**** BEGIN LOGGING AT Fri Mar 20 03:00:02 2009 Mar 20 03:43:29 03Mike Westerhof  07org.openembedded.dev * r295488d84c 10openembedded.git/recipes/wireless-tools/ (wireless-tools/zzz-wireless.if-pre-up wireless-tools_29.bb): Mar 20 03:43:29 wireless-tools: do ifconfig down before the ifconfig up is done. Mar 20 03:43:29 Required to reliably re-initialize the interface if it was running previously. Mar 20 08:11:59 good morning Mar 20 08:14:52 good morging :) Mar 20 08:17:58 morning Mar 20 08:19:00 good morning Mar 20 08:19:35 I always used IMAGE_INSTALL += ... in a recipipe to install additional software into an image. Mar 20 08:19:41 but now bitbake fails: "Cannot find package setterm." Mar 20 08:19:44 What can I do? Mar 20 08:20:25 good morning Mar 20 08:58:21 can oe be distributed and packaged and sold? Mar 20 09:00:50 you mean the build environment, or the results of a build? Mar 20 09:00:58 both Mar 20 09:01:18 well, oe has its own licensing, and each package you can build with oe has its own licensing Mar 20 09:01:26 so you'd need to start reading licenses Mar 20 09:01:43 you can definately do that with GPL so long as you adhere to the terms Mar 20 09:03:14 good morning Mar 20 09:03:20 okay, i think handhelds companies use oe to generate thier file systems. and do these sell oe build environment and filesystem stuff. Mar 20 09:06:24 where can i find details about oe lincencing Mar 20 09:07:29 gx99: OE's metadata is licensed under the MIT license Mar 20 09:07:41 you will find that in the repository Mar 20 09:08:34 gx99: this is exactly like OE read here http://www.kaeilos.com/?q=mit-license Mar 20 09:08:36 gx99: basically the only restriction of the MIT license is that you can't tell, "our company invented the build system we used to generate the root filesystem of our devices" Mar 20 09:10:07 gx99: whatever is made out of OE is not subject to OE's license but the license of the respective package Mar 20 09:10:32 okay Mar 20 09:10:42 are there any ppl selling these Mar 20 09:11:10 gx99: sellin what? Mar 20 09:11:20 gx99: there are some recipes whcih have no source and consists of a bunch of scripts and config files. still this package has an explicit license Mar 20 09:11:23 OE build environment Mar 20 09:11:53 gx99: tar.cfj org.openembedded.dev | ? Mar 20 09:12:04 suppose i add my own meta files to build a custom filesystem Mar 20 09:12:12 gx99: we at KoanSoftware are selling commercial support for OE for example Mar 20 09:12:14 i might want to sell my own distro using oe Mar 20 09:12:57 gx99: you can do it if you distro use the same license Mar 20 09:13:28 i dont understand the licensing things Mar 20 09:13:51 gx99: a FOSS programmer must know licensing stuff Mar 20 09:14:30 gx99: your distro is a result/outcome. it can have whatever license you want (still respecting the license of individual packages) Mar 20 09:16:16 gx99: when it comes to the metadata the situation isn't so easy Mar 20 09:16:38 gx99: the MIT license does not require you to publish it Mar 20 09:17:02 however a packages' license might the be GPL (very likely :) ) Mar 20 09:17:35 and the GPL says, that you need to publish all scripts and whatever was needed to create the binary that a user got from you Mar 20 09:17:55 (so the user has the possibility to create the binary on its own) Mar 20 09:18:13 suffice to say a user can only do this if he/she has complete OE metadata Mar 20 09:18:22 gx99: however IANAL Mar 