**** BEGIN LOGGING AT Wed Mar 23 02:59:57 2011 Mar 23 08:31:06 03Simon Busch 07org.openembedded.dev * rfa7b6692e4 10openembedded.git/recipes/freesmartphone/ (fsoaudiod/fsoaudiod fsoaudiod_git.bb): (log message trimmed) Mar 23 08:31:06 fsoaudiod: several updates to the recipe to be usefull Mar 23 08:31:06 - install daemon initscript Mar 23 08:31:06 - add dependency on alsa-lib Mar 23 08:31:06 - add extra package for daemon config files Mar 23 08:31:06 - add note about deactivate PROVIDER of openmoko-alsa-scenarios and virtual/scenarios Mar 23 08:31:07 - change PV to 0.1.0 and bump PE as version was 0.9.6 before Mar 23 08:31:15 03Simon Busch 07org.openembedded.dev * r2f2447ea3f 10openembedded.git/recipes/freesmartphone/cornucopia.inc: Mar 23 08:31:15 cornucopia: bump SRCREV Mar 23 08:31:15 Signed-off-by: Simon Busch Mar 23 08:31:16 03Simon Busch 07org.openembedded.dev * rb562300eec 10openembedded.git/recipes/freesmartphone/libfsosystem_git.bb: Mar 23 08:31:16 libfsosystem: add initial recipe Mar 23 08:31:16 Signed-off-by: Simon Busch Mar 23 08:31:17 03Simon Busch 07org.openembedded.dev * r72bc76cf4e 10openembedded.git/recipes/freesmartphone/fsodeviced_git.bb: Mar 23 08:31:17 fsodeviced: add libfsosystem as dependency and bump PR Mar 23 08:31:17 Signed-off-by: Simon Busch Mar 23 09:06:56 morning everyone Mar 23 09:08:17 morning Mar 23 09:13:47 03Martin Jansa 07master * r5748e25c59 10openembedded.git/recipes/ (efl1/evas.inc llvm/llvm-native.inc pango/pango.inc): Mar 23 09:13:47 evas, pango, llvm-native: drop empty PACKAGES_DYNAMIC_virtclass-native as it's in native.bbclass now Mar 23 09:13:47 Signed-off-by: Martin Jansa Mar 23 09:13:47 Acked-by: Khem Raj Mar 23 10:34:43 florian: good morning Mar 23 10:35:01 03Martin Jansa 07master * r79d295c31e 10openembedded.git/recipes/efl1/eeze_1.0.0.bb: (log message trimmed) Mar 23 10:35:01 eeze: fix checksums Mar 23 10:35:01 * changed upstream :/ Mar 23 10:35:01 reconfigured with GNU Autoconf 2.68 instead of GNU Autoconf 2.65 Mar 23 10:35:01 bitbake@jama ~/downloads $ diff -rq eeze-old/ eeze-new/ Mar 23 10:35:01 Files eeze-old//eeze-1.0.0/Makefile.am and eeze-new//eeze-1.0.0/Makefile.am differ Mar 23 10:35:02 Files eeze-old//eeze-1.0.0/Makefile.in and eeze-new//eeze-1.0.0/Makefile.in differ Mar 23 10:35:38 good morning Mar 23 10:40:17 hi eFfeM_work Mar 23 10:40:33 hi GNUtoo|bug20, florian, pb__ all Mar 23 10:40:53 I've the classic issue of a -native version taking precendence over a bbclass extended native Mar 23 10:40:58 hi effem_work Mar 23 10:41:05 I founght gmpc issue for a long time Mar 23 10:41:11 digged into gob2 Mar 23 10:41:14 and at the end.... Mar 23 10:41:27 I remember that the solution was to remove gob2-native Mar 23 10:41:33 bad memory Mar 23 10:41:38 what should I do? Mar 23 10:41:59 can I remove gob2-native_2.0.14.bb Mar 23 10:42:05 or move to obsolete? Mar 23 10:42:11 I really want gmpc Mar 23 10:42:15 I need it for bug20 Mar 23 10:42:25 else I can't use mpd Mar 23 10:45:35 basically I fixed locally Mar 23 10:45:39 and forgott about it Mar 23 10:45:59 then the old deleted gob2-native got pulled by git pull I bet Mar 23 10:46:07 and I digged into the problem again... Mar 23 10:46:11 GNUtoo|bug20: sometimes I feel that most of the native recipes should