**** BEGIN LOGGING AT Thu May 24 02:59:57 2007 May 24 03:23:54 , do you can help me with bitbake and oe? May 24 03:24:22 I have a simple question... May 24 03:32:47 ask the question, someone may answer May 24 03:35:31 I changed config of machine. How to build once more opie-image with new machine's config? May 24 03:36:31 I try to remove tmd dir... but it's too clumsy May 24 03:36:55 bitbake thingIwantToBuild -c rebuild May 24 03:36:58 should do it May 24 03:38:57 hm... it does not work May 24 03:41:04 dunno then May 24 03:42:42 I move MACHINE_EXTRA_RRECOMMENDS and module_autoload to bb kernel package... may be it works.... May 24 04:28:38 I need help still... May 24 04:31:52 I define concretely: how to rebuild rootfs? May 24 05:16:38 ok. for rebuilding rootfs is necessary remove rootfs May 24 06:18:47 good morning May 24 06:34:00 , do you can help in automatically installing kernel modules ? May 24 06:36:58 step2back : its easy just add RDEPENDS =+ " kernel-module-blha blah" May 24 06:37:39 I use MACHINE_EXTRA_DEPENDS, is it wrong? May 24 06:38:26 DEPENDS is for bitbake (so it can build the packages) RDEPENDS is for the ipkg system May 24 06:40:34 I should rename MACHINE_EXTRA_DEPENDS to RDEPENDS, should not i? May 24 06:43:54 just add MACHINE_EXTRA_RDEPENDS May 24 06:44:16 , can you get my link to package for simple? May 24 06:44:49 what do you mean ? May 24 06:47:05 please loot at http://cmh.hackndev.com/pastebin/44/ May 24 06:47:28 What should I correct? May 24 06:49:51 , ping May 24 06:50:22 sorry wend to get some coffee :) May 24 06:52:02 oh... I'm sorry. May 24 06:52:31 step2back : change that to MACHINE_EXTRA_RDEPENDS and you should be ok May 24 06:54:12 ok! Thanks! May 24 06:54:47 * step2back start "bitbake opie-image" and waiting for autoloading modules May 24 06:55:02 * * OE Bug has been REOPENED () by sybren@stuvel.eu May 24 06:55:04 * * [Bug 2315] Asterisk misses speex module May 24 06:55:06 * * http://bugs.openembedded.org/show_bug.cgi?id= May 24 07:00:08 good morning May 24 07:03:41 Berbardo Hi ! May 24 07:03:50 hi steliosk May 24 07:04:18 summatusmentis: your drive is mounted either with the 'noexec' or 'user' (which implies noexec) flag, hence the bad interpreter error May 24 07:04:22 good morning all May 24 07:05:27 * Bernardo is trying to find why *.ntp.org won't resolve here in pt... :( May 24 07:05:39 hi koen May 24 07:07:50 * koen wonders why some people insist that a .la file is a library May 24 07:07:54 lol May 24 07:08:03 hi all May 24 07:08:26 maybe they never checked was is the format and the content of a .la file? May 24 07:08:40 Hi May 24 07:09:38 Bernardo: and after being told it's a libtool archive: " I appreciate your insight, but in this case, the libraries belong in the main package." May 24 07:12:36 koen: I'm used to that kind of behaviour every day, unfortunately. One has to try and be, even when we feel the urg to bash the other guy's head in... May 24 07:14:48 hi psokolovsky May 24 07:15:25 anyone can resolve pool.ntp.org, or is it only the pt top dns that is failing? May 24 07:15:26 psokolovsky hi May 24 07:15:42 koen : could you check #2315 May 24 07:16:22 koen : i am not sure how i should answer..... May 24 07:17:18 Bernardo: my dns server resolve fine pool.ntp.org May 24 07:18:18 ok, probably just another bug of the pt top dns server (ns.dns.pt) from which all pt servers inherit May 24 07:19:05 Bernardo: btw, my dns is 80.58.0.33 May 24 07:19:34 strange... even though I have scummvm in org.openembedded.dev/packages, I am getting "No buildable providers available for required build target scummvm" since I updated ten minutes ago May 24 07:19:39 nauj27: thanks May 24 07:19:40 Hi! May 24 07:28:18 03koen 07org.oe.dev * racbf904c... 10/ (5 files in 2 dirs): scummvm: catch up with libmpeg2 removal May 24 07:28:28 Bernardo: that should fix it May 24 07:30:57 koen: thanks - I was already thinking what had I done wrong this time... :) May 24 07:32:04 bbl May 24 07:32:22 * Bernardo is away: Ausente por agora. **** BEGIN LOGGING AT Thu May 24 07:35:47 2007 May 24 08:10:22 * XorA sighs I hate to say this, but looking at bugtracker we are beginning to look like a bunch of retards May 24 08:12:35 XorA: :-( May 24 08:12:49 XorA: Just shear number of bugs? May 24 08:12:55 morning all btw :) May 24 08:13:42 RP: one dev tells someone one thing, they post a patch, which then gets pulled apart for the very things that dev told them to do May 24 08:14:28 Length: 4,514,100 (4.3M) [application/x-redhat-package-manager] May 24 08:14:37 sf.net has some messed up mimetypes May 24 08:14:47 the file is http://puzzle.dl.sourceforge.net/scummvm/scummvm-0.9.1.tar.bz2 May 24 08:15:05 XorA: :/ May 24 08:15:52 * XorA begins to think a mentoring system for new recruits might be needed May 24 08:16:17 RP: I suppose its the problem that each of us has a slightly different style/method May 24 08:19:09 steliosk: http://www.angstrom-distribution.org/repo/?action=details&pnm=asterisk May 24 08:19:25 putting .la files in the main package has nothing to do with style May 24 08:19:50 and noone here writes empty do_install() methods supplemented with do_install_append May 24 08:19:54 koen: I was meaning the fakeroot stuff and the no-chmod patch May 24 08:20:22 the no-chmod patch is utter crap May 24 08:20:34 koen: its not, its a valid solution May 24 08:20:40 it breaks user installs and is unneeded for root installs May 24 08:22:47 koen: bullshit May 24 08:24:05 are we talking about the same patch? May 24 08:24:10 I'm talking about the May 24 08:24:13 -m 644 May 24 08:24:17 +m 444 May 24 08:24:22 patch May 24 08:24:34 its the removing -o root May 24 08:24:39 that is the importat bit May 24 08:24:42 koen : did you uploaded now or was it there ? May 24 08:25:03 steliosk: I ran the sort and index scripts :) May 24 08:25:12 (still need to test your version) May 24 08:25:39 koen: what it actually needs and what I said on channel yesterday was it need the -o root removing without changing the -m May 24 08:25:56 koen: and that avoids any pissing around with fakeroot May 24 08:27:26 XorA: Yes, but we should try and have some standardisation. What kind of things do you mean? May 24 08:28:35 XorA: Ah, I see May 24 08:31:33 03koen 07org.oe.dev * r7c39b426... 10/ (1 packages/scummvm/scummvm_0.9.1.bb): scummvm: add 0.9.1 with default_preference = -1 since the sh3 patch has been dropped May 24 08:50:48 likewise: looks like your libcap commit didn't include the actual patch May 24 08:52:02 * * OE Bug 2315 has been RESOLVED (FIXED) by skoroneos@digital-opsis.com May 24 08:52:04 * * [Bug 2315] Asterisk misses speex module May 24 08:52:06 * * http://bugs.openembedded.org/show_bug.cgi?id=2315 May 24 08:56:09 v8jlene: sorry, will recreate the patch and upload it. May 24 08:56:29 v8jlene: thanks for the heads-up, sorry for breaking your build May 24 08:56:57 likewise: No problem, I'm not actually building it... just noticed the problem in the commit email. May 24 09:10:04 03likewise 07org.oe.dev * r6f9abba2... 