20 09:18:48 gx99: and you should not expect greate excitement among OE developers if you are hiding metadata from users .... Mar 20 09:19:39 gx99: if in doubt get in contact with the compliance lab of the fsf: legal@fsf.org, contacting SFLC may also be good Mar 20 09:42:36 okay, thanks Mar 20 10:00:15 hi Mar 20 10:02:01 why and where is the packages moved ? Mar 20 10:04:22 s/packages/openembedded-packages-directory/ Mar 20 10:05:04 the moved to the 'recipes' dir Mar 20 10:06:05 .. and why ? what was wrong with packages ? Mar 20 10:10:22 because it doesn't really contain the packages (which you ll install on your device) but the receipes to build these Mar 20 10:10:30 http://article.gmane.org/gmane.comp.handhelds.openembedded/21326 Mar 20 10:13:31 ohhh crap .. thats all .. and i have to find on all my buildenvironments packages and replace it ... to keep it running ... damm Mar 20 10:21:50 pwgen: this was discussed for long on the developer mailinglist Mar 20 10:22:00 yepp Mar 20 10:35:26 has the packages directory changed to recipes? Mar 20 10:36:43 yes Mar 20 10:37:57 morning Mar 20 10:46:13 if i want a angrstrom distibution do i need to use meta-angstrom-2007.bb Mar 20 10:49:18 gx99: what is it? Mar 20 10:55:02 gx99: what do you mean with 'need'? AFAIK in your local.conf there is no need to mention any meta* stuff Mar 20 11:29:07 03Jan Lübbe  07fso/milestone5.5 * r567f779c0a 10openembedded.git/ (4 files in 4 dirs): openobex: add version 1.5 Mar 20 11:29:17 03Jan Lübbe  07fso/milestone5.5 * r93086de724 10openembedded.git/ (conf/checksums.ini packages/bluez/bluez4_4.32.bb): bluez4: bump to 4.33 Mar 20 12:41:15 03Jan Lübbe  07fso/milestone5.5 * re367cc2a79 10openembedded.git/packages/bluez/bluez4_4.33.bb: bluez4: state the include files Mar 20 12:41:16 03Jan Lübbe  07fso/milestone5.5 * r400fab11d6 10openembedded.git/ (conf/checksums.ini packages/bluez/obexd_0.9.bb): obexd: new receipe Mar 20 12:41:20 03Jan Lübbe  07fso/milestone5.5 * r790d99300e 10openembedded.git/packages/tasks/task-base.bb: task-base: switch to bluez4 Mar 20 13:33:28 thebohemian: sorry was in a meeting, If i want angstrom image, which .bb file i have to use was my question Mar 20 13:33:40 any Mar 20 13:34:58 okay, are you saying that any image OE builds if my distro is set to angstrom is an angstrom image. Only the contents of the file system will change Mar 20 13:35:38 yes Mar 20 13:36:01 thank you Mar 20 13:36:21 RP: hi! Mar 20 13:36:45 RP: you ancient patch 01699d54e835aca39c96dc2edf8bdc6ca4afadcb to the ancient packages/glibc/glibc_2.3.5+cvs20050627.bb broke it: Mar 20 13:36:57 RP: "--with-headers=${STAGING_INCDIR \" needs a closing "}" Mar 20 13:37:25 hrw: i always thought that when ppl refer say it is a angstrom distribution, they always meant a particular contents of the file system. some meta .bb file which generate the desired file system Mar 20 13:37:30 but im wrong Mar 20 13:37:41 RP: and if you ever repair that, you might as well get rid of FILESDIR= and add Mar 20 13:37:49 FILESPATH = "${@base_set_filespath([ '${FILE_DIRNAME}/glibc-cvs-2.3.5', '${FILE_DIRNAME}/glibc-cvs', '${FILE_DIRNAME}/glibc', '${FILE_DIRNAME}/files', '${FILE_DIRNAME}' ], d)}" Mar 20 13:37:50 instead Mar 20 13:42:52 gx99: yes. the concept of a distro in OE is merely a certain set of preferred package versions Mar 20 13:44:32 oh, are you saying when i say angstrom all it means is im choosing desired version numbers for all these