**** BEGIN LOGGING AT Sat Feb 05 02:59:57 2011 Feb 05 04:03:11 rittk: hmm I think this is it http://lists.linuxtogo.org/pipermail/openembedded-commits/2007-December/011173.html Feb 05 08:10:50 03Martin Jansa  07master * rd1e092f2e5 10openembedded.git/conf/distro/include/efl-from-svn.inc: Feb 05 08:10:50 efl-from-svn.inc: add include to use EFL from svn instead of released tarballs Feb 05 08:10:50 Signed-off-by: Martin Jansa Feb 05 08:17:15 03Martin Jansa  07master * rdb50ac29a9 10openembedded.git/recipes/efl1/efreet.inc: Feb 05 08:17:15 efreet: newer revisions have icon cache enabled by default and dropped old code Feb 05 08:17:15 * option --enable-icon-cache was removed with old code Feb 05 08:17:15 * this file is needed as part of main libefreet1 package Feb 05 08:17:15 Signed-off-by: Martin Jansa Feb 05 08:43:52 help making external toolchain work, built toolchain w/ OE and now want to use it as external rather than internal. Feb 05 12:49:15 gm rp Feb 05 12:49:19 hi effem Feb 05 12:49:38 effem when will you travel? Feb 05 13:04:22 hi woglinde Feb 05 13:04:39 tomorrow morning, brussels is about 1.5 hr driving from where I live Feb 05 13:04:55 planning to leave at around 7.30 am Feb 05 13:09:51 ah Feb 05 13:12:53 um, any particular reason why apt is built without doxygen? Feb 05 13:13:15 it causes the build to fail, because the recipe is attempting to install apt-cache.8 which is not being built Feb 05 13:13:55 two possible fixes - remove references to nonexistant man pages in the recipe or building with doxygen. Which patch would be accepted? Feb 05 13:14:05 there aren many people who uses the .deb support Feb 05 13:14:21 woglinde: sure, that's why I crashed into this :) Feb 05 13:14:47 woglinde: I want to fix it in a way that will be accepted into the OE tree Feb 05 13:19:39 woglinde: hm, does the OE tree contain doxygen? Feb 05 13:19:43 woglinde: it does not seem to Feb 05 13:20:40 dont know offhand Feb 05 13:20:48 woglinde: or maybe I should create a patch with pre-built man pages? Feb 05 13:22:31 hm? Feb 05 13:23:02 woglinde: I could add a patch which would create apt-cache.8 as built by the debian package for example Feb 05 13:23:26 I think makeing a doxygen recipe would be the best solution Feb 05 13:23:30 woglinde: so that we don't need doxygen to build apt and have the man pages Feb 05 13:23:41 woglinde: god no :/ Feb 05 13:24:18 do we need target doxygen or native doxygen to build man pages? Feb 05 13:24:28 imho native should be enough Feb 05 13:24:46 Jay7: I removed the no-doxygen patch and it crashes on not being able to load the perl module Feb 05 13:25:00 Jay7: FATAL: Can't locate DebianDoc_SGML/Format/Alias.pm in @INC Feb 05 13:25:18 Jay7: @INC contains only directories in the build tree Feb 05 13:25:51 filip: middle solution - build w/o man pages and bundle man pages as separate recipe :) Feb 05 13:26:16 right solution is create doxygen recipe Feb 05 13:26:38 Jay7: well doxygen is a bit massive, is it not? Feb 05 13:26:47 quick solution is build w/o man pages :) Feb 05 13:27:00 ehh Feb 05 13:27:05 filip: I haven't used doxygen :) Feb 05 13:27:08 I'm happy :) Feb 05 13:27:37 I'll build with no man pages for now Feb 05 13:28:13 just place no-doxygen patch back :) Feb 05 13:28:34 and I need to fix the recipe refering to the man pages Feb 05 14:05:19 03Martin Jansa  07master * rd313127980 10openembedded.git/conf/distro/include/efl-from-svn.inc: Feb 05 14:05:20 efl-from-svn: make it more usefull Feb 05 14:05:20 Signed-off-by: Martin Jansa Feb 05 14:22:35 hm, I've built a bootstrap image, /bin/sh is from busybox, but there are no other busybox programs Feb 05 14:22:38 how come? Feb 05 14:23:26 no ls even ;p Feb 05 14:31:19 now this is weeird...: Feb 05 14:31:20 builder@jlime:/home/builds/jornada7xx$ find -name sh -type f Feb 05 14:31:21 builder@jlime:/home/builds/jornada7xx$ Feb 05 14:31:30 shouldn't I have it somewhere? Feb 05 15:09:05 filip, /bin/sh is probably a symbolic link to /bin/busybox. Feb 05 15:14:01 mwester: but where are the other links gone? Feb 05 15:14:37 Well, that is what you'll have to figure out! Feb 05 15:15:29 WHOA, -rwsr-xr-x 1 root root 638184 2011-02-04 11:38 busybox Feb 05 15:15:33 suid!? Feb 05 15:15:42 something is fuxored up Feb 05 15:16:09 Are you looking on the device itself, or in your build