**** BEGIN LOGGING AT Fri Jan 27 02:59:57 2012 Jan 27 04:12:28 *yaawn k gn8 all Jan 27 07:46:06 #linux-omap Jan 27 07:46:14 just to let you know.. Jan 27 07:59:20 what's there? some sale or free goodies? Jan 27 08:03:53 good morning Jan 27 08:09:06 ynezz: free goodies with a light sprinkle of sorrows Jan 27 08:26:26 is there a way to gain sun java 1.6 on arm ? Jan 27 08:28:22 hm. I didn't *really* mean to recommend #linux-omap. It's just that weechat threw away /join and left me with just #linux-omap Jan 27 08:28:56 seems I gave Russ high expectations :p Jan 27 09:04:21 rob_w: http://www.oracle.com/technetwork/java/embedded/downloads/javase/index.html and good luck Jan 27 09:07:13 hi all Jan 27 09:09:38 rob_w: check meta-java layer Jan 27 09:16:41 https://wiki.ubuntu.com/LucidLynx/ReleaseNotes/Java6Transition Jan 27 09:37:52 03Richard Purdie  07master * re1b157d263 10bitbake.git/lib/bb/runqueue.py: (log message trimmed) Jan 27 09:37:52 runqueue.py: Add inter setscene task dependency handling Jan 27 09:37:52 This is being added to resolve setscene race issues where we do have Jan 27 09:37:52 particular dependencies required between setscene tasks. This allows Jan 27 09:37:52 specific dependencies to be specified. This allows us to fix the races Jan 27 09:37:52 in sstate with the useradd class in OE-Core. Jan 27 09:37:52 Any tasks being depended upon have their reverse dependencies cleared to Jan 27 09:38:02 03Richard Purdie  07master * r175c5c8a5e 10bitbake.git/lib/bb/ (data.py runqueue.py): Jan 27 09:38:02 data/runqueue.py: Fix commented out test code after variable name changes Jan 27 09:38:02 Signed-off-by: Richard Purdie Jan 27 09:38:02 03Richard Purdie  07master * rb9b5b5129d 10bitbake.git/lib/bb/runqueue.py: Jan 27 09:38:02 runqueue.py: Fix missing setscene dependencies Jan 27 09:38:02 When constructing the setscene inter-dependencies, we need to account for all task, Jan 27 09:38:03 not just the last one found. This patch corrects this oversight and ensures all Jan 27 09:38:03 dependencies are added, not just the first one found. Jan 27 09:38:04 Signed-off-by: Richard Purdie Jan 27 10:21:36 weird, isn't it? http://paste.pocoo.org/show/541482/ Jan 27 10:25:22 JaMa: hmm, we're not doing some hard linking magic there by any chance? Jan 27 10:30:52 that's what I was thinking but still it should behave the same for package* and "package packages-split" imgo Jan 27 10:30:55 imho Jan 27 10:31:44 but it looks like it count's it really only once :) http://paste.pocoo.org/show/541490/ Jan 27 10:51:18 hi folks! Jan 27 10:53:02 i made a pause yesterday until now. and wanted to use the devshell for virtual/kernel. but a make doenst work out as i hoped it exists quite soon with "arm-oe-linux-gnueabi-ld: unrecognized option '-Wl,-O1'" Jan 27 10:54:58 hi again :) launching an ircclient withing the devshells terminal isnt that good of an idea ;) Jan 27 10:55:41 unset LDFLAGS Jan 27 10:56:05 ipaqlor: looks like something is expecting LD to be gcc (-Wl is a gcc option to pass something to the linker) Jan 27 10:56:45 ipaqlor: which is legal, but I don't know why the behaviour would be different within the devshell and building it normally Jan 27 10:57:25 pb_: ok. Jan 27 10:58:22 so i should be able to find some that in a bb. because when i run bitbake virtual/kernel -f -c compile it works and produces kernels. Jan 27 11:05:16 can i see the all configurations that a parsed to build a kernelimage? Jan 27 11:09:42 when grepping over meta* i cant find a recipy that unsets LD flags for the kernel. only eglibc_2.12.bb, u-boot.inc, x-load.inc, eglibc_2.13.bb, eglibc_2.15.bb and net-tools_1.60-23.bb. Jan 27 11:10:00 have unset LD_FLAGS in them Jan 27 11:10:06 I've added my own FTP server as mirror for sources using "SOURCE_MIRROR_URL = "ftp://..." in local.conf. This works fine for sources that are already on the FTP server, but it fails when I add a packet that is not yet on the FTP server. BitBake then fails and says there's a checksum error. I turns out the FTP server returns a HTML file saying '550 File not