**** BEGIN LOGGING AT Fri Feb 11 02:59:57 2011 Feb 11 08:23:28 grg, Jay7 backreading the log, thanks for your messages; actually I proposed this to the board 10 days ago, but never got any feedback on it Feb 11 08:25:26 * eFfeM_work also feels that it would be good if all TSC members actually have a presence in #oe if only to ask about developer opinions and/or pick up issues Feb 11 08:27:48 effem whats the problem? Feb 11 08:28:00 sometimes you seem a little bit to pedantic Feb 11 08:28:59 there is no mubarak in oe-community which put you in a hidden jail Feb 11 08:29:48 hi eFfeM_work , woglinde Feb 11 08:29:57 hi ant Feb 11 08:32:54 hi ant_work, woglinde Feb 11 08:35:11 woglinde: I was somewhat irritated yesterday, but I do feel that the board and the tsc should be more open to the community, after all this is a community project and in order to get the most support it seems best to inform and involve the communjity members as good as possible Feb 11 08:35:28 they will do Feb 11 08:35:40 gm rob_w Feb 11 08:36:14 gm woglinde Feb 11 08:37:28 eFfeM_work: The mails from the board haven been send to the devel list. All this iterim TSC has "done" so far is to setup a ml and negotiate a weekly meeting date Feb 11 08:37:47 eFfeM_work: Most likely monday 9pm GMT on IRC Feb 11 08:38:09 eFfeM_work: The minutes will come timely for this 1-2 days worst Feb 11 08:38:26 eFfeM_work: I will personally look into getting the minutes out even more quickly Feb 11 08:38:46 eFfeM_work: its one of the biggest issues I had before. Not enough info what TSC was doing Feb 11 08:39:07 stefan *g* not info is effem biggest problem Feb 11 08:39:09 eFfeM_work: Now you have the same status on infos as the interim TSC has. :) Feb 11 08:39:12 +enough Feb 11 08:40:25 stefan_schmidt: if you want things to be open why not open up the tsc mailing list or archive for members reading and/or allow (perhaps r/o) access to the irc channel. Feb 11 08:41:07 eFfeM_work: This is something I would like to bring up on the Agenda Feb 11 08:41:34 eFfeM_work: But the interim TSC is not there to fix all the community/people problems but to get the oe-core stuff moving forward Feb 11 08:41:53 Anyway, I have a lecture now. Will be back later Feb 11 08:42:11 woglinde: actually I feel that community members that want to be involved should be involved as much as possible, personally I would even be in favour to involve them into the discussion in the tsc channel (but ony tsc members have a vote when it comes to deciding thigns) Feb 11 08:43:07 effem and you dont believe in the board? Feb 11 08:43:18 I belevie them more than our politicans Feb 11 08:43:21 hi ensc Feb 11 08:44:54 woglinde: I feel this should be as open as possible and let everyone who wants to participate and have a say Feb 11 08:45:04 * JaMa agree, that public irc sessions would be fine Feb 11 08:45:18 and of course at some point things need to be decided and you need a way to get to decisions Feb 11 08:46:30 but I don't share your concerns about current interim TSC.. it's much better to have someone to deal with yocto than to discuss how we can have someone like that Feb 11 08:47:01 but somehow I feel currently too many things are done outside the general public, e.g. why did the board choose the people who expressed interest on the mailing list spontaneously but did not consider those who expressed interest by mail, and why did they not ask if there were others who wanted to participate? Feb 11 08:47:43 as such it does resemble the behaviour of our politicians, although fortunately we have no Wilders in the OE board ;-) Feb 11 08:48:19 for interim stuff there is normaly no asking Feb 11 08:48:47 JaMa: I have no problems with the interim TSC as such only the way it is constructed Feb 11 08:49:10 effem calm down and in 2 months its all over Feb 11 08:50:23 woglinde: perhaps true, but this is one of the most important steps in OE (and if it is an interim stuff then I think it should be interim and last as short as possible (but as long as needed), now I see people being added to the TSC without election for almost a year (assuming the 2 months voting) Feb 11 08:50:31 and be sure I will vote for you in the next voting Feb 11 08:50:39 well, I personally also wonder why it's hidden Feb 11 08:50:51 so dont have to read the thousends emails for not being open Feb 11 08:51:19 to be honest, the fact are contradicting what's written here http://wiki.openembedded.org/index.php/TSCCharter Feb 11 08:51:34 redmine and chilli project, good example Feb 11 08:52:56 https://www.chiliproject.org/projects/chiliproject/wiki/Why_Fork Feb 11 08:55:28 ynezz go fork *g* Feb 11 08:55:49 :) Feb 11 08:56:34 I don't want that same happens to OE, that's all Feb 11 08:57:21 why should this happen? Feb 11 08:59:25 well, the fork is maybe strong word here Feb 11 09:04:30 hm.. what is cheapest HW with ethernet and usb-host? :) Feb 11 09:04:42 arduino? :p Feb 11 09:04:45 03Henning Heinold  07org.openembedded.dev * reae7de3042 10openembedded.git/recipes/libsdl/libsdl-mixer_1.2.11.bb: libsdl-mixer: fix it the same way as the libsdl recipe and bump PR Feb 11 09:04:56 03Henning Heinold  07org.openembedded.dev * ra5d1ff6e86 10openembedded.git/recipes/libsdl/libsdl-directfb_1.2.9.bb: libsdl-directfb: delete unpinned version 1.2.9 Feb 11 09:05:56 Jay7: some WiFi AP probably, depends on flash/ram and cpu/computing power you need Feb 11 09:06:10 ah, hm.. good idea Feb 11 09:06:16 some dlink + *wrt Feb 11 09:06:33 e.g. dir-320 Feb 11 09:07:26 * Jay7 have sven ups with usb and xenserver w/o NUT Feb 11 09:08:42 32MB RAM, hm Feb 11 09:08:52 but the flash seems small Feb 11 09:09:04 so my idea is having some device which will monitor UPS and shutdown xenserver properly :) Feb 11 09:10:32 ah, than it's enough resources Feb 11 09:11:21 hmhm.. I can even still have ethernet+wifi bridge as a bonus :) Feb 11 09:17:04 MeeGone.... Feb 11 09:17:36 another one bites the dust. Feb 11 09:17:58 esbenh *g* Feb 11 09:18:14 the poor suse guy Feb 11 09:18:18 * esbenh sheds a tear over his N900 Feb 11 09:18:20 one year for nothing Feb 11 09:18:28 yeah :/ Feb 11 09:18:34 a shame that they did not manage to take advantage of the work done on the maemo project Feb 11 09:19:08 well, Intel have not wasted the year Feb 11 09:19:13 problem n900 got not ROI Feb 11 09:19:32 you cannt do this many projects Feb 11 09:19:39 not in capitalism Feb 11 09:20:07 having worked in the mobile phone business (Siemens -> BenQ -> Motorola), the missing ROI on N900 is purely a matter of lack of strategic leadership Feb 11 09:20:20 you will never make anything resembling ROI on the first product in a new platform. Feb 11 09:20:58 which is why I had a really hard time to see the vision of the maemo->meego decission seen from Nokia perspective Feb 11 09:21:24 Intel, OTOH have just gained a lot of free work from Nokia to their embedded Linux OS.... Feb 11 09:23:31 hm Feb 11 09:23:38 the problem is Feb 11 09:23:44 what will intel do with it? Feb 11 09:24:21 thats the same problem as nokia has Feb 11 09:24:24 I guess they will continue their strategic spending in MeeGo Feb 11 09:24:54 but intel has more selling fields than nokia Feb 11 09:25:18 * ka6sox fails to see how a company that can't produce an x86 that doesn't have a TDP of less than 7watts can call itself embedded, (having sold off their ARM business) Feb 11 09:25:22 esbenh did you here of any sucessfull meego device? Feb 11 09:25:30 Well, Nokia has much more trouble than Intel. Nokia is loosing their markets fast, while Intel is doing quite well, and can thus decide to keep on spending in R&D Feb 11 09:25:47 esbenh exactly thats what I meant Feb 11 09:26:14 not all embedded systems