packages. if so where are they defined? Mar 20 13:45:05 gx99: basically yes. Mar 20 13:45:25 gx99: in conf/distro/ and below, search for PREFERRED in the various files that that relate to angstrom or get included by it Mar 20 13:45:25 they are defined in the distro config -> conf/distro/angstrom-*.conf Mar 20 13:45:56 it also consists of a few included files. you will understand it when you see it Mar 20 13:46:49 schurig: prepare patch and push it Mar 20 13:47:04 gx99: even better. when saying 'angstrom' it still is not decided whether you build uclibc or glibc. you can chose one but both are angstrom Mar 20 13:47:24 thebohemian: or eglibc even Mar 20 13:47:44 hrw: I don't think that I again got push permission for git. The last time I pushed was two years ago (or more), in the mtn itme. Do I need to send some SSH key somewhere? Mar 20 13:48:26 s/itme/time/ Mar 20 13:49:30 but the angstrom conf files only list version numbers for only some (say 10 or 15) of the packages, not all Mar 20 13:50:47 i will do some more reading and then ask questions Mar 20 13:51:01 are there any docs you recommend Mar 20 13:52:02 schurig: to mickeyl and/or koen Mar 20 13:52:05 gx99: that's why I wrote "or gets included by it", see for example line 26-29 of angstrom-2008.1.conf Mar 20 13:52:36 schurig: or prepare patch and send it to ML (or me) Mar 20 13:52:42 gx99: yeah, for the packages that are not specified in the config OE will choose the highest possible Mar 20 13:53:22 yes, thought so Mar 20 13:53:22 gx99: that can be annoying for users Mar 20 13:53:55 gx99: in poky the number of recipes is vastly reduced. I believe there you have full control about every package (AFAIK) Mar 20 13:54:06 okay Mar 20 13:56:32 got a question, my machine (efika) lists kernel linux-rt, which is 2.6.25, I've been running on this particular platform 2.6.27 with better results (basically better drivers), how can i override the kernel in my image recipe and/or pass different a different .config ? Mar 20 13:56:40 03Robert Schuster  07org.openembedded.dev * r387c5960a5 10openembedded.git/classes/cmake.bbclass: cmake.bbclass: Allow overriding path-to-source argument for cmake call. Mar 20 13:56:46 03Robert Schuster  07org.openembedded.dev * r4b9f02e641 10openembedded.git/classes/cmake.bbclass: cmake.bbclass: Put mandatory cmake option to end Mar 20 13:56:49 03Robert Schuster  07org.openembedded.dev * rf68981e239 10openembedded.git/conf/checksums.ini: checksums.ini: Added swt-gtk 3.4.2 entry. Mar 20 13:56:49 03Robert Schuster  07org.openembedded.dev * ra17689c0c4 10openembedded.git/classes/cmake.bbclass: cmake.bbclass: Allow out of tree builds with cmake. Mar 20 13:56:49 03Robert Schuster  07org.openembedded.dev * rb71dbb35d8 10openembedded.git/recipes/swt/ (swt3.4-gtk-hildon_3.4.2.bb swt3.4-gtk_3.4.2.bb): Mar 20 13:56:51 swt3.4-gtk 3.4.2: New recipe. Mar 20 13:56:53 swt3.4-gtk-hildon 3.4.2: New recipe. Mar 20 13:56:55 03Robert Schuster  07org.openembedded.dev * r968aa56f6f 10openembedded.git/recipes/llvm/ (llvm-2.5/fix-build.patch llvm_2.5.bb): llvm 2.5: New recipe. Mar 20 13:58:19 markos_: you can Mar 20 13:58:31 markos_: you can even provide us patch which will add new kernel Mar 20 13:59:37 sure, give me a pointer on where to start and i'll gladly do that Mar 20 14:00:26 moment Mar 20 14:01:09 markos_: simply add linux-rt_2.6.27.bb (and patches) - your build will use it Mar 20 14:01:29 ok Mar 20 14:01:47 in the machine