**** BEGIN LOGGING AT Sun Jul 31 02:59:57 2011 Jul 31 21:40:21 another collie user in angstrom-users.. Jul 31 21:40:32 with non-working kexecboot Jul 31 21:42:19 yes, we should recompile at least a working 2.6.24 Jul 31 21:42:28 or the latest, but with CF Jul 31 21:43:07 bluelightning: I'm battling again with layers Jul 31 21:43:10 yeah.. Jul 31 21:43:34 now I put klibc_1.5.24 in meta-oe/recipes-devtools Jul 31 21:43:42 meta-oe has priority 6 Jul 31 21:43:57 meta-zaurus (with old klibc_1.5.23) has 7 Jul 31 21:44:13 now, I would expect 1.5.24 is found at least... Jul 31 21:44:25 sigh.. still rebuilding 1.5.23 :/ Jul 31 21:45:13 I would *love* the order reflect the one in the declaration of the variable... Jul 31 21:45:23 from left to right Jul 31 21:45:49 Jay7: we had same issue :) Jul 31 21:45:54 linuxtogo.org is slow like hell Jul 31 21:54:40 * Jay7 have answered that mail just with explaination of situation Jul 31 21:54:51 shitty times.. Jul 31 22:37:27 Jay7, is that were you guys are hosted? Jul 31 22:37:44 ka6sox-farfarawa: yes, we have git there Jul 31 22:37:52 ah... Jul 31 22:39:32 ant__: if klibc is newer in meta-oe why not just remove the old version from meta-zaurus? Jul 31 22:39:43 you want meta-zaurus to have a higher priority... Jul 31 22:40:00 done that Jul 31 22:40:11 still it has higher version.. Jul 31 22:40:22 ??? Jul 31 22:40:26 I don't understand Jul 31 22:40:29 1.5.24 vs 1.5.23 Jul 31 22:40:53 once flattened, first should win Jul 31 22:41:07 are you sure there's no DEFAULT_PREFERENCE at work? Jul 31 22:41:40 any Jul 31 22:41:54 let me see Angstrom... Jul 31 22:42:04 not in the recipes Jul 31 22:42:53 I'm almost done rebasing the patches... Jul 31 22:43:10 I wonder, what does "bitbake-layers show_overlayed" show? Jul 31 22:45:08 http://paste.debian.net/124717/ Jul 31 22:46:33 let me checkout old 1.5.23 ... Jul 31 22:50:04 actually, this is not an overlayed situation, what was I thinking Jul 31 22:50:16 either PREFERRED_VERSION_ or DEFAULT_PREFERENCE will be at work here Jul 31 22:50:30 bitbake -e | grep PREFERRED_VERSION_klibc Jul 31 22:56:33 I'll do. now I see Jul 31 22:56:34 WARNING: klibc: No generic license file exists for: BSD-ADV at /oe/setup-scripts/sources/openembedded-core/meta/files/common-licenses Jul 31 22:56:34 WARNING: klibc: There is also no SPDXLICENSEMAP for this license type: BSD-ADV at /oe/setup-scripts/sources/openembedded-core/meta/files/common-licenses Jul 31 22:56:46 for the rest, 1.5.24 compiles fine now Jul 31 22:57:31 don't worry about that warning Jul 31 22:57:33 for now Jul 31 22:57:45 really? Jul 31 22:58:03 it means the LICENSE you've specified is not one we have in the common list Jul 31 22:58:06 doesn't mean it's wrong Jul 31 22:58:21 I'm not sure it was right at the beginning... Jul 31 22:58:26 for klibc 1.5.24 was it just a case of bitbake not noticing that you had added it? Jul 31 22:58:32 seems that Jul 31 22:58:43 like PREFERRED_VERSION would mask it Jul 31 22:59:08 can be in Angstrom even if I can't see why Jul 31 22:59:47 if it's just started working now magically I suspect it was just a case of bitbake not refreshing its cache Jul 31 22:59:51 as to why, I don't know Jul 31 23:00:01 I thought we ironed out most of the issues there Jul 31 23:00:12 I'm testing now on metal Jul 31 23:02:07 andrea@mizar /oe/setup-scripts $ bitbake -e | grep PREFERRED_VERSION_klibc Jul 31 23:02:07 andrea@mizar /oe/setup-scripts $ Jul 31 23:02:12 yeah Jul 31 23:02:18 it won't be that then Jul 31 23:02:41 andrea@mizar /oe/setup-scripts $ bitbake -e | grep PREFERRED_VERSION | grep klibc Jul 31 23:02:41 andrea@mizar /oe/setup-scripts $ Jul 31 23:02:43 even Jul 31 23:04:21 I think we've established it's a cache related glitch Jul 31 23:04:33 how do I trigger rebuild manually? Jul 31 23:04:49 touch conf/local.conf Jul 31 23:04:55 aditing local.conf e.g. ? Jul 31 23:04:58 :) Jul 31 23:05:02 but really I would like to try reproducing this Jul 31 23:05:15 I can send you the tarball if you mean Jul 31 23:05:21 then I committed Jul 31 23:05:28 locally Jul 31 23:05:44 can you move that new klibc to /tmp, then touch conf/local.conf, then bitbake -p, then move it back and then bitbake klibc and see which version it picks? Jul 31 23:07:33 reparsing Jul 31 23:11:33 ok, cache seems not being rebuuilt after addition of the klibc directory Jul 31 23:11:38 still 1.5.23 Jul 31 23:12:30 I can try to move it to oe-core Jul 31 23:16:33 no, don't do that Jul 31 23:16:49 if you touch conf/local.conf does it fix it? Jul 31 23:17:10 then bitbake -p Jul 31 23:17:27 well actually bitbake -p isn't necessary Jul 31 23:19:06 reparsing Jul 31 23:22:08 still not found Jul 31 23:22:14 still 1.5.23 Jul 31 23:22:28 that's indeed strange Jul 31 23:24:47 ah, fwiw klibc_1.5.24 has lower INC_PR Jul 31 23:24:53 I reset to r0 Jul 31 23:30:45 immediately after I remove it from meta-zaurus klibc appears Jul 31 23:30:52 in meta-oe Jul 31 23:31:02 withou cache rebuild Jul 31 23:31:24 ^^^ Jul 31 23:31:37 so it is somehow masked Jul 31 23:31:51 but is in cache Jul 31 23:43:49 good night **** ENDING LOGGING AT Mon Aug 01 02:59:57 2011