**** BEGIN LOGGING AT Wed Oct 15 02:59:57 2008 Oct 15 07:17:35 03Koen Kooi  07org.openembedded.dev * rb370b6d9ac 10org.openembedded.dev.git/packages/mozilla/fennec_hg.bb: fennec: add super-secret option to speed up rendering in the near future Oct 15 07:23:11 03Dirk Opfer  07org.openembedded.dev * r41520d2ae8 10org.openembedded.dev.git/packages/fuse/curlftpfs_0.9.1.bb: fuse/curlftpfs: Update to 0.9.2 Oct 15 07:34:06 03Dirk Opfer  07org.openembedded.dev * r41520d2ae8 10org.openembedded.dev.git/packages/fuse/curlftpfs_0.9.1.bb: fuse/curlftpfs: Update to 0.9.2 Oct 15 07:52:08 G'Morning Oct 15 08:08:36 bonjour Oct 15 08:09:02 03Koen Kooi  07org.openembedded.dev * rbd558780fa 10org.openembedded.dev.git/packages/mozilla/ (7 files in 3 dirs): firefox 3.1: move from alpha2 to beta1 Oct 15 08:37:55 skrew: good to hear that it worked out. Oct 15 09:15:41 hi lrg Oct 15 09:15:54 hey pb_ Oct 15 09:17:13 pb_: do you know who's looking after the OE git server ? It appears to be down Oct 15 09:20:35 morning Oct 15 09:22:13 lrg: that would be überhacker zecke Oct 15 09:22:23 do you mean the old git server, or the new one? Oct 15 09:22:40 pb_: oh, I'm probably using the old one. Oct 15 09:22:43 * lrg checks Oct 15 09:23:37 pb_: git.openembedded.net Oct 15 09:25:12 lrg: Updated from git://git.openembedded.net/openembedded now just fine Oct 15 09:25:46 pb_, JaMa : looks like out of date. Thanks Oct 15 09:35:23 good morning Oct 15 10:05:31 03Koen Kooi  07org.openembedded.dev * r52161525d0 10org.openembedded.dev.git/packages/sysstat/sysstat_8.1.6.bb: sysstat: add 8.1.6 Oct 15 10:06:10 03Koen Kooi  07org.openembedded.dev * r7d36a7b181 10org.openembedded.dev.git/packages/xorg-driver/xf86-video-omapfb_git.bb: xf86 video omapfb: bump SRCREV Oct 15 10:14:11 mickeyl: good morning Oct 15 10:23:07 hi pb__ Oct 15 10:49:48 03Jan Lübbe  07org.openembedded.dev * r242ff4c599 10org.openembedded.dev.git/contrib/mtn2git/manglenames.py: mtn2git: commit the final version used for the bk&mtn to git conversion Oct 15 10:50:12 just in case anyone is interested, I just found a makefile debugger that proved to be very useful (was trying it from devshell to figure out why some stuff with a totally screwed build system did not work out): http://bashdb.sourceforge.net/remake/ Oct 15 10:51:45 especially nice thing - the possibility to examine the variables used in the makefile Oct 15 11:14:51 Anyone that has some good ideas for a setup that is REALLY fast at building OE? Oct 15 11:15:21 Should I go for a single really powerfull multi core machine, with tons of RAM. Oct 15 11:16:00 Esben: the quadcore with a software raid 0 that I have here performs very well, I believe you could go faster with a hardware raid, got only 4GB RAM though, build time for toolchain, libc, bootable image is around 40 minutes Oct 15 11:16:01 Or perhaps a blade server or a handfull of more standard configured machines? Oct 15 11:16:07 maybe 30, depends.. Oct 15 11:16:21 Cool :-) Oct 15 11:16:24 what distro? Oct 15 11:16:32 an older fedora 7 Oct 15 11:16:49 I mean, is it 30-40 mins for a relatively fat angstrom or poky we are talking about? Oct 15 11:16:58 oh Oct 15 11:17:10 what distro I am building.. I am based on angstrom Oct 15 11:17:22 but use a custom image, so without X11 Oct 15 11:17:37 and I set PARALLEL_MAKE = "-j 8" BB_NUMBER_THREADS = "4" Oct 15 11:17:55 but with directfb and some other stuff, so its more than just minimal Oct 15 11:18:32 but its the time for a clean build (so from scratch, toolchain, libc, kernel, etc.) Oct 15 11:19:53 Esben: fast machine?? Oct 15 11:20:04 Esben: multicore with fast i/o and ram Oct 15 11:20:42 best would be 16-32GB ram to make builds in tmpfs :D Oct 15 11:20:46 hi Cliff Oct 15 11:21:37 hrw: yes, that sounds like a really good idea :-D Oct 15 11:22:09 is DEPENDS also valid for the do_unpack stage? I depend on something -native that I need in do_unpack but the dependency is not being built in time Oct 15 11:22:40 Esben: on my quadcore (intel q6600) with 4gb ram and one sata hdd my nightly buildbot build was going for 8h. it was poky-image-minimal/sato/sdk for qemuarm/c7x0/qemux86 Oct 15 11:23:51 * JaMa has too, quadcore phenom 9600 with 4gb ram, and sw raid0 on 3 sata2 disks and build still doesnt seem fast enough for me :) Oct 15 11:24:12 there is no such thing as enoug cpu/ram/hdd speed Oct 15 11:25:32 yes exactly :) and also REALLY fast is then only fairly fast Oct 15 11:28:56 Jin^eLD: you can set deps for tasks too Oct 15 11:29:06 hrw: how? Oct 15 11:29:56 DEPENDS_do_unpack? or something like that? Oct 15 11:30:08 let me look Oct 15 11:33:20 Jin^eLD: do_unpack[depepnds] += "your-packer-native:do_populate_staging" Oct 15 11:33:25 Jin^eLD: do_unpack[depends] += "your-packer-native:do_populate_staging" Oct 15 11:34:25 Jin^eLD: but this is not wide used in recipes - rather in classes Oct 15 11:34:37 I see.. thanks a lot Oct 15 11:45:43 did the pull instrauctions for git change? Oct 15 11:46:11 git clone git://git.openembedded.net/org.openembedded.dev Oct 15 11:46:36 maybe this changes to git clone git://git.openembedded.net/openembedded Oct 15 11:47:09 stupid net splits Oct 15 11:48:42 hrw, ping Oct 15 11:49:24 pong Oct 15 11:49:34 hrw, has the git url changed? Oct 15 11:49:59 Crofton|work: I used line which Zecke wrote in his mail and it cloned fine Oct 15 11:50:10 I am be looking at getting started Oct 15 11:50:17 Hmm.... I think I will need something like min. 32gb ram, at least 4 x 15k sata hdds running raid0, and 2 x quadcore. Oct 15 11:50:33 let me check, I think I need to update getting started Oct 15 11:50:42 I wonder if there is any worthwhile performance difference between intel and amd cpus these days. Oct 15 11:51:12 he says for commit access, but I guess he means everyone :) Oct 15 11:53:00 hi, I'm trying to add a recipe for qwt scientific library which uses qt4, i am inheriting qmake2 but qt4 library paths are not correct while building. Oct 15 11:55:15 hrw: hello Oct 15 12:05:07 Please check my update to http://wiki.openembedded.net/index.php/Getting_Started#Obtaining_OpenEmbedded_using_GIT Oct 15 12:15:42 Jin^eLD: thanks for the remake link -- I need something like that right now ... Oct 15 12:17:00 Crofton|work: does git.openembedded.org (vs net) work for you? Oct 15 12:17:23 I haven't tried yet Oct 15 12:17:34 It looks like .net is where git is ... Oct 15 12:17:49 CoreDump: nod Oct 15 12:17:53 I'm witing until 8:30 when npr via internet turns off :) Oct 15 12:18:22 urg, I wrote .org Oct 15 12:18:34 cbrake, can you confirm .org does not work? Oct 15 12:18:48 Crofton|work: ping: unknown host git.openembedded.org Oct 15 12:18:59 fixing Oct 15 12:19:35 it's a pity that the same people don't control openembedded.org and openembedded.net domains ... Oct 15 12:19:39 getting started fixed Oct 15 12:21:26 again Oct 15 12:35:20 03Robert Schuster  07org.openembedded.dev * r98cc700911 10org.openembedded.dev.git/packages/classpath/ (classpath-0.97.2/drawpolyline.patch classpath_0.97.2.bb): classpath 0.97.2: Added patch to fix to fix drawPolyline bug. Oct 15 12:48:00 git@git.openembedded.net's password: Oct 15 12:48:05 um, what am I doing wrong Oct 15 12:48:32 I suspect I am missing things in key setup Oct 15 13:09:05 03Marcin Juszkiewicz  07org.openembedded.dev * r14ae1f974c 10org.openembedded.dev.git/MAINTAINERS: MAINTAINERS: I am no longer interested in Tosa Oct 15 13:09:37 03Marcin Juszkiewicz  07org.openembedded.dev * r9b90ffaa4d 10org.openembedded.dev.git/MAINTAINERS: MAINTAINERS: I am no longer interested in OpenZaurus Oct 15 13:09:37 03Marcin Juszkiewicz  07org.openembedded.dev * r40fdfee5f6 10org.openembedded.dev.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Oct 15 13:10:41 ok, I am able to pull/push Oct 15 13:11:43 Sounds like I shoul learn how to use git now ;) Oct 15 13:11:48 ;D Oct 15 13:12:01 I am working on that also Oct 15 13:12:16 florian: clone/pull/push/commit/diff/add/mergetool are most needed Oct 15 13:12:18 I think I need to figure out .ssh.config to manage my identiies Oct 15 13:13:07 363MB repo.. Oct 15 13:13:15 hrw: yes it should be quite easy once you decided for a particular workflow Oct 15 13:13:51 is there a maintainer of the tosa arch? I'm curious how actively it's being tested/maintained... Oct 15 13:14:14 (having to make a few small changes to compile gpe-image - mainly tosa-specific kernel stuff - but of course it'll take 4 more days for the compile to finish) Oct 15 13:14:26 florian: I had one link about it Oct 15 13:14:34 hallyn: tosa machine? Oct 15 13:14:47 I have no idea now Oct 15 13:15:01 florian: http://ktown.kde.org/%7Ezrusin/git/git-cheat-sheet.svg is handy Oct 15 13:15:20 http://zrusin.blogspot.com/2007/09/git-cheat-sheet.html in general Oct 15 13:15:53 http://osteele.com/archives/2008/05/my-git-workflow is worth read Oct 15 13:16:36 * broonie wonders if it's OK to shut down the Wolfson mtn mirror Oct 15 13:17:17 hrw: yeah, the zaurus sl-6000l Oct 15 13:17:49 hallyn: I was interested in it over year ago. recently gave it to MFD maintainer Oct 15 13:17:52 (the latest images don't quite work - gpe-image doesn't have any kernel modules, opie-image has some problems) Oct 15 13:18:06 hrw: what does 'gave it' mean? Oct 15 13:18:21 the MFD maintainer (whatever mfd is) owns hte tosa arch? Oct 15 13:18:58 florian: http://www.andrewmoore.com/public/index.php/My_git_workflow is nicer Oct 15 13:19:24 broonie: I'm considering the same for monotone.nslu2-linux.org Oct 15 13:19:43 (since I've released the updated nslu2-linux makefile) Oct 15 13:19:44 hallyn: MFD is one of subsystems in linux kernel and Tosa contains chip which has a MFD driver not yet fully merged into mainline. Oct 15 13:19:56 I would say: shut them down Oct 15 13:19:59 but thats me Oct 15 13:20:23 hallyn: so I gave him Tosa + serial cable so he is able to test that code on device and fix any problems. Oct 15 13:20:37 hrw: Looks pretty good indeed... thanks! Oct 15 13:20:38 then Tosa will go to one of Tosa kernel hackers Oct 15 13:20:44 hi bin1010 Oct 15 13:21:49 florian: today/tomorrow it will land on ltg planet Oct 15 13:25:47 hrw: very good Oct 15 13:26:21 delicious in feed is good thing Oct 15 13:26:21 hrw: ok,thanks Oct 15 13:26:27 * hallyn goes to look up the mfd maintainer Oct 15 13:27:19 ok, how do I clone git for the case where I need to push ..... Oct 15 13:27:35 hallyn: he does not use OE Oct 15 13:27:48 sigh Oct 15 13:27:57 git clone crofton@git.openembedded.net:openembedded Oct 15 