**** BEGIN LOGGING AT Sun Oct 21 02:59:56 2007 Oct 21 07:58:18 morning Oct 21 07:59:14 mickey|tw : hi ! Oct 21 08:04:02 good morning all Oct 21 08:17:16 mickey|tw: how's taiwan treating you? Oct 21 08:20:28 koen : morning Oct 21 08:20:40 hey steliosk Oct 21 08:20:52 koen : have a look a this. -> http://rafb.net/p/0awCmD14.html Oct 21 08:21:16 I did a fresh build with no image definitions in the local.conf and this is what it comes to Oct 21 08:21:48 it defuals to jffs2 and fails Oct 21 08:25:00 cute Oct 21 08:29:51 steliosk: that's a bug in bitbake Oct 21 08:30:11 steliosk: please file a bug and mtn pull, r c47ba1583ffd46886ad2d44d3587ad3d64affd16 has an ugly workaround Oct 21 08:30:41 * steliosk loves the smell of fresh bugs in the mornings Oct 21 08:33:02 koen : should i put the bug under bitbake or angstrom ? Oct 21 08:33:15 bitbake or OE Oct 21 08:33:19 it's not angstrom specific Oct 21 08:33:35 bbimage is a bitbake function, sadly Oct 21 08:34:17 koen : any idea why it ignores the image settings and defaults to jffs2 ? Oct 21 08:35:11 no idea on that Oct 21 08:39:51 koen : Do you have time to look at why the [456]86 x11-images don't get x started ? Oct 21 08:40:15 koen : i will be looking at it today also Oct 21 08:40:17 not the next few weeks Oct 21 08:40:27 (no x86 hardware I can test on) Oct 21 08:41:10 koen : ok. i'll try and fix the udev error message and try to find out why avahi also fails Oct 21 08:55:38 03koen 07org.oe.dev * r8aff1229... 10/ (1 packages/tasks/task-base.bb): task-base: add DISTRO_EXTRA_DEPENDS (workaround for bitbake bug) Oct 21 08:55:43 03koen 07org.oe.dev * rc47ba158... 10/ (1 conf/distro/angstrom-2007.1.conf): angstrom-2007.1.conf: add workaround for missing mtd-utils-native Oct 21 09:10:36 hi all Oct 21 09:29:53 koen: ok to commit http://en.pastebin.ca/744294 ? I just copied the feed conf, preferred providers and distro extra variables for now. Oct 21 09:33:48 pH5: could you split the feed-arch stuff into a .inc file? That will never change :) Oct 21 09:33:53 the rest looks ok Oct 21 09:34:10 moin Oct 21 09:34:42 hey zecke Oct 21 09:34:48 koen: ok Oct 21 09:34:50 moin, moin :) Oct 21 09:35:18 * zecke has new stuff in his flat... chairs and tables... :( Oct 21 09:37:30 koen: what is include/angstrom.inc? it also has the feed-arch variables (included by angstromoabi.inc, which is not included anywhere afaics) Oct 21 09:37:51 ehm.... Oct 21 09:38:09 angstrom.inc is supposed to be distro independant stuff Oct 21 09:38:16 s/distro/release/ Oct 21 09:38:21 like FEED_ARCH :) Oct 21 09:40:27 and I'm stupid, it is included already Oct 21 09:40:33 "require conf/distro/include/angstrom${ARM_ABI}.inc" Oct 21 09:42:13 and it's even up to date. --> http://en.pastebin.ca/744300 ? Oct 21 09:44:34 great Oct 21 09:45:37 I guess the commented-out SRCDATE lockdown in 2007.1 can go away, too. Oct 21 09:45:49 I think so Oct 21 09:50:11 zecke: congrats Oct 21 09:50:13 mickey|tw: is stefan_schmidt still in .tw? Oct 21 09:51:02 koen: I'm. Meeting right now. Oct 21 09:51:39 stefan_schmidt: I changed the xserver script to "just work" with qvga, could you test it when you get some time later on? (it's in OE already) Oct 21 09:52:11 Ophalen:201 http://au.archive.ubuntu.com gutsy/main libgl1-mesa-dri 7.0.1-1ubuntu3 [13,0MB] Oct 21 09:52:16 13 megs for a lib... Oct 21 09:52:22 mickey|tw: I think I will upload pictures tomorrow... Oct 21 09:52:36 * zecke is even able to prepare pizza now... Oct 21 09:52:41 zecke: cool Oct 21 09:52:50 koen: I've seen already thanks. I'll test it tomorrow. Oct 21 09:53:03 zecke: yay, welcome in our world ;) Oct 21 09:53:35 zecke: :-) Oct 21 09:53:56 and I'm hooked to a gas oven... Oct 21 09:54:00 zecke: sounds like I need to visit soon (this time *with* a present) :) Oct 21 09:56:09 sure Oct 21 09:57:02 pH5: I just had an idea about stamps and rm_work Oct 21 09:57:34 pH5: if someone passes -f have rm_work.bbclass delete the stamps if a rm_work stamp is present Oct 21 09:57:50 I don't know how feasible that is, though Oct 21 09:57:55 * koen looks at zecke and RP Oct 21 09:58:23 zecke: http://bugs.openembedded.org/show_bug.cgi?id=3177#c2 Oct 21 09:58:47 argh... Oct 21 09:59:04 * zecke does something else useful before looking into it... Oct 21 10:14:40 03pH5 07org.oe.dev * ra20ea618... 