tree? Feb 05 15:16:10 filip, no, the SUID feature is set Feb 05 15:16:29 (And of course busybox is setuid root; it needs that for recent features) Feb 05 15:16:30 Tartarus: uh, what's this? Feb 05 15:16:36 filip, re missing links, maybe another problem w/ the deb support? That's always a bit bitrotten as you've seen Feb 05 15:16:44 CONFIG_FEATURE_SUID=y in busybox's config Feb 05 15:17:12 which means we install busybox as suid in the rootfs Feb 05 15:17:39 huh, funny, haven't seen it before Feb 05 15:42:13 Tartarus: do you think problem with symlinks may result from alternatives handling? Feb 05 15:42:50 well, they are done via that mechanism Feb 05 15:42:53 so yes, its possible Feb 05 15:43:21 so it would mean that for deb it's broken... God I don't want to step into this Feb 05 15:44:15 Tried poky? That has rpm working and possibly deb too Feb 05 15:44:19 not 100% sure Feb 05 15:44:31 fray, any clue about deb support in poky? Feb 05 15:45:23 is it an openembedded replacement? Feb 05 15:45:49 It's complicated, heh Feb 05 15:45:57 ah, some kind of an extension Feb 05 15:46:25 well we have decided @ the jlime dev team meeting that we are switching from opkg to dpkg Feb 05 15:46:50 and I as the proposer get to do this ;p Feb 05 15:47:04 Well, check out poky and the various deb classes there Feb 05 15:47:08 and see if it's been fixed up Feb 05 15:47:20 Tartarus: yeah, but why isn't it pushed back to OE then? Feb 05 15:47:43 Well, if this was facebook we'd be in a relationship and it's complicated :) Feb 05 15:47:52 Moving forward, however Feb 05 15:48:05 We're going to be working together to leverage both projects strengths Feb 05 15:48:32 Are the jlime folks out of the loop on the whole meta-openembedded thing then? Feb 05 15:48:55 Tartarus: which idea do you mean :)? Feb 05 15:49:57 Well, using poky as a core and creating layers for the additional bits and all of that Feb 05 15:52:02 hm, last time I've heard it was from rafa, but I don't think this got anywhere Feb 05 15:52:52 what is the way I could get changes from poky to get back into OE? Feb 05 15:53:25 yeah Feb 05 15:56:00 aah, I think this is my fuckup: PREFERRED_PROVIDER_virtual/update-alternatives Feb 05 15:56:21 = "dpkg" Feb 05 15:56:33 where did I come up with this... Feb 05 16:49:08 rick-home, ka6sox-away: ping :) Feb 05 16:53:58 guys, I've reported some bugs and attached fixes to them, how can I convince somebody with rw access to the repo to push them? Feb 05 16:54:42 filip: post patches to ML Feb 05 16:55:11 Jay7: can I just post links to the bugzilla ;)? Feb 05 16:55:40 filip: better send as git-format-patch/git-send-email Feb 05 16:55:47 this will go through patchwork Feb 05 16:56:36 hmmm Feb 05 16:58:34 Jay7: but the last one there is 3 months old... Feb 05 16:59:07 rebase it on top of current OE three Feb 05 16:59:17 Jay7: I mean the content on patchwork Feb 05 16:59:23 hmm.. Feb 05 17:00:29 http://patches.openembedded.org/project/oe/list/ Feb 05 17:00:43 patchwork.oe.org is old.. there was migration Feb 05 17:01:01 I've get link from here: http://www.openembedded.org/index.php/Patchwork Feb 05 17:02:11 Jay7: not sure about the workflow with the ML - I am registered using my gmail account. Should I configure some kind of a pipeable command for it? Feb 05 17:02:14 filip: I'll commit your work if no one will complain against after :) Feb 05 17:02:19 Jay7: or can I just paste it into the web ui? Feb 05 17:02:38 filip: I've configured git-send-email Feb 05 17:02:45 wait, I'll paste config Feb 05 17:05:06 http://pastebin.com/tnskuCA1 Feb 05 17:05:19 I have this in my .git/config Feb 05 17:05:33 may be better idea to move this to ~/.gitconfig Feb 05 17:06:43 filip: try some times by sending to yourself Feb 05 17:07:08 Jay7, pong Feb 05 17:07:24 ka6sox-away: we have strange problem with oestats Feb 05 17:07:39 I need some assistance from db and httpd side :) Feb 05 17:08:01 hola, friendly neighborhood sysadmin at your service Feb 05 17:08:03 whatcha need? Feb 05 17:08:12 cryptk: hey :) Feb 05 17:08:17 go cryptk! Feb 05 17:08:20 haha Feb 05 17:08:22 okay off to my meeting. Feb 05 17:08:50 "ka6sox is away at the moment, if you leave a message after the beep I will do whatever you needed him to do" Feb 05 17:08:51 BEEP! Feb 05 17:09:39 Jay7, I'm here to. What's the problem? Feb 