13:28:07 is getting me a password prompt Oct 15 13:28:15 s/crofton/git Oct 15 13:28:43 Crofton|work: user is git@ - then gitosis check your ssh key to know how to handle you Oct 15 13:29:06 ah, so I need gitosis at this end? Oct 15 13:29:10 no Oct 15 13:29:20 gitosis handles server side Oct 15 13:29:25 hrw: so you're talkinga bout Samuel Ortiz? Oct 15 13:29:26 the git user gets a pw request Oct 15 13:29:28 Crofton, there is one unix user for all of us Oct 15 13:29:31 hallyn: yes Oct 15 13:29:52 Crofton|work: tried password from you ssh key? Oct 15 13:30:07 hrw: ok, thanks. i'll follow up with him Oct 15 13:30:37 hallyn: you rather should contact Dmitriy B. Oct 15 13:30:51 as he is current tosa maintainer (or closest to that title) Oct 15 13:30:54 hrw yes Oct 15 13:31:03 how does git know which key to use? Oct 15 13:31:07 ~seen jlaine Oct 15 13:31:09 i haven't seen 'jlaine', Genesis Oct 15 13:32:35 03Jan Luebbe  07org.openembedded.dev * r869ba9e1fb 10org.openembedded.dev.git/ (4403 files in 4403 dirs): mtn2git: remove .mtn2git_empty files in non-empty directories Oct 15 13:32:40 hrw: he's not in MAINTAINERS in either org.op.. or linux-2.6? Oct 15 13:32:47 Crofton, ssh tells gitosis which key was used via the environment Oct 15 13:33:52 I'm having trobule configuring at my end Oct 15 13:34:04 I made a .ssh/config file Oct 15 13:34:11 Crofton|work: BTW, you can clone anonymously and then push with git@... and it works Oct 15 13:34:19 hallyn: Oct 15 13:34:30 rwhitby, I suspect I will have the same troble Oct 15 13:34:32 hallyn: Dmitry Baryshkov is proper name Oct 15 13:34:56 Crofton|work: yep, was responding to earlier remark, not latest ssh problem Oct 15 13:35:10 anon clone is fine Oct 15 13:36:58 hrw: ah, thanks, got it Oct 15 13:37:26 hallyn: his email is in kernel sources, on lkml, lakml... Oct 15 13:38:00 hrw: yup he's in the kernel maintainers file. thanks. Oct 15 13:39:43 got it, typed in wrong path name for my key file in .ssh/config Oct 15 13:55:24 03Koen Kooi  07org.openembedded.dev * r77c7b94c93 10org.openembedded.dev.git/packages/mozilla/firefox.inc: firefox.inc: add DESCRIPTION Oct 15 14:29:41 03Robert Schuster  07org.openembedded.dev * r4e870c5ec5 10org.openembedded.dev.git/packages/linux/ (linux-2.6.24/atngw100/defconfig linux_2.6.24.bb): Oct 15 14:29:41 linux 2.6.24: Updated defconfig for atngw100 Oct 15 14:29:41 - compile ext2 and ext3 into the kernel (to allow booting from SD) Oct 15 14:37:28 hi, Oct 15 14:37:28 "git config --get user.mail" = thebohemian@openembedded.org Oct 15 14:37:28 still the summaries say "Robert Schuster " Oct 15 14:37:28 any idea what went wrong? Oct 15 14:37:54 thebohemian: user.email Oct 15 14:38:01 and user.name Oct 15 14:38:34 moin Oct 15 14:39:01 time to merge some OM stuff Oct 15 14:39:25 thebohemian: are you going to be at the cbase tonight? Oct 15 14:39:30 rwhitby: ahh, ok Oct 15 14:39:40 zecke: yes, free beer :) Oct 15 14:39:52 did someone say free beer .... ? Oct 15 14:40:01 elektra wants to celebreate madwifi doing ad-hoc correctly Oct 15 14:40:12 thebohemian: ah okay, didn't know that Oct 15 14:40:20 * zecke sees a decrease of OE quality tonight Oct 15 14:40:36 zecke: she sent it through freifunk-berlin@ ML Oct 15 14:42:16 zecke: typo? http://wiki.openembedded.net/index.php?title=GitPhraseBook&diff=700&oldid=699 Oct 15 14:42:55 thebohemian: typo? me? Oct 15 14:43:04 * thebohemian hopes to see plenty of merging between OM and OE :) Oct 15 14:43:18 thebohemian: you still maintain chinook-compat? Oct 15 14:43:32 thebohemian: you might want to use --global on git-config as well Oct 15 14:43:44 thebohemian: I will start with the checksums Oct 15 14:43:46 zecke: I believe this is a bug resulting from building OM with OE's .dev: https://wiki.evolvis.org/jalimo/index.php/Packages#unsatisfied_libgcc-s1_dependency Oct 15 14:43:51 hrw: yes, i do Oct 15 14:44:19 rwhitby: I removed the nslu2 branches, I hope you have a copy of them Oct 15 14:44:31 thebohemian: I have set of maemo recipes to merge from Poky - diablo updates, total cleanup, revised deps etc Oct 15 14:44:31 zecke: they've already moved to git elsewhere Oct 15 14:44:37 16:44:45 up 5 days, 2:05, 28 users, load average: 19.55, 15.82, 13.44 Oct 15 14:44:40 hmmm Oct 15 14:44:45 hrw: I am interested :) Oct 15 14:44:53 zecke: bitbake was just a monotone version of the berlios tree Oct 15 14:45:11 zecke: org.nslu2-linux.dev has moved to http://repo.or.cz/w/nslu2-linux/master.git?a=tree;hb=HEAD Oct 15 14:45:25 do we have rules for user branches on git.oe.net server? Oct 15 14:45:28 zecke: thx for removing them Oct 15 14:45:42 rwhitby: okay, I don't care for details, I just wanted to not leave anything behind :) Oct 15 14:45:48 for example - can I have hrw/merging-from-poky/maemo branch there? Oct 15 14:46:02 hrw: / Oct 15 14:46:03 which I would be able to drop after merging? Oct 15 14:46:12 how do user branches get deleted once the topic is merged? Oct 15 14:46:31 rwhitby: I have to change the hook, currently deleting is just disabled Oct 15 14:46:57 ah, so / can be deleted by ssh key ? Oct 15 14:47:17 rwhitby: not yet, but this would sound sane Oct 15 14:47:19 hrw: http://wiki.openembedded.net/index.php/GitPhraseBook#Feature_branches looks like it is suggesting to use them to do things like that Oct 15 14:47:20 (and created, BTW) Oct 15 14:47:45 and can be multi-level Oct 15 14:47:54 thebohemian: hehe, look at the source... that is how I do stuff for webkit Oct 15 14:48:07 cool Oct 15 14:48:37 * thebohemian is tempted to throw the OE-part of the Jalimo SVN into such a user branch ... Oct 15 14:48:40 rwhitby: yes, but this can also create issues.. e.g. if you create name/foo and then remove that and then add name/foo/bla you will have some issues Oct 15 14:49:03 thebohemian: do you intend them to merge at one point? Oct 15 14:50:06 hail zecke Oct 15 14:50:42 zecke: yes. Jalimo SVN is just an incubator for packages not yet being cleaned up Oct 15 14:51:06 pb__: have you seen the FILE_PR proposal? Oct 15 14:51:33 zecke: however I also add new stuff to it from time to time. so the use case is different from "lets bring blender to build" Oct 15 14:51:45 (for example) Oct 15 14:52:15 thebohemian: long lived rebase stuff is cool as well Oct 15 14:52:33 zecke: this one? http://article.gmane.org/gmane.comp.handhelds.openembedded/19541 Oct 15 14:53:44 Happy good morning all! Oct 15 14:53:55 03Robert Schuster  07org.openembedded.dev * r77c1d91f46 10org.openembedded.dev.git/packages/gizmod/ (files/3.4-build-fix.patch gizmod_3.4.bb): gizmod 3.4: New recipe. Oct 15 14:55:17 pb__: maybe, but also the one in my git branch :) Oct 15 14:57:54 zecke: ah no, I haven't yet ventured into your git branch Oct 15 14:58:06 I'm not even sure how to do that to be honest. Oct 15 14:58:28 my git 5ki11z are fairly lame. Oct 15 15:01:54 hmm now that oe uses git maybe i'll be more inclined to mess with it Oct 15 15:02:25 any of you guys look into tools various distros use to track what patches they havent submitted upstream? Oct 15 15:02:30 i know ubuntu has the UpstreamReport Oct 15 15:02:40 and there's a debian package called vcheck, but i'm not sure if it's actually used Oct 15 15:02:44 now I can have many zecke branches :) Oct 15 15:02:50 and i think thats more for tracking upstream releases than patches to go upstream Oct 15 15:02:57 lightweight branches for the win Oct 15 15:08:06 fyi a rather well done git intro: http://excess.org/article/2008/07/ogre-git-tutorial/ Oct 15 15:10:11 * kergoth plays around with budgeting software Oct 15 15:15:49 kergoth: linux ones? Oct 15 15:16:26 doesn't seem to be a lot of options for finance software for linux in general, other than gnucash or moneydance. nowadays i'm actually using osx on a macbook most of the time though Oct 15 15:16:44 using a windows budgeting app coupled with an iwork spreadsheet to assist, at the moment Oct 15 15:17:00 * kergoth gets juice & caffeine Oct 15 15:23:47 * hrw just looked at 35KB long recipe which generate package from rootfs contents... Oct 15 15:24:19 my nightmare recipe Oct 15 15:25:13 yikes Oct 15 15:25:14 one day I will mangle some names in it and show Oct 15 15:25:37 16mb flash suxx when python has to be in it Oct 15 15:29:40 * mwester read kergoth's comments re: finance software and Linux, and agrees, and wonders why since Linux folks need food and shelter too... Oct 15 15:32:11 bye Oct 15 15:42:25 mwester: it is surprising, i guess no one with the expertise has gotten annoyed enough with the current offerings to decide to start a new project yet Oct 15 15:44:54 I suspect people often just use spreadsheets; there's realtively little need for dedicated finance packages. Oct 15 15:46:39 what a monkey task... Oct 15 15:46:43 :( Unfortunately the tax code here is such that spreadsheets don't suffice, especially if one does consulting sorts of activities as a side job, as many Linux folks here seem to do. Ah well... Oct 15 15:46:53 i use spreadsheets for most of my finance stuff, but for entering all the transactions / importing qif / whatever, it's a hassle. i expect more people nowadays are moving to mint.com, wesabi.com, quicken online now that it's free, yodlee, etc Oct 15 15:46:59 mwester: :( Oct 15 15:47:14 The online services frighten me. Oct 15 15:47:39 I imagine they are nothing but magnets for identity thieves. Oct 15 15:49:17 * mwester imagines Quicken for Linux... Oct 15 15:49:39 i've heard yodlee has good relationships with financial institutions Oct 15 15:49:57 dunno about the others Oct 15 15:50:21 wesabi you have to upload your transactions, so it doesnt hold your financial account information, just the transactions Oct 15 15:50:24 which isn't quite as bad Oct 15 15:50:35 bit less convenient, but.. Oct 15 15:52:33 zecke: wb Oct 15 15:53:50 mwester: would be cool to see an online financial setup that works like clipperz, zero knowledge web app, make the data on the server end encrypted, decrypted by the local javascript, so their servers never see anything other than the big encrypted blob Oct 15 15:54:19 That's a good idea. Oct 15 16:14:04 i have a question about new git version Oct 15 16:14:14 who might help me? Oct 15 16:18:44 btw, about git Oct 15 16:19:04 * Jay7 gets fatal: The remote end hung up unexpectedly Oct 15 16:19:23 In response to what command? Oct 15 16:19:28 Jay7: which remote? Oct 15 16:19:36 git clone git://git.openembedded.net/org.openembedded.dev org.openembedded.dev Oct 15 16:19:56 hm Oct 15 16:20:05 * Jay7 thinks about firewall Oct 15 16:20:11 Jay7: yes? Oct 15 16:20:17 Jay7: where do you get that information from? Oct 15 16:20:18 what ports uses git? Oct 15 16:20:29 zecke: for your amusement, I updated my old collection of benchmark results to include the gta01. http://copper.reciva.com/pb/lmbench.txt Oct 15 16:20:30 from http://wiki.openembedded.net/index.php/Git Oct 15 16:20:30 Jay7: clone git://git.openembedded.net/openembedded Oct 15 16:20:53 ah.. Oct 15 16:20:58 Jay7: fix the wiki page Oct 15 16:21:04 http://gitweb.openembedded.net/?p=openembedded.git;a=summary has a proper URL Oct 15 16:22:29 i want to download only dreambox branch Oct 15 16:22:35 what am i gonna do? Oct 15 16:22:44 for dev branch Oct 15 16:22:56 git clone git://git.openembedded.net/openembedded Oct 15 16:22:57 is ok Oct 15 16:22:59 prokami: git-clone siwtch branches? Oct 15 16:23:03 Why do you not want to download anything else? Oct 15 16:23:14 prokami: git-checkout -b org.openembedded.dreambox origin/org.openembedded.dreambox Oct 15 16:25:15 thanks a lot Oct 15 16:26:24 wiki page was already changed by cbrake Oct 15 16:26:43 but now there is no block with sample commands Oct 15 16:41:09 03Robert Schuster  07org.openembedded.dev * r073f0ef2e1 10openembedded.dev.git/packages/mozilla/firefox.inc: firefox.inc: Fix typo. Oct 15 16:44:22 mwester: so, going back to what we were talking about yesterday, it looks like disabling oabi compatibility gives you about a 10% speedup on system call entry on openmoko. I'm not sure whether you would get the same on nslu2 but I suspect it would be something along the same lines. Oct 15 16:46:32 10%? sounds a lot to me Oct 15 16:46:37 interesting Oct 15 16:46:49 * florian disables oabi compatibility in some kernel configs Oct 15 16:46:53 heh Oct 15 16:46:53 :) Oct 15 16:47:18 admittedly I did also disable frame pointers between the two versions I was comparing, but the null syscall path doesn't touch any C code so that shouldn't have had any impact. Oct 15 16:47:47 in that lmbench link I posted above, the 2.6.24 kernel is the standard openmoko one (from fso-image a few weeks ago) and 2.6.24pb is the same thing with oabi and frame pointers turned off. Oct 15 16:54:19 mickey|coffeebrk: xora was saying something yesterday about having oabi support in the openmoko kernels because of some crazy binary-only gps driver. Oct 15 16:55:24 03Holger Hans Peter Freyther  07org.openembedded.dev * rff30b7fe2e 10openembedded.dev.git/ (7 files in 7 dirs): [squeak] Remove unbuildable old version of the squeak vm Oct 15 16:55:25 pb__: gps userspace application Oct 15 16:55:28 03Holger Hans Peter Freyther  07org.openembedded.dev * rb6c5a08721 10openembedded.dev.git/contrib/site-conf/ (aclocal.m4 configure.ac m4/squeak.m4): [site-conf] Add the tests for the squeak vm Oct 15 16:55:30 03Holger Hans Peter Freyther  07org.openembedded.dev * r9d50508683 10openembedded.dev.git/packages/squeak/ (files/configure-fixes.patch squeakvm_3.10.1.bb): Oct 15 16:55:30 [squeak] Use a regenrated configure file, build against ffi Oct 15 16:55:30 Various fixes to the buildsystem. Sadly we have to use a regenerated Oct 15 16:55:31 configure file as something does not work from within OE (using OE's Oct 15 16:55:33 autoreconf does work though) Oct 15 16:55:35 Change mkconfig.in to use the right tail syntax, build with libffi Oct 15 16:55:37 03Holger Hans Peter Freyther  07org.openembedded.dev * r83a3c665db 10openembedded.dev.git/packages/iotop/iotop_0.2.1.bb: Oct 15 16:55:39 bye bye CIA :) Oct 15 16:55:40 [iotop] Actually remove the site.py and site.pyo from the package Oct 15 16:55:42 Use do_install_append instead of _prepend... to remove the files Oct 15 16:55:44 that are provided by setuptools as well. Oct 15 16:55:46 03Holger Hans Peter Freyther  07org.openembedded.dev * r9357e39aa0 10openembedded.dev.git/ (conf/checksums.ini packages/iotop/iotop_0.2.1.bb): [iotop] Add iotop 0.2.1 to OpenEmbedded/Openmoko Oct 15 16:55:55 03Holger Hans Peter Freyther  07org.openembedded.dev * r2303a5b6c8 10openembedded.dev.git/ (3 files in 3 dirs): Oct 15 16:55:58 [squeak] Attempt to add squeak 3.10-1 Oct 15 16:56:00 - This does configure Oct 15 16:56:01 zecke: right, yeah, that. Oct 15 16:56:02 - But it fails to link, there is something wrong with the Oct 15 16:56:04 autoreconf... Oct 15 16:56:06 03Holger Hans Peter Freyther  07org.openembedded.dev * r472f9f75b3 10openembedded.dev.git/packages/python/python-pyxdg_0.15.bb: [python] Add pyxdg to use fdo specs from python Oct 15 16:56:09 03Holger Hans Peter Freyther  07org.openembedded.dev * rb652b082b0 10openembedded.dev.git/packages/iotop/iotop_0.2.1.bb: [iotop] Make the installed package work Oct 15 16:56:12 03Holger Hans Peter Freyther  07org.openembedded.dev * r9b6657686a 10openembedded.dev.git/ (6 files in 4 dirs): [prelink] Add recent version of prelink Oct 15 16:56:23 03Holger Freyther  07org.openembedded.dev * r9956063810 10openembedded.dev.git/ (3 files in 3 dirs): Oct 15 16:56:26 wvstreams: Add initial compiling and linking version. This is needed for wvdial Oct 15 16:56:28 - Cure some symptoms of the broken buildsystem Oct 15 16:56:30 03Holger Freyther  07org.openembedded.dev * ra6f6def24c 10openembedded.dev.git/ (conf/checksums.ini packages/wvdial/wvdial_1.60.bb): wvdial: Add wvdial 1.6.0 Oct 15 16:56:35 03Carsten Haitzler  07org.openembedded.dev * r7854e5fe62 10openembedded.dev.git/ (5 files in 3 dirs): Oct 15 16:56:38 update to wpa_supplicant 0.6.3 (also add 0.6.0 in) and build packages for Oct 15 16:56:40 it. Oct 15 16:56:42 03freyther@openembedded.org  07org.openembedded.dev * re85c658274 10openembedded.dev.git/packages/wvstreams/wvstreams_4.4.1.bb: wvstreams: Split wvstreams into smaller packages Oct 15 16:56:48 zecke: I think you would be better off ditching oabi support and making a little translator shim for your one lame gps binary. Oct 15 16:56:51 03Julian_chu  07org.openembedded.dev * r9da3205b92 10openembedded.dev.git/conf/checksums.ini: [checksums] Add checksum for mysql-4.1.22.tar.gz Oct 15 16:56:54 03Julian_chu  07org.openembedded.dev * r18d0f1b91a 10openembedded.dev.git/conf/checksums.ini: [checksums] Add checksum for powertop-1.8.tar.gz Oct 15 16:56:57 03Holger Hans Peter Freyther  07org.openembedded.dev * rc7f35945f2 10openembedded.dev.git/conf/checksums.ini: [checksums] Add PyYAML 3.05 Oct 15 16:57:02 03Julian_chu  07org.openembedded.dev * ra4c5a840d0 10openembedded.dev.git/conf/checksums.ini: [checksums] Add checksum for Arkanoid_PSID.sid (fso-sound use it) Oct 15 16:57:05 03Julian_chu  07org.openembedded.dev * r383d9c649f 10openembedded.dev.git/conf/checksums.ini: [checksums] Add checksum for wxWidgets-2.8.9.tar.gz Oct 15 16:57:10 (34 lines omitted) Oct 15 16:58:03 pb__: ooh, right. the gllin situation. I think andrew once posted a just-in-time shim some months ago Oct 15 16:58:22 balrog-kun, that is Oct 15 16:59:14 right, that sounds like the kind of thing you want. it's a bit sad to make everything else run more slowly just for the sake of this one binary. Oct 15 17:00:19 so classpath actually builds in dev and not stable? Oct 15 17:00:50 03Holger Hans Peter Freyther  07org.openembedded.dev * r55e0ea8b49 10openembedded.dev.git/classes/base.bbclass: [base] Autodetect the SCM and add listing of git repositories Oct 15 17:01:01 03Holger Hans Peter Freyther  07org.openembedded.dev * rf4a36e1e11 10openembedded.dev.git/classes/base.bbclass: [base.bbclass] Make it fatal if no entry in checksums.ini exist Oct 15 17:01:57 what is the difference between git pull and git pull --rebase? Oct 15 17:02:24 Crofton|work: git-pull will do a merge by default Oct 15 17:02:35 you want --rebase usually Oct 15 17:02:36 Crofton|work: --rebase will reapply your patches on the old head Oct 15 17:03:06 what does git mean by merge? Oct 15 17:03:23 Crofton|work: with rebase you will have a linear history. Oct 15 17:03:31 pactches are local edits, even un-commited changes? Oct 15 17:04:18 so when you rebase, anychanges you made have their ancestor updated Oct 15 17:08:01 Crofton|work: there's some diagrams that might explain rebasing: http://www.kernel.org/pub/software/scm/git/docs/git-rebase.html Oct 15 17:08:13 Kensan, thanks Oct 15 17:09:36 if you want to retain uncommited patches, you only can use pull Oct 15 17:11:33 03Julian_chu  07org.openembedded.dev * rac2ebfa7ae 10openembedded.dev.git/packages/efl1/edje_svn.bb: Oct 15 17:11:33 [edje-utils] Add recommends to edje-utils Oct 15 17:11:33 edje_decc, a tool of edje-utils, need some modules when decompling edj file. Oct 15 17:11:33 Add those modules as recommends. Oct 15 17:11:33 IMHO you could branch off master then fetch/rebase master and then merge the temporary branch into master. Oct 15 17:11:33 03Holger Hans Peter Freyther  07org.openembedded.dev * rc6ffab30ef 10openembedded.dev.git/packages/efl1/ (ecore.inc ecore_svn.bb): Oct 15 17:11:38 [ecore] Fix DEPENDS to make the build deterministic Oct 15 17:11:40 XCursor, XRandr, Xinerama and Xprint was found and used on the Oct 15 17:11:42 second configure run but not on the first one. Fix that. Oct 15 17:11:44 03Julian_chu  07org.openembedded.dev * rbcc60080c4 10openembedded.dev.git/ (2 files in 2 dirs): Oct 15 17:11:47 [edje-sdk] Add edje-sdk. Oct 15 17:11:49 edje-sdk contains some utils what can be used in toolchain. Oct 15 17:11:53 03Michael 'Mickey' Lauer  07org.openembedded.dev * r876091d0f3 10openembedded.dev.