**** BEGIN LOGGING AT Fri Mar 26 02:59:57 2010 **** BEGIN LOGGING AT Fri Mar 26 13:06:14 2010 **** ENDING LOGGING AT Fri Mar 26 13:09:37 2010 **** BEGIN LOGGING AT Fri Mar 26 13:12:15 2010 Mar 26 14:17:18 * MWelchUK_work is beginning to feel the POETS spirit. Mar 26 14:30:35 who remembers prism2 cards? Mar 26 14:32:25 * XorA does Mar 26 14:33:05 I just found out that minipci wifi card which I got for free (with pci->minipci card) is prism2 based Mar 26 14:33:13 antique Mar 26 14:33:33 wow Mar 26 14:36:09 damn, i remember the prism2 from my 5000d/5500 days Mar 26 14:36:41 * XorA still has a linksys one around for his 5500 Mar 26 14:36:43 I still have cf one Mar 26 14:36:46 pretec Mar 26 14:36:56 I got it from user (as most of my Zaurus stuff) Mar 26 14:37:23 hmm, gcc-runtime did seem like a good idea at the time :/ Mar 26 14:39:55 * mwester has a few old miniPCI prism2 cards as well... Mar 26 14:40:20 * mwester finds one in his paper-clip holder on his desk Mar 26 14:40:33 Patch gcc or try standalone builds... Mar 26 14:40:52 or not care about some build duplication Mar 26 14:42:08 kergoth: I also had an NX5000 :-) Mar 26 14:46:17 * kergoth doesn't envy RP working on that :) Mar 26 14:51:12 kergoth: Its mostly quite simple depending on whether we really want to make gcc-runtime just build the target libs and not gcc itself Mar 26 14:51:43 kergoth: I'm going to just let it make another build for now, fighting the gcc makefile is a pain Mar 26 14:54:18 * kergoth is fighting more rogue task issues with pstage, this time do_deploy in linux .. sometimes its stamp is in pstage, sometimes not, but do_deploy always runs stagefile Mar 26 14:54:43 kergoth: Probably task ordering, do_deploy got moved around in OE Mar 26 14:55:42 RP: its things that just do 'before do_build'.. no guarantees on when that'll run Mar 26 14:55:54 kergoth: yes, that will break it Mar 26 14:56:08 * kergoth thinks we may need to audit tasks before making pstage default Mar 26 14:56:15 kergoth: There used to be some solution for that but it got changed and broke packaged staging Mar 26 14:56:22 hmm Mar 26 14:56:30 Not enough people play with packaged-staging to see the breakage Mar 26 14:56:56 yeah, true, that'd be an advantage to its becoming default, though the people seeing the breakage might not agree ;) Mar 26 14:57:54 kergoth_: right Mar 26 14:58:02 kergoth_: it would make things simpler certainly Mar 26 15:00:50 until i can find and fix the root cause of this, its almost what i was suggesting with do_install. there's a hard boundary. tasks have to be either before or after package_stage Mar 26 15:00:55 heh Mar 26 15:01:31 kergoth_: Wait until you meet the people who want to inject initramfs's into kernels ;-) Mar 26 15:01:51 kergoth_: We have some strange order requirements in places although I think mostly sane solutions have been found now Mar 26 15:02:11 kergoth_: Did you work out khem's problem? Mar 26 15:02:25 not yet, seems to make no sense Mar 26 15:04:41 * RP ponders why libtool would want to link against the lib its currently trying to link Mar 26 15:11:36 how would you name package with kernel headers used for external modules or apps? Mar 26 15:12:09 kernel-dev contains config+system.map+module.symvers Mar 26 15:12:53 and recommends lot of crap: kernel-module-*-dev Mar 26 15:13:17 kernel-headers would be ok? Mar 26 15:13:18 hrw: kernel-external-build ? Mar 26 15:14:07 RP: it suggests support for external building of kernel for me Mar 26 15:15:44 hrw: external-modules ? Mar 26 15:15:54 hrw: headers suggests the wrong thing to me too Mar 26 15:16:07 RP: kernel-headers is result of 'make headers_install' Mar 26 15:18:05 hrw: right Mar 26 15:18:29 can anyone else reproduce khem's issues with bitbake master? Mar 26 15:18:34 it doesn't do that here, a bitbake -p works fine Mar 26 15:21:21 kergoth_: I wonder if he has bitbake from another verison installed on the system? Mar 26 15:21:42 possible, but doesn't bitbake prepend its own lib dir to sys.path? it should