**** BEGIN LOGGING AT Mon Nov 03 02:59:56 2008 Nov 03 04:30:28 mwester_, inc Nov 03 08:37:52 morning Nov 03 08:57:55 mwester: regarding those packages that you set ALLOW_EMPTY on, where is the dependency in question coming from? I don't think anything should be depending on xproto, for example. If the -dev package is the one with the dependency then it would probably be better to suppress that at its source. Nov 03 09:07:06 yo Nov 03 09:26:25 Hello. I am building gcc-cross-4.2.4 do_configure and there is error: Nov 03 09:26:26 | configure: error: GMP 4.1 and MPFR 2.2.1 or newer versions required by fortran Nov 03 09:50:38 in the source dir are gmp-4.2.4.tar.bz2 and mpfr-2.3.1.tar.bz2, how i can install them or check installation? Nov 03 10:09:46 morning Nov 03 10:10:10 Kalemas: which distro you are building? Nov 03 10:17:10 mandriva 2009 Nov 03 10:17:22 bitbake x11-image Nov 03 10:17:27 hi all Nov 03 10:20:03 mmm, angstrom? Nov 03 10:20:36 angstrom 2008.1? Nov 03 10:21:10 2008.1-test-20081103 Nov 03 10:22:55 my build of angstrom-2008.1 for arm target built gmp 4.2.4 and mpfr 2.3.1 Nov 03 10:25:54 i need bitbake gmp anf mpfr again? Nov 03 10:27:26 they should be built automatically Nov 03 10:28:13 and if not? or maybe some thing wrong with gcc-cross? Nov 03 10:29:55 which MACHINE you are building for? Nov 03 10:30:03 and please show your local.conf Nov 03 10:30:15 magician Nov 03 10:30:52 pastebin your local.conf please Nov 03 10:30:54 ~pastebin Nov 03 10:30:55 [~pastebin] A "pastebin" is a web-based service where you can paste anything over 3 lines without flooding the channel. Here are links to a few : http://www.pastebin.com , http://pastebin.ca , http://channels.debian.net/paste , http://paste.lisp.org , http://www.rafb.net/paste Nov 03 10:31:01 http://en.pastebin.ca/1243886 Nov 03 10:34:54 hmm.. Nov 03 10:35:24 Kalemas: does error repeat when you remove all PREFERRED_PROVIDERS lines? Nov 03 11:25:37 hrw: fatal error, nothing chage... Nov 03 11:26:33 ERROR: Build of /home/kalemas/oe/stuff/openembedded/packages/mpfr/mpfr_2.3.2.bb do_fetch failed :^) Nov 03 11:29:00 o thats other thing Nov 03 11:29:23 yes, 2.3.1 seems ok Nov 03 11:32:26 mm, no. building stops at the same moment, | configure: error: GMP 4.1 and MPFR 2.2.1 or newer versions required by fortran. But now i try build mpfr separetely. Nov 03 11:49:00 good morning Nov 03 11:49:29 hi florian Nov 03 11:53:37 it is fails again. :^( Nov 03 12:06:47 zecke: hail zecke Nov 03 12:10:27 hrw: have you any idea for me? Nov 03 12:18:34 1sorry, but nope :( Nov 03 12:47:29 03Felix Domke  07org.openembedded.dev * r70bb8ee097 10openembedded.git/packages/python/python-twisted_2.5.0.bb: Nov 03 12:47:29 twisted-python: fix packaging of twisted-python-protocols Nov 03 12:47:29 patch by Moritz Venn Nov 03 12:47:29 Signed-off-by: Felix Domke Nov 03 12:50:44 03zecke123 * r1117 10bitbake/ (ChangeLog lib/bb/fetch/svn.py): (log message trimmed) Nov 03 12:50:44 [svn] Add @rev to svn checkout command Nov 03 12:50:44 Patch by borgcube@gmx.li Nov 03 12:50:44 Svn tries to be smart about revisions. So, when you check out an older revision Nov 03 12:50:44 of a file it goes to the latest revision (HEAD) and tries to go back to the old Nov 03 12:50:45 file. In this case it was impossible, since the whole thing was moved outside of Nov 03 12:50:49 svn's scope, so svn can't find the file in the HEAD revision. Nov 03 12:55:20 03zecke123 07bitbake-1.8 * r1118 10/ (ChangeLog lib/bb/fetch/svn.py): (log message trimmed) Nov 03 12:55:20 [svn] Add @rev to svn checkout command Nov 03 12:55:20 Patch by borgcube@gmx.li Nov 03 12:55:20 Svn tries to be smart about revisions. So, when you check out an older revision Nov 03 12:55:20 of a file it goes to the latest revision (HEAD) and tries to go back to the old Nov 03 12:55:21 file. In this case it was impossible, since the whole thing was moved outside of Nov 03 12:55:25 svn's scope, so svn can't find the file in the HEAD revision. Nov 03 12:58:13 Hello ...I'd like to propose the removal of fam recipes. Do someone object? Nov 03 12:58:27 I've done it in ml and noone has speaked up; so ... Nov 03 13:33:13 morning Nov 03 13:37:54 hey otavio Nov 03 13:38:06 did you get the Java stuff working? Nov 03 13:38:31 kalev: somewhat; I'd love to get some help on it. Nov 03 13:38:51 kalev: I've written a new recipe for the latest version of it and it somewhat works Nov 03 13:39:17 kalev: however it fails to find the classes so I've probably missing something. Nov 03 13:39:28 kalev: Can I make it available and you take a look? Nov 03 13:40:05 sure thing, but I cannot test it right now Nov 03 13:40:27 http://arquivos.ossystems.com.br/~otavio/sun-jre6_1.6.0.10.bb Nov 03 13:40:32 take a look at it when you can. Nov 03 13:40:52 * otavio is fighting against a xfdesktop segfault right now Nov 03 14:01:35 otavio: it looks ok to me Nov 03 14:02:40 otavio: it might be that you need to change something if you use jdk-6u10-dlj-linux-i586.bin instead of jre-6u10-linux-i586.bin Nov 03 14:03:09 kalev: well, not really Nov 03 14:03:15 kalev: I've done those changes already Nov 03 14:03:21 