**** BEGIN LOGGING AT Wed Sep 19 02:59:57 2007 Sep 19 05:52:57 The line 22 in bluez-utils_3.19.bb needs to have the second half removed since the rfcomm config file is no longer there Sep 19 06:06:55 CM: it's something more than that - rfcomm.conf was in the previous version, and --enable-serial is still set in OECONF Sep 19 06:07:38 rwhitby: Ok. I just wanted to continue the build :) Sep 19 06:08:17 I guess the question is if there should be a default rfcomm.conf file in there or if it gets created somehow Sep 19 06:27:18 CM: there always has been such a default file before Sep 19 06:32:09 *nod* The default on my gentoo system contains only comments though. Sep 19 06:52:29 CM: not if you want to, say, connect to a bluetooth gps via bt serial port ... Sep 19 06:53:46 Of course, and then it's nice to have some default settings to start from. :) Sep 19 07:49:38 yarrrr Sep 19 08:08:29 good morning all Sep 19 08:09:56 morning Sep 19 08:10:26 * koen finally unbroke bluez Sep 19 08:11:29 ~blame mickey|openexpo Sep 19 08:11:30 * ibot blames mickey|openexpo (and Canada) for all the evil in the world Sep 19 08:11:43 morning Sep 19 08:11:44 heh. Sep 19 08:11:48 what did i do? Sep 19 08:12:10 mickey|openexpo: emotion compile segfault Sep 19 08:12:23 mmh. natsy Sep 19 08:12:26 nasty as well Sep 19 08:12:32 we're getting all emotional over that Sep 19 08:12:50 mickey|openexpo: #2992 which you already saw Sep 19 08:13:01 03koen 07org.oe.dev * rb37f72d1... 10/ (4 files in 2 dirs): bluez: fix problems with CONFFILES, rfcomm.conf goes into -compat Sep 19 08:13:09 !oebug 2992 Sep 19 08:13:42 koen: pretty much all poppler using software needs a rev bump, SO_VER has gone from 1->2 Sep 19 08:15:06 http://en.wikipedia.org/wiki/International_Talk_Like_a_Pirate_Day Sep 19 08:15:12 again Sep 19 08:27:02 hi i cant do a build bootstrap-image it fails when building the image with the following error: http://nopaste.biz/19441 Sep 19 08:27:06 any1 got an idea? Sep 19 08:28:12 I don't know, but it seems 'update-modules' is missing Sep 19 08:30:45 gonna check it ... now the question is why it wasnt build automatically? Sep 19 08:34:49 mickey|openexpo: epdf hasnt kept up with poppler :-( Sep 19 08:36:04 bummer Sep 19 08:36:15 i might fix that once i have time Sep 19 08:37:40 poppler_page.cpp:202: error: no matching function for call to 'TextOutputDev::getSelectionText(PDFRectangle*)' Sep 19 08:37:54 mickey|openexpo: http://blogs.gnome.org/bolsh/2007/09/19/osim/ Sep 19 08:39:19 koen: always nice to read stuff from people who have a clue Sep 19 08:39:22 Ari won't like that ;) Sep 19 08:40:16 indeed Sep 19 08:40:45 mickey|openexpo: you know by now you cant possible design an Open system without Nokia at the helm Sep 19 08:40:58 mickey|openexpo: I thought that GUADEC talk had hammered that point home Sep 19 08:41:20 :D Sep 19 08:41:42 silly silly mickey|openexpo Sep 19 08:42:38 hearing the words... not grasping the meaning :D Sep 19 08:48:28 03koen 07org.oe.dev * r8da3c499... 10/ (1 packages/bluez/bluez-gstreamer-plugin_3.19.bb): bluez-gstreamer-plugin: add 3.19 Sep 19 08:48:32 03koen 07org.oe.dev * r088e148a... 10/ (15 files in 6 dirs): poppler dependants: bump PR since poppler 0.6 changed soversion Sep 19 09:15:33 ahah good fortran news Sep 19 09:18:42 or yar me maties, fortrans ship is on the horizon Sep 19 09:19:16 fortran should just die, quietly Sep 19 09:19:33 and it cocks up uclibc builds too Sep 19 09:19:36 RP: what would all the matlabbers do then? Sep 19 09:19:46 use maple ? Sep 19 09:20:50 although I used matlab like for a week and never got the point Sep 19 09:20:58 and Ive never used maple Sep 19 09:22:16 * koen is using both maple and matlab Sep 19 09:22:16 aaarrrr Sep 19 09:22:26 Crofton|home: avast! Sep 19 09:22:51 XorA: Fortran was just a continually source of frustration when I was taught it... Sep 19 09:23:08 shiver me timbers Sep 19 09:23:24 RP, you *taught* FORTRAN? Sep 19 09:23:35 Crofton|home: I was taught it Sep 19 09:23:43 * Crofton|home learned FORTRAN on a punch card system Sep 19 09:23:52 you should be ashamed Sep 19 09:24:20 Crofton|home: They were still teaching it to Physics undergrads in 2003 Sep 19 09:24:55 only to physics undergrads that took the wrong modules Sep 19 09:25:06 I heard a rumour they use it on the local super computer cluster Sep 19 09:25:12 ade|desk: Nope, compulsory. Sep 19 09:25:12 * ade|desk avoided it like the plague Sep 19 09:25:24 where was that ? Sep 19 09:25:29 ade|desk: Durham Sep 19 09:25:35 ah , cold place Sep 19 09:25:43 About the one thing I have against them Sep 19 09:25:44 I work in Durham Hall Sep 19 09:26:41 only been to durham once, visited an all girls college .. never again ... feared for my life Sep 19 09:27:09 ade|desk: Mary's was "interesting". They're all a bit weird there... Sep 19 09:27:27 nice food though Sep 19 09:27:54 There's a lot of good math and physics libs implemented in Fortran, but it's good you don't have to use Fortran to use them. :) Sep 19 09:28:23 ade|desk: I spent most of my time in Mary's in the bar ;-) Sep 19 09:28:55 ade|desk: Usually as part of an every college bar crawl... Sep 19 09:29:36 yay, I am heir to a 48,000,000 GBP fortune Sep 19 09:29:50 lend us a fiver then Sep 19 09:29:57 * ade|desk has no money for lunch Sep 19 09:30:05 RP: you didnt do much crawling, more leaning if you were always in the bar Sep 19 09:30:18 ade|desk: mpfr error? Sep 19 09:30:44 koen sounds about right Sep 19 09:31:59 XorA: With 14 bars and only a few hours it really required running between them rather than crawling :) Sep 19 09:33:36 ade|desk: I have a workaround for that Sep 19 09:34:09 with fortran we can build R too Sep 19 09:34:27 Does anyone have any thoughts on my mail about cross and staging? Sep 19 09:35:17 RP: looks sane, but I think we should think some more about making such changes, so we can do them all well and in one go Sep 19 09:36:12 koen: Yes, I just wanted to write it down whilst I remember it :). I will probably add some of it to poky sooner than later though Sep 19 09:45:24 cross/bin/arm-angstrom-linux-gnueabi-gfortran Sep 19 09:45:26 there we go Sep 19 09:46:01 and now uclibc Sep 19 09:48:02 NOTE: package gcc-cross-4.2.1-r5: task do_compile: started Sep 19 09:50:13 octave on a a780 Sep 19 09:50:16 * koen ponders Sep 19 09:55:24 koen@lieve:/media/hda4/OE/build/tmp-new/angstrom$ ls cross/arm-angstrom-linux-gnueabi/lib/ | grep libgfortran.so.2.0.0 Sep 19 09:55:25 libgfortran.so.2.0.0 Sep 19 09:55:25 koen@lieve:/media/hda4/OE/build/tmp-new/angstrom$ ls deploy/glibc/ipk/* | grep gfor Sep 19 09:55:25 koen@lieve:/media/hda4/OE/build/tmp-new/angstrom$ Sep 19 09:57:39 03koen 07org.oe.dev * rbac543c4... 10/ (7 files in 3 dirs): gcc(-cross): enable fortran for arm/EABI, thanks to Alain for showing and testing the patch Sep 19 10:03:22 anyone got a machine configuration for a lan fonera device Sep 19 10:03:23 ? Sep 19 10:25:06 * koen bumps PR for gcc and hides Sep 19 10:33:53 * Crofton|home larts koen and restarts my build ... Sep 19 10:34:49 03koen 07org.oe.dev * r9d2d4f10... 10/ (13 files in 2 dirs): gcc: change the way to package fortran, reduce spurious variables, make missing fortran files non-fatal Sep 19 10:48:45 hmm Sep 19 10:48:50 grr Sep 19 10:48:55 gcc-r8 fails Sep 19 10:50:07 koen: run while you still got legs :_) Sep 19 11:00:39 03svolpe 07org.oe.dev * r83945abd... 10/ (1 packages/gstreamer/gst-plugins.inc): gst-plugins.inc: Fixed QA Compile Badness with X headers and Libs. Sep 19 11:00:44 03rpurdie 07org.oe.dev * r29237078... 