Jay7: I just cant be bothered with gta02 any more Mar 30 17:14:43 i did not start any solution until i had spoken to someone who is involved to kecexboot Mar 30 17:15:12 XorA: I'll fix xpm stuff yesterday so I'll continue to work :) Mar 30 17:15:21 *FG* maybe shake the menue down with useing the motion sensors .. Mar 30 17:15:35 hm.. Mar 30 17:15:48 pwgen: boot the phone, by "booting the phone" ??? Mar 30 17:15:57 pwgen: when you can show me how to assign sensors to keys ;) Mar 30 17:16:14 ant_work: touchscreen first ;) Mar 30 17:16:19 then voice control :) Mar 30 17:16:49 jay7 : there came raw inputs from event2,3 but shere is an userspace daemon running to identify gesture ( seems trick to get him working with klibc ) Mar 30 17:17:18 hm.. Mar 30 17:17:33 bbl, bye Mar 30 17:18:28 but the aux button for selecting the entries and the powerbutton for starting should be enough, but its event0 and event4 to watch for Mar 30 17:19:33 createing a function for watching more than one event will help or adding an timeoutthread will also help Mar 30 17:19:35 pwgen: is event0 not enough just for keys? Mar 30 17:19:58 I'm thinking already about select() or poll() on input Mar 30 17:20:04 because of some things.. Mar 30 17:20:31 on the gta02 event0 is used for the powerbutton and will send event KEY_POWER and in event 4 came in the aux button eith KEY_AUX Mar 30 17:21:05 ehh.. that is bad.. Mar 30 17:21:42 morning all Mar 30 17:22:07 cbrake: have you had a chance to look at gtk-sharp? Mar 30 17:22:14 yes thas my problem , t can not solve this with changeing on line ...(:-(( Mar 30 17:23:14 pwgen: well.. I'm not sure that I can do it for next release Mar 30 17:23:19 but I'll try :) Mar 30 17:23:23 additional the rootfs creates only /dev/event0 and ther are additianl event interfaces needed Mar 30 17:23:43 garren: no, I have not yet Mar 30 17:24:02 cbrake: ok Mar 30 17:24:25 Jay7: i can write it for the GTA02 and send you my solution .. so you can decide if it helps .. Mar 30 17:27:23 and can someone please uptdate the git repository for kexecboot_git .. Mar 30 17:31:52 pwgen: did you mean solution with polling two input device? Mar 30 17:32:16 Jay7: that will be my first approach Mar 30 17:32:42 pwgen: well, that's good idea. Send me patch then :) Mar 30 17:34:42 Jay7:i am working with kexecboot_git.bb Mar 30 17:36:35 pwgen: that's ok Mar 30 17:36:53 pwgen: what update you will for kexecboot_git.bb? Mar 30 17:37:31 the surc uri is wrong and has moved to git://git.linuxtogo.org/home/groups/kexecboot/kexecboot.git Mar 30 17:38:10 and maybe she SRCREV should be changed to the lates revision Mar 30 17:40:56 ah, yes.. Mar 30 17:41:06 git uri should be changed Mar 30 17:41:19 I'll note ant when he will back :) Mar 30 17:41:36 and maybe add event0-4 to device_table-oldmmc.txt (:-)) Mar 30 17:42:25 i have written a bug report regarding the SRC_URI bug id 5074 Mar 30 17:42:26 well, that can be good idea Mar 30 17:43:16 pwgen: we have separate channel #kexecboot btw Mar 30 17:43:35 but there is silence most of time :) Mar 30 17:45:16 http://kergoth.pastey.net/111232 .. whee, this is fun Mar 30 17:45:43 kergoth: nice! Mar 30 17:46:33 kergoth: PR removal is when it is r0? Mar 30 17:46:43 yep, when it matches what was in bitbake.conf Mar 30 17:47:00 i checked the ConfigParsed event and compared the .conf vars against the d in a task, for a certain few common vars Mar 30 17:47:10 among other things, the FILESDIR/FILESPATH are a bit fancier Mar 30 17:47:21 and the autotools_base, those are recipes that are overriding the config so autoreconf doesn't end up executed Mar 30 17:51:12 like2wise: http://kergoth.pastey.net/111234 Mar 30 17:51:17 that's the current incarnation Mar 30 18:12:41 03Koen Kooi  07org.openembedded.dev * rd7fdcef3d8 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Mar 30 18:12:51 03Roger Monk  07org.openembedded.dev * rf1315f451d 10openembedded.git/recipes/dsplink/ (ti-codec-combos_3.16.bb ti-codec-engine_2.23.bb): ti-codec-combos: Use DVSDK Codec Combos to be aligned with DVSDK Mar 30 18:26:23 any gcc guru around? Mar 30 18:26:38 Im trying to compile a program which contains its own printf Mar 30 18:26:49 gcc of course insist on optimizing it to puts Mar 30 18:27:01 which wont be available once this program starts Mar 30 18:27:15 earlier we could have -fno-builtin-printfs Mar 30 18:27:22 but seems like that is removed Mar 30 18:40:49 kristoffer_: -fno-builtin-printfs remove the trailing s Mar 30 19:03:05 <_cpo_> anybody ever tried building xserver 1.6 with oe? i always run into a libtool error Mar 30 19:24:59 03Koen Kooi  07org.openembedded.dev * r26a4aa3ea3 10openembedded.git/recipes/webkit/webkit-gtk_svn.bb: webkit-gtk: fix PV Mar 30 19:31:13 03Koen Kooi  07org.openembedded.dev * r56ff0ec54b 10openembedded.git/recipes/gcc/ (gcc-4.3.1.inc gcc-4.3.2.inc