**** BEGIN LOGGING AT Wed Oct 18 02:59:56 2006 Oct 18 04:44:31 I added a package, and the package workdir was put to "work/${MACHINE_ARCH}" instead of "work/${TARGET_SYS}", why? Oct 18 04:45:34 this package inherits autotools and pkgconfig Oct 18 06:21:22 03hrw 07org.oe.dev * r98073c8a... 10/ (1 conf/distro/celinux-test.conf): celinux-test: override TARGET_CC_ARCH for omap5912osk - gcc 3.x use other mtune then gcc 4.x Oct 18 06:24:56 03hrw 07org.oe.dev * r0f00f5e1... 10/ (1 conf/distro/celinux-test.conf): celinux-test: added more comments, restructured Oct 18 06:55:50 yo Oct 18 07:06:52 has anyone any sensible remarks about the install location of qtopia? Oct 18 07:07:54 everything in a sandbox, say /usr/qtopia? or stuff in /usr/lib/qtopia and /usr/bin/qtopia? or just put all the libs in /usr/lib? Oct 18 07:08:15 or instead of qtopia just put it in qt4, for compatibility? Oct 18 07:08:30 and should qtopia provide something, like qt4x11? Oct 18 07:08:36 /opt/Qtopia seems sensible to me Oct 18 07:08:54 or, perhaps you mean qtopia core Oct 18 07:09:05 yes that's what i meant, sorry Oct 18 07:09:07 the gpl qtopia core Oct 18 07:09:24 i'm making a qtopia-core_4.2.0.bb Oct 18 07:09:54 why not follow the LSB? Oct 18 07:10:09 yes, where does qtopia go in LSB? Oct 18 07:10:28 /usr/local/Trolltech :) Oct 18 07:10:35 ew Oct 18 07:10:51 i assume since qt is part of LSB Oct 18 07:10:58 now that's retarded :p Oct 18 07:11:24 ya, well. lots of things I can think of are retarded Oct 18 07:11:26 i'm tempted to put libs in /usr/lib/qtopia Oct 18 07:12:05 or, maybe in ${QTOPIADIR}/lib, where QTOPIADIR might be /usr/local/Trolltech ;) Oct 18 07:13:20 mickey pbly has something to say on this ;) Oct 18 07:15:45 giel: I never really understood why Qtopia or Opie went into their own directories Oct 18 07:16:09 giel: if you want to be LSB, then put it into the normal paths, e.g. /usr/lib for the libs, /usr/share/xxx for data files and so on Oct 18 07:16:32 giel: directory separation isn't that needed when you have RPM, DEB or IPK to separate the packages from each other Oct 18 07:16:50 giel: but the Trolltech way to place things used to be /opt/Qtopia Oct 18 07:17:15 and the Opie way way /opt/QtPalmtop AFAIK Oct 18 07:17:22 s/way way/way was/ Oct 18 07:17:28 god mysql hates me Oct 18 07:17:31 schurig: hm, just putting the libs in /usr/lib etc sounds fine to me Oct 18 07:17:54 ggilbert_: I took a liking into PostgreSQL recently :-) Oct 18 07:18:06 I can't even get the fucking thing to start Oct 18 07:18:06 schurig: i'll just make my package that way, put some comment in the .bb as well, submit it and let others figure it out ;) Oct 18 07:19:05 giel: are you making a Qtopia 4.2 BB file? Then do it as you say. If the hard-core Qtians don't like it, they can change it anyway. Oct 18 07:19:35 schurig: yep. Oct 18 07:19:51 schurig: i just need qtopia-core working for me right now, the exact location i can't be bothered with ;) Oct 18 07:20:04 good morning all Oct 18 07:20:06 schurig: i'm also very unhappy with the whole qmake situation Oct 18 07:20:06 giel: AFAIL the used-to-be-opie'ians (like mickey, zecke) are angry at Qt now and don't use it any more. Mickey said to me that he still does Qt, but no more Qt/E but Qt/X11 Oct 18 07:20:25 giel: qmake is nice, what's the problem? Oct 18 07:20:32 schurig: qtopia core as a toolkit is actually quite nice, as it's API compatible with qt4 Oct 18 07:21:29 schurig: that it gets built during the configure ;) of course you can patch that out and depend on qmake2-native or something like that, but it's wrong in the first place imo Oct 18 07:21:37 schurig: building your make tool during configure :S Oct 18 07:22:02 schurig: i didn't bother with changing it like it's done in the current qt4 package, so i let the configure build qmake and use that one Oct 18 07:22:04 giel: yes, the build stuff for Qt is somewhat barock Oct 18 07:22:18 schurig: i'll leave it this way and put a comment again ;) Oct 18 07:22:25 I'm not sure if qmake-qt2 can build qt4. Do you know this? Oct 18 07:23:01 i think qmake2 can do so Oct 18 07:23:16 the current qmake2-native package gets qmake from qt4 Oct 18 07:23:26 and that qmake can pbly build qtopia core 4 fine Oct 18 07:24:10 okay, i'll fix up the install, staging and packaging, put in some comments about my choices, i hope to get stuff working today Oct 18 07:24:22 morning Oct 18 07:50:45 morning Oct 18 07:51:24 hey hrw|work Oct 18 07:53:21 hey hrw|work & XorA Oct 18 07:56:55 morning alli Oct 18 07:57:19 have debian just started to make up version numbers for packages? Oct 18 07:57:29 <[lala]> morning guys Oct 18 07:58:36 XorA: possibly they make up name for packages, ice{dove,weasel} etc Oct 18 07:58:57 dhcdbd upstream as far as I can tell is 1.14 debian has a 2.0 Oct 18 08:02:33 ahah, upstream changed authors Oct 18 08:04:04 ade|desk: ice-embedded! Oct 18 08:04:09 ade|desk: icebake! Oct 18 08:04:40 03koen 07org.oe.dev * rfa15ef2d... 10/ (1 conf/distro/celinux-test.conf): celinux-test: typos Oct 18 08:04:40 and after yesterday icevidia Oct 18 08:08:16 hrw|work: hi hrw Oct 18 08:08:16 iceL4hurd ? Oct 18 08:08:16 koen: goedemorgen koen Oct 18 08:08:17 likewise: Are you successful in using icecc with OE? Oct 18 08:08:26 hey likewise Oct 18 08:08:33 good morning everyone. Oct 18 08:08:34 I should go to sleep Oct 18 08:08:37 hey Laibsch Oct 18 08:08:41 but right now I just want to drink heavilly Oct 18 08:08:47 howdy koen Oct 18 08:09:02 Laibsch: almost. I fixed icecc.bbclass to package up the toolchain (more) correctly, but cross-compiling still fails on the clients... Oct 18 08:09:16 ggilbert_: I recommend some dark rum for that Oct 18 08:09:49 Laibsch: the cross compiler is correctly extracted at the client's /var/cache/icecc/ directory, but I get "error writing to cached socket" failures. Oct 18 08:10:20 Laibsch: also, in the examples I sometimes see libbfd in the tar file, but I have no such beast. Oct 18 08:10:40 I don't even get that far. It bails out right when configuring or compiling, I cannot remember now. But AFAICT nothing ever gets distributed. Oct 18 08:10:57 Laibsch: did you use my corrected icecc.bbclass? Oct 18 08:10:58 "GNU libbfd, the binary file descriptor library from GNU " Oct 18 08:11:12 Yes, I pulled it from http yesterday. Oct 18 08:11:16 koen: yes, but I do not see which exe depends on it. Oct 18 08:11:42 Laibsch: did you INHERIT += " icecc " in your local.conf? Oct 18 08:12:04 likewise: isn't libbfd part of binutils? Oct 18 08:12:09 Laibsch: Also, did you set export ICECC_PATH? Oct 18 08:12:35 koen: yes, so only maybe 'as' would depend on it Oct 18 08:13:17 on the bright side, the problem seems to be that the current mysql release just doesnt work on this machine. Oct 18 08:14:59 koen: there is no libbdf in build/tmp/cross/ (or staging) Oct 18 08:15:58 03hrw 07org.oe.dev * re9e4facf... 10/ (1 conf/distro/celinux-test.conf): celinux-test: one more typo missed Oct 18 08:16:47 Laibsch: ping! What gives a ls -al build/tmp/{staging,cross}/ice/ ? Oct 18 08:17:49 moin zusammen Oct 18 08:18:37 s/moin zusammen/good morning ;-) Oct 18 08:22:06 I inherited but I did not set icepath. What do you have for that? I will paste the requested output. Oct 18 08:22:16 moin florian_kc Oct 18 08:23:08 morgen florian_kc Oct 18 08:25:23 likewise: when bitbaking binutils (non-cross): NOTE: Multiple libraries (libbfd-2.17.50.0.5.so, libopcodes-2.17.50.0.5.so) found and LEAD_SONAME not defined Oct 18 08:25:35 likewise: I guess http://rafb.net/paste/results/8bfyiz56.html is what you meant Oct 18 08:25:40 03koen 07org.oe.dev * rd6769507... 