kalev: but I have two issues ... Nov 03 14:03:46 kalev: 1) if I try to run java by hard it fails to load java.Lang (dunno why) Nov 03 14:03:51 kalev: 2) plugin fails Nov 03 14:04:04 kalev: it is probably due the same root cause but I failed to find the reason Nov 03 14:04:24 kalev: could you give it a try and see if you can come up with a possible solution for it? Nov 03 14:04:43 kalev: you looks to be more experienced then me on that stuff since I'm not a java guy :P Nov 03 14:08:37 otavio: I'm not much of a Java guy either, and this recipe is at least 2 years old :) Nov 03 14:08:54 kalev: wow .. god dammit :P Nov 03 14:09:35 otavio: I could take a look, but maybe on Friday or so Nov 03 14:09:53 lo, somthing new about ppc64-linux target? Nov 03 14:10:12 kalev: please do; I'll try to solve it in meanwhile but if I fail it would be great if you could take a look and help me on that Nov 03 14:10:19 ok Nov 03 14:10:54 I have a very pressing deadline with another project. Sorry that I'm not much help here right now. Nov 03 14:20:27 hi woglinde Nov 03 14:24:27 Did someone already got 'Stale NFS file handle' inside of a rootfs using ext2? Nov 03 14:31:06 kalev: No problem; I know the feeling ;-) Nov 03 14:42:13 pb__: The dependency for xproto is coming from xlib. The dependency of xproto-dev on xproto is coming from the packaging code itself, it hard-codes a dependency on "foo ( =1.2.3)" for each "foo-dev". I agree that this is somewhat silly, but that last bit is what's coded into the class itself. And the ALLOW_EMPTY is what has been suggested as a solution to this problem (unsatisfied dependencies) on the OE mailing list previously. Nov 03 14:43:09 mwester: with recent (last few weeks) metadata from git, you shouldn't be getting a dependency of foo-dev on foo if foo is empty. Nov 03 14:43:09 This is a wronkaround in my POV, not a solution... Nov 03 14:43:34 mwester: (search for "main_is_empty" in package.bbclass for the code which is meant to deal with that case) Nov 03 14:43:54 if that isn't working, I think we should just fix package.bbclass so that it does, rather than generating a load more empty packages. Nov 03 14:44:09 pb__: I fully agree on that Nov 03 14:44:23 I agree but I have an immediate problem, and my packaging/python foo is inadequate. Nov 03 14:44:30 * otavio agrees even though he is not a core member Nov 03 14:44:52 mwester: you can solve your immediate problem by putting the ALLOW_EMPTY lines in your local.conf, or your own DISTRO.conf. Nov 03 14:45:50 if you start checking that kind of thing in then it tends to snowball: anybody else who has a similar problem will say "ah, mwester fixed it like this and he knows what he's doing, so that must be the right thing to do". Nov 03 14:47:02 not that it's a terribly big deal, but it would be nice to avoid too much of this stuff if possible. empty packages do have a cost (albeit a small one) in terms of Packages file clutter, opkg memory footprint and speed, and so on. Nov 03 14:47:11 Right. Or perhaps someone who understands how that all works can see from my emails to the list that it is broken, and can take a more active role in fixing it. :) Nov 03 14:47:18 right :-) Nov 03 14:47:46 if you are sure that it is broken in the current head, I will have a look at it this afternoon. Nov 03 14:47:51 pb__: are you going to take a more active role? :P Nov 03 14:48:09 pb__: already answered my question :P LOL Nov 03 14:48:26 But either way, the fact remains that DEBPV does not currently take DISTRO_PR into account, and packages that use it to write dependencies will write incorrect dependencies. Nov 03 14:48:27 heh Nov 03 14:48:39 right, that is a separate issue Nov 03 14:48:56 I did think that one was already fixed as well (see zecke's comments) but I can believe that it might not be. Nov 03 14:49:26 mwester: It is fixed in HEAD Nov 03 14:49:37 otavio: I'm using HEAD. Nov 03 14:49:42 Or do you mean bitbake? Nov 03 14:49:46 mwester: test current git and comment on it. I and RP was involved on the fix. Nov 03 14:50:15 mwester: but the biggest issue is that just rebuilding the package doesn't solve it; I've done a from scratch build to be sure of it Nov 03 14:50:15 oh, heh. the empty package thing is also caused by DEBPV Nov 03 14:50:29 My last git pull on that tree was no later than two days ago. Nov 03 14:50:42 mwester: you doesn't look to be using it since DEBPV doesn't exist anymore Nov 03 14:50:45 mwester: can you try adding an appropriate DISTRO_PR expression to line 462 of package.bbclass and see if that clears up your problem with empty packages? Nov 03 14:51:10 it should be fairly obvious from the contents of that line what sort of thing is required Nov 03 14:51:25 mwester: there's a missing DEVPV in package.bbclass but it has been change in bitbake.conf Nov 03 14:52:17 So this is a bitbake problem. Nov 03 14:53:25 pb__: Adding the DISTRO_PR to Line 462 is EXACTLY the patch I put in the email. Nov 03 14:53:48 oh, heh Nov 03 14:54:25 That fixes it so that the dependency on "package" written out into "package-dev" is correct, but it still requires that "package" exist -- hence the ALLOW_EMPTY is also needed. Nov 03 14:54:48 mwester: er, really? the expression in line 462 is never written into any package. Nov 03 14:54:53 all it is used for is to _remove_ dependencies. Nov 03 14:55:07 pb__: I believe I fixed it locally Nov 03 14:55:10 pb__: let me push it Nov 03 14:55:18 otavio: ok, great Nov 03 14:55:49 mwester: the dependency of -dev on the main package is done in a completely different place, that one comes from bitbake.conf. Nov 03 14:55:54 03Otavio Salvador  07org.openembedded.dev * rd14249e378 10openembedded.git/ (3 files in 3 dirs): replace remaining DEBPV variables with EXTENDPV Nov 03 14:56:01 mwester: give it a try Nov 03 14:56:16 pb__: Line 480 rewrites the dependency. Nov 03 14:56:29 mwester: right, but only by removing things. Nov 03 14:56:48 mwester: follow the code, look at what "remstr" is used for. the only place it is mentioned is in a removal operation. Nov 03 14:56:56 moin Nov 03 14:57:01 hi zecke Nov 03 14:57:39 pb__: not for me. It does not remove the dependency if main is empty; and it changes the dependency instead. My python foo is weak, but I know what I see when I look at the "control" file written into the packages... Nov 03 14:57:53 mwester: I think you are mixing up two different things. Nov 03 14:57:54 mwester: use the git head of now Nov 03 14:57:58 But I'll have to work on this later, right now I must go off to a meeting. Thanks for the help. Nov 03 14:58:01 mwester: it had a missing replace Nov 03 14:58:14 mwester: I think it is fixed now Nov 03 14:59:16 03Otavio Salvador  07org.openembedded.dev * rf23ab98cc7 10openembedded.git/packages/vnc/ (libvncserver/autoreconf.patch libvncserver_0.9.1.bb): Nov 03 14:59:16 libvncserver_0.9.1: fix building using libtool 2.2.4 Nov 03 14:59:16 To allow building using new libtool the whole source needed to be Nov 03 14:59:16 updated. To make it easier we've updated all building scripts using Nov 03 14:59:16 autoreconf and dropping previous included libtool m4 script from Nov 03 14:59:20 acinclude.m4. Nov 03 14:59:22 03Otavio Salvador  07org.openembedded.dev * redbb2e0541 10openembedded.git/packages/xfce/thunar_0.9.0.bb: thunar: depends on gamin to use inotify Nov 03 15:01:22 otavio: right, that looks like the correct kind of thing Nov 03 15:01:56 pb__: yes. It was not running since it was not matching. I can't test it right now but it was obviously broken Nov 03 15:02:21 mwester: Revision: d14249e378 should do it for you Nov 03 15:07:23 hi kergoth Nov 03 15:12:42 huh, xf86-input-evdev has runtime calibration and axis swapping? wonder how its calibration works Nov 03 15:22:50 03Michael 'Mickey' Lauer  07org.openembedded.dev * r8156545fea 10openembedded.git/packages/tasks/task-cli-tools.bb: task-cli-tools: ship utils that require python seperately (for system with low space requirements) Nov 03 15:23:00 03Michael 'Mickey' Lauer  07org.openembedded.dev * r4037943fc2 10openembedded.git/packages/images/ (fso-console-image.bb fso-image.bb): fso-[console-]image: WIP towards using more tasks Nov 03 15:23:01 03Michael 'Mickey' Lauer  07org.openembedded.dev * rf5d962579b 10openembedded.git/packages/ (python/python-fugrep_0.50.bb tasks/task-python-everything.bb): Nov 03 15:23:01 python-fugrep 0.50: new recipe; a python library for fuzzy pattern matching. Nov 03 15:23:01 * add to task-python-everything Nov 03 15:32:44 khem` ? Nov 03 15:36:25 03Holger Hans Peter Freyther  07org.openembedded.dev * rc0fcc31726 10openembedded.git/: Merge branch 'holger/enterprise' into org.openembedded.dev Nov 03 15:36:26 03Holger Hans Peter Freyther  07org.openembedded.dev * r0d2c29eaff 10openembedded.git/ (53 files in 17 dirs): (log message trimmed) Nov 03 15:36:26 bluemchen wiese: Allow to disable certain ugly packages globally Nov 03 15:36:26 Adopt the notion of gstreamer and allow to easily not build certain Nov 03 15:36:28 ugly packages. People are constraints to not live on a bluemchen Nov 03 15:36:30 wiese can set ENTERPRISE_DISTRO="1" in their config and get ugly Nov 03 15:36:32 packages removed. Nov 03 15:36:34 This is implemented by a set of packages in the BBMASK (as a safety Nov 03 15:36:36 03Holger Hans Peter Freyther  07org.openembedded.dev * rf0573933b8 10openembedded.git/: Nov 03 15:36:39 Merge branch 'org.openembedded.dev' of git+ssh://git@git.openembedded.net/openembedded into org.openembedded.dev Nov 03 15:36:41 Conflicts: Nov 03 15:36:43 packages/images/fso-image.bb Nov 03 15:38:57 *sigh* Nov 03 15:39:03 http://gitweb.openembedded.net/?p=openembedded.git;a=commitdiff;h=aed79502dc797275531dd1cb37dcf5c6f013034a broke C++ in SDKs Nov 03 15:39:20 by deciding C++ isn't necessary apprently Nov 03 15:39:49 tartarus hm Nov 03 15:39:54 (And not moving it to another task-sdk) Nov 03 15:41:02 perhaps adding task-sdk-c++ to task-sdk-extra.bb would be "right" Nov 03 15:41:17 Tartarus: hehe, patches welcome Nov 03 15:42:21 * Tartarus hmms Nov 03 15:42:42 Tartarus hm try and please send a patch Nov 03 15:43:12 Actually Nov 03 15:43:48 task-sdk-bare has it, now.. Nov 03 15:44:11 side-effect of how i've got my sdk setup i guesss Nov 03 15:44:15 sorry