go, if there is an 1.X that uses bbclass extend, but also an 1.X-native that I'd say is definitely an error, if there is a 1.Y using bbclass extend and a 1.X-native (X < Y) that is also somewhat fishy, and if there is a 1.Y without extend and a 1.X-native (X which made me loose time Mar 23 10:46:31 there is an old -native Mar 23 10:46:39 and more recent bbclass-extended ones Mar 23 10:47:00 but the old -native takes precedence Mar 23 10:47:13 so how to remove that recipe? Mar 23 10:47:19 I think it is probably ok to delete 2.0.14-native. Mar 23 10:47:23 git rm? Mar 23 10:47:31 or should I move it to obsolete? Mar 23 10:47:33 I feel that this is a situation that could/should be fixed, but I gave up working in it Mar 23 10:47:50 the "new" one is only 3 point releases different, and it's only a native package anyway, so it doesn't seem like that minor upgrade will do anybody any harm. Mar 23 10:48:10 since I've bad memory I will keep trying to fix it and loose my time if I don't fix it Mar 23 10:48:15 pb__: you've beaten me on that one, I was about to check what was there Mar 23 10:48:33 pb__, the old one has huge issues: Mar 23 10:48:54 from #gmpc: Mar 23 10:48:57 GNUtoo|bug20: I t hink you need to update gob2 Mar 23 10:48:57 there was a bug in one of the releases Mar 23 10:48:57 ok Mar 23 10:48:57 that it created the wrong namespaces because of a stupid buffer overflow Mar 23 10:49:32 so I move or simply git rm? Mar 23 10:50:10 new vresion is 2.0.17,given what you pasted I'd say rm Mar 23 10:50:16 ok Mar 23 10:50:18 thanks a lot Mar 23 10:50:22 btw just for information: Mar 23 10:50:23 openembedded/recipes$ ls */*native*bb | wc -l Mar 23 10:50:23 224 Mar 23 10:50:40 I remove the normal non-native one too? Mar 23 10:50:44 yes please Mar 23 10:50:48 the gob2_2.0.14.bb Mar 23 10:50:49 ok Mar 23 10:50:51 thanks Mar 23 10:50:54 q Mar 23 10:51:48 actually I added 2.0.17 on dec 5, 2010 because it 2.0.14 did not build properly on ubuntu 10.04 Mar 23 10:52:01 see git log gob2_2.0.17.bb Mar 23 10:52:18 so somehow I am the cause of yoru problem as at that point I didn't remove the old one yet Mar 23 11:03:28 I would like to add a multiple mirrors to a SRC_URI that is used by multiple recipes. e.g. there is KERNELORG_MIRROR = "http://kernel.org" set in conf/bitbake.conf however it only adds one mirror like *_MIRROR in same file. how to go about adding multiple ? Mar 23 11:04:33 eFfeM_work, pb__ thanks a lot!!! I will push very soon,as soon as it's tested Mar 23 11:04:56 dcordes: create DCORDES_MIRROR variable in conf/bitbake.conf and use it Mar 23 11:04:56 great, thanks! Mar 23 11:05:15 (that was to GNUtoo|bug20) Mar 23 11:05:18 ok Mar 23 11:06:58 hrw: I am wondering cause the *_MIRROR have only one URI . my aim is to provide bitbake with alternative SRC_URI if one fails Mar 23 11:07:47 dcordes: check classes/mirrors.bbclass ;) Mar 23 11:10:22 hrw: lib/bb/fetch/git.py ? Mar 23 11:11:07 dcordes: for mirrors variable you need to check mirrors.bbclass Mar 23 11:12:09 hrw: no such file Mar 23 11:12:34 dcordes: ok, maybe my copy is old Mar 23 11:12:35 1.12.0 Mar 23 11:12:38 dcordes: OE Mar 23 11:12:42 dcordes: not bitbake Mar 23 11:12:57 ok Mar 23 11:13:46 crumbs, zecke seems to be hacking on sctp. eleet. Mar 23 