10/ (3 files in 3 dirs): Oct 21 10:14:40 angstrom-2007.1.conf: already uses include/angstrom.inc, Oct 21 10:14:40 no need to duplicate the FEED_ARCH variables Oct 21 10:14:46 03pH5 07org.oe.dev * rdf62641d... 10/ (1 conf/distro/angstrom-2008.1.conf): Oct 21 10:14:46 angstrom-2008.1.conf: stop including angstrom-2007.1.conf, Oct 21 10:14:46 which would made 2008.1 follow the version lockdown. copy the Oct 21 10:14:46 relevant variables instead. Oct 21 10:15:22 s/would // Oct 21 10:20:40 koen: not sure if we have a bitbake variable so tasks can tell whether they were called forcedly, but my problem was the other way around (bitbake rebuilding stuff when unneeded stamps are missing) Oct 21 10:24:30 ~ping RP Oct 21 10:24:31 pong RP Oct 21 10:28:21 zecke : Any idea why bb.data.setVar('PARALLEL_MAKE' , ' ', d) will not override the PARALLEL_MAKE setting from a local.conf file ? Oct 21 10:28:41 steliosk: where do you set that? Oct 21 10:28:54 in icecc.bbclass Oct 21 10:29:12 it used to work, but now noticed that its not anymore Oct 21 10:29:54 i use it in icecc.bbclass to override the PARALLEL_MAKE settings for packages that should not/can not use icecc Oct 21 10:29:59 hmm, stupid question. You don't set it in do_configure() and expect it to work in do_compile()? Oct 21 10:31:27 steliosk: woglinde's favorite tool is "bbread" (now in bitbake itself). Use that to take a look Oct 21 10:31:59 its set both in configure and compile Oct 21 10:32:24 steliosk: hmm, it should work... Oct 21 10:33:15 zecke : it used to with bitbake 1.6 Oct 21 10:33:55 haven't paid much attention to it with 1.8, but now that i found some free time to work on getting SDK packages to use icecc, i noticed it Oct 21 10:33:59 steliosk: the package in question doesn't have something like PARALLEL_MAKE_foo? Oct 21 10:34:29 steliosk: you can do something like bb.data.update_data("PARALLEL_MAKE", d)\nthe old set data Oct 21 10:34:36 and check if that makes a difference.. Oct 21 10:35:00 will try Oct 21 10:44:41 zecke : It's getting even worst -> http://rafb.net/p/DACeE116.html Oct 21 10:44:59 ;) Oct 21 10:45:31 I said do it the other way around, but that doesn't really matter :} Oct 21 10:47:14 zecke : You meant to do the bb.data.update_data before changing the var ? Oct 21 10:47:25 yes, but the result should be the same Oct 21 10:48:43 EXTRA_OEMAKE_prepend_task_do_compile = "${PARALLEL_MAKE} " Oct 21 10:49:18 you could use the anonfunc (called when the bb file got loaded) to filter that variable, or just set to something else Oct 21 10:49:33 better filter it (check the openssl recipes on how to filter stuff) Oct 21 10:50:15 zecke : But i'll have to change all problematic files by hand ? Oct 21 10:52:40 steliosk: no, you already know the problematic PNs? have an anonfunc "python () {}" in icecc.bbclass and filter there Oct 21 10:53:57 steliosk: alternatively EXTRA_OEMAKE_prepend_task_do_compile_pn-PROBLEMATIC_ONE_1 = "" do something like this in either icecc.bbclass (don't know if it is working there) or an .inc Oct 21 10:55:06 zecke : Well the idea was to remove PARALLEL_MAKE from .bb files that get excluded from been build with icecc, eitherwise you end up with -j20 running on a single cpu, which actually slows things down Oct 21 10:55:45 zecke : Wouldn't that bug effect other classes also ? Oct 21 10:56:02 steliosk: probably ;) Oct 21 10:56:32 I hate the excuse but I need to do some other things now :} Oct 21 10:57:24 zecke : not a nice picture -> http://rafb.net/p/DduGZb44.html Oct 21 11:00:38 steliosk: I doubt that it is broken in general. You might see a nice/weird order that.... Oct 21 11:39:14 Happy Sunday ;) Oct 21 11:41:18 Has anyone seen this error when compiling gcc-cross-initial 4.2.2 ? Oct 21 11:41:22 checking for unistd.h... configure: error: C compiler cannot create executables Oct 21 11:41:38 check the config.log to see why it fails Oct 21 11:42:05 (yes, the config.log is in a non-obvious place with gcc, stupid gnu people) Oct 21 11:44:24 Well, that's the bloody problem :( I seem to be unable to find the correct one :( Oct 21 11:44:38 At least You gave me a hint, I'll start finding and grepping ;) Oct 21 11:45:11 IIRC it's buried somewhere in the builddir (not the sourcedir!) Oct 21 11:45:33 ~seen florian Oct 21 11:45:36 florian was last seen on IRC in channel #oe, 1d 13h 42m 31s ago, saying: 're'. Oct 21 11:46:05 Oh, another side question, was removing of xlsfonts, mkfontscale, xkbcomp and xkbutils a final decision or can I get them back by submitting working .bb files to bugzilla ? Oct 21 11:46:27 you can get them back by submitting fixed recipes to bugzilla Oct 21 11:46:58 great, I'll do that. I try to locate the offending config.log and then I might be back for some gcc advice since compiling this beast is not my speciality :( Oct 21 11:50:29 hey mickey|tw - you