config right? Mar 20 14:03:11 or my recipe? Mar 20 14:03:16 no need to change machine config Mar 20 14:03:21 add new recipe Mar 20 14:04:16 so just adding linux-rt_2.6.27.bb (plus patches) to IMAGE_INSTALL, will do? Mar 20 14:06:03 no Mar 20 14:06:07 to metadata Mar 20 14:06:12 OE will use newer ver Mar 20 14:06:24 ok Mar 20 14:11:46 <_cpo_> I build the xserver-xorg package from latest dev. the problem is when running the xserver i cant use my mouse/keyboard Mar 20 14:11:57 <_cpo_> is there any proto missing? Mar 20 14:12:44 03Stanislav Brabec  07org.openembedded.dev * r899f35e8f7 10openembedded.git/recipes/ (rpm/rpm_4.4.2.3.bb yum/files/paths2.patch): rpm, yum: Reverted back python "site-recipes" to "site-packages". Mar 20 14:12:45 03Stanislav Brabec  07org.openembedded.dev * ra2f5dab7c8 10openembedded.git/recipes/rpm/ (files/rpmconfigdir.patch rpm_4.4.2.3.bb): rpm: Fixed build if libdir != prefix/lib. Mar 20 14:12:45 03Stanislav Brabec  07org.openembedded.dev * raad4de5b0e 10openembedded.git/recipes/rpm/rpm-native_4.4.2.3.bb: rpm-native: Requires gettext-native. Mar 20 14:13:06 i think you have to set XSERVER += "xf86-driver-input-mouse xf86-driver-input-keyboard" for this Mar 20 14:13:31 <_cpo_> well i installed xf86-input-mouse/keyboard Mar 20 14:13:34 <_cpo_> makes no difference Mar 20 14:13:45 <_cpo_> strange thing is, when changing to xserver 1.4 it works again Mar 20 14:13:55 <_cpo_> and yes, i recompiled the packages for the new serverversion Mar 20 14:18:55 hi, I'd like to move my deploy dir outside of tmp by setting DEPLOY_DIR in my local.conf but it simply ignores it and always use tmp/deploy ... Mar 20 14:19:13 whereas setting DEPLOY_DIR_PSTAGE works... Mar 20 14:19:16 mpoullet|work: check where it is set Mar 20 14:21:03 mpoullet|work: you use angstrom? Mar 20 14:21:43 hrw: yes Mar 20 14:22:04 edit oe/conf/distro/include/angstrom.inc and change DEPLOY_DIR = to ?= Mar 20 14:22:57 hrw: oh yes, thanks! Mar 20 14:42:08 03Stanislav Brabec  07org.openembedded.dev * r2fbac192f1 10openembedded.git/recipes/beecrypt/ (beecrypt_3.1.0.bb files/x64fix.patch): beecrypt: Fixed x64fix.patch to follow standard libdir. Mar 20 14:43:08 03Koen Kooi  07org.openembedded.dev * rd094b05396 10openembedded.git/ (conf/checksums.ini recipes/sqlite/sqlite3_3.6.5.bb): sqlite3: update to 3.6.5 Mar 20 14:43:10 03Koen Kooi  07org.openembedded.dev * rb3f902c514 10openembedded.git/recipes/mozilla/ (fennec/jsautocfg-dontoverwrite.patch fennec_hg.bb): fennec: update from alpha to beta release, work around latest mozilla bugs (filed upstream) Mar 20 14:43:11 03Koen Kooi  07org.openembedded.dev * r5cc2f40929 10openembedded.git/conf/distro/include/angstrom-2008-preferred-versions.inc: angstrom 2009.X: bump gtk+ and glib-2.0, runtime tests don't show obvious regressions Mar 20 14:43:12 03Koen Kooi  07org.openembedded.dev * rba0746e14d 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Mar 20 15:08:07 how can I checkout all OE trees from main repo ? Mar 20 15:08:43 or how can I know the list of available branches in OE main repo? Mar 20 15:09:07 if I do git clone git://git.openembedded.net/openembedded I get only one branch Mar 20 15:09:23 http://cgit.openembedded.net/?r=openembedded Mar 20 15:09:42 I see a list of all branches with 'git branch' Mar 20 15:10:02 but I'm using git 1.4.4.4 on the server where i'm compiling Mar 20 15:10:15 ClashTheBunny: thx ;-) Mar 20 15:10:20 tsjsieb: $ git branch Mar 20 