are that power critical Feb 11 09:26:26 take fx. in-vehicle systems Feb 11 09:26:58 their are clearly embedded systems, but does have the benefit of a significantly better power budget than fx. mobile phones. Feb 11 09:27:00 esbenh, okay I shouldn't talk, my first embedded linux project is still powered by an x86 part. Feb 11 09:27:20 embedded != small Feb 11 09:28:05 hm ka6sox btw. Feb 11 09:28:10 or good night? Feb 11 09:28:14 Do nokia already announced which platform they will be using? Feb 11 09:28:33 funny thing is that we had a 10" touchscreen and an x86...still works 12yrs later Feb 11 09:28:35 I fear they will choose Win Phone Feb 11 09:28:37 M$ Feb 11 09:28:45 they did Feb 11 09:28:55 sh$t Feb 11 09:29:25 Choosing my next smartphone just go slightly more challenging Feb 11 09:29:30 esbenh, I guess with the rush to handheld devices I forgot the non-power critical sector. Feb 11 09:29:33 yeah.. Feb 11 09:29:40 seems HTC.. Feb 11 09:29:40 perhaps an HP WebOS phone... Feb 11 09:29:50 or HP, yes Feb 11 09:30:01 Jay7: http://www.engadget.com/2011/02/11/nokia-and-microsoft-enter-strategic-alliance-on-windows-phone-b/ Feb 11 09:30:10 having played with the new ones last night they are better than before. Feb 11 09:30:14 (the HP ones) Feb 11 09:30:27 ka6sox: I am mainly working on deeply-embedded devices for industrial usage. Typically without any input devices and with lots of power. Feb 11 09:31:00 esbenh, I have to make things work on 1ma of power if I can. Feb 11 09:31:27 its a challenge to make a logger that can go for 6months on drycells. Feb 11 09:35:19 the Pre3 has a better keyboad and they fixed the slider Feb 11 09:35:40 the "veer" is almost small enough to call a "watch phone" Feb 11 09:42:04 I just hope that we will see danish operator support for WebOS phones... Feb 11 09:42:48 does O2 operate in Denmark? Feb 11 09:42:59 nope Feb 11 09:43:28 we have: TDC, Telenor, Telia and 3 Feb 11 09:43:45 so mainly nordic based operators :-( Feb 11 09:43:50 ah Feb 11 09:44:08 I'll have to ask Sconix what carrier he is on... Feb 11 09:44:12 he has a pre2 Feb 11 09:44:19 but he is north of you. Feb 11 09:44:42 03Koen Kooi  07org.openembedded.dev * r98ceb8ba47 10openembedded.git/recipes/rfkill/ (2 files in 2 dirs): Feb 11 09:44:42 rfkill: fix install and remove bogus DEPENDS, it's not using libnl nor pkgconfig(-native) Feb 11 09:44:42 Signed-off-by: Koen Kooi Feb 11 09:50:10 woglinde, I should be in bed but I have a term wrong in a VHDL project and I'm chasing the logic...clockcycle by clockcycle... Feb 11 09:51:03 ka6sox oha Feb 11 09:51:29 the sim is missing it... Feb 11 09:51:46 porr sim Feb 11 09:52:34 ah right Feb 11 09:52:36 ups Feb 11 09:54:30 its painful... Feb 11 09:54:58 thats why I stay away from vhdl Feb 11 09:55:06 or used mitrion-c Feb 11 09:55:19 not a system-C fan. Feb 11 09:55:42 mitrion-c has nothing to do with c Feb 11 09:55:58 I'm pretty sure is a 1 cycle glitch but I can't find the condition. Feb 11 09:56:04 its a functional programming language which abstract a lot of stuff Feb 11 09:56:16 and produces vhdl file in the end Feb 11 09:56:16 ah Feb 11 09:56:47 I like vhdl very much but sometimes the tools I think leave a lot to be desired. Feb 11 09:57:07 ughs Feb 11 09:57:16 they gone bankrot Feb 11 09:57:25 pretty much Feb 11 09:57:42 synthesis seems to be a black art. Feb 11 09:58:08 it was really fun to programm in it Feb 11 09:58:50 the problem I'm having is a deadlock. so thats bad. Feb 11 10:00:01 I'm trying to synthesize a DP RAM part I can no longer get. Feb 11 10:00:23 http://www.mitrionics.com/?page=FPGAs_in_HPC_5 Feb 11 10:03:11 I give up for tonight...this rowlock deadlock is giving me fits. I guess I could just make it non-blocking and be done with it. Feb 11 10:03:53 nn Feb 11 10:07:20 gn ka6sox Feb 11 10:08:18 thanks, I wanted it identical but I'm just going to make it "better" and hope its not too fast for the app. Feb 11 10:27:51 good morning Feb 11 10:28:27 http://linuxconfau.blip.tv/file/4719842/ Feb 11 10:48:19 Looks like Elop is right though, the man has jumped off of the burning platfform, but what he didn't mention is that the sea is cold and deep.. and the man can't swim, so now the man will sink.. *G* Feb 11 10:50:34 nice effect of using lxc is ability to freeze building process at any time Feb 11 10:51:55 NOTE: package lzma-native-4.65-r5.1: task SRC_DISTRIBUTECOMMAND: Failed Feb 11 10:51:57 ERROR: Function 'SRC_DISTRIBUTECOMMAND' failed (see /dev/null for further information) Feb 11 10:51:59 hehe.. Feb 11 10:52:08 may be other nice effect of unfreezing.. Feb 11 10:52:44 but I've seen this already when was doing build with 12 bb threads Feb 11 11:06:20 Jay7: that's my old friend ;) Feb 11 11:07:20 http://lists.linuxtogo.org/pipermail/openembedded-devel/2010-October/026076.html Feb 11 11:07:50 but the /dev/null is recent stuff iirc Feb 11 11:08:01 I saw it first 2 weeks ago Feb 11 12:33:29 Hello ... Feb 11 12:33:50 we've built a sdk and we're having issues using pkg-config on it Feb 11 12:34:19 the .pc file seems fine but prefix and exec_prefix do not have the "sdk path" defined Feb 11 12:38:55 hi obi Feb 11 12:39:40 hi! Feb 11 12:44:42 otavio: libtool 2.4? Feb 11 12:45:12 eFfeM_work: it seems PKG_CONFIG_SYSROOT_DIR is not working Feb 11 12:45:14 03Andreas Oberritter  07master * r7ae30cf534 10openembedded.git/recipes/libdca/libdca_0.0.5-3.bb: Feb 11 12:45:15 libdca-0.0.5-3: initial recipe (Debian) Feb 11 12:45:15 Signed-off-by: Andreas Oberritter Feb 11 12:45:25 otavio: strange, did not see that yet Feb 11 12:45:27 03Andreas Oberritter  07master * r53b3d8cd38 10openembedded.git/recipes/libmms/libmms_0.6.bb: Feb 11 12:45:27 libmms-0.6: initial recipe Feb 11 12:45:27 Signed-off-by: Andreas Oberritter Feb 11 12:45:46 ...ded-linux/openembedded/recipes/libinih% echo $PKG_CONFIG_SYSROOT_DIR Feb 11 12:45:46 /opt/sdk/osnetboot-terminal/i586/i586-oe-linux Feb 11 12:46:03 BUT Feb 11 12:46:18 ...ded-linux/openembedded/recipes/libinih% pkg-config inihcpp --variable=prefix Feb 11 12:46:23 empty Feb 11 12:47:37 pkg-config 0.23 Feb 11 12:49:52 hm, strange, i don't think I have build an sdk recently so can't say too much about it Feb 11 12:50:44 eFfeM_work: it seems it is ignoring the sysroot env Feb 11 12:51:16 in libtool 2.4 there is an = or ~ or so that should be before the path in the .pc file if I recall correctly Feb 11 12:54:38 03Andreas Oberritter  07master * rd9c70aed82 10openembedded.git/recipes/upstart/upstart.inc: Feb 11 12:54:38 upstart: create directory before installing files into it Feb 11 12:54:38 * do_install installs files into ${sysconfdir}/event.d, so the Feb 11 12:54:38 directory must be created first. Feb 11 12:54:38 Signed-off-by: Andreas Oberritter Feb 11 12:54:48 03Andreas Oberritter  07master * rc08301653c 10openembedded.git/recipes/linux/linux.inc: Feb 11 12:54:49 linux: don't patch unifdef.c if it doesn't exist Feb 11 12:54:49 * Fixes build with Linux 2.6.18. Feb 11 12:54:49 Signed-off-by: Andreas Oberritter Feb 11 12:55:12 obi wohaha 2.6.18 Feb 11 12:55:31 yes, unfortunately Feb 11 12:55:42 but i'm working on 2.6.38-rc ;) Feb 11 12:55:54 hehe Feb 11 12:56:03 what device? Feb 11 12:57:13 all broadcom based dreamboxes. dm500hd, dm800, dm800se and dm8000 Feb 11 12:57:49 03Koen Kooi  07org.openembedded.dev * rd2a74c0c9b 10openembedded.git/recipes/xbmc/libmms_0.5.bb: libmms 0.5: move to libmms dir Feb 11 13:00:17 oh, i didn't notice that existing recipe Feb 11 13:00:25 are they still sold? Feb 11 13:01:01 of course :) Feb 11 13:01:53 those are the most recent models Feb 11 13:02:30 cbrake: will there be a testing-next this week or are we suspending that in favour of testing-release-2011-03 Feb 11 13:02:57 (in