**** BEGIN LOGGING AT Sun Nov 18 02:59:56 2007 Nov 18 03:36:04 * * OE Bug 3332 has been created by cwixon(AT)usa.net Nov 18 03:36:06 * * Abiword 2.5.x: workaround for UI issues on small screens Nov 18 03:36:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3332 Nov 18 03:41:01 morning Nov 18 05:27:32 ~seen likewise Nov 18 05:27:42 likewise was last seen on IRC in channel #oe, 8h 9m 3s ago, saying: 'Jin^eLD: that way, we can work rather independently (because we are too busy to work in tandem).'. Nov 18 07:00:53 03mwester 07org.oe.dev * r10a42628... 10/ (3 files in 2 dirs): Nov 18 07:00:53 gtk+: undo erroneous changes to gtk+ that were probably inadvertantly committed along with squashfs changes. Nov 18 07:00:53 Bump PR to force the missing packages to be rebuilt. Nov 18 08:09:23 asking somthing obvious, which parameter stands for "arm-angstrom-linux-gnueabi"? Nov 18 10:01:27 03pH5 07org.oe.dev * rd99aa6ab... 10/ (1 conf/distro/angstrom-2008.1.conf): angstrom-2008.1.conf: sync with 2007.1, avahi and openssh-sftp-server in RRECOMMENDS Nov 18 10:01:32 03pH5 07org.oe.dev * r1d6353fe... 10/ (1 conf/distro/angstrom-2008.1.conf): angstrom-2008.1.conf: sync with 2007.1, use linux-handhelds-2.6 version 2.6.21-hh18 Nov 18 10:01:38 03pH5 07org.oe.dev * r82cf726f... 10/ (1 conf/distro/angstrom-2008.1.conf): angstrom-2008.1.conf: use gcc 4.2.2 Nov 18 10:01:43 03pH5 07org.oe.dev * r18ea0ed5... 10/ (1 conf/distro/angstrom-2007.1.conf): angstrom-2007.1.conf: remove xrandr preferred versions, they are defined in angstrom-2007-preferred-versions.inc Nov 18 10:01:47 03rwhitby 07org.oe.dev * r5218b291... 10/ (1 conf/distro/include/sane-srcrevs.inc): sane-srcrevs: Bump libgsmd to match mickey's changes to gsmd.inc - tested by CM Nov 18 11:01:30 03koen 07org.oe.dev * r8e25ce12... 10/ (1 packages/gsm/gsmd.inc): gsmd: repair breakage for people not using bleeding edge svn versions Nov 18 12:22:11 gm Nov 18 12:34:03 re Nov 18 12:34:37 guys, any idea what machine conf is for a cpu: e300, chip: MPC83xx? Nov 18 12:34:52 I need to build something for that hw so trying to figure out what machine conf to base on Nov 18 12:37:56 koen@bitbake:~/OE/monotone/org.openembedded.dev/packages$ ls ../conf/machine/ | grep mpc Nov 18 12:37:57 mpc8313e-rdb.conf Nov 18 12:37:57 mpc8323e-rdb.conf Nov 18 12:38:11 Jin^eLD: you could try one of those two Nov 18 12:38:22 I wonder why I missed them yesterday, I must have been blind Nov 18 12:38:23 Jin^eLD: jeremy_laine, steliosk and likewise will know more about those Nov 18 12:38:28 but I guess I was too obsessed grepping for ppc Nov 18 12:38:48 thanks Nov 18 12:39:53 Jin^eLD: Me and Jeremy added the MPC83xx processor and e300c2 and e300c3 core lately. Steliosk and another developer (?) have add bits'n'pieces lately. Nov 18 12:40:17 likewise: how to figure out if its a c3 or c2? Nov 18 12:40:25 and whats the difference there? Nov 18 12:40:33 Jin^eLD: which processor do you have? c3 = w/ FPU, c2 = w/o FPU Nov 18 12:41:05 likewise: on idea... there is a NAS box from Promise and I am trying to port my upnp server for their firmware (i.e. matching their glibc version) so users could use it as an addon Nov 18 12:41:25 http://www.nslu2-info.de/showthread.php?t=5609 Nov 18 12:41:36 if you scroll down a little, there is a cat /proc/cpu output and some other things Nov 18 12:41:50 no idea about the FPU though Nov 18 12:42:14 is there a way to figure out? the user whom I am helping has shell access Nov 18 12:42:23 Jin^eLD: MPC8343 Nov 18 12:44:09 Jin^eLD: seems to be "e300" without c2 or c3. I'm looking up the diff for that. Nov 18 12:45:01 thanks Nov 18 12:46:11 Jin^eLD: http://www.freescale.com/files/32bit/doc/ref_manual/e300coreRM.pdf