05 17:09:39 cryptk: well.. oestats save only first package sent (quilt-native mostly) Feb 05 17:09:51 * mwester furiously shuffles through his stack of "to-do" papers to find all the ones with ka6sox's name on them Feb 05 17:09:52 I suspect some unique index in mysql Feb 05 17:10:00 or may be http throttling Feb 05 17:10:10 not http throttling. We fixed that Feb 05 17:10:25 rick-home was created some indexes on db Feb 05 17:10:32 yep, this is one for rick-home Feb 05 17:10:41 yes I did, but I thought, I didn't create any unique ones... Feb 05 17:10:51 but if I did, them I'm an evil person Feb 05 17:10:59 hmmmm Feb 05 17:11:05 rick-home: I've looked to scheme you posted to pastebin, there was no unique Feb 05 17:11:09 right Feb 05 17:11:29 but you was created 1-2 after iirc Feb 05 17:11:32 mwester, did you have something that needed to be done as well? Feb 05 17:11:36 can you please have look? Feb 05 17:11:49 and not only that, but in looking at sql logs, the oestats does each save individually Feb 05 17:11:53 Jay7: thanks for the config :) Feb 05 17:11:55 there's no "bulk" updates. Feb 05 17:12:04 hm.. Feb 05 17:12:07 cryptk, yeah, but it's best that it wait for ka6sox (autobuilder load discussions) Feb 05 17:12:20 mwester, I am in on all of those discussions as well ;) Feb 05 17:12:23 rick-home: have you mysql query log? Feb 05 17:12:30 check the emails, I am CC'd on all of them, lol Feb 05 17:12:43 mwester, cryptk = ka6sox for purpose of those discusions Feb 05 17:12:57 is it possible to share last day's log wrt oestats for me? Feb 05 17:13:01 ka6sox-away, go to your meeting, lol Feb 05 17:13:07 I have several Jay7 -- I don't leave query log on for long... but I have occasionally turned it on for an hour or so to debug the problems. Feb 05 17:13:31 cool. :) Then I'll just mention that at some point in the next month or so, I'd like to re-enable the SlugOS autobuilder. Feb 05 17:13:47 * Jay7 scratching his head.. Feb 05 17:14:05 Jay7, the stats you are refering to update throughout the day don't they, this isn't some overnight batch job... Feb 05 17:14:28 mwester, lets wait until we get the current load under control, but that is a possibility Feb 05 17:14:56 Yep, it's not a critical issue. For the moment, I can manually build when things get out-of-date. Feb 05 17:15:20 cool Feb 05 17:15:47 I'd just like for the unstable feeds to remain at least useful, as we change the kernel and libs for the core SlugOS distro in the upcoming weeks. :) Feb 05 17:16:09 Jay7, how about I do a log run -now- and send it to you? Feb 05 17:16:22 rick-home: well.. may be good idea Feb 05 17:16:28 how long do you want, or a better question is, how big a file do you want? Feb 05 17:16:44 build should take about 2 hours Feb 05 17:16:47 1hr =145MB Feb 05 17:16:48 I'll run one Feb 05 17:17:00 heh.. Feb 05 17:17:21 other idea - setup it here and try to reproduce.. Feb 05 17:17:35 mwester, after the server shuffle, I don't see it being a problem, neither does ka6sox Feb 05 17:17:36 heading out...bbl...SCALE meeting Feb 05 17:17:37 rick-home: well.. can you please send me latest db schema? Feb 05 17:17:56 I'll try to do some tests at home Feb 05 17:18:03 Jay7: so I should git send-email to openembedded-devel@lists.openembedded.org? Feb 05 17:18:17 filip: yes, from subscribed email Feb 05 17:18:43 this is funny how a proper tool (bugzilla) is discarded in favor of a tool dedicated for something else (ML) Feb 05 17:18:47 how did this happen? Feb 05 17:19:05 filip: nobody cares about bugzilla, but all people are reading mail.. Feb 05 17:19:11 I think that was reason Feb 05 17:19:56 Jay7, -- I'm going to see if I can filter out just the oe_stats tables for you and send that to you as well. Feb 05 17:20:02 for a query log Feb 05 17:20:11 rick-home: that would be great.. Feb 05 17:21:13 rick-home: other Q - may you give me r/o access to some mysql tool (phpmysqladmin) to oestats db? Feb 05 17:21:22 then no need for logging queries Feb 05 17:21:26 I'll just look into db Feb 05 17:22:24 Jay7: I've just sent one, does this look ok? Feb 05 17:22:53 havent received yet.. I'll look as it appears Feb 05 17:24:11 Jay7: bitbake patches should be sent to the same address? Feb 05 17:24:16 Jay7, I think that we can get you a cert into phpmyadmin. You have mysql passwords that