10/ (1 packages/libcap/libcap-1.10/syscall.patch): libcap: Add missing syscall.patch. Thanks v8jlene for spotting. May 24 09:36:39 koen: where can I find a ready to download angstrom neo image? May 24 09:37:09 zecke: behind the "download" link on the angstrom website May 24 09:38:37 zecke: upgrading your phone? May 24 09:41:07 I would like to... May 24 09:41:23 zecke: you will be amazed at the speed of Angstrom May 24 09:41:28 and my angstrom-uclibc build still needs firefighting on libiconv May 24 09:42:03 libiconv raises my bloodpressure May 24 09:42:13 * koen eats a banana to raise bloodsugar May 24 09:42:43 koen: this fucking *-link.m4 hardcodes/forces rpath on, the code to disable rpath in this very .m4 is broken May 24 09:42:56 and on autoreconf I get failures :} May 24 09:43:10 * koen hands zecke some bananavla May 24 09:44:29 zecke: one caveat: the openmoko developers moved some files around between packages, so you probably have to use -force-overwrite :( May 24 09:45:09 koen: well, I would reflash now... May 24 09:46:02 the "one hour parsing time" reports scare me a bit May 24 09:46:47 where? openmoko reports? this is the nature of srcdate=now May 24 09:46:57 yes May 24 09:47:08 that makes the patch pretty unmergable as-is May 24 09:47:33 and I would really like to get rid of SRCDATE and use SRCREV instead in package-names May 24 09:47:33 I don't agree. It is up to people to decide to use or not use srcdate=now May 24 09:50:58 koen: I think we should split out the kernel patches and do them properly in OE as well, they way they do them breaks the svn fetcher May 24 09:51:24 koen: and why fetch all the sources, then tar them, then untar them all in same operation May 24 09:51:29 XorA: yeah, but I was too lazy to do that May 24 09:51:47 that will also protect us from the daily breakage May 24 09:52:07 can pull the ASoC stuff straight from ALSA now :-) May 24 09:54:52 morning May 24 09:55:06 hey hrw May 24 10:02:29 Anyone seen this before? "...staging/i686-linux/bin/install-sh: ./-m755 does not exist." May 24 10:02:37 yes May 24 10:02:53 replace '-m755' with '-m 755' in the offending package May 24 10:03:20 oe's install.sh doesn't like the option without the space May 24 10:03:30 right, I found out this too... but this worked before. May 24 10:04:02 only by luck May 24 10:04:14 I sent Philippe a ton of patches to fix it for gpe about a year ago :) May 24 10:04:38 oh May 24 10:04:52 he applied all of them May 24 10:05:08 s/gpe/gpe packages I build/ May 24 10:06:43 hum well, wouldn't it better to fix the install script? we will run into this over and over... May 24 10:07:33 install allows this too and I can't see a reason why it should be invalid. May 24 10:10:00 koen: Thanks for fixing scummvm. I was sure I had rgrep'd for libmpeg before. Not sure how I missed it. May 24 10:10:42 hi Laibsch May 24 10:10:53 Laibsch: openembedded-essential is handled by the same guy which handle bitbake for debian iirc May 24 10:12:01 hrw: Is there an official package now? May 24 10:12:13 I still get it from openembedded.org/dl May 24 10:12:33 Laibsch: hm. mail him please about status May 24 10:12:38 Hi florian. How is it going. Settled the X31 issues? May 24 10:13:19 hrw: IOW, no more releases of oe-essential from oe itself? May 24 10:13:41 hrw: nuttin in debian itself, yet May 24 10:14:25 Laibsch: yes, the third one I goe was ok. :-) May 24 10:14:46 Laibsch: I do not current status of it May 24 10:21:17 florian: Sounds like a lot of trouble and uncertainty May 24 10:23:03 Laibsch: yes... but finally I got a really nice device quite cheap May 24 10:27:49 whats an X31 May 24 10:28:26 ibm thinkpad May 24 10:28:30 Thinkpad X31 May 24 10:28:37 -> thinkwiki.org May 24 10:28:42 Nice machine May 24 10:28:48 12" beast May 24 10:29:00 just like any X series Thinkpad May 24 10:29:11 * Laibsch pads the X24 on his desktop May 24 10:29:16 :-) May 24 10:29:17 hi, all! May 24 10:29:25 hi step2back May 24 10:29:38 * XorA wants a 12inch laptop with 1600x1200 screen, 4 cores, and a decent GeForce, where do I get one :-) May 24 10:29:43 How to build angstrom with all locales? May 24 10:29:57 bitbake angstrom-x11-image May 24 10:30:26 # If ENABLE_BINARY_LOCALE_GENERATION is set to "1", you can limit locales ... May 24 10:30:39 I don't want to limit, :) May 24 10:30:44 step2back: then dont May 24 10:30:51 step2back: by default it builds all May 24 10:31:35 But they don't include to image, :( May 24 10:34:22 If I set ENABLE_BINARY_LOCALE_GENERATION to 0, than I get all locales inside image? May 24 10:34:41 step2back: no, you get locale sources which have to generated on machine May 24 10:35:10 It's not I want. May 24 10:37:03 I can write list of all locales by hand in GLIBC_GENERATE_LOCALES = "", but it's very hard. ;) May 24 10:38:00 set IMAGE_LINGUAS variable May 24 10:38:15 conf/distro/angstrom-2007.1.conf is a good place to look May 24 10:41:36 ~curse udev May 24 10:41:51 May the fleas of a thousand camels infest your most sensitive regions, udev ! May 24 10:41:53 I connect hdd with 3 partitions and only sdh is created May 24 10:44:10 hrw: sdh, you confused it with too many drives :-) May 24 10:45:32 XorA: I had sdk yesterday ;D May 24 10:45:57 hrw: I have trouble with getting a working machine and now you just flaunt it May 24 10:46:01 ~prod hrw May 24 10:46:15 * ibot zaps hrw with a high voltage cattle prod May 24 10:47:05 yo mmp_ May 24 10:47:38 XorA: hello:) May 24 10:48:41 hello there May 24 10:52:50 can someone tell me howto bitbake obexpush, I am trying to bitbake opie-image and obexpush complains, that there is no libusb. I found that problem also in bug1113, but I did not find a solution, which I could understand May 24 10:56:02 * * OE Bug 2372 has been created by pinguozz@netsons.org May 24 10:56:04 * * totem package fail when gpe-image building for hx4700 May 24 10:56:06 * * http://bugs.openembedded.org/show_bug.cgi?id=2372 May 24 10:58:02 anyone with zlib knowledge around? May 24 10:58:05 !oebug 1113 May 24 10:58:06 * * Bug 1113, Status: ASSIGNED, Created: 2006-06-20 11:30 May 24 10:58:07 * * loft(AT)nc.rr.com: Looks like a libusb dep in obexpush-1.0 .0 is missing for obexclient May 24 10:58:08 * * http://bugs.openembedded.org/show_bug.cgi?id=1113 May 24 11:02:01 * * OE Bug 2372 has been RESOLVED (WONTFIX) by koen@dominion.kabel.utwente.nl May 24 11:02:03 * * [Bug 2372] totem package fail when gpe-image building May 24 11:02:05 * * http://bugs.openembedded.org/show_bug.cgi?id=2372 May 24 11:02:41 what does "needed to -Wl,-rpath,${STAGING_LIBDIR} to make libusb find its dependencies" mean? Mickey wrote this, but I do not know what to do May 24 11:03:00 shouldn't that be rpath-link? May 24 11:10:33 where should I add -rpath option or -rpath-link option? May 24 11:17:13 ok I got it, just taking newest obexpushXXXX.bb file from dev-branch, thank you May 24 11:21:09 * Crofton hates -rpath ... May 24 11:27:39 * hrw starts to dislike nfs May 24 11:33:18 which package contains the install-sh that goes to staging? May 24 11:34:12 install-native May 24 11:35:00 hrw: thanks May 24 11:46:19 argh.. May 24 11:46:28 fscking nfs May 24 11:47:13 I can mount share but cant create files on it.. May 24 11:47:21 /srv/nfs/sources/ 192.168.1.0/24(rw,sync,all_squash) May 24 11:47:24 rootsquash? May 24 11:48:39 ~lart openslug for old stuff May 24 11:48:40 * ibot readies the nuke launcher and fires some rounds at openslug for old stuff May 24 11:49:22 The