in Taiwan or? Oct 21 11:51:41 yep Oct 21 11:55:28 Ok, I located the offending file. since make -j 2 was confusing the logs it was here: gcc-4.2.2/build.i686-linux.i586-linux/i586-linux/libgomp/config.log. The root cause of the error however is /home/anttix/OpenEmbedded/bleeding/build/tmp/cross/i586-linux/bin/ld: crt1.o: No such file: No such file or directory Oct 21 11:55:41 The entire log is here http://pastebin.com/m6fb55f Error is at line 106 Oct 21 11:56:18 is it possible, that crt1.o has not been built yet? Thus gcc-cross-initial fails when PARALLEL_MAKE is set? Oct 21 11:56:52 anttix : are you using icecc.bbclass for the build ? Oct 21 11:57:31 Hmm, I don't know ;) If it's the new default, then yes ;) Oct 21 11:58:04 nop, its not set by default Oct 21 11:58:11 then I shouldn't Oct 21 11:58:52 I have those set: PREFERRED_PROVIDERS += " virtual/${TARGET_PREFIX}libc-for-gcc:glibc-intermediate"PREFERRED_PROVIDERS += " virtual/${TARGET_PREFIX}libc-initial:glibc-initial" Oct 21 11:58:52 And also glibc for virtual/libc, libiconv and libintl Oct 21 11:59:40 anttix : what i am asking is more related to the PARALLEL_MAKE and if you use icecream for distributed compilling Oct 21 11:59:54 not the gcc itself Oct 21 12:00:13 steliosk: ok, since I don't know about it, I probably don't use it ;) Oct 21 12:00:47 steliosk: I do have PARALLEL_MAKE = "-j 2" Oct 21 12:01:46 Gcc 4.1.1 however seems to work ok with my setup. When I set PREFERRED_VERSION_gcc-cross-initial = "4.1.1" gcc and glibc build flawlessly Oct 21 12:01:49 anttix : I don't think that PARALLEL_MAKE could be the issue in your case, but try it without it Oct 21 12:02:03 ok, I'll try a build without it and see what happens. Oct 21 12:02:15 aah Oct 21 12:08:23 join /#opensimpad.org Oct 21 12:08:28 sorry Oct 21 12:28:08 steliosk you were right. disabling PARALLEL_MAKE didn't make any difference :( Oct 21 12:35:39 ~ping CIA Oct 21 12:35:40 pong CIA Oct 21 12:35:51 ~CIA Oct 21 12:35:52 cia is probably a CVS/Subversion commit notification bot run on #commits among other channels, originally by scanline, http://cia.vc/, or the Central Intelligence Agency, but the bot is more important. ;) Oct 21 12:37:01 03mickeyl 07org.oe.dev * r8c39fbca... 10/ (1 packages/gstreamer/gst-meta-base_0.10.bb): gst-meta-base: fix typo in RDEPENDS. please watch your diffs more closely Oct 21 12:37:09 03pH5 07org.oe.dev * r1e94b210... 10/ (1 conf/distro/angstrom-2008.1.conf): Oct 21 12:37:09 angstrom-2008.1.conf: use preferred versions includes (e, gpe2.8, openmoko) Oct 21 12:37:09 * in some cases these preferred versions upgrade, not downgrade packages Oct 21 12:37:09 due to DEFAULT_PREFERENCE settings (for example libmimedir in gpe2.8) Oct 21 12:37:15 03pH5 07org.oe.dev * r0186939e... 10/ (3 files in 3 dirs): Oct 21 12:37:15 angstrom-2007.1.conf: move wpa-supplicant pref version into -preferred-versions.inc Oct 21 12:37:15 instead of setting it there to 0.4.9 and overwriting to 0.5.5 in .conf Oct 21 12:38:44 now angstrom-2008.1 has 0.4.9 and angstrom-2007.1 has 0.5.5 - why does wpa-supplicant 0.5 have DEFAULT_PREFERENCE="-2" set but angstrom uses it anyway? Oct 21 12:39:34 no idea Oct 21 12:39:42 we should merge 0.5.8 from poky anyway Oct 21 12:39:49 that makes network-manager work :) Oct 21 12:40:55 I think there are quite some packages flying around with DEF_PREF="-1" where people just forgot to remove it. Oct 21 12:41:04 * * OE Bug 3183 has been created by autobuild(AT)openembedded.org Oct 21 12:41:06 * * maxima-5.13.0-r0-do_configure Oct 21 12:41:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3183 Oct 21 12:41:23 This is quite surprising in cases where you remove the preferred version and suddenly older packages are used. Oct 21 12:42:39 Actually I interpreted the results wrong. It did make a difference, it now dies with: | In file included from /home/anttix/OpenEmbedded/bleeding/build/test/tmp/cross/i586-linux/include/stdlib.h:438, Oct 21 12:42:39 | from /home/anttix/OpenEmbedded/bleeding/build/test/tmp/work/i586-linux/gcc-cross-initial-4.2.2-r1/gcc-4.2.2/libiberty/regex.c:128: Oct 21 12:42:39 | /home/anttix/OpenEmbedded/bleeding/build/test/tmp/cross/i586-linux/include/sys/types.h:100: error: two or more data types in declaration specifiers Oct 21 12:43:02 I'm all out of ideas. Maybe I should file a bug report? Oct 21 12:43:10 who maintains gcc anyway? Oct 21 12:43:17 I mean gcc recipies in OE. Oct 21 12:49:56 mickey|away: not going to head down to Australia while you are in this part of the world? Oct 21 12:54:04 * * OE Bug 3183 has been RESOLVED (INVALID) by Oct 21 12:54:06 * *  maxima-5.13.0-r0-do_configure Oct 21 12:54:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3183 Oct 21 12:56:23 03koen 07org.oe.dev * r2548d574... 