10/ (10 files in 2 dirs): gcc-cross-initial: Factor common code into an inc file, don't run make install for do_install since its pointless (from poky) Sep 19 11:00:53 03rpurdie 07org.oe.dev * r318b0b2a... 10/ (3 files in 2 dirs): gcc-cross-initial: Use common inc file for a couple more files (from poky) Sep 19 11:04:07 re Sep 19 11:04:36 RP: "ERROR: Could not include required file gcc-cross-initial.inc while parsing" is that down to your last commits ? Sep 19 11:04:54 ade|desk: mtn pull ;-) Sep 19 11:06:36 ah thats a very recent addition then Sep 19 11:07:33 XorA: r8 isn't mine :) Sep 19 11:10:41 03rpurdie 07org.oe.dev * r0ba4bbe5... 10/ (1 packages/gcc/gcc-cross-initial.inc): gcc-cross-initial: Add missing file Sep 19 11:21:35 mickey|openexpo: what are your thought on marking gtkrc and themefiles as CONFFILE so changes will get preserved over package upgrades? Sep 19 11:21:54 sounds reasonable Sep 19 11:22:02 anyone aware of a goot howto for building machine.conf files? with jffs options etc? Sep 19 11:24:05 koen, has anyone mentioned to you that libgfortran is not compilng in gcc-cross? Sep 19 11:24:30 thos: I have it compiling over here Sep 19 11:25:09 koen, I get a problem with kinds.h being missing Sep 19 11:25:37 koen, http://pastebin.ca/703110 Sep 19 11:27:27 koen: Note thos is using 4.1.2, not 4.2.1 Sep 19 11:28:00 I didn't specify any version myself... Sep 19 11:28:04 I built 4.1.2 and 4.2.1 with fortran this morning Sep 19 11:28:42 let me see if I can reproduce it Sep 19 11:30:29 libgfortran/mk-kinds-h.sh: Unknown type Sep 19 11:30:42 I had that error like 2 years ago, don't know how it went away Sep 19 11:38:41 thos: cross-initial built ok? Sep 19 11:39:26 koen, yeah Sep 19 11:40:20 ok Sep 19 11:41:22 I'm rebuilding gcc-cross 4.1.2 on two machines, one with and one without rebuilding gcc-cross-initial Sep 19 11:41:30 koen: I didn't change the stamps on cross-initial anyway since the changes have no effect on staging/cross Sep 19 11:42:02 and people probably don't want to rebuild gcc initial after gcc-cross anyway since it has potential to screw things up Sep 19 11:42:09 cross-initial share PR with -cross Sep 19 11:42:24 ok Sep 19 11:43:33 NOTE: package gcc-cross-4.1.2-r8: task do_rm_work_all: completed Sep 19 11:43:42 for the one without rebuilding cross-initial Sep 19 11:43:54 * koen waits for the other machine to finish Sep 19 11:45:09 koen, by the way *please please please* can we remove the "got here" patch from dbus Sep 19 11:45:35 thos: only if you can prove it isn't needed anymore Sep 19 11:46:00 it solved a reproducable segfault Sep 19 11:46:18 Does poky use that patch? Sep 19 11:46:19 koen, we've not used it in poky for months (years?) and never had a problem Sep 19 11:46:36 yes, poky Sep 19 11:46:38 not OE Sep 19 11:46:55 koen, it was removed from openmoko OE for a while and never had a problem Sep 19 11:47:16 koen: Is there a simple test to see whether this segfault is present? Sep 19 11:48:59 RP: yes: https://bugs.freedesktop.org/show_bug.cgi?id=7675 Sep 19 11:49:55 if some with an *OE* build can run that testcase and it doesn't segfault, then and only then the patch can go Sep 19 11:50:17 on the autobuilder: NOTE: package gcc-cross-4.1.2-r8: task do_build: completed Sep 19 11:51:28 thos: next build, remove that patch, type "dbus-monitor --system" in the image and then ask koen to remove the patch ;-) Sep 19 11:52:04 RP: only when running dbus-daemon under gdb Sep 19 11:53:08 I don't want to have something like dbus break in a non-obvious way :/ Sep 19 11:53:49 koen: That patch just moves some kind of race or memory corruption, its still broken Sep 19 11:54:10 RP: what's the alternative? Sep 19 11:54:27 koen: Fix it properly Sep 19 11:54:43 RP: I don't know how and the dbus dudes don't know either Sep 19 11:54:53 right way, wrong way, and the homer way :) Sep 19 11:55:16 removing a patch that allows the daemon to run isn't an option Sep 19 11:55:56 koen: no segfault under gdb here either Sep 19 11:56:20 RP: well, as soon as gcc-cross compiles again, I will ;-) Sep 19 11:56:21 koen: The --nofork option doesn't exist now so the testcase no longer even fully applies ;-) Sep 19 11:56:25 RP: ok, so that patch can