10/ (1 conf/distro/angstrom-2007.1.conf): angstrom: set the correct binutils version Oct 18 08:25:49 My directory structure seems to differ from yours. Oct 18 08:26:32 Laibsch: and what if you do a find ../tmp/openzaurus/{staging,cross}/ice/ Oct 18 08:27:03 Well, cross isn't there. Oct 18 08:27:08 aha Oct 18 08:27:11 03koen 07org.oe.dev * r5714dc17... 10/ (10 files in 3 dirs): binutils: add 2.17.50.0.5 with patches from uclibc buildroot Oct 18 08:27:15 cu in a min Oct 18 08:27:34 likw|bbl: And apart from the directories there do not seem to be any files. Oct 18 08:27:51 will help you in a min, have a short meeting now... Oct 18 08:28:03 Sure. Oct 18 08:30:11 03xora 07org.oe.dev * rab0313a1... 10/ (3 files in 3 dirs): Oct 18 08:30:11 dhcdbd_2.0.bb : new upstream version, also added patch to disable sending Oct 18 08:30:11 -x to dhclient (which is a patch dhclient needs if it is to understand Oct 18 08:30:11 this new arg) Oct 18 08:30:56 03koen 07org.oe.dev * rc4b26d4a... 10/ (1 packages/tasks/task-base.bb): task-base: bump PR Oct 18 08:32:13 ah, right someone still need to fix networkmanager in OE Oct 18 08:33:06 hm Oct 18 08:33:39 koen: I am working with mmp to fix it up Oct 18 08:34:04 XorA: coolness Oct 18 08:34:40 koen: good progress last night, I got it to get a dhcp lease Oct 18 08:34:50 koen: but there is no public AP at wolf Oct 18 08:35:05 XorA: I couldn't get it to work with wpa-supplicant :( Oct 18 08:35:33 koen: /etc/init.d/wpa stop, then it will work Oct 18 08:35:58 /etc/init.d/wpa??? Oct 18 08:36:12 hrw|work: thats what wpa-supplicant installs here Oct 18 08:36:26 wpa-supplicant was started by 'ifup wlan0' on OZ Oct 18 08:36:58 XorA: my problem was that the default wpa-supplicant.conf made wpa-supplicant barf Oct 18 08:37:21 koen: I have 0.5.1 installed, that works Oct 18 08:38:04 XorA: 0.4.x is the default in .dev, right? Oct 18 08:38:27 drat Oct 18 08:38:30 koen: probably, but I figure wpa-supplicant in .dev and oz got desynced Oct 18 08:38:41 I don't have to OE patches for teh dbus security stuff anymore Oct 18 08:41:32 ah, 0.5.1 even uses gnutls Oct 18 08:42:07 koen: later than 0.5.1 though uses autoconf 1.6 Oct 18 08:42:25 2.60 I mean Oct 18 08:42:36 XorA: is rxvt-unicode still being pink for new versions? Oct 18 08:42:56 koen: havent seen if there is any new upstreams, but its still pick for me Oct 18 08:45:24 from 7.8: "INCOMPATIBLE CHANGE: this version will always read ~/.Xdefaults," Oct 18 08:45:37 so yo don't have to logout for colour changes anymore Oct 18 08:52:31 NOTE: package rxvt-unicode-7.9: completed Oct 18 08:52:49 XorA: way ahead of you ;) Oct 18 08:53:01 03koen 07org.oe.dev * r3469f732... 10/ (4 files in 3 dirs): rxvt-unicode: add 7.9 Oct 18 08:53:19 haven't tested it yet, though Oct 18 08:53:38 what happened to test before add policy? Oct 18 08:53:43 * koen rebuilds from scratch to rule out bintutils troubles Oct 18 08:53:55 XorA: DEFAULT_PREFERENCE = "-1" Oct 18 08:55:58 koen: I know I put that there, but there is little point increasing the crap levels Oct 18 08:56:10 NOTE: package bootstrap-image-1.0: completed Oct 18 08:56:20 celinux-test/omap5912osk works Oct 18 08:56:28 Downloading http://ewi546.ewi.utwente.nl/tmp/rxvt-unicode_7.9-r0_armv5te.ipk Oct 18 08:56:29 Upgrading rxvt-unicode on root from 5.6-r3 to 7.9-r0... Oct 18 08:57:12 koen: you can remove that -1 now :-) Oct 18 08:57:48 no pink here Oct 18 08:59:00 XorA: http://dominion.kabel.utwente.nl/koen/images/Screenshot-1.png Oct 18 09:00:32 koen: pink strikes back Oct 18 09:00:55 I think it's because my Xdefaults has non-standard stuff Oct 18 09:03:12 koen: http://www.xora.org.uk/Screenshot-1.png Oct 18 09:06:40 03koen 07org.oe.dev * rc3193d48... 10/ (1 packages/rxvt-unicode/rxvt-unicode_7.9.bb): rxvt-unicode: remove def-pref -1 from 7.9 Oct 18 09:06:53 koen: wonder if its a theme problem Oct 18 09:07:17 that's a possibility Oct 18 09:08:15 7.9 does feel a lot faster as 5.6 Oct 18 09:08:15 koen: it does doesnt it Oct 18 09:08:15 someone familiar with mainstone? Oct 18 09:08:19 * koen points to lrg Oct 18 09:09:29 * XorA summons lrg Oct 18 09:09:57 I think lrg is done with the mainstone now he has the monahan board :) Oct 18 09:10:41 morning all Oct 18 09:12:49 XorA: i'm impressed :-) Oct 18 09:12:54 hi lrg Oct 18 09:13:10 koen: there is monahan update for mainstone Oct 18 09:13:40 hey florian_kc Oct 18 09:14:46 lrg: there are rumours that you are familiar with mainstone board(s) Oct 18 09:15:23 hrw|work: I've got a few ;) Oct 18 09:15:52 lrg: which MACHINE_FEATURES should it get (except kernel26 ofcourse)? Oct 18 09:18:36 hrw|work: PCMCIA, touchscreen, audio, mmc, ethernet, usb (gadet & host are supported in hw), 32Mb Nor Flash, Irda Oct 18 09:19:10 hrw|work:there is also a keyboard Oct 18 09:19:45 hrw|work: oh, and a SIM socket Oct 18 09:19:51 thx Oct 18 09:20:19 20 MACHINE_FEATURES = "kernel26 pcmcia usbhost usbgadget pcmcia touchscreen keyboard alsa irda" Oct 18 09:20:22 lrg: pci? Oct 18 09:20:30 hrw|work: no Oct 18 09:21:10 hrw|work: 2 UARTS Oct 18 09:24:50 Laibsch: I am back. First, disable INHERIT += " icecc " and make sure your build completes without icecc. Oct 18 09:25:20 hrw|work: oops, forgot some:- camera, memory stick Oct 18 09:25:55 likewise: It does. Oct 18 09:26:22 Laibsch: OK, next step: you got ICECC_PATH exported in the shell that runs bitbake? Oct 18 09:26:44 Laibsch: ICECC_PATH=`which icecc` Oct 18 09:29:49 lrg: I will push first version Oct 18 09:31:46 hrw|work: are you working on Mainstone hw ? Oct 18 09:31:51 lrg: no Oct 18 09:32:32 lrg: but I took it as target for build so prefer to task-base it Oct 18 09:33:03 hrw|work: ah Oct 18 09:34:54 03hrw 07org.oe.dev * rd4e60c43... 10/ (1 conf/machine/mainstone.conf): mainstone: prepare for task-base stuff Oct 18 09:36:42 ~lart screen over ssh chain Oct 18 09:36:42 * ibot slaps screen over ssh chain upside and over the head with one freakishly huge killer whale named hugh Oct 18 09:41:05 I would launch a ncurse software on my devboard trough the serial line , but that failed : Oct 18 09:41:12 Screen must be at least 25x80 !! Oct 18 09:41:29 i think ncurse look after some terminal capabilities Oct 18 09:41:31 Genesis: change terminal emulator to something better? Oct 18 09:41:42 yeap , how to do that ? Oct 18 09:41:51 i don't remember where is it Oct 18 09:42:15 i think the kernel probes the size of the terminal in fonction of the screen on the board Oct 18 09:42:53 but via the serial line, i've a better resolution on my host terminal :) Oct 18 09:43:45 Genesis: which app you use to connect via serial? Oct 18 09:43:54 minicom? hyperterminal? Oct 18 09:44:02 minicom Oct 18 09:44:25 Genesis: echo $LINES; echo $COLUMNS Oct 18 09:44:46 empty variables Oct 18 09:45:02 can i try to export them ? Oct 18 09:45:13 you can Oct 18 09:46:51 yeap ! Oct 18 09:47:17 that works , thanks Oct 18 09:47:23 ettercap works :) Oct 18 09:47:41 Genesis: then you have to find out why they were not set. maybe wrong TERM set Oct 18 09:48:37 noticed that on my todo :) Oct 18 09:50:07 for the previous questoin i ask you another day , i find a good url to learn dd in the same application as me : http://melbourne.wireless.org.au/wiki/?minitar Oct 18 09:50:22 a good pointer for noobies. Oct 18 09:52:58 <[lala]> hrw|work: i found some empty dirs in .oz354x - is it possible that you drop them? Oct 18 09:55:01 Genesis: hardcore machine.. 