Constitution gives every American the inalienable right to make a damn fool of himself. May 24 11:49:38 Hopefully, this quote helps you understand Americans :) May 24 11:49:59 uf.. May 24 11:50:52 from http://www.quotationspage.com/quotes/John_Ciardi May 24 11:51:04 permissions on slug were wrong May 24 12:21:21 rwhitby: I assume you've seen: http://www.regdeveloper.co.uk/2007/05/24/slug_linux_server/ May 24 12:46:09 hi all May 24 12:49:25 copying DL_DIR to NFS take eons May 24 12:54:09 hi hrw, did you have time to build a collie kernel with suspend patches and test it? May 24 12:54:17 nope May 24 12:55:29 but you know that I attached the patch for the bb to bug 2323? May 24 12:56:20 !oebug 2323 May 24 12:56:21 * * Bug 2323, Status: NEW, Created: 2007-05-15 01:45 May 24 12:56:22 * * skliarieh(AT)gmail.com: angstrom collie: no suspend support May 24 12:56:23 * * http://bugs.openembedded.org/show_bug.cgi?id=2323 May 24 12:56:48 thesing: whats left to fix on collie? May 24 12:57:07 audio, usb client... May 24 12:57:33 audio sux :-D May 24 12:58:50 btw, x11 bitbake jffs2 generated image it's about 20mb and do not fit in the rom May 24 12:59:18 I mean, for angstrom collie May 24 12:59:24 so? May 24 12:59:51 collie now is in beta state where console-image is better May 24 12:59:56 XorA: yes is sux. But its the next item on my todo-list. But first I have to get some time ;) May 24 13:00:40 thesing: do you already have a ticket for the LinuxTag? May 24 13:00:45 yep ^_^U May 24 13:01:19 zecke: LinuxTag? When and where? May 24 13:01:35 where: Berlin (Messe) when: soon May 24 13:02:58 03koen 07org.oe.dev * ra73ff29a... 10/ (3 files in 3 dirs): xserver-kdrive-common: fix typo May 24 13:03:19 zecke: no. I don't have one yet. But this sounds interesting. And it's not far away ;) May 24 13:03:54 thesing: I will send you an invitation once I'm home May 24 13:04:31 zecke: Thanks. May 24 13:08:50 thesing: Im kidding because I work for Wolfson and am sick of audio :-) May 24 13:11:09 XorA: So you are familiar to asoc? May 24 13:11:15 thesing: yes May 24 13:13:04 Let's say there is a package classpath and one classpath-gtk where the latter is supposedy targetted to gtk GUI. All the $PN stuff now breaks for classpath-gtk and has to be reset explicitly. I was wondering if doing a PN = "classpath" might have any bad sideeffects. It seems to be the easier solution. May 24 13:16:31 XorA: Great! AFAIK collie has a ucb1200 connected to the sa1100 mcp port (buzzer and micro )and a pcm1741 to the sa1100 ssp (headphone). So I would have to write codec drivers for pcm1741 and ucb1200, a machine driver for the sa1100 and a dai driver for mcp. Rigth? May 24 13:17:12 thesing: yes May 24 13:18:01 XorA: Any advice where to start? May 24 13:18:25 thesing: stealig as much code as possible :-) May 24 13:18:39 thesing: Id start with machine driver, then do dai, then attach the codecs May 24 13:18:55 thesing: as to test the dai you can always dummy the codecs May 24 13:20:23 XorA: Thanks. I will try it this way. May 24 13:21:31 How can I debug what my FILES_${PN} ends up being? May 24 13:22:05 I have cruft remaining which should have been picked up by the default FILES... May 24 13:24:28 thesing: ucb and mcp code is in mainline May 24 13:24:47 thesing: also AFAIK the collie has something like a ucb1200 but not from phillips May 24 13:25:36 zecke: exactly May 24 13:28:17 zecke: I know. I didn't discover any relevant differences in the datasheet yet, but I did only look at the ADC. May 24 13:28:59 /amsg bye all! ;) May 24 13:48:22 http://book.efika.org/ May 24 13:56:31 CM: no I hadn't - thanks for the pointer May 24 13:58:27 rwhitby: I really feel like I have to get one too now. It was a really positive review :) May 24 14:00:08 I've added it to our articles page at http://www.nslu2-linux.org/wiki/Main/Articles May 24 14:01:08 Ah, I'll remember to update that wiki if I find something related May 24 14:06:15 rwhitby: can recent OE be used to build packages for openslug 3.10? May 24 14:07:07 hrw: dunno - can't remember if there has been an incompatible toolchain change or not. May 24 14:07:29 just run slugos/be head :-) May 24 14:07:44 or checkout the oe snapshot from svn ;) May 24 14:07:46 rwhitby: thats the option too May 24 14:07:55 NAiL: no! May 24 14:08:06 why not? May 24 14:08:09 NAiL: I am tired enough with openmoko obsolete snapshot May 24 14:08:21 aha May 24 14:08:43 hrw: this snapshot is 3.10 source archive only - it's not meant to be used for current building. May 24 14:08:43 NAiL: I do not plan to limit myself with 'what is known to work with openslug' - feeds already contain it May 24 14:09:05 for example I need hid2hci for my bt dongle May 24 14:09:26 (i.e. we make an svn snapshot as an easy way to manage a set of stable feeds which are guaranteed to continue to work with that firmware release) May 24 14:09:38 and openslug 3.10 does not have it. and it use 2.25 bluez... May 24 14:09:53 hrw: right - you should use slugos head May 24 14:10:01 I am building slugos/be now May 24 14:10:13 rwhitby: does it work? May 24 14:10:17 sure May 24 14:10:26 (it's built on OE, isn't it?) May 24 14:10:53 * mwester harumphs May 24 14:11:19 rwhitby: if something builds then it does not mean that it works ;D May 24 14:11:30 hrw: As of about two weeks ago, it built and booted ok on my test NSLU2. May 24 14:11:59 NOTE: multiple providers are available for virtual/libintl (glibc, gettext); May 24 14:11:59 NOTE: consider defining PREFERRED_PROVIDER_virtual/libintl May 24 14:11:59 NOTE: multiple providers are available for virtual/libiconv (libiconv, glibc); May 24 14:11:59 NOTE: consider defining PREFERRED_PROVIDER_virtual/libiconv May 24 14:12:06 thats with slugos distro... May 24 14:12:27 can you guys define those? May 24 14:12:57 but non-determinism adds spice to life! May 24 14:13:03 hi, all! May 24 14:13:09 sure - you can add them if you like in slugos.inc (you'd probably know better what to set them to) May 24 14:13:28 anybody knows of some debian .dsc file parser in Perl? May 24 14:13:36 or python? May 24 14:14:17 hrw: seriously, if it's something like that which is an obvious general OE thing and not slugos specific, feel free to just fix it May 24 14:15:24 * mwester can't find which providers were actually used in his recent SlugOS builds... :( May 24 14:15:52 rwhitby: slugos.inc is used by ucslung too May 24 14:15:55 hi May 24 14:16:42 hrw: hmm. officially ucslugc is no longer supported, but unofficially some guys are trying to resurrect it. May 24 14:17:19 rwhitby: http://pastebin.ca/507339 is ok? May 24 14:18:15 yeah, that works. do you know what values they would need to have for uclibc? May 24 14:18:43 PREFERRED_PROVIDER_virtual/libiconv = "libiconv" May 24 14:18:43 PREFERRED_PROVIDER_virtual/libintl = "libintl" May 24 14:18:57 this is from generic-uclibc May 24 14:19:23 second one should point to gettext May 24 14:19:54 some people should be forbidding to talk to users as OE representative.... May 24 14:20:12 CoreDump|home: where this time? May 24 14:20:45 ah just a bug report w/ proposed patch. "This is crap" isn't exactly an enlightening answer. May 24 14:24:33 03hrw 07org.oe.dev * rf497aad6... 