**** BEGIN LOGGING AT Tue Jan 02 02:59:57 2007 Jan 02 03:14:30 03mwester 07org.oe.dev * rac6e8d15... 10/ (1 conf/distro/debianslug.conf conf/distro/openslug.conf): openslug/debianslug: add via-velocity NIC driver module to the rootfs Jan 02 03:41:51 * chouimat|ibook is away: Gone away for now. Jan 02 03:42:02 * chouimat|ibook is back. Jan 02 07:46:02 hi people. my package manager wants to upgrade monotone to 0.32, is that version compatible with OE metadata? Jan 02 07:54:40 ArjanS: yes, network protocol is unchanged Jan 02 07:59:46 rwhitby: thank you. my current version is 0.29, do I need a mtn migrate or something like that? Jan 02 08:00:25 btw my mtn is not working. Everything is installed. any idea? Jan 02 08:05:27 ArjanS: there will be a migrate step (or maybe two) in your working area. See the monotone site UPGRADE page for details. Jan 02 08:09:36 rwhitby: thanks! Jan 02 08:53:34 morning all Jan 02 09:04:07 hi ade|desk Jan 02 09:04:08 hi zecke Jan 02 09:07:09 hey master! Jan 02 09:07:15 happy new year, success and health Jan 02 09:19:11 morning Jan 02 09:19:51 zecke, where may we corespond regarding the mq200? Jan 02 09:20:07 (I am that guy with the simpad..) Jan 02 09:20:19 are registered at oesf? Jan 02 09:22:23 no I'm not at oesf Jan 02 09:22:50 the code and doc is there where it has been for the last ~2 years and 'mr_nice' already pointed you to it Jan 02 09:23:14 pb_ long time no see, what are you up to these days ? Jan 02 09:24:51 zecke, I bag your pardon, where is the code ? Jan 02 09:25:09 I d/l the .git from ~zecke Jan 02 09:25:22 niv_one_three: do me a favor, check the log from yesterday Jan 02 09:25:23 what is the name I should find ? Jan 02 09:25:44 I will try, but this irssi , is full Jan 02 09:25:47 niv_one_three: then do `find . -name '*mq200*` Jan 02 09:25:52 10x Jan 02 09:26:25 niv_one_three: but the mq200 'driver' is where normal video drivers are located Jan 02 09:27:35 niv_one_three: e.g. http://handhelds.org/cgi-bin/cvsweb.cgi/linux/kernel26/drivers/video/mq200/ for older copy... Jan 02 09:27:39 well I get no hits from the find Jan 02 09:27:50 I will try mr_mice l8r Jan 02 09:27:55 oh Jan 02 09:27:56 ok Jan 02 09:28:48 niv_one_three: make sure to search at the right spot... Jan 02 09:29:05 oh, I just saw the hh repository via that web cgi Jan 02 09:29:10 ok Jan 02 09:29:13 thanks Jan 02 09:29:28 sweet, I see the code Jan 02 09:30:30 it is based on a driver from lineo. as the datasheets were not 'googable' we used mem dumps to set it up Jan 02 09:32:01 I will look in to it in 10 hours, thanks Jan 02 09:32:30 btw, I got a basic C knowladge, hope the code is easy to grasp Jan 02 09:33:46 btw, is the a public log collected from this chat? Jan 02 09:34:11 yes Jan 02 09:34:14 ~log Jan 02 09:34:26 well, log is as piece of wood, or a record, or the opposite of exponentiation Jan 02 09:34:27 10x Jan 02 09:34:30 ~help Jan 02 09:34:36 I will Jan 02 09:34:53 ~logs Jan 02 09:34:56 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" Jan 02 09:39:15 good morning all Jan 02 09:39:21 Morning Jan 02 09:43:13 morning Jan 02 09:54:59 zecke: thanks. a happy new year to you as well. Jan 02 09:58:49 pb_ : hi I am new here, are you a DEV? Jan 02 09:59:01 not THe dev but a dev Jan 02 09:59:04 :-) Jan 02 10:02:25 niv_one_three: not really, but zecke is Jan 02 10:03:55 lol Jan 02 10:04:03 is diet-x11 failing for anyone else (makekeys.c:34:19: X11/X.h: No such file or directory Jan 02 10:04:03 ) ? Jan 02 10:04:06 niv_one_three: pb_ is our shy master Jan 02 10:04:44 morning all Jan 02 10:05:13 rwhitby: Try cleaning it and rebuilding it. It does that when you try to rebuild it iirc. Fixes welcome :) Jan 02 10:05:26 RP: this is with tmp blown away Jan 02 10:05:29 hey RP Jan 02 10:05:46 rwhitby: ah :-/ Jan 02 10:06:02 * rwhitby is in the middle of a complete nslu2-linux feeds rebuild ... Jan 02 10:06:11 and it's looking a bit scary. Jan 02 10:11:51 koen: good morning, thanks for the davinci ml link. Jan 02 10:12:39 likewise: goedenmorgen! Jan 02 10:12:50 * likewise rubs eyes Jan 02 10:14:42 XorA|gone: it really starting to look like that .la problem you highlighted a while back Jan 02 10:15:23 XorA|gone: OE needs to sed out anything matching TMPDIR/work Jan 02 10:15:32 Anyone know when bitbake -k stopped doing what it should? Jan 02 10:15:48 rwhitby: which version of bitbake do you use? Jan 02 10:15:53 1.6 head Jan 02 10:16:02 eek, no idea then :} Jan 02 10:16:19 1.6.3 to be exact Jan 02 10:16:25 * rwhitby hopes that is 1.6 head Jan 02 10:16:54 sorry I was not aware of that Jan 02 10:17:12 no problem. that's what testers are for :-) Jan 02 10:18:15 looking at my schedule for january, it is best to file a bug report :} Jan 02 10:28:01 zecke: what's the easiest way to get bitbake to tell me which package is depending on diet-x11, so I can expurgate it and allow my feed rebuild to continue? Jan 02 10:28:45 rwhitby: you ask the tough questions Jan 02 10:29:22 rwhitby: use a bitbake where -g is working and then look for edges pointing to virtual/x11 (diet-x11) Jan 02 10:30:12 thx Jan 02 10:30:44 hmm - 1.6.3 is not such a version :-) Jan 02 10:31:01 tough questions at the morning Jan 02 10:31:28 9:30pm here :-) Jan 02 10:31:40 I've had all day to think of them. Jan 02 10:31:54 you seem to have used the time visely Jan 02 10:31:59 rwhitby: what is broken? Jan 02 10:32:14 preferred_versions[pn] = bb.providers.findBestProvider(pn, pkg_pn)[2:4] Jan 02 10:32:15 TypeError: findBestProvider() takes at least 3 arguments (2 given) Jan 02 10:32:29 okay that one is reported already, try 1.4 :} Jan 02 10:34:37 03koen 07org.oe.dev * r7df4ca6e... 10/ (1 packages/linux/linux-rp_2.6.19+git.bb): linux-rp 2.6.19+git: s/DOSRC/RPSRC/ for kexec patch Jan 02 10:38:43 zecke: I think that findBestProvider is fixed in 1.6 svn (1.6.3). We should release it at some point Jan 02 10:39:15 hmm. or is 1.6.3 broken too? :-/ Jan 02 10:39:30 the one I'm using reports itself as 1.6.3 Jan 02 10:39:34 I can update from svn just to check Jan 02 10:40:29 yep, matches svn Jan 02 10:40:41 rwhitby: ok, thanks. I'll take a look Jan 02 10:41:43 RP: rebuild of diet-x11 definitely doesn't work. Note that the box I'm building on has no host X include files installed if that matters. Jan 02 10:41:58 OE has its own x includes Jan 02 10:42:06 yeah, that's what I figured. Jan 02 10:42:23 rwhitby: I'm thinking of a different problem with makedevs Jan 02 10:50:51 03rpurdie 07bitbake-1.6 * r720 10/bin/bitbake: bin/bitbake: Fix findBestProvider parameters (close #1715) Jan 02 10:51:59 rwhitby: That should solve the bitbake -g problem Jan 02 10:53:20 <[cc]smart> |--- git.orig/drivers/video/backlight/hx2750_bl.c 