2M flash, 8M ram Oct 18 09:55:12 [lala]: I can do that Oct 18 09:55:37 <[lala]> let me check which ones ... Oct 18 09:55:39 [lala]: do you remember best automatic way to find all of them? Oct 18 09:56:17 <[lala]> i'll give you a list Oct 18 09:56:30 ok Oct 18 09:58:32 hey guys.. Oct 18 09:58:38 hey tkp Oct 18 09:58:49 I've just done an update on my dev branch checkout... it's been a while Oct 18 09:59:08 and? Oct 18 09:59:11 when I try to cook some recipies, I'm getting complaints about get_kernelmajorversion Oct 18 09:59:17 ERROR: name 'get_kernelmajorversion' is not defined while parsing /home/tom/stv/oe-stv/org.openembedded.dev/packages/linux/linux-geodelx_2.6.11.bb Oct 18 09:59:19 etc etc Oct 18 09:59:25 (lots more similar errors) Oct 18 09:59:32 hrw|work, : it's the hack that is interessant Oct 18 09:59:44 anyideas what this is about? Oct 18 10:00:04 tkp: one bbclass in classes/ changed Oct 18 10:00:14 tkp: cd classes;mtn list unknown Oct 18 10:00:37 http://melbourne.wireless.org.au/wiki/?MinitarHacking Oct 18 10:02:22 hrw|work: none of them have changed... I do have a couple of my own classes in my own bb collection tho Oct 18 10:02:34 03koen 07org.oe.dev * r09ebb399... 10/ (1 packages/stage-manager/stagemanager-native_0.0.1.bb): stagemanager-native: add empty do_stage Oct 18 10:02:58 oh hang on... missed the last part of the error... Oct 18 10:02:58 Please install the git utility Oct 18 10:03:24 ( http://rafb.net/paste/results/OTJEVL76.html ) Oct 18 10:04:28 likewise: I guess I can just set ICECC_PATH in local.conf as well? Oct 18 10:04:41 Trying that now. Oct 18 10:04:50 tkp: Have you installed git? Oct 18 10:04:59 I have now! Oct 18 10:05:13 * tkp tries to re-cook his recipie Oct 18 10:05:14 tkp: *kernel* classes are changed Oct 18 10:05:22 Laibsch: no I would not bet on that. An exported ICECC_PATH is different than one set in local.conf, AFAIK! Oct 18 10:06:09 likewise: OK, it did not work ;-) So where should I set it? Just in the calling bash? Oct 18 10:06:20 Laibsch: yes Oct 18 10:06:34 Righty. Oct 18 10:07:36 got any ideas why this might be: http://rafb.net/paste/results/JOVfFZ34.html ? Oct 18 10:07:57 tkp: yes, "See `config.log' for more details." Oct 18 10:07:57 koen: where is config.log? Oct 18 10:08:05 ${S} Oct 18 10:09:27 ta Oct 18 10:10:10 likewise: Seems to work better now. But still does not finish. I tried to compile nunome. Maybe it does not handle icecc compilation well. Can you tell me about a small package that you can compile? Oct 18 10:11:16 Laibsch: First thing to make sure is that 1) staging/ice and cross/ice directories are created Oct 18 10:11:27 OK Oct 18 10:11:53 Laibsch: compile any native or non-native package, doesn't matter which one. Oct 18 10:12:13 Laibsch: Second thing to make sure is that 2) cross/ice contains a tar file. Oct 18 10:12:47 Both dirs are there and there is a tar under cross. Oct 18 10:13:45 Let me paste the console. This does seem to be relevant. Oct 18 10:15:55 likewise: http://rafb.net/paste/results/9KJ9lN28.html. After all is said and done /export/interim/openembedded/tmp/openzaurus/cross/ice/arm-linux-openzaurusarm-linuxhard4.1.1X24.tar.bz2 is there now. Maybe it will work now. Oct 18 10:17:43 No, it still complains that the file is not there when in fact it is. Oct 18 10:28:16 likewise: Tried to run this on a different computer and I get a different error ;-) http://rafb.net/paste/results/F8KaAo95.html But it completes on that computer. Without distributing compilation to other computers, though. This computer also runs the icecc scheduler for the network. Oct 18 10:47:10 Hi! Oct 18 10:47:19 hey psokolovsky Oct 18 10:47:56 koen: that OE logo rendition in angstrom image indeed sucks ;-) what happened with anti-aliasing? Oct 18 10:48:21 I have a bunch of bb files I need to add to oe that are all closely related, can I put them in a directory under packages? Oct 18 10:48:29 koen: But well, refershing change, anyway. now at least it is branded Oct 18 10:48:46 psokolovsky: I think the image script doesn't like transparency Oct 18 10:49:20 psokolovsky: or psplash scales it Oct 18 10:49:21 koen, but damn, now it doesn't recognize mouse in qemu and wants to calibrate TS ;-( Oct 18 10:49:28 koen: I see Oct 18 10:50:44 psokolovsky: poky loads the wacom tablet driver for that :) Oct 18 10:51:25 koen, well, I swear that angstrom worked with qemu OOB before (just like you suggested to my older post to GPE list) Oct 18 10:52:00 it does use the mouse when I load it on my x86 box :) Oct 18 10:52:29 I don't recall changing anything in X for the past few weeks Oct 18 10:52:30 koen: I wonder, if that's the caveat ;-) will investigate Oct 18 10:52:34 ok Oct 18 10:54:36 koen, aha, it wants to calibrate, but if kill xtscal, it works w/mouse afterwards Oct 18 10:54:56 weird Oct 18 10:55:16 you should be able to calibrate with your mouse as well iirc Oct 18 10:56:54 koen, hm. IIRC, xtscal uses exactly xcalibrate ext, and thing IIRC skips even tslib when it works (as it produces calib file for it), so, it seems liek pretty hardcoded for TS. I may mix something however. Oct 18 10:58:02 koen, anyway, that's build with binutils 2.17.50.0.5. Mine from yesteday, though. Yet need to look at your recipes. I found you reused some 2.16.* .bb for .1, insteda of using 2.17.bb. Oct 18 10:58:25 koen, at least gtk+ was rebuilt with *.50.0.5 here. Oct 18 11:04:13 "System Info" still segfaults Oct 18 11:07:46 'still'? Oct 18 11:08:06 Is there any specific section on the OE wiki/docs about setting the parallel make process up? Oct 18 11:08:18 PARALLEL_MAKE = "-j s/"/>"/ Oct 18 11:08:39 koen: With recommendations for ? Oct 18 11:09:02 What about BB_NUMBER_THREADS? Oct 18 11:09:04 number of cpu's *2 Oct 18 11:09:23 koen, yep, segfaulted for me in qemu all the time. Familiar's works in qemu. don't remember state on real device. Oct 18 11:09:53 pi? Oct 18 11:11:35 psokolovsky: I haven't seen it crash on a device in a long time Oct 18 11:31:38 lardman: and make sure quilt is installed on your system, otherwise the threaded build will need it before it is built :-) Oct 18 11:33:49 likewise: Do you happen to know what http://rafb.net/paste/results/F8KaAo95.html is about and why I see no distribution of compilation tasks to the clients on my LAN? Oct 18 11:36:14 Laibsch: yes. See the icecc.bbclass. The create_env function tries to make a tar file containing the compiler binaries. Oct 18 11:37:13 Laibsch: It does something like "objdump -x $CC | grep NEEDED" to find out which libc library it needs. I suggest you run this by hand on both your native compiler and cross compiler to see if you have objdump installed. Oct 18 11:40:32 it feels like libtool is extracting _l paths from library objexts, does this make sense? Oct 18 11:42:56 likewise: objdump or ${TARGET_PREFIX}objdump ? Oct 18 11:43:24 * koen kick CIA-4 Oct 18 11:43:34 03koen 07org.oe.dev * re5d4b002... 