**** BEGIN LOGGING AT Sat Oct 11 02:59:57 2008 Oct 11 06:48:10 <[X]Spot> Hi guys Oct 11 07:20:10 03  07org.openembedded.dev * r471b152a57 10org.openembedded.dev.git/ (3 files in 2 dirs): python 2.5 add python-bz2 and python-elementtree (from hrw), add symbol.py to python-lang Oct 11 07:20:16 03  07org.openembedded.dev * r4c177da7c1 10org.openembedded.dev.git/conf/distro/include/sane-srcrevs.inc: snae-srcrevs.inc: bump openmoko kernels Oct 11 08:06:50 <[X]Spot> where can I find some docs about configuring, building and installing packages with OE (for angstrom, beaglebord) Oct 11 09:44:14 morning all Oct 11 10:11:52 morning Oct 11 11:15:47 hello; anyone uses OE to build uClibc images for powerpc boards? Oct 11 11:16:27 I'm trying to build one for MPC8315, but it looks like uClibc gets misconfigured, as .config contains TARGET_alpha=y Oct 11 11:26:53 i build for ublibc/powerpc Oct 11 11:26:58 oopc uclibc too Oct 11 11:27:56 hmm let me look in oe to make sure that bit is the same as the foonas repo still.... Oct 11 11:28:45 mmp look in packages/uclibc/uclibc-0.9.29/ - all the machines have a config, you need to make one Oct 11 11:30:05 i guess symlink to efika will do Oct 11 11:34:00 timtimred: hmm, it's quite possible, I was just looking that the uClibc.machine file is empty Oct 11 11:35:55 looks like there already are files for other boards.. Oct 11 11:37:52 Hey Oct 11 11:38:29 I just tried to find out where the kernel is included in the config files but I was unable to find it. Does MACHINE_TASK_PROVIDER play a role? Oct 11 11:46:49 ah, got it :) Oct 11 12:11:40 mickey|tw|zzZzz, ping Oct 11 12:11:55 anyone familiar with packaging python handy? Oct 11 12:14:45 what's up Oct 11 12:14:50 fwiw, i added symbol to python-lang Oct 11 12:18:15 heh Oct 11 12:18:17 thanks :) Oct 11 12:18:22 pyhton-lang Oct 11 12:18:30 awesome Oct 11 12:18:41 if you need to add more, add them in generate-manifest-2.5.py Oct 11 12:18:49 then call it to regenerate the manifest file Oct 11 12:18:51 I tried and failed :) Oct 11 12:18:53 heh Oct 11 12:18:56 ah Oct 11 12:19:00 I added to inc Oct 11 12:19:14 ok, time to test Oct 11 12:19:20 ok Oct 11 12:19:24 if you need more, ping me Oct 11 12:19:29 this stupid gnuradio stuff is importing everything ..... Oct 11 12:19:32 i'm boarding the plane to germany soon tho Oct 11 12:19:42 will be back tomorrow Oct 11 12:19:43 ok Oct 11 12:20:09 did you have a good trip? Oct 11 12:20:14 absolutely Oct 11 12:20:17 was amazing this year Oct 11 12:20:24 great Oct 11 12:20:25 got tons of stuff done Oct 11 12:20:30 and also seen a couple of sites Oct 11 12:20:36 three weeks passed like a breeze Oct 11 12:20:39 even better :) Oct 11 12:20:41 yep Oct 11 12:20:58 we need to take care about next OEDEM soon Oct 11 12:21:02 even though we can do a lot of work via email/irc Oct 11 12:21:10 it is a good thing to get together in person Oct 11 12:21:13 right Oct 11 12:21:25 much more efficient if you have to deal w/ complex things Oct 11 12:21:36 like, cross layer power management or stuff Oct 11 12:21:46 also helps with festering people issues :) Oct 11 12:22:10 even if you got no work done, I think productivity would increase in the weeks following Oct 11 12:22:36 correct Oct 11 12:22:52 urg Oct 11 12:23:00 updated the wrong machine Oct 11 12:23:10 maybe that is why my efforts yesterday failed :) Oct 11 12:23:48 no, did both Oct 11 12:24:46 so installing python-lang will make "import symbol" work? Oct 11 12:25:22 latest build, ya Oct 11 12:25:26 added that a couple of hours ago Oct 11 12:25:26 good Oct 11 12:25:34 I'll fire off a build and get coffee Oct 11 12:25:38 good speed Oct 11 12:25:41 thanks Oct 11 12:25:43 packing here Oct 11 12:25:44 have a good trip Oct 11 12:25:45 cu Oct 11 12:25:46 thanks Oct 11 12:25:48 l8r Oct 11 12:55:58 Do we have any plans for when or where OEDEM might be? Oct 11 12:56:20 I haven't heard Oct 11 12:56:49 would your new corporate overlords be supportive of you going? Oct 11 12:57:04 * Crofton|work loves the phrase "corporate overlords" Oct 11 