go :) Sep 19 11:56:50 koen: This is on poky, not OE but I fail to see why it would be much different for OE Sep 19 11:57:08 thos: Just disable fortran for now Sep 19 11:57:19 thos: Its not like you need it... Sep 19 11:57:58 RP: the patch stays in till someone proves it isn't needed on an OE build Sep 19 11:58:15 RP: it's not like it's been need since the 70s either ;-) Sep 19 11:58:30 thos: quite ;-) Sep 19 11:58:42 http://www.math.vt.edu/people/sturler/ Sep 19 11:58:48 search this page for fortran Sep 19 11:58:49 koen: thos will do that shortly, once his builds work again Sep 19 11:59:00 they let EE's use the pool and beach also Sep 19 11:59:24 just for the record, gcc-cross is barfing for me on a clean build Sep 19 12:00:39 Crofton|home: thos said "need", not "use" ;-) Sep 19 12:01:28 I'd even buy the argument that fortran was more accurate than C with floating point until the last 10 years or so but it isn't really needed now... Sep 19 12:01:51 its just moldy old guys who do not want to rewrite their code Sep 19 12:02:07 any clues on disabling fortran in gcc then? Sep 19 12:04:02 thos: mtn pull && mtn up Sep 19 12:07:13 * Crofton|home should setup OE on vt cluster so we can test builds really fast ... Sep 19 12:07:41 RP, rewrite bitbake in FORTRAN Sep 19 12:08:00 * RP pushes Crofton|home down the plank Sep 19 12:09:29 man the poop deck! Sep 19 12:10:16 gcc-cross? Sep 19 12:10:20 barfing? Sep 19 12:11:09 shiver koens timbers :-) Sep 19 12:13:36 Don't tase me bro' Sep 19 12:19:01 03koen 07org.oe.dev * r0cd4358a... 10/ (1 packages/gcc/gcc_4.1.2.bb): gcc 4.1.2: disable fortran as it doesn't build for everybody (it works fine on the autobuilder and my home machine) Sep 19 12:23:01 I wish this guy had said "I'm being repressed!" Sep 19 12:23:24 Crofton|home: your out of line, please sit down before I tase you Sep 19 12:23:43 * Crofton|home tasered himself once Sep 19 12:24:00 yarr, Ill make yah timbers shiver with me magic box of lightning Sep 19 12:24:13 it'll make you jump Sep 19 12:24:33 * XorA has only thankfully electircuted himself from the collapsing current of coils Sep 19 12:24:55 that is likely how a taser works .... Sep 19 12:25:15 * Crofton|home guesses Sep 19 12:25:18 I thought they were jacobs ladder things Sep 19 12:26:00 I am guessing they interupt current flow in an inductor to get high voltage pulses Sep 19 12:26:17 v = L di/dt Sep 19 12:26:22 Crofton|home: time for wikipedia :-) Sep 19 12:27:46 Crofton|home: nah, they are a standard transformer, nothing interesting Sep 19 12:28:03 autobot or decepticon? Sep 19 12:28:06 Crofton|home: or jacobs ladder Sep 19 12:28:18 Decepticon Sep 19 12:28:32 the autobots only ever won by luck Sep 19 12:29:06 shoudl gcc build atm? Sep 19 12:29:24 gcc waz a might ship, yarrr Sep 19 12:47:50 nik0n : there are a few machines that use jffs2 in OE. One i worked on is magicbox Sep 19 12:48:44 * steliosk wished dollar to keep falling vs euro till my US trip is over Sep 19 12:52:25 stop complaining, you are already getting a good deal .... Sep 19 12:52:59 ~change 1 usd to eur Sep 19 12:53:05 i got some error trying that Sep 19 12:53:16 ~change 1 usd to euro Sep 19 12:53:21 i got some error trying that Sep 19 12:54:23 Crofton|home | its 1.398 since i checked in the morning Sep 19 12:55:59 * Crofton|home cries Sep 19 12:59:45 Crofton|home : Don't :) it will change before OEDEM Sep 19 13:25:53 morning Sep 19 13:38:57 gcc 4.1.2 is broken isn't it ? Sep 19 13:39:14 Yay, cross in poky taken from 100MB in 1011 files to 24MB in 72 files :) Sep 19 13:39:32 nice Sep 19 13:39:37 Two less make install cycles for glibc too so faster Sep 19 13:40:31 nud: It shouldn't be if you mtn up Sep 19 13:40:39 yeah Sep 19 13:40:49 it looks like it is working for me now Sep 19 13:40:51 * nud mtn up again Sep 19 13:40:57 maybe I was just in the wrong timeframe Sep 19 13:41:07 yeah Sep 19 13:41:14 bad morning Sep 19 13:42:57 RP: switch to uclibc, that builds in like 3 minutes :) Sep 19 13:43:49 :) Sep 19 13:44:33 hm do someone know where I could find quickcapture specs from intel ? /me desperate Sep 19 13:44:56 compiling gcc still fails Sep 19 13:45:53 http://rafb.net/p/xjh8bB72.html <-- the error message Sep 19 13:46:12 nud, I restarted by wiping tmp Sep 19 13:46:24 wiping tmp ? Sep 19 13:46:47 * nud doesn't want to drop all his tmp directory, it will take ages to build it again Sep 19 13:47:34 * mwester reads the log... Sep 19 13:47:54 Somebody tell me, please, that we didn't bump the version on gcc again, did we? Sep 19 13:48:37 only the rev number Sep 19 13:48:38 nud: Are you sure you did mtn pull and mtn up? Sep 19 13:48:47 RP: certain Sep 19 13:49:13 nud: It shouldn't be building fortran for gcc 4.1.2 now Sep 19 13:49:29 I rm'ed -rf /stuff/build/tmp/work/armv5te-angstrom-linux-gnueabi/gcc-cross-4.1.2-r8/ Sep 19 13:49:37 now let's try again Sep 19 13:49:50 nud: bitbake gcc-cross -c clean first Sep 19 14:14:47 bother Sep 19 14:15:20 my autobuild croaked horribly Sep 19 14:15:59 * cbrake ponders X startup issues again ... Sep 19 14:17:16 i'm having problems with "ttf-liberation_0.2.bb", because wget tries to access a https-url without having the certificate nor -no-check.. Sep 19 14:19:32 hi, all! do anybody remembers how to re-generate sysdeps/.../configure files within glibc? Sep 19 14:20:33 root configurde is re-generated ok Sep 19 14:21:34 osas : it looks like i will miss your presentation :( My flight arrives at 9:15 don't think i'll be at the venue before 11 Sep 19 14:21:52 RP: I think packages/gcc needs to get a good makeover Real Soon(TM) instead of tweaking the current cruft Sep 19 14:22:02 I will make another one for you :) Sep 19 14:22:10 do you have a hotel? Sep 19 14:22:22 RP: same applies to glibc Sep 19 14:23:09 osas : yeah. i am staying at the conference hotel. The astricon people were kind enought to help out Sep 19 14:23:16 koen: Agreed, I kind of started on that with the -initial gcc changes ;-) Sep 19 14:23:17 nice Sep 19 14:24:27 osas : I am going over my schedule. There is one more interesting presentation at 3:30 Porting asterisk to IBM Power risck Processor Sep 19 14:25:12 we will both attend then Sep 19 14:26:09 RP: I often think "this looks very wrong, but I'll leave it in" Sep 19 14:28:03 koen: There is a lot of potential for breaking things too though :/ Sep 19 14:28:29 that's why I leave it in :/ Sep 19 14:35:40 RP, koen : Could i make a suggestion ? Would it be possible to hold all gcc changes for after OEDEM ? Sep 19 14:37:51 steliosk: What's the concern? Sep 19 14:38:50 RP, koen : It would make sense in order to get a stable tree, from the existing code base, before moving on Sep 19 14:42:36 steliosk: The tricky bit then is everyone switches to this stable tree and nobody tests .dev :/ Sep 19 14:45:03 RP : I don't think this will happen Sep 19 14:45:50 RP : we have a lot of adventure seeking people in OE :) Sep 19 14:46:14 steliosk: :) Sep 19 14:46:49 steliosk: I don't really want to ban touching the toolchain for 3-4 weeks until OEDEM Sep 19 14:47:11 steliosk: I wouldn't make any change like the staging layout until after then though Sep 19 14:47:51 to be fair the last changes to gcc were tested by two people on three machines and worked, it was just bad luck they fail on some machines Sep 19 14:47:52 to quote mickey|openexpo "to we need a .really-dev?!?!" Sep 19 14:48:10 we need to attract the less adventurous types Sep 19 14:48:18 the ones with loose wallets :) Sep 19 14:48:34 Crofton: Angstrom-stable is a topic for OEDEM Angstrom meeting Sep 19 14:48:38 RP : I am not talking about a general ban just to avoid big changes , like those discussed on the ml ;) Sep 19 14:49:02 I need to write down my thoughts on stable and how to deal with third party groups Sep 19 14:49:12 koen, :) Sep 19 14:49:13 steliosk: Like the one I just