the latter case I need to stop the autobuild of testing-next) Feb 11 13:03:08 eFfeM_work we are suspending in favour of testing-release .... Feb 11 13:03:11 hm with 2.6.38 they are really cool Feb 11 13:03:13 eFfeM_work: I'll send a note to the list Feb 11 13:03:14 ok Feb 11 13:06:51 any reason why we didn't update to newer pkgconfig? Feb 11 13:07:00 0.25 was released on march Feb 11 13:10:37 dont ask me Feb 11 13:10:40 ask on the ml Feb 11 13:11:48 otavio: probably because no one cared to update (but maybe e.g. khem would know) Feb 11 13:12:10 otavio: you can ofc add 0.25 locally, test and submit a patch :-) Feb 11 13:13:21 khem: ^ Feb 11 13:13:35 otavio: for what it is worth: poky is on 0.25 Feb 11 13:13:37 khem: did you ever see the problem with PKG_CONFIG_SYSROOT_DIR? Feb 11 13:27:58 03Steffen Sledz  07org.openembedded.dev * rb295a16b5c 10openembedded.git/contrib/testing/testscript.sh: Feb 11 13:27:58 testscript.sh: be able to use branches other than testing-next Feb 11 13:27:58 Signed-off-by: Steffen Sledz Feb 11 13:27:58 Acked-by: Paul Menzel Feb 11 13:27:58 Acked-by: Frans Meulenbroeks Feb 11 13:34:51 anyone remember Stefan Schmidts nick? Feb 11 13:35:25 stefan_schmidt? Feb 11 13:35:31 you are welcome crfoton Feb 11 13:36:02 woglinde, thanks Feb 11 13:36:06 I thought that was the case Feb 11 13:36:25 khem: patchwork seems to ignore some patches with utf8 encoding. probably an unfortunate situation for Eric Bénard ;). there was a similar problem with the previous installation and i think there was a patch available to fix this Feb 11 13:39:37 khem: maybe this could help: http://patchwork.ozlabs.org/patch/78682/ Feb 11 13:40:25 obi are you working on upstart? Feb 11 13:40:30 newer version? Feb 11 13:40:52 yes Feb 11 13:40:54 0.6.7 Feb 11 13:41:05 cool Feb 11 13:41:06 okay Feb 11 13:41:13 one task I can delete from my list Feb 11 13:41:28 :) Feb 11 13:44:14 gm kergoth Feb 11 13:44:56 hey Feb 11 14:07:19 woglinde: lol Feb 11 14:07:55 khem: also, otavio's replies to patches 521 and 527 didn't show up in patchwork, maybe because the mailinglist was CC'd only Feb 11 14:09:07 cc' ml is bad Feb 11 14:09:21 hm gtk3.0 released Feb 11 14:09:24 woglinde: why? :) Feb 11 14:09:52 mail is from ml Feb 11 14:09:57 so reply to ti Feb 11 14:10:05 and cc whom ever you want Feb 11 14:10:40 no, i sent the patch to the ml with cc to otavio Feb 11 14:10:55 so he got it doubled Feb 11 14:11:02 and answered the wrong one Feb 11 14:11:06 * woglinde runs Feb 11 14:11:38 i guess that mailman doesn't send mails to recipients already on cc Feb 11 14:11:53 however... Feb 11 14:12:16 if the mailman archive gets the reply, then patchwork should get it too ;) Feb 11 14:12:41 everything else doesn't feel right Feb 11 14:13:36 yes Feb 11 14:13:45 looks like a bug in patchwork Feb 11 14:13:55 and should be easy to fixing Feb 11 14:24:21 hi GNUtoo|laptop & stefan_schmidt Feb 11 14:24:30 eFfeM_work, hi Feb 11 14:24:57 hi eFfeM_work, GNUtoo|laptop Feb 11 14:25:04 hi Feb 11 14:53:59 re jay7 Feb 11 15:04:04 RP: http://cgit.openembedded.net/cgit.cgi/bitbake/commit/?id=3372d84fa90ba51e8b2cd4557ee5ce501d0c6bf7 is likely to conflict with http://cgit.openembedded.net/cgit.cgi/openembedded/commit/?id=2bdfe8519eda8067845019a699acdf19a21ba380 Feb 11 15:11:53 Question: Can I modify u-boot environment that boot from MMC card would be disabled? I changed bootcmd to "run nandboot", but it still loads everything and boots from MMC. **** BEGIN LOGGING AT Fri Feb 11 15:15:53 2011 Feb 11 15:31:24 obi: Not conflicting exacty, it needs to support "no" though Feb 11 15:31:41 woglinde: hi Feb 11 15:32:23 woglinde, JaMa|Off_: You discussed about gnu/stubs-32.h in an icedtea build yesterday? Feb 11 15:33:23 woglinde, JaMa|Off_: I now have a report and log file here that shows a compile error with this header file missing on a amd64 ubuntu 10.10 system Feb 11 15:33:30 smells a bit Feb 11 15:33:47 Especially as I do all my builds on 64bit and never run into it Feb 11 15:42:13 03Andreas Oberritter  07master * rb9e21ac5e6 10openembedded.git/recipes/wpa-supplicant/wpa-supplicant-0.7.3/ (5 files): Feb 11 15:42:13 wpa-supplicant-0.7.3: remove unused files Feb 11 15:42:13 Signed-off-by: Andreas Oberritter Feb 11 15:42:13 Acked-by: Stefan Schmidt Feb 11 15:42:21 03Andreas Oberritter  07master * r5bf221337b 10openembedded.git/recipes/wpa-supplicant/wpa-supplicant-0.7.inc: Feb 11 15:42:21 wpa-supplicant-0.7.3: enable madwifi support Feb 11 15:42:21 Signed-off-by: Andreas Oberritter Feb 11 15:42:21 Acked-by: Stefan Schmidt Feb 11 15:42:23 03Andreas Oberritter  07master * refb86c94ef 10openembedded.git/recipes/wpa-supplicant/wpa-supplicant-0.7.inc: Feb 11 15:42:23 wpa-supplicant-0.7.3: depend on openssl Feb 11 15:42:23 * let defconfig decide which crypto library to use Feb 11 15:42:23 Signed-off-by: Andreas Oberritter Feb 11 15:42:23 Acked-by: Stefan Schmidt Feb 11 15:42:25 03Andreas Oberritter  07master * r33d60583aa 10openembedded.git/recipes/wpa-supplicant/ (wpa-supplicant-0.7.3/defconfig-gnutls wpa-supplicant-0.7.inc): Feb 11 15:42:25 wpa-supplicant-0.7.3: rename defconfig-gnutls to defconfig Feb 11 15:42:25 * Use a more useful name for overrides in case gnutls is not to be used. Feb 11 15:42:25 Signed-off-by: Andreas Oberritter Feb 11 15:42:29 Acked-by: Stefan Schmidt Feb 11 15:42:36 03Andreas Oberritter  07master * r00b6dea99b 10openembedded.git/recipes/wpa-supplicant/wpa-supplicant-0.7.inc: Feb 11 15:42:36 wpa-supplicant-0.7.3: install only required files for dbus support Feb 11 15:42:36 Signed-off-by: Andreas Oberritter Feb 11 15:42:36 Acked-by: Stefan Schmidt Feb 11 15:45:38 03Andreas Oberritter  07master * r2fcf3125c7 10openembedded.git/recipes/busybox/ (busybox.inc busybox_1.1x.inc files/inetd files/inetd.conf): (log message trimmed) Feb 11 15:45:38 busybox: create busybox-inetd package Feb 11 15:45:38 * Empty package unless CONFIG_INETD=y. Feb 11 15:45:38 * Use update-alternatives with priority 50. Feb 11 15:45:38 * Install initscript and inetd.conf. Feb 11 15:45:38 * RPROVIDE inetd Feb 11 15:45:39 Signed-off-by: Andreas Oberritter Feb 11 15:46:27 busy busy morning :) Feb 11 15:46:41 03Tom Rini  07org.openembedded.dev * r3c52ace035 10openembedded.git/ (8 files in 3 dirs): (log message trimmed) Feb 11 15:46:41 postgresql: Update from 8.4.4 to 8.4.7, drop 8.1.x, 8.2.4 Feb 11 15:46:41 We update SHR to pinning 8.4.7 (safe upgrade) and drop Feb 11 15:46:41 8.1.x (EOL) and 8.2.x (un-pinned and EOL this year). We also Feb 11 15:46:41 un-pin SHR (at Martin's suggestion). Feb 11 15:46:41 Signed-off-by: Tom Rini Feb 11 15:46:42 Acked-by: Khem Raj Feb 11 16:21:34 Hi. I'm using OE on a i686 host and building for a i585 target. Is a path like tmp/sysroots/i686-linux/usr/i586/i586-oe-linux/ expected? Or did I screw my tmpdir up? Feb 11 16:27:34 No, that's expected Feb 11 16:28:01 It means $TMPDIR/sysroots/$HOST_SYS/usr/$TARGET_ARCH/$TARGET_SYS Feb 11 16:28:12 i586-oe-linux is what your i586 target is called Feb 11 16:28:39 And we put things like the cross toolchain there so that actual target packages (busybox) can be shared between say 32bit and 64bit hosts Feb 11 16:28:50 But you would still get a rebuild of the cross tools Feb 11 16:30:33 I see. Thanks for the info, Tartarus. I'm trying to build chicken eggs (http://call-cc.org) on OE. It seems to be a bit tricky. Feb 11 16:30:46 hmm. Feb 11 16:31:27 thinking about it, i don't think i like the new persist_data api fully.. its an improvement, certainly, but i think we can drop one layer of indirection cleanly Feb 11 16:31:31 hmmmm Feb 11 16:33:46 One of the tricky parts is that chicken is a Scheme->C compiler, and it has a compiler driver which compiles Scheme->native