10/ (1 conf/distro/slugos.conf): slugos: set PREFERRED_PROVIDERS for libiconv/libintl/libc May 24 14:25:27 hrw: thx May 24 14:28:17 np May 24 14:28:37 hrw: wouldn't it be easier to get angstrom running instead of trying to fix slugos? May 24 14:28:46 angstrom runs on my ixp machine quite well May 24 14:29:10 koen: root=/dev/sda1? May 24 14:29:32 no, it's using internal flash May 24 14:29:42 haven't tried the cf slot yet May 24 14:29:52 koen: nslu2 flash is too small for me May 24 14:30:20 koen: angstrom does run, and slugos is not broken May 24 14:33:23 (and yes, as a seasoned OE developer, hrw might be more useful in helping CoreDump|home get altboot support working to the point where Angstrom + altboot can replace the turnup functionality in SlugOS) May 24 14:34:07 * CoreDump|home had to trash his slug buildenv due to temporary space constraints May 24 14:34:17 rwhitby: now I mostly need slug as machine which will work (tftp, nfs) then experiment on it May 24 14:34:21 CoreDump|home: you have an account on nudi ... May 24 14:34:28 I have? May 24 14:35:37 yep May 24 14:36:48 * rwhitby notices that it's midnight, and goes to bed. May 24 14:37:12 gn8! May 24 14:37:57 rwhitby: seems to be in some partial timezone :-) May 24 14:40:13 Australia ftw May 24 14:43:26 has anyone here experience with merging an SVN tree into a monotone DB ( and keep updating changes in SVN to mtn)? May 24 14:43:57 CoreDump|home: good luck :( May 24 14:44:06 heh May 24 14:44:11 CoreDump|home: that was zecke's speciality wasnt it May 24 14:44:23 zecke: ping? May 24 14:44:27 the tool that zecke used is broken now May 24 14:44:33 hmmm too bad May 24 14:44:46 time for some l33t haxoring then =) May 24 14:44:51 svn supports commit hooks does it not? May 24 14:45:05 it should May 24 14:45:10 * CoreDump|home looks at CIA-4 May 24 14:45:24 just need to script up on commit for the diff to be committed to mtn as well May 24 14:45:37 right May 24 14:45:47 30 secs work for a genius like you :-) May 24 14:46:13 morning May 24 14:46:16 * CoreDump|home chuckles May 24 14:46:55 CoreDump|home: yes? May 24 14:47:25 zecke: I heard you had some experiences w/ automating a sync between an SVN tree and a monotone DB? May 24 14:47:37 http://progetti.arstecnica.it/tailor/search?q=monotone&wiki=on&changeset=on&ticket=on May 24 14:48:00 CoreDump|home: let me know if anything works for you. I am also interested May 24 14:48:03 CoreDump|home: check the wiki. I had used tailor + patches for some of your commits May 24 14:48:24 steliosk: will do May 24 14:48:46 zecke: I didn't realise there was information about this topic on the wiki, sorry May 24 14:54:17 CoreDump|home: I don't know if my tailor scripts are there though May 24 14:55:18 CoreDump|home: why do you want svn? May 24 14:55:47 my own repo is in SVN and I kinda like it ;) May 24 14:56:06 its "interface" is very intuitive May 24 14:56:08 CoreDump|home: the working copy operations with OE will be really slow :) May 24 14:56:55 hmm, it is more complicated than I thought May 24 14:57:04 as far as i got with tailor was to do the initial pull, and then it failed May 24 14:57:18 i think it was because the changes in mtn log May 24 14:57:34 but i am not mtn expert... May 24 14:57:52 mtn has a cvssync branch, so you could mtn <> cvs and then use cvs2svn? May 24 14:58:07 that would be fun :) May 24 14:58:23 yuck May 24 14:58:25 =) May 24 14:58:40 probably too much work for a few files May 24 14:59:05 OE overlay ftw =D May 24 15:00:05 with a bit of scripting you might will be able to keep the code in sync, but i don't think that the log/commit messages could be synced easily May 24 15:00:37 SVN -> OE would be very easy to script May 24 15:00:53 I'm moe concerned about SVN <- OE May 24 15:01:50 that should be easy too May 24 15:02:23 but the problem is how do you get the commit comments, May 24 15:02:30 make SVN and monotone use the same working dir, I dont think their status files clash May 24 15:02:45 hi. Since upgrading my OE checkout recently, I have noticed that do_populate_staging() doesn't seem to be run automaticly for a package by simply adding it to DEPENDS. May 24 15:02:52 s May 24 15:03:43 XorA : The problem is that svn adds the .svn files even to the temp files generated by MTN and mtn *really* dislikes that May 24 15:04:06 steliosk: ah thats a pity May 24 15:04:11 :\ May 24 15:04:11 tkp: wow, I'm really out of the loop, I think it is only called do_stage May 24 15:04:19 tkp: (nowadays) May 24 15:05:17 hi. sorry, I know someone just responded... but I'm having trouble with my IRC client and I can't read the response... trying to fix now.. May 24 15:05:48 17:04 < zecke> tkp: wow, I'm really out of the loop, I think it is only called do_stage May 24 15:06:02 zecke: g'day sir May 24 15:06:07 zecke: hi, ok... irc working now :) May 24 15:06:13 zecke: be a pall and check my bugs, or are you busy? May 24 15:06:23 gm Zero_Chaos May 24 15:06:26 zecke: I made all the requested improvements May 24 15:06:32 Crofton: morn' May 24 15:07:08 maybe I could hack MokoMakefile into a HentgesMakefile using the overlay but that would be less than ideal May 24 15:07:09 so, do I have to do something other than add to DEPENDS to ensure do_stage is called at the right time? May 24 15:07:18 Zero_Chaos: good, once I'm home I will comment May 24 15:07:45 tkp: no, just provide do_stage (I don't know if I talk bullshit...) May 24 15:07:53 zecke: if you can now, can you look at it and make sure it is satifactory first? May 24 15:08:08 not now, I'm at work May 24 15:08:16 the package in question is libvx, where ]]] May 24 15:08:27 it got built by adding to DEPENDS in my package May 24 15:08:32 but do_stage was not called May 24 15:08:35 pastebin the entire bb file and I can do a quick look May 24 15:08:39 hmmmmm May 24 15:08:42 tkp: call your method do_stage May 24 15:08:52 explecitly building libxv causes the do_stage to be called... just not via DEPENDS May 24 15:08:58 Anyone with commit access want to look at 4 of my bugs/additions and tell me if they are satisfactory. May 24 15:09:25 tkp: now it is do_stage? what happened to populate_staging? May 24 15:10:02 koen: bluez-libs-3.11 and bluez-utils-3.11 are out. May 24 15:10:05 zecke: sorry. I don't think I have been clear (as usual)... May 24 15:10:27 tkp: well, don't worry I just try to understand May 24 15:10:35 Zero_Chaos, what is an easy bug number? May 24 15:10:59 adding libXV to DEPENDS get it built - but not all stages seem to have been run. do_package_write() & do_populate_staging() did not get run May 24 15:11:01 2361,2368,2369,2370 May 24 15:11:09 Crofton: nothing that complex to be honest. May 24 15:11:16 Crofton: I'm not that good ;-) May 24 15:11:18 although they run if I manually 'bitbake libxv' May 24 15:11:31 tkp: using bitbake -i? May 24 15:12:14 holtmann: 'Add initial version of serial port support.' == hciattach successor? May 24 15:13:33 XorA: any good docs about how to use bitbake -i: I never saw that before May 24 15:13:46 Kati tetio 8a prepei na sou bgazei -> http://rafb.net/p/hQxnMq41.html May 24 15:13:49 tkp: no, but I had wierd shit with bitbake -i May 24 15:14:03 oh I see. no I'm not using -i May 24 15:14:11 sorry guys wrong window May 24 15:14:15 I just ran 'bitbake mypackage' May 24 15:14:45 and it make all the dependencies, but didn't run do_populate_staging() for libxv May 24 15:15:03 bitbake 1.8.2? May 24 15:15:05 I'm sure it is todo with a change in the latest bitbake May 24 15:15:07 koen: yes May 24 15:15:20 try 1.8.3 May 24 15:15:30 koen: ok, thanks May 24 15:17:02 1.8.3 you mean svn? 1.8.2 seems to be the latest release May 24 15:17:16 svn good! May 24 15:18:35 03koen 07org.oe.dev * rc8c229b1... 