page 1-36 Nov 18 12:49:31 so c1 then? Nov 18 12:49:39 Jin^eLD: Start with the 603e core, it is the predecessor to the e300(c1). Should work. c2 and c3 do not apply. Nov 18 12:49:49 i.e. copy or use machine/storcenter.conf Nov 18 12:50:20 ok, thanks Nov 18 12:50:24 will try that Nov 18 12:52:39 Just found there is already a e300c4... Nov 18 12:53:25 where do the built images end up after a bake? Nov 18 12:57:52 Stephmw: depends on your configuration, and then in the deploy directory Nov 18 12:58:42 Jin^eLD: cheers :) Nov 18 12:59:20 np Nov 18 13:02:46 koen: I use the following patch to make gcc-cross-initial 4.2.x compilation work http://en.pastebin.ca/779760. Is there a better way to avoid the conflicting psignal definitions between glibc and libiberty? Nov 18 13:08:03 pH5: did you rebuild from scratch? http://archives.devshed.com/forums/development-94/gcc-4-2-branch-tiny-issue-with-libiberty-2385157.html Nov 18 13:08:59 pH5: no idea, try catching khem, he's a libc mastah Nov 18 13:11:28 koen: 'k Nov 18 13:12:13 likewise: I think I also had this problem when building from scratch. today I just did a rebuild of gcc-cross-initial because somebody pointed me to this issue again after I upped the gcc version in angstrom-2008.1.conf Nov 18 14:44:23 03mickeyl 07org.oe.dev * re563507d... 10/ (3 files in 3 dirs): Nov 18 14:44:23 openmoko-sound-system2: Nov 18 14:44:23 * supply --no-cpu-limit to pulseaudio to prevent shutting down on overload Nov 18 14:44:23 (which is more likely to happen on embedded systems). closes OM #1010 - thanks Robert Schuster. Nov 18 14:44:23 * rework start script to use startstopdaemon. closes OM #785 - thanks Kristian Mueller. Nov 18 15:56:33 hmm, I am trying to build an app using the nylon distro (it matches the glibc version that I need), and it downloads glibc and all that but for some reason it then also tries to get eglibc Nov 18 15:56:39 any idea what makes it do so? Nov 18 15:56:47 there is no mention of eglibc in the whole nylon conf Nov 18 16:01:47 any ideas? I'm trying to figure out how to not make it use this eglibc thing Nov 18 16:02:07 you need to look through all the includes, machine config etc... Nov 18 16:02:32 aah I probably missed the PREFERRED_PROVIDER thing Nov 18 16:05:27 yeah, seems like that was it Nov 18 16:29:01 koen, ping? Nov 18 16:30:47 did the autbuilder create a u-boot for the OSK? Nov 18 16:31:04 I suspect the stock u-boot may not have the correct machine ID ... Nov 18 16:35:28 woah, you have an autobuilder now? Nov 18 16:37:51 koen does builds of a number of machines Nov 18 16:45:35 can anyone help me with bitbake collections? Nov 18 16:45:52 CoreDump, the docs are pretty good? Nov 18 16:45:57 what is the problem? Nov 18 16:46:14 http://www.openembedded.org/user-manual&dpage=special_bitbake_collections <- the docs suck Nov 18 16:46:15 is ${TARGET_SYS} the right variable for arm-angstrom-linux-gnueabi? Nov 18 16:46:39 BB does not honor the configured priorities and I do not know why Nov 18 16:46:40 I define the docs as goos since I got it working without to much trouble .... Nov 18 16:47:03 I think you need to update BBPATH also Nov 18 16:47:21 If I have two identical .bb's, one in .dev, one in the overlay, the one with the higher priority should be selected. It isn't here :\ Nov 18 16:47:38 Crofton|home: I did that already Nov 18 16:47:51 hmmm Nov 18 16:48:02 I do not have that setup, but I have heard of it working Nov 18 16:48:22 I had it working in the past Nov 18 16:48:41 ah BBFILES includes collection? Nov 18 16:48:50 yes Nov 18 16:50:59 http://rafb.net/p/AHUv1m77.html Nov 18 16:51:49 pH5: http://www.angstrom-distribution.org/unstable/autobuild/ Nov 18 16:52:33 