10/ (1 conf/distro/include/angstrom-2007-preferred-versions.inc): angstrom-2007-preferred-versions.inc: remove some more svn packages Oct 21 12:57:49 mickey|tw|dinner: if you ever end up in Adelaide, please do look me up Oct 21 12:58:49 zecke, mickey|tw|dinner: http://article.gmane.org/gmane.comp.version-control.monotone.devel/12807 Oct 21 13:17:32 03freyther 07org.oe.dev * re2ef6c7c... 10/ (9 files in 5 dirs): Oct 21 13:17:32 packages/webkit: Build the Qt/Qtopia port of WebKit as well. Oct 21 13:17:32 Unifiy the WebKit recipe and allow to build the Gtk+, Qt Oct 21 13:17:32 and Qtopia port. Add classes/qtopia4core.bbclass to assist Oct 21 13:17:32 building for Qtopia. Change the WebKit.pro to currently not Oct 21 13:17:33 build the testkjs and DumpRenderTree test utilities. Oct 21 13:20:29 Bleh! Does anybody know why libx11-native_1.0.1.bb actually downloads and uses libx11 1.1.1 sources? Oct 21 13:20:46 Somebody forgot to increase version number in filename? Oct 21 13:20:50 Bug? Oct 21 13:23:54 feature? undocumented? Oct 21 13:24:34 zecke: what do you think about moving to a more recent srcrev for webkit* ? Oct 21 13:25:24 koen: for WebKit/Gtk+ it should compile and work(tm) (there were massive changes in trunk, a branch got merged) Oct 21 13:27:55 koen : i fixed the dbus problem, but avahi still fails to load on its own. If i manually do an ifup then avahi starts. any ideas ? Oct 21 13:28:58 which 'avahi' do you mean? Oct 21 13:29:05 avahi-daemon or something else? Oct 21 13:31:53 avahi-daemon is started by sysvinit Oct 21 13:32:02 avahi-autoipd is started by ifup Oct 21 13:32:59 avahi-daemon Oct 21 13:33:25 I still suspect a broken kernel config Oct 21 13:33:28 but i just noticed that the init scripts are missing network, ifplugd and a few others Oct 21 13:33:39 since avahi-daemon works on all my devices, but not on e.g. gumstix Oct 21 13:34:04 i don;t think its a kernel, if i manualy do an ifup eth0 and then /etc/init.d/avahi start it starts Oct 21 13:34:54 weird Oct 21 13:37:45 koen : is there any specific reason that only this init scripts get installed ? -> http://rafb.net/p/dWlrQM68.html Oct 21 13:37:58 network is missing for example Oct 21 13:38:13 no idea Oct 21 13:39:20 koen : can you post which ones are used on Angstrom now Oct 21 13:39:35 that depends on the image Oct 21 13:39:36 from a working machine Oct 21 13:39:45 x11-image Oct 21 13:40:37 from an openmoko-image: http://rafb.net/p/4KnEWB82.html Oct 21 13:47:53 koen : yep, init scripts are missing from the run levels Oct 21 13:48:12 koen : added S20networking and avahi-daemon starts Oct 21 13:56:05 * * OE Bug 2714 has been RESOLVED (FIXED) by antti.andreimann(AT)mail.ee Oct 21 13:56:07 * *  lock mtd-utils to a specific git version so builds will not break unexpectedly Oct 21 13:56:09 * * http://bugs.openembedded.org/show_bug.cgi?id=2714 Oct 21 13:56:17 * * OE Bug 3184 has been created by antti.andreimann(AT)mail.ee Oct 21 13:56:19 * * mate xorg-app/mkfontscale recipies Oct 21 13:56:21 * * http://bugs.openembedded.org/show_bug.cgi?id=3184 Oct 21 13:56:25 morning Oct 21 13:56:29 * * OE Bug 3185 has been created by antti.andreimann(AT)mail.ee Oct 21 13:56:31 * * mate xorg-app/xkbutils recipie Oct 21 13:56:33 * * http://bugs.openembedded.org/show_bug.cgi?id=3185 Oct 21 13:56:41 * * OE Bug 3186 has been created by antti.andreimann(AT)mail.ee Oct 21 13:56:43 * * mate xorg-app/xlsfonts recipie Oct 21 13:56:44 03ifaistos 07org.oe.dev * rb150bd2e... 10/ (1 classes/icecc.bbclass): classes/icecc.bbclass : Add support for the SDK packages Oct 21 13:56:45 * * http://bugs.openembedded.org/show_bug.cgi?id=3186 Oct 21 13:56:56 03ifaistos 07org.oe.dev * r380d0bd7... 10/ (1 packages/dbus/dbus.inc packages/dbus/dbus_1.0.2.bb): Oct 21 13:56:56 packages/dbus/dbus.inc : Add group netdev during installation (if it does not exists). Oct 21 13:56:56 Closes #3179 Oct 21 13:56:56 packages/dbus/dbus_1.0.2.bb : Bump PR Oct 21 14:10:03 koen if You have time, please consider applying patches from #3184, #3185, #3186, #3187 and #3188. Thank you *blink*. Oct 21 14:11:04 * * OE Bug 3179 has been RESOLVED (FIXED) by skoroneos(AT)digital-opsis.com Oct 21 14:11:06 * *  dbus fails to start on x86/Angstrom Oct 21 14:11:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3179 Oct 21 14:17:17 koen : any idea what could be fscking the init scripts for x86 ? Oct 