11:13:49 hrw: awesome that's what I was looking for, thanks Mar 23 11:14:18 pb yes Mar 23 11:14:27 for gsm stuff Mar 23 11:14:43 ah, very good Mar 23 11:15:09 but he now is on vacation Mar 23 11:15:25 dcordes: use grep ;D Mar 23 11:15:37 he seems to be sending patches despite his vacation :-} Mar 23 11:15:40 hrw: I gir greped but with lack of patience it seems Mar 23 11:15:49 or maybe his mail is just lagging Mar 23 11:16:24 hm Mar 23 11:16:37 seems he is cheating Mar 23 11:17:35 hm on which list he send it? Mar 23 11:18:26 netdev Mar 23 11:19:33 hm okay this was a cheap patch Mar 23 11:19:36 *g* Mar 23 11:20:00 yeah :-) Mar 23 11:20:29 http://marc.info/?l=linux-sctp&m=129535071412417&w=3 Mar 23 11:21:19 right Mar 23 11:23:11 hm there is linux-sctp Mar 23 11:23:58 03Denis 'GNUtoo' Carikli 07org.openembedded.dev * r4f6643a771 10openembedded.git/recipes/musicpd/gmpc_0.20.0.bb: Mar 23 11:23:58 gmpc 0.20.0 : bump PR after gob2 fix Mar 23 11:23:58 The commit 5c856dbb35d02f91d8b535b8a538cef082b3a0b0 Mar 23 11:23:58 ( gob2 and gob2-native: delete version 2.0.14 ) Mar 23 11:23:58 fixes GMPC compilation, so we need to bump GMPC PR Mar 23 11:23:58 to make it rebuild and not fail. Mar 23 11:23:59 Signed-off-by: Denis 'GNUtoo' Carikli Mar 23 11:24:02 03Denis 'GNUtoo' Carikli 07org.openembedded.dev * r5c856dbb35 10openembedded.git/recipes/gob2/ (gob2-native_2.0.14.bb gob2_2.0.14.bb): (log message trimmed) Mar 23 11:24:03 gob2 and gob2-native: delete version 2.0.14 Mar 23 11:24:03 Without that the old(2.0.14) version get chosen for Mar 23 11:24:03 native, which makes gmpc fail to compile: Mar 23 11:24:03 From #gmpc on Freenode on irc: Mar 23 11:24:04 Mar 23 12:33:02 GNUtoo|bug20: I t hink you need to update gob2 Mar 23 11:24:04 Mar 23 12:33:09 there was a bug in one of the releases Mar 23 11:31:55 Hi all Mar 23 11:32:40 Does someone can tell me if vlc is building well for him? because here I have a collect2: ld returned 1 exit status Mar 23 11:36:54 GarthPS whichv ersion? Mar 23 11:37:37 woglinde: i justd do "bitbake vlc" and it builds vlc_1.1.4.1.bb Mar 23 11:37:59 latest dev? Mar 23 11:38:50 woglinde: hum? dev what ? Mar 23 11:38:57 oe master Mar 23 11:39:00 latest rev Mar 23 11:39:55 woglinde: ah :) . yep Mar 23 11:40:18 what's the actual linker error? Mar 23 11:40:35 * woglinde starts a build here Mar 23 11:43:03 pb__: http://pastebin.com/raw.php?i=wK4Sxbv9 Mar 23 11:43:36 ah Mar 23 11:43:44 lib from host sneaked in Mar 23 11:44:36 what the heel Mar 23 11:44:45 they are doing at do_configure in vlc Mar 23 11:45:30 --enable-live555 Mar 23 11:45:42 and I cannt see liblive in the DEPENDS Mar 23 11:45:44 dman Mar 23 11:45:49 that recipe is stinky Mar 23 11:47:54 we should delete the older versions too Mar 23 11:48:54 03Martin Jansa 07master * r80c7a0a06f 10openembedded.git/recipes/freesmartphone/cornucopia.inc: Mar 23 11:48:55 cornucopia: bump SRCREV a bit for fsodeviced build fix Mar 23 11:48:55 Signed-off-by: Martin Jansa Mar 23 11:49:11 woglinde: :) Mar 23 11:49:30 GarthPS, we need more details,....ah ok you already posted the details Mar 23 11:49:51 GNUtoo teh recipe is at lousy state sorry Mar 23 11:55:10 woglinde: do you think you will do something? because it think it is beyond my skills.. Mar 23 12:01:44 I dont