10/ (1 packages/angstrom/task-angstrom-x11.bb): task-angstrom-x11: add rxvt and autostarter back Oct 18 11:58:31 likewise: Thanks Oct 18 12:00:37 okay, my qtopia .bb seems to work reasonably well Oct 18 12:00:50 qtopia core that is Oct 18 12:01:24 ~qtopia core Oct 18 12:01:26 qtopia core is probably qt/embedded. please don't support stupid marketing tricks. Oct 18 12:01:38 no, it's different from qt/e Oct 18 12:01:51 qt/e doesn't implement the qt api fully, right? Oct 18 12:01:53 qtopia core does Oct 18 12:02:11 till a few weeks ago it was called qt/e Oct 18 12:02:15 then they renamed it Oct 18 12:02:23 it's their right Oct 18 12:02:49 sure, but it still is a stupid marketing trick Oct 18 12:02:55 whatever Oct 18 12:03:50 they even retroactively renamed it Oct 18 12:03:54 03koen 07org.oe.dev * rc7eba125... 10/ (5 files in 2 dirs): gpe-session-scripts: remove unused patches Oct 18 12:04:01 it's going to be confusing if you keep calling something which is renamed by the old name ;) Oct 18 12:04:08 so if you were using qt/e 2.x in the past you now use qtopia core Oct 18 12:04:22 i'm talking about 4 Oct 18 12:04:33 even if you burned it into a ROM 2 years ago, tt will claim it's qtopia core Oct 18 12:04:53 now *that* is confusing Oct 18 12:05:10 agreed. Oct 18 12:05:13 i didn't know that Oct 18 12:05:23 i always thought qtopia core was the qt/e of qt4 Oct 18 12:05:58 not according to trolltech :) Oct 18 12:06:09 okay, so what is it called then? qtopia core 4? Oct 18 12:06:14 but yeah, if you download qtopia core you will get qt4/e Oct 18 12:06:34 and my bb file is called qtopia-core_4.2.0.bb ;) Oct 18 12:07:31 it'd be a nice addition to OE Oct 18 12:08:25 i'm not sure about a number of things, but where i have my doubts i've put comments, so i'll do one more test build and run and then put this version in bugzilla Oct 18 12:08:48 koen: objdump Oct 18 12:08:54 i'm happy to try to act on any remarks, and people who know more about the subject than i do are welcomed to modify everything Oct 18 12:09:15 i know i've cut some corners, but i needed this working badly ;) Oct 18 12:15:19 any guidelines for entering a new .bb in bugzilla? Oct 18 12:15:36 attach it instead of pasting it inline Oct 18 12:16:01 koen, I have a bb organizational question Oct 18 12:16:17 I need a to add several bb files, that are closely related Oct 18 12:16:32 can a create a directory in packages and put bb fiels in there? Oct 18 12:16:40 and build them seperately? Oct 18 12:17:06 03koen 07org.oe.dev * rc7eba125... 10/ (5 files in 2 dirs): gpe-session-scripts: remove unused patches Oct 18 12:17:06 03koen 07org.oe.dev * rc7eba125... 10/ (5 files in 2 dirs): gpe-session-scripts: remove unused patches Oct 18 12:17:33 Crofton|home: souns like a plan Oct 18 12:17:39 good :) Oct 18 12:17:58 I don't want to clutter packages with a bunch of stuff people won't recognize :) Oct 18 12:18:45 03koen 07org.oe.dev * rc7eba125... 10/ (5 files in 2 dirs): gpe-session-scripts: remove unused patches Oct 18 12:18:45 03koen 07org.oe.dev * rc7eba125... 10/ (5 files in 2 dirs): gpe-session-scripts: remove unused patches Oct 18 12:19:06 * koen kicks CIA-4 Oct 18 12:19:06 ow Oct 18 12:19:20 * koen kicks CIA-4 again Oct 18 12:27:42 03koen 07org.oe.dev * r782a9f37... 10/ (1 packages/angstrom/task-angstrom-x11.bb): task-angstrom-x11.bb: also ship matchbox-keyboard Oct 18 12:32:22 hrw|work: can we remove ipkg-link from *ipk.bbclass? Oct 18 12:34:12 likewise: The error seems to be gone now. I assume since the tar has been created and is in place now. But still no distribution of compilation tasks. Oct 18 12:34:41 Laibsch: step by step pls, I am not there myself yet. Inspect the tar contents. Oct 18 12:35:04 OK, I see. Sorry to jump ahead. Oct 18 12:37:59 koen: think about those guys which want to use flash as / instead of card Oct 18 12:38:16 Laibsch: Inspect with find again, then tar tjf the tar file Oct 18 12:38:33 Laibsch: make sure you have both the native and cross compiler in the ice/ dirs Oct 18 12:44:40 03hrw 07org.oe.oz354x * r5c3a3f57... 10/ (6 files in 5 dirs): dropped few empty dirs Oct 18 12:46:37 03hrw 07org.oe.oz354x * r5c3a3f57... 10/ (6 files in 5 dirs): dropped few empty dirs Oct 18 12:54:11 likewise: There are three files in there: g++, gcc and as. Inspecting g++ and gcc with objdump tells me they both need libc.so.6 Oct 18 12:55:22 Laibsch: not good, in icecc.bbclass the create_env function does not seem to work well for you. It should add several more files to the tar. Oct 18 12:55:41 Laibsch: such as cc1 cc1plus, libc Oct 18 12:58:07 hrw|work: I don't want ipkg-link in angstrom, because it just doesn't work Oct 18 12:58:36 koen, what's wrong with it? Oct 18 12:58:37 koen: then write RFC about it Oct 18 12:59:09 koen, how about upgrading handhelds-pza-2.6 to 2.6.16-hh7? Oct 18 13:00:39 psokolovsky: what's the difference between 6 and 7? Oct 18 13:01:16 RFC: ipkg-link is distro thing, not an inherent ipkg thing Oct 18 13:01:18 likewise: And I guess that is where the previous error comes in. It just did not find /lib/libc.so.6\nlibc.so.6 Oct 18 13:01:39 * Laibsch inspecting http://www.sidebranch.com/openembedded/icecream/icecc.bbclass further Oct 18 13:02:14 koen, no 6 at all ;-) hh7 has sound working for h4000 and asic3_mmc driver fixes which preclude possible data loss on all asic3 devices (hx4700, h4000, h1910, etc.) Oct 18 13:02:46 psokolovsky: ok, no weird breakages? Oct 18 13:02:47 Laibsch: Do this (fill in your own arch instead of armeb-linux-): objdump -x tmp/cross/armeb-linux/bin/armeb-linux-gcc | grep NEEDED Oct 18 13:02:55 Laibsch: then see where you libs are on your host. Oct 18 13:02:57 koen, no, stable branch. Oct 18 13:03:33 likewise: Did that. It says libc.so.6 Oct 18 13:03:50 Laibsch: as you see, the /lib/ path is hardcoded in the bbclass. It might be wrong for your system? Oct 18 13:04:15 Laibsch: Oct 18 13:04:17 # which libc does the compiler need? Oct 18 13:04:18 libc = "/lib/"+os.popen("objdump -x `which %s` | sed -n 's/.*NEEDED *//p'" % cc).read()[:-1] Oct 18 13:04:30 psokolovsky: what do you think about extracting the diffs per machine and adding those to a mainline tarball, instead of pulling from cvs? Oct 18 13:04:51 likewise: Checking ATM. Oct 18 13:04:57 psokolovsky: I'm getting a bit tired of trying to get people to submit patches upstream Oct 18 13:06:17 koen, that want work that nice at all. hh.org has large bunch of *common* code reused by many machines, which keeps improving contantly. that's why per-machine doens't work, and hh.org kernel cvs in nice thing actually. Oct 18 13:06:27 koen, no worries, we work on that ;-) Oct 18 13:06:38 koen, and CVS is nice staging area Oct 18 13:06:39 psokolovsky: FWIW Giorgio never spoke to me much about ASoC. Oct 18 13:06:42 morning Oct 18 13:06:51 psokolovsky: I've been hearing 'working on that' for the past 3 years Oct 18 13:07:16 psokolovsky: and RP has managed to get a zillion machines and common code upstream Oct 18 13:07:16 koen, working with CVS (even with CVS! ;-) ) is much easier than with 8mb on of patches Oct 18 13:07:43 koen, well, you hear, new people now speaking that, lol ;-) Oct 18 13:07:56 RP, I got idea, that was Liam. anyway, that's on todo Oct 18 13:07:58 koen, who is getting OE working for the EFIKA? Oct 18 13:08:25 Crofton: the person that gets their efika first :) Oct 18 13:08:25 heh