work for all your tables... so I see no reason why that can't happen. Wait a minute. Feb 05 17:24:44 filip: bitbake patches better to send to bitbake ML.. I don't know it's address Feb 05 17:24:59 filip: ask kergoth or RP about Feb 05 17:25:24 rick-home: I haven't any password :) Feb 05 17:26:41 Jay7: ok, thanks for the hints, maybe this will work out better :) Feb 05 17:29:23 filip: well.. I hope we will push all your changes :) Feb 05 17:29:30 or at least propose something other :) Feb 05 17:29:54 Jay7: these are minor ones, nothing to debate about :) Feb 05 17:30:38 * mwester warms his hands over his monitor, whilst looking about his office to see if he can turn on any more heat-producing devices Feb 05 17:31:21 mwester: run OE build on all computers :) Feb 05 17:31:43 do whatever you do on emerald... on all devices, lol Feb 05 17:31:47 that should build some heat Feb 05 17:31:56 I moved all my big build machines into the basement, Jay7 -- that's why it's so cold in here now! Feb 05 17:32:09 filip: I have received your patch Feb 05 17:32:30 filip: looking good but better to paste problem descritions around bugzilla link Feb 05 17:32:39 to have all info at once :) Feb 05 17:32:58 mwester: that was your fault! :) Feb 05 17:34:38 filip: now let's wait for monday's evening and if no one will complain I (or someone else) will push it Feb 05 17:35:07 sure, but I hardly can imagine about any complaints Feb 05 17:35:26 does anybody know anything about softmac driver for orinoco cards? Feb 05 17:35:31 filip: may be some burocracy :) Feb 05 17:35:48 e.g. wrong commit message or something like :) Feb 05 17:36:08 damn Feb 05 17:36:58 http://openembedded.org/index.php/Commit_Policy Feb 05 17:37:03 this describe rules Feb 05 17:37:27 Jay7: yeah I've read them about a year ago ;) Feb 05 17:37:44 btw, I've remember something about gnu-configize from Tartarus or JaMa Feb 05 17:37:54 may be your patch was already reinvented :) Feb 05 17:38:32 the sooner the better - I'd just like to move on :) Feb 05 17:38:47 :) Feb 05 18:12:47 Jay7, I forget your email Feb 05 18:13:52 rick-home: PM'ed Feb 05 18:22:47 has anybody managed to actually draw a bitbake -g graph :P? Feb 05 18:22:53 dot segfaults on me Feb 05 18:23:11 filip: what version you using/ Feb 05 18:23:15 dot: width (132216 >= 32768) is too large. Feb 05 18:23:28 Jay7: of bitbake? BitBake Build Tool Core version 1.10.2, bitbake version 1.10.2 Feb 05 18:23:53 Jay7: I am trying to figure out why X libraries are being built for the bootstrap image Feb 05 18:24:20 and cups, and other weird stuff Feb 05 18:24:24 iirc, bb -g was fixed already.. Feb 05 18:24:27 try git master Feb 05 18:26:16 um, git master? bitbake site says something about svn Feb 05 18:27:01 ok I see it in the OE repo Feb 05 18:27:11 filip: git clone git://git.openembedded.org/bitbake Feb 05 18:30:52 from bb.server.process import ProcessServer, ServerCommunicator Feb 05 18:30:52 ImportError: No module named process Feb 05 18:30:54 whoops Feb 05 18:31:58 yeah, this file does not exist Feb 05 18:33:40 fixed Feb 05 18:35:01 Hi Feb 05 18:35:24 anyone know the trick to build an image with openssh instead of dropbear ? Feb 05 18:35:45 tried DISTRO_SSH_DAEMON = "openssh" in my image recipe Feb 05 18:35:49 hey, who maintains OE's websites? like the wiki and such? Feb 05 18:35:50 but it did not work Feb 05 18:35:58 do you guys have someone? or do you have us do it? Feb 05 18:36:09 cryptk? Feb 05 18:36:22 cryptk: I thought ka6sox doing this from tech side :) Feb 05 18:36:25 ok Feb 05 18:36:45 I need to make sure all of your stuff is setup correctly then for the HA mysql setup I am doing Feb 05 18:37:14 hehe.. Feb 05 18:37:16 ? Feb 05 18:37:22 yes? Feb 05 18:37:25 Jay7: whoa, fancy progressbar in the git bitbake Feb 05 18:37:35 filip: yeah, parallel parsing :) Feb 05 18:37:42 * cryptk thinks woglinde_ is wondering who this cryptk guy is, lol Feb 05 18:37:49 right Feb 05 18:38:00 cryptk: ask ka6sox-away about this Feb 05 18:38:04 he should be right person Feb 05 18:38:12 for all intents and purposes, as far as the sysadmin stuff go's cryptk==ka6sox Feb 05 18:38:26 I am the "other guy" that plays on the servers Feb 05 18:38:49 * cryptk also happens to do server virtualization for a living Feb 05 18:38:59 good to hear Feb 05 18:39:05 and you are from