sent? Sep 19 14:49:24 XorA : A stable release assumes a stable toolchain Sep 19 14:49:29 Crofton: Please do Sep 19 14:50:04 I don't pretend to know everything, but we need a starting place Sep 19 14:50:05 RP : I was reading yesterday mails, haven't looked at today post Sep 19 14:50:16 copying back and forth to poky drives me nuts :) Sep 19 14:50:43 Crofton : this would be a good thing Sep 19 14:50:48 yeah Sep 19 14:50:56 must finish SDR Forum paper first Sep 19 14:51:24 Crofton: Any viewpoints are welcome on the issue, particularly one I've not heard before :) Sep 19 14:51:34 heh Sep 19 14:51:50 steliosk: I just prposed a 33 files changed, 75 insertions(+), 406 deletions(-) patch to several bits of core toolchain ;-) Sep 19 14:52:10 howdy - i essentially copied the angstrom-2007.1 config, set my TARGET_OS to linux-gnueabi, and all is well until libtool-native. in the compile phase, it first does a "/bin/sh ./config.status --recheck", which bombs out with: "configure: error: source directory already configured;" Sep 19 14:52:10 RP : How about "Please don't make changes to the toolchain and i'll give you a big bottle of Ouzo at OEDEM" ? Sep 19 14:52:19 hmmm Sep 19 14:52:28 bring a case Sep 19 14:52:51 what flavour is Ouzo again? Sep 19 14:52:51 that stuff is like drinking candy Sep 19 14:52:55 licorice Sep 19 14:53:05 HopsNBarley: you have a symlink in your path Sep 19 14:53:10 i compared the non-eabi and eabi config.status'es of libtool-native, and their is one wierd difference: a block of sed code is missing in the eabi one. Sep 19 14:53:18 koen, uh, lemme check that. Sep 19 14:53:47 I dud Ouzo and Irn Bru, the blacksmith cocktail, we should make some at OEDEM Sep 19 14:53:50 koen, how'd you know that? (-; Sep 19 14:53:54 s/dud/dub/ Sep 19 14:54:11 HopsNBarley: it's FAQ :p Sep 19 14:54:25 steliosk: One of my biggest fears is I go and clean poky's toolchain up and then can't merge the cleanup back to OE Sep 19 14:54:44 thanks ynezz! Sep 19 14:54:53 or get left having to merge the cleanup in one go rather than step by step Sep 19 14:57:02 crap, I just deleted my auto build script Sep 19 14:57:37 koen, can you pastebin your script again ... Sep 19 14:57:54 this time I call it something different from the auto build tmp dir .... Sep 19 14:58:12 RP: I think we should hold major changes to the toolchain for after OEDEM. I understand that it will probably require more work to catch up with Poky, but for i believe a stable and tasted toolchain is need it. Sep 19 14:58:32 Crofton: http://rafb.net/p/z30gdb47.html Sep 19 14:58:50 what does OEDEM have to do with a stable toolchain? Sep 19 14:59:03 steliosk: As long as you understand it will kill my interest in fixing it Sep 19 14:59:12 fair play Sep 19 14:59:37 RP : That's where the Ouzo comes in :) Sep 19 14:59:46 maybe its time again for the discussion to move glibc/binutils/gcc out of .dev to .dev-toolchain Sep 19 14:59:56 koen: steliosk wants the current toolchain in the stable release rather than something I/we change before then Sep 19 15:00:15 RP: your cross-must-go plan would benefit uclibc as well Sep 19 15:00:35 RP: you can branch of at any point in the history with mtn Sep 19 15:00:40 03koen 07org.oe.dev * raef79819... 10/ (2 files in 2 dirs): matchbox-keyboard-im: apply small-screen patch Sep 19 15:01:36 steliosk: It doesn't sadly. I have a tiny amount of time now to look at this and I can't come back to it in a month when I won't remember why/what I was doing and will have to remember it :( Sep 19 15:02:14 koen: I'm not totally out to remove cross. I can now see why it needs to be there and its probably ok in the reduced form, if we can get the compiler using sane paths Sep 19 15:02:44 koen: Removing the lib and include files is fair game to me though as I'd class it as a bug currently Sep 19 15:05:15 mmh I can't link my c++ code anymore, looks like libstdc++ has vanished Sep 19 15:06:06 cbrake: Well written post again :) Sep 19 15:28:10 [embedder@parterredecrocus] armv5te-angstrom-linux-gnueabi $ find /stuff/build/tmp/staging/ -name "libstd*.so" Sep 19 15:28:10 [embedder@parterredecrocus] armv5te-angstrom-linux-gnueabi $ Sep 19 15:28:13 CM: thanks Sep 19 15:28:25 why isn't libstdc++ installed anymore ? Sep 19 15:31:27 maybe its because nothing need it by default? Sep 19 15:32:02 I've got a c++ project of my own in my image, and it doesn't compile anymore... Sep 19 15:32:19 well, actually it does, but it fails at link time Sep 19 15:32:42 it worked fine yesterday... Sep 19 17:26:51 RP: marcin was working on the zaurus defconfigs, right? Sep 19 17:46:59 koen: did you have any further thoughts on the "right way" to deal with my libgd size & option selection issues? No urgency as I am using the bb colections feature to get what I need. Just curious whether you consider it a bug or a feature :-) Sep 19 17:54:52 koen: yes Sep 19 17:56:05 minor bitbake question, koen ? Sep 19 18:36:33 :/ kdrive 1.4's console "keyboard" driver hangs while reenabling with a second USR1 signal, but the evdev keyboard driver doesn't work yet without xkb support. Sep 19 19:08:31 03khem 07org.oe.dev * r0de79e77... 10/ (1 conf/distro/include/sane-srcrevs.inc): Update SVN revision of eglibc. Sep 19 19:11:48 I've got a question about an odd error with OE I'm seeing from ipkg-cl not being able to find any of the packages that have been built, when trying to make an image. Anyone heard of something like this before? Sep 19 19:16:54 03khem 07org.oe.dev * r0de79e77... 10/ (1 conf/distro/include/sane-srcrevs.inc): Update SVN revision of eglibc. Sep 19 19:18:32 03khem 07org.oe.dev * r0de79e77... 10/ (1 conf/distro/include/sane-srcrevs.inc): Update SVN revision of eglibc. Sep 19 19:41:26 yay, angstrom-console-image build dir is only 5G again -- graphical dependencies must be back under control :-) Sep 19 19:46:20 we need to support the iphone Sep 19 19:46:21 http://www.computerworld.com/action/article.do?command=viewArticleBasic&articleId=9037398&intsrc=hm_list Sep 19 19:55:41 zecke! Sep 19 19:55:47 zecke: how did you spend the night? Sep 19 19:55:58 (don't say "sleeping") Sep 19 20:02:40 mickey|openexpo: I did sleep, in a christian hostel (shelter.nl) in the redlight district Sep 19 20:03:11 oops...redlight tss... Sep 19 20:05:53 I have been hitting this error here http://pastebin.ca/703749 Sep 19 20:05:59 any ideas ? Sep 19 20:14:57 mickey|openexpo2: But I'm back in eindhoven now, I don't know what I love about Amsterdam Sep 19 20:15:31 i like the bridges.. Sep 19 20:17:08 zecke: if you are on trevel it is worth to take a look at http://couchsurfing.org/ Sep 19 20:31:28 i just did a mtn pull & update from a repo which was a week old and now bitbake starts almost over from scratch Sep 19 20:32:36 sounds like people worked on the toolchain Sep 19 20:32:43 that happens sometimes Sep 19 20:35:17 too bad .. i was almost done, it stopped at missing gtk-webcore-ncrit or such then i thought i could try a simple update Sep 19 20:36:04 well, you can try bitbake -c rebuild Sep 19 20:36:11 this doesn't honor dependencies Sep 19 20:36:39 does this work against bitbake gpe-image ? Sep 19 20:37:04 well i keep it rebuilding over night . Its ok Sep 19 21:00:02 anyone out there built rhythmbox? there are two versions out there. Just wondering if anyone got them to build. It fails configure looking for Gstreamer and says it doesn't find anything newer than .8 even though I have .10 built. Sep 19 21:01:55 mickey|zzZZzz: ping? Sep 19 21:03:47 hmm why is this : " No package 'gtk-webcore-nrcit' found" Sep 19 21:08:12 I don't see rhythmbox in the repo either Sep 19 21:18:02 feels like osb-nrcit misses osb-jscore and osb-nrcore as dependency Sep 19 21:20:28 koen: did you build the browser yet? Sep 19 21:21:10 NOTE: package gpe-mini-browser-0.21: completed :-) Sep 19 21:23:14 this task foo magic