gcc-4.3.3.inc): gcc 4.3.x: disable thumb mode for now, glibc builds don't like it Mar 30 19:33:41 re Mar 30 19:48:31 I have a question regarding the move that has been done from packages to recipes. How can I browse the old log info at http://cgit.openembedded.net/ for a specific recipe? With the move, the log info of any file just includes the rename message and precious info about past log doesn't seem to be accessible any more. Mar 30 19:51:20 gregoiregentil: you can use at least tag mv-packages-to-recipes-pre Mar 30 19:51:31 this was tagged right before Mar 30 19:53:38 Jay7: Very good! That's exactly what I needed. Many thanks for the answer. I definitely appreciate. Mar 30 19:53:58 gregoiregentil: :) Mar 30 19:55:17 likewise or like2wise, ping Mar 30 19:56:04 eFfeM: pong Mar 30 20:02:57 gregoiregentil: or "git log --follow filename" should handle rename too Mar 30 20:03:59 but I'm not sure how to tell cgit/gitweb to follow the rename... Mar 30 20:25:53 03Stanislav Brabec  07org.openembedded.dev * r9fdd2e01e2 10openembedded.git/recipes/mc/ (mc.inc mc_4.6.2.bb): mc: Use system libexecdir. Mar 30 20:55:07 03Jrg Bakker  07org.openembedded.dev * r164dfba075 10openembedded.git/recipes/linux/ (files/linux-2.6-limits.patch linux-epia_2.6.19.2.bb): linux-epia: Add patch to include in scripts/mod/sumversion.c Mar 30 20:57:30 03Stanislav Brabec  07org.openembedded.dev * ra021ada0a7 10openembedded.git/ (3 files in 3 dirs): gst-ffmpeg: Updated to version 0.10.7. Mar 30 20:57:41 03Stanislav Brabec  07org.openembedded.dev * re65b621d52 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Mar 30 20:58:02 03Tom Rini  07org.openembedded.dev * rd24f5daa7f 10openembedded.git/recipes/linux/ (2 files): linux-nokia: Add linux-2.6-limits.patch, bump PR Mar 30 20:58:03 03Tom Rini  07org.openembedded.dev * r90cebee2ea 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Mar 30 20:58:09 Damnit Mar 30 20:58:15 Forgot I hadn't run git config ;( Mar 30 20:58:25 Anyway to fix that up or too late? Mar 30 21:01:42 rofl Mar 30 21:01:47 awesome domain Mar 30 21:01:53 thanks Mar 30 21:02:25 * Tartarus spent too much time naming everything in his network to fit the theme Mar 30 21:37:12 * XorA|gone mehs as kexectools fails to compile Mar 30 21:40:17 do we have a current recipe best practices document or documents? Mar 30 21:50:27 * kergoth tries to think of other best practices that can be tested from a task Mar 30 21:51:39 XorA|gone: what is wrong with kexectools? Mar 30 21:52:43 Jay7: cant find mach/memory.h when compiling Mar 30 21:53:45 XorA|gone: hm.. is it in current .dev? Mar 30 21:53:56 Jay7: yes Mar 30 21:55:11 XorA|gone: I can test for akita Mar 30 21:55:18 03Andrea Adami  07org.openembedded.dev * rc7b72aee55 10openembedded.git/ (4 files in 2 dirs): Mar 30 21:55:18 kexecboot: clean recipes Mar 30 21:55:18 - fix SRC_URI and bump SRCREV for _git recipe Mar 30 21:55:18 - minor cleanings to the recipes Mar 30 21:55:18 - bump PR Mar 30 21:55:20 - remove old kexecboot-0.3 checksums Mar 30 22:05:46 XorA|gone: there are fresh kexec-tools out in the wild Mar 30 22:07:52 yeah debian seems to have some dated snapshot Mar 30 22:41:09 RP, mickeyl: ping Mar 30 22:48:22 kergoth: your pastey signales I can get rid of PR in a recipe. Is a bit unclear to me why... Mar 30 22:48:54 btw it was right..I removed the unuseful S = blah blah Mar 30 22:49:05 you can if it is PR = "0" Mar 30 22:49:25 ok, and if no PR ? I added "1" ;-) Mar 30 22:49:34 it reports you can remove PR if you're setting it to the default, at least that should be the case Mar 30 22:49:36 although, I kind of like using PR= "0" so people leave it in the asame place when it is updated Mar 30 22:50:00 was kexecboot Mar 30 22:50:02 coll Mar 30 22:50:04 well, its just a first pass checker, i don't expect it to be a list of everything to fix Mar 30 22:50:46 * kergoth thinks it's a good example of the types of problems he expects would go away with more individual responsibility wrt recipe maintanance Mar 30 22:51:13 kergoth: I feel guilty now... I added a PR Mar 30 22:55:01 i don't think there's any documented set of best practices for recipes.. it would be nice to write something up Mar 30 22:58:51 kergoth: anyway, when we'll bump git SRCREV for the recipe, we'll have to PR-bump th ekernels containing the binary in initramfs too...this can't be automated atm Mar 30 23:06:24 'nite Mar 30 23:28:44 hmm, wonder if itd be good to create something like COLLECTIONS = space separated list of paths to the root of metadata repositories, in priority order, and use that to automatically populate BBFILES, BBFILE_COLLECTIONS, and the collection patterns and priority variables Mar 30 23:39:53 good night all Mar 31 01:06:25 sdlmame is really being a pain to get to compile under OE. x.x **** ENDING LOGGING AT Tue Mar 31 02:59:57 2009