for the false alarm :) Nov 03 16:03:51 hehe, of course, task-sdk.bb is just ${LIBC}, ${LIBC}-dev and libgcc, for task-sdk-bare, but task-sdk-bare.bb (which gets used) is glibc-specific Nov 03 16:06:52 * Tartarus looks for a bug Nov 03 16:24:22 RP: r1090 of bitbake broke mine SSH agent to work; Could you take a look on that? Nov 03 16:32:10 RP: if I revert it, it goes back to work properly Nov 03 16:32:29 RP: mine local bitbake is svn HEAD with r1090 reverted. Nov 03 16:41:31 Bah Nov 03 16:41:54 How do I, someone without commit access, show a local git commit in a way I can post on the bugzlla? Nov 03 16:42:08 Should I just post a diff from previous to my rev? Nov 03 16:58:22 OK, all that sdk stuff in a patch on a bug Nov 03 16:58:27 4797 Nov 03 16:58:49 * Tartarus has a uclibc build going now to see if there's anything that should get added in on that side of the equation Nov 03 17:12:06 Tartarus, any chance you played with g++ code under OE? Both gcc 4.2.4 and 4.1.1 on ppc630e get an infinite recursion segfault calling floor(). DOe snot happen when compiled with gcc Nov 03 17:14:02 waite, only in so far as the SDK stuff wasn't right for me at one point :) Nov 03 17:15:23 This is even with the internal to OE compiler. Not even treading into the meta-toolchain/sdk stuff Nov 03 17:15:23 hi woglinde Nov 03 17:24:41 hi guillaum1 Nov 03 17:29:11 hi Nov 03 17:29:25 how can i change the boot splash , on angstrom ? Nov 03 17:29:42 boris_OmegA: change psplash package Nov 03 17:31:50 thanks otavio Nov 03 17:32:17 but are you sure ? it's the kernel that display this splash while booting Nov 03 17:33:18 ok thanks otavio :) Nov 03 18:01:20 ~ping Crofton Nov 03 18:01:21 pong Crofton Nov 03 18:04:16 zecke: good evening Nov 03 18:09:30 hmm Nov 03 18:09:56 hi Nov 03 18:14:06 http://wiki.openembedded.net/index.php/New_Dev should probably be updated or deleted Nov 03 18:15:20 hmm, OpenEmbedded could really use a Contributing page. where would, say, someone who wants to improve our qa or documentation, go on the site to get info on where to go? Or someone who wants to contribute financially Nov 03 18:16:02 um money Nov 03 18:16:13 Send all beer to.... Nov 03 18:43:30 re Nov 03 18:48:23 wow, finally xfdesktop is fixed :-D Nov 03 18:48:49 otavio: cool Nov 03 18:49:00 :-) Nov 03 18:49:36 otavio: so after years since AMD^H^H^H CosmicPenguin added xfce and my updates you got it updated and running Nov 03 18:50:15 hehe Nov 03 18:50:17 gratz Nov 03 18:52:36 hey otavio Nov 03 18:52:53 That bug mwester found, was the fix doing something so that ${DISTRO_PR} ends up in ${DEBPV} all the time? Nov 03 18:54:08 Tartarus: it needs to be in DEBPV; well, in EXTENDPV now since it was renamed Nov 03 18:54:28 hrw: hahah Nov 03 18:54:37 otavio, k Nov 03 18:54:51 Then yes, that should fix that section :) Nov 03 18:54:55 * Tartarus deletes the thread now Nov 03 18:55:19 * mwester will test after work today. Nov 03 18:55:48 But I am interested if the usage of DEBPV in the two bb recipes should be replaced with EXTENDPV as well. Nov 03 18:56:12 Tartarus: but his bug are two in reality. One is the DEBPV^W EXTENDPV that needed fixes and this was done early last week and finished today since few places were not updated to use it. Another is the ALLOW_EMPTY issue. Nov 03 18:56:28 otavio, which ALLOW_EMPTY issue? Nov 03 18:56:43 Packages shouldn't need ALLOW_EMPTY anymore, except for tasks and such Nov 03 18:56:48 mwester: yes, they need to be replaced since DEBPV was remove. If it should use another value then this is another bug. Nov 03 18:56:50 Hmm.. I thought that pb__ was arguing that if the EXTENDPV fix works, then the ALLOW_EMPTY should not be required? Nov 03 18:57:13 Where mwester's patch was, was to the area I got added like a month ago to strip the erronious dep out Nov 03 18:57:14 mwester: It should fix this issue as well but if it doesn't then it is another bug. Nov 03 18:57:40 I'll make a note to revert the ALLOW_EMPTY changes, rebuild, and test that too, then. Nov 03 18:57:46 Thanks. :) Nov 03 18:57:56 mwester: please do that and let people know. Nov 03 19:24:17 zecke: mind changing git-log and git-barnch to git log and git branch in base.bbclass for me? don't think i have git push access yet. Nov 03 19:24:33 kergoth: git 1.6 user :) Nov 03 19:24:39 :) Nov 03 19:24:42 kergoth: send a key Nov 03 19:24:46 k Nov 03 19:25:03 * kergoth wonders where which backup drive has his key Nov 03 19:25:50 bye Nov 03 19:27:00 bye hrw Nov 03 19:30:19 zecke: will there be use of signed/acked/etc? i see some people signing off and some not, in our git tree currently Nov 03 19:30:54 kergoth: no idea, maybe. we have that in the wiki but no defined workflow Nov 03 19:31:01 kergoth: e.g. I would Ack your change Nov 03 19:31:15 kergoth: did you send your key yet? you could pastebin it as well Nov 03 19:31:43 yeah. i like it, acked and reviewed by especially. SOB isnt as big a deal yet for us, since its primarily to show show patches flowed through, and right now we don't break out subsystems or machines by repository or anything Nov 03 19:31:53 s/show show/show who/ Nov 03 19:32:37 