21 14:17:38 i can't see to find anything strange Oct 21 14:21:35 my guess would be a malfunction update-rc.d Oct 21 14:24:48 what's the minimum rc2.d and rc5.d should contain ? Oct 21 14:25:04 * * OE Bug 3187 has been created by antti.andreimann(AT)mail.ee Oct 21 14:25:06 * * mate xorg-app/ xkbcomp recipie and add xkbcomp-native Oct 21 14:25:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3187 Oct 21 14:25:16 * * OE Bug 3188 has been created by antti.andreimann(AT)mail.ee Oct 21 14:25:18 * * ibx11-native package Oct 21 14:25:20 * * http://bugs.openembedded.org/show_bug.cgi?id=3188 Oct 21 14:27:52 question: OE will build repository (yum repo) if rpm packaging is choosed ? Oct 21 14:29:12 it will likely blow up Oct 21 14:29:18 Real People don't use rpm Oct 21 14:29:24 *gig* Oct 21 14:30:03 A question about bitbake. It used to be possible to instruct bitbake to download all sources at once by issuing bitbake -fc fetch whatever-image Oct 21 14:30:05 koen: s/real/sane/ :) Oct 21 14:30:15 http://blog.haerwu.biz/2006/10/17/rpm-is-really-poorly-maintained/ Oct 21 14:30:24 Good morning all Oct 21 14:30:31 now it doesn't work anymore :( Is it a bug or is there a better way now? Oct 21 14:30:40 -c fetchall Oct 21 14:31:55 Yipikayee! It works ;) Thanx koen! Oct 21 14:32:29 I'm trying to understand an error I'm seeing. do_patchcleancmd syntax error near unexpected token 'else' Oct 21 14:32:51 1.) there are no patches being applied Oct 21 14:33:06 2.) I don't see an else anywhere in that method: do_patchcleancmd() {if [ -n "`quilt applied`" ]; then quilt pop -a -R -f || exit 1; fi} Oct 21 14:35:03 you problably have a leading space somewhere in local.conf Oct 21 14:35:18 Is there anything obviously wrong with this recipe? Oct 21 14:35:20 http://pastebin.ca/744445 Oct 21 14:35:33 koen: local.conf in OE? Oct 21 14:36:13 no local.conf in ubuntu Oct 21 14:36:40 koen, sorry i don'r understand if is a 'yes' or 'no' .... Oct 21 14:36:57 real people don't use rpm ... real gremlin do ! :P Oct 21 14:38:09 hmm. This is FC6 not Ubantu, but I gather you're addressing gremlin[it] not me Oct 21 14:39:02 * mwester is pleased to notice that he must not be a real person either... (that explains *so much more* than just his use of rpm) :p Oct 21 14:40:04 good morning (and a special good morning to my fellow RPM users) Oct 21 14:40:53 heh Oct 21 14:40:57 hillct: you should probably read that as "yes, of course local.conf in OE, where else did you see a local.conf lately" Oct 21 14:41:04 * * OE Bug 3189 has been created by antti.andreimann(AT)mail.ee Oct 21 14:41:06 * * akeroot package to fix download failure Oct 21 14:41:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3189 Oct 21 14:42:09 pH5: I suppose so. I need more morning coddee before I can reliable interpret passive agressive sarcasm over IRC successfully Oct 21 14:42:20 er Morning Coffee Oct 21 14:44:09 is it just me or have ftp.info-zip.org download failures became a standard? I mean when was the last time somebody was able to download zip from there? Oct 21 14:45:02 not sure of the last time I tried Oct 21 14:49:52 heh, no local.conf at all Oct 21 14:50:00 where else might I look? Oct 21 15:12:04 03mickeyl 07org.oe.dev * rf8424d0f... 10/ (1 conf/distro/include/sane-srcrevs.inc): sane-srvrevs: bump openmoko-mediaplayer2 and openmoko-theme-standard2 Oct 21 15:20:21 03mailto:antti.andreimann 07org.oe.dev * rab000972... 10/ (4 files in 3 dirs): xkbcomp: reinstate recipe, closes #3212 Oct 21 15:20:27 03mailto:antti.andreimann 07org.oe.dev * r9839cfc4... 10/ (1 packages/xorg-app/xlsfonts_1.0.1.bb): xlsfonts: reinstate it, closes #3186 Oct 21 15:20:35 03mailto:antti.andreimann 07org.oe.dev * r2bf48b18... 10/ (1 packages/xorg-app/xkbutils_1.0.1.bb): xkbutils: reinstate it, close #3185 Oct 21 15:20:42 03antti.andreimann 07org.oe.dev * r288863eb... 10/ (3 files in 2 dirs): mkfontscale-native: reinstate it, closes #3184 Oct 21 15:20:52 * koen coughs Oct 21 15:21:04 heh Oct 21 15:21:29 one bugnumber wrong and 3 addresses Oct 21 15:21:45 sounds like you should go sleeping for today Oct 21 15:21:56 it's 23.21 anyways Oct 21 15:22:13 or inhale less smoke Oct 21 15:22:25 * koen accidentally set a pan of oil on fire today Oct 21 15:22:59 hey lrg Oct 21 15:23:17 hey koen Oct 21 15:27:08 yo Oct 21 15:32:13 koen : any idea if the "default" angstrom x server settings would work on an ati radeon card on a pc ? Oct 21 15:32:18 03koen 07org.oe.dev * r8315fee7... 