code, transparently, using a C compiler behind the scenes. Feb 11 16:34:28 Oh my, heh Feb 11 16:34:38 is the Scheme->C compiler a binary or a perl script/etc? Feb 11 16:35:07 It is a binary, compiled from Scheme using chicken itself. Feb 11 16:38:40 good luck Feb 11 16:38:47 So, I need to use the Scheme compiler from the cross tools area (I've successfully build chicken and chicken-cross) and make it link agains the libs from the target area. Chicken provides a tool (chicken-install) which is more or less like bitbake (oe emerge): it fetches sources, builds and installs exptensions (which are dynamically linked libraries). Feeding chicken-install the right paths is my current problem (I GUESS). Feb 11 16:39:11 heh, I remember chicken well from my days as a debian-arm porter. Feb 11 16:39:54 pb_: do you use chicken? Or were you just packaging it? Feb 11 16:42:11 hiho pb Feb 11 16:44:03 hi woglinde Feb 11 16:44:18 pb whats with your house? Feb 11 16:44:23 mario-goulart: just packaging it, there were a bunch of things in debian that needed it Feb 11 16:44:30 crofton asked it too between Feb 11 16:44:43 woglinde: screeding the floors on monday, electricians nearly finished. Feb 11 16:45:10 ah Feb 11 16:45:15 so you see an end Feb 11 16:45:15 pb_: Cool. Do you remember what version was it when you were packaging chicken? Feb 11 16:45:32 pb will you install some solarpanels? Feb 11 16:46:27 woglinde: yes, at some point. the heating system is set up for them but we don't actually have the panels yet. Feb 11 16:47:00 ah so only for heating not for electricity Feb 11 16:47:20 possibly electricity as well, that's easy to fit later. Feb 11 16:47:35 but, right now the solar PV ones are still quite expensive for what they do. Feb 11 16:47:48 yes Feb 11 16:48:01 you need 15 years to get break event Feb 11 16:48:08 heating-wise, we have a 1000 litre accumulator tank in the utility room which is ready to receive input from the panels. :-} Feb 11 16:48:28 pb_, have wookey come over and make you a debian controller :) Feb 11 16:48:29 mario-goulart: sorry, I don't recall. this must have been nearly 10 years ago now. Feb 11 16:48:35 03Martin Jansa  07master * rd917bd637e 10openembedded.git/recipes/xorg-app/ (3 files in 2 dirs): Feb 11 16:48:35 xkbcomp: upgrade from 1.2.0 to 1.2.1 Feb 11 16:48:35 Signed-off-by: Martin Jansa Feb 11 16:49:14 pb_: so probably very early chicken versions. Very interesting. :-) Feb 11 16:49:47 1.x, I'd guess. Feb 11 17:11:43 03Chris Larson  07master * r34c7ff67d7 10bitbake.git/lib/bb/ (4 files in 3 dirs): Feb 11 17:11:43 persist_data: drop SQLData indirection Feb 11 17:11:43 The common case (if not the only case) is to only use a single domain. The Feb 11 17:11:43 only reason SQLData exists is to make it easier to delete a domain. Yet, Feb 11 17:11:43 there's no need for us to delete a domain if SQLTable knows how to clear Feb 11 17:11:43 itself out. So, add clear() to the table and pass the domain to persist(). Feb 11 17:11:44 Signed-off-by: Chris Larson Feb 11 17:11:46 03Chris Larson  07master * rc197043717 10bitbake.git/lib/bb/persist_data.py: Feb 11 17:11:46 persist_data: use better items/values methods for SQLTable Feb 11 17:11:46 Signed-off-by: Chris Larson Feb 11 17:15:35 what's the policy regarding wifi card firmwares? Feb 11 17:15:50 can we provide a downloader script in postinstall? Feb 11 17:18:04 nah, chicken and egg problem Feb 11 17:24:40 downloader script seems perfectly reasonable to me Feb 11 17:25:20 well not without wifi Feb 11 17:25:57 hmm, yes, guess it could be conditional on also having ethernet at the time of install, but ick, that is ugly indeed :\ Feb 11 17:26:37 obi_: I have patched the patchwork should work better with utf hereon Feb 11 17:26:44 * mwester seems to recall that the zd wifi firmware is downloaded and packaged in OE, and made a dependency of the zd kernel module (at least for the IXP4XX kernels) Feb 11 17:26:47 thanks for reporting Feb 11 17:26:57 hmmmmm Feb 11 17:35:53 hi all Feb 11 18:35:36 hey ho .. anybody ever brought the camera interface on a ompa3530 to life ? expierence ? Feb 11 18:35:46 not I Feb 11 18:38:59 khem ping Feb 11 18:39:13 woglinde: yes Feb 11 18:39:34 khem uclibc git update? Feb 11 18:39:56 make and memoryleak fix? Feb 11 19:00:03 woglinde: will do Feb 11 19:00:21 woglinde: today or tomorrow Feb 11 19:16:12 03Denys Dmytriyenko  07master * r92caf6e4ce 10openembedded.git/recipes/udev/ (udev-141/init udev_141.bb): Feb 11 19:16:12 udev-141: backport the initscript depmod call from recent versions Feb 11 19:16:12 Signed-off-by: Denys Dmytriyenko Feb 11 19:48:25 anybody wants to ack this? http://thread.gmane.org/gmane.comp.handhelds.openembedded/42484 Feb 11 19:50:09 denix: Böhmische Dörfer Feb 11 19:51:12 :) Feb 11 19:51:43 btw, thanks for the lesson in German... :) Feb 11 19:53:01 hi denix and nschle Feb 11 19:53:26 woglinde: hey Feb 11 19:54:21 denix, i hope you can google "Böhmische Dörfer" Feb 11 19:55:50 denix: http://de.wikipedia.org/wiki/B%C3%B6hmisches_Dorf Feb 11 19:56:41 denix: sorry there is only a german entry: This means: I do not understand anything Feb 11 19:57:33 woglinde: hi you are the Navit , I am right ? Feb 11 19:58:40 evening Feb 11 19:59:09 first run of release testing is done Feb 11 19:59:14 let's see Feb 11 20:00:09 angstrom-2008.1/collie was failed on lzma Feb 11 20:00:25 tosa was failed on lzma as well Feb 11 20:00:35 but with angstrom-2010.x Feb 11 20:01:12 SRC_DISTRIBUTECOMMAND Feb 11 20:01:31 hm.. Feb 11 20:01:52 does anybody know ( i am to foul) who maintains E17 (Illime ...) Feb 11 20:02:00 (illunme) Feb 11 20:02:06 nschle85: try to ask JaMa|Off Feb 11 20:02:19 JaMa|Off: is off Feb 11 20:02:22 but look into MAINTAINERS file before Feb 11 20:02:27 nschl jama is the maintainer Feb 11 20:02:41 but I am contributor to navit it self Feb 11 20:03:13 woglinde: i asked about E17 not navit Feb 11 20:03:33 nschle85: yes, I got it the first time :) Feb 11 20:04:10 woglinde: JaMa|Off is the maintainer of E17 ? Feb 11 20:05:29 nschle85: are you looking for E17 maintainer or E17 recipes maintainer? :) Feb 11 20:07:03 denix: E17 recipes maintainer first. Feb 11 20:13:18 nschle85: http://pastebin.com/8xbHLuRU :) Feb 11 20:17:12 I istalled oe and bitbake in /oe. After carrying out "bitbake nano" as described in "getting started", it show permission denied. I tried sudo bitbake nano, it show "sudo: bitbake: command not found", then, I type "sudo -s", then "bitbake nano", it finally end up seems doing something. At the end of process, I got FAtal: Do not use Bitbake as root. what is going on here? Feb 11 20:17:19 denix: i undestood Feb 11 20:17:35 bing never run it with root rights Feb 11 20:17:48 denix: ok now I need a E17 Maintainer Feb 11 20:18:04 lol Feb 11 20:18:30 because it is in /oe, if I do'nt run it as root, how should I run it? Feb 11 20:18:53 bing chmod is your friend Feb 11 20:18:59 ok Feb 11 20:19:24 thanks, trying it again Feb 11 20:20:04 args Feb 11 20:20:05 the new arm atomics are not compatible with thumb mode Feb 11 20:20:06 chown Feb 11 20:20:15 khem gcc 4.6? Feb 11 20:20:21 woglinde: no uclibc Feb 11 20:20:31 we moved the nptl atomics over to be generic Feb 11 20:20:43 and now I see that the implementation on works on thumb2 and arm modes Feb 11 20:20:53 not on thumb1(old thumb) mode Feb 11 20:20:58 in OE thats what we use Feb 11 20:21:05 for <=armv5 arches Feb 11 20:21:13 hmmm Feb 11 20:21:23 I guess I need to implement them Feb 11 20:21:31 nschle85: http://www.enlightenment.org/p.php?p=contact Feb 11 20:21:42 khem hm Feb 11 20:21:51 now update the git rev Feb 11 20:21:57 *g* Feb 11 20:22:06 nschle85: raster used to hang out here before, haven't seen him for quite some