2006-12-30 15:03:10.000000000 +0000 Jan 02 10:53:20 <[cc]smart> |+++ git/drivers/video/backlight/hx2750_bl.c 2006-12-30 15:29:55.000000000 +0000 Jan 02 10:53:46 [cc]smart: That backlight patch needs to be after the hx2750 backlight driver patch Jan 02 10:54:33 RP - that seems to have fixed the problem, but I get lots of "ERROR with provider" messages now, and no dot output :-( Jan 02 10:54:50 * rwhitby guesses that slugos-packages dependencies are royally screwed. Jan 02 10:55:14 rwhitby: FWIW, that code is totally rewritten in bitbake 1.7... Jan 02 10:55:26 rwhitby: I don't think so (slugos-packages deps) Jan 02 10:55:39 first one is "NOTE: ERROR with provider: patch-native Jan 02 10:55:39 " Jan 02 10:55:54 rwhitby: It doesn't handle anything in ASSUME_PROVIDED Jan 02 10:56:12 oh - I see the output files now. sorry. Jan 02 10:57:27 aha - python->tk->x11 Jan 02 10:57:52 that's gonna kill a lot of packages. Jan 02 10:59:32 so diet-x11 is working for everyone else who is rebuilding after deleting tmp today? Jan 02 11:00:32 03rpurdie 07org.oe.dev * r9f89b3c9... 10/ (1 packages/linux/linux-rp_2.6.19+git.bb): linux-rp-2.6.19+git: Fix patch order Jan 02 11:01:23 * RP hasn't wiped tmp for a month or so since I made those changes locally... Jan 02 11:02:04 * koen wipes tmp/ almost daily Jan 02 11:02:20 which OE package is meant to stage the X11 includes? Jan 02 11:03:32 hmm - the include files are in staging/i686-linux/include/X11, but diet-x11 makekeys.c is not finding them Jan 02 11:03:35 xorg-headers-native Jan 02 11:04:01 rwhitby: Keep in mind makedevs is compiled for the native system, not cross compiled iirc Jan 02 11:04:33 I don't see the connection to makedevs? Jan 02 11:04:46 (I'm talking diet-x11 and makekeys) Jan 02 11:05:32 ok, the includes are in staging/armeb-linux/include/X11 too Jan 02 11:08:19 rwhitby: I meant makekeys - its not a cross compiled util so perhaps it uses the system cflags Jan 02 11:12:45 that might make sense - I don't have any X11 includes installed on the host Jan 02 11:13:41 rwhitby: Have a read of the libx11 .bb file and see what it does to makekeys. I suspect you need to add some headers to the compile command Jan 02 11:13:45 re Jan 02 11:13:53 that must be a change since last time we compiled the feed though, cause it's the same host config Jan 02 11:14:57 That .bb is ugly :-( Jan 02 11:18:44 /nick hrw Jan 02 11:18:50 morning Jan 02 11:18:58 u-boot CVS needs to be removed Jan 02 11:19:30 Laibsch: I will clean that perl config in next few days Jan 02 11:24:03 hi hrw Jan 02 11:25:06 hey hrw Jan 02 11:25:40 Crofton: can we replace uboot_LABEL*.bb SRC_URI with archive from ftp.denx.de? it will give us one recipe with normal fetching more Jan 02 11:26:29 ~lart 128kbps connections Jan 02 11:26:30 * ibot offers 128kbps connections some herring Jan 02 11:28:22 man I hate perl Jan 02 11:28:43 patching file drivers/video/backlight/hx2750_bl.c Jan 02 11:28:43 Hunk #1 FAILED at 95. Jan 02 11:28:43 1 out of 1 hunk FAILED -- saving rejects to file drivers/video/backlight/hx2750_bl.c.rej Jan 02 11:29:01 RP: the linux-rp+git still seems to have a problem Jan 02 11:29:01 03rpurdie 07org.oe.dev * rb154b78e... 10/ (4 files in 3 dirs): Add qemux86.conf from Poky (and update qemu rootfs sizes) Jan 02 11:29:08 03rpurdie 07org.oe.dev * r67fe520d... 10/ (1 classes/package_deb.bbclass): Add package_deb.bbclass from Poky Jan 02 11:29:26 koen: I seem to have badly broken it :-/ Jan 02 11:29:29 RP: i586 or i686 qemux86? Jan 02 11:29:38 hrw, is there a recipe? Jan 02 11:29:45 hrw: i586 Jan 02 11:29:58 Crofton: only src_uri need to be changed to same as in 1.1.x ones Jan 02 11:30:01 hrw: What was the reason for i686? Jan 02 11:30:12 ok Jan 02 11:30:26 RP: progear, guiiness are already i686 angstrom targets Jan 02 11:30:28 I will try and test that and the new kernel in the OSK this week Jan 02 11:30:49 Crofton: 1.1.6 u-boot is also available - maybe will be even better then the one with label Jan 02 11:30:58 ok Jan 02 11:31:26 Crofton: I will drop uboot_cvs.bb soon - uboot_git.bb with DEF_PREF=-1 will get added Jan 02 11:31:36 hrw: Arguably this means we have more testing coverage then ;-) Jan 02 11:31:54 RP: ;D Jan 02 11:32:42 ok Jan 02 11:32:52 what are your thought on OE code sprints? Jan 02 11:32:57 Is uboot cvs still valid even? Jan 02 11:33:09 koen: local face to face? Jan 02 11:33:12 Crofton: with srcdate <20060930 it is Jan 02 11:33:29 e.g. we decide to concentrate on e.g. on thing (.dep support, packaged-staging, tinderbox) for a week Jan 02 11:33:32 old stuff :) Jan 02 11:33:47 zecke: ideally, but I think we'll have to settle for distributed sprints Jan 02 11:35:04 it would be nice if tinderbox would be able to show machine/distro combo in boxes view Jan 02 11:35:31 koen: Its a nice idea. It would rely on people having time windows to commit to it though :-/ Jan 02 11:38:49 03rpurdie 07org.oe.dev * rc87c6565... 10/ (1 classes/rootfs_deb.bbclass): Add rootfs_deb.bbclass from Poky Jan 02 11:39:46 FWIW, .deb package generation works fine. The rootfs generation is a bit unpredictable :} Jan 02 11:43:08 RP: will get solved one day Jan 02 11:43:56 hrw: Its not far off, just needs time. It basically works now Jan 02 11:47:35 koen, [cc]smart: This time the kernel should really work ;-) Jan 02 11:47:42 03rpurdie 07org.oe.dev * r52d8fae9... 10/ (1 packages/linux/linux-rp_2.6.19+git.bb): linux-rp-2.6.19+git: Really fix the backlight patch Jan 02 11:48:42 NOTE: package linux-rp-2.6.19+git-r4: task do_compile: started Jan 02 11:48:51 hrw where do you see uboot 1.1.6? Jan 02 11:50:26 I should also mention, ipk and deb package generation can run in parallel, just INHERIT = "package_ipk package_deb". The order is important as the first one is used for rootfs generation... Jan 02 11:51:12 (you can override that with IMAGE_PKGTYPE = "deb" etc. Jan 02 11:51:22 NOTE: package linux-rp-2.6.19+git: completed Jan 02 11:51:44 koen: What speed machine is that? :) Jan 02 11:51:54 * koen hints RP at the documentation branch Jan 02 11:52:02 RP: dual opteron with fast discs Jan 02 11:52:31 koen: nice. Mine might get there in about 10 minutes :-/ Jan 02 11:53:00 uf. got screen back Jan 02 11:53:04 Crofton: denx.de Jan 02 11:53:16 If I touch the documentation branch, I'll get asked where the poky documentation is :-/ Jan 02 11:53:27 Crofton: ftp.denx.de/pub/u-boot/ Jan 02 11:53:40 RP: in the documentation branch of course :) Jan 02 11:53:42 I see the tag, but there is no tarball listed on the sf page ... Jan 02 11:54:07 Crofton: ignore sf.net Jan 02 11:54:12 Crofton: they do not use it Jan 02 11:54:18 * RP wishes life was simple