10/ (4 files in 2 dirs): bluez: update libs and utils to 3.11 May 24 15:20:54 | unwind-pe.c:(.text+0x91bd8): undefined reference to `__fixsfsi' May 24 15:20:54 | collect2: ld returned 1 exit status May 24 15:20:57 fsck May 24 15:21:11 glibc 2.3.5 strikes back again May 24 15:22:04 Zero_Chaos, in 2368, what happens if ARCH isn't ARM? May 24 15:22:35 Crofton: that was my question from yesterday May 24 15:22:40 :) May 24 15:22:41 I know May 24 15:22:49 Crofton: as said in my response, if arch is arm, it does "make arm", everything else is just "make". I know it works on arm and x86, the rest is untested. May 24 15:22:51 Zero_Chaos: could you please name bug numbers again May 24 15:23:06 Zero_Chaos: but why is it only needed for arm? May 24 15:23:34 zecke: because with arm, extra cflags are recommended (not required) May 24 15:23:45 zecke: 2361,2368,2369,2370 May 24 15:23:53 Zero_Chaos: why don't you mark patches as 'patch' in bugzilla? May 24 15:23:59 so i builds and works for arm and x86? May 24 15:24:26 koen: because I'm used to submitting new programs, but since I'm submitting as mtn diff i guess that makes everything a patch, doesn't it? May 24 15:24:27 why not do_compile_arm () {} May 24 15:24:31 koen: my bad May 24 15:24:42 with generic do_compile() {} for everything else May 24 15:24:51 Crofton: test for those two, yes. May 24 15:24:56 the stuff before and after can be in do_compile_append/prepend May 24 15:24:59 Zero_Chaos: and stop naming everything 'mtn diff' or 'another patch' May 24 15:25:07 XorA: I have no idea how to do that, and I took the patch from 1711. May 24 15:25:29 koen: I thought my comments were the description, I'll name things better in the future. May 24 15:25:30 Zero_Chaos: http://bugs.openembedded.org/attachment.cgi?id=1522 no entry to checksums.ini, why do you force O3 as compilation flags? May 24 15:26:09 cu May 24 15:26:34 * Crofton makes a note to visit Briancon one day May 24 15:27:03 zecke: -03 is in the makefile and desired by the author, although after previous discussions I suppose that is rather pointless. and the checksums.ini entry is in a comment because I cannot make a bunch of diffs of the same file (checksums.ini) May 24 15:27:11 MAKE_TARGETS="arm" May 24 15:27:11 } May 24 15:27:11 do_compile() { May 24 15:27:11 export KERNDIR=${STAGING_KERNEL_DIR} May 24 15:27:12 module_do_compile May 24 15:27:14 } May 24 15:27:39 Zero_Chaos: what do you mean with "ASPj is drunk" ? May 24 15:27:48 XorA: and what does this do for non-arm? I know the way I have it works. May 24 15:27:58 zecke: the author's description of his program. May 24 15:28:20 Zero_Chaos: for non arm it wont set MAKE_TARGETS May 24 15:28:33 updted May 24 15:28:43 updated bitbake to svn... same problem May 24 15:28:51 XorA: be a pal and throw that in the comments. if it is more correct I'm happy to update. May 24 15:29:28 it does do_fetch, and do_unpack for libxv (which is in DEPENDS). Then it moves straight onto do_compile for mypackage May 24 15:29:43 which is going to fail since do_stage didn't get run for libxv May 24 15:30:00 Zero_Chaos: done May 24 15:30:11 XorA: thanks, I'll update it in a few. May 24 15:30:13 tkp: is this against .dev? May 24 15:30:19 koen: yes May 24 15:30:24 hmmm May 24 15:30:33 base.bbclass is up to date? May 24 15:30:35 Zero_Chaos: you are almost there, probably one more iteration May 24 15:30:43 Zero_Chaos: its what I would do, not sure if others would do the same May 24 15:30:48 koen: it's all 1 week old May 24 15:31:15 XorA: I'll look at which one is uglier... but like I said, that part was stolen. May 24 15:31:47 zecke: just the arm change suggested by XorA is what you mean, right? and then happy? May 24 15:31:50 I'll update and try again May 24 15:33:30 Zero_Chaos: well, I can live with the checksums.ini sniplets in the bug tracker. But there a re some weird FILES_DEV-${PN} lines, some weird unexplained fakeroot do_install() methods... and what XorA and koen said in their comments (here and in the bugtracker) May 24 15:34:12 zecke: I explained the fakeroot in the bugtracker, but it is not only proper but it is covered in the OE FAQ. May 24 15:34:42 zecke: the odd FILES_DEV line is commented out and really there more for me to figure how than anything else. May 24 15:35:45 off for bus May 24 15:36:43 Zero_Chaos: adding empty fakeroot do_install() {} is in the FAQ? Then please point me to the FAQ entry May 24 15:37:18 zecke: the install routine is in the programs makefile, wrapping fakeroot around it is how you fix chown and chmod errors. May 24 15:37:49 Zero_Chaos: point me to the FAQ entry showing this construct. You said it is in the FAQ May 24 15:38:15 are these setuid binaries? May 24 15:39:05 zecke: I don't believe the specific construct is in the faq, but reguardless, it works. May 24 15:39:22 zecke: I disabled setuid I believe. May 24 15:39:37 Zero_Chaos: well, it was you who said it is in the FAQ May 24 15:39:56 Zero_Chaos: but an empty fakeroot do_install() {} is not bringing you far (that is the point of the comment) May 24 15:40:11 zecke: the install script wants to -o root and OE barfs on that because it's not running as root. the solution is in the FAQ and the way I did it WORKS May 24 15:40:14 * Zero_Chaos fails to see the issue May 24 15:40:51 Zero_Chaos: again, please point me to the FAQ entry! May 24 15:41:04 zecke: yes I *could* retype the entire "make install" routine into the fakeroot do_install() but that is pointless. it works fine empty and saves 50kb May 24 15:41:19 Zero_Chaos: again, please point me to the FAQ entry! May 24 15:41:20 Zero_Chaos: again, please point me to the FAQ entry! May 24 15:41:20 Zero_Chaos: again, please point me to the FAQ entry! May 24 15:42:01 wow May 24 15:42:12 zecke: why are you arguing about something that works? koen won't let me "hack" into the makefile and remove the -o root because that's ugly, and said use fakeroot. If I'm not using fakeroot right, then tell me how to use it, but it WORKS FINE May 24 15:42:33 zecke: you prefer I add fakeroot do_install () { oe_runmake "DSTROOT=${D}" install } May 24 15:42:33 ? May 24 15:42:54 zecke: does the same thing except explicitly calls something for no reason. May 24 15:43:04 and I'd likely drop the pointless DSTROOT anyway May 24 15:43:05 Zero_Chaos: sorry, you lost. I won't repeat myself more than three times. May 24 15:43:14 :} May 24 15:43:24 zecke: lost what? the reciepe works and I don't know what I did wrong? May 24 15:43:55 Zero_Chaos: if you are refering to the "FAQ" and are asked to print the uri and you don't do it you... May 24 15:44:26 zecke: you need the uri to the FAQ? under Build, number 1: http://www.openembedded.org/oefaq May 24 15:44:52 zecke: it has some extra stuff in there that wasn't required, and last time I explicitly called something