koen, the OSK build should include u-boot Nov 18 16:52:42 Crofton: I see no difference to my setup. The basic overlay works, only the priorities are not honored Nov 18 16:52:51 Crofton|home: it uploads everything that gets built Nov 18 16:52:57 koen, ok Nov 18 16:53:00 I'll check Nov 18 16:53:13 maybe I dropped u-boot at some point when I was have git fetcher issues Nov 18 16:53:27 although the OSK only needs u-boot from tar ball Nov 18 16:55:20 mickey_away: -resample-method=trivial is already in /etc/pulse/daemon.conf Nov 18 16:56:31 koen, how many sets of build results will you keep in unstable? Nov 18 16:56:53 Crofton|home: that's up to the mentors Nov 18 16:57:01 they can delete stuff they don't want/need Nov 18 16:57:05 ok Nov 18 17:00:07 brb Nov 18 17:00:35 CoreDump, I need to start playing with the stuff from gumstix, which uses overlays Nov 18 17:00:51 I'll see if I can confirm the behavior you are seeing when I do ... Nov 18 17:01:47 I had it working in the past with a different setup so I _know_ it works. BB just makes it utterly fucking impossible to debug the problem w/o hacking python Nov 18 17:02:46 hi all! Nov 18 17:04:34 lumag: hey Nov 18 17:08:58 Crofton|home: could you ping rishi again about sending us the omap24xx OSKs? Nov 18 17:09:09 I can try Nov 18 17:09:24 Have you had any contact with him? Nov 18 17:09:50 * Crofton|home wishes TI would get their act together Nov 18 17:09:54 nope Nov 18 17:10:25 they must exist because khem added support Nov 18 17:10:50 there is a TI guy claiming some new more affordable dev boards are in the pipeline Nov 18 17:11:49 yeah, for the dv iirc Nov 18 17:11:58 OMAP line also Nov 18 17:12:43 http://linux.omap.com/pipermail/linux-omap-open-source/2007-November/012046.html Nov 18 17:19:02 sakoman_, in your gumstix stuff, where do you set TMPDIR? Nov 18 17:19:54 koen: are there any simple way to duplicate packages, e.g. libfftw and libfftw3 Nov 18 17:20:22 xjqian: what do you mean exactly? Nov 18 17:22:04 * chouimat just bought his niece xmas gift ... a nice box of 832 crayons ;) Nov 18 17:22:10 koen: PACKAGES =+ "lib${PN} lIb{PN}3" Nov 18 17:22:21 xjqian: fftw is already in OE Nov 18 17:22:46 koen, while they should have exactly the same files, etc Nov 18 17:24:25 koen: nevermind Nov 18 17:26:22 Crofton: at the moment I don't -- I let it use the default Nov 18 17:27:36 Crofton: also if you update today you'll find I set up auto.conf, and site.conf as well as update the extras/profile stuff to match those changes Nov 18 17:28:20 what is the default? At one time it was the workign dir Nov 18 17:28:39 which meant I started a new tmp every I ran bb from a different place Nov 18 17:28:52 Crofton: I have found my problem. Someone thought it funny to set a PREFERRED_VERSION for a package that only ever had a single version (and will probably never ever get a version bump besides PR). Apparently PREFERRED_VERSION has a higher priority than configured overlay priorities.... Nov 18 17:29:04 ah Nov 18 17:29:04 Crofton: yes, that is the default Nov 18 17:29:19 sakoman_, that will drive people crazy :) Nov 18 17:29:31 Crofton: suggestion? Nov 18 17:29:41 I use ~/oe/tmp Nov 18 17:30:33 Crofton: OK, I'll pick something suitable for the default gumstix directory setup Nov 18 17:30:54 that should help newbies Nov 18 17:31:09 Crofton: probably ~/gumstix/gumstix-oe/tmp Nov 18 17:31:23 yeah, that is what I would pick Nov 18 17:31:54 * Crofton|home spent a few weeks wondering why the toolchain randomly rebuilt .... Nov 18 17:33:37 Crofton: good suggestion, I'll check that in later today after a quick test Nov 18 17:38:57 koen: while building octave, I have to symbolic link gfortran to g77. I'm not sure