found' and apparently the checksum of this file is saved. What can I do? Jan 27 11:10:07 what am i missing? Jan 27 11:10:56 kobe_: sounds like a broken ftp server. Jan 27 11:13:53 kobe_: id see if i could turn off this "feature" or use a ftp server that behaves better == not returning "something" but nothing + errorcode Jan 27 11:17:37 a ok, its because its unset witha bunch of others. ok. if i am right here: meta/classes/kernel.bbclass Jan 27 11:18:50 kernel_do_compile() { Jan 27 11:18:50 unset CFLAGS CPPFLAGS CXXFLAGS LDFLAGS MACHINE Jan 27 11:18:59 ok, now i got it :) Jan 27 11:25:17 bluelightning: the behaviour is different because kernel_do_compile is not run when running -c devshell on the kernel Jan 27 11:25:38 right, that makes sense Jan 27 11:26:26 the devshell sets up the environment as it would have for running tasks, however it does not actually run them Jan 27 11:26:38 hence the unset line you quoted above doesn't happen Jan 27 11:31:19 hi, all! Jan 27 11:32:10 can anybody help me with python cross-compilation in oe-core? Jan 27 11:32:52 I don't get C-based modules with python build, line _curses.so and IN Jan 27 11:33:22 build logs are http://ossfans.org/py/ Jan 27 11:33:58 build is from scratch Jan 27 11:34:10 buildhost is debian testing amd64 Jan 27 11:37:56 slapin: line 8 and 9 in the python-configure sure dont look like its supposed to be that way. elese i got no idea :) sorry. Jan 27 11:38:48 maybe ACLOCAL_AMFLAGS is missing the $ACLOCAL_AMFLAGS ? Jan 27 11:40:47 ipaqlor: IIRC python doesn't use automake Jan 27 11:40:52 let me check Jan 27 11:43:16 yeah, bunch of Makefile.am is there so it uses automake Jan 27 11:44:40 so something in OE automake running scripts misses to set ACLOCAL_AMFLAGS? Jan 27 11:47:36 slapin: "sh: ACLOCAL_AMFLAGS: command not found" clearly looks like there is a $ missing somewhere :) because i guess not many people have a command named ACLOCAL_AMFLAGS in their PATH. Jan 27 11:48:28 not even hardcore bit letter filename msdosusers :) because of the 8+3 ;) Jan 27 11:48:39 s/bit/big/ Jan 27 11:55:21 slapin: check who is maintaining python recipes in OE and ask directly Jan 27 11:55:26 imho this is shorter way Jan 27 11:55:28 or ask in ML Jan 27 11:56:03 iirc, kergoth` or khem anknowledged with python in OE but they are on other side of Earth ;) Jan 27 11:59:09 Ok, thanks. I'll check the settings Jan 27 12:02:58 Jay7: thanks a lot! Jan 27 13:06:35 Hi, why was boost-jam-native removed from openembedded? Jan 27 13:33:14 I found one of the causes of my python problems, but I don't know how to fix it properly. Jan 27 13:34:19 if I symlink Lib/plat-linux2 to Lib/plat-linux3 most of stuff is installed properly. Jan 27 13:34:42 that is for native Jan 27 13:34:54 but I don't know if it is proper way Jan 27 13:35:07 probably another simlink is needed in package Jan 27 13:35:29 or renaming post-processing is needed Jan 27 13:38:28 slapin: maybe you need something like this? http://git.openembedded.org/openembedded-core/commit/?id=4b7e7b004dacb698ed637f35661a60d2402c00cd Jan 27 13:38:57 slapin: python works for me fine but for 2.7 I was always using linux3 Jan 27 13:40:36 JaMa: no, regen works Jan 27 13:41:14 JaMa: if you can, could you fire-up your python binary and do 'import IN' inside Jan 27 13:41:57 JaMa: if it works when built on linux3 I want to see your logs and patches Jan 27 13:43:44 in which subpackage it should be provided? Jan 27 13:44:01 python-misc Jan 27 13:44:16 JaMa: python-misc Jan 27 13:46:55 seems to work even on linux2 build http://paste.pocoo.org/show/541584/ Jan 27 13:50:31 JaMa: a problem is on linux3 as build host Jan 27 13:52:40 JaMa: well, I think, I have an idea for temprorary workaround - will build just these modules and add custom package Jan 27 13:53:04 donno how to fix core python package Jan 27 14:03:22 ah wow Jan 27 14:03:30 I found a workaround Jan 27 14:04:02 oe_runmake MACHDEP=linux2 solves all problems completely Jan 27 14:04:40 so, how to put that into receipe properly... Jan 27 14:05:07 EXTRA_OEMAKE ? Jan 27 14:06:55 ant_work: yeah Jan 27 14:07:06 ant_work: thanks Jan 27 14:07:13 yw Jan 27 15:34:16 hmmm Jan 27 15:34:17 hi Jan 27 15:34:41 the firefox of meta-mozilla does illegal-instruction Jan 27 15:34:43 for me Jan 27 15:34:59 so I gdbserver+gdb it as usual Jan 27 15:35:13 but instead of finding x86 stuff or arm for the wrong arm version Jan 27 15:35:16 I find that: Jan 27 15:35:18 0x2c050044 <+540>: ; instruction: 0xffffd3d4 Jan 27 15:35:56 I guess I need an up to date arm manual Jan 27 15:35:59 and look it up? Jan 27 15:36:05 strange isn't it Jan 27 16:02:49 so, for once I'm actually going to be at a conference. I'll be at the yocto dev day and the rest of ELC this year Jan 27 16:02:56 (first time since going to ols in like '04 or something.. sad) Jan 27 16:03:07 heh Jan 27 16:03:10 very good Jan 27 16:03:28 hopefully i can talk them into sending me to elc-e too, that'd be nice, never know Jan 27 16:04:34 bleh, so glad the week is just about over Jan 27 16:04:39 kergoth where do you live nomrally? Jan 27 16:05:13 scottsdale, az Jan 27 16:05:47 ahh.. so it's not that bad of a trip for you.. Jan 27 16:06:00 I still have to book my airline tickets and hotel.. mgmt is being slow with final approval Jan 27 16:06:11 but I'll be there as well for the Yocto day and ELC Jan 27 16:06:14 nah, couple hour flight, probably barely hit cruising altitude before heading back down Jan 27 16:06:16 cool Jan 27 16:06:25 little longer trip from Minneapolis Jan 27 16:06:45 ah, I didn't realize you lived in minneapolis. pretty much all of my family is there, thats where i grew up Jan 27 16:07:26 * kergoth` goes to get some caffeine before he falls asleep in his chair Jan 27 16:08:02 kergoth`: hi Jan 27 16:08:25 is that normal in gdb: 0x2c050044 <+540>: ; instruction: 0xffffd3d4 Jan 27 16:08:27 kergoth`: I was told you can help me with my python-2.7.2 troubles Jan 27 16:08:32 or is that the cause of my illegal instruction? Jan 27 16:09:03 slapin: what troubles? Jan 27 16:09:59 kergoth`: might you be kind and look at oe-core mailing list? I fixed one part with setting MACHDEP=linux2 in EXTRA_OEMAKE, but _ncurses.so is still not built. Jan 27 16:10:34 ah, your issues are with building the target python.. I'm no expert in crosscompiling python, though I doubt anyone is, with the possible exception of mickeyl Jan 27 16:10:59 valhalla, also did some stuff with python Jan 27 16:11:12 kergoth`: python-2.7.2 build with linux-3.x build host Jan 27 16:11:16 mickeyl, now do more vala than python tough Jan 27 16:11:32 frameworkd was in python Jan 27 16:11:48 but now it is replaced by cornucopia(not totally tough) which is in vala Jan 27 16:12:11 denisATeukrea: that was with oe classic, however, and mostly related with sugar Jan 27 16:12:18 ok Jan 27 16:12:23 I wonder how basic problems went missing, it seems nobody uses python these days Jan 27 16:13:00 so maybe git log on python recipes to see who touched it the last Jan 27 16:13:17 plenty of people use python, but don't know how many use it on the target.. Jan 27 16:13:30 also grep for inherit distutils and such things? Jan 27 16:13:34 s/?// Jan 27 16:13:51 valhalla: could you help me with python cross-compile woes? Jan 27 16:13:56 valhalla, did you migrate to oe-core yet? Jan 27 16:14:30 denisATeukrea: I'm trying, with not much success (mostly in the area of what goes where) Jan 27 16:14:41 ok Jan 27 16:14:50 what are the issues? Jan 27 16:15:03 slapin: I'm a bit rusty in oe lately Jan 27 16:15:09 for what its worth, in an oe-core migration, i'd suggest copying someone elses setup/layout so you have a known working baseline. use poky, or angstrom-setup-scripts, or the shr setup scripts, or similar Jan 27 16:15:49 part of this problems is even with building python-native, IN module is not built/installed. when I set MACHDEP=linux2 it all worked. But it still reports missing ncurses stuff, while it is installed. Jan 27 16:16:08 No errors during build at all, just reports missing stuff in the end Jan 27 16:18:48 valhalla: http://ossfans.org/py/ build logs Jan 27 16:21:20 slapin, To find the necessary bits, look in setup.py in detect_modules() for the module's name. Jan 27 16:21:28 did you do that? Jan 27 16:23:10 denisATeukrea: yeah, probably I missed something, but everything seems to be in place Jan 27 16:23:53 what's the build system of python? Jan 27 16:24:02 autootools?