europe Feb 05 18:39:16 nope, I am from the USA Feb 05 18:39:20 hm uhm Feb 05 18:39:22 okay Feb 05 18:39:30 San Antonio TX Feb 05 18:39:33 server only form france Feb 05 18:39:40 Jay7: anyway dot still segfaults, so I think I'll have to get to it in some other way Feb 05 18:40:40 jay7 cannt cryptk looking at oestats? Feb 05 18:41:02 woglinde_: we are looking with rick-home Feb 05 18:41:09 ah okay Feb 05 18:41:12 he is another ka6sox' clone ;) Feb 05 18:52:04 Jay7, http://www.nas-admin.org/oe_tinderbox.sql.gz Feb 05 18:52:23 there's a sqldump for you Feb 05 18:56:10 Jay7, let me know when you are done downloading that so I can remove it Feb 05 18:58:15 is there any other way to show dependencies except -g? Feb 05 18:58:26 like in some indented tree or something Feb 05 18:59:15 Jay7, hardly a ka6sox clone... a semi-competant dba at best. Mostly I'm a user-interface designer Feb 05 19:01:35 Jay7, ping? Feb 05 19:03:46 filip no Feb 05 19:03:55 bzt -g has all infos you need Feb 05 19:05:11 Jay7, I'm not sure, but looking at the query log for oe_stats, you might want to go find the guy who wrote it and apply a hammer to his fingers..... Feb 05 19:10:46 woglinde: I know, but .dot does not exceed in readability and I have trouble rendering it with graphviz Feb 05 19:11:02 oh, ps output seems to have worked Feb 05 19:11:39 filip? Feb 05 19:11:43 you dont need to render it Feb 05 19:11:59 you can pick up all information from the files Feb 05 19:12:03 isnt hard to learn Feb 05 19:12:36 woglinde: I know I can read it, it's just very uncomfortable :P Feb 05 19:12:49 filip for what you need it? Feb 05 19:12:51 woglinde: need to jump all over the file to find out the dep chain Feb 05 19:12:56 woglinde, can you describe for me what's actually broken in oe_stats ??? Feb 05 19:13:10 woglinde: I am building a bootstrap-image and wanted to find out what pulls in X libraries, pango etc Feb 05 19:13:26 rick-home I heard that sometimes the stat goes through and sometimes not Feb 05 19:13:37 I dont use it for a while Feb 05 19:13:41 kk Feb 05 19:14:34 rick-home in the same bitbake call Feb 05 19:22:29 Jay7, oe_stats general query log sent to your email. Feb 05 19:22:59 woglinde, if you see Jay7 please ask him to grab the sqldump he asked for so we can delete it. Feb 05 19:23:12 I'm off to run errands and will not be back for several hours. Feb 05 19:24:42 bye rick-home Feb 05 19:25:46 rick-home: here Feb 05 19:27:05 rick-home, cryptk: I'm downloading Feb 05 19:27:14 cool, lemme know when you finish Feb 05 19:28:17 according to wget - after 35 min :) Feb 05 19:28:36 cryptk: is that full mysqldump of oestats db? Feb 05 19:28:51 oe_tinderbox is what rick told me you wanted Feb 05 19:29:28 and yes, full mysqldump Feb 05 19:29:33 cool Feb 05 19:44:27 hm.. oestats is disabled w/o notification in log Feb 05 19:44:45 oestats.id is empty in TMPDIR Feb 05 19:44:53 wtf?.. Feb 05 19:45:13 * Jay7 is going to setup local oestats-server Feb 05 19:48:07 Jay7, I'm back -- wife delayed our leaving. Feb 05 19:48:22 what is it you're seeing? Feb 05 19:48:22 happens :) Feb 05 19:48:47 rick-home: oestats-client store build id in file Feb 05 19:49:05 when it can't continue it stores empty file Feb 05 19:49:05 yes. Feb 05 19:49:21 seems something is happen but I see nothing in build log Feb 05 19:49:41 btw, dump was downloaded Feb 05 19:49:52 cryptk, rick-home: you can unshare it :) Feb 05 19:49:57 ok, thanks cryptk will get it out of the way then Feb 05 19:50:02 did you get the query log? Feb 05 19:50:19 ah, yes, I see it Feb 05 19:50:27 removed Feb 05 19:50:35 it sounds, tho, like the problem may be on the script side. Feb 05 19:51:05 well, thanks, people Feb 05 19:51:13 I'll investigate Feb 05 19:51:14 welcome Feb 05 19:51:22 let is know if there's something we can do for you Feb 05 19:51:53 ok Feb 05 20:07:47 hm, I have trouble figuring out the dependency tree... is it normal that gnome is being built for bootstrap-image? Feb 05 20:07:58 for example NOTE: Running task 1885 of 3372 (ID: 1705, /home/users/builder/openembedded/recipes/gnome/gnome-vfs_2.24.1.bb, do_install) Feb 05 20:08:26 filip: some of natives may require it Feb 05 20:08:40 but this is wrong imho Feb 05 20:08:52 Jay7: yeah, it considerably slows down development Feb 05 20:15:17 garnet + emerald is killing me... Feb 