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Oct 15 17:12:00 03Michael 'Mickey' Lauer  07org.openembedded.dev * r18e31101a8 10openembedded.dev.git/packages/mozilla/firefox.inc: firefox.inc: fix syntax Oct 15 17:12:29 03Koen Kooi  07org.openembedded.dev * r201e05c72b 10openembedded.dev.git/packages/linux/linux.inc: linux.inc: add switch to turn of OABI compat for arm kernels Oct 15 17:12:30 03Koen Kooi  07org.openembedded.dev * r0f14c351dd 10openembedded.dev.git/conf/distro/include/angstrom.inc: angstrom: turn of OABI compat in EABI kernels Oct 15 17:13:39 why does classpath require gtk+?? Oct 15 17:14:09 echelon: AWT requires a toolkit? Oct 15 17:14:23 it could just as well use X11 Oct 15 17:14:26 echelon: and java should integrate with the desktop, e.g. have similar dialogs... Oct 15 17:14:35 echelon: there is no stock File dialog in X11 Oct 15 17:14:44 echelon: there is no stock way to use icon themes with X11.. :) Oct 15 17:14:58 right, in fact there is nothing really recognisable as a modern toolkit in x11. Oct 15 17:15:07 i just need it to run j2me/midp applications :\ Oct 15 17:15:17 you could indeed reimplement awt on top of raw xlib but it would be a pile of work and the resulting apps would look and behave nothing like any others. Oct 15 17:16:07 echelon: gtk isn't that big, I think you should just grit your teeth and install it. Oct 15 17:16:30 echelon: then don't use classpath Oct 15 17:16:30 anyway, cocao-initial's bitbake is broken in stable Oct 15 17:16:40 i now have both branches :P Oct 15 17:16:46 echelon: classpath is the whole thing Oct 15 17:17:12 echelon: used midpath (or whatever the gpl version/fork is called) for only j2me Oct 15 17:18:24 zecke: what vm did you use? Oct 15 17:18:53 hi folks, now that monotone is out the window and git is here, is there a way to grab the tree without using git:// as corp firewall blocks it ? Oct 15 17:19:05 echelon: I think cacao work with midpath but Robert and Henning can tell you more Oct 15 17:19:14 03Koen Kooi  07org.openembedded.dev * r6840b8c693 10openembedded.dev.git/packages/linux/linux-rp.inc: linux-rp.inc: add switch to disable OABI compat like linux.inc has Oct 15 17:19:27 ade|desk: yes, go to repo.or.cz, ask for things to be mirrored, grab the result with http Oct 15 17:19:41 ah right Oct 15 17:21:00 zecke: is there such a service available for subversion? Oct 15 17:22:02 echelon: not that I know of Oct 15 17:22:31 k Oct 15 17:27:22 echelon: svn is centralized scm. It's not that easy (if at all possible without quite some effort) to "mirror"/clone the repo. Oct 15 17:27:38 Kensan: svn can easily do that :) Oct 15 17:28:06 zecke: how? Oct 15 17:28:34 Kensan: svk Oct 15 17:28:43 I wanted to type svk :) Oct 15 17:29:02 zecke: ah well svk is not svn ;) Oct 15 17:29:29 zecke: one could also do it with git-svn I guess *heh* Oct 15 17:29:47 one must be insane to switch to svn now :) Oct 15 17:30:02 ERROR: libxml2-2.7.2: ftp://xmlsoft.org/libxml2/libxml2-2.7.2.tar.gz has no entry in conf/checksums.ini, not checking URI Oct 15 17:30:19 I guess we now force files to have an entry in checksums.ini? Oct 15 17:30:27 ya Oct 15 17:30:32 came over the mailing list some days ago Oct 15 17:30:38 zecke: well those people are all "stupid & ugly" according to some "dude on youtube" *heh* Oct 15 17:30:49 Crofton|work: I just pushed the change :) Oct 15 17:31:00 Crofton|work: in tmp/checksums.ini is a entry ready to copy and paste Oct 15 17:31:29 you are fixing atm? Oct 15 17:32:12 Crofton|work: no, I share the fixing with you guys now Oct 15 17:32:16 ok Oct 15 17:32:19 Crofton|work: I just tell you what to do Oct 15 17:32:20 I will fix Oct 15 17:32:31 md5sum source file, add to checksums? Oct 15 17:32:31 cool :) Oct 15 17:32:55 I just did not want to test two people making identical commits Oct 15 17:32:57 Crofton|work: tmp/checksums.ini in your build dir, just copy the three lines to conf/checksums.ini Oct 15 17:32:59 Well, assuming you trust the source file you have in your possession of course ;) Oct 15 17:33:08 it will take me a few minutes Oct 15 17:33:20 I assume I downloaded it so I coould cechk it ... Oct 15 17:37:46 now I need to read the git phrasebook Oct 15 17:38:14 mwester: trust? :) Oct 15 17:41:00 * mwester has an email folder full of correspondence with his ISP that (apparently) document a major failing of TCP checksumming. :( Oct 15 17:41:33 urg Oct 15 17:41:36 good point Oct 15 17:41:48 well, if I have a bad file, people will bitch :) Oct 15 17:42:22 Big thumbs up for the required md5sums; and I now download large files multiple times until I get a pair with identical md5sums. Oct 15 17:43:12 right we have git://repo.or.cz/openembedded.git and http://repo.or.cz/r/openembedded.git as mirrors for us behind firewalls Oct 15 17:43:18 03Philip Balister  07org.openembedded.dev * r4a9b373df5 10openembedded.dev.git/conf/checksums.ini: libxml2-2.7.2.tar.gz : Add entry in conf/checksums.ini Oct 15 17:43:26 ok, let me know if I wrecked git Oct 15 17:43:50 Crofton|work: please do git-config user.email again... the phrasebook was wrong :) Oct 15 17:43:51 yep Oct 15 17:43:56 forgot to set who I am Oct 15 17:44:08 sorry Oct 15 17:44:22 Crofton|work: no, I could blame you, but my phrasebook was wrong :) Oct 15 17:44:29 well, you can Oct 15 17:44:38 because I skipped it entirely Oct 15 17:44:39 :-D The final email exchange in that folder is where a manager at that ISP told me I should not be using download protocols that do not do their own integrity checks. So I replied and asked him if that meant that I was free to use bittorrent despite what the TOS says. Oct 15 17:44:42 is it fixed now? Oct 15 17:45:15 mwester: you could do git clone via https. Oct 15 17:45:28 Jay7: that has been removed from http://wiki.openembedded.net/index.php/Git Oct 15 17:45:47 mwester: but that sounds rather strange... Oct 15 17:46:09 can I use the fake oe address for email Oct 15 17:46:14 or a real email address? Oct 15 17:50:05 Crofton|work: real... Oct 15 17:50:18 Crofton|work: but you can set up a fake one Oct 15 17:50:34 does oe.org forward? Oct 15 17:50:59 for some people, e.g. not me Oct 15 17:51:14 I can't remember if mine does ... Oct 15 17:51:20 how can one delete a wikipage? Oct 15 17:51:27 Crofton|work: you can test it :) Oct 15 17:51:35 i am Oct 15 17:51:46 I think there is a delete button somewhere? Oct 15 17:52:12 hmm Oct 15 17:52:16 but I cannot find it Oct 15 17:52:47 boing Oct 15 17:55:28 does the checksum test apply to things from git and svn repos? Oct 15 17:56:46 03Koen Kooi  07org.openembedded.dev * ra3125e86b1 10openembedded.dev.git/conf/checksums.ini: checksums: add more checksums, sort it Oct 15 17:57:49 Oh wow! Koen's email re Atheros releasing the HAL so we don't need the wackelf crap anymore! And we have git. And the Openmoko engineering news. I think I must still be asleep and dreaming.... Oct 15 17:57:55 zecke, I think you need page delete rights Oct 15 17:58:08 mwester: no, drugs Oct 15 17:59:32 Crofton|work: no and not even to the "stashed" tarballs... Oct 15 17:59:47 wew Oct 15 17:59:51 I was a little worried Oct 15 18:01:53 03Koen Kooi  07org.openembedded.dev * r20805ba4b3 10openembedded.dev.git/conf/checksums.ini: checksums.ini: add even more checksums Oct 15 18:08:22 man git pull is slow .... Oct 15 18:09:50 Crofton|work: ? Oct 15 18:10:33 well, it sat there for a bit asking for my pass phrase Oct 15 18:10:47 I expect it to get hammered over the next few days Oct 15 18:10:55 Crofton|work: compared to mtn it is like lightning Oct 15 18:11:00 :) Oct 15 18:11:21 03Koen Kooi  07holger/hash-style * r52161525d0 10openembedded.dev.git/packages/sysstat/sysstat_8.1.6.bb: sysstat: add 8.1.6 Oct 15 18:11:23 03Koen Kooi  07holger/hash-style * r7d36a7b181 10openembedded.dev.git/packages/xorg-driver/xf86-video-omapfb_git.bb: xf86 video omapfb: bump SRCREV Oct 15 18:11:23 03Jan Lübbe  07holger/hash-style * r242ff4c599 10openembedded.dev.git/contrib/mtn2git/manglenames.py: mtn2git: commit the final version used for the bk&mtn to git conversion Oct 15 18:11:24 03Robert Schuster  07holger/hash-style * r98cc700911 10openembedded.dev.git/packages/classpath/ (classpath-0.97.2/drawpolyline.patch classpath_0.97.2.bb): classpath 0.97.2: Added patch to fix to fix drawPolyline bug. Oct 15 18:11:27 03Marcin Juszkiewicz  07holger/hash-style * r14ae1f974c 10openembedded.dev.git/MAINTAINERS: MAINTAINERS: I am no longer interested in Tosa Oct 15 18:11:30 03Marcin Juszkiewicz  07holger/hash-style * r40fdfee5f6 10openembedded.dev.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Oct 15 18:11:33 03Marcin Juszkiewicz  07holger/hash-style * r9b90ffaa4d 10openembedded.dev.git/MAINTAINERS: MAINTAINERS: I am no longer interested in OpenZaurus Oct 15 18:11:38 03Holger Hans Peter Freyther  07holger/hash-style * r9d50508683 10openembedded.dev.git/packages/squeak/ (files/configure-fixes.patch squeakvm_3.10.1.bb): Oct 15 18:11:41 [squeak] Use a regenrated configure file, build against ffi Oct 15 18:11:43 Various fixes to the buildsystem. Sadly we have to use a regenerated Oct 15 18:11:45 configure file as something does not work from within OE (using OE's Oct 15 18:11:47 autoreconf does work though) Oct 15 18:11:51 Change mkconfig.in to use the right tail syntax, build with libffi Oct 15 18:11:53 03Koen Kooi  07holger/hash-style * r77c7b94c93 10openembedded.dev.git/packages/mozilla/firefox.inc: firefox.inc: add DESCRIPTION Oct 15 18:11:53 see what I mean :) Oct 15 18:11:56 03Holger Hans Peter Freyther  07holger/hash-style * r472f9f75b3 10openembedded.dev.git/packages/python/python-pyxdg_0.15.bb: [python] Add pyxdg to use fdo specs from python Oct 15 18:11:59 03Holger Hans Peter Freyther  07holger/hash-style * rb652b082b0 10openembedded.dev.git/packages/iotop/iotop_0.2.1.bb: [iotop] Make the installed package work Oct 15 18:12:27 03Jan Luebbe  07holger/hash-style * r869ba9e1fb 10openembedded.dev.git/ (4403 files in 4403 dirs): mtn2git: remove .mtn2git_empty files in non-empty directories Oct 15 18:12:30 03Julian_chu  07holger/hash-style * r9da3205b92 10openembedded.dev.git/conf/checksums.ini: [checksums] Add checksum for mysql-4.1.22.tar.gz Oct 15 18:12:30 03Holger Hans Peter Freyther  07holger/hash-style * r9b6657686a 10openembedded.dev.git/ (6 files in 4 dirs): [prelink] Add recent version of prelink Oct 15 18:12:30 03Holger Hans Peter Freyther  07holger/hash-style * rfe6b416f8d 10openembedded.dev.git/conf/checksums.ini: [checksums] Add alternative fbreader source Oct 15 18:12:32 03Julian_chu  07holger/hash-style * r383d9c649f 10openembedded.dev.git/conf/checksums.ini: [checksums] Add checksum for wxWidgets-2.8.9.tar.gz Oct 15 18:12:35 03freyther@openembedded.org  07holger/hash-style * re85c658274 10openembedded.dev.git/packages/wvstreams/wvstreams_4.4.1.bb: wvstreams: Split wvstreams into smaller packages Oct 15 18:12:38 03Holger Freyther  07holger/hash-style * ra6f6def24c 10openembedded.dev.git/ (conf/checksums.ini packages/wvdial/wvdial_1.60.bb): wvdial: Add wvdial 1.6.0 Oct 15 18:12:39 hmm our git-notify needs work :) Oct 15 18:12:41 03Julian_chu  07holger/hash-style * r18d0f1b91a 10openembedded.dev.git/conf/checksums.ini: [checksums] Add checksum for powertop-1.8.tar.gz Oct 15 18:12:48 03Julian_chu  07holger/hash-style * rabfa317081 10openembedded.dev.git/conf/checksums.ini: [checksums] Add checksum for yaml-0.1.1 Oct 15 18:12:51 03Julian_chu  07holger/hash-style * re002d8faef 10openembedded.dev.git/conf/checksums.ini: [checksums] Add checksum for gdb-6.7.1 Oct 15 18:12:54 03Holger