**** BEGIN LOGGING AT Thu Mar 08 02:59:58 2012 Mar 08 06:56:49 <_tasslehoff_> morning. can anyone spot what's wrong here: http://pastebin.com/c6mFP8EK ? Mar 08 06:58:27 <_tasslehoff_> I added a small patch in Lib/Queue.py, and bumped the revision to PR = "${INC_PR}.4". Can't understand how that should upset it so, even if my patch was bad. Mar 08 08:29:27 good morning Mar 08 08:33:37 hi mckoan Mar 08 08:47:54 morning Mar 08 10:28:48 I'm trying to build opkg-utils_svn.bb but do_fetch fails since svn.openmoko.org has been down for 3 days now, any way to work around this? Mar 08 10:32:15 I notified the admins in #openmoko yesterday about the svn server being unavailable Mar 08 10:34:20 it's in git now Mar 08 10:37:06 JaMa: So the bb package needs to be updated? Mar 08 10:38:20 it's already updated in oe-core and there is patch for oe-classic on ML Mar 08 10:44:07 JaMa, I'm using the master branch @ git://github.com/openembedded/openembedded.git Mar 08 10:45:51 that's oe-classic Mar 08 10:46:35 JaMa, ok, where is oe-core? Mar 08 10:47:19 JaMa, nm, sorry, googled it :) Mar 08 10:48:11 JaMa, thanks for the help! Mar 08 11:08:20 hi jama Mar 08 11:44:12 woglinde: hi Mar 08 12:37:27 ping woglinde Mar 08 12:44:42 whats up pwgen? Mar 08 12:46:37 i try to get the openjdk6 compiled for oe-core and i586, but i got mostly segfaults with qem, and when its not qemu i have strange errors with ( rpData + null ) Mar 08 12:48:35 hm it should not use qemu for i586 Mar 08 12:48:37 hm hm Mar 08 12:49:05 but that is not build in yet Mar 08 12:49:49 i want building it on a 64 bit machine for a 32 bit machine with i586 target Mar 08 12:50:01 and to be honest I am not in the mood to change it quickly Mar 08 12:50:14 hm Mar 08 12:50:26 you normaly do not need it Mar 08 12:50:35 there is build in stuff that it is using -m32 Mar 08 12:50:43 dont know why the qemu-line kicks in Mar 08 12:51:29 but I am open for patches Mar 08 12:53:29 after i set the "QEMU= .... -s 1048576" option to 2048576 in openjdk-6-common.inc the qemu went through, but it fails with these "Native.putLong(pData + null, v)" errors complaining that + is not valid for null pointers Mar 08 12:54:44 uhm I tought I used more mem for qemu already Mar 08 12:54:47 hm hm Mar 08 12:55:01 maybee I didnt checked it in some ages ago Mar 08 12:55:53 for which target platforms do you could compiled/tested openjdk / Mar 08 13:48:09 pwgen for arm Mar 08 14:02:08 ok then i will try it for an arm architecture .. Mar 08 14:10:48 hmmm Mar 08 14:11:10 does someone know which recipe I should make to get 'gss_OID' Mar 08 14:11:32 I didn't found the header on search engines Mar 08 14:11:43 *corresponding header Mar 08 14:11:49 it's related to gss API Mar 08 14:16:40 ah I think it's kerberos Mar 08 14:16:43 I'll try Mar 08 14:16:48 denisATeukrea try kerberos Mar 08 14:16:49 yes Mar 08 14:16:50 lol Mar 08 14:16:59 thanks a lot Mar 08 14:18:37 since there is no kerberos in oe-core I'll just disable it Mar 08 14:36:00 Anybody having problems to compile python-setuptools? My build fails when installing. Build log: http://pastebin.com/CdPeFb5N Mar 08 14:48:36 Nobody using python-setuptools? Mar 08 18:57:30 is anyone familiar with get_subpkgedata_fn in meta/oe/lib/packagedata.py? Mar 08 18:59:15 get_subpkgedata_fn is not working the multilib case Mar 08 18:59:39 its defaulting the PKGDATA_DIR instead of the right multilib directory for pkgdata Mar 08 19:06:11 msm write to the oe-core ml Mar 08 19:06:42 woglinde: yea, i think i can fix this bit… i'll propose a patch Mar 08 19:17:09 hi woglinde Mar 08 19:17:25 do