sometimes Jan 02 11:54:28 Crofton: git is hosted on denx.de, archives too Jan 02 11:54:42 they still link there ... there is 1.1.5 on sf, but not 1.1.6 which came a week later Jan 02 11:54:51 just confusing Jan 02 11:55:24 "Tarballs can be found at the DENX FTP server." Jan 02 11:55:42 <[cc]smart> patching file drivers/video/backlight/hx2750_bl.c Jan 02 11:55:42 <[cc]smart> Hunk #1 FAILED at 95. Jan 02 11:56:19 <[cc]smart> but... updateing monoton i saw no difference, so do i have to remove the cached patch first ? Jan 02 11:56:48 [cc]smart: No, you should have pulled a change in which would make it use a new backlight patch Jan 02 11:57:12 <[cc]smart> leme recheck Jan 02 11:57:45 <[cc]smart> yes. i missed that on first sight: mtn: modifying packages/linux/linux-rp_2.6.19+git.bb Jan 02 11:58:58 hrw, I just skimmed the changelog, looks like 1.1.6 should be good for OSK Jan 02 11:59:08 Crofton: good Jan 02 11:59:16 <[cc]smart> i have a reject file, but also: Applied patch backlight_parameter_fix-r0.patch (forced; needs refresh) Jan 02 11:59:32 [cc]smart: It should be using -r1 Jan 02 11:59:42 RP: did you bumped PR? Jan 02 12:00:04 hrw: No, since it didn't work before. Jan 02 12:00:10 [cc]smart: clean it and try again Jan 02 12:00:35 zecke: http://bash.org/?18477 Jan 02 12:00:39 <[cc]smart> i rmoved the directory of linux-xxx and the stamps of same, iirc that should do it Jan 02 12:01:24 [cc]smart: That sounds like a lot more effort than "bitbake virtual/kernel -c clean" ;-) Jan 02 12:01:39 Presuming you have it as the default kernel... Jan 02 12:01:52 [cc]smart: bitbake -crebuild.... Jan 02 12:04:35 argh... putty -> ewi is getting too much drops Jan 02 12:08:38 looks like it is ewi problem - I had to cut off tinderclient usage because I was unable to do any builds Jan 02 12:11:30 hrw: there is some weird interaction with the lab firewall and ewi Jan 02 12:11:42 hrw: it only seems to show up when using ssh over here Jan 02 12:12:10 but my crappy cable connections makes it hard to tell when it happens and when my cable is just crappy Jan 02 12:13:17 koen: no plans on moving tinderbox to LTG? Jan 02 12:13:34 hrw: plans yes Jan 02 12:13:40 hrw: waiting on zecke for that Jan 02 12:13:51 hrw: or we could move it to the new OE server Jan 02 12:14:55 ok Jan 02 12:15:51 :} Jan 02 12:16:00 koen: where does new machine will be hosted? Jan 02 12:16:20 hrw: next to ftp.nl.debian.org :) Jan 02 12:17:06 it will probably become openembedded.snt.utwente.nl Jan 02 12:17:28 hm.. collie build is fscked a bit Jan 02 12:17:34 need to again set of config fixes Jan 02 12:18:30 hrw, do we know what other machines use u-boot? Jan 02 12:18:53 gta01 for one Jan 02 12:19:10 mickey|zzZZzz: pïng Jan 02 12:22:10 hm.. just 87G free left... Jan 02 12:22:37 Crofton: amsdelta omap5912osk for sure Jan 02 12:23:04 ok Jan 02 12:23:17 just curious what might be broken by an update :) Jan 02 12:23:22 Crofton: some more probably too but does not have it in extra depends Jan 02 12:23:49 Crofton: add new versions and inform about it on ML with request for testing Jan 02 12:24:00 will do Jan 02 12:24:08 niv_one_three: hi Jan 02 12:24:17 Crofton: it would be best to have two uboot versions in OE: latest one + git one Jan 02 12:25:58 hrw, I agree Jan 02 12:27:28 allo - python depends on tk which depends on libX11 Jan 02 12:27:34 this is not great for headless systems Jan 02 12:27:49 mr_mice: hi Jan 02 12:28:04 libX11 can be used to run clients on headless systems which display on remote servers Jan 02 12:28:20 (a number of people do such things with the nslu2 today) Jan 02 12:28:24 blaster8: I think that is only for building. Installing python does not pull in X libs. Jan 02 12:28:52 that's more reasonable Jan 02 12:28:53 blaster8: I've used python on several headless projects in the last year. Jan 02 12:28:59 I run python on the osk and it works fine wo a display Jan 02 12:29:19 niv_one_three: did you get the git tree? Jan 02 12:29:45 mr_mice: It burned all night Jan 02 12:29:50 blaster8: where do you see a libx11 dependency: http://www.angstrom-distribution.org/repo/?action=details&pnm=python ? Jan 02 12:30:02 I am now at work and sshing how shows it was d/l Jan 02 12:30:30 koen: I think blaster8 is referring to the diet-x dependency that tk has (where tk is a dependency of python) Jan 02 12:30:33 1011M handhelds.org/~zecke" Jan 02 12:30:36 ok? Jan 02 12:30:50 (build-time dependency) Jan 02 12:31:04 niv_one_three: with wget -r or? Jan 02 12:31:06 blaster8: building python creates lot of python packages. some of them needs tk, some of them needs x11, some needs other stuff... but no one force you to install all python packages Jan 02 12:31:10 rwhitby: I was hinting at blaster8 to check claims before sprouting nonsense Jan 02 12:31:26 night all Jan 02 12:31:43 niv_one_three: then you will need to delete the index.html stuff in it Jan 02 12:31:48 koen: as I understand it, the python meta-package depends on libX11 however indirectly Jan 02 12:32:01 blaster8: meta-package or recipe? Jan 02 12:32:06 blaster8: it doesn't Jan 02 12:32:10 ok, fine Jan 02 12:32:15 need to check our build system then Jan 02 12:32:18 niv_one_three: git clone not wget... Jan 02 12:32:37 zecke: wget is also ok Jan 02 12:32:47 blaster8: http://openembedded.org/oemyths Jan 02 12:32:53 zecke: he had a problem with curl support in his git tree Jan 02 12:33:03 we need to clean up this page and promote to mainpage somewhere Jan 02 12:33:22 niv_one_tree: here is a description about the git stuff http://opensimpad.org/index.php/SLACKpad_from_scretch:playground#Zeckes_Git_Kernel Jan 02 12:33:33 <[cc]smart> i have - ${RPSRC}/backlight_parameter_fix-r0.patch;patch=1;status=pending - in linux-rp_2.6.19+git.bb. so it seems despite an update i did not get to fix-r1 Jan 02 12:33:38 python depends on ncurses + python-core, python-core depends on libpython, libpython depends on libc6 Jan 02 12:33:41 no x11 there Jan 02 12:33:50 not even indirectly Jan 02 12:33:54 hrw: that would be fine, but diet-x11 is broken for us here :) Jan 02 12:34:15 <[cc]smart> RP: i have - ${RPSRC}/backlight_parameter_fix-r0.patch;patch=1;status=pending - in linux-rp_2.6.19+git.bb. so it seems despite an update i did not get to fix-r1 Jan 02 12:35:15 when this connection will die I will go into gone status Jan 02 12:35:43 niv_one_three: you will need to clone the downloaded dir to get a working git tree Jan 02 12:35:44 blaster8: so your problem is diet-x11 not python Jan 02 12:36:04 primarily it's diet-x11 Jan 02 12:36:17 zecke: I recompiled git with curl supprt Jan 02 12:37:03 I can run git again, so it will be at home when I return Jan 02 12:37:17 