05 20:15:21 makes me sadfaced Feb 05 20:23:07 why do people even bother with stuff like gpe-mini-browser when chromium exists? Feb 05 20:33:18 filip: how many gigabytes does it take to build chromium? Feb 05 20:34:07 pupnik: probably A LOT, but build machines aren't storage restrained, right? Feb 05 20:34:29 I fear something was changed in OE internals what break oestats.. Feb 05 20:36:54 that's a lot of code to audit Feb 05 20:41:25 NOTE: oestats: error sending task, disabling stats Feb 05 20:41:35 bb 1.10 show this Feb 05 20:41:43 bb master doesn't Feb 05 20:51:00 rick-home, cryptk: are you sure that you are not throttling http posts? :) Feb 05 20:51:17 yep Feb 05 20:51:22 6th post is constantly failed :( Feb 05 20:55:13 ERROR: Task 12 (/home/users/builder/openembedded/recipes/images/bootstrap-image.bb, do_rootfs) failed with exit code '1' Feb 05 20:55:54 uh, where can I find some debugging output? Feb 05 20:56:04 it doesn't seem to reside in an usual location Feb 05 20:58:17 filip? Feb 05 20:58:23 ah rootfs Feb 05 20:58:25 hm Feb 05 21:10:17 belgacom.be seems official sponsor of fosdem :) Feb 05 21:10:29 re rp Feb 05 21:14:43 woglinde: any hints beside stracing :)? Feb 05 21:14:59 filip hm not in the moment Feb 05 21:15:09 I remember too it was hard to debug Feb 05 21:26:53 * Jay7 found assert in client that breaks things Feb 05 21:26:59 now I should understand why.. Feb 05 21:29:21 hehe Feb 05 21:29:54 for key in vars: Feb 05 21:29:55 assert vars[key] Feb 05 21:30:03 this ^ Feb 05 21:30:16 how can this happens?.. Feb 05 21:30:32 ah.. empty value Feb 05 21:31:14 Jay7, if I was trying to schedule some downtime for the mysql database, who would I need to talk to? Feb 05 21:31:34 announce here and in OE ML Feb 05 21:32:19 cool, thanks man Feb 05 21:32:27 I will work with ka6sox to get it announced Feb 05 21:32:48 I basically just need a little downtime to get the backup database synced and turn on master-master replication Feb 05 21:33:34 http://openembedded.org/index.php/Infrastructure Feb 05 21:33:44 here is list of responsible people Feb 05 21:35:59 thx Feb 05 21:46:30 Jay7, I find it funny that I am looking for "someone responsible" and the page of responsible people lists me, HAHA Feb 05 21:46:41 well, it mentions nas-admin.org staff, lol Feb 05 21:46:45 :)) Feb 05 21:54:38 haha Feb 05 22:00:48 can someone tell me how to replace dropbear with openssh in an angstrom image Feb 05 22:01:10 setting DISTRO_SSH_DAEMON did not work Feb 05 22:01:29 ISTRO_SSH_DAEMON = "openssh" Feb 05 22:01:32 hm Feb 05 22:01:49 what distribution and image did you use? Feb 05 22:02:27 I created a angstrom-custom.conf image file in an overlay, with includes angstrom-2008.1 and override the DISTRO_SSH_DAEMON variable Feb 05 22:03:21 what image? Feb 05 22:03:38 and I use a custom image as well which depens on task-base-extended and task-proper-tools Feb 05 22:07:27 try run bitbake -e your_image Feb 05 22:07:41 and look what the SSH_DAMEON says Feb 05 22:09:57 DISTRO_SSH_DAEMON="openssh" Feb 05 22:10:52 my old images were using dropbear, maybe I need to clean my image and try again ? Feb 05 22:10:55 well.. I hope I've fixed it Feb 05 22:11:35 bitbake -c clean my_image doesn't seem sufficient though Feb 05 22:11:56 woglinde, do you know what could be the problem ? Feb 05 22:15:40 julianpid try build from scratch Feb 05 22:17:43 it's gonna take decades, but I think that's the best I can do :) Feb 05 22:20:31 buy a faster machine Feb 05 22:22:22 :) Feb 05 22:22:24 btw Feb 05 22:22:52 do you know which alternative to tinylogin I could put as IMAGE_LOGIN_MANAGER ? Feb 05 22:22:52 well.. I'll run testbuilder with fix applied and look tomorrow on results Feb 05 22:23:05 if all will be good I'll commit this fix Feb 05 22:24:21 Jay7, are you talking about a fix for the SSH daemon stuff ? Feb 05 22:24:34 no Feb 05 22:24:38 oe_stats Feb 05 22:24:40 julianpid_: no, about fix for oestats-client :) Feb 05 22:24:48 oh, sorry then Feb 05 22:24:52 np Feb 05 22:25:51 btw Jay7, I think I had a discussion with you about multilib toolchains a while ago Feb 05 22:26:21 ah, I've remember something Feb 05 22:26:56 julianpid_ btw. be sure you have ${DISTRO_SSH_DAEMON} under IMAGE_INSTALL = Feb 05 22:27:29 after some investigations, it turns out I was not looking for multilib