15:10:20 * org.openembedded.dev Mar 20 15:10:47 I think it has changed in the newer git version Mar 20 15:11:08 mckoan: 'git branch -r' Mar 20 15:11:44 hrw: correct thx Mar 20 15:12:17 how can I update OE user manual ? Mar 20 15:12:30 mckoan: submit patch to OE ML Mar 20 15:13:24 hey hrw Mar 20 15:13:35 can you please send me the package list from your rpm-working ubuntu setup? Mar 20 15:14:27 Tartarus: ok Mar 20 15:15:38 thanks Mar 20 15:16:12 Tartarus: it will take some time anyway - I have to reboot cause vbox refuse to start ;( Mar 20 15:43:31 hi, all! where could I buy beagle in Germany? Mar 20 15:44:10 digikey ships to Germany? Mar 20 15:45:55 denix, well, when I visit Germany? Mar 20 15:47:50 denix: they ship to Poland so should also to Germany ;D Mar 20 15:49:18 but they won't to Russia Mar 20 15:49:59 slapin: Iron Courtain alive Mar 20 15:50:50 yeah Mar 20 15:51:12 so that's why I'm interested if there's somewhere I could pick it up Mar 20 15:51:33 digikey says something about export restrictions Mar 20 15:52:16 but they allow for companies Mar 20 15:52:25 which makes me wonder Mar 20 15:52:49 slapin: try to buy EVBeagle instead Mar 20 15:52:50 that's what I know from my local TI representative Mar 20 15:53:47 hrw, what's it? Mar 20 15:54:27 German clone of beagleboard Mar 20 15:55:37 hrw, any links? my google-foo failed me Mar 20 15:55:44 s/foo/fu Mar 20 15:56:35 link is on elinux.org/beagleboard Mar 20 16:09:32 hrw, thanks a lot! Mar 20 16:11:54 03Jan Lübbe  07fso/milestone5.5 * rdce5046627 10openembedded.git/ (conf/checksums.ini packages/tinymail/libtinymail_0.0.9.bb): libtinymail: add 0.0.9 Mar 20 16:48:39 how do I fetch origin/hrw/poky-rpm ? Mar 20 16:53:13 slapin: DigiKey ? Mar 20 17:17:47 anyone here that uses ppc603e (like lite5200, efika, n1200, storcenter) or any of the other 'powerpc' architectures with oe? Mar 20 17:26:19 I am seeing a problem fetching cvs repository for bintutils, though it says module binutils it ends of checking out the whole src tree Mar 20 17:26:56 SRC_URI = "cvs://anoncvs:anoncvs@sources.redhat.com/cvs/src;module=binutils;method=pserver;localdir=src Mar 20 17:27:04 this is the src_uri Mar 20 17:27:38 if I do not specify localdir then it does try to checkout the module but then fails because there is a binutils directory inside src too Mar 20 17:27:48 I am suspecting thats confusing the fetcher Mar 20 17:28:22 All I want is cvs -d :pserver:anoncvs@sources.redhat.com:/cvs/src co binutils Mar 20 17:28:55 can someone beat fetcher to do that Mar 20 17:31:06 ugh, i hate that cvs, it checks out differently than you ask it to Mar 20 17:31:08 very annoying Mar 20 17:31:19 * kergoth remembers trying to add some _cvs recipes and being quite annoyed Mar 20 17:31:53 yeah I work on bleeding edge so sometimes I need hot from oven Mar 20 17:32:00 :) Mar 20 17:36:41 kergoth: why is fetcher looking for directory with same name as module Mar 20 17:36:52 * kergoth shrugs, don't remember Mar 20 17:36:57 I tried briefly to look at cvs.py Mar 20 17:37:26 the whole fetch module should be scrapped imo Mar 20 17:37:28 for immediate need I decided to fix the recipe Mar 20 17:37:30 hi all Mar 20 17:37:30 it's terrible Mar 20 17:37:37 hmm Mar 20 17:38:28 api sucks (can't request a download to anywhere but DL_DIR easily, for example), structure sucks... and why aren't we using urllib2 or python-curl to fetch http/ftp instead of wget? :) Mar 20 17:38:47 yeah curl is