damnit, i can't find my main ssh key. /me generates another set Nov 03 19:32:55 hehe Nov 03 19:33:09 wb kergoth Nov 03 19:33:21 wait, i lied, its on my usb key, i switchd to using just one key instead of per machine Nov 03 19:33:23 :P Nov 03 19:33:42 s/usb key/usb stick/ Nov 03 19:33:46 gods, i can't think/type today Nov 03 19:34:16 i should get a new usb stck at some point, this one has the same storage as my ipod shuffle, getting old :) Nov 03 19:34:21 course its not like i use it much Nov 03 19:34:26 ♥ Dropbox (getdropbox.com) Nov 03 19:36:43 zecke: sent Nov 03 19:36:56 zecke: please take a look at http://arquivos.ossystems.com.br/~otavio/bitbake_ssh_support_for_git_fetcher.diff Nov 03 19:37:31 zecke: this is what I've been using for a while without a problem; I'd like to merge it so I could use "upstream" code ;-) Nov 03 19:38:23 zecke: does OpenEmbedded have a branch/tag policy right now? Nov 03 19:38:41 kergoth: kergoth/* is your domain Nov 03 19:39:13 otavio: you don't need. SRC_URI = "git://git.foo.com;protocol=ssh" should do the trick Nov 03 19:39:21 cool. what about for staging areas for machine/distro maintainers? machine/*, distro/*, etc? could see that being useful, for stuff that isnt quite ready for primetime Nov 03 19:39:24 otavio: e.g. we use it on the moblin servers to use http Nov 03 19:39:46 kergoth: no policy right now Nov 03 19:39:51 k Nov 03 19:40:22 kergoth: I'm used to shared/foo-work but that is no policy... or that is my "Use my way or you bold, gray and stupid"-policy Nov 03 19:40:26 * kergoth thinks we need better maintainer policy in general, people should be able to very easily determine the contact points for current machine or distro maintainers Nov 03 19:40:29 hehe Nov 03 19:40:53 zecke: I'll try it; hold one Nov 03 19:40:55 on Nov 03 19:42:11 zecke: git ls-remote git://projetos.ossystems.com.br/git/liboslicense.git master failed with signal 1, Nov 03 19:42:18 zecke: does not work Nov 03 19:45:33 Hello. I have following problem during building gcc-cross_4.2.4: configure: error: GMP 4.1 and MPFR 2.2.1 or newer versions required by fortran, but some thing wrong, the newer gmp and mprf are builded yet. What i need to check? Nov 03 19:49:02 Kalemas can you paste the config.log from inside gcc? Nov 03 19:50:01 otavio: show me the bb file Nov 03 19:50:04 kergoth: give it a try Nov 03 19:51:16 god, i just went to ssh-add the kye, and tried to git add it instead, out of habit Nov 03 19:51:19 that's sad Nov 03 19:51:20 http://en.pastebin.ca/1244255 Nov 03 19:51:54 zecke: http://pastebin.com/f15da23cd Nov 03 19:54:29 zecke: hmm, public key auth not working here. the user/host is git@git.openembedded.net? Nov 03 19:55:29 kergoth: yes Nov 03 19:56:07 kergoth: ssh -vvv do you offer the key? Nov 03 19:56:15 yep Nov 03 19:56:29 wait, hmm Nov 03 19:56:38 ah, yes, it is Nov 03 19:56:44 debug1: Offering public key: /Users/kergoth/.ssh/id_rsa Nov 03 19:56:44 debug2: we sent a publickey packet, wait for reply Nov 03 19:56:51 debug1: Next authentication method: password Nov 03 19:57:03 hehe rsa key Nov 03 19:57:20 otavio: I have a ssh connection here Nov 03 19:58:34 zecke: ouch! I forgot to revert my changes in git fetcher Nov 03 19:58:41 zecke: shame on me! Nov 03 19:59:00 zecke: i'll give it another try once I've finished firefox built Nov 03 19:59:15 otavio: :) Nov 03 19:59:21 kergoth: again? Nov 03 20:02:58 grr Nov 03 20:03:01 * kergoth scratches head Nov 03 20:03:23 kergoth: another key? Nov 03 20:03:53 is this a weak debian key? Nov 03 20:04:06 zecke *g* Nov 03 20:11:44 kergoth: okay Nov 03 20:12:11 okay, now it just hates me Nov 03 20:12:18 Seeker% git fetch ssh Nov 03 20:12:18 Connection closed by 80.81.242.146 Nov 03 20:12:18 fatal: The remote end hung up unexpectedly Nov 03 20:12:24 hehe Nov 03 20:12:28 key works now though Nov 03 20:12:58 kergoth: paste the output of gitosis Nov 03 20:13:26 Seeker% ssh git@git.openembedded.net Nov 03 20:13:26 Connection closed by 80.81.242.146 Nov 03 20:13:37 no output at all that i can see Nov 03 20:13:43 kergoth: weird Nov 03 20:13:49 see, it hates me :) Nov 03 20:14:05 kergoth: it is still sshd level Nov 03 20:14:40 good point, i think it is after double checking the -vvv output Nov 03 20:14:52 sends the publickey packet, then an immediate disconnect Nov 03 20:15:00 doesn't fall back to password, just immediate dc Nov 03 20:15:04 never allocates a tty or anything Nov 03 20:15:22 * kergoth doesn't think he's seen this before Nov 03 20:16:36 totally OT, but anyone have an opinion on good mind mapping tools, either app or webapp? Nov 03 20:18:55 hm didnt work some time ago with a windows tool Nov 03 20:19:08 but google should find quick a good one Nov 03 20:19:16 * kergoth likes it for brainstorming and keeping track of certain things.. outlining is so limited Nov 03 20:19:19 let me examine apt-cache Nov 03 20:19:32 google finds lots, i'm just curious about which of them people like :) Nov 03 20:19:46 kergoth: pastebin another key? Nov 03 20:20:03 third time's the charm? Nov 03 20:20:19 kergoth: sure Nov 03 20:21:11 this new key 'randomart image' is so silly Nov 03 20:22:47 