use that first Mar 26 15:26:38 kernel.bbclass/do_stage is scary Mar 26 15:26:43 kergoth_: offhad can you remember what its a sign of when libtool sets postdeps instead of letting the linker? Mar 26 15:26:59 not recognising the gnu linker? Mar 26 15:27:28 not offhand, been too long since i've touched its code, happily Mar 26 17:03:36 kergoth: hello Mar 26 17:09:41 kergoth: bitbake -b recipes/fbset/fbset-modes.bb Mar 26 17:09:47 reproduced it for me Mar 26 17:14:01 kergoth: If I remove the SRC_URI in that recipe and just touch the file so to have it, it works though I still get bitbake messages Mar 26 17:16:21 kergoth: http://pastebin.com/mUWJDm86 Mar 26 17:16:43 this recipe does not have PV in its name so I tried to append PV but it did not help Mar 26 17:41:30 kergoth: problem seems to be there when recipe has file:// in SRC_URI Mar 26 17:41:45 if I have a patch or a tar file or a url its ok Mar 26 17:49:24 kergoth: I think I see the reason for the problem Mar 26 17:50:05 its likely related to FILESPATH then, but I don't see how Mar 26 17:50:27 kergoth: http://pastebin.com/AQxh1FmM Mar 26 17:50:35 I have this in my local.conf Mar 26 17:50:58 I don't see anything wrong there, that should work Mar 26 17:51:14 hmm Mar 26 17:52:39 for a moment I punted it out Mar 26 17:52:45 and right now I am reparsing Mar 26 17:52:46 lets see Mar 26 17:52:49 k Mar 26 17:57:48 kergoth: yes it works without this Mar 26 17:58:09 so it seems to be the problem Mar 26 18:03:16 k, will try to figure out why.. Mar 26 18:03:42 kergoth: thanks Mar 26 18:05:16 kergoth: recipes/netbase/netbase_4.21.bb or recipes/fbset/fbset-modes.bb are the ones which you can use to reproduce it Mar 26 18:06:27 03Khem Raj  07org.openembedded.dev * r76619b80b5 10openembedded.git/recipes/ (5 files in 5 dirs): recipes: Get rid of redundant FILESPATH Mar 26 19:01:03 bye Mar 26 19:16:44 hi kergoth Mar 26 19:16:49 florian: good evening Mar 26 19:28:55 hmm Mar 26 19:28:56 hey pb_ Mar 26 19:29:16 bleh, moving offices is a pain Mar 26 19:29:39 heh, very true Mar 26 19:48:06 hello, can anybody help me with http://effik.pastebin.com/iYRpp8Fi ? Mar 26 21:05:20 03Antonio Ospite  07org.openembedded.dev * ra63c035d2d 10openembedded.git/files/device_table-ezx.txt: Mar 26 21:05:20 ezx: update device_table-ezx.txt Mar 26 21:05:20 Update mux nodes to be used with the ts27010mux driver. Mar 26 21:05:20 Fix names for ttyIPC0 and video0 nodes. Mar 26 21:05:20 Signed-off-by: Antonio Ospite Mar 26 21:05:20 Acked-by: Stefan Schmidt Mar 26 21:05:22 03Antonio Ospite  07org.openembedded.dev * rb7436506ca 10openembedded.git/ (3 files in 3 dirs): (log message trimmed) Mar 26 21:05:23 ezx: update ezxd version and ezxd.init script Mar 26 21:05:23 Newer kernels have drivers for BP interaction built as modules, add the Mar 26 21:05:23 loading sequence in ezxd.init. Mar 26 21:05:23 Fix also the recipe so that the CC and CROSS variables do not interfere Mar 26 21:05:23 with the crosscompiler from OE. Mar 26 21:05:24 Signed-off-by: Antonio Ospite Mar 26 21:05:24 03Antonio Ospite  07org.openembedded.dev * rde0ef33e79 10openembedded.git/ (2 files in 2 dirs): Mar 26 21:05:25 ezx: update openezx-kernel version Mar 26 21:13:35 03Antonio Ospite  07org.openembedded.dev * rc4b15bc76b 10openembedded.git/recipes/numptyphysics/numptyphysics_svn.bb: Mar 26 21:13:35 numptyphysics: update SRC_URI Mar 26 21:13:35 The old SRC_URI doesn't work anymore, use vcs.maemo.org instead of Mar 26 21:13:35 garage.maemo.org as stated on https://garage.maemo.org/scm/?group_id=638 Mar 26 21:13:35 Signed-off-by: Antonio Ospite Mar 26 21:13:36 Acked-by: Stefan Schmidt Mar 26 23:52:04 I've got a bit of an issue wherein gettext-native-0.17 is trying to link the host system's libxml2 (directly by path, in fact) instead of the staged libxml2-native and failing because the staged libz does not have gzopen64. This happens regardless of the recipe I'm using, so I don't think it's a distro or machine problem. **** ENDING LOGGING AT Sat Mar 27 02:59:56 2010