**** BEGIN LOGGING AT Fri Aug 09 02:59:59 2013 Aug 09 07:53:11 morning all Aug 09 07:53:50 morning bluelightning Aug 09 07:54:28 hi mihai Aug 09 08:01:38 morning Aug 09 08:02:07 hi JaMa Aug 09 08:08:01 JaMa: btw, I think the libpostproc dependency in libav is actually bogus, there's no reference to it in the libav 9+ source Aug 09 08:19:22 bluelightning: interesting, sorry I wasn't looking into it, I was just commenting about difficulties with current recipes Aug 09 08:19:34 JaMa: no worries Aug 09 08:19:38 so that first version in oe-core can be done right on first try Aug 09 08:19:50 sure, and I'm glad you pointed out the difficulties Aug 09 08:21:13 so this being the case I'll probably not add libpostproc to OE-Core, but I still can uprev and clean up the recipe if the latter is needed Aug 09 08:27:49 actually, doesn't look like any uprev is needed Aug 09 08:28:11 recipe is mostly clean as well other than SUMMARY vs DESCRIPTION Aug 09 08:28:21 and it builds successfully Aug 09 09:18:20 bluelightning: SRC_URI was indented correctly :) Aug 09 09:27:58 morning jackmitchell Aug 09 09:28:24 rburton: morning Aug 09 09:30:01 JaMa: convention as far as I am aware is to continue the spacing level from the previous line Aug 09 09:30:25 JaMa: but if that's the only issue with those patches I'll be happy :) Aug 09 10:30:13 hi there, how can i disable sanity checker in Yocto? Aug 09 10:30:26 Krz: what one in particular? Aug 09 10:30:49 ERROR: OE-core's config sanity checker detected a potential misconfiguration. Aug 09 10:30:55 this one Aug 09 10:31:34 i complains about pseudo not working correctly Aug 09 10:31:39 that's pretty serious Aug 09 10:31:42 what's the error? Aug 09 10:32:58 (pseudo not working means you can't build images) Aug 09 10:34:29 yes, you absolutely need pseudo Aug 09 10:34:53 so 1. i'm doing it under cygwin, 2. i created my own pseudo as bash script (exec $@) 3. i just want to do bitbake -c populate_sdk Aug 09 10:35:09 I hope i dont full pseudo functionality Aug 09 10:35:50 if you don't have pseudo, files will end up with the wrong permissions Aug 09 10:36:35 where ? i just need xcompiler, i don't care about permissios Aug 09 10:36:36 also, separately, rpm needs to chroot as part of its process of installing packages, so it relies on the fakechroot-type functionality also present in pseudo Aug 09 10:37:12 * rburton is scared to ask why running poky under cygwin Aug 09 10:37:54 that's just a project requirement to have windows/linux xcompiler Aug 09 10:38:56 i didn't try hard yet with canadian on linux host (not supported as well for cygwin target) Aug 09 10:39:21 bluelightning: do you know if do_fetch task checksum includes SRC_URI checksums? Aug 09 10:40:03 bluelightning: now when fetcher is able to rename download with wrong archive, it would be nice to force do_fetch execution when recipe is updated with different checksums Aug 09 10:40:05 JaMa: I'm not sure, I'll check Aug 09 10:40:15 Krz: might be easier to extend oe-core to support win32 as a host for the sdk, instead of making oe-core run on cygwin. Aug 09 10:41:51 really? what about all native tools? Aug 09 10:42:16 that would be total rework of Yocto Aug 09 10:42:45 well, win32+cygwin i guess. Aug 09 10:44:31 my point is only to create sdk for win32, i don't want to build rootfs/kernel/anything else Aug 09 10:45:08 Krz: it shouldn't be a total rework - possibly even less work than what you're trying now Aug 09 10:45:22 Krz: FYI, someone was asking about doing this same thing on the oe-devel mailing list the other day Aug 09 10:45:48 Krz: http://lists.openembedded.org/pipermail/openembedded-devel/2013-August/091545.html Aug 09 10:46:15 on oe-devel i remember question about linux host canadian for mingw32 as target, which is close Aug 09 10:46:31 yeah, this one Aug 09 10:47:03 its been a long time since i used it, but mingw32 was definitely the better option compared to cygwin Aug 09 10:47:24 i talked to the author, he is still at the same place, didnt solve the problem Aug 09 10:47:59 the beauty of cygwin is we can give people binaries + cygwin.dll and it is 'native' then Aug 09 10:49:27 the problem is I suspect it's going to be a considerable amount of pain to get the build system fully working on top of cygwin Aug 09 10:49:34 what you're hitting now is probably just the start Aug 09 10:52:25 ok, so the easiest way would be to use linux as host and try canadian, as author of oe mail? Aug 09 10:52:45 that's what I'm thinking, yes Aug 09 11:51:50 bluelightning: it's hard to configure vim to understand which line is in shell task and which line in in python task or even multiline variable :) Aug 09 16:01:11 bluelightning: I'll let QA_ERROR change in poky conf for someone using poky, OK? Aug 09 16:02:46 JaMa: certainly, I wasn't implying you would need to do that - that's up to us :) Aug 09 16:02:58 just noting it needed to be done Aug 09 16:03:40 ok, thanks for noting it :) Aug 09 16:04:38 funnily enough Scott was just about to add BUILDHISTORY_CHECKVERBACKWARDS to the manual on the same day you sent that patch :) Aug 09 16:07:32 hehe Aug 09 16:08:03 bluelightning: there is one pending change for buildhistory (recording real package size instead of .ipk) Aug 09 16:08:26 bluelightning: you wanted to improve it, will you have time for it? Aug 09 16:08:31 JaMa: ah right, I forgot about that one Aug 09 16:10:29 * bluelightning looks at it now Aug 09 16:15:20 ah yes, now I remember Aug 09 16:31:27 Has anyone out there ever had a clean build of qemuarm? I have nothing but bunches of error messages. 