you know if gpsd 2.96 and gpsd 3.x have the same api? Mar 08 19:17:39 is it worth to fight against scons to create a new(3.x) recipe? Mar 08 19:17:58 the advantages would be ubx support and hooks support Mar 08 19:20:18 it seem that slapin already tried but got no response when he was blocked so he abandonned Mar 08 19:21:22 GNUtoo: if you get 3.4 working please let me know, im facing a very similar issue and scons is pissing me off to no end Mar 08 19:21:31 hell any of the 3.x builds would work Mar 08 19:22:46 yes but did the API for apps changed? Mar 08 19:23:09 for instance will apps built for 2.96 still work on 3.x? Mar 08 19:23:39 since they moved to jason, i doubt it Mar 08 19:23:52 ok thanks Mar 08 19:24:07 maybe I'll look later on gpsd 3.x Mar 08 19:24:12 I'll write the gpsd app right now Mar 08 19:24:17 its hard to say, all the recent work iv done with is the 3.x series Mar 08 19:24:28 but at some point we want to switch om-gta02 to gpsd Mar 08 19:24:36 GNUtoo: i have an updated scons build and gpsd build, though i suspect i did somthing wrong as i suck at writing oe builds Mar 08 19:24:45 ah nice Mar 08 19:24:49 can you pastebin it Mar 08 19:24:49 ? Mar 08 19:24:55 gpsd refues to build... because it claims i have scons 1.2 Mar 08 19:25:12 even though its A. not installed, B. not installed on the build system and C. scons 2.1 builds and installs Mar 08 19:25:15 is there a scons bbclass? Mar 08 19:25:24 GNUtoo: if it gets my issues fixed sure... Mar 08 19:25:40 GNUtoo: not sure... im an idiot whos too dumb to give up... started learning oe/bb 2 weeks ago Mar 08 19:25:44 also submit infos permitting me to make a commit Mar 08 19:25:51 ill share what i have with the hope of getting the issue fixed Mar 08 19:25:57 like who should I credit for the initial recipe Mar 08 19:26:23 GNUtoo: i just ripped off the old ones and tried updating to new build locations and hashs's Mar 08 19:26:26 thats about as far as i got Mar 08 19:26:32 what is the old one? Mar 08 19:26:35 the 3.x old one? Mar 08 19:26:37 or the 2.x Mar 08 19:26:38 2.95 is the last i have Mar 08 19:26:43 ok Mar 08 19:26:52 im using the gumstix anstrom build Mar 08 19:26:52 so maybe I should start from the 3.x one posted on ml Mar 08 19:26:56 thats recomended off their website Mar 08 19:26:56 ok Mar 08 19:27:05 gimi a min ill post what i have Mar 08 19:27:06 is it core based? Mar 08 19:27:14 no :( i dont think so Mar 08 19:27:18 ah ok Mar 08 19:27:28 I'll have to do a core recipe tough Mar 08 19:27:36 because shr switched to oe-core long time ago Mar 08 19:28:14 GNUtoo: i followed this... http://gumstix.org/software-development/open-embedded/61-using-the-open-embedded-build-system.html Mar 08 19:28:18 i dont know if its core or not... Mar 08 19:28:48 it's not core based Mar 08 19:28:51 i do know that what ever it is, i had to build it on a 10.x ubuntu system as it wont run on my normal day to day gentoo box, to many failed packages that i didnt want to fix Mar 08 19:29:07 great..... Mar 08 19:29:16 so similar issue and i cant use the fix... oh well Mar 08 19:29:23 let me get you what i have... maybe it will help Mar 08 19:30:13 that cant be too old... it builds me a 3.0 kernel Mar 08 19:30:38 the kernel and the build system are a separated thing Mar 08 19:30:45 maybe a 3.2 kernel doesn't exist yet for your device Mar 08 19:31:03 for instance with SHR we have a device that has a 2.6.24 Mar 08 19:31:04 may not Mar 08 19:31:20 i was implying more that it still gets updated Mar 08 19:31:22 still the build system(openembedded) is up to date Mar 08 19:32:40 gpsd.inc -> http://slexy.org/view/s21O2nFsqN Mar 08 19:33:40 my 3.4 build Mar 08 19:33:42 http://slexy.org/view/s2qSiJxUoO Mar 08 19:34:34 scons 2.1 http://slexy.org/view/s201DuvJV2 Mar 08 19:34:46 hmmm you say that 3.4 is out Mar 08 19:34:55 but wikipedia still says 3.00 Mar 08 19:35:17 yeah wikipedia is prob wrong Mar 08 19:35:18 ah indeed Mar 08 19:35:23 they had to move the site Mar 08 19:35:25 yes releases have 3.4 Mar 08 19:35:28 but everything points to the old one Mar 08 19:35:33 ok Mar 08 19:35:48 scons native build http://slexy.org/view/s2C1Iu5dw5 Mar 08 19:35:52 thats what i have... Mar 08 19:35:55 hope it helps Mar 08 19:36:04 let me know how it goes maybe i can make it work for my stuff Mar 08 19:36:31 have to meet with client on sunday... was going to try to have somthing by then... 2.95 is rather unstable for some reason Mar 08 19:36:39 I think I'll go from the one in meta-oe Mar 08 19:36:50 cool Mar 08 19:36:58 the links and hashes should help though Mar 08 19:37:03 those pull from the correct site Mar 08 19:37:14 they changed license? Mar 08 19:37:58 cant say i really checked... Mar 08 19:38:10 i normaly look at that after things work Mar 08 19:38:23 you have put GPL as license Mar 08 19:38:38 im pretty sure that was in the orginal 2.95 build file Mar 08 19:38:43 ah ok Mar 08 19:38:45 then it's an error Mar 08 19:38:48 because it's BSD Mar 08 19:40:07 yeah orginal inc file says GPL Mar 08 19:41:07 ok Mar 08 19:50:40 * Tartarus dusts off, posts a recipe Mar 08 19:55:23 TheX1le, btw do you know how to write gpsd apps? Mar 08 19:55:46 because I'm just interested in doing that: Mar 08 19:56:00 I just want to look if there is incomming data on the GPS Mar 08 19:56:01 not more Mar 08 19:56:06 not what the data is etc... Mar 08 19:56:31 GNUtoo: iv written some python ones, but never c Mar 08 19:56:42 I need to write it in python Mar 08 19:56:47 so I did that: Mar 08 19:56:48 daemon = gps.gps ( host="localhost",port=2947,mode=gps.WATCH_ENABLE|gps.WATCH_NMEA,verbose=1) Mar 08 19:57:13 and then I can do daemon.next() and daemon.response Mar 08 19:57:23 GNUtoo: id be happy to share the code i wrote but i dont know if it will help you Mar 08 19:57:27 but I wonder if there is something that suits more my use case Mar 08 19:57:45 I just want something that tells if some incommin...ahh wait a second Mar 08 19:57:47 i wrote somthing that logs everything out to a database and makes things more reable Mar 08 19:57:50 maybe xgps has something Mar 08 19:58:20 if I remember well xgps had something to know if there was incomming data Mar 08 19:59:53 it does gobject watch Mar 08 19:59:54 hmmm Mar 08 20:00:48 btw if I fail to do the recipe because of the lack of time, look at the mailing list Mar 08 20:00:52 there is a thread named: Mar 08 20:01:03 [oe] trying to compile recent gpsd (one with scons) with OE Mar 08 20:01:12 that one contain a better scons bbclass Mar 08 20:01:38 GNUtoo: i sent you code, see pm Mar 08 20:01:43 ok Mar 08 20:05:13 yeah see i have no problem getting scons 2.1 to build.... however i cant get the gpsd to detect its there... Mar 08 20:06:20 let's see because it's rebuilding the toolchain here Mar 08 20:17:34 GNUtoo: http://slexy.org/view/s2viYgCuJ7 Mar 08 20:17:40 this is the error i cant seem to get around Mar 08 20:21:45 ok Mar 08 20:24:44 for me just inherit scons make it go further Mar 08 20:24:54 it fails at compilation tough Mar 08 20:25:16 | gpsutils.c:19:21: fatal error: QDateTime: No such file or directory Mar 08 20:25:23 + cross compile badness Mar 08 20:25:55 TheX1le, look if there is a meta-gumstix