for no reason I was told to remove it. May 24 15:45:50 well if "fakeroot do_install() {}" is calling make install then this is a bitbake bug May 24 15:46:10 do it as in the FAQ, or set the varflag of do_install to fakeroot... May 24 15:46:42 so the problem is, it might work today, but will stop working when the bug is fixed? May 24 15:46:47 zecke: no that makes sense, I said it works, I tested it (as always), if it's a bitbake bug that it works, that is well outside my knowledge. May 24 15:47:01 zecke: I'll add oe_runmake install May 24 15:47:11 * Zero_Chaos had no idea he found a bug May 24 15:47:24 that could have been mentioned. May 24 15:48:20 well, if files get installed then it is a bug in BitBake's parser May 24 15:48:52 then it is a bug in bitbake's parser, because the files were installed, and properly I might add. As I said, I tested it and it works. May 24 15:53:56 is there a table of hardware somewhere for OE? May 24 15:54:00 supported devices? May 24 15:57:30 ls conf/machines :) May 24 15:57:40 I thought so, thanks May 24 15:57:51 of course, not all may work :) May 24 15:58:10 Hi koen! May 24 15:58:17 Nice work on xorg packages. May 24 15:58:19 hey anttix May 24 15:58:28 Can I ask another small favor? May 24 15:59:28 Who maintains mesa package? May 24 16:00:17 It does not build on FC6 :( I have details figured out so I was wondering if anyone can take a look if my fix is good enough to be commited. May 24 16:00:18 http://bugs.openembedded.org/show_bug.cgi?id=2363 May 24 16:00:21 noboday afaik May 24 16:01:53 It's a strange bug and actually not OE-s fault, but it's easier to work around it than fix sed on host system. May 24 16:04:02 * * OE Bug 2363 has been RESOLVED (FIXED) by koen@dominion.kabel.utwente.nl May 24 16:04:04 * * [Bug 2363] mesa fails to build on Fedora Core 6 with May 24 16:04:06 * * http://bugs.openembedded.org/show_bug.cgi?id=2363 May 24 16:05:50 03antti.andreimann 07org.oe.dev * r6de428a8... 10/ (1 packages/mesa/mesa-mesa.inc): mesa: fix build on FC6, closes #2363 May 24 16:06:30 zecke: I believe Zero_Chaos was not really referring to the FAQ but to http://www.openembedded.org/user-manual&dpage=recipes_fakeroot May 24 16:07:00 It does not say to produce an empty routine, though May 24 16:07:04 Thanx koen, I owe You another beer in case we ever meet in person ;) May 24 16:07:44 Laibsch: thanks, we need to consider to fix that May 24 16:07:49 Laibsch: that too. like I said it worked empty, but since it *shouldn't*, I'm happy to fix. I just couldn't understand so much complaint of something that worked. May 24 16:09:29 Zero_Chaos: and to the DEPENDS += "fakeroot-native" it is almost correct (for bitbake < 1.8) May 24 16:09:51 Zero_Chaos: By googling I found http://www.openembedded.org/user-manual&dpage=fakeroot I have not read it. But maybe it could be interesting for you May 24 16:10:00 * Laibsch seriously does not know May 24 16:10:04 Laibsch: if you have some time, could you file bug reports to mention the FAQ and the UserManual May 24 16:10:20 zecke: I don't really understand the problem. May 24 16:10:27 I only skimmed the backlog May 24 16:10:39 backlog=scrollback? or whatever it is called May 24 16:10:54 zecke: so now I *shouldn't* add it to the depends, even though it is in the manual and fakeroot-native is required or my reciepe will fail? May 24 16:11:14 Zero_Chaos: I think you might eb in the best position to file these bugs. May 24 16:11:27 Laibsch: sure why not. May 24 16:11:44 Zero_Chaos: we are not perfect yet. The manual is almost right in reagrd to DEPENDS += "fakeroot-native" but not totally correct for BitBake1.8 May 24 16:11:47 brb May 24 16:11:56 Don't worry if they are not 100% accurate (you never should, despite some contrary and unrespectful comments), together we will figure it out May 24 16:12:14 zecke: see you May 24 16:12:30 zecke: can you define "not totally correct"? what *should* I do? May 24 16:12:46 it's barely noon and I so need a drink. May 24 16:13:00 Zero_Chaos: I wanted to make sure you understand that nobody expects bug reports to be poetry May 24 16:13:10 If there are mistakes that is OK May 24 16:13:18 Zero_Chaos: well do_install depends on fakeroot-native:do_populate_staging May 24 16:13:21 You are willing to learn, that is what really counts for me May 24 16:13:37 Laibsch: you added an extranious word, "nobody" doesn't seem entireley accurate ;-) May 24 16:13:39 Much more than delivering perfection at the first time around May 24 16:14:06 zecke: so how might I phrase that in the reciepe? May 24 16:14:27 Zero_Chaos: He can be a nice guy, too. But if you never met him in person, it is sometimes hard to tolerate his bitching May 24 16:14:32 Zero_Chaos: Bitch back!! May 24 16:14:37 That is what I do ;-) May 24 16:14:47 Zero_Chaos: classes/image.bbclass:18:do_rootfs[depends] += "makedevs-native:do_populate_staging fakeroot-native:do_populate_staging" May 24 16:14:52 Zero_Chaos: is an example May 24 16:14:59 ;) May 24 16:15:11 Laibsch: I don't pull punches, you may have noticed that in the comments when I respond to a few things.... May 24 16:15:42 zecke: holy shit that confuses me. read classes/image.bbclass:18 and learn? May 24 16:16:13 I think I kinda get it actually May 24 16:16:17 wow, that's sick May 24 16:16:24 Zero_Chaos: just this line. It says task do_rootfs depends on the task do_populate_staging of fakeroot-native May 24 16:16:39 Zero_Chaos: See? I have not gotten zecke's last comment. You'll soon be master! ;-) May 24 16:17:01 But OTOH I don't need to get it ;-) May 24 16:17:22 Laibsch: hopefully "master" comes with a little respect.... it seems only some people here give me any... May 24 16:17:45 zecke: kick, give me a minute and I'll write it up right. May 24 16:18:02 * * OE Bug 1923 has been RESOLVED (FIXED) by bugs.openembedded.org@rolf.leggewie.biz May 24 16:18:04 * * [Bug 1923] Reorder the dependencies of gtkmm and glibmm May 24 16:18:06 * * http://bugs.openembedded.org/show_bug.cgi?id=1923 May 24 16:19:46 zecke : Any idea why some python code (its in the oplinux-packages.bb) run fine on bitbake 1.6 but complains for indentantion on 1.8 ? May 24 16:20:09 steliosk: yes May 24 16:20:18 zecke: is 'do_install[depends] += "fakeroot-native:do_populate_staging"' required, or can I drop the ":do_populate_staging" part? May 24 16:20:37 Zero_Chaos: the do populate is crucial May 24 16:20:55 steliosk: well, for speed reasons we compile a lot of code at once (the anon functions) May 24 16:21:01 zecke: I'm guessing this isn't doumented anywhere as to why? :-) May 24 16:21:35 Zero_Chaos: the best thing is to do mtn log classes/image.bbclass and to read Richard's commit message May 24 16:21:48 zecke : ok. i see. how can i fix it ? May 24 16:21:49 steliosk: indention of anonfuncs are 4 spaces May 24 16:21:56 does anyone know of a board that is supported by u-boot 1.2.0 that boots only from NAND flash? May 24 16:22:02 zecke : ok May 24 16:22:48 zecke: what if people don't want to build an image, but only packages? May 24 16:23:07 I don't even have +w and I'm expected to read "mtn log"... well, at least I'm learning ;-) May 24 16:23:15 Crofton: the fic-gta01 boots from nand, but uses u-boot git May 24 16:23:22 thanks May 24 16:23:25 