since some weeks is really scary ! how is this done ? Sep 19 21:31:40 rob_w, where do you get that error? Sep 19 21:31:43 ipkg-cl in making an image? Sep 19 21:32:36 Tartarus, not was at gpe-mini-browser at task_configure -> building this other osb-XX by hand fixed it Sep 19 21:33:03 doh, k Sep 19 21:33:18 i'm having this problem where when making my image, ipkg-cl can't find any of the ipk files Sep 19 21:48:50 | genext2fs: couldn't allocate a block (no free space); NOTE: package gpe-image-1.0-r0: task do_rootfs: failed huuh ?? Sep 19 22:00:06 rob_w: Read the message.. are you out of disk space? Sep 19 22:01:27 NAbyss, my host certainly not but the i686-generic target filesystem "ext2.gz" probably is .. using only "tar" as filesystem is fine Sep 19 22:05:38 incredible 131MB is gpe-image these days .. wow Sep 19 22:11:37 zecke, would you be interested in committing a recipe for PyWebKitGtk? http://bugs.openembedded.org/show_bug.cgi?id=3028 Sep 19 22:15:06 LICENSE = "GPL" is that what you really want? Sep 19 22:16:08 hmm that's what PyWebKitGtk has in COPYING Sep 19 22:16:57 tuukkah: right, I just wonder about it. Wrapping a partly BSD/LGPL project using h2def.py and saying the result is GPL is strange, but the recipe is okay in this respect Sep 19 22:17:24 tuukkah: you don't want to hardcode /usr though Sep 19 22:18:35 zecke, is there some variable i should be using instead? Sep 19 22:19:25 tuukkah: let me search an example Sep 19 22:20:46 autotools.bbclass has "--prefix=${prefix}" Sep 19 22:22:05 tuukkah: http://www.openembedded.org/repo/org.openembedded.dev/packages/python/python-openmoko_0.1.bb Sep 19 22:22:26 tuukkah: I think the answer is distutils-base and you want to include hrw's patch for your configure.ac Sep 19 22:23:10 hey ppl. Sep 19 22:23:19 i tried to use distutils-base but the resulting packages didn't look right. i didn't check the patch though Sep 19 22:23:50 poll : how long for building a minimal image from the ground up ? Sep 19 22:23:57 tuukkah: instead of /usr/lib you want to use ${libdir} for sure (check conf/bitbake.conf for available paths) Sep 19 22:24:28 tuukkah: but mickeyl is our python master, so you might want to ask him as well Sep 19 22:24:39 I have some quirks in my work directory, and it would take some effort (thus time) to fix, but a complete build is sooo long Sep 19 22:24:55 zecke, ok, thanks! Sep 19 22:27:22 tuukkah: and consider changing the license Sep 19 22:27:56 yes, i'll ask about it Sep 19 22:29:10 zecke, would LGPL be good? Sep 19 22:29:48 or should it be some mixture of BSD and LGPL? Sep 19 22:30:35 tuukkah: LGPL would be good, the usage of WebKit should not be limited by a license and I think a call to h2def.py is not really copyrightable :) Sep 19 22:30:58 tuukkah: http://projects.openmoko.org/plugins/scmsvn/viewcvs.php/python/src/openmoko-ui/openmokouimodule.c?rev=1&root=bindings&view=markup, you want to add the PyErr_Print() to your module.c as well Sep 19 22:31:13 tuukkah: it is nice to find out why loading didn't work Sep 19 22:37:54 ok, good tip Sep 19 23:22:21 03daniel 07org.oe.dev * r59ac76a9... 10/ (1 conf/checksums.ini packages/navit/navit_0.0.1.bb): navit: Add checksums for navit tgz. Improve performance on arm (thanks cp15) Sep 20 00:57:02 Hi. When I run "bitbake gpe-image" on latest org.openembedded.dev tree, it fails building gpe-mini-browser-0.21 with an error "No package 'gtk-webcore-nrcit' found". Any ideas? Sep 20 02:09:21 i want to use directfb as the backend but x11. but after i modified DEPENDS in pango.inc & cairo.inc, the depends for xrender and x11 package is not removed. Sep 20 02:10:13 why it act as this, and how to correct the recipe? **** ENDING LOGGING AT Thu Sep 20 02:41:57 2007 **** BEGIN LOGGING AT Thu Sep 20 02:42:11 2007 **** ENDING LOGGING AT Thu Sep 20 02:44:12 2007 **** BEGIN LOGGING AT Thu Sep 20 02:44:35 2007 **** ENDING LOGGING AT Thu Sep 20 02:59:57 2007