toolchains, but more "bi-arch" toolchain (using the gcc jargon) Feb 05 22:28:31 I've been doing a small patch for it Feb 05 22:28:53 I implemented as a machine feature Feb 05 22:30:06 I don't know if it's worth sending a RFC for it on the oe mailing-list though, as it is for very special use cases Feb 05 22:31:16 and here I am wondering why I get no logs from do_rootfs: 17279 unlink("/home/builds/jornada7xx/tmp/work/jornada7xx-oe-linux-gnueabi/bootstrap-image-1.0-r0/temp/log.do_rootfs.17279") = 0 Feb 05 22:31:20 why!? Feb 05 22:32:58 julianpid_: it does no bad if you will send it :) Feb 05 22:32:58 woglinde, I have no DISTRO_SSH_DAEMON in my IMAGE_INSTALL, but it is supposed to be taken care of by the task-base-extended task Feb 05 22:33:51 filip: hehe.. ant have seen even more interesting bug - log name was /dev/null :) Feb 05 22:34:31 Jay7, I will, I have so much patches pending various opensource projects :) I need to reserve a few hours to empty my queue at some point :) Feb 05 22:35:00 it happens Feb 05 22:35:35 Jay7: ok I know where it explodes, around line 1090 in runqueue.py Feb 05 22:35:57 Jay7: this line probably: bb.build.exec_task(fn, taskname, the_data) Feb 05 22:36:01 wow that told me a lot :/ Feb 05 22:36:30 well.. at least now oestats-client is reporting more than quilt-native Feb 05 22:36:56 woglinde, would it help if I do ./bb -c clean task-base-extended ? Feb 05 22:37:39 god I hate space indentations in python Feb 05 22:38:32 add brackets #{ #} :) Feb 05 22:39:22 yeah.. python isn't very good for collaborative development :) Feb 05 22:39:22 julianpid_: no, I like tab indentations :P Feb 05 22:39:38 it's just four spaces which makes me throw up ;p Feb 05 22:39:42 because of different 'working styles' Feb 05 22:40:01 ok, I got the thing that blows up my build Feb 05 22:40:03 http://pastebin.com/VMh66EXL Feb 05 22:40:08 any suggestions about this one :)? Feb 05 22:40:58 I could downgrade bitbake or fix this ;P Feb 05 22:41:59 julianpid maybee Feb 05 22:42:55 pyshlex complains about the "$((" separator... Feb 05 22:44:05 sh#t... oestats is disabled again Feb 05 22:47:46 is using $((..)) in the recipes a bug? Feb 05 22:48:00 it seems like bitbake can't parse it Feb 05 22:48:00 filip: try grepping over recipes Feb 05 22:48:14 if no one is using it then it is Feb 05 22:48:33 oh.. bb 1.10 is sooo slooow.. Feb 05 22:48:39 * Jay7 waiting for parsing Feb 05 22:49:42 woglinde, cleaning the task worked Feb 05 22:49:49 thx for your help Feb 05 22:53:51 julianpid_ you mostly solved it yourself Feb 05 23:00:13 ant__: hello Feb 05 23:01:01 hi Feb 05 23:01:29 still working on framebuffer? Feb 05 23:01:51 maybe there is a kernel bug :/ Feb 05 23:02:33 ant__: no, today I'm fixing oestats :\ Feb 05 23:04:48 oh, go ahead, pls Feb 05 23:05:13 Jay7: I've sent another patch for consideration :) Feb 05 23:29:07 Jay7, any progress? Feb 05 23:29:24 rick-home: this is client-side issue Feb 05 23:29:26 at least one Feb 05 23:29:43 thought so from what I was seeing. Anything I can do to help? Feb 05 23:30:04 no, thanks Feb 05 23:30:16 kk -- if there is, let me know Feb 05 23:30:24 ok Feb 05 23:31:14 HTTP POST isn't very reliable way to report lot of data.. Feb 05 23:33:10 well, unless you post it in really big xml chunks. But if it's atomic, little transactions, then no. Feb 05 23:33:43 it's atomic little transactions :) Feb 05 23:34:02 every package status change is notified by post Feb 05 23:34:09 oh dear god. Feb 05 23:34:27 that's what I was afraid of from looking at the queries. Feb 05 23:34:49 yeah.. Feb 05 23:35:02 that would have been fine back when the builder was building 4 or 5 target machines and a few dozen packages. but now..... Feb 05 23:35:02 imagine what is doing http server :) Feb 05 23:35:23 design didn't evision scaling. Feb 05 23:35:27 envision Feb 05 23:35:57 * Jay7 is thinking about single server which listen on some port and have persistent DB connection Feb 05 23:36:13 old'n'plain tcp :) Feb 05 23:36:45 * mwester wonders if a temporary solution might be a local daemon which buffers the local POST data, and sends fewer, larger, transactions up to the oestats server. Feb 05 23:36:58 WARNING: oestats: exception Feb 05 23:37:00 WARNING: oestats: error sending task, disabling stats Feb 05 23:37:10 mwester: yes, that is good idea too