koen: this is why it needs to be added to the package bb file? May 24 16:23:35 koen: and I pointed him to an example? May 24 16:23:46 I need to read up n git againg May 24 16:24:05 I having to create u-boot support for the SFFSDR May 24 16:25:58 * chouimat is bored May 24 16:26:39 chouimat: you could comment on all my bugs, I'm sure I've done something else wrong ;-) May 24 16:30:48 * chouimat is listening to Prelude to descent by Paradise Lost on In Requiem [Amarok] May 24 16:31:17 zecke: is this okay 'fakeroot do_install () { oe_runmake install }' or do I need the DSTROOT (for some unknown reason)? May 24 16:32:18 you will need DSTROOT May 24 16:32:29 you are unlikely installing directly to /usr May 24 16:32:58 zecke: the install routine should have all that covered though... shouldn't it? May 24 16:33:19 zecke: seems like an extra override. May 24 16:33:37 Zero_Chaos: do it as in the FAQ May 24 16:34:35 zecke: at the risk of being an uber pain in the ass, I don't understand, can you explain why? again, it works without. May 24 16:35:50 well. a) I don't trust any human saying it works(tm). Don't be angry May 24 16:36:07 b) with normal packages we configure them with ./configure --prefix=/usr May 24 16:36:25 c) doing a make install would try to install to /usr (right?) May 24 16:36:38 zecke: I understand, first rule of any support roll (doctor, tech, whatever) is that the person you are talking to is lying ;-) May 24 16:37:01 luckily most sane packages respect setting another destination (probably DSTROOT) May 24 16:37:04 zecke: c) no, make install takes the prefix and appends the bin and sbin as appropriate in this case. May 24 16:37:20 which still would be /usr/bin/foo-binary May 24 16:37:36 zecke: which is where the program should go... May 24 16:37:46 no? May 24 16:38:12 Zero_Chaos: pal, abstract. Sure when you install a package you want it on /usr/bin May 24 16:38:29 but when compiling do you want to overwrite your host ls with one for ARM? May 24 16:39:33 zecke: likely not, but in this case that doesn't happen. I guess I understand why it may be needed at this point, i just know that in this case it's not. but to save a possible accident let's do it your way to be safe ;-) May 24 16:39:41 zecke: I'm such a nice guy, aren't I? :-) May 24 16:41:20 Zero_Chaos: right, for packaging purposes we ask the app to install in another directory... May 24 16:46:27 zecke: considering the conversation we just had, shouldn't oe run in a chroot or something? May 24 16:47:37 chroot needs root permissions May 24 16:47:48 Thats the whole reason why OE was invented in the first place May 24 16:48:07 heh, fair enough. May 24 16:48:13 virtual chroot ? ;-) May 24 16:48:20 fakechroot has its problems May 24 16:48:35 chroot is not the solution - we have proved conclusively, time, after time that a chroot is not needed May 24 17:00:38 zecke : that was it. Thanks ! May 24 17:01:05 * chouimat goes trying the OSX support of icecream May 24 17:01:18 zecke : btw have you looked at how things like hyperthreading etc effect the 1.8 ? (better/worst execution time) May 24 17:04:45 * steliosk wishes he had a multi cpu / multi core machine to run 1.8 May 24 17:05:11 steliosk: no idea :) May 24 17:09:03 wow I'm about to use sed to fix libiconv May 24 17:11:08 I think we need a concentrated effort in pushing our fixes to upstream projects. Hell, the quality of build files is really in everybodys best interes May 24 17:15:27 anttix: agreed, at least I mail (most of) my changes upstream May 24 17:16:26 03coredump2 07org.oe.dev * r2e1d9e05... 10/ (8 files in 5 dirs): Add The Hentges ROM 2.x ( based on Ansgtrom ) to OE. Very early state. May 24 17:16:31 Maybe we could have a sort of a database, maybe wiki page for a start, with upstream bugzilla links so we can track progress? May 24 17:16:32 03coredump2 07org.oe.dev * r525da879... 10/ (3 files in 3 dirs): Fix license May 24 17:22:02 zecke : one more for you :) May 24 17:23:04 zecke : I am trying to set PARALLE_MAKE from with in icecc.bbclass to "" for packages that are not supposed to use icecc May 24 17:23:47 steliosk: yes, you know where to do it? May 24 17:24:14 zecke : nop, that what i am trying to find out. whatever i tried till now failed May 24 17:29:02 steliosk: python () { import bb\n bb.data.setVar('PARALLEL_MAKE' , '', d) } ? May 24 17:29:24 steliosk: after the bb file has been parsed this method will be invoked May 24 17:35:43 robtaylor: do you work for collabora? May 24 17:43:05 03freyther 07org.oe.dev * r7efadd62... 10/ (4 files in 3 dirs): (log message trimmed) May 24 17:43:05 packages/libiconv: Battle libiconv to not use a rpath (and fix the uclibc build) May 24 17:43:05 -Update libiconv to 1.11 (the latest) May 24 17:43:05 -We do not package the LD_PRELOADable libiconv stub May 24 17:43:05 -Try hard to disable the rpath issue and. May 24 17:43:05 --disable-rpath doesn't disable the roath May 24 17:43:07 --disable-rpath=no (wasn't tested and if it works....) May 24 18:09:03 * * OE Bug 2352 has been RESOLVED (WONTFIX) by bugs.openembedded.org@rolf.leggewie.biz May 24 18:09:04 * * [Bug 2352] Unable to create gstreamer . xml plugin May 24 18:09:07 * * http://bugs.openembedded.org/show_bug.cgi?id=2352 May 24 18:28:45 Any distro and task-base expert around? May 24 18:29:26 I have a distro (OE port of the thinstation) which uses busybox init instead of sysvinit. May 24 18:30:11 Currently I am not using task-base, because it used to include sysvinit. May 24 18:32:37 However now there is a way to exclude sysvinit May 24 18:32:54 but, initscripts is still included despite the fact that sysvinit is not selected. May 24 18:32:59 Is it intentional? May 24 18:37:58 is ifplugd depreciated? May 24 18:40:22 Can't tell if it's depreciated in OE, but it's been quite stale for a long time in upstream though. May 24 18:41:48 bbl May 24 18:49:59 anttix: yeah that's what I've noticed, and it's not building for me :( May 24 18:50:12 I didn't know if there was something newer and better out there May 24 18:51:20 Well Gnome is pushing a thing called NetworkManager maybe You should take a look May 24 18:51:52 The base daemon shouldn't require gnome libs only dbus and friends. May 24 18:52:09 Don't know if it's in OE though, and too lazy to check right now :P May 24 18:53:01 yup I've heard of networkmanager May 24 18:53:30 I just got ifplugd to compile with the upstream kernel-types.patch for nylon May 24 18:54:22 well, file a bug report with a patch then, and bug package maintainer about it ;) May 24 18:57:57 i think freyther's last push created a second head in MTN May 24 18:58:09 could someone check May 24 18:58:45 pulling May 24 19:00:05 Who maintains cups? May 24 19:01:11 it looks like only one head amt May 24 19:01:14 atm May 24 19:01:41 And who maintains xterm? Koen You? ;) May 24 19:01:44 hmmm so probably is my db May 24 19:02:25 This ancient xterm compile bug still needs fixing :P http://bugs.openembedded.org/show_bug.cgi?id=1352 May 24 19:02:32 Just tested on FC6 May 24 19:02:40 CROSS COMPILE Badness: /usr/include in INCLUDEPATH: /usr/include/freetype2 May 24 19:41:16 03ifaistos 07org.oe.dev * rad877be1... 