did the wget give partial d/l? Jan 02 12:37:36 need I run git to get the stuff again? Jan 02 12:38:14 !oebug 1739 Jan 02 12:38:16 * * Bug 1739, Status: NEW, Created: 2007-01-02 05:08 Jan 02 12:38:17 * * openembedded(AT)hrw.one.pl: strace does not build for mipsel Jan 02 12:38:18 * * http://bugs.openembedded.org/show_bug.cgi?id=1739 Jan 02 12:39:32 well, I am keeping this chanenl to monitor, and getting back to work Jan 02 12:39:36 koen: http://www.openembedded.org/repo/org.openembedded.dev/packages/python/python_2.4.4.bb Jan 02 12:39:43 DEPENDS = "python-native readline zlib gdbm openssl tcl tk" Jan 02 12:39:47 mr_mice: are you a dev here? Jan 02 12:39:48 and tk depends on libX11 Jan 02 12:39:50 blaster8: so? Jan 02 12:39:59 blaster8: you said package Jan 02 12:40:05 niv_one_three: no read http://opensimpad.org/index.php/SLACKpad_from_scretch:playground#Getting_Zeckes_git_tree Jan 02 12:40:07 blaster8: DEPENDS is BUILDTIME dependency not RUNTIME Jan 02 12:40:10 may we get in to the mq200 , in 7 hours? Jan 02 12:40:18 blaster8: the python package doesn't depend on x11, not even indirectly Jan 02 12:40:25 niv_one_three: no I am no dev here Jan 02 12:41:00 mr_mice: what comes after Getting_Zeckes_ ? Jan 02 12:41:09 in the URL you gave? Jan 02 12:41:39 niv_one_three: http://shortlink.org/1062 Jan 02 12:41:42 oh ok I get the URL Jan 02 12:42:18 ok I am leaving the git , do its thng Jan 02 12:42:21 thing Jan 02 12:42:35 niv_one_three: you did already a wget -r or? Jan 02 12:42:52 niv_one_three: then cd to the directory Jan 02 12:43:06 niv_one_three: find ./ -name index.html* -delete Jan 02 12:43:19 niv_one_three: to delete all index.html stuff Jan 02 12:43:20 yaI did the wget Jan 02 12:43:40 niv_one_three: and then do a git clone linux-2.6/.git/ ../Zeckes.Kernel.git.tree Jan 02 12:43:48 fsck.. my angstrom/collie build tries to build orinoco 0.13e... Jan 02 12:44:57 niv_one_three: in Zeckes.Kernel.git.tree is then the complete tree as if you had done a git clone http://handhelds.org/~zecke/simpad/linux-2.6/.git Zeckes.Kernel.git.tree Jan 02 12:45:01 ok I did it Jan 02 12:45:30 thanks for sharing your knowladge Jan 02 12:45:33 niv_one_three: then the best is to make a copy of it Jan 02 12:45:39 niv_one_three: no problem Jan 02 12:46:03 niv_one_three: we have the same target Jan 02 12:46:10 ARGH Jan 02 12:47:12 I should have my new fuji S9000 next week Jan 02 12:47:18 will post better shots Jan 02 12:47:36 hope to have the code runing Bluetooth by then Jan 02 12:47:50 niv_one_three: what about your problem you have written on oesf forum? Jan 02 12:48:03 ? Jan 02 12:48:18 hrw: do you have a couple of minutes (regarding your company)? Jan 02 12:48:40 will come back here Jan 02 12:48:47 monitoring .. Jan 02 12:49:06 pgfeller: january is booked for my transofrmations, leaving old company.. Jan 02 12:49:14 sorry mr_mice, need to pay them taxs my coutry charges :-) Jan 02 12:49:19 pgfeller: then I'm booked by one OE related company Jan 02 12:49:25 niv_one_tree: np Jan 02 12:49:50 hrw: I see - cool :-) ... will come back to you in feb. if I was not able to handle things myself ... good luck for your transition Jan 02 12:50:18 koen: in the topic s/no/not Jan 02 12:50:22 lol too late Jan 02 12:50:54 ls Jan 02 12:51:43 ade|desk: . Jan 02 12:51:46 ade|desk: .. Jan 02 12:51:56 grr wrong window Jan 02 12:52:10 ~lart self Jan 02 12:52:11 * ibot squishes self like a bug Jan 02 12:52:20 ~lart me Jan 02 12:52:20 * ibot gives ade|desk a good seeing to Jan 02 12:56:59 koen: pong Jan 02 12:57:09 hi mickeyl Jan 02 12:57:10 morning guys, happy new year for all I didn't see previously Jan 02 12:58:07 we need to pester AMD guys to check geode*x support... Jan 02 12:58:43 and I need to get cash and buy laptop with linux for me Jan 02 12:58:52 <[cc]smart> RP: now i received -r1... Jan 02 13:03:13 koen: MACHINE_KERNEL_VERSION = 2.6 is useless in angstrom.conf Jan 02 13:03:24 koen: distro is loaded AFTER machine Jan 02 13:03:40 right Jan 02 13:04:17 collie part in -oabi.conf would need to be written ;( Jan 02 13:10:24 ping kergoth Jan 02 13:11:39 Am I right in assuming that something like http://edu.kde.org/kiten/ would not be possible on an embedded device? Jan 02 13:12:15 why not? Jan 02 13:12:26 (ignoring KDEs broken buildsystem from hell) Jan 02 13:12:49 Laibsch: depends how much storage will you get Jan 02 13:13:27 since kprintf("hello world\n"); drags in 200MB of kdelibs Jan 02 13:16:12 * mrz80 is away: Metaphysically I'm still here; physically, I'm elsewhere Jan 02 13:16:45 koen: at least with cmake, their system is more sane Jan 02 13:17:10 and I'm sure lorn will provide KDE4 integration with OpieII Jan 02 13:18:49 Well, I got a suggestion for using kiten for the vocab flash card at the gakusei.sf.net project. I rejected it with the argument that it would not run well if at all on an embedded device. I guess that holds true. Jan 02 13:21:16 Laibsch: with a day or two porting it could run (refering to pi-sync.net stuff...) Jan 02 13:22:35 ~emulate zautrix Jan 02 13:22:46 this is exactly what the GPL is about. grab the stuff and run away Jan 02 13:24:46 <[cc]smart> RP: linux compilationcompleted. thx alot. Jan 02 13:25:21 Well, I still rather focus on jards and granule which are close to completion. granule already compiles and upstream even bought a collie just for improving his code. He made a special build target so we can make some GUI changes for PDA. Those need a little bit of more work. jards was always targetted at the Z. It compiles. Upstream wants to continue development for Sharp ROM but perl currently breaks :-( Jan 02 13:34:33 cu Jan 02 13:34:50 ~seen Ifaistos Jan 02 13:34:53 ifaistos was last seen on IRC in channel #oe, 4d 20h 54m 21s ago, saying: 'koen happy birthday :)'. Jan 02 13:36:40 do I have to use a special flag during build, so that the runtime linker will find pthread lib @ exec. time? Jan 02 13:46:25 03koen 07org.oe.dev * rbfde8722... 