ping khem Nov 03 20:23:28 kergoth: try again Nov 03 20:23:31 kergoth hm I thought google finds this too Nov 03 20:23:44 kergoth I just tested kdissert now Nov 03 20:23:47 seems okay Nov 03 20:23:49 perhaps, i'll look again Nov 03 20:24:12 mindmeister and mind42 aren't bad, webapps, but they both limit the # of maps for the free incarnation Nov 03 20:24:15 kergoth: the java mindmapper is pretty classic Nov 03 20:24:53 yes its java Nov 03 20:24:57 kdissert is kde Nov 03 20:38:03 i ♥ git clean Nov 03 20:41:04 i have no any idea how to fix it... Nov 03 20:41:13 mickeyl: awake? Nov 03 20:41:39 * zecke has enough of ssh over GSM... and declares gitosis broken Nov 03 20:44:39 hehe Nov 03 20:44:52 maybee I should take over Nov 03 20:46:29 kergoth: try again Nov 03 20:47:48 zecke: works great now, thannks Nov 03 20:48:24 zecke: any reason the org.* weren't renamed when we moved to oe? seems a bit redundant Nov 03 20:48:32 s/^org.openembedded.//g Nov 03 20:49:01 anybody can help? Nov 03 20:49:46 huh, gmane's view of the oe-devel list is awfully slow to update Nov 03 20:50:19 Kalemas you didnt paste the config.log from inside gcc Nov 03 20:50:50 so we cant see why configure is failing at checking for any version of mpfr.h... no Nov 03 20:51:04 03Chris Larson  07org.openembedded.dev * r3ff5572d2b 10openembedded.git/classes/base.bbclass: base.bbclass: use 'git log' & 'git branch', not git-log & git-branch. Nov 03 20:53:28 zecke: wonder if there's any way to get CIA to mangle emails when it dumps to irc Nov 03 20:53:39 zecke: I'd hate to see anyones real addresses in web irc logs from people's bots Nov 03 20:53:44 in general, not just OpenEmbedded Nov 03 20:53:46 heh Nov 03 20:54:09 guess we could do it in the hook, just cia.vc wouldn't have the real info Nov 03 20:54:16 that might be best Nov 03 20:54:18 kergoth yes good idea Nov 03 20:55:04 * kergoth 's kergoth.com addy is generally for personal mail, people emailing him directly, give another addy to companies to try to limit spam through the former Nov 03 21:08:56 woglinde: http://en.pastebin.ca/1244316 it is located /home/kalemas/oe/stuff/build/tmp/work/armv5te-angstrom-linux-gnueabi/gcc-cross-4.2.4-r3/gcc-4.2.4/build.i686-linux.arm-angstrom-linux-gnueabi/config.log Nov 03 21:13:23 woglinde: it is only one "config.log" in /home/kalemas/oe/stuff/build/tmp/work/armv5te-angstrom-linux-gnueabi/gcc-cross-4.2.4-r3 Nov 03 21:19:01 kalemas hm Nov 03 21:19:04 Im puzzeld Nov 03 21:19:07 /usr/bin/ld: cannot find -lc Nov 03 21:19:19 /usr/bin/ld? Nov 03 21:19:32 yes Nov 03 21:19:36 gcc-cross Nov 03 21:19:41 this is right Nov 03 21:31:54 denix: whats up Nov 03 21:35:49 is someone using tightvnc-viewer? Nov 03 21:36:04 i'm having trouble with it since it required a helvetica font Nov 03 21:45:49 khem: re Nov 03 21:46:27 khem: what do you think about ticket #4798? Nov 03 21:47:52 it start building all again, i'll back after few days... Nov 03 21:48:19 khem: it fixes glibc for i686/pentiumpro/pentium3 archs, uses your patch for eglibc :) Nov 03 21:50:03 denix: You can apply it to glibc too Nov 03 21:50:19 khem: that's what I did Nov 03 21:50:19 khem *g* thats what he is wrtiting Nov 03 21:50:29 khem please put it in Nov 03 21:50:45 ok Nov 03 21:51:00 woglinde_: thanks Nov 03 21:52:17 khem, do you look at SDK stuff? 4797 :) Nov 03 21:54:34 Tartarus: patch looks sane to me Nov 03 21:56:14 Tartarus: may be you should also add LIBC_PKGS_linux-gnu = "${GLIBC_PKGS}" Nov 03 21:56:47 Do we have any of those targets? Nov 03 21:57:09 trini@carnation:~/cur-oe/openembedded/conf/distro$ grep -r -- -gnu . | grep -v eabi Nov 03 21:57:10 ./include/insane-srcrevs.inc:SRCREV_pn-gnuradio ?= "${AUTOREV}" Nov 03 21:57:10 ./include/angstrom-eglibc.inc:#TARGET_OS = "linux-gnuspe" Nov 03 21:57:10 ./include/sane-srcrevs.inc:SRCREV_pn-gnuradio ?= "9809" Nov 03 21:58:21 Tartarus: usually not but its nice to have it sometimes linux-gnu might be used Nov 03 22:00:07 I dunno, I think a buncha other stuff goes unhappy first :) Nov 03 22:00:15 Such as sane.bbclass target mapping Nov 03 22:00:20 (or is that in insane, I forget) Nov 03 22:01:01 its just a nice to have Nov 03 22:02:19 maybe its used in autoconf, but not in oe. maybe if you run into a Linux machine that has no gnu software someday.. Nov 03 22:11:37 kergoth: yes thats right, I think there are not so many differences. GCC used this triplet to identify some startup code in configs but that might also not be used anymore Nov 03 22:12:17 even if it does, it's likely to be a distro decision, which would be used, and it could be manged in the packages affected by it Nov 03 22:12:23 it neednt be a new TARGET_OS Nov 03 22:12:30 * kergoth gets caffeine Nov 03 22:13:10 kergoth: yes Nov 03 22:14:57 one area where I think we need to work on is toolchain staging. as toolchain has some intermediate components it becomes hard when you have a completely built temp and you do a change to say gcc or glibc then deps dont come out right Nov 03 22:15:59 We need to tell to bitbake that some packages are obsoleted by another package if available like gcc-initial is not required to be rebuilt