why gfortran was not recognized as the default fortran compiler. gnu-configize fault? Nov 18 17:39:22 aren't g77 and gfortran differrent things? Nov 18 17:39:54 koen, gfortran is fully backward compatible to g77 Nov 18 17:40:10 ah Nov 18 17:40:27 so it would be safe to teach gcc-cross.inc to symlink it Nov 18 17:41:13 koen: yes. I would recommend that instead of finding out what's wrong with gnu-configize or autotools Nov 18 17:41:48 koen: it's 100% safe and the practical thing to do Nov 18 17:42:56 great Nov 18 17:49:27 RP, mickey_away: http://qemu-arm-eabi.wiki.sourceforge.net/ Nov 18 17:54:48 hrw|gone, I've updated the unified updater.sh in the bug#1526 Nov 18 19:35:21 NOTE: package octave-2.9.17: completed.:) please let me know your comments for octave.inc @ http://pastebin.ca/781085. Changes in fftw are needed. Recommend to change package name from fftw to fftw3 to comform to debian naming. fftw patches will follow. Nov 18 20:18:45 crap, I remember I had such an error half a year ago and now I forgot how I solved it mhm, any hints :) build/promise/cross/lib/gcc-lib/powerpc-linux/3.3.4/../../../../powerpc-linux/bin/ld: cannot find -lgcc_eh Nov 18 20:18:50 where is this gcc_eh supposed to be coming from? Nov 18 20:18:57 I am trying to build glibc 2.3.2 Nov 18 20:20:48 Jin^eLD: hmm, dunno. Nov 18 20:21:21 I wish I had written it down somewhere Nov 18 20:27:44 isn't gcc_eh for static linking? Nov 18 20:27:59 I have no idea Nov 18 20:28:45 well, I removed it manually from the make script Nov 18 20:28:49 seems to continue building Nov 18 20:29:25 not sure of the implications, I actually do not even need to run it, I only try to match the versions so I can compile a couple of packages as addon to some existing firmware Nov 18 20:30:50 the manual hack did the trick, still odd it did not work out of the box Nov 18 20:51:19 btw when I bitbake some app, all dependencies are compiled and in the end I get the ipk for that app, but not the ipk's for all the deps, they seem to be only staged Nov 18 20:51:30 can I somehow say that I want all the deps of that app packaged too? Nov 18 20:58:13 Jin^eLD: hey, hmmm I'm still out of the loop but that is some kind of unexpected behaviour Nov 18 20:58:44 Jin^eLD: short term cheating would be building a image with your app... then package_write gets executed as well but RP knows better :} Nov 18 20:59:23 well yes, I kind of wanted to avoid the image thing.. I did it by hand then, were only couple of depedencies Nov 18 20:59:31 and indeed I think it worked before Nov 18 21:22:32 What does a single : mean inside a do_install() body? Nov 18 21:22:54 I think it lets the method parse with no conetent Nov 18 21:23:08 to prevent the default code from executing ? Nov 18 21:29:52 Jin^eLD, zecke: It depends what you define the correct behaviour as :} Nov 18 21:30:01 :) Nov 18 21:30:14 well, I was more asking for a way to do such a thing without having to create an image Nov 18 21:30:36 RP: hehe, I used the word unexpected :) Nov 18 21:31:01 or is "bt" actually a manufacturer Nov 18 21:31:03 RP: I'm pretty sure that the package_write (or how that is called) is not in the depends tree so bitbake is not doing anything wrong Nov 18 21:31:04 oops wrong window :) Nov 18 21:31:32 zecke: Its not unexpected to me ;-) Nov 18 21:32:00 f00d time :> brb Nov 18 21:32:06 zecke: Our dependencies used to be such it happened to package all dependencies ;-) Nov 18 21:32:07 Crofton: So it's the empty statement? Nov 18 21:32:12 Jin^eLD: bon appetit Nov 18 21:32:15 pretty sure Nov 18 21:32:56 thx :) Nov 18 21:33:40 Crofton: Its just emptiness Nov 18 21:34:04 * * OE Bug 3234 has been REOPENED by joshua(AT)willowisp.net Nov 18 