Hans Peter Freyther  07holger/hash-style * rc7f35945f2 10openembedded.dev.git/conf/checksums.ini: [checksums] Add PyYAML 3.05 Oct 15 18:12:57 03Julian_chu  07holger/hash-style * r48b45d4beb 10openembedded.dev.git/conf/checksums.ini: [checksums] Add checksum for systat-8.1.5.tar.gz Oct 15 18:13:00 03Holger Hans Peter Freyther  07holger/hash-style * r5e4b4c2c2a 10openembedded.dev.git/conf/checksums.ini: [checksums] Add SDL image 1.2.6 Oct 15 18:13:03 03Holger Hans Peter Freyther  07holger/hash-style * r5bfb0c33f0 10openembedded.dev.git/conf/checksums.ini: [checksums] Add xwininfo to the checksums Oct 15 18:13:10 03Holger Hans Peter Freyther  07holger/hash-style * rca2fa92e11 10openembedded.dev.git/conf/checksums.ini: [checksums] Add asterisk 1.4.17 Oct 15 18:13:13 03Holger Hans Peter Freyther  07holger/hash-style * rc70919dd6e 10openembedded.dev.git/conf/checksums.ini: [checksums] Add bluez-hcidump 1.42 Oct 15 18:13:16 03Holger Hans Peter Freyther  07holger/hash-style * r2d2729552d 10openembedded.dev.git/conf/checksums.ini: [checksum] Add checksum of elfutils Oct 15 18:13:21 night all Oct 15 18:13:21 03Holger Hans Peter Freyther  07holger/hash-style * rdc3780eb11 10openembedded.dev.git/conf/checksums.ini: [checksums] Add xournal 0.4.0.1 Oct 15 18:13:24 03Julian_chu  07holger/hash-style * rd07fc56a2d 10openembedded.dev.git/conf/checksums.ini: [checksums] Add checksum for x11vnc-0.9.4.tar.gz Oct 15 18:13:27 03Holger Hans Peter Freyther  07holger/hash-style * r7d7f2e0364 10openembedded.dev.git/conf/checksums.ini: [checksums] Add my sums for for wpa_supplicant Oct 15 18:13:36 03Holger Hans Peter Freyther  07holger/hash-style * rc5391cfb3f 10openembedded.dev.git/conf/checksums.ini: [checksums] Add libmodplug 1.8 Oct 15 18:13:39 03Holger Hans Peter Freyther  07holger/hash-style * ra0b49a2a42 10openembedded.dev.git/conf/checksums.ini: [checksums] Add checksums for hal Oct 15 18:13:42 03Holger Freyther  07holger/hash-style * r5255d67147 10openembedded.dev.git/conf/checksums.ini: [checksums] Add my sums for libtool Oct 15 18:13:47 03freyther@openembedded.org  07holger/hash-style * rc4c838db5a 10openembedded.dev.git/conf/checksums.ini: checksums: iproute2 2.6.22-070710 Oct 15 18:13:50 03Julian_chu  07holger/hash-style * r83ceb1bf17 10openembedded.dev.git/conf/checksums.ini: [checksum] Add checksum for midori-0.0.17 Oct 15 18:13:53 03Holger Hans Peter Freyther  07holger/hash-style * rf0ae264dee 10openembedded.dev.git/conf/checksums.ini: [checksums] Add checksums for navit Oct 15 18:14:00 03freyther@openembedded.org  07holger/hash-style * r275768c948 10openembedded.dev.git/conf/checksums.ini: checksums: Add popt 1.13 Oct 15 18:14:03 03Holger Hans Peter Freyther  07holger/hash-style * r7f920fe087 10openembedded.dev.git/conf/checksums.ini: [checksums] Add checksums for portaudio Oct 15 18:14:06 03Holger Hans Peter Freyther  07holger/hash-style * r126994d5b2 10openembedded.dev.git/conf/checksums.ini: [checksums] Add pygame 1.8.1 Oct 15 18:14:11 03Holger Hans Peter Freyther  07holger/hash-style * ra64ba1e11c 10openembedded.dev.git/conf/checksums.ini: [checksum] Add prelink checksum for the normal mirror Oct 15 18:14:14 03Robert Schuster  07holger/hash-style * r4e870c5ec5 10openembedded.dev.git/packages/linux/ (linux-2.6.24/atngw100/defconfig linux_2.6.24.bb): Oct 15 18:14:17 linux 2.6.24: Updated defconfig for atngw100 Oct 15 18:14:19 - compile ext2 and ext3 into the kernel (to allow booting from SD) Oct 15 18:14:23 03Holger Freyther  07holger/hash-style * r29b68e1044 10openembedded.dev.git/conf/checksums.ini: [checksums] gpsd 2.36 Oct 15 18:14:26 03Holger Hans Peter Freyther  07holger/hash-style * re4fe3ae0de 10openembedded.dev.git/conf/checksums.ini: [checksum] Add checksum for pyxdg, matches the debian source Oct 15 18:14:29 (56 lines omitted) Oct 15 18:16:05 03Holger Freyther  07holger/hash-style * re4f5585fe2 10openembedded.dev.git/conf/ (4 files in 2 dirs): Oct 15 18:16:05 [commit] Allow the distribution to set the hash-style and enable it for angstrom Oct 15 18:16:05 With binutils >= 2.17 we can use --hash-style=both and this will work Oct 15 18:16:05 with all versions of glibc. glibc >= 2.5 can take advantage of the new hash Oct 15 18:16:06 style and should give better application startup. Oct 15 18:16:25 03Holger Hans Peter Freyther  07holger/hash-style * rb82be9b7e8 10openembedded.dev.git/ (3276 files in 1160 dirs): Oct 15 18:16:25 [PR] Change PR to FILE_PR and intro PR as FILE_PR + DISTRO_PR Oct 15 18:16:25 Allow a distribution to globally bump the PR of every package Oct 15 18:16:25 this is useful when there was a change in the toolchain and Oct 15 18:16:25 every package is going to be different. Oct 15 18:16:27 find packages/ \( -name '*.bb' -or -name '*.inc' \) -exec sed -i s/"^PR="/"FILE_PR ="/ {} \; Oct 15 18:16:31 find packages/ \( -name '*.bb' -or -name '*.inc' \) -exec sed -i s/"^PR ="/"FILE_PR ="/ {} \; Oct 15 18:20:57 when is RP back from the conference? Oct 15 18:25:54 03Philip Balister  07org.openembedded.dev * re765346792 10openembedded.dev.git/conf/checksums.ini: numpy-1.1.1.tar.gz : Update checksums.ini Oct 15 18:26:24 zecke: Did you add my key yet? Oct 15 18:27:06 alphaone: maybe Oct 15 18:27:19 alphaone: I might have another brain fart and mispelled you :) Oct 15 18:27:20 zecke: I get ERROR:gitosis.serve.main:Repository read access denied Oct 15 18:27:50 alphaone: could you pastebin your key? Oct 15 18:28:33 http://pastebin.com/md9819b0 Oct 15 18:29:15 03Marcin Juszkiewicz  07org.openembedded.dev * r4211f6e483 10openembedded.dev.git/conf/checksums.ini: Oct 15 18:29:15 checksums.ini: re-sorted to remove duplicates Oct 15 18:29:15 Please do not sort this file by hand - there is a tool written specially Oct 15 18:29:15 for this: contrib/source-checker/oe-checksums-sorter.py Oct 15 18:29:16 03Marcin Juszkiewicz  07org.openembedded.dev * r83955fe2a5 10openembedded.dev.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Oct 15 18:29:19 03Marcin Juszkiewicz  07org.openembedded.dev * r82a7db80b6 10openembedded.dev.git/conf/checksums.ini: checksums.ini: merged with Poky Oct 15 18:30:17 wuahah Oct 15 18:30:22 oe git spam Oct 15 18:32:41 moin Oct 15 18:33:06 zecek hm Oct 15 18:33:19 I will look at you squeak attempts Oct 15 18:33:35 03Michael 'Mickey' Lauer  07org.openembedded.dev * r76640097cf 10openembedded.dev.git/packages/ttf-fonts/ttf.inc: ttf.inc: INHIBIT_DEFAULT_DEPS Oct 15 18:33:35 03Michael 'Mickey' Lauer  07org.openembedded.dev * r7a4455feaa 10openembedded.dev.git/packages/tasks/task-fonts-truetype.bb: add task-fonts-truetype Oct 15 18:34:50 zecke do we have /var/run/network as dir in oe? Oct 15 18:34:59 no idea Oct 15 18:35:36 wpa_supplicant needs a patch Oct 15 18:35:47 to run porper with udev and ifupdown Oct 15 18:36:01 ohloh is making sense of OE now Oct 15 18:36:02 https://www.ohloh.net/projects/3928 Oct 15 18:38:51 hi woglinde Oct 15 18:39:10 hi pb :) Oct 15 18:41:18 hi florian Oct 15 18:41:38 zecke: http://pastebin.com/md9819b0 Oct 15 18:41:46 03Holger Hans Peter Freyther  07org.openembedded.dev * rdd3582d219 10openembedded.dev.git/packages/connman/connman_git.bb: Oct 15 18:41:46 [connman] Use the http protocol to be able to git-clone/git-fetch connman Oct 15 18:41:46 rsync: link_stat "projects/connman.git/packed-refs" (in repos) failed: No such file or directory (2) Oct 15 18:41:46 rsync error: some files could not be transferred (code 23) at main.c(1385) [receiver=2.6.9] Oct 15 18:41:48 fatal: Could not run rsync to get refs Oct 15 18:41:50 03Julian_chu  07org.openembedded.dev * raed1f4ad67 10openembedded.dev.git/ (3 files in 3 dirs): Oct 15 18:41:53 [connman] Upgrade connman to latest revision. Oct 15 18:41:55 connman use native _nm_ to list symbols. Oct 15 18:41:57 However, it may cause problem when cross compiling. Oct 15 18:41:59 Add a patch to use cross-tools if set host=$SOMETHING Oct 15 18:42:01 when do configure. Oct 15 18:42:03 03I-Fan, Chen  07org.openembedded.dev * rffdec57267 10openembedded.dev.git/packages/connman/connman_git.bb: Oct 15 18:42:06 [connman] modify the git server from moblin.org to git.moblin.org Oct 15 18:42:08 thanks Dietmar Friede info the setting of servers changed. Oct 15 18:42:10 03Holger Hans Peter Freyther  07org.openembedded.dev * rc438f9d0b3 10openembedded.dev.git/packages/libgdbus/libgdbus_git.bb: Oct 15 18:42:13 [naming policy] If we build a specific revision from a SCM have a "r" after the SCM name Oct 15 18:42:19 Clean up PV naming and make it consistent. if we build a revision from a Oct 15 18:42:21 scm system then have svnr/gitr in the package name. Oct 15 18:42:23 03Holger Hans Peter Freyther  07org.openembedded.dev * r3fa03adfe1 10openembedded.dev.git/packages/libgdbus/libgdbus_git.bb: Oct 15 18:42:26 [libgdbus] Use the right git server, use the right protocol Oct 15 18:42:28 - moblin.org => git.moblin.org Oct 15 18:42:30 - really use http as protocol Oct 15 18:44:06 zecke hm why you need popt 0.13 when there is 0.14? Oct 15 18:44:51 woglinde: need? we have popt 0.13 in the metadata Oct 15 18:44:59 alphaone: hmmm Oct 15 18:45:04 alphaone: paste your error please Oct 15 18:45:34 zecke: http://pastebin.com/m2b629ba6 Oct 15 18:45:44 Or is that not the correct address? Oct 15 18:46:31 alphaone: hmmm Oct 15 18:49:07 alphaone: try again please Oct 15 18:49:31 http://pastebin.com/m7e04be16 Oct 15 18:51:17 alphaone: you are daniel_willmann? Oct 15 18:51:41 zecke: define "you are" Oct 15 18:51:46 *g* Oct 15 18:52:01 Daniel Willmann is my name, yes Oct 15 18:52:17 alphaone is my username on thebe Oct 15 18:52:24 But that shouldn't matter, should it? Oct 15 18:52:47 now it's workinf Oct 15 18:52:49 g Oct 15 18:53:01 alphaone: I don't figure out if gitosos looks at the filename and or the content Oct 15 18:53:12 it doesn't iirc Oct 15 18:53:14 alphaone: I did change nothing, must be a user issue :) Oct 15 18:53:24 [looking at the content] Oct 15 18:53:25 zecke: Only filename of key matters iirc Oct 15 18:53:35 zecke: I didn't either :-) Oct 15 18:53:54 Just repeated last command Oct 15 18:54:02 * zecke had to change the key content for the initial repo... hmm might be a special case of gitosis-init Oct 15 19:00:47 zecke: ping Oct 15 19:00:55 rschuster: pong, should I come over? Oct 15 19:01:08 zecke: sure :) Oct 15 19:01:27 how long do you stay? Oct 15 19:01:29 there is still plenty of free beer :) Oct 15 19:01:40 ~midnight Oct 15 19:01:41 Midi/Karaoke player and C++ midi library. URL: http://midnight.linuxbox.com, or some