/,distutils? Jan 27 16:24:08 s#/## Jan 27 16:25:36 denisATeukrea: autotools + custom stuff Jan 27 16:25:51 denisATeukrea: custom stuff is written in python Jan 27 16:25:52 * denisATeukrea fears the custom stuff Jan 27 16:25:55 ok Jan 27 16:25:59 ouch Jan 27 16:44:25 muriani1 Jan 27 16:44:55 ... Jan 27 17:25:24 Hi. Current OE-Core + meta-oe, postgresql building: http://paste.debian.net/153811/ . How to resolve this? Jan 27 17:30:39 You shCamera will not have any use case related to the OCMEM. Hence the corresponding memory sector(192KB) can be assigned to needed area. Jan 27 17:30:42 ack Jan 27 17:30:49 sorry Jan 27 18:00:41 heyo Jan 27 19:49:00 flo_lap: hey! la= nbp? ^:p Jan 27 19:49:08 +p Jan 27 19:49:59 flo_lap: anyways I read through some stuff, do I got it right when I think that we need to port dga in addition to exa? Jan 27 20:11:27 re Jan 27 20:12:55 cs_nbp: no not yet :) its still my little thinkpad Jan 27 20:14:09 flo_lap: I'm using much ssh -X atm for missing applications and it is surprisingly fast Jan 27 20:14:33 flo_lap: also found an experimental release of nxclient for arm to play around w Jan 27 20:15:02 flo_lap: could be an interesting recipe for oe anyways ;) Jan 27 20:16:39 cs_nbp: yes indeed Jan 27 20:17:16 * flo_lap needs to get oe building again Jan 27 20:19:16 * flo_lap shoots pixman Jan 27 20:19:33 update meta-oe ;) Jan 27 20:19:46 and oe core Jan 27 20:19:53 had that tonight Jan 27 20:21:01 version of pixman changed Jan 27 20:24:11 flo_lap: btw I believe I fixed irda - just enabled an irda serial console option in kernel and it stopped complaining Jan 27 20:28:24 no hw to test it though, or can serial console be established over the irda of a dvbt stick? Jan 27 21:01:06 ah dam a friends laptop doesnt start, gota take a look (suspect cr2023 bios battery) Jan 27 21:07:34 wrong chan sry Jan 27 21:08:25 ERROR: Nothing PROVIDES '"base-passwd' Jan 27 21:08:26 NOTE: Runtime target 'systemd' is unbuildable, removing... Jan 27 21:08:26 Missing or unbuildable dependency chain was: ['systemd', '"base-passwd'] Jan 27 21:08:29 anyone seeing it Jan 27 21:09:33 I think someone posted a patch for that on the list. Jan 27 21:10:53 ok Jan 27 21:11:51 seems that RP checked in a bogus patch to useradd.bbclass Jan 27 21:13:35 Does he know about it ? Jan 27 21:14:28 commit 63d006b2d3fc2223c74f81b91f70f5c841108c80 Jan 27 21:14:28 Author: Richard Purdie Jan 27 21:14:28 Date: Fri Jan 27 16:59:37 2012 +0000 Jan 27 21:14:28 useradd: Ensure dependencies are only added for target recipes, not native or nativesdk Jan 27 21:14:31 this one ? Jan 27 21:14:38 yeah Jan 27 21:14:56 if he has been reading the list, he should know about it. I don't know whether he has or not. Jan 27 21:15:24 I guess he might be off duty, which is fair enough being Friday evening Jan 27 21:18:44 hmm it seems the patch is tied to bitbake Jan 27 21:18:55 since I upadated both I got bitten Jan 27 21:19:56 khem: there's a patch for it on the list, missing trailing quote Jan 27 21:21:42 kergoth: quick link ? Jan 27 21:21:50 I dont have my mail handy here Jan 27 21:22:11 let me fix it by hand Jan 27 23:27:16 khem: , pb_ : is in this sed a bashism? >tz=$(sed -e 's:#.*::' -e 's:[[:space:]]*::g' -e '/^$/d' "${src}") Jan 27 23:27:44 ant__: looking ok Jan 27 23:27:59 the double square bracket scares me Jan 27 23:28:06 it's sed syntax Jan 27 23:28:09 not shell Jan 27 23:28:52 [[:space:]] match any whitespace symbol Jan 27 23:29:01 well..all this work to read out /etc/timezone ? Jan 27 23:29:43 this doing something with $src and saves result in $tz :) Jan 27 23:30:00 saves output Jan 27 23:31:12 I think I've finally simplifies the script: [[ ]] are found in bash and ksh afaik Jan 27 23:31:23 here for sed are ok, then **** ENDING LOGGING AT Sat Jan 28 02:59:59 2012