Oct 18 13:08:29 my project depends on that project :) Oct 18 13:08:41 how many oe people are on the list? Oct 18 13:09:06 iirc 4 Oct 18 13:09:07 ifaistos, likewise, you and me Oct 18 13:09:20 cool Oct 18 13:09:31 sometime in November they start shipping? Oct 18 13:09:49 Crofton: slightly before hell freezes over Oct 18 13:10:02 likewise: /lib/libc.so.6 is there. But the error message said that /lib/libc.so.6\nlibc.so.6 was not found. So it looks like the \nlibc.so.6 is superfluous. I do not see where it is coming from in your python line, though. I do not know much python which I think is a bummer. Oct 18 13:10:03 yeah, it seems like that Oct 18 13:10:03 Crofton: we've been waiting >6 months now Oct 18 13:10:25 I need a powerpc for some stuff I am interested in Oct 18 13:10:38 get a mac ;) Oct 18 13:10:40 there are xilinx FPGA's with ppc's onboard Oct 18 13:11:41 * koen cleans up desk Oct 18 13:13:25 koen: why have updated the binutils-cross again. I cleaned my cache ysterday and rebuild everything over night, and now i have to do it again. Was there something wrong with the previoues one? Oct 18 13:16:30 hi Oct 18 13:16:34 hi CoreDump|home Oct 18 13:16:38 lo Oct 18 13:17:26 goxboxlive, well, you reported problems with it. people rushed to fix it! ;-) Oct 18 13:17:26 anyone else having problems building gnome-vfs-dbus-2.8.4.4-r1 in .oz? Oct 18 13:20:09 :-) Oct 18 13:21:43 Is there a way that "cu" could be added to the OE dev? We use it to try out the phone on htc universal. Oct 18 13:22:12 With familiar i am able to use the debian package , but that didnt succed with Angstom Oct 18 13:23:08 03koen 07org.oe.dev * r813a6226... 10/ (1 classes/rootfs_ipk.bbclass): Oct 18 13:23:08 rootfs_ipk.bbclass: remove ipkg-link from IPKG_INSTALL Oct 18 13:23:09 * if you want it add it via a different way, some people don't want ipkg-link in the rootfs Oct 18 13:25:58 brb Oct 18 13:27:52 03koen 07org.oe.dev * r813a6226... 10/ (1 classes/rootfs_ipk.bbclass): Oct 18 13:27:52 rootfs_ipk.bbclass: remove ipkg-link from IPKG_INSTALL Oct 18 13:27:52 * if you want it add it via a different way, some people don't want ipkg-link in the rootfs Oct 18 13:29:29 goxboxlive: write recipe and add it into OE Oct 18 13:29:34 * koen wonders what's wrong with CIA Oct 18 13:31:22 i would if i knew how, but i dont have a clue about the rdepends and depends to it. Oct 18 13:32:21 goxboxlive: then now is the time to learn! Oct 18 13:32:42 hmm Oct 18 13:34:04 goxboxlive: Depends: libc6 (>= 2.3.6-6) Oct 18 13:34:12 goxboxlive: so no DEPENDS at all Oct 18 13:34:24 k Oct 18 13:34:42 should it be in the linux folder? Oct 18 13:34:48 own one Oct 18 13:37:14 goxboxlive: it even does not need nothing to get built Oct 18 13:40:49 ok, then i just have to find the source. thx Oct 18 13:40:58 goxboxlive: get from debian Oct 18 13:41:14 ok Oct 18 13:55:33 Laibsch: http://www.mockies.de/main.php?seite=crosscompiling&lang=en Oct 18 14:05:39 03koen 07org.oe.dev * r813a6226... 10/ (1 classes/rootfs_ipk.bbclass): Oct 18 14:05:39 rootfs_ipk.bbclass: remove ipkg-link from IPKG_INSTALL Oct 18 14:05:39 * if you want it add it via a different way, some people don't want ipkg-link in the rootfs Oct 18 14:05:52 CIA-4: we already know that Oct 18 14:11:04 03koen 07org.oe.dev * r56cef8ae... 10/ (1 packages/gcc/gcc-package-cross.inc): gcc-package-cross.inc: libg2c ins't always present, so don't error out if strip has a fit Oct 18 14:15:04 03koen 07org.oe.dev * r813a6226... 10/ (1 classes/rootfs_ipk.bbclass): Oct 18 14:15:04 rootfs_ipk.bbclass: remove ipkg-link from IPKG_INSTALL Oct 18 14:15:04 * if you want it add it via a different way, some people don't want ipkg-link in the rootfs Oct 18 14:17:57 ~irclogs Oct 18 14:19:30 03koen 07org.oe.dev * r56cef8ae... 10/ (1 packages/gcc/gcc-package-cross.inc): gcc-package-cross.inc: libg2c ins't always present, so don't error out if strip has a fit Oct 18 14:19:49 ~kick Oct 18 14:21:54 ~logs Oct 18 14:21:55 apt/ibot/jbot/purl all log to http://ibot.rikers.org// where channelname is html encoded ie: %23debian | lines that start with a space are not shown | some channels have stats at http://ibot.rikers.org/stats/.html.gz, or updated "nightly" Oct 18 14:23:34 hrw|work, :). thanks. Oct 18 14:23:49 hrw|work, how about the new distro, celinux ? Oct 18 14:28:43 leoncamel: its distro for CELF Oct 18 14:29:40 hrw|work, :). wow. but, who supply it ? Oct 18 14:30:44 in which meaning? Oct 18 14:31:39 hrw|work: what are celf's plans for a uclinux based image? Oct 18 14:31:50 no idea now Oct 18 14:32:50 hrw|work, I mean who give you celinux-test distro ? Oct 18 14:35:16 leoncamel: I created celinux-test distro as part of my work for celf Oct 18 14:36:20 hrw|work, OH.. I got it. but, it seems another kind of omap5912osk board. Oct 18 14:37:04 leoncamel: omap5912osk is one of their targets so distro config has everything which is needed to support it Oct 18 14:38:10 hrw|work, OK. Oct 18 14:40:34 cu Oct 18 14:40:45 lecture in 1 hour Oct 18 14:44:41 bitbake -g is meant to build a depgraph, right? Oct 18 14:49:20 is it broken at the moment? Oct 18 15:05:07 furlongm_: for bitbake trunk, it works Oct 18 15:05:37 just checked it out :) Oct 18 15:05:56 likewise: would that also fix gcc-cross not being recognised as a dep? Oct 18 15:06:03 furlongm_: I have uploaded a patch that enables regexp support for --ignore-deps. Oct 18 15:06:16 furlongm_: I don't know really. Oct 18 15:06:37 furlongm_: hm, I only see libtool-cross in my visual depgraph Oct 18 15:08:49 ~seen thejapa Oct 18 15:09:04 thejapa was last seen on IRC in channel #opie, 36d 16h 31m 20s ago, saying: 'hrw|gone: sorry, tried to get online earlier. I'm replacing a fellow teacher at my work. Maybe I can get you online tomorrow.'. Oct 18 15:09:59 ~seen zecke Oct 18 15:09:59 zecke was last seen on IRC in channel #oe, 1d 2h 41m 46s ago, saying: 'wb woglinde '. Oct 18 15:15:55 likewise: trying to build uclibc fails for me as there's no gcc Oct 18 15:17:44 furlongm_: sorry, don't know... Oct 18 15:20:43 hmm, and bitbake from svn seems to givea a lot of errors like "ERROR: Error Method already seen: yada yada" Oct 18 15:21:04 will go back to 1.4.2 and see if that helps Oct 18 15:24:09 don't Oct 18 15:24:15 1.6 is the recommended version Oct 18 15:26:03 koen: want to see if i can get something that builds with it first, it worked with 1.4.2 before so just gonna see if i can pinpoint if that's the prob Oct 18 15:27:30 try actually reporting the problem with 1.6.x Oct 18 15:27:41 1.4.4 is old and mostly unsupported Oct 18 15:29:29 03crofton 07org.oe.dev * rfedad6b5... 