10/ (1 conf/distro/include/sane-srcrevs.inc): sane-srcrevs: bump some more srcrevs for openmoko, tested on gta01/angstrom Oct 21 15:32:23 03koen 07org.oe.dev * rdefba6ec... 10/ (1 conf/distro/include/sane-srcrevs.inc): sane-srcrevs: add note about ordering Oct 21 15:32:36 steliosk: dunno, I never used xserver-xorg from OE Oct 21 15:32:39 only kdrive Oct 21 15:32:51 kdrive also fails to load Oct 21 15:33:12 if you have a framebuffer, kdrive should work Oct 21 15:33:31 i do. radeonfb but it does not work :( Oct 21 15:33:37 and i am running out of ideas Oct 21 15:34:41 xserver-xorg gives me a strange error message -> _XSERVTransSocketOpenCOTSServer: Unable to open socket for inet6 Oct 21 15:35:01 can't remember what the kdrive message is Oct 21 15:35:35 that's harmless and means you don't have ipv6.ko loaded Oct 21 15:36:46 mickey|tw: I didn't bump dialer2, since trunk can't accept calls ( http://bugzilla.openmoko.org/cgi-bin/bugzilla/show_bug.cgi?id=949 ) Oct 21 15:37:25 really? how unfortunate. Oct 21 15:37:31 ~lart regressions Oct 21 15:37:31 * ibot crushes regressions with a full height scsi disk Oct 21 15:38:06 the pimlico stuff stopped crashing with later revs, so I bumped those as well Oct 21 15:42:52 good morning Oct 21 15:43:00 hey sakoman Oct 21 15:43:15 hey koen! Oct 21 15:43:20 sakoman: I just realised you 8x10 is 8x10 inch, not cm :) Oct 21 15:43:48 koen: correct :-) Oct 21 15:43:56 a really big sucker! Oct 21 15:44:46 koen: hmm . . . did something change in the image class? Did a pull update and a fresh build resulted in: run.image_cmd.21330: line 402: mkfs.jffs2: command not found Oct 21 15:46:10 even with my workaround? Oct 21 15:46:25 koen: what workaround might that be :-) Oct 21 15:46:37 http://monotone.openembedded.org/revision/info/c47ba1583ffd46886ad2d44d3587ad3d64affd16 Oct 21 15:47:01 I don't know what is causing mtd-utils-native not to get built, though Oct 21 15:49:53 koen: looks like I messed that change -- I will do another pull/update Oct 21 15:50:09 koen: that camera is older than I am Oct 21 15:50:26 Thanx koen for commiting my recipies. I have created two new bugs though for Your consideration: #3189 and #3190 ;) Oct 21 15:50:51 koen: I think it dates from the 1940's Oct 21 15:50:58 that's quite old Oct 21 15:52:26 the lens is fairly new though, got it new around 1990 Oct 21 15:52:28 sakoman: I was looking for ways to get 6x9, 6x12 pto 6x24 panoramic pictures and stumbled across a 6x17 graflock back Oct 21 15:52:52 gonna buy it? Oct 21 15:52:58 nah Oct 21 15:53:12 that back needs a field camera Oct 21 15:53:33 of course, that's what collecting is all about :-) Oct 21 15:54:29 I gave up because it can become an obsession Oct 21 15:54:46 There's an endless universe of cool old cameras Oct 21 15:54:50 I know what you mean :) Oct 21 15:54:56 blast! I now realized that mkfontscale packages were put into wrong directory. They are currently under packages. But they should be under xorg-app. Must be my typo. Can You move the files for me koen? Sorry for the blunder! Oct 21 15:55:27 anttix: I wait till someone complains and have that person move them :) Oct 21 15:55:32 03koen 07org.oe.dev * r8983d5fa... 10/ (1 conf/distro/include/angstrom-2007-preferred-versions.inc): conf/distro/include/angstrom-2007-preferred-versions.inc: and more fixes.... Oct 21 15:58:03 * * OE Bug 3190 has been created by antti.andreimann(AT)mail.ee Oct 21 15:58:05 * * kfontdir-native download location Oct 21 15:58:07 * * http://bugs.openembedded.org/show_bug.cgi?id=3190 Oct 21 16:00:38 koen: this was my favorite medium format camera: http://www.kenrockwell.com/tech/plaupix.htm Oct 21 16:02:01 I love the look of folders Oct 21 16:02:26 same here. I really regret selling it. Oct 21 16:02:47 http://cgi.ebay.nl/Plaubel-Makina-67-with-Nikkor-2-8-80mm-lens_W0QQitemZ280164353803QQihZ018QQcategoryZ710QQrdZ1QQssPageNameZWD1VQQcmdZViewItem Oct 21 16:02:49 $2k Oct 21 16:03:49 sakoman: I'm waiting for a 6x9 pinhole to clear the customs Oct 21 16:04:15 going to load some slide film and find some nature Oct 21 16:05:07 those are great fun! Oct 21 16:05:40 I did some pinhole shots with my 8x10. Very long exposures! Oct 21 16:07:00 I got inspired by http://flickr.com/photos/zebandrews/tags/pinhole/ Oct 21 16:09:32 koen: I can see why you were inspired, those are really nice images Oct 21 16:14:04 * * OE Bug 3187 has been RESOLVED (FIXED) by antti.andreimann(AT)mail.ee Oct 21 16:14:06 * *  reanimate xorg-app/ xkbcomp recipie and add xkbcomp-native Oct 21 16:14:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3187 Oct 21 16:14:16 * * OE Bug 3184 has been RESOLVED (FIXED) by antti.andreimann(AT)mail.ee