time... Feb 11 20:22:19 woglinde: that guy was so pushy in getting that change in and I was not able to test it Feb 11 20:22:28 blindvt pushed it anyway Feb 11 20:22:32 and here we are Feb 11 20:22:43 hey denix Feb 11 20:22:48 howdy Feb 11 20:22:58 attending ELC this year ? Feb 11 20:22:58 khem so release stalled Feb 11 20:23:11 woglinde: yeah I have not blasted the bomb yet Feb 11 20:24:03 even if I use blx its going to break for armv4t Feb 11 20:24:23 * khem thinks of abandoning thumb1 mode Feb 11 20:24:36 * mwester likes thumb. Feb 11 20:24:47 heh Feb 11 20:25:09 mwester: not in general but only for compiling uclibc may be a compromise Feb 11 20:25:22 * khem is tired of fixing thumb now Feb 11 20:25:24 ah, ok. Feb 11 20:25:58 uclibc itself will be only few kb in size difference between arm and thumb Feb 11 20:26:13 I wonder if its worth keeping the thumb port uptodate Feb 11 20:26:15 khem: planning to attend, hopefully both ELC and ELCE... :) Feb 11 20:26:16 mwester you dont like thumb1 Feb 11 20:26:19 khem: you? Feb 11 20:26:36 denix: I will attend ELC USA but probably not ELCE yet Feb 11 20:26:56 I am still alien in US and leaving and coming back is a bit hassle Feb 11 20:26:57 I will go to elce Feb 11 20:27:03 isnt that far from berlin Feb 11 20:27:16 woglinde: where is it next time Feb 11 20:27:17 and we have a train to prauge Feb 11 20:27:20 woglinde: prague, afaik Feb 11 20:27:20 args Feb 11 20:27:28 thanks denix Feb 11 20:27:28 hmm I have been to Prague Feb 11 20:27:29 woglinde: elce ? Feb 11 20:27:37 bribed two policemen on the way Feb 11 20:27:50 khem: :) Feb 11 20:28:06 yeah I drove in a Car with German license place Feb 11 20:28:10 khem: are you on GC or H1? Feb 11 20:28:15 H1 still Feb 11 20:28:18 http://events.linuxfoundation.org/events/embedded-linux-conference-europe Feb 11 20:28:22 but GC in final stages Feb 11 20:28:54 khem: good luck! it's becoming more difficult every year... Feb 11 20:28:55 denix: may be this summer I should get GC Feb 11 20:29:03 denix: yeah I know Feb 11 20:29:07 I applied in 2006 Feb 11 20:29:24 and right now I have 485 filed Feb 11 20:29:24 hm 350$ Feb 11 20:29:30 mine took 6 years... Feb 11 20:29:32 I have to think oit over Feb 11 20:29:49 denix: oh ok then I still have till 2012 :) Feb 11 20:30:07 gc? Feb 11 20:30:22 woglinde: green card - permanent residency in us Feb 11 20:30:22 denix: did you emigrate from a *friendly* country Feb 11 20:30:31 gc= green card Feb 11 20:30:55 I friend of mine emigrated from Ukraine and it was not fun Feb 11 20:31:06 his application sat in FBI check for 2 years :) Feb 11 20:31:13 khem: Russia/Ukraine... Feb 11 20:31:20 there you go Feb 11 20:31:26 my friend is from Kiev Feb 11 20:31:28 khem: driving in a cheap german car in czech is very "difficult" Feb 11 20:31:39 hmm.. Feb 11 20:31:46 ELC-E this time in Prague Feb 11 20:31:49 that is fine Feb 11 20:31:52 nschle85: I had a cheap Opel :) Feb 11 20:32:19 I hope this time visa application will be easier than UK's one :) Feb 11 20:32:42 yeah this visa stuff is hassle too Feb 11 20:32:46 jay7 350$ Feb 11 20:32:51 if paying now Feb 11 20:32:53 it costs like 400-500 USD here Feb 11 20:32:56 form may 400 Feb 11 20:33:02 khem: ok you should drive Zafira, its a family car, not vectra :-) Feb 11 20:33:03 from august 550 Feb 11 20:33:06 khem: russians are stuck in fbi checks, indians are stuck in huge lines - many of my indian office-mates complain... Feb 11 20:33:21 yes I am in line Feb 11 20:33:32 woglinde: I hope I'll go with presentation :) Feb 11 20:33:45 jay7 about? Feb 11 20:33:52 kexecboot! ;) Feb 11 20:33:56 Jay7: even with presentation you will have to pay .eu to enter their borders Feb 11 20:34:01 Jay7: way to go! :) Feb 11 20:34:27 I've had Indian coworkers who were denied entry into Italy for customer visits.. (these are perminent US residents who are attempting to get US citizen ship -- not even GC) Feb 11 20:34:31 I did not go to GCC summit for same reason as I have to go to LA for Canadian visa Feb 11 20:34:37 khem: yeah.. Feb 11 20:35:01 I hate all this travel documentation Feb 11 20:35:03 visa is shitty Feb 11 20:35:08 will never leave the eu Feb 11 20:35:36 woglinde: +1 to shitty visa :) Feb 11 20:35:52 Germans govt was nice they offerred me that if I keep coming to germany for work reasons they may extend my visa Feb 11 20:36:03 khem: my wife is from estonia, she visited me years ago (tallinn, stockholm, munich... then we wento to ski via poland to szech republic... boarder controls are funny :-) Feb 11 20:36:06 but last time I enterteed germany was 2006 Feb 11 20:36:29 nschle85: yeah Feb 11 20:36:33 khem: now you can drive opel here :-) Feb 11 20:36:39 gc's are now pink Feb 11 20:36:41 * Jay7 had chance to migrate to Germany one time.. Feb 11 20:36:43 althogh I did not have that bad experiences in western europe Feb 11 20:37:05 Crofton|work: they should rename them to PC :) Feb 11 20:37:26 Crofton|work: heh thats true Feb 11 20:37:55 Crofton|work: and its interesting that they check you for same old diseases that were prevalent in early 1900 Feb 11 20:38:13 and anyway it should be done when alien enters the contry Feb 11 20:38:20 not after being here for 2 yeard Feb 11 20:38:37 by the time if he has some communicable disease he already has spread it Feb 11 20:39:06 oh laws are funny Feb 11 20:39:25 03George Alexandru Dorobantu  07org.openembedded.dev * rbdc4fc779b 10openembedded.git/recipes/libtomcrypt/libtomcrypt_1.17.bb: Feb 11 20:39:25 libtomcrypt: updated download link Feb 11 20:39:25 * updated the download link; libtomcrypt has moved to google code and the old link is dead Feb 11 20:39:25 Signed-off-by: George Alexandru Dorobantu Feb 11 20:39:25 Signed-off-by: Tom Rini Feb 11 20:39:51 Tartarus: around Feb 11 20:39:57 pong Feb 11 20:40:12 03AJ ONeal  07org.openembedded.dev * r81e0297498 10openembedded.git/recipes/nodejs/ (5 files in 2 dirs): Feb 11 20:40:12 nodejs: updated to 0.2.6 Feb 11 20:40:12 * added version number to patches Feb 11 20:40:12 * libev-cross-cc_0.2.6.patch - turns off execution so that cross-compile succeeds Feb 11 20:40:12 * node-cross-cc_0.2.6.patch - enables cross-compiling of v8 for x86, x86_64, arm, and mips Feb 11 20:40:13 Signed-off-by: AJ ONeal Feb 11 20:40:14 Signed-off-by: Tom Rini Feb 11 20:40:15 Tartarus: Is it possible your autobuilder to kick in a build on say when a commit happens Feb 11 20:40:46 so may be keep polling with git pull --rebase every hour or so Feb 11 20:41:03 If we had a much smaller subset of targets, yes, I could try and set that up Feb 11 20:41:16 Tartarus: so say all qemus Feb 11 20:41:25 one distro doesnt matter which one Feb 11 20:41:44 Well, it depends on what we're going to be checking for Feb 11 20:42:01 yeah incremental build with proper build deps would be more efficient Feb 11 20:42:23 Tartarus: how often do u run the autobuilder now Feb 11 20:42:26 * fray notes he does "git remote update" and "git rebase ..." manually instead of letting pull do it Feb 11 20:42:32 03Denys Dmytriyenko  07master * rf447637269 10openembedded.git/recipes/udev/ (15 files in 6 dirs): Feb 11 20:42:32 udev: also use /proc/devices to invalidate devcache Feb 11 20:42:32 * Besides checking for kernel version, bootargs or cmdline changes between Feb 11 20:42:32 reboots, also consider changes in /proc/devices to invalidate device cache. Feb 11 20:42:32 Signed-off-by: Denys Dmytriyenko Feb 11 20:42:33 Acked-by: Koen Kooi Feb 11 20:42:39 khem, weekly, since I've pushed the config matrix up to, I think, 2 days worth of stuff Feb 11 20:42:45 I'll know for sure Monday how long it takes :) Feb 11 20:42:52 Tartarus: ok Feb 11 20:42:56 (206 combinations of stuff, and then each on 3 diff hosts) Feb 