much better Mar 20 17:49:31 i tried adding a urllib2 fetcher, but it turns out it used syscalls which aren't sane when run after fork but before exec, or something, so the whole deal would need reworking to use it Mar 20 17:53:23 hi, how is the new fennec supposed to be cross-compiled...it still has the same problem than the old one: | /bin/sh: autoconf2.13: command not found Mar 20 17:53:32 and I do not see autoconf 2.13 in oe Mar 20 17:56:23 Gnutoo: by any chance can you get a autoconf upgraded version or ask the fennec people for it Mar 20 17:57:04 khem, ok i'll look at the bug report again to see if there are progress... Mar 20 17:57:26 * Gnutoo must find this bugreport again Mar 20 18:00:41 generally we patch anything we add to work with current autotools Mar 20 18:00:46 not worth trying to maintain multiple versions Mar 20 18:01:36 khem, thanks for the response...by the way I bet that having 2 versions of autotools components in staging isn't great because of the globing(*) in recipes Mar 20 18:01:54 s/components/components at the same time/ Mar 20 18:38:12 Gnutoo: though 2.13 can co exist with newer autoconf we should avoid it if possible Mar 20 18:38:54 Gnutoo: other option could be that you do not regenerate config files at all for your package in that case it wont require autconf Mar 20 18:39:16 oh well you still need to run configure Mar 20 18:40:13 03Marcin Juszkiewicz  07org.openembedded.dev * r0c549f852e 10openembedded.git/classes/package_rpm.bbclass: Mar 20 18:40:13 package_rpm.bbclass: files needs to start from / Mar 20 18:40:13 git-svn-id: https://svn.o-hand.com/repos/poky/trunk@5116 311d38ba-8fff-0310-9ca6-ca027cbcb966 Mar 20 18:54:35 03Khem Raj  07org.openembedded.dev * rc3c3af2895 10openembedded.git/recipes/binutils/ (7 files in 2 dirs): (log message trimmed) Mar 20 18:54:35 binutils_cvs: Various fixes to get it building. Mar 20 18:54:35 * Port patched needed from binutils 2.19. Mar 20 18:54:35 * Refresh the old patches. Mar 20 18:54:35 * Cover for bitbake fetcher lameness where it Mar 20 18:54:37 checks out the while src tree when we only Mar 20 18:54:39 asked for binutils module. So we have to build Mar 20 18:54:43 03Khem Raj  07org.openembedded.dev * ra7b83866a1 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Mar 20 18:58:05 x86_64-linux-libtool: link: `lzo1c_9x.lo' is not a valid libtool object Mar 20 18:58:09 any thoughts on this? Mar 20 18:58:49 khem, ok I thought about it too...I hope that we don't need to patch the configure.[ac|in] Mar 20 18:59:03 khem, thanks a lot Mar 20 18:59:33 urg Mar 20 18:59:49 I am in a git branch and I need to pull dev updates trough Mar 20 19:00:05 and i forgot the magic git command Mar 20 19:01:37 got it Mar 20 19:01:40 in phrasebook Mar 20 19:02:16 Crofton|work: you need to pull Mar 20 19:02:40 heh Mar 20 19:02:45 yeah, just did Mar 20 19:03:06 I don't remember all the git foo :) Mar 20 19:07:08 Crofton|work: you can set a branch by default to merge from origin/org.openembedded.dev Mar 20 19:07:59 so you can set the default for git pull based on what branch you are in? Mar 20 19:08:08 Crofton|work: yes Mar 20 19:08:12 XorA: could tou share how, so that I can forget again? :-) Mar 20 19:08:16 tou=you Mar 20 19:08:24 gimme a sec to boot git machine :-) Mar 20 19:09:30 I need to finish my part of a conference abstract Mar 20 19:09:36 * Crofton hates writing Mar 20 19:09:41 [branch "xora/angstrom-sqeeezy"] Mar 20 19:09:41 