**** BEGIN LOGGING AT Wed Jun 01 02:59:56 2011 Jun 01 08:02:16 RP__: ping Jun 01 08:02:29 qhe: pong Jun 01 08:03:27 RP__: I met some issues when trying to add a BBCLASSEXTEND_append in the bitbake.conf Jun 01 08:03:44 qhe: What kind of issues? Jun 01 08:03:58 it's OK at the parsing, but when the task is about to run, a no key error is reported Jun 01 08:04:32 at lib/bb/cache.py line 304: return bb_data[virtual] Jun 01 08:04:39 KeyError: '...' Jun 01 08:05:35 qhe: Its sounds like somehow the virtual part of the filename became '...' which is wrong :/ Jun 01 08:05:41 RP__: If the BBCLASSEXTEND is written in the individual recipe, such error won't happen Jun 01 08:05:56 RP__: it's actually lib32, I just omitted it Jun 01 08:06:14 KeyError: 'lib32' Jun 01 08:06:14 qhe: ah, ok Jun 01 08:06:55 qhe: It should work the same as if it were in a recipe. This sounds like a bug in bitbake Jun 01 08:06:56 it happens within rq.execute_runqueue() Jun 01 08:09:40 qhe: I'm trying to reproduce locally Jun 01 08:10:04 RP__: ok, thank you. If you need my back trace or something I can help Jun 01 08:29:37 morning all Jun 01 09:12:12 qhe: I have a fix for you - remove the conditional on line 377 of ast.py ("if not onlyfinalise or "default" in onlyfinalise:") Jun 01 09:12:42 RP__: got it, I'll have a try Jun 01 09:13:09 RP__: and have another question Jun 01 09:13:56 do you think it's cleaner if gcc-cross-lib32 is built as well? Though it costs more time, of course Jun 01 09:46:17 RP__: ping Jun 01 10:33:41 ant_work: pong Jun 01 10:58:22 RP__: with your arm hat, what to do now that CONFIG_MTD_PARTITIONS is prohibited? Jun 01 10:59:05 ant_work: It being prohibited is the first I've heard of it Jun 01 10:59:07 http://www.linux-mtd.infradead.org/index.html Jun 01 10:59:16 btw ?? pls bump mtd-utils Jun 01 10:59:27 s/??/^^ Jun 01 10:59:35 ant_work: btw, pls send a patch ;-) Jun 01 11:01:08 ant_work: Sounds like someone needs to write some patches for the mtd partitions stuff Jun 01 11:03:47 it's just about the 'mtdparts' cmdline arg if I understand it Jun 01 11:04:07 ant_work: no, if you read the actual email from Artem its about making it on by default Jun 01 11:04:16 ant_work: so not really anything to worry about Jun 01 11:04:33 ah, ok, still possible to dynamically resize Jun 01 11:04:38 pheew Jun 01 11:05:04 do you think we could get rid of the bootloaders and move to devicetree's? Jun 01 11:05:20 for arm is still WIP though Jun 01 11:05:32 ant_work: you're answering your own question ;-) Jun 01 11:05:36 he he Jun 01 11:05:46 I hoped you had some insider knowledge ;) Jun 01 11:06:03 well, there are some good people working on arm device tree stuff Jun 01 11:06:35 good to know you're still following that stuff Jun 01 11:07:19 these days I'm fighting agains breakages wich removed some of your code (corgi_kbd et al.) Jun 01 11:08:28 ant_work: I wish I had time to looks after the zaurus devices but I simply don't :( Jun 01 11:08:42 All that code has its purposes :/ Jun 01 11:31:41 RP__: about mtd-utils, bumped Jun 01 11:32:30 ant_work: Perhaps I wasn't clear, please could you send a patch? ;-) Jun 01 11:32:50 ;) Jun 01 11:36:16 about this, no-complains makes me think nobody bothers with jffs2 in oe-core nowadays Jun 01 11:57:02 ant_work: People use mmc cards a lot more these days Jun 01 13:33:04 Hey folks, does anyone know whether the "POKYMODE" toolchain selection mechanism is still good for oe-core? Jun 01 13:56:40 Think I just answered my own question... no, it isn't. Is there a replacement mechanism, does anyone know? Jun 01 14:53:12 gm all Jun 01 16:28:12 thhp: It became a generic TCMODE Jun 01 16:32:24 RP__: thanks, I discovered this in defaultsetup.conf a little while earlier Jun 01 16:32:44 RP__: just trying to educate myself about the oe-core approach to distro definition now Jun 01 18:02:28 RP__: ping Jun 01 18:02:35 sgw: pong Jun 01 18:03:19 Would I be correct that a change to something like autoconf, should rebuild the world (or close to it)? Jun 01 18:07:24 sgw: if basichash is turned on Jun 01 18:07:59 Ah, ok which it is not by default, since I just tried and only a limited set build. Jun 01 18:10:09 sgw: correct Jun 01 19:10:24 sgw_away: it seems gnutls doesn't compile anymore Jun 01 19:11:20 sgw_away: it fails with missing reference to dlopen and like Jun 01 19:45:43 sgw_away: I fixed gnutls here; will post the fix soon Jun 01 19:46:05 RP__: since it is a build failure, want me to put it somewhere for you to push it? Jun 01 20:37:27 anyone seeing python_2.6.6 install failures? Jun 01 20:37:42 * dvhart thinks it might be kernel 3.0.0-rc1 related... Jun 01 20:57:24 otavio: under what conditions does it fail, I have multiple successful builds Jun 01 21:03:04 sgw: look at the patch I sent to ml; pb and I are discussing about it there Jun 01 21:26:18 otavio: is it possible that you have a site file that sets ac_cv_func_dlopen or ac_cv_dlopen_self? This problem occurred recently with ppc build and that was the fix. Jun 01 21:27:11 sgw: seems not **** ENDING LOGGING AT Thu Jun 02 02:59:56 2011