Feb 05 23:37:35 run separate bb thread which will collect data and send it to server Feb 05 23:38:05 but I know only 2 person who can do this - RP and kergoth :) Feb 05 23:38:05 i'd just store it in a local db and dispatch at BuildCompleted/BuildFailed event firing Feb 05 23:38:14 pretty straightforward, i'd think Feb 05 23:38:30 yes, it's ok as well Feb 05 23:38:51 the point is, it's a refactoring of the posting code in the builder/stats pkg. Feb 05 23:38:59 good luck with that Jay7 Feb 05 23:39:24 yeah.. Feb 05 23:39:36 * Jay7 is looking at 'urllib2.HTTPError' Feb 05 23:40:01 what to do with it?.. Feb 05 23:40:28 well.. let's try to resend then.. Feb 05 23:41:22 Jay7, the problem is that the code on the OTHER side is also expecting atomic transactions, you can't store them up in a local temp table until the build succeeds or fails and then post them in a batch... Feb 05 23:41:44 rick-home: so we should change server-side part as well Feb 05 23:41:55 they have to coordinate... Feb 05 23:41:59 heh, thats what i was thinking -- wouldn't want the code executing at build completed time to have to send with a shitload of individual posts Feb 05 23:42:25 unrelated - https://gist.github.com/811790, this seem sane, output/interface wise? Feb 05 23:42:30 kergoth: may you write some draft from bb side? Feb 05 23:42:44 i don't see it having anything to do with bb, its class code, event handlers Feb 05 23:43:03 well.. then point to right handlers :) Feb 05 23:43:06 might be able to just use persist_data to avoid having to touch a db directly Feb 05 23:43:31 it already has handlers for the tasks, just need to have it save the info rather than sending it, and add a handler at BuildCompleted event and BuildFailed event Feb 05 23:43:35 I need draft, I can't write it from scratch :( Feb 05 23:43:44 but I can fill draft with real code Feb 05 23:44:32 ah.. hm Feb 05 23:44:40 I have seen that Feb 05 23:45:00 oestats-client even listen for BuildCompleted and BuildFailed Feb 05 23:45:09 to close build Feb 05 23:45:15 the only part that isn't already there, really, is how you save/load the info about what happened with the task Feb 05 23:45:20 to some external file Feb 05 23:45:36 kergoth: do bb have some methods to work with sqlite? Feb 05 23:45:45 bb.persist_data, i just mentioned it ;) Feb 05 23:45:55 its what's used to store the revision incrementing stuff for AUTOREV Feb 05 23:46:17 well.. interesting.. Feb 05 23:46:20 it acts like a key/value store within a domain. domain == table, a string identifier Feb 05 23:47:19 persist = bb.persist_data.PersistData(d); persist.addDomain('task_results'); persist.setValue('task_results', 'foo', 'bar') Feb 05 23:47:24 that's all there is to it Feb 05 23:47:31 and of course there's getValue, etc Feb 05 23:47:39 current bitbake has a better API, but OE doesn't require that version yet ;) Feb 05 23:48:00 current: persist = bb.persist_data.persist(d); persist['task_results']['foo'] = 'bar' Feb 05 23:48:34 of course, if that's too simple, not enough, you'd need to use sqlite3 directly Feb 05 23:48:42 but you can look at the persist_data code for an example of that, if you need to Feb 05 23:49:00 one of these days i want to replace the output of the pkgdata text files with persist_data or sqlite3 usage Feb 05 23:50:28 well Feb 05 23:50:43 * Jay7 written all this to his todo wiki Feb 05 23:50:47 :) Feb 05 23:50:54 good practice if your python is rusty ;) Feb 05 23:52:37 yeah.. Feb 05 23:56:04 i might be able to take a look on monday, depends, doubt i'll have time today or tomorrow, errands / cleaning galore Feb 05 23:56:51 btw, we should cleanup old tinderbox-client code Feb 05 23:56:58 zecke was Ack'ed it :) Feb 06 00:45:46 * Jay7 -> sleep Feb 06 00:46:07 will commit fixed version tomorrow Feb 06 00:54:17 hi Feb 06 00:54:41 how can I find out what package directly depends from another package? Feb 06 00:55:06 i am trying to build an image, and one of the dependencies is python-numpy, which has been completely broken in OE for a while Feb 06 00:55:26 so I want to remove that dependency, but I need to find out which package requires python-numpy Feb 06 01:19:26 anyone here that works with emerald? Feb 06 02:16:03 dv_, bitbake -g will output the dependency graph in dot format. **** ENDING LOGGING AT Sun Feb 06 02:59:57 2011