remote = origin Mar 20 19:09:41 merge = refs/heads/org.openembedded.dev Mar 20 19:09:59 I assume there is some magic you can set to fix those from git-config Mar 20 19:10:22 XorA: its from .gitconfig Mar 20 19:10:43 khem: .git/config yes Mar 20 19:11:00 I normally ending using vi :-) Mar 20 19:18:37 I'll generate the configure using autotools from my laptop's distro...because there is no ./configure inside the mozilla-central directory... Mar 20 19:18:56 the makefile makes you generate it Mar 20 19:26:47 bad idea...I think i'll try autoconf-2.13 Mar 20 19:27:23 * XorA is scared to do a git-pull --rebase Mar 20 19:30:19 XorA, why? Mar 20 19:30:41 Crofton|work: mass rename Mar 20 19:30:46 ah Mar 20 19:30:54 packages-> recipes? Mar 20 19:31:31 XorA fears "all your rebase are belong to us" name confusion. Mar 20 19:32:04 I wonder how many of the youngsters here will get that reference :) Mar 20 19:32:24 * Crofton|work misses AYB Mar 20 19:32:32 wait - thats a old person reference now? Mar 20 19:33:47 hmm, only 7-8 years old Mar 20 19:34:12 re Mar 20 19:35:17 Ancient reference to to late 20th century (Zero Wing on Mega Drive, 1991). Mar 20 19:36:33 http://en.wikipedia.org/wiki/Zero_Wing#.22All_your_base_are_belong_to_us.22 Mar 20 19:38:45 I rarely pull Mar 20 19:38:50 usually git remote update Mar 20 19:38:51 <_cpo_> gggghmmm Mar 20 19:38:57 <_cpo_> the minimal distribution works quiet fine Mar 20 19:43:01 CosmicPenguin - a lot of the peeps here think Gen X is ooooold. Mar 20 19:45:13 presumably the same people who use the word peeps Mar 20 19:45:58 XorA: a pull --rebase is safe, git/rebase is smart enough to handle the rename, at least it did here for all my commits Mar 20 19:53:15 * kergoth wonders why the gnu_hash variable in insane.bbclass isn't used anywhere Mar 20 20:05:06 CosmicPenguin - I'm probably older than 95% of the people in here. Mar 20 20:08:16 You know what they say, you are only as old as your vocabulary Mar 20 20:16:03 CosmicPenguin - what would Maynard G. Krebs say about that? Mar 20 20:20:48 Thats a bit old for me, but I am familiar with Shaggy from Scooby Doo if that helps Mar 20 20:25:44 The Scobby Doo characters were modeled on the characters from Doby Gillis. Mar 20 20:26:19 err, "Scooby" Mar 20 21:10:03 Hi ! I downloaded manually a file in the oe downloads dir, but bitbake wan't to fetch it anyway. How to fix this ? Mar 20 21:10:40 try to create a filename.md5 Mar 20 21:12:40 gremlin[it]: It's the first thing I tried. In fact you need the .lock file too ... Mar 20 21:12:42 than Mar 20 21:12:44 ks Mar 20 21:13:31 you welcome Mar 20 21:13:49 good morning Mar 20 21:14:42 I'm trying to put my DEPLOY_DIR outside of TMP_DIR, but the kernel.bbclass seems to have a problem with that Mar 20 21:15:00 03Angus Ainslie  07fso/milestone5.5 * rd8de3851dc 10openembedded.git/packages/efl1/ecore-native_svn.bb: ecore-x needs to be enabled for edje-sdk to compile Mar 20 21:15:02 it stops at do_deploy Mar 20 21:15:11 03Angus Ainslie  07fso/milestone5.5 * rf9649d0279 10openembedded.git/conf/distro/include/sane-srcrevs.inc: Catch up srcrevs for fso-abyss, libgsm0710 and libgsm0710mux Mar 20 21:15:13 my local.conf is here: http://pastebin.com/d748446b1 Mar 20 21:16:54 the problem happens when do_deploy called package_stagefile_shell Mar 20 21:17:08 because src and dest are the same Mar 20 21:18:21 I mean this call: package_stagefile_shell ${DEPLOY_DIR_IMAGE}/${KERNEL_IMAGE_BASE_NAME}.bin