**** BEGIN LOGGING AT Fri Aug 01 02:59:56 2008 Aug 01 03:23:06 * * OE Bug 4464 has been created by raj.khem(AT)gmail.com Aug 01 03:23:08 * * RRECOMMEND for libgcc in glibc/eglibc Aug 01 03:23:10 * * http://bugs.openembedded.net/show_bug.cgi?id=4464 Aug 01 06:31:00 morning Aug 01 06:31:25 morn Aug 01 06:33:40 morning Aug 01 06:37:18 morning Aug 01 07:13:34 I am trying to bitbake a gpe-image,. and it failed... How can I fix that? Aug 01 07:14:03 by hand Aug 01 07:14:04 where did it fail? Aug 01 07:14:45 on compiling a package "ODBM_File.o" error 1 Aug 01 07:15:37 there is no package on that name Aug 01 07:16:35 package: perl-native-5.8.8-r13/perl-5.8.8/ext/ODBM_File Aug 01 07:17:06 perl-native is package Aug 01 07:17:15 use pastebin to show error log Aug 01 07:17:40 what is pastebin? Aug 01 07:19:46 ~pastebin Aug 01 07:19:47 [~pastebin] A "pastebin" is a web-based service where you can paste anything over 3 lines without flooding the channel. Here are links to a few : http://www.pastebin.com , http://pastebin.ca , http://channels.debian.net/paste , http://paste.lisp.org , http://www.rafb.net/paste Aug 01 07:22:06 check pastebin.com on nighthawk paste... Aug 01 07:26:20 give me url Aug 01 07:27:01 http://pastebin.com/d38b30287 Aug 01 07:28:30 # Aug 01 07:28:30 | ODBM_File.xs: In function 'XS_ODBM_File_DESTROY': Aug 01 07:28:30 # Aug 01 07:28:31 | ODBM_File.xs:124: error: too few arguments to function 'dbmclose' Aug 01 07:28:40 thats what you should start with ;D Aug 01 07:29:08 how do I fix that? Aug 01 07:29:27 looks like perl-native does not have --disable-dbm or such for configure.. Aug 01 07:30:31 do I need to run it with that switch? how do I fix that? Aug 01 07:31:14 moment.. let me look at Aug 01 07:36:19 check tmp/work/*/perl-native*/temp/log*configure* file and tell me does it found dbmclose() Aug 01 07:38:24 dbmclose() found. Aug 01 07:38:46 ~curse perl for crappy configure scripts Aug 01 07:38:47 May the fleas of a thousand camels infest your most sensitive regions, perl for crappy configure scripts ! Aug 01 07:39:04 nighthawk08: uninstall libdbm-dev or how it is called on you host system please Aug 01 07:41:05 I am using suse 11,. how can I uninstall the libdbm-dev? Aug 01 07:45:33 hrw? Aug 01 07:46:00 no idea - i never used suse Aug 01 07:49:14 there is a bug already in bugzilla for that Aug 01 07:49:29 and a patch if i recall Aug 01 07:49:50 s/dbmclose/xdbmclose/ Aug 01 07:56:10 good morning Aug 01 07:57:27 hrw: what exactly is that library? I found libdb-4_5-devel is that it? Aug 01 07:58:03 nighthawk08: XorA told that patch is in bugzilla - please look and test it Aug 01 07:58:30 thx gos that no one tries to use Poky on suse Aug 01 07:58:37 s/gos/gods Aug 01 07:58:54 o/ Aug 01 07:59:09 http://bugs.openembedded.net/show_bug.cgi?id=4396 Aug 01 07:59:23 not a patch, but workaround Aug 01 08:00:35 where do I need to change the line to xdbmclose()? Aug 01 08:01:07 * XorA points out he doesnt use suse either Aug 01 08:01:12 just remebered the bug Aug 01 08:10:17 XorA: hi Aug 01 08:10:30 yo ant|work Aug 01 08:10:48 XorA: do you know about dbus changes? Aug 01 08:10:59 ant|work: no Aug 01 08:31:03 seems that the patch worked properly... Aug 01 08:32:15 yes, but perl-native remains broken...look at bug 3403 Aug 01 08:36:05 XorA: how many gsmd exists in openmoko world? Aug 01 08:36:27 XorA: om2007.2 gsmd, asu gsmd2 or there are more? Aug 01 08:38:00 florian: good morning Aug 01 08:38:02 hrw: probably Aug 01 08:38:05 hrw: I lose track Aug 01 08:38:16 hrw: but FSO will take over at least 2007.2 one Aug 01 08:38:22 good morning Aug 01 08:38:24 hrw: community is working on that now Aug 01 08:42:34 * ant|work hopes the two communities will finally agree on a gps daemon Aug 01 08:44:01 XorA: navit_svn + gpsd + dbus = no go. Is openmoko any better? Aug 01 08:44:25 can I borrow some