10/ (1 packages/meta/oplinux-packages.bb): May 24 19:41:16 packages/meta/oplinux-packages.bb : Fix "identation error" when used with bitbake 1.8 (thanks to zecke for the info) May 24 19:41:16 Varius typos fixed May 24 19:52:55 03ifaistos 07org.oe.dev * r3524b1de... 10/ (1 classes/icecc.bbclass): May 24 19:52:55 classes/icecc.bbclass : Make sure that PARALLEL_MAKE is set to null if the package is not to use the icecc scheduler. May 24 19:52:55 In this way you don't end up with -jX to be handled by a single cpu/machine May 24 19:52:55 Remove glibc and add glibc-intermediate in the list of packages not use the icecc scheduler May 24 20:25:02 * * OE Bug 1925 has been RESOLVED (INVALID) by koen@dominion.kabel.utwente.nl May 24 20:25:04 * * [Bug 1925] Remove cairo/libsvg, May 24 20:25:06 * * http://bugs.openembedded.org/show_bug.cgi?id=1925 May 24 20:30:18 hi florian May 24 20:31:01 re May 24 20:55:06 koen: Does gpe-bootsplash truly depend on libsvg-cairo or could it depend on librsvg instead? May 24 20:55:31 That is the way I understood bug 1925 May 24 20:56:17 it depends on libsvg-cairo till someone ports it to librsvg May 24 20:56:28 OK May 24 20:57:14 heh May 24 20:57:19 LAK is funny May 24 20:57:24 oh May 24 20:57:38 "use buildroot" "no, use OE" "no, use ptxdist" May 24 20:58:31 anyone brok glibc-intermediate lately? May 24 20:58:44 do *not* use a montavista toolchain ;) May 24 20:58:50 I built it for i586 yesterday May 24 20:59:21 I'm working with a guy to build a tol chain so he can build a davnici kernel May 24 20:59:34 so he is building angstrom/osk May 24 20:59:48 ERROR: log data follows (/stuff/tmp/work/armv5te-angstrom-linux-gnueabi/glibc-intermediate-2.5-r6/temp/log.do_configure.4850) May 24 20:59:48 | FATAL: no configure script found May 24 21:00:07 that one is weird May 24 21:00:14 that looks like it failed to extract May 24 21:00:53 anyway May 24 21:00:56 'night all May 24 21:01:06 good night May 24 21:12:07 morning May 24 21:12:08 :-) May 24 21:24:06 likewise: Hi May 24 21:25:07 evenin' chaps May 24 21:26:39 likewise or someone else: I have a newbi question and I believe that you can help me: I'm trying to build a lot of packages that are not available in their original repositories... what variable can I set in my distro conf to download from http://www.angstrom-distribution.org/unstable/sources/ ? May 24 21:28:41 vivijim: sorry, I don't know May 24 21:28:49 vivijim: I don't know from heart. SOURCE_MIRROR or something... May 24 21:29:12 vivijim: It should do that automatically May 24 21:29:33 my turn for a question... anyone know offhand how big the kdrive (and various optional libs) are? or where I could find that info? May 24 21:29:36 and if the sources are not available as specified in the bb file than that is a bug that needs fixing May 24 21:32:34 hi May 24 21:34:02 * * OE Bug 972 has been RESOLVED (WONTFIX) by bugs.openembedded.org@rolf.leggewie.biz May 24 21:34:04 * * [Bug 972] hostap .pc/ipaq_compat.patch is wrong! May 24 21:34:06 * * http://bugs.openembedded.org/show_bug.cgi?id=972 May 24 21:35:02 * * OE Bug 979 has been RESOLVED (WONTFIX) by bugs.openembedded.org@rolf.leggewie.biz May 24 21:35:03 * * [Bug 979] at76c503a-modules-0.0+cvs fails compiling... May 24 21:35:06 * * http://bugs.openembedded.org/show_bug.cgi?id=979 May 24 21:40:02 * * OE Bug 2373 has been created by bugs.openembedded.org@rolf.leggewie.biz May 24 21:40:04 * * gsoap-native fails do_compile May 24 21:40:06 * * http://bugs.openembedded.org/show_bug.cgi?id=2373 May 24 21:40:06 Laibsch: if this is a bug: is a fix to change the bb file of package pointing to http://www.angstrom-distribution.org/unstable/sources/ where we are sure that this sources are available? May 24 21:40:30 vivijim: No, that is not the answer May 24 21:40:37 What bb file are you talking about? May 24 21:41:17 http://distro.ibiblio.org/pub/linux/distributions/gentoo/distfiles/elfutils-0.108.tar.gz for example... May 24 21:42:07 Pyrex-0.9.3.1.tar.gz; tk8.4.11-src.tar.gz; expat-1.95.7.tar.gz, etc... May 24 21:42:44 you should file bug reports that do_fetch fails May 24 21:42:53 one bb file - one report May 24 21:43:16 ok... got it.. I'll do that... May 24 21:43:59 Extra points for including where the file can be gotten instead. May 24 21:44:28 If it is an old version gone away, then we can update the bb file version (aka PV) May 24 21:45:40 uhm got it... Thanks! May 24 21:45:47 elfutils for example needs an update to version 127 May 24 21:47:17 Try "mv packages/elfutils/elfutils_0.108.bb packages/elfutils/elfutils_0.127.bb" and see if that fixes your problem (fetch should work, compilation might fail) May 24 21:47:40 I will try and make the necessary commit later. but I am currently struggling with something else. May 24 21:48:26 uhm... I have a list here and I believe that most of them needs an update... I'll report these bugs tomorrow... Now I'm leaving... for a while I was downloading from angstron repository to sources dir... May 24 21:48:50 You could just make one single bug report in this case, I think May 24 21:49:03 Let me have the list and I will try to fix as many as I can May 24 21:49:22 * Bernardo is away: Ausente por agora. May 24 21:53:28 Laibsch: I tried a pvt but I receive a lot of auto-reply in germam, and I didn't undestand... :) May 24 21:56:06 Laibsch: the list is here: http://pastebin.ca/508171 May 24 21:56:10 thanks for your help May 24 21:56:11 bye May 24 22:12:25 anyone got the email address of thomas kunze? May 24 22:12:40 ah found it May 24 22:39:02 ibot: botmail for vivijim: most of the packages in your list fetch just fine. Make sure you use the default SOURCEFORGE_MIRROR variable. IOW, do *not* specify it explicitly in local.conf. May 24 22:39:33 ibot: botmail for vivijim: most of the packages in your list fetch just fine. Make sure you use the default SOURCEFORGE_MIRROR variable. IOW, do *not* specify it explicitly in local.conf. May 24 22:41:02 Laibsch: is there an echo in here?? May 24 22:41:21 No, I sent the message twice by accident May 24 22:41:30 hvontres|poodle: Still no cables BTW May 24 22:41:37 Laibsch: hmmmmmm May 24 22:42:06 Laibsch: I did confirm that I sent it on the 9th... May 24 22:42:38 Laibsch: unfortunately, I didn't think to add tracking.... :( May 24 22:47:13 In fact, every single one of those "unfetchable" reports was bogus :-/ May 24 22:47:39 hvontres|poodle: I am sure they will arrive eventually. Maybe they are indeed being "shipped" May 24 22:47:47 That would take 2-3 months May 24 22:47:57 Or they are stuck in customs May 24 22:48:05 n8 May 24 22:50:20 Laibsch: exept that they were sent "Global Priority Mail" which should arive in 6-9 days...:( May 24 22:50:49 Laibsch: but customs is probably a good guess. May 24 22:54:08 Wait a minute. May 24 22:54:16 6 to 9 days from the 9th? May 24 22:54:34 I was gone from the 15th to the 20th May 24 22:55:12 But I did not find a notice that a package was waiting for me (that is normally the case if they cannot deliver it to you in person) May 24 22:55:22 Sometimes they deliver to the neighbours. May 24 22:55:29 I will ask around tomorrow May 24 22:57:07 Usually they also have to leave a notice if they deliver to the neighbours May 25 00:00:56 RP, You awake, could use some quick feedback **** ENDING LOGGING AT Fri May 25 02:59:57 2007