horse in World of Warcraft, or a middle of the night, 0:00 Oct 15 19:01:45 i need to code :) Oct 15 19:02:06 rschuster: or actually merge some patches... Oct 15 19:02:31 yeah, i've seen it already. really nice work! Oct 15 19:03:17 now hal to get rid of a lot of memleaks Oct 15 19:06:05 03Holger Hans Peter Freyther  07org.openembedded.dev * rbabeb73693 10openembedded.dev.git/packages/hal/hal_0.5.9.bb: [hal] Remove commented out things we have a SCM to recover it Oct 15 19:06:07 03Holger Hans Peter Freyther  07org.openembedded.dev * r20e32f99b6 10openembedded.dev.git/packages/hal/ (hal.inc hal_0.5.9.bb): Oct 15 19:06:07 [hal] Move almost everything into the hal.inc file Oct 15 19:06:07 hal_0.5.9.bb is the only user of hal.inc and we can safely move everything Oct 15 19:06:08 into the common include file. Oct 15 19:06:10 03Holger Hans Peter Freyther  07org.openembedded.dev * rb0252872d7 10openembedded.dev.git/packages/hal/ (hal-info.inc hal-info_20070618.bb hal-info_20080313.bb): [hal] Create hal-info.inc and share it with the releases Oct 15 19:06:14 03Holger Hans Peter Freyther  07org.openembedded.dev * rf629930ce8 10openembedded.dev.git/ (5 files in 3 dirs): Oct 15 19:06:17 [hal] Add latest hal and hal-info Oct 15 19:06:19 hal-0.5.11 should fix many many leaks compared to hal-0.5.9 Oct 15 19:06:21 and upgrade hal-info on the way as well Oct 15 19:06:59 I might not be able to merge as much as this every day... :} Oct 15 19:07:42 zecke:it seems to me that you are in the 'flow' Oct 15 19:08:11 git mojo Oct 15 19:08:12 so its better not to disturb you :) Oct 15 19:08:13 rschuster: yes, I will probably not end up in the cbase tonight Oct 15 19:08:36 * zecke is reviewing and merging webkit patches :} Oct 15 19:08:37 zecke has recovered from jetlag, it seems :) Oct 15 19:09:35 mwester: regarding OM stability: fixable tasks look events/0 taking all the CPU, wlan in itself, xglamo Oct 15 19:10:17 mwester: you found a sad workaround for #1024, we have some esoteric suspend/resume issues that I don't understand... Oct 15 19:10:31 anyone has the balls to say yes to my FILE_PR change? Oct 15 19:11:16 Has anyone objected? :) Oct 15 19:11:28 mwester: it touches every file :) Oct 15 19:12:30 Yes, but after a git migration, shouldn't people be expecting to do a full rebuild? Seems like good timing to me, and nothing but good comments on the ML. (And I too like the idea). Oct 15 19:13:18 I did most of the conversion with sed... and a fresh rebuild for me would take a day... Oct 15 19:13:20 heh! "sad workaround" <-- good way to phrase it. Oct 15 19:13:46 ~curses evolution Oct 15 19:14:52 or fedora for not packaging enigmail ... Oct 15 19:17:14 i'm trying to build midpath from dev branch, but i keep getting this error.. ERROR: '/tmp/openembedded/org.openembedded.dev/packages/libsdl/libsdl-image_1.2.6.bb' failed Oct 15 19:17:57 echelon: yes? but that is not the only message you get Oct 15 19:18:35 * mwester reads pb_'s email, and just bets that SlugOS is still on glibc 2.4 :-D Oct 15 19:18:52 1 sec i tried to bitbake libsdl-image by itself this time Oct 15 19:19:31 Nope, I'm wrong -- glibc 2.5 for slugos. Today is a particularly good day, it seems... Oct 15 19:19:40 lol Oct 15 19:20:14 zecke: http://rafb.net/p/P0TuO037.html Oct 15 19:20:52 wow, ohloh has the commit timeline Oct 15 19:20:58 i've been away from oe longer than i was with oe Oct 15 19:21:18 ah koen Signed-Off my change Oct 15 19:21:18 still on the top 4 contributers though # of commit wise Oct 15 19:21:19 hehe Oct 15 19:23:31 NOTE: package libsdl-gfx-2.0.17: started Oct 15 19:23:31 ERROR: This autoconf log indicates errors, it looked at host includes. Oct 15 19:23:31 Rerun configure task after fixing this. The path was '/tmp/openembedded/tmp-zipit2-angstrom/work/arm-oabi-angstrom-linux/libsdl-image-1.2.6-r0/SDL_image-1.2.6' Oct 15 19:25:47 echelon: okay, then take a look at config.log? Oct 15 19:26:19 re Oct 15 19:26:25 hi hrw Oct 15 19:26:34 Kensan: speaking of ohloh.. we need to change git location there Oct 15 19:26:41 wb Oct 15 19:27:07 config.log shows.. configure: exit 0 Oct 15 19:28:39 http://rafb.net/p/kDdSfs40.html Oct 15 19:28:41 there Oct 15 19:28:49 echelon: so it only has one line entry? Oct 15 19:29:01 echelon: fix the error yourself (and this is meant to be friendly) Oct 15 19:29:53 kergoth: speaking of ohloh.. we need to change git location there Oct 15 19:30:07 echelon: we have an article in our news section explaining this class of errors Oct 15 19:30:48 hm Oct 15 19:30:50 git add Oct 15 19:31:46 the manual read Oct 15 19:34:21 log.configure Oct 15 19:34:56 no errors Oct 15 19:35:39 ugh, i'll just go back to stable branch Oct 15 19:36:17 echelon: even searching for error finds something... Oct 15 19:36:48 the log.configure file shows that everything's fine Oct 15 19:41:40 i don't see any news section Oct 15 19:45:59 hi guys Oct 15 19:46:22 Hey Oct 15 19:46:47 i ve a small problem compiling bitbake task-base or anything else, it fail on shasum-native build, is it something known ? Oct 15 19:46:57 baelo know Oct 15 19:47:00 am I the only one who can't pull changes? (fatal: the remote end hung up unexpectedly) Oct 15 19:47:08 woglinde: hey, could you do me a favor? Oct 15 19:47:09 shasum-native is one of the first packages Oct 15 19:47:15 zecke what? Oct 15 19:47:21 I am watching eli stone Oct 15 19:47:34 woglinde: yes i know, that's why i ask Oct 15 19:47:42 * zecke_ just crashed apple's svn server :} Oct 15 19:47:52 zecke hahaha Oct 15 19:48:47 good zecke !!!! Oct 15 19:49:03 who was the guy that said they were able to run opera-mini with midpath? Oct 15 19:49:11 apple should use git Oct 15 19:49:17 me Oct 15 19:49:58 woglinde: so you didn't get any errors while building midpath/libsdl-image? Oct 15 19:50:05 i'm on the dev branch now Oct 15 19:50:26 echelon: focus. search for error or BAD in config.log Oct 15 19:50:26 I am using the qt backend so I do not need sdl stuff Oct 15 19:50:36 here is my console output http://pastebin.ca/1228134 Oct 15 19:51:35 03Holger Hans Peter Freyther  07org.openembedded.dev * rd07c292757 10openembedded.dev.git/conf/ (3 files in 2 dirs): Oct 15 19:51:35 [commit] Allow the distribution to set the hash-style and enable it for angstrom Oct 15 19:51:35 With binutils >= 2.17 we can use --hash-style=both and this will work Oct 15 19:51:35 with all versions of glibc. glibc >= 2.5 can take advantage of the new hash Oct 15 19:51:37 style and should give better application startup. Oct 15 19:51:55 zecke whats with uclibc? Oct 15 19:52:01 03Holger Hans Peter Freyther  07org.openembedded.dev * raa03004ace 10openembedded.dev.git/ (3277 files in 1160 dirs): (log message trimmed) Oct 15 19:52:01 [PR] Change PR to FILE_PR and intro PR as FILE_PR + DISTRO_PR Oct 15 19:52:01 Allow a distribution to globally bump the PR of every package Oct 15 19:52:01 this is useful when there was a change in the toolchain and Oct 15 19:52:01 every package is going to be different. Oct 15 19:52:02 find packages/ \( -name '*.bb' -or -name '*.inc' \) -exec sed -i s/"^PR="/"FILE_PR ="/ {} \; Oct 15 19:52:06 find packages/ \( -name '*.bb' -or -name '*.inc' \) -exec sed -i s/"^PR ="/"FILE_PR ="/ {} \; Oct 15 19:52:18 woglinde: could you see if it supports GNU_HASH? Oct 15 19:52:34 hm how do I look it up? Oct 15 19:54:19 is someone able to git pull the openembedded repository? Is it down? Oct 15 19:54:57 http://wiki.openembedded.net/index.php/GitPhraseBook#Git_Phrase_Book Oct 15 19:55:49 zecke hm I split up libiconv Oct 15 19:55:59 into a .inc file and version Oct 15 19:56:16 is renamed: libiconv_1.12.bb -> libiconv.inc Oct 15 19:56:22 modified: libiconv_1.12.bb Oct 15 19:56:25 okay as commit? Oct 15 19:56:35 woglinde: you could try to search for ".gnu.hash" Oct 15 19:56:57 woglinde, I'm experiencing the problem on a build host, which worked yesterday - so git is fine? Oct 15 19:57:21 quickdev didnt you see the commits drom zecke every minute Oct 15 19:57:26 zecke_: better search for SHT_GNU_HASH; the section name is not significant in itself Oct 15 19:57:37 pb_: SHT or DHT? Oct 15 19:57:43 hm I guess it will not have it Oct 15 19:57:48 oh, DT I suppose Oct 15 19:58:22 pb_: I tried to find that symbol but my diskio is quite slow :) Oct 15 19:58:23 woglinde: it wouldn't be very hard to add, you can basically steal the code from glibc's dynamic linker Oct 15 19:58:42 given your prowess in porting linuxthreads it should be a walk in the park for a man of your skills Oct 15 19:58:57 pb lol Oct 15 19:59:03 I only copied 2 files Oct 15 19:59:04 * mwester admires pb_'s skill in sales Oct 15 19:59:36 what does this mean.. CROSS COMPILE Badness: /usr/include in INCLUDEPATH: /usr/include/SDL Oct 15 19:59:40 cc1: internal compiler error: in add_path, at c-incpath.c:371 Oct 15 19:59:52 echelon: you have badness in your cflags Oct 15 19:59:57 echelon need a translation f the error message? Oct 15 20:00:17 that'd be nice Oct 15 20:00:31 you passed -I/usr/include/SDL to the compiler Oct 15 20:00:39 when you cross compile no stuff of the host should sneak in Oct 15 20:00:42 thats it is Oct 15 20:00:48 which contains "/usr/include", which the compiler knows is not a good path to include when cross compiling Oct 15 20:00:52 otherwise you can have unreliable results Oct 15 20:00:59 hence, badness Oct 15 20:01:33 ah, so i should make adjustments in the makefile? Oct 15 20:02:01 echelon: look a bit up in this file Oct 15 20:02:11 echelon: you will see a call to /usr/bin/sdl-config Oct 15 20:04:56 configure:12252: found /usr/bin/sdl-config Oct 15 20:05:21 zecke_: out of interest, did you benchmark the speed of --hash-style=both compared to sysv? Oct 15 20:05:27 or rather, the speed of the resulting binaries Oct 15 20:05:37 pb_: not yet, I'm still compiling Oct 15 20:05:43 ok, cool Oct 15 20:05:47 it'd be interesting to know the results Oct 15 20:05:53 pb *g* Oct 15 20:06:08 pb_: the only benchmark I could think of is actually system boot... Oct 15 20:06:16 found it! CFLAGS = -isystem/tmp/openembedded/tmp-zipit2-angstrom/staging/arm-oabi-angstrom-linux/usr/include -fexpensive-optimizations -frename-registers -fomit-frame-pointer -Os -I/usr/include/SDL -D_GNU_SOURCE=1 -D_REENTRANT Oct 15 20:06:16 zecke_: right, I think that is a good benchmark Oct 15 20:06:22 pb did your colleuge use readahead for the speed up? Oct 15 20:06:22 is that the line i should change? Oct 15 20:06:30 of booting your device Oct 15 20:06:43 no, we don't do anything with