**** BEGIN LOGGING AT Fri Jun 03 02:59:57 2011 Jun 03 14:38:27 bitbake is having an issue with the u-boot package, which I feel is probably a semi-critical one to have working :-) Apparently a file expected is not found... http://paste.pocoo.org/raw/400075/ Jun 03 14:52:16 jefferai: it's looking for the license file (specified as COPYING in u-boot recipe it would seem) in order to check if it has changed Jun 03 14:52:24 Sure Jun 03 14:52:40 And isn't finding the file, or it's in an incorrect format, or some such thing Jun 03 14:52:44 question is, what do I do about it :-) Jun 03 14:52:51 presumably this file has moved in a recent version of u-boot Jun 03 14:53:03 at least since the recipe was last built successfully... Jun 03 14:53:27 hmm, or something else went wrong in an earlier build Jun 03 14:53:31 heh Jun 03 14:53:47 I'm happy to assist in fixing, but I'm really a scathingly new n00b to Yocto Jun 03 14:53:56 so I'd need some handholding Jun 03 14:53:56 have you had failures building u-boot prior to this or used ctrl+c to break out? Jun 03 14:54:03 sure, happy to help if I can Jun 03 14:54:16 well, it took several days to compile all packages Jun 03 14:54:24 and I've noticed it bailing on uboot before Jun 03 14:54:28 but it finished the rest of the packages Jun 03 14:54:40 so when it was finished I reran the bitbake and it still fails on this Jun 03 14:54:43 but this is all within the last week Jun 03 14:54:59 ok, let me try up to do_configure for u-boot here Jun 03 14:55:05 OK Jun 03 14:55:17 (is it easy to manually run specific tasks for specific packages?) Jun 03 15:00:48 yes, bitbake -c configure packagename Jun 03 15:00:59 that will do everything up to and including configure Jun 03 15:13:00 jefferai: hmm, having some trouble here (for other reasons..) Jun 03 15:13:35 FYI sorry to be a pain but I think you would be well advised to use the 5.0.1 release rather than 5.0, (it looks like that's what you're currently building) Jun 03 15:14:51 jefferai: before trying an upgrade though can you try "bitbake -c cleanall u-boot" then "bitbake u-boot" ? Jun 03 15:15:21 you shouldn't have to do that but it would at least exclude any "broken previous build state" issue Jun 03 15:22:28 bluelightning: sure, sec Jun 03 15:23:40 bluelightning: same problem Jun 03 16:39:51 Hi, I have an error with latest bitbake cache refractory in a recipe in meta-xilinx and can't locate how to fix this error: http://pastebin.com/fA1F1t6L Jun 03 16:43:13 jefferai: hmm, well I just tried it here with 5.0.1 and it did work, so I would recommend an upgrade Jun 03 16:43:35 jefferai: in theory you can keep your same build directory Jun 03 16:47:33 aalonso: argh I think that is my fault... temporarily you can revert 58bdecbc43390796e7050ad9e0033176716ec423 Jun 03 16:51:50 bluelightning: thanks, let me try that Jun 03 17:01:38 aalonso: btw why use bitbake -b? that's really only for debugging Jun 03 17:02:44 oh, I just read that same error on a mailing list Jun 03 17:03:12 bluelightning: umm reverting spot another error: http://pastebin.com/7rffUcSz Jun 03 17:03:44 right, that's another problem someone else reported Jun 03 17:05:07 aalonso: can you not just "bitbake u-boot-xilinx"? Jun 03 17:05:43 ok, let me read on the ml and see what can I find to solve it Jun 03 17:05:49 naturally we want to fix bitbake -b but it shouldn't be necessary to use it Jun 03 17:05:54 in this case Jun 03 17:06:01 I get "NoProvider: u-boot-xilinx_git.bb" Jun 03 17:06:31 aalonso: when you run what command exactly? Jun 03 17:06:58 I thought -b was to specify a recipe; Running bitbake u-boot-xilinx_git.bb Jun 03 17:07:08 le me get ou the _git.bb Jun 03 17:07:55 Same result just running "bitbake u-boot-xilinx" Jun 03 17:09:24 that should work Jun 03 17:09:50 when specifying a target it's just the package name, no version or .bb i.e. u-boot-xilinx Jun 03 17:10:01 log error: http://pastebin.com/hkLTpUQA Jun 03 17:10:32 aalonso: I know u-boot.inc has some logic that skips the whole package if UBOOT_MACHINE is not defined... is that defined for the MACHINE you have configured? Jun 03 17:11:05 you can find out if it's beeing skipped by running bitbake -DDD u-boot-xilinx Jun 03 17:11:58 btw we are aiming to improve this kind of thing for the next yocto project release Jun 03 17:13:14 I saw that new u-boot.in includes some logic to set UBOOT_MACHINE and in meta-xilinx there is some there is a class xilinx-boot.bbclass that does the same, maybe there is some conflicts Jun 03 17:14:36 I think I need to rework this recipe since this the only recipe in meta-xilinx layer that spots this type of error Jun 03 17:16:25 aalonso: if you want to see if UBOOT_MACHINE is set then you can do "bitbake -e | grep UBOOT_MACHINE" Jun 03 17:22:15 bluelightning: I should , but I got the same error "NoProvider: u-boot-xilinx" Jun 03 17:22:34 s/I should/It should/ Jun 03 17:24:05 aalonso: which? for bitbake -DDD u-boot-xilinx or bitbake -e ? Jun 03 17:26:13 bluelightning: For "bitbake -e u-boot-xilinx"; bitbake -e doesn't show anything Jun 03 17:26:52 right, I suspected bitbake -e u-boot-xilinx would fail too early to give anything meaningful, that's why I suggested bitbake -e on its own Jun 03 17:27:25 so it sounds to me that you have a problem in that UBOOT_MACHINE is not set correctly and that's probably why u-boot-xilinx isn't showing up (being skipped) Jun 03 17:27:42 bluelightning: correction it does but greping for UBOOT_MACHINE returns nothing Jun 03 17:28:22 yeah, so it's not set globally and I assume it should be Jun 03 17:29:06 gotta go, but it should be easy to fix by setting in local.conf or your machine conf file Jun 03 17:29:28 ok, so definitely need to refractor the recipe since I do the uboot machine setting at pre-configure state and this check is happening earlier Jun 03 17:29:59 right, it could be a timing thing, I'm not sure yet at this point Jun 03 17:30:16 although actually it won't be since otherwise bitbake -e would still show it Jun 03 17:30:49 bluelightning: Yep, but since my target is fpga I can't assume which arch, thanks for your help Jun 03 17:31:34 no worries Jun 03 18:01:10 interesting....I'm doing a build from sstate and linux-libc-headers failed during the rpm packaging step - due to a missing bzip2 .so. So, I... Jun 03 18:01:28 am trying bitbake bzip2-native -c cleanall and rebuilding it, but the tasks refuse to run Jun 03 18:01:44 It just reports NOTE: Tasks Summary: Attempted 0 tasks of which 0 didn't need to be rerun and 0 failed. Jun 03 18:09:35 zenlinux_laptop: I think just bzip2, not bzip2-native Jun 03 18:10:11 jefferai, bzip2-native is a valid recipe - it uses BBCLASSEXTEND Jun 03 18:10:24 ok Jun 03 18:11:10 thanks for suggesting something, though :) Jun 03 18:29:24 zenlinux_laptop: :-) Jun 03 18:29:43 if you know about BBCLASSEXTEND -- which I don't -- you know much more than I do, so I can't help you any further :-) **** ENDING LOGGING AT Sat Jun 04 02:59:56 2011