10/ (5 files in 4 dirs): Oct 18 15:29:29 Create a directory for all packages specific to the OSSIE project and move Oct 18 15:29:29 the ossiecf package into the new directory. Oct 18 15:32:44 Crofton: OE doesn't support a three level tree Oct 18 15:33:27 I just found that out Oct 18 15:33:37 Crofton|oe: fixed Oct 18 15:33:42 I assume I can have multiple bb files in the ossie dir though Oct 18 15:34:02 koen, what did you fix? Oct 18 15:34:07 ossiecf :) Oct 18 15:34:16 hmm Oct 18 15:34:23 I was fixing also ... Oct 18 15:34:52 03koen 07org.oe.dev * r1ea5da9a... 10/ (4 files in 3 dirs): ossiecf: put it in a place where OE can find it Oct 18 15:36:27 I made the mistake of commiting b4 testing ... Oct 18 15:39:49 good evening Oct 18 15:40:11 tkp: FYI: you need to inherit linux-modules-base now if you want to call get_kernelmajorversion() Oct 18 15:42:44 Will anyone be terribly offended if the omniORB-native package deletes the omniORB librarires from the build staging area? Oct 18 15:43:22 koen: ok, same thing happens with bitbake 1.4.4: "i486-linux-uclibc-gcc: No such file or directory", so this is a problem with the .bb dependencies? Oct 18 15:43:31 I know this is a bad hack, I will comment the reason :) Oct 18 15:43:58 Crofton, wouldn't it work not to install them? Oct 18 15:44:03 well Oct 18 15:44:24 when I build something that uses the libraires built for the target Oct 18 15:44:39 libtool tries to link against the host staging libs and fails :) Oct 18 15:44:52 I've spent two days trying to figure out why ... Oct 18 15:45:15 and I need to make some forward progress Oct 18 15:45:29 hm. probably, there's better way to fix that, but if it couldn;t be found in 2 days... Oct 18 15:45:40 :) Oct 18 15:45:48 I am almost certain there is Oct 18 15:46:27 but I can't think of a case were anyone will actually need those libs, when they do, I will revisit the issue Oct 18 15:47:15 who worked on qtopia/core 4.bb ? Oct 18 15:47:29 it needs to go into /usr/lib, nowhere else Oct 18 15:48:15 and re. qt/e, it implements the whole Qt by design. You can _build_ it to omit some of the features, but you don't need to. Oct 18 15:48:25 mickeyl: giel is the qtopia core man Oct 18 15:48:31 k Oct 18 15:48:41 and hi mickeyl :) Oct 18 15:50:15 hi Oct 18 15:51:38 mickeyl: sadly /opt is part of fhs :( Oct 18 15:53:00 a+ Oct 18 15:54:01 mickeyl: ok, thanks for the tip Oct 18 15:54:29 I seem to be having some trouble building xcalibrateext-0.0+git20060814 Oct 18 15:54:37 it's failing on do_fetch Oct 18 16:00:23 Is the openembedded.org filebrowser still being updated? The .dev branch is showing stuff that is a week old... Oct 18 16:01:16 hvontres|poodle: http://www.openembedded.org/filebrowser/org.openembedded.dev/ Oct 18 16:01:29 hvontres|poodle: what's the timestamp behind the MAINTAINERS file? Oct 18 16:01:57 03koen 07org.oe.dev * r51bc3649... 10/ (1 conf/distro/include/angstrom.inc): angstrom: check for sanity Oct 18 16:02:26 is there any way to get it to retrieve these sources without using git? Oct 18 16:03:32 koen: ok...wierd... nevermind Oct 18 16:03:38 tkp: bitbake git-native, then try again Oct 18 16:12:52 mickeyl: yeah, that did it... guess git-native needs to be added to DEPENDS in xcalibrateext then Oct 18 16:13:56 tkp: it actually means your host git install is busted Oct 18 16:14:08 tkp: and git is in the RequiredSoftware page Oct 18 16:14:56 koen:I have never used git before, but it is on my host Oct 18 16:15:04 and seems to open ok Oct 18 16:15:28 it is were ok, you wouldn't have to build git-native :) Oct 18 16:15:45 how come you rely on the host having git instead of using a native recipie? Oct 18 16:16:04 03koen 07org.oe.dev * rababfa8b... 10/ (1 conf/machine/netbook-pro.conf): netbook-pro: finish conversion to task-base Oct 18 16:17:16 yeah, my local git was using gitfm rather than git-scm Oct 18 16:19:35 unfortunately we have no method to make a fetcher depend on something Oct 18 16:19:41 that would be handy at times Oct 18 16:20:25 mickeyl: that sounds like a good chicken and egg issue tho...I need git to fetch git to build git.... Oct 18 16:21:34 hi guys Oct 18 16:25:46 I'm trying to build gcc for sh3 processor(jornada6xx) but I get this error: Openembedded config's sanity checker detected a possible misconfiguration. And then it tells me to modify some variables. Of which file is it talking about? Oct 18 16:27:59 local.conf Oct 18 16:28:12 in your build-directory/conf Oct 18 16:32:52 woglinde: now it just says please install the git utility :( Oct 18 16:35:18 what has "git" to do with gcc? Oct 18 16:35:58 you git code to compile with gcc Oct 18 16:36:53 TheProphet[S], OE is complete system for building distro, and there's only one place to check, so it checks all which *may* be needed. Oct 18 16:38:49 sorry guys another ERROR: Nothing provides dependency /home/.../Desktop/gcc_3.3.3.bb Oct 18 16:42:16 do I need to include the path for local.conf? Oct 18 16:45:47 in BBPATH .... Oct 18 16:46:11 theprophet did you read "getting started"? Oct 18 16:47:33 How do I get the svn fetcher to update the code? Oct 18 16:47:51 crofotn? Oct 18 16:48:26 I updated a repo with a fix, but bb seems to want to use the stuff it pulled earlier today Oct 18 16:49:19 Crofton: I thought BBPATH was for the packages? Oct 18 16:49:57 that's BBFILES Oct 18 16:50:19 Crofton, kill cached tarball Oct 18 16:50:35 hi andersee Oct 18 16:51:24 Crofton: set SRCDATE = "now" Oct 18 16:51:49 tkp, I bet that is my problem Oct 18 16:51:55 and I always do - export BBPATH=/stuff/build:/stuff/org.openembedded.dev - before bitbake Oct 18 16:52:01 I made the change I need a few minutes ago Oct 18 16:53:43 I actually followed the getting started guide to do that Oct 18 16:54:06 fetcher vomits on SRCDATE = "now" Oct 18 16:54:55 svn? Oct 18 16:54:59 yes Oct 18 16:55:18 remove rev= from SRC_URI if you have it there Oct 18 16:55:30 not there ... Oct 18 16:55:51 hmm... Oct 18 16:56:10 In mine I have also set SRCNAME too Oct 18 16:56:17 ( to the name of the module) Oct 18 16:59:04 Am I not meant to point BBPATH to /stuff/build:/stuff/org.openembedded.dev ? Oct 18 17:00:58 seems to be a few problems with xserver-kdrive-X11R7.1-1.1.0... Oct 18 17:01:23 http://rafb.net/paste/results/DimCl511.html Oct 18 17:01:57 also... I couldn't get it to build at all without modifying the recipe a little: Oct 18 17:02:03 EXTRA_OECONF_append_stv = " --disable-w100 " Oct 18 17:02:43 I did that because it failed here: Oct 18 17:02:43 http://rafb.net/paste/results/FVe0Ff34.html Oct 18 17:10:02 * mrz80 is away: Metaphysically I'm still here; physically, I'm elsewhere Oct 18 17:14:01 *nod* Oct 18 17:14:11 ok, i guess we should do it the other way round Oct 18 17:14:14 --disable-w100 by default Oct 18 17:14:30 enable it only for arm or c7x0 Oct 18 17:15:10 mickeyl: psst: hx4700 Oct 18 17:23:38 yep Oct 18 17:23:42 better arm then Oct 18 17:25:14 yeah I agree.. don't really see why it should be enabled by default Oct 18 17:28:23 hmm Oct 18 17:28:26 FAT: cp437 not found Oct 18 17:28:29 mount: invaliad argument Oct 18 17:28:37 would mount really fail because it can't find a codepage? Oct 18 17:28:42 or could it be something else? Oct 18 17:29:42 it would raelly fail Oct 18 17:30:12 really* Oct 18 17:30:51 hmmm Oct 18 17:31:05 strip got some stricter type checking with the new binutils Oct 18 17:36:30 that's good isn't it? Oct 18 17:36:39 sort off Oct 18 17:36:57 03koen 07org.oe.dev * rcf30bb94... 