10/ (4 files in 3 dirs): opentapi: add (untested) initscript Jan 02 13:48:28 * mrz80 is back (gone 00:32:15) Jan 02 14:03:49 how do I build a specific version of u-boot Jan 02 14:03:59 without setting up preferred version Jan 02 14:04:06 just testing 1.1.6 build? Jan 02 14:04:28 bitbake uboot- Jan 02 14:05:15 this dash/underscore stuff confuses me :) Jan 02 14:05:18 thanks Jan 02 14:14:34 u-boot 1.1.6 builds for OSK Jan 02 14:14:45 now building 2.6.19-omap1 Jan 02 14:32:38 Rebuilding glib-2.0 fails with "/usr/bin/autoreconf: unrecognized option `--exclude=autopoint'" here. What now? Jan 02 14:34:36 autoreconf is version 2.60. It belongs to package autoconf which is at version 2.60-1 on an up-to-date edgy. Jan 02 14:35:25 oe uses its own auto(re)conf Jan 02 14:35:46 so "bitbake -c rebuild autoconf"? Jan 02 14:35:58 OE should use its own autoreconf (/usr/bin/autoconf indicates something else) Jan 02 14:38:44 sounds like you're using bitbake -b too much. check whether STAGING_BINDIR/autoconf is present Jan 02 14:40:32 "At work, I finally got my workstation installer going. It's a netbootable Debian image that sets up the disk, unpacks a tarball of Fedora Core 6 (ironic, huh?) onto the machine, makes sure it will be bootable, and then writes some machine specific config. " Jan 02 14:40:34 heh Jan 02 14:40:44 (http://davyd.livejournal.com/204234.html) Jan 02 14:42:12 03awilcox 07org.oe.dev * rc2ec81fe... 10/ (1 packages/jhead/jhead_2.6.0.bb): jhead: add do_install Jan 02 14:45:32 03awilcox 07org.oe.dev * r1a9ae927... 10/ (1 packages/meta/openprotium-packages.bb): openprotium_packages: remove non-building libdvb Jan 02 14:49:23 HopsNBarley: hny andy Jan 02 14:49:41 hi leon, happy new year! Jan 02 14:50:33 koen: did I miss your birthday recently? if so, hiep hiep hoera! Jan 02 14:53:07 if I compile a module (.so) for tslib with OE devshell the module only works if I do not use pthread functions in it - otherwhise it won't load (no reason displayed why - nothing in dmesg & logread) Jan 02 14:54:02 if I open the .so in a hex editor I find references to the OE stage dirs /stuff .../ - do I have to do something else to get an .so which can be linked @ runtime? Jan 02 14:54:14 likewise: 28-12 Jan 02 14:54:24 HopsNBarley: HNY Jan 02 14:55:13 koen, thanks! HNY back to you! Jan 02 14:55:32 koen: congrats! Still in the twenties? Jan 02 14:55:45 likewise: yep, 24 :) Jan 02 14:55:46 pgfeller: if your username is != drepper use objdump or nm Jan 02 14:56:24 koen: sigh. I had to dial-in when I was 24 :-) Jan 02 14:58:03 zecke: I do not know mr. drepper :-) ... thanks for the hint - will make myself familiar with those two Jan 02 15:00:24 zecke: DSO only works on x86, since every other archs is either annoying or broken Jan 02 15:00:34 Jan 02 15:00:53 ~pastebin Jan 02 15:00:59 pastebin is, like, a place to paste your stuff without flooding the channel - try http://pastebin.ca, or http://channels.debian.net/paste Jan 02 15:02:49 zecke: which one to check? dynamic reloc looks reasonable to me (as far I can tell ...): http://pastebin.ca/301086 Jan 02 15:03:04 pgfeller: SO_NEEDED is always interesting Jan 02 15:03:19 pgfeller: and as you suspect threads, search for pthread? Jan 02 15:08:16 zecke: in "Dynamic Section:" I have libpthread.so.0, libgcc_s.so.1, libc.so.6 => http://pastebin.ca/301092 - all of them present on the device Jan 02 15:08:41 vnc crap :) Jan 02 15:09:46 pgfeller: man dlopen, man dlerror, grep dlopen | wc -l, grep dlerror | wc -l, add dlerror usage to tslib Jan 02 15:12:40 zecke: yep - vnc :-) ...now almost done -> uinput works now (keyboard & mouse) - so I only have to inject the events somewhere. I figured that tslib would the best (used by opie & gpe, device independant) Jan 02 15:12:57 zecke: thanks for the hints - will read trough them Jan 02 15:14:24 pgfeller: please remind me why Qts VNC wasn't good enough Jan 02 15:15:09 zecke: I did not manage to "share" the screen - but "only" to invoke a 2nd opie instance Jan 02 15:15:41 zecke: unfortunately to go into the code of qte is not practicable for me Jan 02 15:16:07 hehe and instead of writing a working QWS_DISPLAY variable you spend weeks/month on the other thing. Call me lazy but I would not have done this :) Jan 02 15:16:07 zecke: and I figured that a GUI independent VNC server would be a nice thing for linux handhelds Jan 02 15:17:02 zecke: tried several things with QWS_DISPLAY ... but never managed to have the same screen on the pda & on the vnc client :-/ Jan 02 15:17:02 pgfeller: nice initial thought, but either you use directFb/QtE or a network transparent protocol like X anyway :) Jan 02 15:18:19 zecke: I'm still glad for a working QWS_DISPLAY variable :-) Jan 02 15:18:51 pgfeller: I won't do any public QtE until TT behaves again Jan 02 15:20:01 zecke: :-) np ... just another one in favor of vnc: for qpe users they would not need an X server if they want to remote control from their atari, amiga, c64 ... or what ever - jsut a working java vm :-) Jan 02 15:20:14 zecke: not qpe - meant gpe Jan 02 15:20:35 pgfeller: X teleportation! Jan 02 15:21:25 zecke: ? Jan 02 15:22:02 pgfeller: with X/GPE you are able to teleport single applications to another display Jan 02 15:22:16 pgfeller: e.g. teleport the addressbook from your pda display to your PC Jan 02 15:22:27 pgfeller: and then copy and paste data between these two clients :) Jan 02 15:22:45 zecke: but you need an x-server on the other computer - right? Jan 02 15:25:57 pgfeller: yes Jan 02 15:26:21 zecke: & this way a learn a lot about programming & linux :-) ... but now I have to go; thanks again for your help - it is much apreciated Jan 02 15:26:55 mickeyl: I hardly ever use bitbake -b. I think I might be seeing the effect of http://permalink.gmane.org/gmane.comp.handhelds.openembedded/14355 Jan 02 15:28:24 oh right Jan 02 15:28:32 i rebuilt everything once i read that mail Jan 02 15:29:16 I will probably do the same now. It will take me about 24-48 hours before being able to do any real compilation again :-( Jan 02 15:34:39 I'm working through a procedure to install grub to a CF card for a OE build x86 distro. Currently I'm just running grub from the host system, but I suspect it is using some pieces from the host version of grub. Has anyone worked through this? Jan 02 15:44:27 why use grub? Jan 02 15:44:54 the problem with grub is that it requires the scripts be run with the actual physical media in hand Jan 02 15:45:01 that doesn't lend itself very well to a universal image Jan 02 15:45:03 CosmicPenguin: I personally believe bootloaders must have tab support! Jan 02 15:45:57 syslinux does not require the physical media on hand to be installed - though it does require a FAT filesystem Jan 02 15:46:12 cbrake: I have done GRUB installs to a CF image, off-target. Jan 02 15:46:15 which leads into the need for multiple partitions Jan 02 15:46:42 I've been meaning to port extlinux for what - three decades now, and I'm sure I'll get around to it before Halley's comet returns Jan 02 15:47:08 hehe Jan 02 15:52:38 cu Jan 02 15:55:31 * cbrake looks at syslinux ... Jan 02 16:03:54 * cbrake does not like the FAT partition requirement for syslinux Jan 02 16:36:02 * RP watches OABI packages segfault on an EABI system :} Jan 02 16:39:18 03rpurdie 07org.oe.dev * r1664f078... 