Oct 21 16:14:18 * *  reanimate xorg-app/mkfontscale recipies Oct 21 16:14:20 * * http://bugs.openembedded.org/show_bug.cgi?id=3184 Oct 21 16:14:28 btw, what happens/will happen to xserver-common Oct 21 16:14:28 * * OE Bug 3185 has been RESOLVED (FIXED) by antti.andreimann(AT)mail.ee Oct 21 16:14:30 * *  reanimate xorg-app/xkbutils recipie Oct 21 16:14:32 * * http://bugs.openembedded.org/show_bug.cgi?id=3185 Oct 21 16:14:39 can't we pull that into OE and sync with xserver-kdrive-common somehow? Oct 21 16:14:41 * * OE Bug 3186 has been RESOLVED (FIXED) by antti.andreimann(AT)mail.ee Oct 21 16:14:43 * *  reanimate xorg-app/xlsfonts recipie Oct 21 16:14:45 * * http://bugs.openembedded.org/show_bug.cgi?id=3186 Oct 21 16:16:01 pH5: you have my vote for that :) Oct 21 16:20:39 funny, in today's 2008.1 x11-image the on-screen keyboard in gpe-dm doesn't emit any keypresses Oct 21 16:22:28 matchbox-keyboard works, looks like xkbd is broken with kdrive 1.4 Oct 21 16:23:51 Formula 1 GP last competition is going to be very interesting. Currently: Hamilton 12, Alonso 3, Räikkonen 2 Oct 21 16:24:12 55 laps to go. Oct 21 16:24:42 Hamiltons' car just stalled for 30 seconds. Almost like it was rebooted over the wire ;) Oct 21 16:24:57 I guess they don't use stable things like OE in F1-s ;) Oct 21 17:42:32 koen: I was able to solve the problem by romoving an errant comment from my recipe. Thanks Oct 21 17:44:27 Another question: I have a recipe where I've defined do_configure_prepend () and do_install__append() within each of which I perform operations specific to the main package but also some operations specific to sub-packages. Is there a concept of _append_${Sub-Package_name}() ? Oct 21 17:44:49 where I should be doing sub-package related things? Oct 21 17:45:08 hillct: What kind of things? Oct 21 17:45:26 hillct: "sub packages" are something that don't exist until the packaging stage... Oct 21 17:45:46 my build process depends upon sox-native to perform conversion of audio files, which are packaged into a sub-package callweaver-sounds Oct 21 17:45:54 ah Oct 21 17:46:04 ok, so I'll leave it as is then Oct 21 17:47:54 I was thinking I may be able to adjust the dependencies such that sox-native was a dependency of callweaver-sounds rather than callweaver, and the packaging operations that require sox, could be done in the packaging phase specifically for the callweaver-sounds package Oct 21 17:47:59 RP: is it the case that comments cannot appear after variable definitions in bitbake recipes? Oct 21 17:48:18 e.g. S = "foo" # foo is the new value for S Oct 21 17:48:37 rwhitby: Not sure offhand, that probably depends what the parser does with such things... Oct 21 17:48:53 RP: see line 416 of http://pastebin.ca/744614 Oct 21 17:49:21 (that's what caused hillct's syntax error) Oct 21 17:50:05 rwhitby: When you're inside a function definition, the parsing rules mean it will pass things verbatim Oct 21 17:50:29 RP: http://pastebin.ca/744445 was the recipe - line 5 Oct 21 17:51:31 rwhitby: Right, we don't support that then ;-) Oct 21 17:52:07 I hadn't looked at the docs recently, but as I recall that wasn't explicitly mentioned when comments are discussed Oct 21 17:52:11 stray space, the lack of spaces and comments are the bane or recipe writes :) Oct 21 17:52:14 in any case, god to know Oct 21 17:52:27 ere good to know Oct 21 17:53:15 in fact, nowhere can I see any mention of comments in http://bitbake.berlios.de/manual/ Oct 21 17:53:32 so strictly speaking, comments in general are not officiall supported :-) Oct 21 17:53:37 I think it was in a wiki somewhere Oct 21 17:54:04 rwhitby: I don't think half the fetchers or the task dependency logic is official either :/ Oct 21 17:54:40 so, my next question is in reference to this recipe: http://pastebin.ca/744644 Oct 21 17:55:18 zecke: Your cross compile badness patch doesn't work with sysroot :/ Oct 21 17:55:32 is there anything I should do in order to break out the dependency on sox-native to make it a dependency only of callweaver-sounds? Does it make sense to do anything at all in this regard? Oct 21 17:55:52 RP: is bitbake -s supposed to look like this: Oct 21 17:55:53 hostap-conf 0:1.0-r8 1:.-0 Oct 21 17:55:54 hostap-daemon 0:0.4.4-r2 0:.-4 Oct 21 17:55:54 hostap-modules 0:0.4.7-r4 0:.