21:34:06 * *  bluetooth not working under h2200 with CSR radio (bcsp) Nov 18 21:34:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3234 Nov 18 21:42:31 koen|away, ping Nov 18 21:57:30 hi all Nov 18 21:57:47 hi flo_lap Nov 18 21:57:59 hey Nov 18 21:58:30 bye all! Nov 18 22:11:52 zecke, Jin^eLD: http://www.rpsys.net/openzaurus/temp/buildall.patch Nov 18 22:25:46 RP: how would one use it? bitbake foo -cbuildall? Or do we inject buildall into every of our (R)DEPENDS? (I think we inject... wow I need to work more on bitbake) Nov 18 22:32:04 zecke: We could change BB_DEFAULT_TASK or people call -c buildall when the mean buildall... Nov 18 22:33:37 -cbuildall is okay Nov 18 22:33:39 :) Nov 18 22:33:56 * zecke is watching http://video.google.com/videoplay?docid=-8002801113289007228&q=user%3A%22Google+engEDU%22&total=572&start=60&num=10&so=1&type=search&plindex=7 Nov 18 22:34:01 pretty bad speaker though Nov 18 22:35:25 03likewise 07org.oe.dev * r2a0cadd6... 10/ (5 files in 4 dirs): pango: Added 1.18.3 with DEFAULT_PREFERENCE -1. Nov 18 22:37:34 zecke: I got a patch from Matthew/Tim implementing the md5sum business btw. I can't see what it fixes :/ Nov 18 22:38:03 hehe Nov 18 22:38:18 we can add it in a dead codepath :) Nov 18 22:39:06 zecke: ;-) Nov 18 22:42:42 * RP ponders bug 3321 and whether we need to improve bitbake or not... Nov 18 22:52:50 RP: can you ponder on #3177 instead? Nov 18 23:02:38 koen, did my key make it to the Angstrom list Nov 18 23:03:44 it did Nov 18 23:09:03 I'm trying to understand why "non-dev package contains symbolic link" is a QA issue Nov 18 23:09:20 xjqian: Specifically the .so symlink Nov 18 23:09:31 RP: yes Nov 18 23:09:47 xjqian: You only need the for development purposes... Nov 18 23:10:49 RP: seems I lack some basic development knowledge. I will read a bit more. Thanks Nov 18 23:11:01 Is bugs.openembedded.org working for anyone else? Nov 18 23:12:36 koen: Which one was that since I can't access the bugzilla? Nov 18 23:12:55 that sepukku can't attach logs anymore due to type mismatches Nov 18 23:13:29 rendering the autobuild infrastructure pretty useless Nov 18 23:20:41 koen: There is an error somewhere within seppuku_eventhandler() in the .bbclass file, probably towards the end. I suggest commenting out lines until it works and find the faulty line Nov 18 23:21:16 sorry need to sleep Nov 18 23:21:28 'night zecke Nov 18 23:21:29 RP: ok, thanks for looking Nov 18 23:21:40 'night zecke Nov 18 23:21:59 koen: I'd bet on the last line, the #% might be meant to be #%s Nov 18 23:22:27 and the #% a few lines further up... Nov 18 23:23:58 RP: I get the feeling seppuku_create_attachment isn't working Nov 18 23:24:33 of course I can't check now, since bugzilla is down .... Nov 18 23:25:48 koen: I think two two typos I mentioned might do it... Nov 18 23:26:14 Alternatively its some python syntax I've not seen before :} Nov 18 23:26:31 the two typos are after the thing it fails to do, namely attach the logs Nov 18 23:26:50 but, we'll see :) Nov 18 23:27:00 bitbake is magic enough for such things to work Nov 18 23:27:55 It could the attachement code is failing too. At least this way you'll get a nice message :) Nov 18 23:31:46 RP: I just thought of a nice gimmick for poky: have ross create ICC profiles for the displayes Nov 18 23:32:05 RP: argyll can't be crosscompiled, but that's what remote X is for :) Nov 18 23:34:15 koen: Its an interesting idea :) Nov 18 23:35:11 The local photo club also has a spyder, but we seem to have misplaced it Nov 19 00:15:33 which package contains an uncrippled version of mktemp? (can't figure out how to get busybox version to work) **** ENDING LOGGING AT Mon Nov 19 02:59:56 2007