know your skills Mar 23 12:16:59 woglinde: :) yeah , they are not sufficient that is why I am asking if you could take care about this vlc recipes.. Mar 23 12:19:06 the enable-live555 is there for > 2 years Mar 23 12:19:45 this seems a good example why it is useful to build packages only with those packages staged that are in DEPENDS Mar 23 12:20:15 eFfeM_work I doubt there are many people who builds vlc Mar 23 12:20:26 true Mar 23 12:20:54 which makes that issues like this are unnoticed for quite a long time Mar 23 12:21:35 yes Mar 23 12:22:51 I had the idea that was also what Esben was aiming at in the 'eliminating dependency race conditions" thread, but didn't get to diving into the details of it Mar 23 12:23:09 anyway, sounds like an appealing idea Mar 23 12:25:08 03Michael 'Mickey' Lauer 07org.openembedded.dev * rbcc83ea3d5 10openembedded.git/recipes/vala/ (vala_0.11.7.1.bb vala_0.11.7.2.bb): vala[-native]: 0.11.7.1 -> 0.11.7.2 Mar 23 12:25:49 effem using a lib from host is no race condition Mar 23 12:26:41 woglinde: fully true Mar 23 12:27:29 actually I brought up the race condition thread as it creates inconsistency due to order in which are things are build (which is the more general problem) Mar 23 12:27:56 if live555 is buildl before GarthPS would probalby not have this problem Mar 23 12:28:08 eFfeM_work ahe sure Mar 23 12:28:14 but its not in the DEPENDS Mar 23 12:28:16 so Mar 23 12:28:18 but recipes should ideally not use things like /usr/lib etc Mar 23 12:28:21 I know it isn't Mar 23 12:29:07 and if only DEPENDS packages would be staged, it would be noticed a long time before as e.g. my autobuilder has fairly few host packages Mar 23 12:30:08 o.O? Mar 23 12:30:17 it depends a lot on the m4 scripts too Mar 23 12:30:25 if you have --enable-foo Mar 23 12:30:40 and libfoo isnthere and the macrosays okay disable foo Mar 23 12:30:42 yeah. I don't think there is any doubt that having only the packages in DEPENDS staged into the sysroot would be a good thing. it's just a question of somebody writing the code :-} Mar 23 12:30:46 you will never hit it Mar 23 12:31:19 pb__: I didn't gist all the details, but I thought that is what Esben did Mar 23 12:31:34 well, esben wrote the code for oe-lite. Mar 23 12:31:38 woglinde: true, it does not solve all Mar 23 12:32:21 pb__: I know & it needs work, but i'm not sure if the responses in the thread were very favourable to such an activity Mar 23 12:32:33 I don't think his implementation could be trivially dropped into OE itself. Mar 23 12:34:08 yeah, I think it is true that his approach is not universally popular. RP clearly has some concerns about the whole idea of using native packages to populate the sysroot. I am not so bothered about that but I do have some concerns about the idea of changing the way that dependencies have to be specified. Mar 23 12:35:02 esben clearly feels quite strongly that both of these are key advantages and he doesn't seem to have much interest in solving "just" the per-recipe sysroot problem. Mar 23 12:35:18 using native packages to populate sysroot is something that can be fixed by having an appropriate -native recipe (like tar-native) Mar 23 12:35:54 yeah. it does also require making native builds generate output packages (which they don't at present) and that opens up a small can of worms around the toolchain packages. Mar 23 12:36:39 pb__: actually there are two things: staging per package instead of staging per recipe, the other is only stage from DEPENDS Mar 23 12:37:03 the native build thing might be nasty indeed Mar 23 12:37:47 the most obvious issue with generating output for toolchain packages is that right now they have the "wrong" architecture. I think toolchain-desuck might have a fix for that but I don't remember if it is complete. Mar 23 12:38:20 maybe a good way would be as an intermediate have a flag in the recipe that says PACKAGE_BASED_STAGING or so and still rely on DEPENDS to fidn what needs to be changed, (and in the end of course the flag could go) Mar 23 12:38:26 eFfeM_work, pb__, woglinde : all your discution is way above my head :) . just let me know if you want me to test something or if you will do something about this Mar 23 12:38:58 eFfeM_work: yeah, that would be ok. or, as we discussed on the list, a modification to the syntax of DEPENDS. Mar 23 12:39:10 there clearly are ways of solving that problem, I don't think it needs to be a show-stopper. Mar 23 12:39:30 GarthPS: add "live555" to depends of the vlc recipe, and test that, if it works please mail a patch to the list Mar 23 12:40:08 I am not personally all that sold on the benefits of doing per-package rather than per-recipe dependencies; doing them per-recipe is enough to make the results deterministic and that's the main thing that I am personally concerned about. Mar 23 12:40:37 eFfeM_work: hmm I just build live555 and then did a -cl clean vlc plus a bitbake vlc and it is still the same issue Mar 23 12:40:38 pb__: true, but I guess it would stimulate activity if we feel as a community that this is the way forward, and RPs response didn't make it too clear for me that it is the way Mar 23 12:40:44 the thing with extraneous bits getting installed in the sysroot seems like a slightly more academic issue. I guess it might be a real problem in some circumstances but it doesn't seem like it would be a big deal. Mar 23 12:41:06 GarthPS: strange, then the issue is deeper rooted Mar 23 12:41:13 eFfeM_work: yeah, rp clearly does not like the target package business. Mar 23 12:41:18 eFfeM_work: i don't know it is suppose to do the same thing that you just told me Mar 23 12:41:37 that again is sort of orthogonal to the sysroot construction thing, in that you could perfectly well construct sysroot from sstate. Mar 23 12:41:42 pb__: I agree that per recipe is probably good enough Mar 23 12:42:07 so, to that extent, rp's thing is not an objection to the whole concept, just part of it. Mar 23 12:42:17 GarthPS: if you did a build of live555 then a -cclean of vlc then rebuild vlc adn the problem is still there it is deeper rooted and my suggestion is not oging to help Mar 23 12:42:25 but I agree, from esben's point of view it perhaps does not seem very encouraging. Mar 23 12:42:40 pb__: haven't peeked into sstate details Mar 23 12:43:05 eFfeM_work: ok thanks . I am doing what you said any way Mar 23 12:43:07 * pb__ lunchtime now Mar 23 12:43:21 for me mainly the determinism is important (and I guess it would also be important to the people that produce