10/ (1 packages/linux/linux-rp_2.6.19+git.bb): linux-rp-2.6.19+git: Update to ASoC 0.13pre1, fix compile failures and pronounce spitz working with this version Jan 02 16:48:47 u-boot 1.1.6 works on the OSK Jan 02 16:51:23 RP: If I want to do some testing on .dev imagages for poodle, should I try the 2.6.19 kernels or should I stick wthis 2.6.17 for now? Jan 02 16:52:40 hvontres|poodle: I'd be interested in reports on the 2.6.19+git kernel Jan 02 16:52:51 hvontres|poodle: Its what will become 2.6.20 Jan 02 16:53:07 2.6.19-omap1 oops loading ALSA for me :( Jan 02 16:54:50 RP: OK. I'll have to rebuild staging on my build box tonight so I might have some results later this week. Those two P-II's just aren't that fast :) Jan 02 16:55:37 RP: btw, 2.6.18 crashes on suspend/resume... so I guess good riddance to it :) Jan 02 16:57:16 hvontres|poodle: Hmm. It shouldn't crash :-/ Jan 02 16:57:55 hvontres|poodle: I'm planning to drop 2.6.18 anyway. If I fix anything it will be 2.6.19+git Jan 02 16:58:29 RP: I know. I just wanted to give you a +1 on dropping 2.6.18 :) Jan 02 16:58:38 hvontres|poodle: :) Jan 02 17:00:38 03rpurdie * r721 10bitbake/lib/bb/fetch/svn.py: svn fetcher: Don't have a date in the filename for specific svn revisions Jan 02 17:08:45 03rpurdie 07bitbake-1.6 * r722 10/lib/bb/data.py: bitbake-1.6 data.py: Backport emit_var fixes and add unexport flag support Jan 02 17:11:41 koen: There should be support for setting MACHINE in the environment for .dev now with both bitbake 1.6 svn and 1.7 Jan 02 17:11:57 nice Jan 02 17:12:11 did you merge the poky fixes? Jan 02 17:12:30 koen: Yes, the image one came in yesterday with the other changes Jan 02 17:13:09 The rest came in a while ago Jan 02 17:13:12 time to rewrite my buildscripts afain Jan 02 17:13:39 hi koen Jan 02 17:13:41 koen: Test it first to make sure I've not missed anything. If I have, we can grab the fixes from poky though Jan 02 17:14:11 hey ssvb Jan 02 17:14:54 grrr Jan 02 17:15:03 svn built by OE doesn't have http(s) support Jan 02 17:15:16 :-/ Jan 02 17:15:32 koen: can you do some quick test on xscale box? (if you are not too busy): http://www.oesf.org/forums/index.php?showtopic=22280&view=findpost&p=149865 Jan 02 17:15:49 ssvb: see my comment about svn :) Jan 02 17:15:56 checking out the tree on another box right now Jan 02 17:17:33 meh, i wish qemu had a hostfs Jan 02 17:17:54 kergoth: nfs is quite close Jan 02 17:19:25 not when you're on a tightly controlled machine where you lack the root access necessary to create a tun/tap device Jan 02 17:19:28 :P Jan 02 17:21:08 anyone know why u-boot had a patch for navman.. in the bb file? Jan 02 17:21:54 Crofton: because upstream lacks support for that machine Jan 02 17:22:13 I wonder if 1.1.6 supports it? Jan 02 17:22:30 I created a 1.1.6 bb file, but left that patch out Jan 02 17:24:11 try putting it in Jan 02 17:25:09 a pox on people who do not send patches upstream Jan 02 17:27:40 koen, do you know who can test the addition of the patch? Jan 02 17:28:02 I will add the line to the bb file Jan 02 17:28:08 if it applies it should be allright Jan 02 17:29:02 the SRC_URI for the patch looks like it only runs if you are building for that machine ... Jan 02 17:29:43 I did look in u-boot source and it does not look like the patch was applied upstream Jan 02 17:34:00 ssvb: what does the output of unquantize tell me? Jan 02 17:34:41 root@h2200:/data/site/maemo-mplayer/libavcodec/tests# ./test-unquantize Jan 02 17:34:41 dct_unquantize_h263_helper_c time=0.04372 usec per element, or 10.9 cycles (250MHz), 18.2 cycles (416MHz) Jan 02 17:34:41 dct_unquantize_h263_helper_armv5te time=0.01935 usec per element, or 4.8 cycles (250MHz), 8.1 cycles (416MHz) Jan 02 17:41:07 patch fails Jan 02 17:41:32 koen: thanks, that matches the results from arm926, what optimization options did you use, is it possible to tweak them to get better performance for C version? Jan 02 17:41:33 I'll push without the patch line and the navman people will have to fix ... Jan 02 17:45:09 root@hx4700:~$ ./test-unquantize Jan 02 17:45:09 dct_unquantize_h263_helper_c time=0.02691 usec per element, or 6.7 cycles (250MHz), 11.2 cycles (416MHz) Jan 02 17:45:10 dct_unquantize_h263_helper_armv5te time=0.01179 usec per element, or 2.9 cycles (250MHz), 4.9 cycles (416MHz) Jan 02 17:45:22 ssvb: -march=armv5te -mtune=xscale -fexpensive-optimizations -fomit-frame-pointer -frename-registers -Os Jan 02 17:51:06 mreimer: happy new year! Jan 02 17:51:16 koen|away: thanks! to you too! Jan 02 17:54:56 03crofton 07org.oe.dev * r35295539... 10/ (1 packages/uboot/u-boot_1.1.6.bb): u-boot_1.1.6.bb : Add new bb file to build u-boot-1.1.6. Jan 02 18:00:37 koen: one last question, what is the clock frequency on that device? is it higher than 416MHz? :) Jan 02 18:01:38 ssvb: h2200 is 400MHz, hx4700 is 624MHz Jan 02 18:02:03 the hx4700 is a pxa270 Jan 02 18:04:06 koen: ok, thanks Jan 02 18:05:05 i think i will add an option to automatically detect clock frequency for a less confusing report Jan 02 18:09:38 koen: just to clarify, was iwmmxt only introduced in pxa270? Jan 02 18:09:59 ssvb: yes Jan 02 18:11:39 I see, so mplayer still can be optimized (much?) better for older devices Jan 02 18:12:12 most older devices are pxa255 Jan 02 18:12:42 or arm920t based (armv4t) Jan 02 18:12:52 also it is interesting what chip will be used in Nokia 770 successor, if it will be ARM11, it should have some sort of SIMD instructions which are useful for multimedia Jan 02 18:13:07 ? Jan 02 18:13:12 arm11 doesn't add simd IIRC Jan 02 18:14:26 but it uses armv6 instruction set, and there seem to be some SIMD instructions according to arm docs Jan 02 18:15:23 arm docs are always a bit ehm.. "enthousiastic" Jan 02 18:15:58 http://arm.com/products/CPUs/ARM1176.html Jan 02 18:16:45 "SIMD (Single Instruction Multiple Data) media processing extensions deliver up to 2x performance for video processing" Jan 02 18:18:04 But OMAP2 chips also have some dedicated video processing hardware, so maybe software decoding with mplayer will not be required there :) Jan 02 18:18:20 the IVA? Jan 02 18:19:19 yes, I wonder what thing is that? Jan 02 18:20:13 is it flexible enough to accelerate any codecs, or it specialized only in a few of them? Jan 02 18:20:43 I suspect it's just a c54x dsp Jan 02 18:21:28 omap3 seems to have dsp + iva2 Jan 02 18:22:40 anyway, arm920t devices are much less lucky, they don't have fast multiply enhanced dsp instructions which were introduced in armv5te Jan 02 18:24:12 later guys Jan 02 18:24:13 I think the vfp in most arm11 is the best treat Jan 02 18:24:20 even if it's "optional" Jan 02 18:26:52 koen: it also adds unaligned memory access support in hardware (it is a big stopping factor when porting apps to arm) Jan 02 18:27:05 indeed Jan 02 18:57:46 ehm Jan 02 18:57:53 who just pushed 184 csets? Jan 02 18:58:33 sorry, 177 Jan 02 19:00:37 is adding a custom inittab in packages/sysvinit/sysvinit// the best way to get a getty on tty1? Jan 02 19:01:05 koen: is that why my pull has been taking forever ... Jan 02 19:01:30 it seems someone at nslu2-linux was a bit enthousiastic with merges Jan 02 19:02:24 rwhitby: if nslu2 automerges, could you turn that off? Jan 02 19:02:46 cbrake: yep, pulling in merges is quite slow Jan 02 19:05:11 koen: still can't mtn pull from oe.org ... Jan 02 19:05:50 try ewi546.ewi.utwente.nl or opensource.wolfsonmicro.com Jan 02 19:05:59 koen: ok Jan 02 19:19:36 hmm -- grub2: Cross-platform installation which allows for installing GRUB from a different architecture. Jan 02 19:19:43 have to check that out at some point ... Jan 02 19:22:28 koen: ok, I didnt ever code a .bbclass fix, I just deleted the junk by hand Jan 02 19:23:45 sed should be able to fix it, right? Jan 02 19:26:55 03koen 07org.oe.dev * rc6043a6b... 10/ (1 conf/distro/angstrom-2007.1-oabi.conf): angstrom oabi: COMPATIBLE_MACHINE is a bit unreliable in a distro.conf, so remove it Jan 02 19:27:10 indeed Jan 02 19:27:24 someone pushed 170+ merges Jan 02 19:27:30 koen: yeah, it If I remember right, you can in fact hard code the replacement with ${STAGING_LIBDIR} and thats all thats needed Jan 02 19:27:44 and the last merge is from nslu2-linux Jan 02 19:28:27 hey gerwinin Jan 02 19:29:15 hey koen Jan 02 19:29:34 I am sitting here with a tulband around my head :) Jan 02 19:29:52 koen: or ${CROSSDIR/blah/blah/lib} but I dont know which variable are used for that Jan 02 19:30:11 gerwinin: everything went ok? Jan 02 19:30:24 Koen: yes Jan 02 19:30:46 Koen : didn't have to stay in the hospital Jan 02 19:30:49 dependency_libs=' -Ltmp/cross/arm-angstrom-linux-gnueabi/lib/ -lgcc_s -lc -lgcc_s -lm -lgcc_s -lc -lgcc_s' Jan 02 19:30:55 koen: is I think all you need Jan 02 19:33:32 dependency_libs=' -Ltmp/cross/arm-angstrom-linux-gnueabi/lib/ -lgcc_s -lc -m', right? Jan 02 19:33:37 -lm Jan 02 19:34:09 koen: yeah, I wasnt sure if the repeats were to work out some wierd bugs in linker Jan 02 19:34:48 could you code up a sed rule that replace TMPDIR/work with crossdir? Jan 02 19:38:50 koen: is that ${CROSSDIR} ?? Jan 02 19:41:12 CROSS_DIR = "${TMPDIR}/cross" according to bitbake.conf Jan 02 19:44:02 sed -e "s|dependency_libs\s*=\s*.*|dependency_libs = ' -L${CROSS_DIR}/${HOST_SYS}/lib|" |" Jan 02 19:44:24 miss out that last |" cnp error Jan 02 19:46:12 s:-L${TMPDIR}/work*:-L${CROSS_DIR}/${HOST_SYS}/lib:g Jan 02 19:46:26 some .la files have more than one reference to tmp/work Jan 02 19:47:03 Ive compiled xcalibration tools but they dont contain the xcalibration tool, what package might I find that one in? Jan 02 19:47:21 sed -e "s|dependency_libs\s*=\s*.*|dependency_libs = ' -L${CROSS_DIR}/${HOST_SYS}/lib -lgcc_s -lc -lgcc_s -lm -lgcc_s -lc -lgcc_s '|" Jan 02 19:47:46 koen: that should replace the ones for gcc cleanly Jan 02 19:48:03 gcc isn't the only one infected :( Jan 02 19:48:07 koen: I hate the paths with ..\..\..\..\ and such Jan 02 19:48:29 koen: gcc is the one that starts the infection, fix that, clean /tmp and all else should be clean Jan 02 19:48:59 koen: as far as I can tell, libtool just concatenates all dependency_libs lines to make .la file for library currently being built Jan 02 19:49:17 that would explain it Jan 02 19:50:23 koen: so I suspect something along the lines of what I put when building gcc would probably clean it forever, I dont have the time right now to fiddle with the oeclasses Jan 02 19:50:47 could you reply to the list with your sed magic? Jan 02 19:56:40 koen: done Jan 02 19:57:54 thanks Jan 02 20:06:32 * XorA|gone returns to being gone for now Jan 02 20:20:32 koen: our monotone.nslu2-linux.org server does not automerge. our autobuilder automerges, but doesn't push. so it had to be a user (could be an nslu2-linux user) Jan 02 20:23:16 * mwester has been committing stuff via nslu2-linux -- but certainly didn't do any (manual) merges... unless the "make push" script automerged and did it... :( Jan 02 20:32:16 lpotter: was the "Qtopia 4.2.0" announcement on Dec. 14 really "Qtopia Platform 4.2.0"? Jan 02 20:49:39 hmm, why aren't I picking up the 2.6.19 omap1 kernel, after I drop the DEFAULT_Preference entry Jan 02 21:04:26 oops, found prefered version .... Jan 02 21:06:36 tn1: not. it is qtopi aphone without drm, sxe and telephony Jan 02 21:08:04 ljp: I see. The naming is a bit confusing. So will Qtopia Platform replace "Qtopia Phone-Lite" later? Jan 02 21:13:58 not sure what is going to replace the light sdk in 4.2.0 Jan 02 21:17:20 morning Jan 02 21:17:25 hey hrw Jan 02 21:17:27 morning hrw Jan 02 21:18:54 03koen 07org.oe.dev * r155f936f... 10/ (11 files in 10 dirs): Jan 02 21:18:54 basefiles: updates: Jan 02 21:18:54 * issue.* for angstrom Jan 02 21:18:54 * usbfs automounts for akita, htcuniversal, netbook-pro, nokia770, spitz, tosa Jan 02 21:18:54 Should fix #1741 Jan 02 21:20:32 time to pull 2 weeks.. Jan 02 21:20:59 but someone killed mtn server? Jan 02 21:22:11 hrw, I had to add another patch to 2.6.19-omap1 Jan 02 21:22:23 Crofton: I know - I read linux-omap Jan 02 21:22:25 uboot 1.1.6 is ok Jan 02 21:23:03 I'll push changes once the board boots on all new stuff Jan 02 21:23:04 Crofton: nice. can we drop all version other then 1.1.6/1.1.4? Jan 02 21:23:18 Crofton: 1.1.4 is preferred by amsdelta machine Jan 02 21:23:23 for OSK at least :) Jan 02 21:23:47 rwhitby: which users can commit with the nslu2-linux@nslu2-linux.org key? Jan 02 21:24:45 viewmtn upgraded to 0.32 as well Jan 02 21:26:45 ~seen tjabo Jan 02 21:27:09 ~seen tapo Jan 02 21:27:14 tjabo <~tapo@p50928FE7.dip0.t-ipconnect.de> was last seen on IRC in channel #handhelds.org, 835d 23h 36m 16s ago, saying: 'hi flo'. Jan 02 21:27:20 tapo was last seen on IRC in channel #kde, 40d 14h 11m 39s ago, saying: 'oh, another question: can I deactivate spell-checking in konquerors textareas? I only see setup dialogs, but no checkbox to just deactivate it.'. Jan 02 21:27:47 root@omap5912osk:~$ uname -a Jan 02 21:27:47 Linux omap5912osk 2.6.19-omap1 #1 Tue Jan 2 15:34:15 EST 2007 armv5tejl unknown Jan 02 21:27:50 :) Jan 02 21:28:07 Crofton: uboot 1.1.6 + fresh kernel? Jan 02 21:28:11 yep Jan 02 21:28:16 Crofton: cool Jan 02 21:29:24 and new bootstrap-image Jan 02 21:36:56 cbrake: no need for custom inittab. please use SERIAL_CONSOLE (or so, please check MACHINE confs) Jan 02 21:40:05 hrw, changes pushed Jan 02 21:40:18 I do not have the nohelloworld patch in uboot Jan 02 21:40:30 03crofton 07org.oe.dev * r61f078c7... 