12:57:17 :-D Oct 11 12:57:27 good phrase to recover from a hangover to Oct 11 12:57:30 Crofton|work: I think they would let me go, yes Oct 11 12:58:19 ah, but would they pay :) Oct 11 12:58:44 my corporate overlord would be OK with me going Oct 11 12:58:55 as long as I made arrangement for the food and water Oct 11 12:59:05 and made sure someone opened the door for him Oct 11 13:00:13 Crofton|work: I think they realise the importance of my relationship with OE and the importance of the meeting Oct 11 13:00:25 cool Oct 11 13:01:45 The future of Poky should soon become clear :) Oct 11 13:02:57 hmm, is OE capable of handling compressed (gzip/bzip2) patches? Oct 11 13:03:05 mmp: yes Oct 11 13:03:15 * RP -> afk for a bit Oct 11 13:06:50 * Crofton|work peers into a crystal ball Oct 11 13:33:18 Crofton|work: btw, meant to congratulate you on the high quality security at airports in USA Oct 11 13:33:29 thanks Oct 11 13:33:32 it is awesome Oct 11 13:33:54 Crofton|work: not only did the TSA not allow my ex to serperate her liquids from hand luggage, they also allowed her through with a 500ml bottle Oct 11 13:34:05 when you come through security after arriving from foreign parts, you go through again Oct 11 13:34:14 hahaha Oct 11 13:34:32 my wife carried a camelback full of water through once Oct 11 13:34:52 anyway they re-screen you and confiscate all your duty free booze Oct 11 13:34:58 they must have raging parties Oct 11 14:25:28 hello, it looks like there is buggy patch (packages/glib-2.0/glib-2.0-2.16.1/configure-libtool.patch Oct 11 14:27:00 it fails, unable to find library suffix, because it tries to execute ./powerpc-linux-uclibc-libtool, which is not in glib's directory Oct 11 14:27:37 the errorneous line in patch seems to be +shrext_cmds=`./$host_alias-libtool --config | grep '^shrext_cmds='`; newer glib version seems to have this fixed already Oct 11 14:52:32 any idea why bitbaking bootstrap-image for generic-uclibc distribution also compiles things like gstreamer? :) Oct 11 14:53:43 from its bbfile it looks like it's quite minimal, but something pulled such dependencies in, and I have no idea what was that Oct 11 14:54:09 (compiling image for mpc8315e-rdb machine) Oct 11 14:55:24 bitbake -g bootstrap-image Oct 11 14:56:02 That will create a depends.dot file, which you can then examine manually, or convert to a pretty graphic that will show the dependency graph (and allow you to find what is pulling the unwanted stuff in) Oct 11 15:03:02 * mmp has bad feeling about leaving 'dot' running on output of "bitbake -g bootstrap0image" for few minutes, with still no output produced Oct 11 15:04:47 * mmp now knows why there was that bad feeling Oct 11 15:05:15 I think you are better off reading the dot file :) Oct 11 15:07:30 yes; it's in the end much more readable :) Oct 11 15:08:52 (and also, the victim pulling in all the gstreamer-mess was probably bluez-utils) Oct 11 15:12:32 which in turn seems to be unconditional contents of task-base :-( Oct 11 17:15:53 Is bitbake capable of handling dates such as 2008-07-16 as package versions? (e.g. comparing them, etc.) Oct 11 17:17:54 nothing anymore; I found a way to use reasonable dates :) Oct 11 17:17:58 it doesn't understand them as dates, but it should work fine so long as they are written in a reasonable format. Oct 11 17:18:17 i.e. not month-day-year, but iso format ought to be ok Oct 11 17:20:04 or didn't find the proper way... Oct 11 17:20:21 pb_: yes, they are written in YYYY-MM-DD format Oct 11 17:20:53 something doesn't like the shell-like expansions I'm doing in the bbfile... Oct 11 17:21:05 e.g. VER_YEAR = $(echo ${PV} | sed -re "s/([0-9]{4})([0-9]{2})([0-9]{2})/\1/") Oct 11 17:21:41 when used in SRC_URI, I'm just getting MalformedUrl exception Oct 11 17:22:28 argh, I need to use these hacks anyway, as the URL contains the date also in directory structure... Oct 11 17:22:34 er, do you mean you're trying to reference $VER_YEAR from SRC_URI? Oct 11 17:22:42 you can't have shell expressions in variables like that Oct 11 