binary feeds) Mar 23 12:43:26 pb__: enjoy! Mar 23 12:43:33 thanks Mar 23 12:43:34 later all Mar 23 12:43:55 eFfeM_work: if you wanted to follow up to the mailing list thread then that would probably be a fine thing Mar 23 12:44:00 * pb__ -> Mar 23 12:44:53 pb__: will see if I can come up with something, actually hesitated because often proposals from my side are often not considered welcome Mar 23 12:45:13 actually I figured that by not saying anything in the thread I had a higher chance of having it happen ;-) Mar 23 12:47:33 eFfeM_work: ok i did what you said and it is exacly the same Mar 23 12:48:02 yes, that is what I expeced, adding to DEPENDS just forces that live555 is build first Mar 23 12:48:13 which you manuall already did Mar 23 12:48:49 eFfeM_work: should I make a mail to the ML to point out this vlc issue ? not sure to be able to say relevant things. Mar 23 12:48:51 guess there is some deeper patch needed, but that is autotools related, and I am not a wiz in that area (nor do I have much time atm) Mar 23 12:49:14 GarthPS: please send the mail and/or file a bug in bugzilla Mar 23 12:49:25 * eFfeM_work is nto sure about the status of bugzilla atm Mar 23 12:49:40 what is weird is that Otavio Salvador just commit things about vlc last night so he should be able to build it I think.. Mar 23 12:49:49 state the error and that a host lib is used for a build Mar 23 12:50:03 GarthPS: i think it will build properly if you do not have live555 natively installed Mar 23 12:50:23 as a test you could try to uninstall lvie555 from your host (if you can) and then rebuild Mar 23 12:52:35 eFfeM_work: I just checked and I don't have liblivemedia or live555 installed on my ubuntu host Mar 23 12:52:53 hm, how then did you get the lib on the host ???? Mar 23 12:53:52 peeked again at your log, actually the issue is differently, your liglviemedia.a is not ok Mar 23 12:54:17 so it is probably related to the live555 recipe Mar 23 12:54:23 * eFfeM_work has no experience in that areay Mar 23 12:58:19 eFfeM_work: hmm..ok . I will fill a bug to the bugzilla with my output and I will see Mar 23 13:26:19 althought I set IMAGE_ROOTFS_SIZE_ext3 = "819200" the resulting rootfs.ext3 is 329944kb in size ( no extra free space ) Mar 23 13:26:24 what's wrong ? Mar 23 13:26:58 what if you change IMAGE_ROOTFS_SIZE without ext3 override? Mar 23 13:27:13 similar problem was discussed on yocto ML in last 2 days Mar 23 13:27:56 * dcordes registers to yocto ML Mar 23 13:28:03 hi risca Mar 23 13:30:41 hi woglinde Mar 23 14:28:40 GarthPS: thanks, saw your msg on the list added a little bit to it Mar 23 14:34:11 eFfeM_work: ok thanks to you! Mar 23 15:01:05 03Martin Jansa 07master * r655f557609 10openembedded.git/recipes/freesmartphone/msmcomm.inc: Mar 23 15:01:05 msmcomm: bump SRCREV Mar 23 15:01:05 Signed-off-by: Martin Jansa Mar 23 15:39:48 khem: fennec works on htcleo Mar 23 15:40:47 khem, using a toolchain from oe/tmp to build u-bbot, the final link segfaults on me Mar 23 16:42:19 03Kartik Mohta 07master * rb1f2e1501c 10openembedded.git/recipes/cmake/cmake_2.8.3.bb: (log message trimmed) Mar 23 16:42:19 cmake: add `ncurses` to `DEPENDS` (for `ccmake`) Mar 23 16:42:19 Task compile fails with the following error message with `angstrom-2010.x`, `minimal` and `minimal-uclibc` for `MACHINE = "beagleboard"`. Mar 23 16:42:19 cd Mar 23 16:42:19 /oe/build-minimal-uclibc/minimal-uclibc-dev/work/armv7a-oe-linux-uclibceabi/cmake-2.8.3-r2.1/cmake-2.8.3/Source Mar 23 16:42:19 && Mar 23 16:42:20 /oe/build-minimal-uclibc/minimal-uclibc-dev/sysroots/i686-linux/usr/armv7a/bin/arm-oe-linux-uclibceabi-g++ Mar 23 16:56:23 Crofton|work: interesting Mar 23 16:56:37 Crofton|work: what binutils version are u using Mar 23 16:57:04 angstrom 2010 Mar 23 16:57:26 not a crisis, just seems weird Mar 23 17:49:13 03Paul Menzel 07master * r7aeb138d8c 10openembedded.git/recipes/xfce-base/ (xfce4-session_4.6.1.bb xfce4-session_4.6.2.bb): (log message trimmed) Mar 23 17:49:13 xfce_session_4.6.{1,2}: explicitely pass the path of `iceauth` Mar 23 17:49:13 Xfce Session build depends on `iceauth`, which fails when cross compiling for a different architecture. Mar 23 17:49:13 As suggested by Khem Raj [1] pass the value, `${bindir}/iceauth` is taken as suggested by Koen Kooi [2], to the configure script so that the test is avoided. Mar 23 17:49:13 Therefore `iceauth` is removed from `DEPENDS` and the PR variable is incremented. Mar 23 17:49:13 This issue has been reported upstream as ticket #7420 [3]. Mar 23 17:49:14 [1] http://lists.linuxtogo.org/pipermail/openembedded-devel/2011-March/031115.html Mar 23 19:07:35 evening Mar 23 19:07:46 evening, Jay7 Mar 23 19:08:25 hi anarsoul Mar 23 19:08:37 how is your builds? Mar 23 19:08:56 built console-image for zipit Mar 23 19:09:30 but did not perform any tests yet, still have to fix my udc cable for zipit Mar 23 19:09:39 ah Mar 23 19:09:48 interesting Mar 23 19:09:59 I've not ordered zipit :\ Mar 23 19:10:24 Jay7: and there're no more zipits for $10 Mar 23 19:10:36 np Mar 23 19:10:45 I have nanonote and lot of zauruses :) Mar 23 19:10:58 nanonote has no wifi and usbhost Mar 23 19:11:19 but akita does :) Mar 23 19:12:53 well.. today I should finish my customers work with redundant iSCSI setup Mar 23 19:13:57 hm.. Mar 23 19:14:23 dear TSC members, please read https://wiki.yoctoproject.org/wiki/Yocto_Buildbot_Autobuilder_Discussions before next meeting Mar 23 19:14:57 (wrt infrastructure changes) Mar 23 19:15:58 03Denis Carikli 07org.openembedded.dev * rb3aa538b81 10openembedded.git/conf/machine/bug20.conf: Mar 23 19:15:58 bug20: add wifi and bluetooth to MACHINE_FEATURES Mar 23 19:15:58 Signed-off-by: Denis 'GNUtoo' Carikli Mar 23 19:15:58 Acked-by: Stefan Schmidt Mar 23 19:19:24 03Martin Jansa 07master * r3e7790be2b 10openembedded.git/ (classes/sanity.bbclass conf/sanity.conf): (log message trimmed) Mar 23 19:19:24 sanity.bbclass: add check for changed bb_persist_data.sqlite3 location Mar 23 19:19:24 * 4f883343de0427769de3df96ed8adaa589279112 could move Mar 23 19:19:24 bb_persist_data.sqlite3 from default CACHE tmp/cache/machine/ Mar 23 19:19:24 to global tmp/cache/ Mar 23 19:19:25 * because I have 1st report of wrong SRCPV behavior, add sanity check from Mar 23 19:19:26 http://patches.openembedded.org/patch/823/ Mar 23 19:23:10 Jay7: can you reply to the agenda email w/ that please? Mar 23 19:23:22 if you haven't already Mar 23 19:24:05 Tartarus: good idea, I will do Mar 23 19:24:30 need reboot.. my mouse gone mad.. eject/insert