in the meantime Mar 08 20:26:06 because you'll have to switch to core based oe at some point Mar 08 20:26:18 -I/usr/include/qt4/QtNetwork Mar 08 20:26:22 ahd that's why Mar 08 20:27:55 GNUtoo: yeah thats odd i have the inherit scons in the gps.inc Mar 08 20:28:44 I replaced autotools by scons with inherit scons too Mar 08 20:28:51 but as I said we don't have the same oe versions Mar 08 20:28:59 yeah... Mar 08 20:29:08 you use the old unmaintained one where no one develops for it anymore Mar 08 20:29:23 so the question is not if to switch or not, it's when to switch Mar 08 20:29:41 yeah... which means i have to port all my other custom code... Mar 08 20:29:43 damn it Mar 08 20:29:53 i hate embeded dev lol Mar 08 20:29:56 you can just do a layer Mar 08 20:30:23 i have to read how all of this works Mar 08 20:30:34 it's not very different Mar 08 20:30:46 you just have licenses checksums that were added Mar 08 20:30:52 There's the rub with OE -- if you can't spend the numerous hours per week to keep up with it, it's darn near impossibl eto catch up (as I'm finidign) Mar 08 20:30:57 and do_stage was removed Mar 08 20:31:09 and some other minor details Mar 08 20:31:22 also if you didn't use .bbclass before you should learn about it Mar 08 20:31:25 oops Mar 08 20:31:26 sorry Mar 08 20:31:32 also if you didn't use .bbappend before you should learn about it Mar 08 20:35:10 yeah im gonna have to read up on how to build the env, i know so little about embeded stuff its not funny Mar 08 20:35:21 so learning a new way to creat images is gonna take some time Mar 08 20:35:29 but updates are good and im going to need it Mar 08 20:35:38 would love the 3.3 kernel or at least its wireless patches Mar 08 20:36:15 use compat-wireless for this Mar 08 20:36:35 yeah, that doesnt exist in the curent verions of oe im using Mar 08 20:36:54 than make a recipe Mar 08 20:39:07 woglinde: well since the stuff im using is no longer support my time would be better spent porting my stuff to the oe-core Mar 08 20:39:15 as it seems others are working on it as well Mar 08 20:39:57 TheX1le, is there a meta-gumstix for oe-core? Mar 08 20:40:18 http://www.openembedded.org/wiki/LayerIndex says that there is Mar 08 20:40:21 so you're good Mar 08 20:41:08 GNUtoo: http://gumstix.8.n6.nabble.com/overo-on-oe-core-environment-td668072.html Mar 08 20:41:36 yes you should try that Mar 08 21:06:22 gnutoo ping Mar 08 21:06:31 pong Mar 08 21:06:52 gnutoo you played with the akmd free variant? Mar 08 21:06:58 yes of course Mar 08 21:07:00 okay private its off topic Mar 08 21:07:12 maybe come in #replicant rather Mar 08 21:27:16 * kergoth sighs, meta-oe still seems broken wrt systemd Mar 08 21:27:18 Missing or unbuildable dependency chain was: ['busybox-httpd', 'systemd', 'udev-systemd', 'consolekit', 'virtual/libx11'] Mar 08 21:27:38 libx11 sucked in by busybox.. Mar 08 21:28:44 lol Mar 08 21:28:45 why? Mar 08 21:29:00 I thought koen has enough time now to fix these problems Mar 08 21:29:11 * kergoth hasn't been following the systemd threads, shrugs Mar 08 21:31:35 kergoth: meta-oe+what? Mar 08 21:31:45 Been building angstrom fine, of course Mar 08 21:31:58 or at least, systemd-image builds fine Mar 08 21:32:02 anything without x11 in DISTRO_FEATURES will see the above, as far as i can tell Mar 08 21:32:09 there's n o way to disable systemd usage, still, as far as i can tell Mar 08 21:32:30 which is ridiculous, it should never have been merged in this state Mar 08 21:44:15 is there an easy way in a recipe to set a variable to a value only if this value is not empty ? (to be