11 20:43:20 Tartarus: thats fairly large service Feb 11 20:43:25 Yes :) Feb 11 20:43:26 Tartarus: I have first results.. should fill table Feb 11 20:43:32 Jay7, yay Feb 11 20:43:47 big scary table.. Feb 11 20:43:52 I will start builds too but garnet is crippled Feb 11 20:43:53 Jay7, and start posting to the ML about failures? :) Feb 11 20:44:00 so need to share my dev laptop Feb 11 20:44:02 yes, exactly :) Feb 11 20:44:15 because failure starting to be very repeatable Feb 11 20:44:25 Jay7, repeatable is good Feb 11 20:44:25 I mean lzma-native Feb 11 20:44:39 http://tinderbox.openembedded.net/packages/1953751/ Feb 11 20:44:47 http://tinderbox.openembedded.net/packages/1947369/ Feb 11 20:45:36 Jay7: no logs ? Feb 11 20:45:44 khem: I have local Feb 11 20:45:57 so what does it report Feb 11 20:45:58 I'll paste, moment Feb 11 20:47:42 khem: http://pastebin.com/fam0YzkS Feb 11 20:47:49 Tartarus, I'd rather link the release stuff from the developer section Feb 11 20:47:50 * khem lunch Feb 11 20:47:51 this is bb log Feb 11 20:48:03 looking strange Feb 11 20:48:15 Crofton|work, yeah, i saw. I think so long as it's linked somewhere prominatly i don't have a big opinion Feb 11 20:48:25 ok Feb 11 20:48:35 distros are more than a collection of recipes :) Feb 11 20:52:55 Crofton|work: ???????? Feb 11 20:54:53 distros include toolchain policy, possibly binary feeds, support paths etc :) Feb 11 20:57:44 Crofton|work: puhhh, ok shr is a disribution :-) Feb 11 20:57:50 right Feb 11 21:13:30 http://www.openembedded.org/index.php/Main_Page Feb 11 21:13:32 thoughts Feb 11 21:18:00 looking Feb 11 21:18:24 looks good to me Feb 11 21:18:25 thanks Feb 11 21:22:18 np Feb 11 21:30:43 Jay7: that message is not helpful Feb 11 21:43:20 khem: sorry, I have no other Feb 11 21:52:06 Jay7, can you pastebin your local.conf? Feb 11 21:56:16 Tartarus: moment Feb 11 21:56:53 http://pastebin.com/SGue67kJ Feb 11 21:57:25 hm.. lot of vars from env Feb 11 21:58:20 Tartarus: if you need to know value of any variable - ask me :) Feb 11 21:58:52 03Filip Zyzniewski  07org.openembedded.dev * rc875f74890 10openembedded.git/recipes/initscripts/initscripts-1.0/devices: Feb 11 21:58:52 initscripts: handle busybox version of makedevs properly. Feb 11 21:58:52 Makedevs from busybox has a different command line syntax. Feb 11 21:58:52 Signed-off-by: Filip Zyzniewski Feb 11 21:58:52 Signed-off-by: Tom Rini Feb 11 21:58:52 Well, what's the other config file look like? Feb 11 21:58:56 or what distro/etc is being set? Feb 11 21:59:16 this is my testbuilder script from contrib/testing Feb 11 21:59:35 distro is changing Feb 11 21:59:50 angstrom-2008.1, angstrom-2010.x, minimal Feb 11 22:00:07 machines are collie, tosa+akita, efikamx, ben-nanonote Feb 11 22:00:10 and that only fails on minimal right? Feb 11 22:00:34 no, both on angstrom, but different versions and machines Feb 11 22:00:43 http://tinderbox.openembedded.net/builders/Jay7-tb/ Feb 11 22:00:44 look here Feb 11 22:00:49 um Feb 11 22:01:09 hm Feb 11 22:01:14 ant_ have seen this too Feb 11 22:01:39 Ah, there we go, my grep was off.. Feb 11 22:01:54 Is this a recent change? Feb 11 22:02:13 what exactly you mean? Feb 11 22:02:15 the failure Feb 11 22:02:20 btw, this is bitbake master Feb 11 22:02:35 hm.. may be last week Feb 11 22:02:44 ant_: when have you seen this first time? Feb 11 22:02:51 can you reproduce a passing case, without changing bitbake versions? Feb 11 22:02:57 This needs a bisect Feb 11 22:03:05 the obvious bits haven't changed in a while Feb 11 22:03:13 I'm not chaning bb versions, always is master here Feb 11 22:03:49 Yeah Feb 11 22:03:52 but presumably you update Feb 11 22:03:55 and master does change Feb 11 22:04:01 ah, yes Feb 11 22:04:05 So Feb 11 22:04:06 http://tinderbox.openembedded.net/builds/111288/ Feb 11 22:04:08 I've updated bb before build Feb 11 22:04:09 is when you had that build last Feb 11 22:04:18 Try current bitbake master and that OE Feb 11 22:04:20 see if it works or fails Feb 11 22:04:27 So we know where you need to git bisect :) Feb 11 22:04:50 (and you should just be able to -c src_distribute_local or so rather than building all of lzma-native, not that it's big to start with) Feb 11 22:05:47 Tartarus: check Jay7's builds also ;) Feb 11 22:06:29 so, minimal is affected too Feb 11 22:06:31 yes Feb 11 22:06:42 all that inherit src_distribute_local are broken for you Feb 11 22:06:42 http://tinderbox.openembedded.org/builds/123533/ Feb 11 22:06:48 but that code hasn't changed since march Feb 11 22:06:54 So... Any working case is fine Feb 11 22:07:02 You just need to run the test Feb 11 22:07:05 well, I'll try to build it Feb 11 22:07:08 So you know where to start off with git bisectr Feb 11 22:07:10 *bisect Feb 11 22:07:40 Have you used git bisect before? :) Feb 11 22:07:54 yeah.. Feb 11 22:08:50 k Feb 11 22:09:14 So you know it doesn't matter too much if you have the absolute last time it worked, just a working case, esp w/ a small testcase like this :) Feb 11 22:15:08 Jay7: Tartarus: it was last october: http://lists.linuxtogo.org/pipermail/openembedded-devel/2010-October/026076.html Feb 11 22:16:47 DISTRO=angstrom-2008.1 MACHINE=collie bitbake lzma-native Feb 11 22:17:01 got it Feb 11 22:17:19 will investigate Feb 11 22:17:30 yay :) Feb 11 22:17:42 yeah.. that was easy :) Feb 11 22:18:09 filip, ping Feb 11 22:19:25 this time w/o error Feb 11 22:19:41 what about grg's patch? Feb 11 22:19:43 race really Feb 11 22:20:00 -addtask distribute_sources before do_build after do_fetch Feb 11 22:20:02 +addtask distribute_sources before do_build after do_unpack Feb 11 22:23:07 But, er, what part of the code relies on the unpack part rather than fetch? Feb 11 22:23:47 there have been changes to unpack last year iirc Feb 11 22:24:13 yeah, but the src_distribute code I mean Feb 11 22:24:21 it's just for the fetched stuff Feb 11 22:24:25 Or does it also do patches? Feb 11 22:24:55 sure looks like it just does source archives+md5 Feb 11 22:25:19 fwiw sourc eis very small Feb 11 22:25:35 lzma465.tar.bz2 │ 288122 Feb 11 22:25:46 yeah distribute_sources doesn't use or require unpack, it operates directly against fetched items from SRC_URI Feb 11 22:26:15 how can that fix halp then? Feb 11 22:26:49 doubt it would do much of anything, personally Feb 11 22:26:50 * kergoth_ shrugs Feb 11 22:26:55 03Andreas Müller  07org.openembedded.dev * r47fd3c8109 10openembedded.git/recipes/clutter/ (clutter-1.4.2/fix-lib-location.patch clutter_1.4.2.bb): (log message trimmed) Feb 11 22:26:55 clutter: fix broken clutter-gtk Feb 11 22:26:55 * Since clutter 1.4.2 clutter-gtk fails for machines/archs configuring Feb 11 22:26:55 * flavour=eglx with (see clutter.inc) Feb 11 22:26:55 * Feb 11 22:26:55 * | Making all in clutter-gtk Feb 11 22:26:55 * | make[2]: Entering directory `/home/Superandi/tmp/work/armv7a-angstrom-linux-gnueabi/clutter-gtk-0.10.8-r2.0/clutter-gtk-0.10.8/clutter-gtk' Feb 11 22:27:05 besides, what's that obscure hash? Reverting a6a23f73 seems to make it Feb 11 22:27:05 disappear. Feb 11 22:27:05 03Andreas Müller  07org.openembedded.dev * rb4265bfdb1 10openembedded.git/recipes/clutter/clutter-gst_1.3.4.bb: (log message trimmed) Feb 11 22:27:05 clutter-gst add 1.3.4 Feb 11 22:27:05 * clutter 1.4.2 broke clutter-gst and made it necessary to have a recipe from Feb 11 22:27:05 * development snapshot sources Feb 11 22:27:06 * Feb 11 22:27:06 * Build tested with Feb 11 22:27:07 * BB_VERSION = "1.10.2" Feb 11 22:28:32 Tartarus: what's up? Feb 11 22:28:36 grg talked with kergoth about that Feb 11 22:28:37 http://ibot.rikers.org/%23oe/20101028.html.gz Feb 11 22:29:10 do_distribute_sources is started before do_unpack is finished Feb 11 22:29:18 00:12.23 