10/ (1 conf/distro/angstrom-2007.1.conf): angstrom: use binutils-cross 2.17 till we figure out how to teach OE to handle the newer and stricter strip Oct 18 17:37:03 OE now fails left and right because it tries to strip *directories* and strip now errors out over that Oct 18 17:37:11 aye Oct 18 17:39:57 so angstrom now uses a buggier binutils till OE gets fixed Oct 18 17:40:13 * koen puts is somewhere at the bottom of his todo Oct 18 17:40:36 s/is/it/ Oct 18 17:42:16 seems like the svn fetcher gets the earliest rev availble on the current day Oct 18 17:42:27 which is a problem if I made a change today ... Oct 18 17:43:34 Crofton: SRCDATE = "now" is failing? Oct 18 17:43:40 yes Oct 18 17:43:40 * JustinP sticks his tongue out Oct 18 17:43:44 it worked before.... Oct 18 17:43:55 maybe I did it wrong ... Oct 18 17:43:56 how is it failing? Oct 18 17:44:05 oh, it might be broken.... Oct 18 17:44:11 what version of bitbake, BTW? Oct 18 17:44:32 um 1.4.2 ,,,, Oct 18 17:44:37 1.4.2?? Oct 18 17:44:40 try 1.6 Oct 18 17:44:42 heh Oct 18 17:44:43 yeah Oct 18 17:44:55 I was just thinking that after running the rev command Oct 18 17:44:56 sorry, 1 too many ? there Oct 18 17:44:58 * Crofton|oe blushes Oct 18 17:45:29 1.4.2 is technically a stable release, I think, but 1.6 is also stable Oct 18 17:45:47 IIRC 1.4 is still used by familiar Oct 18 17:46:41 koen: fixing it to only do files would be pretty straightforward.. Oct 18 17:46:47 oe that is, when stripping Oct 18 17:47:18 * koen looks at http://www.graphicsmagick.org/ Oct 18 17:47:37 kergoth: it wasn't entirely obvious when I looked at it 5 minutes ago :) Oct 18 17:48:16 any link to what is better about bb 1.6.1? Oct 18 17:48:38 dunno Oct 18 17:49:02 some stuff in changelog Oct 18 17:50:53 looks like now is an addtion since 1.4.2 Oct 18 17:51:47 koen: line 278 of package.bbclass, change if not os.path.islink(file) and isexec(file): to if not os.path.islink(file) and not os.path.isdir(file) and isexec(file): Oct 18 17:52:19 afaict anyway Oct 18 17:52:21 havent tested Oct 18 17:54:15 SRCDATE = "now" does work :) Oct 18 18:02:23 why is this happening sudenly? : http://pastebin.ca/208566 Oct 18 18:04:04 scroll back Oct 18 18:04:13 koen discussed a little while ago Oct 18 18:04:58 thx Oct 18 18:07:28 kergoth: testing it now Oct 18 18:15:59 jo Oct 18 18:34:41 * mrz80 is back (gone 01:24:40) Oct 18 18:43:56 does anyone know if gcc-cross-initial should be built before uclibc-initial? Oct 18 18:44:51 trying to build generic-uclibc with no modifications, and it keeps failing at uclibc Oct 18 18:56:49 Hi guys Oct 18 18:57:43 any idea why I get this error when trying to cross-compile gcc? error: #error "glibc cannot be compiled without optimization" Oct 18 18:58:01 I use bitbake Oct 18 19:00:16 TheProphet[S], still sh3/sh4 targets, right? Oct 18 19:01:18 is there a lot of work to get the backlight work again for the htc universal? Is there alot to rewrite? Oct 18 19:01:55 goxboxlive, very little, actually. have a look at h4000_bl.c for example Oct 18 19:02:29 and compare it with htcuniversal_bl.c ? Oct 18 19:02:48 goxboxlive, I'm on #htc-linux Oct 18 19:07:16 hi everyone Oct 18 19:10:01 hey nicolasfr Oct 18 19:10:14 psokolovsky: what timezone are you in? Oct 18 19:11:36 koen, lol ;-) european, GMT+3 specifically. But I get used to live per Californian time a bit ;-) Oct 18 19:12:13 ah, that explains why I though you where living in the us :) Oct 18 19:12:34 qt-x11-free depends on mysql :/ and mysql won't build today because of missing pthread_kill definition. Why the hell does qt needs mysql ... Oct 18 19:13:37 hi Oct 18 19:13:41 hrw: hi Oct 18 19:13:47 hi all Oct 18 19:13:50 nicolasfr: because someone who use qt3 need mysql support? Oct 18 19:14:57 hrw: I guess, in fact my question was more why won't it build while it was yesterday (and I have not seen any commit concerning qt-x11-free). Oct 18 19:16:15 I can run build to check Oct 18 19:16:53 maybe pthread.h needs to be staged when building pth. So mysql needs dependency on pth. I guess? Oct 18 19:17:08 nicolasfr: you use uclibc? Oct 18 19:17:13 no Oct 18 19:18:18 heh.. cant connect to build machine ;( Oct 18 19:19:18 hrw|tv: ;( Oct 18 19:22:24 psokolovsky: yes sorry it is sh3 Oct 18 19:45:11 TheProphet[S], well, you better catch some sh3 wizard. IIRC, that's v8jlene. Oct 18 19:46:59 TheProphet[S], In the meantime, please google for "how to ask smart question". Be assured that OE works - perfectly, and for many people. So, any issue you have are likely due ti peculiarities of your (mis)configuration. Please be ready to comminicate them, as noone here tele-telnets at your system. anyway, just friendly suggestion. Oct 18 19:53:52 * v8jlene is back Oct 18 20:07:43 I have some files that are installed but not packaged Oct 18 20:07:56 I need to get them packaged, how do I go about making this happen? Oct 18 20:08:56 Crofton: http://www.openembedded.org/user-manual&dpage=ch02s04 :) Oct 18 20:09:22 bother Oct 18 20:09:22 Crofton: the bottom paragraphs Oct 18 20:09:42 I was actually looking for some other info there earlier :) Oct 18 20:09:53 but that will wait until tomorrow Oct 18 20:10:12 yep weirdpath Oct 18 20:10:36 can I wildcard the filelist? Oct 18 20:14:30 excellent wildcards work ... Oct 18 20:14:56 Crofton: look at the existing FILES_ defintions in conf/bitbake.conf, that shows you what the defaults are and the sort of things you can specify... Oct 18 20:16:47 simple wildcarding is working for me Oct 18 20:16:49 thanks Oct 18 20:21:56 Hi, does anyone know to build gpe image in Mac OS X Oct 18 20:37:11 03mickeyl 07org.oe.dev * r93a3d840... 10/ (1 packages/python/python-pycairo_1.2.2.bb): python-pycairo: inherit pkgconfig Oct 18 20:41:01 Any Mac OS X user in there Oct 18 20:48:13 likewise: I do not understand enough python to get to the root cause of that extra \nlibc.so.6 in http://rafb.net/paste/results/F8KaAo95.html Maybe you have an idea? For example what are the % libc and % cc part doing on lines 92 and 94 of http://www.sidebranch.com/openembedded/icecream/icecc.bbclass ? Oct 18 20:50:03 Laibsh: thos are variables that get substituted for the %s in the preceding string...kinda like pintf Oct 18 20:51:17 cu guys Oct 18 20:54:09 hvontres|poodle: Thank you for the explanation. Oct 18 20:54:36 * Laibsch scratches head as to where that extra \nlibc.so.6 is coming from. Oct 18 20:55:39 koen: actually /opt is part of fhs: http://www.pathname.com/fhs/pub/fhs-2.3.html#PURPOSE14 Oct 18 20:57:42 ljp: that's what I said Oct 18 20:58:02 it's just that neither Trolltech nor Qtopia seems to be a LANANA registered provider name :D Oct 18 20:59:59 so /opt is part of fhs, but TT gets crap for using it anyway Oct 18 21:00:59 *shrug* we can put it wherever we like. we build the software. Oct 18 21:02:16 kergoth: the problem is in module-strip.bbclass, which is shell :( Oct 18 21:02:43 modules="`find ${WORKDIR}/install/$p/lib/modules -name \*${KERNEL_OBJECT_SUFFIX}`" Oct 18 21:02:43 if [ -n "$modules" ]; then Oct 18 21:02:44 ${STRIP} -v -g $modules Oct 18 21:03:07 bah, its annoying that we have two places where stripping happens Oct 18 21:04:33 s/the problem is in/the problem is also in/ Oct 18 21:05:31 * koen|tv tries to think shell Oct 18 21:05:44 && ![ -d $foo] ? Oct 18 21:07:42 yeah, that should work. might need a space between ! and [ Oct 18 21:07:56 thx mickey for the fix Oct 18 21:08:07 np Oct 18 21:08:23 statusicon works in matchbox! Oct 18 21:08:40 good work! Oct 18 21:08:44 make us a screeny when yo can Oct 18 21:09:32 heh Oct 18 21:09:40 ok Oct 18 21:09:48 module_strip is vulnerable to path_max as well Oct 18 21:10:16 gotta bitbake gpe-scap Oct 18 21:15:41 03mickeyl 07org.oe.dev * rad1166f6... 