Nov 03 22:16:10 or may be toolchain should be staged separately Nov 03 22:18:02 agreed. our depends are kind of limited in general, doesn't describe all the relationships fully Nov 03 22:19:27 We need someway to denote this chicken egg problem Nov 03 22:27:58 hi robert Nov 03 22:28:27 yay, bitbake has finally worked Nov 03 22:28:47 oliford good Nov 03 22:29:02 after several complication resulting from my work's proxy deciding that the gpe-theme-clearlooks tarball was porn or something Nov 03 22:29:17 hi rschuster Nov 03 22:30:22 oliford: lol Nov 03 22:30:38 i remember at my old job their stupid filter blocked wikipedia for like 3 weeks Nov 03 22:30:41 hi kgilmer Nov 03 22:30:42 they had some unhappy engineers Nov 03 22:31:05 unfortunately the proxy returns http ok, which meant all I saw was MD5 errors Nov 03 22:31:19 hey wogline_ Nov 03 22:32:08 i'm gonna put this on the ml but why not here too. Working on OE/Eclipse integration plugins: http://community.buglabs.net/kgilmer/posts/3-OpenEmbedded-Tools-for-Eclipse Nov 03 22:32:57 cool! Nov 03 22:33:07 kgilmer: nice Nov 03 22:33:49 thx guys. still pretty basic but usable. Nov 03 22:35:17 any reason why it might have not set /bin/su suid root? Nov 03 22:35:27 (angstrom gpe-x11-image) Nov 03 22:35:35 kgilmer: very cool Nov 03 22:35:57 thx kergoth :) Nov 03 22:36:13 kgilmer: great stuff. Eclipse integration is a very nice thing! Nov 03 22:37:02 yeah there are some nice things out-of-the-box, like visual diff. Nov 03 22:37:32 hope to have it really polished up for fosdem. Nov 03 22:38:14 am I doing smth wrong if my kernel headers are in staging/${MACHINE}-linux-gnueabi, but apps look for them in staging/${ARCH}-linux-gnueabi? Nov 03 22:38:25 I've just started developing with gtk in eclipse and eventually want to build an app for my ipaq, having OE in eclipse would be awsome Nov 03 22:39:14 oliford, the plugins now support using oe crosscompiler. it's not wrapped up yet in a nice UI but we'll have that shortly. Nov 03 22:39:43 cool, wd Nov 03 22:40:07 basically you need to set the path to the crosscompiler and the root of include dirs. I already have this in the bitbake work i've done, just needs to be integrated so it 'just works' out of the box. Nov 03 22:40:38 thx oliford Nov 03 22:41:27 kgilmer: this reminds me, that we (= OE) should ask for a developer room and not only a booth Nov 03 22:42:04 zecke_: any idea what the checkonly bit in the fetchers is? Nov 03 22:42:55 kergoth: for git? Nov 03 22:43:01 in general. Nov 03 22:43:09 it's passed to the fetch method or something, uses CHECKCOMMAND Nov 03 22:43:23 passed to the go() Nov 03 22:43:24 kergoth: nope, I think it is there from bk times? Nov 03 22:43:30 hmm could be Nov 03 22:43:32 rschuster: sounds good. also if I can be of help on Jalimo let me know. Nov 03 22:43:39 are you planing on a talk? Nov 03 22:43:40 zecke_: testing a fetcher that uses urllib Nov 03 22:43:41 heh Nov 03 22:45:01 denix: the kernel headers should be installed in staging/${ARCH}-linux-gnueabi Nov 03 22:49:05 khem: thanks, I suspected so, but cannot find why they are installed in staging/${MACHINE}-linux-gnueabi/kernel... Nov 03 22:50:19 denix: Do you have build logs handy ? Nov 03 22:52:52 denix: linux-libc-headers package provides them so check if thats built correctly Nov 03 22:54:15 khem: aha, that would explain it. linux-libc-headers is not built, but provided by external-toolchain... Nov 03 22:57:41 mickeyl: good evening Nov 03 22:58:40 hi pb Nov 03 22:58:51 hi woglinde_ Nov 03 23:03:10 hmm, why does bitbake get upset when you move the TMPDIR? Nov 03 23:03:31 (i mean move it and change the TMPDIR var accordingly) Nov 03 23:03:41 bitbake itself shouldn't care, though oe might. Nov 03 23:03:47 what's the actual issue you're seeing? Nov 03 23:03:53 err, yea, the sanity checker does Nov 03 23:04:06 khem: so, with external-toolchain, there needs to be extra -I${SDK_PREFIX}/.../usr/include, right? Do I need to set TARGET_CPPFLAGS_append? Nov 03 23:04:10 hmm, this is odd Nov 03 23:04:12 * kergoth kicks python Nov 03 23:04:22 line 29, in Nov 03 23:04:22 import sys Nov 03 23:04:22 ImportError: cannot import name urlopen Nov 03 23:04:35 kergoth: can i join in in the kicking? Nov 03 23:04:36 * kergoth scratches head Nov 03 23:04:37 kergoth: that is rather weird Nov 03 23:04:38 it said "Error, TMPDIR has changed location. You need to either move it back or rebuild" Nov 03 23:04:39 please do Nov 03 23:04:52 denix: yes Nov 03 23:05:03 oliford: ah, right. I think that's because various things in TMPDIR contain absolute path references to other things in staging. Nov 03 23:05:16 that's an oe issue rather than a bitbake one, though Nov 03 23:05:22 oh, i'll leave it be then Nov 03 23:05:32 yeah, probably best Nov 03 23:05:42 fixing it would be quite hard, almost certainly not worth the effort Nov 03 23:05:48 pb_: this imports fine from a commandline interactive shell, just not from a fetch in lib/bb Nov 03 23:05:49 oddly Nov 03 23:06:33 kergoth: strange. no idea what would cause that though, I guess this would be a task for mickeyl :-} Nov 03 23:06:47 kergoth: by the way, congratulations, I see you updated your blog! Nov 03 23:06:55 heh Nov 03 23:06:56 hehe, indeed Nov 03 23:07:20 that's almost one post a month, a pretty respectable average Nov 03 23:08:07 nite guys Nov 03 23:08:23 zecke_: later Nov 03 23:21:05 nite Nov 03 23:21:06 question! Nov 03 23:21:10 okay Nov 03 23:21:12 hubar Nov 03 23:21:13 hi wog! Nov 03 23:21:15 last questin Nov 03 23:21:20 :0 Nov 03 23:21:36 I am deleting a patch applied to apache2. Nov 03 23:21:54 What files do I need to modify? Nov 03 23:22:02 Other than a2.bb? Nov 03 23:22:20 aeh Nov 03 23:22:23 none? Nov 03 23:22:26 I whacked the line with the patch in a2.bb. Nov 03 23:22:48 hmm, but I always get this error Nov 03 23:22:49 should the patch go upstrea? Nov 03 23:23:11 NOTE: make DESTDIR=/home/wweng/tmp/work/i486-linux/apache2-2.2.3-r5/image install Nov 03 23:23:13 make: *** No rule to make target `install'. Stop. Nov 03 23:23:14 FATAL: oe_runmake failed Nov 03 23:23:26 hm Nov 03 23:23:26 what do you mean "go upstream"? Nov 03 23:23:33 hubar into oe Nov 03 23:23:47 this seems an error in the apache Makefiles Nov 03 23:24:10 hmmm. Nov 03 23:24:19 and I have no time yet looking into it Nov 03 23:24:22 good nite Nov 03 23:24:25 How can I find out? Nov 03 23:24:27 okay! Nov 03 23:24:37 woglinde_: but if we call install? Nov 03 23:24:37 Thanks for the help already! Nov 03 23:24:49 looking in the topdir Makefile of apache? Nov 03 23:25:17 apache/apr buildsystem suckz Nov 03 23:25:21 to much autotools abuse Nov 03 23:25:38 And I can not build apache. Nov 03 23:25:51 Though the package is there. Nov 03 23:25:57 zecke till tomorrow Nov 03 23:27:14 TARGET_FPU = "" Nov 03 23:27:16 ERROR: Nothing PROVIDES 'apache' (but '[]' DEPENDS on or otherwise requires it) Nov 03 23:27:22 bye :) Nov 04 00:30:46 hello, i am flashing a gumstix (connex 400) over ssh/ethernet by following the guide at http://docwiki.gumstix.org/index.php/Replacing_the_filesystem_image#Flashing_over_SSH_.2F_ethernet Nov 04 00:31:44 i have updated u-boot with the latest one that i could find - u-boot-connex-400-r1604.bin Nov 04 00:32:32 the guide asks to flash the rootfs and kernel u-image using # echo && flashcp -v /tmp/rootfs.arm_nofpu.jffs2 /dev/mtd1 && flashcp /tmp/uImage /dev/mtd2 && echo > /dev/watchdog Nov 04 00:32:43 but i don't have /dev/mtd2 Nov 04 00:32:55 * abbyz is all confused Nov 04 01:16:37 urg Nov 04 01:16:41 RP, mickeyl: ping Nov 04 01:17:18 they may be in bad .... Nov 04 01:17:29 bed Nov 04 01:17:32 well Nov 04 01:17:46 * kergoth doesn't really recall the time difference offhand Nov 04 01:17:46 they are 6-7 hours from here Nov 04 01:18:02 ah Nov 04 01:18:16 rp is one hour closer Nov 04 01:18:21 apparently some syscalls shouldn't be run after a fork but before an exec Nov 04 01:18:37 and our runqueue does a whole lot of that Nov 04 01:18:41 http://www.worldtimeserver.com/ Nov 04 01:18:49 so you're limited on what you can safely do in your python tasks Nov 04 01:18:51 weird Nov 04 01:18:57 on Mac OS X, it actually yells at you when you do it Nov 04 01:19:01 more picky, i guess Nov 04 01:19:13 see posix, "fork-safe" on functions Nov 04 01:19:28 anyway, we should probably rework it to use threads or something Nov 04 01:19:28 hmmm Nov 04 01:19:39 from a correctness standpoint, anyway Nov 04 01:20:03 yeah Nov 04 01:20:23 * kergoth was trying to make a fetcher that uses python's urllib/urllib2 modules instead of wget, but it freaks Nov 04 01:33:23 I am having trouble building a C++ application on powerpc using OE and I get an infinite recursion problem calling floor(). Nov 04 01:33:37 ANyone here a guru with g++ configurations? Nov 04 02:15:53 cbrake, thanks for the weekly changelog, again Nov 04 02:29:07 Crofton: sure, its fun to see everything that is happening Nov 04 02:29:15 Crofton: wish I had more time right now ... Nov 04 02:29:37 every little bit helps Nov 04 02:30:29 I've been on the road for two weeks Nov 04 02:30:40 and will be on the road until the holiday Nov 04 02:30:41 vacation? Nov 04 02:30:45 work Nov 04 02:31:09 not great fun Nov 04 02:31:16 but it will cover lots of bills :) Nov 04 02:31:18 the glamor of business travel fades rather quickly. hope it's not too long a trip. Nov 04 02:34:01 should be back by Thanksgiving Nov 04 02:34:08 the time zones are the worst Nov 04 02:34:23 at least the difference is down to two hours now Nov 04 02:34:57 home weekends or working weekends? Nov 04 02:35:43 getting home is rough Nov 04 02:35:53 spend most of it traveling Nov 04 02:36:12 in two weeks, I leave here at 2PM and get home at midnight Nov 04 02:36:24 Yikes! Nov 04 02:36:28 if the last plane from Charlotte to Roanoke actually goes Nov 04 02:36:46 then to get back leave sunday around 2 and get here around 8:30 Nov 04 02:37:05 it sucks using a small airport Nov 04 02:37:12 I guess I shouldn't complain about O'Hare so much. It's always delayed, but at least it has a lot of flights. Nov 04 02:37:17 hhe Nov 04 02:37:18 yeah Nov 04 02:37:47 direct flights make a difference Nov 04 02:37:52 ok, need to find some food Nov 04 02:37:54 l8r **** ENDING LOGGING AT Tue Nov 04 02:59:57 2008