grg do_fetch and do_unpack are racing in src_distribute... i cant see any reason why its causing failures, but it is Feb 11 22:29:37 filip, are you a jlime maintainer now? Feb 11 22:29:54 Tartarus: well according to jlime-2010.1.conf I've been him for quite a while Feb 11 22:30:01 Tartarus: but the main guy is kristoffer Feb 11 22:30:08 filip, heh, can you update MAINTAINER too please? :) Feb 11 22:30:31 Tartarus: he's been pretty busy with his med school lately, but it should get better soon Feb 11 22:30:34 That makes it easier to say jlime maintainer posting jlime patch, apply :) Feb 11 22:30:45 (I mean, add yourself) Feb 11 22:30:48 Tartarus: well I am there Feb 11 22:30:57 Tartarus: last maintainer line Feb 11 22:31:07 I mean ./MAINTAINERS Feb 11 22:31:19 http://pastebin.com/8VTY3xzA Feb 11 22:31:30 I think I'll ask kristoffer for that, it would be a bit arrogant to do it myself :P Feb 11 22:31:39 heh, ok Feb 11 22:31:51 btw, i'm going to move http://patches.openembedded.org/patch/485/ to rejected, based on the rest of the comments :( Feb 11 22:32:36 * ant_ still dreams about USEFLAGS Feb 11 22:32:40 Tartarus: sure, I disagree, but I understand the reason Feb 11 22:32:40 http://patches.openembedded.org/patch/555/ doesn't apply btw Feb 11 22:33:18 damn, I'll look into it tomorrow. Sorry. Feb 11 22:33:28 03Filip Zyzniewski  07org.openembedded.dev * r8ee80af14d 10openembedded.git/recipes/initscripts/initscripts-1.0/jornada7xx/ (checkroot devices): Feb 11 22:33:28 initscripts: removed jornada7xx specific scripts. Feb 11 22:33:28 Stock ones work better (remounting / rw, makedevs calling). Feb 11 22:33:28 Signed-off-by: Filip Zyzniewski Feb 11 22:33:28 Signed-off-by: Tom Rini Feb 11 22:33:30 probably mixed with my not-ready-yet stuff Feb 11 22:33:31 tomorrow's saturday it's not work time! Feb 11 22:33:40 demigod2k: jlime is my hobby :P Feb 11 22:34:05 filip, k, talk to you tomorrow Feb 11 22:34:46 the dpkg perl patch looks OK, I think Feb 11 22:34:56 just a little confused, moving stuff into -perl that normally was in the main package? Feb 11 22:35:06 it wasn't packaged at all Feb 11 22:35:07 hm.. who is setting SRC_DISTRIBUTECOMMAND? Feb 11 22:35:21 that was the reason for the patch Feb 11 22:36:04 ah, I see it in src_distribute_local.bbclass Feb 11 22:36:17 Jay7: it looks like the issue are the two run.base_do_fetch Feb 11 22:36:42 filip, ok Feb 11 22:36:44 thanks Feb 11 22:36:57 i'll wait and see if anyone else comments and push it tomorrow Feb 11 22:37:06 ant_: I suspect other.. SRC_DISTRIBUTECOMMAND is initialized at wrong time Feb 11 22:37:26 kergoth_: have you any idea about this? Feb 11 22:38:22 Jay7: i'm not seeing why its a problem that SRC_DISTRIBUTECOMMAND runs before do_unpack finishes.. Feb 11 22:38:49 red herring? Feb 11 22:38:51 problem that it is uninitialized at this time... i.e. var is not set as I understand Feb 11 22:40:05 I can't understand at which time initialization happens.. Feb 11 22:40:48 i'm not understanding what's confusing about it Feb 11 22:41:04 if you inherit src_distribute_local, SRC_DISTRIBUTECOMMAND is set at finalize time by an anonymous python function Feb 11 22:42:08 Ah ha? Feb 11 22:42:13 Stupid annoying anon python functions? Feb 11 22:42:37 Since they're seeing it not be set, and be tried as that literal Feb 11 22:42:39 still not seeing how this is a problem. they're always run before executing the task itself Feb 11 22:42:51 Tartarus: yes, looks like Feb 11 22:43:01 kergoth_, but they're seeing the literal SRC_DISTRIBUTECOMMAND try and be executed Feb 11 22:43:06 And that change moves the race around Feb 11 22:43:10 htf can that happen? Feb 11 22:43:26 no idea, but anonymous python functions aren't magic Feb 11 22:44:24 may be you have idea how to debug this? Feb 11 22:44:27 also, if SRC_DISTRIBUTECOMMAND is unset when executing the distribute sources function, it will immediately error out talking about it Feb 11 22:44:28 Jay7: btw, do you have a .pc dir after rm_work in lzma-native workdir? Feb 11 22:44:49 kergoth_: http://tinderbox.openembedded.net/packages/1947369/ Feb 11 22:44:54 as you see.. Feb 11 22:46:00 ant_: right now I have tree after that error Feb 11 22:46:01 i'm not seeing much. is there a log somewhere there that i'm missing? Feb 11 22:46:09 so, no rm_work was done yet Feb 11 22:46:21 kergoth_: no... log is /dev/null Feb 11 22:46:34 all it says is that it failed to run SRC_DISTRIBUTECOMMAND, not *WHY* Feb 11 22:46:39 run.SRC_DISTRIBUTECOMMAND.25212 Feb 11 22:46:40 so i'm not seeing how we're all assuming its undefined Feb 11 22:46:56 0 size.. Feb 11 22:47:17 ah.. stop.. there is run.do_distribute_sources.25212 Feb 11 22:48:05 now, i do see one obvious problem when looking at the class. it iterates over the urldata, but this can't be relied upon, which is why do_unpack doesn't do it anymore Feb 11 22:48:16 it should be changed to iterate over the SRC_URI and get the urldata of each url, the way do_unpack does Feb 11 22:48:52 kergoth_: any quick hack to test? Feb 11 22:48:53 the urldata method acts very oddly when overrides of SRC_URI are involved, it ends up holding them all Feb 11 22:49:00 just copy/paste the lines from do_unpack Feb 11 22:49:20 its a simple loop over the urls to get the urldata, nothing to it more than a couple lines of code you can copy over Feb 11 22:49:29 I bet that's it then Feb 11 22:49:35 lzma does have a semi special SRC_URI Feb 11 22:50:25 i don't see why this would cause this particular problem, but who knows, its the obvious problem that exists that should be fixed, whether or not it resolves this Feb 11 22:52:21 (the urldata being user/recipe visible is one of the problems that should really be fixed in bb.fetch2, if it isn't yet, it's quite ugly) Feb 11 22:52:27 * kergoth_ thinks Feb 11 22:52:46 i.e. like this? Feb 11 22:52:47 -for url, urldata in urldatadict.iteritems(): Feb 11 22:52:49 +for url in src_uri: Feb 11 22:52:50 + urldata = srcurldata[url] Feb 11 22:53:10 no Feb 11 22:53:15 its urldatadict, not srcurldata Feb 11 22:53:17 as you can see Feb 11 22:53:23 you didn't change the variable name to match Feb 11 22:53:28 but otherwise yes Feb 11 22:53:45 ah, yes Feb 11 22:53:48 I see now Feb 11 22:54:12 urldata = urldatadict[url] Feb 11 22:54:37 (urldata is really too generic a term, which is why there's these silly names -- 'urldata' -- its unclear from this name whether its referring to all the data for all the cached urls, or for one particular url Feb 11 22:54:41 ) Feb 11 22:55:06 * Jay7 can reasily read php and perl but fails to parse python :( Feb 11 22:55:35 * kergoth_ doesn't see php and perl being much different, its just another programming language, same concepts, slightly different syntax and details Feb 11 22:55:54 true, but.. Feb 11 22:56:09 well.. let's see Feb 11 22:56:47 NameError: global name 'src_uri' is not defined Feb 11 22:56:55 so define it Feb 11 22:56:58 you didn't copy/paste that line Feb 11 22:57:06 src_uri =... Feb 11 23:00:13 well.. passed this time Feb 11 23:01:32 * Jay7 is trying with different bb threads num Feb 11 23:02:05 free Egypt ! Feb 11 23:04:42 well.. I've started builds set with bb threads from 2 to 13 to check Feb 11 23:04:55 but at first look it is fixed now Feb 11 23:09:19 Jay7: great Feb 11 23:09:23 kergoth: thx Feb 11 23:09:44 Tartarus: next one to check is the damned curl-native Feb 11 23:09:59 kergoth_: will you have time to create proper fix or should I commit mine? Feb 11 23:10:36 the proper fix *is* copied and pasted lines from unpack, so i doubt mine would differ very much :) Feb 11 23:11:03 Jay7: to avoid flames later, collect some acks now Feb 11 23:11:44 kergoth: please review then http://pastebin.com/WAtADCHg Feb 