readahead at the moment Oct 15 20:06:50 hehe Oct 15 20:07:02 echelon: that is a log file, changing it does not change anything :) Oct 15 20:07:09 we should try that, though I don't know how much it will help Oct 15 20:07:17 echelon: proper way. Change configure.ac of sdl image to not search for sdl-config in /usr/bin Oct 15 20:07:24 echelon the fix is to use sdl-config from oe Oct 15 20:07:34 zecke_: that's from the Makefile Oct 15 20:07:52 and we should find out why its not found first Oct 15 20:09:24 pb_: or I could use Jakub's original benchmark Oct 15 20:09:31 what was that? Oct 15 20:09:39 zecke_: i don't see a configure.ac file Oct 15 20:10:29 pb_: http://sources.redhat.com/ml/binutils/2006-06/msg00418.html click on stats Oct 15 20:10:40 echelon: configure.in then? Oct 15 20:11:14 aha, right Oct 15 20:11:22 args damnit Oct 15 20:11:28 I forget the commit message Oct 15 20:11:31 I think system boot is probably a better benchmark for oe than running openoffice.org Oct 15 20:11:35 *sigh* Oct 15 20:12:17 pb_: e.g. when we have a sleep 5 in a init script... we have other things to optimize :) Oct 15 20:12:25 or maybe you could use kde as your benchmark, I guess that has plenty of shared libraries Oct 15 20:12:26 hmah hehe Oct 15 20:12:28 git commit --amend Oct 15 20:12:30 is cool Oct 15 20:12:40 zecke_: heh, do we have an initscript like that? gosh, we suck. Oct 15 20:12:57 zecke_: i don't see anything related Oct 15 20:13:09 woo sleep 5 why this? Oct 15 20:13:21 grep -rn sleep * | wc -l Oct 15 20:13:21 12 Oct 15 20:13:27 doh Oct 15 20:13:30 configure.in: http://rafb.net/p/9FWoJP40.html Oct 15 20:13:31 what the fuck Oct 15 20:13:32 in the init.d folder Oct 15 20:13:43 thus suckz Oct 15 20:13:56 woglinde, git clone git://git.openembedded.net/org.openembedded.dev openembedded - what's wrong with this cmd? Oct 15 20:14:08 woglinde: hopefully all in the restart only Oct 15 20:14:15 quickdev: you didn't read? Oct 15 20:14:19 udev: sleep 60 Oct 15 20:14:22 Ick Oct 15 20:14:38 mwester: yah, udev is the enemy of performance Oct 15 20:14:42 mwester: not in my version Oct 15 20:15:19 pb__, that's why you pre-seed instead :) Oct 15 20:16:11 zecke_, sorry if I missed something, I really tried to solve it by myself, what should I read? Oct 15 20:16:28 quickdev: what did you read so far? Oct 15 20:16:51 03woglinde  07org.openembedded.dev * r4717426a86 10openembedded.dev.git/packages/libiconv/ (libiconv.inc libiconv_1.12.bb): libinconv: split up into .inc and version file Oct 15 20:16:53 so my first commit Oct 15 20:17:33 args Oct 15 20:17:34 damnit Oct 15 20:18:26 :) Oct 15 20:18:45 ok, so nobody have idea about problem with shasum-native build failed Oct 15 20:19:20 pb_: my low hanging fruits are prelink, hash-style, hwcaps to only look for one lib, we do use PACKAGE_ARCHS for that kind of things anyway Oct 15 20:19:46 03woglinde  07org.openembedded.dev * r8f676f4eb8 10openembedded.dev.git/packages/libiconv/libiconv_1.12.bb: libiconv: fix the wrong merge Oct 15 20:20:32 woglinde: why did you merge at all? Oct 15 20:20:45 zecke -rebase did it Oct 15 20:20:51 zecke_: right, those all sound good. I agree about hwcaps, see the mail I sent to the list earlier today. Oct 15 20:21:02 hm why my commt messages looks difrent from yours Oct 15 20:21:25 you might want to consider the other search path things i mentioned as well, I think they should be fairly trivial to do and dentry lookups with jffs2 are relatively expensive. Oct 15 20:21:45 ^_- Oct 15 20:21:51 pb_ we will switch to ubitfs soon Oct 15 20:21:51 i don't have sdl-config Oct 15 20:21:56 *g* Oct 15 20:22:17 woglinde: yah, that is probably a good idea as well Oct 15 20:22:50 I think we will switch to ubifs too Oct 15 20:22:58 or cramfs on platforms where we have enough ram Oct 15 20:23:29 right now though, our two most common systems only have 32MB RAM and almost all of it is in use, no way to use cramfs there. Oct 15 20:23:32 I will test it on my simpad Oct 15 20:24:11 pb_: I assume s3c24xx nand.c is slow as well Oct 15 20:24:38 only need sane initrd Oct 15 20:24:39 for making the fs Oct 15 20:24:39 args Oct 15 20:24:39 damnit Oct 15 20:24:39 the bootloader cant read from it Oct 15 20:24:39 should have better tought about it Oct 15 20:24:46 zecke_: yeah. not really the s3c24xx's fault though, the nand itself is just inherently slow. Oct 15 20:24:47 hm who is writing ubifssupport for uboot? Oct 15 20:26:03 zecke_, I read irc backlog and Git_Phrase_Book, but there no clone command with a specific branch like the way I use it Oct 15 20:26:10 zecke_: suddenly the days of 32-bit strataflash seem like a golden age Oct 15 20:27:12 quickdev: could you paste your command again? Oct 15 20:28:27 zecke_, git clone git://git.openembedded.net/org.openembedded.dev openembedded, it fails on several boxes and worked yesterday on several boxes.. Oct 15 20:29:51 quickdev: so again, where did you get this command from? Oct 15 20:29:51 fatal: The remote end hung up unexpectedly Oct 15 20:29:59 quickdev: an what do you think is a branch name? Oct 15 20:30:25 zecke_, I got it from: http://shr.bearstech.com/Makefile - which was created by rwhitby Oct 15 20:30:46 hrm, what package is libvolume_id in? Oct 15 20:30:49 * pb_ stabs hal Oct 15 20:30:56 I have to correct myself, major parts have been took from another makefile Oct 15 20:31:25 libvolume_id is from e2fsprog*, I think. Oct 15 20:31:31 aha Oct 15 20:31:33 mwester: thanks Oct 15 20:31:37 * pb_ stabs hal again, anyway Oct 15 20:31:37 quickdev: that Makefile hasn't been updated yet. Oct 15 20:31:38 quickdev: was it updated recently? Oct 15 20:31:47 pb_: why? Oct 15 20:31:56 quickdev: I will work on that today. Oct 15 20:31:57 why does it have to be updated? :) Oct 15 20:32:07 zecke_: checking for VOLUME_ID... configure: error: Package requirements (libvolume_id >= 0.77) were not met: Oct 15 20:32:16 quickdev: cause the git repo went from trial to production at a new path Oct 15 20:32:18 Requested 'libvolume_id >= 0.77' but version of libvolume_id is 0.63.0 Oct 15 20:32:24 quickdev: because it was written before OE used git Oct 15 20:32:47 I guess I need a newer e2fsprogs Oct 15 20:32:52 quickdev: please return this thread to #openmoko-cdevel Oct 15 20:33:18 ok, thanks zecke_ :) Oct 15 20:33:32 rwhitby: BTW: did you see that julian started to rsync the build config? Oct 15 20:34:00 hm libvolume-id0 source udev Oct 15 20:34:04 zecke_: I didn't, but that's great. Oct 15 20:34:26 http://downloads.openmoko.org/build Oct 15 20:37:06 hm for package update the same procedure? Oct 15 20:37:07 doh Oct 15 20:37:10 * pb_ stabs e2fsprogs Oct 15 20:37:12 as for mtn? Oct 15 20:37:23 woglinde: maybe Oct 15 20:37:23 NOTE: Applying patch 'no-hardlinks.patch' (/home/pb/oe/org.openembedded.dev/packages/e2fsprogs/e2fsprogs/./no-hardlinks.patch) Oct 15 20:37:23 NOTE: Task failed: Error: /home/pb/oe/org.openembedded.dev/packages/e2fsprogs/e2fsprogs/./no-hardlinks.patch not found. Oct 15 20:37:23 NOTE: package e2fsprogs-1.41.2-r0: task do_patch: failed Oct 15 20:37:25 hehe Oct 15 20:37:29 git-mv cp git-add? Oct 15 20:37:43 aeh Oct 15 20:37:50 pb thats why I set it to -1 Oct 15 20:37:55 I guess Oct 15 20:38:17 woglinde: ah. what version works? Oct 15 20:38:39 38 Oct 15 20:38:47 hm, but apparently that's too old for hal Oct 15 20:39:09 okay I will fix it Oct 15 20:39:10 mom Oct 15 20:39:28 oh, wait, maybe it's e2fsprogs-libs that I need anyway Oct 15 20:39:30 let me try that one Oct 15 20:39:47 hehe Oct 15 20:39:50 thats Oct 15 20:39:56 up to date nearly Oct 15 20:40:00 and compile Oct 15 20:40:03 I tested it Oct 15 20:40:05 haha Oct 15 20:40:07 very good Oct 15 20:40:07 and fixed Oct 15 20:40:10 MKINSTALLDIR Oct 15 20:40:36 hm, no, that didn't ship libvolume_id Oct 15 20:40:43 pb its come from udev Oct 15 20:40:51 thats what my debian says Oct 15 20:41:06 lol Oct 15 20:41:13 oh, seems mwester fed me some duff info there Oct 15 20:41:19 hm libvolume-id0 source udev Oct 15 20:41:23 drat Oct 15 20:41:27 I posted a few minutes ago Oct 15 20:41:33 * mwester is ashamed Oct 15 20:41:40 Yes, it is udev Oct 15 20:41:55 okay :-) Oct 15 20:41:58 * pb_ try new udev Oct 15 20:42:36 woglinde: oh yes, sorry, I didn't notice you said that already Oct 15 20:42:49 I guess I had e2fsprogs tunnel vision Oct 15 20:43:47 okay, excellent. udev_124 has libvolume_id 0.85 Oct 15 20:43:52 surely that must be good enough for hal Oct 15 20:44:09 * pb_ stabs hal one more time Oct 15 20:46:28 hal needs inherit gettext anyway Oct 15 20:46:34 doh Oct 15 20:46:54 well, that doesn't bother me, I laugh at internationalisation Oct 15 20:47:06 at least as far as my mythfront-image is concerned :-} Oct 15 20:47:37 pb_ I dont want the locales packages Oct 15 20:47:44 when I set USE_NLS=no Oct 15 20:48:05 yah, fair enough Oct 15 20:49:13 damnit my wireless laser mice is out of battery again Oct 15 20:49:51 oh dear, presumably that leaves you vulnerable to all kinds of alien attack Oct 15 20:50:05 he yes Oct 15 20:50:11 where is my light sabre Oct 15 20:50:40 I think zecke has it. Oct 15 20:51:23 you have it, but I use my force to make you not see it anymore Oct 15 20:52:53 nono its masked bomb Oct 15 20:55:03 hm I should update gmp Oct 15 20:55:24 'morning Oct 15 20:56:47 hi ant Oct 15 21:04:37 hi ant Oct 15 21:06:18 hey pb_ Oct 15 21:06:38 hm how is git diff working? Oct 15 21:13:22 03Koen Kooi  07holger/hash-style * r201e05c72b 10openembedded.dev.git/packages/linux/linux.inc: linux.inc: add switch to turn of OABI compat for arm kernels Oct 15 21:13:24 03Michael 'Mickey' Lauer  07holger/hash-style * r18e31101a8 10openembedded.dev.git/packages/mozilla/firefox.inc: firefox.inc: fix syntax Oct 15 21:13:24 03Koen Kooi  07holger/hash-style * r0f14c351dd 10openembedded.dev.git/conf/distro/include/angstrom.inc: angstrom: turn of OABI compat in EABI kernels Oct 15 21:13:25 03Michael 'Mickey' Lauer  07holger/hash-style * r876091d0f3 10openembedded.dev.