Mar 20 21:20:13 oh package_stagefile_shell() does only copy something if the packaged-staging is used which is my case Mar 20 21:24:01 so package_stagefile_shell() wants to replace ${TMPDIR} with {PSTAGE_TMPDIR_STAGE} Mar 20 21:24:33 but it gets ${DEPLOY_DIR}/images/beagleboard as input Mar 20 21:25:30 and in this case there's nothing to replace, and boom! cp src dest with src=dest... Mar 20 21:36:24 I've solved it with a quick hack in base.bbclass but I'd like to understand the problem better Mar 20 21:38:12 pretty sure i reported that in the tmpdir restructuring email thread :) Mar 20 21:38:17 not sure what the best soultion is though Mar 20 21:45:20 mpoullet - what's your "quick hack" look like? Mar 20 21:48:02 simply looks if src == dest before it copies Mar 20 21:48:13 that's probably wrong Mar 20 21:48:31 but I'd like to have it through the build though... Mar 20 21:51:33 my first build for packaged-staging is done now Mar 20 21:51:52 this time I've secured the pstage AND the deploy dirs Mar 20 21:52:17 let's do remove tmp and start the second build Mar 20 21:56:18 hggfggdcdczx Mar 20 21:56:51 sorry , kids :D Mar 20 21:56:57 and don't forget to remove the staging-opkg-native package first... Mar 20 21:58:54 * kergoth can't seem to reproduce the opkg-cl error anymore for some reason Mar 20 22:02:25 kergoth: it happens to me again today, both at work and at home Mar 20 22:02:57 kergoth: but deleting it and letting it be rebuilt works Mar 20 22:05:27 sh***t Mar 20 22:05:39 second build fails at the kernel Mar 20 22:06:00 because it wants to compile it again Mar 20 22:06:19 and doesn't find the cross compile Mar 20 22:11:37 :-( I think I need to secure the cross dir in tmp too Mar 20 22:11:51 * mpoullet|home OE drives me crazy Mar 20 22:16:56 anyone tried a gcc-4.3.2 toolchain with libstdc++? I get libstdc++.so: undefined reference to `__gcc_qmul@GCC_4.1.0' when linking - not sure where the 4.1.0 reference is coming from Mar 20 22:17:10 well that's strange, during the first build, I get a dir called cross/armv7a/bin/ in the tmp dir which contains arm-angstrom-linux-gnueabi-gcc Mar 20 22:17:41 but during the second build, I get this structure: cross/arm-angstrom-linux-gnueabi/ Mar 20 22:18:02 and no cross compiler there... Mar 20 22:29:32 running into the same libstdc++.so: undefined reference to `__gcc_qmul@GCC_4.1.0' for gcc-4.3.3 Mar 20 22:49:22 jo Mar 20 22:54:05 second build done, but I needed to put the tmp/cross dir from the first build back so that it can compile the kernel, don't know why Mar 20 23:20:48 gb Mar 21 00:49:40 hmm Mar 21 00:49:53 can't push. wonder if mickey ever got my current public key added Mar 21 00:49:55 * kergoth scratches head Mar 21 00:50:00 i didd Mar 21 00:50:14 weird Mar 21 00:50:32 paste it in a query and let me double check Mar 21 00:51:22 can connect to other ssh servers with this same key in their authorized keys Mar 21 00:51:23 k Mar 21 00:53:58 that's correct Mar 21 00:54:36 it's git@, right? Mar 21 00:56:10 ya, git@git.openembedded.org:openembedded Mar 21 01:01:13 i don't get it, i can ssh to work machines with that pubkey in their authorized keys just fine Mar 21 01:01:43 from the same machine, same terminal session Mar 21 01:02:34 * kergoth 'll have to mess with it another time, heads out to go pick up the gf from work Mar 21 01:09:23 have fun Mar 21 01:09:27 * mickeyl going to sleep **** ENDING LOGGING AT Sat Mar 21 02:59:57 2009