clear I want to change LD only when ld-is-gold is set in DISTRO_FEATURE) Mar 08 21:46:47 use base_contains in an inline python snippet Mar 08 21:49:28 ok I wanted to avoid the python snippet but failed to do so ;) Mar 08 21:49:40 thanks Mar 08 21:49:57 there's PACKAGECONFIG, but that only lets you control depends and EXTRA_OECONF, not other vars Mar 08 22:35:17 TheX1le, slapin are you still there? Mar 08 22:36:30 good nite Mar 08 22:36:36 good night Mar 08 22:36:41 ha he has quit Mar 08 23:12:26 khem: here gcc-cross-initial_4.6.bb, do_patch) failed with exit code '1' (patch gcc-4.6-linaro-r106740.patch) Mar 08 23:40:20 * mr_science shakes his fist at the TI arago people Mar 08 23:48:40 mr_science: problems? Mar 08 23:49:25 yeah, switching branches inside the arago repos seems to confuse bitbake Mar 08 23:49:50 my setenv and local.conf are ignored by git Mar 08 23:50:46 but after switching branches it stops respecting SCRATCH and starts dumping downloads and tmp in the wrong place Mar 08 23:51:02 very irritating Mar 08 23:51:22 hmm, which branches? works fine for next and master Mar 08 23:52:02 setenv and local.conf were removed from under SCM control to preserve for this kind of switches Mar 08 23:52:16 we're making local branches off one particular TI dev's tags Mar 08 23:52:25 there are .sample versions of those files to copy from. unless you switch to something really-really old Mar 08 23:52:41 i started from the samples Mar 08 23:52:51 they're in .gitignore Mar 08 23:53:03 what tree/branch/tag are those? Mar 08 23:53:09 so i'm not sure why it gets confused Mar 08 23:57:32 http://paste2.org/p/1932585 Mar 08 23:58:33 brb Mar 09 00:23:12 back... Mar 09 00:23:55 denix: that paste is the answer to your question Mar 09 00:24:58 checking it now Mar 09 00:25:51 so for now i'm giving up on splitting things up Mar 09 00:26:09 i had i working nicely but it got confused again today Mar 09 00:27:02 hmm, besides using a WIP c6a8 tree instead of the mainline arago, everything else looks fine Mar 09 00:27:33 and it's been working for my daily branch switching between next and master Mar 09 00:28:03 can you check your shell env for SCRATCH as well as what bitbake thinks SCRATCH is through bitbake -e Mar 09 00:38:52 shell had SCRATCH correct Mar 09 00:39:55 didn't know about the other one Mar 09 00:40:26 yup, bitbake has the wrong value Mar 09 00:42:32 so it's not respecting the "export SCRATCH=foo" in setenv but if i change it in local.conf to match what setenv sets, then it really freaks out and starts getting parse errors on all of the git-based packages Mar 09 00:42:57 what am i missing? Mar 09 00:43:20 brb Mar 09 00:54:16 mr_science: try uncommenting back BB_PRESERVE_ENV in setenv Mar 09 00:55:17 bbl Mar 09 00:56:48 mr_science: i assume you'r eadding that to BB_ENV_EXTRAWHITE? Mar 09 00:59:25 kergoth, adding the BB_PRESERVE_ENV piece? Mar 09 00:59:50 haven't looked at the branch, just figured it was an obvious thing to check. bitbake filters the env, after all Mar 09 01:00:32 yeah, but i'm still not sure what you're asking Mar 09 01:01:12 this is what i have now: BB_ENV_EXTRAWHITE="OEBASE MACHINE META_SDK_PATH http_proxy https_proxy ftp_proxy no_proxy GIT_PROXY_COMMAND" Mar 09 01:02:08 but i can certainly set BB_PRESERVE_ENV=1 and see what happens Mar 09 01:03:08 well, that seems to preserve scratch correctly Mar 09 01:07:42 okay, it seems happy now Mar 09 01:08:23 thanks, kergoth and denix Mar 09 01:23:00 * mr_science hits the trail... Mar 09 01:36:03 nerdboy: how do you set remove **** ENDING LOGGING AT Fri Mar 09 02:59:58 2012