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Oct 15 21:13:27 03Koen Kooi  07holger/hash-style * r6840b8c693 10openembedded.dev.git/packages/linux/linux-rp.inc: linux-rp.inc: add switch to disable OABI compat like linux.inc has Oct 15 21:13:30 03Koen Kooi  07holger/hash-style * ra3125e86b1 10openembedded.dev.git/conf/checksums.ini: checksums: add more checksums, sort it Oct 15 21:13:33 03Philip Balister  07holger/hash-style * r4a9b373df5 10openembedded.dev.git/conf/checksums.ini: libxml2-2.7.2.tar.gz : Add entry in conf/checksums.ini Oct 15 21:13:36 03Philip Balister  07holger/hash-style * re765346792 10openembedded.dev.git/conf/checksums.ini: numpy-1.1.1.tar.gz : Update checksums.ini Oct 15 21:13:43 03Koen Kooi  07holger/hash-style * r20805ba4b3 10openembedded.dev.git/conf/checksums.ini: checksums.ini: add even more checksums Oct 15 21:13:46 03I-Fan, Chen  07holger/hash-style * rffdec57267 10openembedded.dev.git/packages/connman/connman_git.bb: Oct 15 21:13:49 [connman] modify the git server from moblin.org to git.moblin.org Oct 15 21:13:51 thanks Dietmar Friede info the setting of servers changed. Oct 15 21:13:53 03Holger Hans Peter Freyther  07holger/hash-style * rc438f9d0b3 10openembedded.dev.git/packages/libgdbus/libgdbus_git.bb: Oct 15 21:13:56 [naming policy] If we build a specific revision from a SCM have a "r" after the SCM name Oct 15 21:13:58 Clean up PV naming and make it consistent. if we build a revision from a Oct 15 21:14:00 scm system then have svnr/gitr in the package name. Oct 15 21:14:02 03Holger Hans Peter Freyther  07holger/hash-style * r20e32f99b6 10openembedded.dev.git/packages/hal/ (hal.inc hal_0.5.9.bb): Oct 15 21:14:11 [hal] Move almost everything into the hal.inc file Oct 15 21:14:13 hal_0.5.9.bb is the only user of hal.inc and we can safely move everything Oct 15 21:14:15 into the common include file. Oct 15 21:14:17 03Holger Hans Peter Freyther  07holger/hash-style * rbabeb73693 10openembedded.dev.git/packages/hal/hal_0.5.9.bb: [hal] Remove commented out things we have a SCM to recover it Oct 15 21:14:22 03Michael 'Mickey' Lauer  07holger/hash-style * r7a4455feaa 10openembedded.dev.git/packages/tasks/task-fonts-truetype.bb: add task-fonts-truetype Oct 15 21:14:25 03Marcin Juszkiewicz  07holger/hash-style * r83955fe2a5 10openembedded.dev.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Oct 15 21:14:26 woglinde: it gives you the diff between any two commits or trees etc. Oct 15 21:14:28 03Michael 'Mickey' Lauer  07holger/hash-style * r76640097cf 10openembedded.dev.git/packages/ttf-fonts/ttf.inc: ttf.inc: INHIBIT_DEFAULT_DEPS Oct 15 21:14:35 03Holger Hans Peter Freyther  07holger/hash-style * rd07c292757 10openembedded.dev.git/conf/ (3 files in 2 dirs): Oct 15 21:14:38 [commit] Allow the distribution to set the hash-style and enable it for angstrom Oct 15 21:14:40 With binutils >= 2.17 we can use --hash-style=both and this will work Oct 15 21:14:42 with all versions of glibc. glibc >= 2.5 can take advantage of the new hash Oct 15 21:14:44 style and should give better application startup. Oct 15 21:14:48 (39 lines omitted) Oct 15 21:14:54 hm all in hash-style Oct 15 21:15:44 woglinde: you can use the output as a patch. so it's unified diff by default I think. Oct 15 21:16:17 kensan hm git diff . is not doing the same as monotone or svn Oct 15 21:16:19 woglinde: hash or branch-name or all the other ways you can reference these objects in git. Oct 15 21:16:38 it showed me nothing Oct 15 21:16:56 woglinde: by default it shows you the diff between the index and the head. Oct 15 21:17:14 woglinde: git-show to see a revision Oct 15 21:17:23 woglinde: so it should show you the things you are about to commit. Oct 15 21:17:47 woglinde: well actually modified not necessarly about to commit... Oct 15 21:22:24 hm.. dillo 2.0 released Oct 15 21:23:04 hms Oct 15 21:23:45 woglinde: are you going to be at the uni tomorrow? Oct 15 21:25:39 zecke nope Oct 15 21:30:50 "Contributed by Philip Blundell , 2001." Oct 15 21:30:58 master :) Oct 15 21:31:46 heh, where did you find that? Oct 15 21:33:08 glibc Oct 15 21:33:34 thread stuff Oct 15 21:33:37 for arm Oct 15 21:33:39 I guess Oct 15 21:33:39 ah yes Oct 15 21:35:23 drat, stupid cat wants to rub her head on my keyboard again Oct 15 21:35:28 * pb_ fights the beast off Oct 15 21:35:42 lol Oct 15 21:35:42 wifes cat? Oct 15 21:36:29 yeah, a stray that she adopted Oct 15 21:37:09 I'm quite fond of her (the cat) but she can be annoying Oct 15 21:37:19 not the one that housed in your car? Oct 15 21:37:36 heh, no, that one belongs to our neighbours Oct 15 21:37:37 make her happy (both) otherwise they get rid of you :) Oct 15 21:38:56 the worst thing is when I try to use my laptop, she keeps putting her front paws on the keyboard, licking my hands, and sticking her tail in my face. Oct 15 21:38:59 very annoying Oct 15 21:39:04 anyway, heh Oct 15 21:39:11 hi everybody Oct 15 21:39:19 pb_: cat? Oct 15 21:39:23 yah Oct 15 21:40:04 hehe.. Oct 15 21:40:04 hi thesing Oct 15 21:40:15 I had the same problem some minutes before ;) Oct 15 21:40:38 my wife wants a cat too.. :) Oct 15 21:41:15 I like dogs better. Oct 15 21:41:16 pb_, http://fifisramblings.blogspot.com/2008/04/cat-is-not-helping.html Oct 15 21:44:47 sadly, he is in the cat hospital for failing kidneys, again :( Oct 15 21:48:04 Crofton: how old is the cat? Oct 15 21:48:35 15 or so Oct 15 21:49:54 wow. my last cat was 18. Oct 15 21:50:35 I don't think he'll make 18 Oct 15 21:50:46 but he might make another year Oct 15 21:50:49 just depends Oct 15 21:51:01 wow is the glibc code obfusicated Oct 15 21:52:53 Crofton: you never know. I thought for five years that my dog would not make another year. Oct 15 21:54:02 yeah Oct 15 21:54:23 we have another cat that had anemia issues for a few years, but she has been fine since Oct 15 21:57:26 My two cats are quite young 2 years and 4 months (mother and daughter). Its fun watching them play. Oct 15 21:58:07 yeah Oct 15 21:58:15 cats are always funny Oct 15 21:59:34 * rschuster has two cats too, siblings which are inseparable Oct 15 22:02:21 http://rafb.net/p/4whOlf84.html I have forward ported a old patch Oct 15 22:02:37 03woglinde  07org.openembedded.dev * raaed254a21 10openembedded.dev.git/ (5 files in 3 dirs): Oct 15 22:02:37 gmp: update to version 4.2.4 Oct 15 22:02:37 * noexecstack and gnuinline patch were merged upstream Oct 15 22:02:37 * configure patch was refreshed Oct 15 22:04:13 args gcc-4.3.2 has no checksum Oct 15 22:10:15 zecke_: I'm not sure that even FAST_MULT is an important hwcap nowadays Oct 15 22:10:25 basically every arm cpu you are likely to use has that capability. Oct 15 22:11:32 iirc, arm610/arm710 were the only processors built without it, and the chances of encountering one of those in oe is fairly low (not negligible, since arm7500 is based on arm710, but still fairly low) Oct 15 22:11:57 03woglinde  07org.openembedded.dev * rcc05e851d8 10openembedded.dev.git/conf/checksums.ini: checksum.ini: add checksums for gcc-4.3.2 Oct 15 22:13:02 regarding /etc/ld.so.nohwcap, remember that checking for this file involves an extra syscall and dentry lookup, so it isn't completely free. Oct 15 22:13:27 if you plan on shipping that file as default in your DISTRO then you would probably be better off just patching glibc to disable hwcaps altogether. Oct 15 22:14:25 pb_: right, I was just starting to read this hwcaps code Oct 15 22:14:46 pb_: and I didn't even look at RPs patch :} Oct 15 22:16:01 zecke_: are you admin for git? Oct 15 22:16:46 zecke is masters of the universe Oct 15 22:17:13 no Oct 15 22:17:18 and I'm tired :) Oct 15 22:18:10 ok. who then? because the gitweb urls in the commit mails are wrong. Oct 15 22:19:09 thesing: still? Oct 15 22:20:46 yes: http://gitweb.openembedded.net/?p=openembedded.dev.git&a=... instead of http://gitweb.openembedded.net/?p=openembedded.git;a=... Oct 15 22:21:30 thesing: RP said he fixed that... I can take a look tomorrow Oct 15 22:21:51 ok. Thanks. Oct 15 22:52:56 03Jeremy Lainé  07org.openembedded.dev * r4268f1db1f 10openembedded.git/conf/checksums.ini: dtc-v1.1.0.tgz: update checksums.ini Oct 15 22:57:18 Question, assume the target machine is standard pc structured, is it possible to build a Virtual Machine based on OE? Oct 15 22:57:22 he Oct 15 22:57:33 uclibc hash HASH_GNU support Oct 15 22:57:40 they even check the binutils for it Oct 15 22:58:20 hubar if you mean that after bitbake bllalal a vmware image is build? Oct 15 22:59:47 but it is not activated in our uclibc configs yet Oct 15 22:59:52 so I am going to bed Oct 15 22:59:53 nite Oct 15 23:00:21 woglinde: well Oct 15 23:00:38 woglinde: doesn't matter what process I have to go through. Oct 15 23:00:58 woglinde: As long as I can create a VM image. :) Oct 15 23:01:32 hubar no is not support yet Oct 15 23:01:45 Hubar[TD]: you can build ext2 images. Iit should be possible to create an VM image from that. (with external tools) Oct 15 23:01:51 night all Oct 15 23:02:29 yes thesing is right Oct 15 23:02:46 but I do not know in which shape it is Oct 15 23:02:49 The thing is that I need to create a VM image with OE, so we can test the thing to see if it is alright. Oct 15 23:02:55 set machine in your local.conf to vmware Oct 15 23:03:00 and just try Oct 15 23:03:23 okay bye Oct 15 23:05:16 03Jeremy Lainé  07org.openembedded.dev * r04c3feecc8 10openembedded.git/conf/checksums.ini: locomo_kbd_tweak-r1a.patch: update checksums.ini Oct 15 23:34:53 03Jeremy Lainé  07org.openembedded.dev * rd211081be3 10openembedded.git/ (9 files in 3 dirs): Oct 15 23:34:53 linux-*.bb: use KERNELORG_MIRROR consistently (fixes checksum errors) Oct 15 23:34:53 * remove trailing slash in KERNELORG_MIRROR definition Oct 15 23:34:53 * add slash after KERNELORG_MIRROR in *.bb where needed Oct 15 23:39:57 03Jeremy Lainé  07org.openembedded.dev * rc6786fbc46 10openembedded.git/conf/checksums.ini: checksums.ini: remove duplicate slashes in kernel.org urls Oct 15 23:56:32 03Jeremy Lainé  07org.openembedded.dev * r35a6bddb9a 10openembedded.git/conf/checksums.ini: checksums.ini: add entry for squid-2.6.STABLE14.tar.bz2 Oct 16 00:10:08 03Jeremy Lainé  07org.openembedded.dev * r8b195623bb 10openembedded.git/ (2 files in 2 dirs): offlineimap_4.0.14.bb: fix duplicate slash in URL Oct 16 02:01:47 03Jeremy Lainé  07org.openembedded.dev * re8848c25d6 10openembedded.git/ (5 files in 3 dirs): Oct 16 02:01:47 uicmoc4-(native|sdk): update to version 4.4.3 Oct 16 02:01:47 * drop qapplication.diff, included upstream Oct 16 02:01:47 * update URL in uicmoc4-native.inc, qtopia is gone Oct 16 02:15:50 03Jeremy Lainé  07org.openembedded.dev * rc6975e41c8 10openembedded.git/packages/qmake/ (qmake2-native.inc qmake2-native_2.10a.bb qmake2-sdk_2.10a.bb): qmake2-(native|sdk): update to qt 4.4.3 Oct 16 02:58:58 03Jeremy Lainé  07org.openembedded.dev * r040a1dda0c 10openembedded.git/packages/qt4/ (12 files in 2 dirs): qt-embedded: update to version 4.4.3 **** ENDING LOGGING AT Thu Oct 16 02:59:58 2008