17:23:36 in general, the only expansions you can have in variables are the ones that the bitbake core knows how to do for itself, and $(random shell commands) is not one of those. Oct 11 17:24:20 well, so this is going to be hard then... Oct 11 17:24:26 if this is just an issue of building the right path for SRC_URI, see gnome.bbclass for a way to do what I think you are looking for Oct 11 17:24:36 I thought the excerpt above just makes normal bitbake variable Oct 11 17:24:52 it does, but it makes a normal bitbake variable which still contains the literal text that you wrote. Oct 11 17:25:00 nothing in bitbake knows how to execute the shell to perform that expansion. Oct 11 17:25:18 so, you will end up trying to pass that whole mess to the url fetcher, and it will indeed choke on it. Oct 11 17:25:59 aha... I stard to understand the reaction of fetcher :) Oct 11 17:26:19 but the gnome trick seems to be quite nice :( Oct 11 17:26:22 :-) Oct 11 17:37:48 hello people, i didn't find a way to make an image without the kernel installed. Do someone knows a way to avoid the kernel to get installed? Oct 11 17:43:58 otavio: you should be able to simply omit the kernel from PACKAGE_INSTALL. Oct 11 17:44:50 (or IMAGE_INSTALL, whichever is appropriate for your case) Oct 11 17:45:55 pb_: humm i see Oct 11 17:47:53 pb_: works at glance, thanks :) Oct 11 17:58:52 hello (once again:), TARGET_LDFLAGS should be passed directly to gcc, or directly to 'ld' ? Oct 11 17:59:12 I have one case which suggests the former, but performs the latter Oct 11 18:07:57 humm Oct 11 18:08:29 pb_: there's a way to "drop" an rdepends? I wouldn't like to rewrite task-boot to remove the kernel from it Oct 11 18:47:48 There's any way to identify why a package is being installed? procps is being installed but I fail to see why. Oct 11 19:12:57 is there any way to specify package dependencies for packages generated by FILES_${PN}-something ? Oct 11 19:14:12 ahh, DEPENDS_${PN}-something :) Oct 11 19:36:40 03  07org.openembedded.dev * r1879878159 10org.openembedded.dev.git/ (2 files in 2 dirs): python : Add python-pkgutil package. Oct 11 19:39:25 As a general rule, every variable works like that Oct 11 20:29:27 otavio: no, you can't suppress an RDEPENDS on a package which already has it. If task-boot doesn't do what you want then your best options are: (a) don't use it, or (b) supply an empty kernel package to keep it happy. Oct 11 20:30:32 pb_: right ...; another question ... I have no idea why libc6-dev is being included at the image. How I could discover it? Oct 11 20:32:11 otavio: try "ipkg whatdepends libc6-dev" Oct 11 20:32:16 or inspect the Packages file manually Oct 11 21:14:57 03  07org.openembedded.dev * r820f68801d 10org.openembedded.dev.git/packages/jpeg/jasper_1.900.1.bb: jasper: Depend on unzip-native to unpack. Oct 11 23:41:30 So, my kernel is panicking. :P Oct 12 01:27:25 03  07org.openembedded.dev * r88237a5be2 10org.openembedded.dev.git/conf/distro/include/slugos.inc: Oct 12 01:27:25 SlugOS: slugos.inc - Set preferred version for gettext; enable testlab Oct 12 01:27:25 (The newer version of gettext causes install failures for e2fsprogs) Oct 12 01:27:35 03  07org.openembedded.dev * r9341cee9a6 10org.openembedded.dev.git/: Oct 12 01:27:35 merge of '3985bcd7ec541c61234ed284c4d7d7ee263a3a9f' Oct 12 01:27:35 and 'eb9fcfc3252827f84d56edd36426fb39c72f1eed' Oct 12 01:31:56 wow - quite the lag... Oct 12 02:27:55 MUAHAHHAw Oct 12 02:28:03 psplash has been killsoresd Oct 12 02:30:41 oookay Oct 12 02:30:54 rootfs on a fvat fs loopback mounted ... is baaad mmkay Oct 12 02:34:40 03  07org.openembedded.dev * rccec743761 10org.openembedded.dev.git/packages/mumpot/ (5 files in 2 dirs): add mumpot: courtesy of Andreas Kemnade. close bug# 4714 Oct 12 02:53:19 03  07org.openembedded.dev * r95629ec1ac 10org.openembedded.dev.git/packages/vala/ (vala-native_0.3.4.bb vala_0.3.4.bb): Oct 12 02:53:19 vala: bump to 0.3.5. close bug# 4708 Oct 12 02:53:19 * package devhelp to doc **** ENDING LOGGING AT Sun Oct 12 02:59:57 2008