doesn't help Mar 23 19:31:48 hrm, oe currently requires >=1.10 right? Mar 23 19:32:15 kergoth: BB_MIN_VERSION = "1.10.2" Mar 23 19:32:47 ah, right, thanks Mar 23 19:33:00 guess we can drop this <= 1.9.0 code in base.bbclass :P Mar 23 19:33:38 conf/distro/celinux-test.conf Mar 23 19:33:38 78:BB_MIN_VERSION = "1.6.0" Mar 23 19:33:41 .... reeeally? Mar 23 19:33:44 1.6.0? Mar 23 19:33:48 i really doubt that Mar 23 19:49:40 * kergoth sighs Mar 23 19:49:56 being bit in the ass by cycles in bitbake module imports again Mar 23 19:50:07 what's most annoying is how useless python's error messages are Mar 23 19:50:10 in this context Mar 23 19:50:45 the infamous "module object has no attribute" Mar 23 19:56:42 * khem has been benchmarking a simple loop in python Mar 23 19:56:51 and list iteration seems to be fastest Mar 23 19:57:55 Crofton|work: Can you use binutils 2.21 Mar 23 19:58:00 and see if it still happens Mar 23 19:58:12 not easily Mar 23 19:58:22 did you update binutils "recently" Mar 23 19:58:32 ok then can you collect all objects and libs Mar 23 19:58:38 and make it self containing Mar 23 19:58:54 I'll send you an email Mar 23 19:58:55 * kergoth grumbles Mar 23 19:58:59 Crofton|work: ok Mar 23 19:59:07 basically clone u-boot, onfig for overo, make Mar 23 19:59:17 I just need to remember how to do that Mar 23 19:59:32 I et CROSS_COMPILE to point into the systrrot Mar 23 19:59:37 ok so u took u-boot from its git Mar 23 19:59:42 right Mar 23 19:59:42 and used OE sdk Mar 23 19:59:48 or oe internal toolchain Mar 23 19:59:51 well, oe sysroot :) Mar 23 19:59:57 ok Mar 23 19:59:59 internal toolchain Mar 23 20:00:07 u-boot does not need much from sysroot Mar 23 20:00:30 and it was for overo which mean armv7/arm mode Mar 23 20:00:40 right Mar 23 20:00:59 did binutils version change in the past week? Mar 23 20:01:02 ok send me info I will be if I am able to reproduce it Mar 23 20:01:06 no Mar 23 20:01:14 binutils havent changed for long time Mar 23 20:01:41 You could have uneathed a bug Mar 23 20:02:05 it worked "last week" Mar 23 20:02:36 ok Mar 23 20:02:50 source code is unchanged ? Mar 23 20:03:10 yeah Mar 23 20:03:15 Macth 19 it built Mar 23 20:03:23 there may be one slitght change Mar 23 20:07:22 which one Mar 23 20:47:45 hi, last JaMa|Off commit added a check for persistency Mar 23 20:47:53 here's the message: Mar 23 20:48:17 Error, persistent cache file 'bb_persist_data.sqlite3' exists in old location '/home/gnutoo/embedded/oe/oetmps/angstrom2010/cache/eglibc/bug20', please migrate it to new location '/home/gnutoo/embedded/oe/oetmps/angstrom2010/cache' and merge them together if you have one for each MACHINE. Mar 23 20:48:22 the problem for me is the last part Mar 23 20:48:27 that is to say Mar 23 20:48:36 "merge them together if you have one for each MACHINE." Mar 23 20:48:43 how do I do that? Mar 23 20:48:51 is there an sqlite3 script for that Mar 23 20:49:33 the arch of the machines are different Mar 23 20:49:44 eee701 and bug20 and bug Mar 23 20:51:57 of course JaMa|Off is off Mar 23 20:52:04 not yet Mar 23 20:52:08 ok Mar 23 20:52:11 there isn't script Mar 23 20:52:15 so what should I do? Mar 23 20:53:42 basically you want _rev from machine which was built latest, and max from _count Mar 23 20:53:52