-4 Oct 21 17:55:56 ? Oct 21 17:56:20 koen: Probably not Oct 21 17:56:27 koen: How did you break it? Oct 21 17:57:10 by setting DISTRO=angstrom-2007.1 Oct 21 17:57:14 it gets even better Oct 21 17:57:23 File "/data/build/oe/bitbake/lib/bb/cooker.py", line 153, in showVersions Oct 21 17:57:23 prefstr = pref[0][0] + ":" + pref[0][1] + '-' + pref[0][2] Oct 21 17:57:24 IndexError: string index out of range Oct 21 17:57:27 hillct: Not really Oct 21 17:57:41 K. thanks Oct 21 17:57:49 RP: a few places in bitbake do something like newver = ver[1] + ver[2] Oct 21 17:57:52 and crashes Oct 21 17:58:03 s/+/'='/ Oct 21 17:58:06 gah Oct 21 17:58:19 RP: it is not too bad Oct 21 17:58:28 %s=%s % ver[1], ver[2] Oct 21 17:58:57 koen: That would only crash if they weren't strings Oct 21 17:59:13 koen: If they're not strings thats likely to be the real problem Oct 21 18:01:42 * RP -> afk (food) Oct 21 18:08:49 RP: hotplug-qcop 0:cvs-r0 c:v-s Oct 21 18:09:03 something is going very wrong in the field splitting Oct 21 18:28:13 RP: hi ! Oct 21 18:32:53 * koen reads http://www.atoker.com/blog/2007/10/21/a-good-week-for-text-and-graphics/ Oct 21 18:55:42 03koen 07org.oe.dev * r785f4833... 10/ (1 conf/distro/include/angstrom-2007-preferred-versions.inc): angstrom-2007-preferred-versions.inc: and even more fixes Oct 21 19:27:53 koen: That doesn't look good :/ Oct 21 19:52:31 RP: it has been that way for a few weeks, but since adding the lock-down file bitbake -s actually crashes Oct 21 19:55:36 03ifaistos 07org.oe.dev * r7bfeaee7... 10/ (1 conf/machine/x86.conf): Oct 21 19:55:36 conf/machine/x86.conf : Set default xserver to kdrive-vesa as its supported by most Oct 21 19:55:36 pc drivers Oct 21 19:55:43 03rpurdie 07org.oe.dev * r68cafb7a... 10/ (7 files in 2 dirs): glibc-intermediate: Factor common code into a .inc file Oct 21 19:55:48 03rpurdie 07org.oe.dev * rc8d1c0f6... 10/ (9 files in 2 dirs): glibc-initial: Factor common code into a .inc file Oct 21 19:58:22 koen: It only broke recently? Oct 21 19:58:39 * RP wonders what changed as I don't think it was bitbake... Oct 21 19:59:00 * RP -> afk again :/ Oct 21 19:59:07 bitbake -s has spat out bogus info for some weeks now Oct 21 19:59:23 only after adding a file with a zillion DEF_PREFS it started crashing Oct 21 19:59:44 koen: Can you pin down which entry does it? :) Oct 21 20:00:01 I did, and after that it crashed in a different place Oct 21 20:00:25 lib/bb/providers.py:230 used to be the bad one Oct 21 20:00:45 What I mean is I'm much more likely to look into it if I have something I have a line I can add to poky which makes it crash... Oct 21 20:00:47 now it is cooker.py:153 Oct 21 20:00:56 * RP really afk Oct 21 20:02:03 RP : whenever you find time could you look at #3182 also. It's not critical but its strange Oct 21 22:29:23 03rpurdie 07org.oe.dev * r5b721c26... 10/ (3 files in 2 dirs): openobex: Fix staging function Oct 21 22:29:29 03rpurdie 07org.oe.dev * ra760a71f... 10/ (1 packages/libxml/libxml2_2.6.29.bb): libxml: Stage binconfig file to the correct place (from poky) Oct 21 22:29:33 03rpurdie 07org.oe.dev * r4c79780f... 10/ (1 packages/bison/bison-native_2.0.bb): bison-native: Use STAGING_DATADIR (from poky) Oct 21 22:49:26 anyone play half life on the PS2? Oct 21 22:52:58 hmm.. what exactly does mtn do on "mtn db regenerate_caches" ? Oct 21 22:53:14 it just hogs one cpu and seems not to use much io Oct 21 22:56:02 roh: yep, very CPU intensive. Leave it overnight - do not expect it to finish within a couple of hours. Oct 21 23:11:24 03rpurdie * r977 10/ (4 files in 4 dirs): providers.py: Fix bogus preferred_version return values Oct 21 23:12:28 koen|away: That should help your problem a bit Oct 21 23:18:17 rwhitby slowly i really get the urge to hit some mtn developers with heavy and sharp things Oct 21 23:33:08 03rpurdie * r978 10/ (4 files in 4 dirs): taskdata.py: Fix 'depends' flag splitting Oct 21 23:48:16 roh: I feel the same. There has been considerable speculation as to what on earth mtn might possibly be doing with so much CPU for so relatively small data. :( Oct 22 00:53:34 mwester henryk is doing some profiling on normal usecases like pull and up.. Oct 22 00:54:19 mwester http://pastebin.ca/745053 Oct 22 00:55:16 mwester: http://pastebin.ca/745053 shows the top lines of mtn pull with 208 revisions. i especially like line 6 Oct 22 00:55:59 basically it says: slow but not because of one thing, because of all of it together Oct 22 00:57:45 right, there's no single hot spot. but maybe one should take a deeper look. for example i strongly suspect that lines 6, 7 and 9 belong together and to some redblack tree handline code. maybe it's possible to form clusters of functionality (e.g. "cryptographic verification") and then see how much they contribute. **** ENDING LOGGING AT Mon Oct 22 02:59:56 2007