12 failed tasks last time I tried bitbaking. Aug 09 16:32:05 cfo215: my build machine should be doing one every night Aug 09 16:32:12 I do one every couple days or so Aug 09 16:32:19 cfo215: what's the error? Aug 09 16:32:20 clean build of what, specifically Aug 09 16:35:24 My goal is to build Angstrom with Qt for BeagleBone Black ARM Ti Sitara am335x proc. I can't get past the bitbake -k core-image-sato using qemuarm for a machine. Aug 09 16:35:46 that's what i build Aug 09 16:35:51 you should pastebin the failures Aug 09 16:36:40 ERROR: Task 5258 (/media/ToshibaExt/poky/meta/recipes-devtools/cmake/cmake-native_2.8.10.2.bb, do_compile) failed with exit code '1' Aug 09 16:37:56 error log files work for pastebin? I have lots of those. Aug 09 16:39:34 @rburton: are you for hire? I seriously need help from someone who knows what they are doing. Aug 09 16:39:56 cfo215: sadly not Aug 09 16:40:01 cfo215: several people here are though Aug 09 16:40:21 rburton: Please send them my way. Aug 09 16:41:20 my go-to contractor man sadly is on holiday Aug 09 16:43:45 cfo215: they means? Aug 09 16:44:22 cfo215: email tomas@sleepfive.com, tell him i sent you from here :) Aug 09 16:44:55 rburton: thanks Aug 09 16:45:12 cfo215: https://www.yoctoproject.org/ecosystem/yocto-project-participants lists the official participants of the yocto project, several of those are contractors for hire Aug 09 16:45:47 rburton: thanks again Aug 09 16:46:01 np. we need to make this more obvious somehow. Aug 09 16:51:15 just a suggestion, but you should put the MD5 or SHA checksums on your download pages. I was trying to follow along with the Using Pre-Built Binaries and QEMU section of the doc and the sh command must have been corrupt: "Extracting SDK...tar: Skipping to next header Aug 09 16:51:20 bzip2: Data integrity error when decompressing. Aug 09 16:51:25 Input file = (stdin), output file = (stdout) Aug 09 16:51:28 " Aug 09 17:25:38 Hi guys. Quick question. Is there GMA600 OpenGL ES support for minnowboard? Aug 09 17:52:24 moin Aug 09 18:12:22 great, now i get to explain why underlinking is not good Aug 09 18:34:39 what would happen if I sent a patch to rename meta-yocto to meta-poky? Aug 09 18:36:40 that'd break an awful lot of clones, unless you set up something for compatiblity Aug 09 18:36:46 git repo renames havec to be approached carefully Aug 09 18:42:07 yeah Aug 09 18:42:21 I am merely expressing frustration at the confusion that causes :) Aug 09 18:42:56 arg, http://schedule.armtechcon.com/session-id/71 Aug 09 18:46:19 khem, +1 for having a talk at ARM Techcon Aug 09 18:50:00 * kergoth nods Aug 09 18:50:37 nice, i didn't know he was doing a talk there Aug 09 18:50:46 pfft, baserock. platform for selling hardware Aug 09 18:55:29 rofl Aug 09 18:55:38 yeah, I came to the same solution Aug 09 18:55:47 mind you, OE sells intel servers :) Aug 09 18:55:55 heh :) Aug 09 18:56:07 although I erally want someone to do OE builds on arm servers Aug 09 18:58:26 kergoth: those cheeky baserock guys are pretty good at badmouthing yocto Aug 09 18:59:00 kergoth: my favourite was "a full yocto build can take days" Aug 09 18:59:06 haha Aug 09 18:59:20 apparently he meant hours but it came out wrong in the vodcast thingy they did, and edited. Aug 09 18:59:28 ... Aug 09 18:59:30 anyway, dinner Aug 09 19:36:39 Crofton|work: hush you :P Aug 09 19:39:42 speakigng of bad mouthing, I am trying to use some Xilnx tools Aug 09 19:59:10 kergoth: I was thinking about importing buildstats-summary into our layer, but meta-mentor doesn't have LICENSE file, can I assume MIT? Aug 09 20:03:28 hmm, thought i added one Aug 09 20:05:37 fwiw I was looking here: http://git.yoctoproject.org/cgit/cgit.cgi/meta-mentor/tree/ Aug 09 20:08:15 * kergoth nods Aug 09 21:52:10 how can i get a list of available targets for bitbake? Aug 09 22:01:07 bitbake -s is the easiest way, though it also shows versions Aug 10 00:24:17 guys Aug 10 00:24:41 I have 2 versions of a recipe.. one with a predefined version and another with git version Aug 10 00:25:21 I built first the one with a specific version.. then, later, the one from git commit Aug 10 00:26:11 the problem is that recipes that RDEPENDS on this recipe is requiring the minimum version of the package to be the specific version instead of the git one Aug 10 00:26:23 was I clear enough? Aug 10 00:28:00 let say I have two versions of libfoo Aug 10 00:28:10 libfoo_1.0 and libfoo_git Aug 10 00:28:16 some recipes RDEPENDS on it Aug 10 00:29:20 when I build an image that intalls recipes that RDEPENDS on libfoo.. this recipe is trying to install libfoo 1.0 instead of libfoo git, altough I built libfoo_git after the libfoo_1.0 Aug 10 01:42:08 oi **** ENDING LOGGING AT Sat Aug 10 02:59:58 2013