10/ (3 files in 3 dirs): linux-omap1-2.6.19-omap1 : Add patch to fix kernel oops starting ALSA. Jan 02 21:40:33 can you check that you can build u-boot without it? Jan 02 21:40:37 03crofton 07org.oe.dev * rc8313ad7... 10/ (1 conf/distro/openomap.conf): openomap.conf : Select kernel version automatically. Jan 02 21:40:42 03crofton 07org.oe.dev * rf3d1266d... 10/ (1 conf/machine/omap5912osk.conf): omap5912osk : Select u-boot version automatically. Jan 02 21:40:59 Crofton: can Jan 02 21:41:31 koen: can you next time split your commit into two? Jan 02 21:42:14 Crofton: first have to pull all those merges Jan 02 21:43:25 ~lart people that use nslu2-linux@nslu2-linux.org to commit 170 bogus merges Jan 02 21:43:25 * ibot judo chops people that use nslu2-linux@nslu2-linux.org to commit 170 bogus merges Jan 02 21:43:55 koen: the way monotone works makes that happen Jan 02 21:44:02 koen,hrw: I have an interesting oe question: It seems that the future for flash-poor devices is running them with root on SD. I have also figured out that kexec can be used as a "bootloader". What would be the best way to set up OE to compile two versions of the kernel with different commandline parameters at the same time? Jan 02 21:44:24 XorA|gone: no, the way there admin^^H^H^H scripts works makes that happen Jan 02 21:44:33 hvontres|poodle: separate recipe for smaller kernel? Jan 02 21:44:41 mickey|tv: ahh, should have looked at the bb closer -- thanks! Jan 02 21:45:09 XorA|gone: 'mtn merge' requires to be typed in Jan 02 21:45:22 it doesn't magically happen by staring at the screen Jan 02 21:45:49 koen: I have ended up in the same situation when I had committed some stuff locally I wasnt ready to push, everytime you pull you have to merge. When you eventually are ready to push you have hundreds of the bloody merges queued Jan 02 21:45:53 hrw: hmm...that could work too. I was hoping to help CoreDump to come up with some sort of "rescue partitition" for flashing and the full image for SD Jan 02 21:46:12 koen: monotone should detect if they are truly usless merges and drop them Jan 02 21:47:06 XorA|gone: I think that such offline commits,pull,merge,pull,merge,commit,pull,merge,push is normal thing and thats what for DSCM exists Jan 02 21:47:42 hrw: then why are people complaining about the correct operation Jan 02 21:47:58 XorA|gone: people or just koen? Jan 02 21:48:03 XorA|gone: if a *person* committed 170 merges + cset it would be ok Jan 02 21:49:15 XorA|gone: this is a script that does 'pull ; merge', but not push, and someone did 'push' Jan 02 21:49:26 or the script got changed Jan 02 21:49:40 re Jan 02 21:50:08 that surprised me the other day after a commit - I thought I had really flubbed it. At the least, monotone should divide up the commit message a bit. Jan 02 21:51:04 anyway, I need go pay attention to better half :-) Jan 02 21:51:52 hvontres|poodle: have a look at kboot for the kexec bootloader thingy Jan 02 21:53:47 giel: actually, I got the "bootloader" working inside of altboot right now. I was just wondering what the best way would be to crate the two kernels in OE (one for flash and one for the rootfs) Jan 02 21:54:17 hvontres|poodle: okay... it's just that kboot is exactly that: using kexec as bootloader ;) Jan 02 21:55:19 giel: well the simplest form of that would be kecex -l $KERNEL_YOU_WANT_TO_BOOT;kecec -e :) Jan 02 21:56:27 cu guys Jan 02 21:56:54 Crofton: I will check uboot/kernel tomorrow - going to sleep Jan 02 21:57:15 gn Jan 02 21:57:25 Does anybody know where CoreDump|home went? he hasn't been around in ~2 weeks Jan 02 21:58:36 hvontres|poodle: actually, we are working on getting something usuable in ~5MB flash with angstrom Jan 02 22:04:03 koen: ok. But I was also thinking that this could be a good way to deal with braindead boot-loaders. The flashed kernel would only need upgrading if something changed drasticly and the one on SD could even be updated by ipkg or it's replacement Jan 02 22:04:27 hvontres|poodle: we have something like this on h2200 and h4000 Jan 02 22:04:36 LAB - linux as bootloader Jan 02 22:04:57 lab scans the media (flash, sd, cf) for kernels and labrun files Jan 02 22:05:21 RP: what does http://svn.o-hand.com/view/poky/trunk/meta/packages/xtscal/xtscal/change-cross.patch?rev=1111&view=markup substantially do? Jan 02 22:30:30 03koen 07org.oe.dev * r4a394f77... 10/ (7 files in 5 dirs): basefiles: add fstab for efika and ep93xx Jan 02 22:46:42 can we make efika images with oe yet>? Jan 02 22:47:19 koen|away: I'd like to talk with you about what things to seperate into different features in task-base in Angstrom. Jan 02 23:00:38 heeelp (anyone knowing python autoconf stuff). AM_PATH_PYTHON puts in the native staging include dir, not the target one. Is that a known (possibly already fixed?) problem? Jan 02 23:38:33 koen|away: Honestly, I don't know. I'm on holiday atm. I'm hoping it fixes the screen corruption I've seen in xtscal before Jan 02 23:42:07 Is it a valid assumption that "bitbake world" should work for distro=angstrom, machine=ixp4xxle ? Jan 02 23:42:35 (DISTRO="angstrom-2007.1" actually) Jan 02 23:43:29 rwhitby: bitbake world should always work. Sadly, we don't have many people testing it and a lot of bugs in various dusty corners of the tree :-/ Jan 02 23:43:57 ok, I'll make that the default target for angstrom builds on our nslu2-linux autobuilder then. Jan 02 23:44:14 rwhitby: It only makes sense with the continue upon error mode of bitbake Jan 02 23:44:22 heh - when that works ;-) Jan 02 23:44:39 rwhitby: :-/ Jan 03 00:04:37 good night Jan 03 00:06:29 'night florian Jan 03 00:07:40 are there instructions floating around for how to use multimachine? Jan 03 00:08:06 mreimer: Just inherit it and change MACHINE Jan 03 00:08:17 wow, that's easy Jan 03 00:08:17 thanks RP Jan 03 00:10:09 is PARALLEL_MAKE safe to use? Jan 03 00:10:21 mreimer: Basically, yes Jan 03 00:10:34 ok Jan 03 00:32:27 anyone who knows gdb well here? Jan 03 00:32:37 I want to get a backtrace for a problem... Jan 03 00:33:10 the problem happens when I send a SIGINT (Ctrl-c) to a program...but gdb catches it.... Jan 03 00:36:23 JustinP: handle SIGINT pass Jan 03 00:36:41 maybe... :) Jan 03 00:39:56 RP: thanks, that did it :-) Jan 03 00:45:04 did ASoC get into mainline? Jan 03 00:45:19 mreimer: Its on its way Jan 03 00:45:23 great! Jan 03 00:50:22 i have a question on kbdd - i have some breaindead keyboard here i think i finally figured out how the protocoll works Jan 03 00:50:34 who is the maintainer for that? **** ENDING LOGGING AT Wed Jan 03 02:59:58 2007