10/ (79 files in 4 dirs): add xorg-headers-native to enable building X11 on stations without X11-devel installed Oct 18 21:16:26 mickeyl: Do you have the notes you made at OEDEM handy? Oct 18 21:16:48 RP: yes and no Oct 18 21:16:53 let me see if I can serialize them for you Oct 18 21:16:59 i'll do it tomorrow Oct 18 21:17:12 heh Oct 18 21:17:22 mickeyl: you know, you could have added a tarball :) Oct 18 21:17:33 koen|tv: i could... Oct 18 21:17:37 * koen|tv looks at the zillion .h files ;) Oct 18 21:17:42 nah, it's not so much Oct 18 21:17:51 eventually we find out which ones are not needed and yank 'em Oct 18 21:18:00 it would be better to grab them from xorg anyway Oct 18 21:18:23 but this solution works at last :) Oct 18 21:18:32 mickeyl: ok, thanks. The alternative is you reply to a mail with any points I've missed :) Oct 18 21:19:28 03koen 07org.oe.dev * r9bc80e3b... 10/ (1 classes/module_strip.bbclass classes/package.bbclass): Oct 18 21:19:28 module_strip.bbclass: filter out directories to stop recent binutils from barfing Oct 18 21:19:28 package.bbclass: likewise Oct 18 21:19:43 aah Oct 18 21:19:47 easier than i thought Oct 18 21:20:08 RP: where to? Oct 18 21:20:24 mickeyl: rpurdie@rpsys.net please Oct 18 21:20:30 kergoth: http://lists.linuxtogo.org/pipermail/openembedded-commits/2006-October/001284.html Oct 18 21:20:38 03koen 07org.oe.dev * r72be69ab... 10/ (1 conf/distro/angstrom-2007.1.conf): angstrom: more binutils pingpong :) Oct 18 21:21:00 RP: sent Oct 18 21:21:04 evening zecke Oct 18 21:21:16 can anyone shed some light on the various xserver-xorg packages... Oct 18 21:21:19 koen|tv: !!! you broke XHTML 1.0 with your css update... Oct 18 21:21:35 xvfb ( I thionk I know what that one is ), print and utils Oct 18 21:21:46 zecke: right, I've been asking people for ages to fix the OE html Oct 18 21:21:52 what is xserver-xorg-xprint? Oct 18 21:22:04 ~oe snapshots Oct 18 21:22:06 from memory, oe snapshots is see http://ewi546.ewi.utwente.nl/OE/ Oct 18 21:22:29 I have been using kdrive in the past, but some work I'm doing requires more of a fully fledged xserver implementation Oct 18 21:22:32 ibot: no, oe snapshots is see http://www.openembedded.org/snapshots/ Oct 18 21:22:34 mickeyl: okay Oct 18 21:23:27 nicolasfr: btw, XorA|gon1 is working on improving the networkmanager recipe in OE Oct 18 21:23:30 ~oe oe Oct 18 21:24:08 mickeyl: thanks. I'd just missed one point :) Oct 18 21:24:16 koen|tv: fine, it was segfaulting when I tried last week Oct 18 21:24:58 bleh Oct 18 21:25:06 I'm starting to appreciate python Oct 18 21:25:11 koen|tv: btw, I compiled an angstrom image yesterday and installed it, and its slow (3/4 seconds pause on every app launch...) weird... Oct 18 21:25:36 nicolasfr: weird Oct 18 21:25:38 zecke: You can blame yesterdays bugzilla outage on django :p Oct 18 21:25:41 Laibsch: "% libc" means that the first %s in the print format is replaced with the variable libc Oct 18 21:26:12 Laibsch: I just am learning python on a "need to know basis" Oct 18 21:26:26 I am going to try to adapt this for OE: http://www.emperorlinux.com/images/contrib/wirelessscan.png Oct 18 21:27:10 wifiradar++ Oct 18 21:27:22 likewise: Thanks. hvontres|poodle was already nice enough to explain that to me. Where the extra \nlibc.so.6 coming from or am I getting ahead of the game? Oct 18 21:27:51 koen: yes, I like it more than wifi-radar Oct 18 21:28:20 also some fun stuff to port to OE: http://midnightresearch.com/projects/wicrawl/ Oct 18 21:28:26 zecke: hi Oct 18 21:28:29 likewise: I want to learn it that way too. Never got around to it. But just read a few chapter of "Byte of Python" Quite some time ago I had read another introduction but never got around to using python much so I do not know how to use it really. Oct 18 21:28:50 MAybe this time it will stick better. Oct 18 21:29:12 Good nicht. Oct 18 21:29:16 Laibsch: it is equal to printf("%s bla\n", libc); if you know C (at least I think it is) Oct 18 21:29:16 Good night. Oct 18 21:29:26 Laibsch: CU Oct 18 21:30:09 likewise: but no printf statement in your code so where does that error of mine come from? Oct 18 21:31:09 Laibsch: forget the 'printf', python allows formatted strings everywhere ("I guess" applies here as well :-) ) Oct 18 21:34:44 http://saltycom.net/screenshot-1.png <-- the "I" icon on the lower right is made with three lines of Python with pygtk2.10 Oct 18 21:35:23 woot! Oct 18 21:35:26 ~praise PyGtk Oct 18 21:35:30 All hail PyGtk! Oct 18 21:35:48 ;) Oct 18 21:36:44 wifi applet with wpa support!! many applets! Oct 18 21:37:10 nicolasfr: did you ask something about NetworkManager I didnt see? Oct 18 21:37:27 * koen follows nicolasfr|Zzzz example Oct 18 21:37:29 'night all Oct 18 21:37:41 XorA: koen said you were working on it, I just told him I tried to install it last week and it segfaulted... Oct 18 21:37:51 'night koen Oct 18 21:38:15 nicolasfr|Zzzz: yeah, its a bit unstable, but I did make it get a dhcp lease from a WEP WAP last night Oct 18 21:38:44 XorA: oh great, I should try that tomorrow Oct 18 21:39:32 nicolasfr|Zzzz: it requires some fiddling, which I can tell you tomorrow Oct 18 21:39:32 law has made some work on the card applet, we should very soon have nice GUI tools for network + card management! Oct 18 21:39:52 XorA: yes please, I'd like to test. Oct 18 21:41:20 also have a look at http://www.emperorlinux.com/images/contrib/wirelessscan.png, shoudl be great as a complementary wifi scanner (wifi radar uses 100% CPU on the zaurus :/ and this one has better support for WPA management) Oct 18 21:41:31 night all. Oct 18 21:42:26 * RP finally posts the bitbake session summary for oedem :) Oct 18 21:46:32 RP: excellent :) Oct 18 21:46:34 g'night Oct 18 21:47:33 'night mickey|zzZZzz Oct 18 21:47:39 whole lotta snoring here tonight Oct 18 21:49:55 How can I find out what is the lib directory of the development host? I.e. lib32 or lib64 or /lib? Oct 18 21:50:36 likewise: ls /lib/* Oct 18 21:50:46 for each of the above :-) Oct 18 21:51:42 XorA: I want to ask the system for the absolute location of libc. I should question the dynamic linker I assume. Oct 18 21:52:29 likewise: ah, toolchains freakiness, sorry dont know Oct 18 21:53:24 * XorA waits for machine to cool down a bit Oct 18 21:54:08 ldconfig -p | grep -e libc.so.6$ Oct 18 21:54:17 but I assume this isn't reliable Oct 18 21:54:25 XorA: you were baking eggs? Oct 18 21:54:31 XorA: or bits? Oct 18 21:54:34 likewise: my amd64 overheats Oct 18 21:55:02 XorA: my amd64 really runs cool here, 'only' 38 degrees Celcius or so Oct 18 21:55:40 likewise: 65 a minute ago Oct 18 21:57:17 31C here on my AMD X2... Oct 18 21:57:58 * XorA probably wont get a miniITX next time Oct 18 22:04:45 XorA: or check for proper heat conducting paste Oct 18 22:05:17 Any sed pro's in here? Oct 18 22:06:21 pros? Oct 18 22:06:31 I've hacked a few build with sed.... Oct 18 22:06:42 likewise: its not the cpu, the MB overheats :-( Oct 18 22:06:49 anyway time for me to zzZZzz Oct 18 22:06:59 JustinP: I need to extract the abolsute path to libc from this output: Oct 18 22:07:01 ldconfig -p | grep -e libc.so.6$ Oct 18 22:10:32 likewise: ldconfig -p | grep -e libc.so.6$ | sed 's:[^/]*/:/:' Oct 18 22:13:43 JustinP: thanks. I don't get the sed but it works. Oct 18 22:14:20 Replace anything before the last / to /, right? Oct 18 22:15:04 likewise: replace 0 or more non-"/" followed by a "/" with "/" Oct 18 22:15:34 JustinP: thanks Oct 18 23:04:14 good nite all **** ENDING LOGGING AT Thu Oct 19 02:59:57 2006