11 23:13:21 * ant_ reboots now from Ubuntu to Gentoo, brb Feb 11 23:17:48 ant__: in gentoo you have longer underscore ;) Feb 11 23:17:56 heh Feb 11 23:18:32 now I'll rebuild from here with the same metadata Feb 11 23:18:55 2 builds left Feb 11 23:19:02 I hope to be able to find whats wrong Feb 11 23:20:33 Jay7: seems fine to me Feb 11 23:20:41 cool Feb 11 23:21:05 kergoth: may I add your Ack? Feb 11 23:21:09 yep Feb 11 23:21:54 who will provide second Ack? ;) Feb 11 23:22:37 well.. all builds are finished w/o errors Feb 11 23:25:02 I'm still rebuilding, can't really say yet, though add my ack Feb 11 23:25:24 * Jay7 is thinking on commit message.. Feb 11 23:25:45 iterate over src_uri directly, not over urldata. or just read the logs on do_unpack, since that change was already done there Feb 11 23:30:09 Jay7: http://www.openembedded.org/index.php/Commit_Policy Feb 11 23:30:15 be careful... Feb 11 23:30:19 :P Feb 11 23:30:43 well.. I can go through ML if needed Feb 11 23:30:57 in those cases is silly Feb 11 23:31:01 is a fix Feb 11 23:35:07 Jay7: lzma-native still not built, the strange news is curl-native built fine on this run :/ Feb 11 23:35:23 hehe.. :) Feb 11 23:35:45 I'll try with *lot* of threads, to expose it again Feb 11 23:36:11 I can waste 100% CPU here :) Feb 11 23:37:27 try to build lzma-native before ;) Feb 11 23:37:34 with patch applied Feb 11 23:37:42 doing that Feb 11 23:37:59 iirc is one of the last tasks Feb 11 23:38:17 here is about 91 task for build lzma-native Feb 11 23:39:10 here is 2074 Feb 11 23:39:21 (do_compile) Feb 11 23:39:44 done, just fine Feb 11 23:39:56 I'll add your Ack then Feb 11 23:40:04 sure Feb 11 23:40:42 good nite Feb 11 23:42:17 03Yuri Bushmelev  07master * r03de2481dc 10openembedded.git/classes/src_distribute.bbclass: (log message trimmed) Feb 11 23:42:17 src_distribute.bbclass: iterate over SRC_URI instead of using urldata Feb 11 23:42:17 Do not use bb.fetch.urldata in do_distribute_sources() for traversing Feb 11 23:42:17 SRC_URIs like was already done for do_unpack(). See 661658f62b for Feb 11 23:42:17 additional info. Feb 11 23:42:17 This fixes race issue with lzma-native: Feb 11 23:42:18 http://lists.linuxtogo.org/pipermail/openembedded-devel/2010-October/026076.html Feb 11 23:43:05 Jay7: wow, no races today for console-image ! Feb 11 23:43:21 ant__: it is fixing :) Feb 11 23:43:22 4 bb threads Feb 11 23:43:32 kergoth: thank you for help! :) Feb 11 23:43:36 let see with 8 Feb 11 23:44:18 np Feb 11 23:44:23 glad to hear it worked Feb 11 23:44:28 i still don't know why exactly, but what the hell Feb 11 23:44:29 :) Feb 11 23:44:42 hmm scary Feb 11 23:49:43 when replacing dropbear w/ openssh shouldn't /etc/passwd and /etc/group be updated with the sshd user/group ? I am not seeing after IMAGE_INSTALL += "openssh" what am I missing? Feb 11 23:50:37 ant__: I've found error in oestats-client wrt qa log Feb 11 23:50:50 ah, good Feb 11 23:51:00 for key in files: Feb 11 23:51:01 if not vars[key]: continue Feb 11 23:51:09 he, you'll learn python soon :) Feb 11 23:51:14 should be files[key] instead of vars[key] Feb 11 23:51:28 yeah :) Feb 11 23:51:49 also, that will only work if key is *in* files, just with a false/none value. itll raise a KeyError if its not in files at all, of course Feb 11 23:52:04 kergoth: there is for key in files around Feb 11 23:52:16 good good Feb 11 23:52:21 I don't know how it is possible :) Feb 11 23:54:15 I don't know how it is possible to have empty files['qalog'] as well.. Feb 11 23:54:23 Jay7: btw 8 threads and -j5 is not slow on quad... Feb 11 23:54:34 I'll left it as new default :) Feb 11 23:54:47 *let Feb 11 23:55:05 00:53:54 up 40 min, 3 users, load average: 6.52, 4.71, 3.83 Feb 11 23:55:09 ant__: I have time benchmarking in my todo ;) Feb 11 23:55:22 I could easily emerge smthg Feb 11 23:55:40 in the meanwhile Feb 11 23:55:53 very responsive, I'm surprised Feb 11 23:56:22 ant__: do you have 4 cores or 8? Feb 11 23:56:33 or 4 + HT? Feb 11 23:56:47 xeon 4 cores Feb 11 23:57:58 ant__: on what distro/machine/image you encountered that oestats error? Feb 11 23:58:58 I built 4 images, console, opie, x11, x11-gpe Feb 11 23:59:11 opie and x11 have surely QA Feb 11 23:59:20 (desktop files, mostly) Feb 12 00:00:28 uups 00:59:15 up 46 min, 3 users, load average: 11.22, 7.35, 5.04 Feb 12 00:01:29 ant__: try to apply http://pastebin.com/1kijwrtH Feb 12 00:01:42 I'll test here too Feb 12 00:02:07 soon, still rebuilding...lzma-native ok Feb 12 00:02:28 curl-native in progress... Feb 12 00:02:38 done Feb 12 00:02:40 :) Feb 12 00:02:52 today is a great day :)) Feb 12 00:05:12 true :) Feb 12 00:12:18 hello. has anybody tried openembedded on a digital picture frame? I have access to my DPF's bootloader and I am about to take the plunge Feb 12 00:14:07 b0nafide: what processor does it run Feb 12 00:14:18 i believe it's a little endian mips Feb 12 00:14:25 200 mhz Feb 12 00:14:33 do u know the name of SoC Feb 12 00:14:38 ASM3161 Feb 12 00:14:49 btw. OE supports mipsel pretty well Feb 12 00:14:53 cool Feb 12 00:17:10 i figure I'm on my own on this one, it's currently running a modified eCos and it's a Sony product. Sony has not been helpful. Neither has ASMedia, the producer of the SoC Feb 12 00:18:41 sure Feb 12 00:18:56 if you frame your questions well you will get enough help here Feb 12 00:19:22 * Jay7 -> sleep Feb 12 00:20:41 thanks, i have a lot of reading to do about oe but I have a background in gentoo and openwrt so the learning curve shouldn't be too deadly Feb 12 00:22:32 i'll be back with questions at some point, just checking to see if anybody else has a DPF with OE already Feb 12 00:31:09 khem: any idea about http://tinderbox.openembedded.net/public/logs/task/22744431.txt Feb 12 00:31:12 ? Feb 12 00:31:22 is some kind of race... Feb 12 00:41:17 #ifndef HAVE_SOCKET Feb 12 00:41:18 #error "We can't compile without socket() support!" Feb 12 00:49:52 ant__: your host is busted Feb 12 00:50:33 so, having many bb threads implies curl-native peeks at buildhost and not sysroot? Feb 12 00:50:44 ant__: build machine yes Feb 12 00:50:46 all is fine with 2 threads Feb 12 00:51:02 hmmm can you post config.log for this Feb 12 00:51:17 ok, I'll make it fail Feb 12 00:52:05 strangely it built once with 8 threads... Feb 12 00:52:13 atm once :/ Feb 12 00:53:35 well socket support comes from libc Feb 12 00:54:07 khem: I've already seen that if I build curl-native from scratch it doesn't link to ldap while if I build console-image from scratch it does Feb 12 00:54:16 so I dont know what breaks it with -j Feb 12 00:54:23 may be some other header race Feb 12 00:54:28 ah, -j5 here Feb 12 00:54:41 ant__: yes those issues will always be there Feb 12 00:54:55 welcome the beast called autoconf Feb 12 00:55:00 heh, same PV Feb 12 00:55:05 it knows more than you know about your system Feb 12 00:55:11 ^_^ Feb 12 00:55:37 autoconf has tests which fail first and due to some other dep ldap gets built Feb 12 00:55:51 and bingo when you rebuild he knows to configure with ldap Feb 12 00:56:15 yea, that's why Gentoo has USE flags Feb 12 00:56:45 that can solve it to a certain extent yes Feb 12 00:56:46 it's more predictable than current situation Feb 12 00:56:54 but autoconf is pretty huge Feb 12 00:57:08 best is to generate config.h and cahce it Feb 12 00:57:13 and update it when needed Feb 12 01:05:37 ok, I'll upload config .log Feb 12 01:12:41 http://www.filedropper.com/configlog Feb 12 01:13:01 sorry for captchas..but filebin.ca is dead Feb 12 01:21:35 khem: better here http://www.scribd.com/full/48671778?access_key=key-in83ly405qx37wt8g4s Feb 12 01:22:50 very late here, good night **** ENDING LOGGING AT Sat Feb 12 02:59:57 2011