patches? Aug 01 08:45:26 * XorA has never played with GPSD Aug 01 08:45:38 ach Aug 01 08:46:15 ant|work: I think that openmoko guys likes to reinwent everything and write it in python Aug 01 08:46:39 hrw: and o-h ? http://folks.o-hand.com/iain/gypsy/why-not-gpsd.html Aug 01 08:46:41 ogpsd which is gypsy/gpsd in python Aug 01 08:46:53 ant|work: gypsy has sense compared to gpsd Aug 01 08:47:28 ogpsd is python rewrite of gypsy just to spend^Wwaste some time instead of patching gypsy Aug 01 08:47:39 I see... Aug 01 08:47:56 from what I know they wrote ogpsd just because gypsy does not supported ublox gps units... Aug 01 08:48:13 so why add support to existsing daemon if you can write new one? Aug 01 08:48:21 he he Aug 01 08:49:33 :-| Aug 01 08:50:28 here it's why use audiofile when you can put vlc to just read mp3 =) Aug 01 08:51:03 Genesis: or mplayer Aug 01 08:51:15 * hrw wants mplayer based bootsplash Aug 01 08:51:25 yeap , it's no screen application , i suggessted mpd ... Aug 01 08:51:50 they ask me cli skype Aug 01 08:51:51 etc Aug 01 08:52:14 and they don't know what tarball means. Aug 01 08:53:44 ok, enough joking Aug 01 08:55:54 ^^ Aug 01 08:56:07 you're , they not ;-) Aug 01 09:37:10 hi lrg, woglinde Aug 01 09:37:13 hi Aug 01 09:37:39 hey pb_ Aug 01 09:39:57 hi pb Aug 01 09:54:12 woglinde: I have a little doubt about speech-dispatcher Aug 01 09:54:53 ant tell me Aug 01 09:55:06 flite or flite-alsa? Aug 01 09:56:17 I see flite depends from alsa-oss which depends from alsa-lib Aug 01 09:59:42 why this flite-alsa then? Is only for the EXTRA_OECONF = "--with-audio=alsa --enable-shared" ? Aug 01 10:02:10 woglinde: I correct myself, to be precise the doubt is about flite (depended by speech-dispatcher) Aug 01 10:03:51 ant dont knoow Aug 01 10:04:00 it was in the recipe Aug 01 10:04:39 found it ? http://projects.linuxtogo.org/pipermail/openembedded-commits/2008-January/011388.html Aug 01 10:08:35 ~seen xjqian Aug 01 10:08:35 xjqian was last seen on IRC in channel #opie, 10d 14h 42m 1s ago, saying: 'hi bluelightning'. Aug 01 10:10:30 ant doesnt it work? Aug 01 10:10:40 I only test compiling Aug 01 10:10:47 woglinde: thx anyway, in stable was working, now I see why .dev does not Aug 01 10:10:59 default to oss Aug 01 10:11:13 hm provide a patch I will apply Aug 01 10:11:34 I'll try to depend on flite-alsa Aug 01 10:11:47 and see Aug 01 10:13:58 woglinde: speech-dispatcher was stalling, then I tried flite, one symptom was: 'flite_time 12:23' -> 'oss_audio: failed to open audio device /dev/dsp' Aug 01 10:14:25 oss_audio was already suspect Aug 01 10:15:21 just wondering how sane the dependencies of flite are... Aug 01 10:20:12 morning Aug 01 10:20:39 hrw: btw it seems I miss 'aplay', alsa-utils was installed, though Aug 01 10:21:56 alsa-utils-aplay is in feed Aug 01 10:22:22 hrw: it's even in my console-image build Aug 01 11:27:37 morning all Aug 01 11:28:33 hi thesing Aug 01 11:29:55 hrw: my mainboard broke again. I guess K8N boards are just crappy. Aug 01 11:31:05 But fortunately this time I have a laptop Aug 01 11:37:41 thesing: if you switch to intel, badaxe is old but stable (and is ok with OSx86 too) Aug 01 11:38:10 thesing: did btw your C menu-proggy compile? Aug 01 11:38:41 a kosme Aug 01 11:38:49 sorry Aug 01 11:42:45 ant|work: I was busy installing linux on my laptop yesterday. Aug 01 11:44:11 I can imagine...:-/ Aug 01 11:46:01 Fortunately my brother has a pc too. Otherwise there wouldn't be a way to access the code of the program. Aug 01 11:47:14 urgh, sources not lost at least Aug 01 11:48:50 I now commited stuff to a repository. Aug 01 11:58:38 ping Aug 01 11:58:51 pong Aug 01 11:59:07 Any objections if I update EFL_SRCDATE to "20080802" in sane_srcdate? Aug 01 11:59:16 eet is broken in builds before tht Aug 01 11:59:18 that Aug 01 12:00:54 alphaone: efl is still cvs? Aug 01 12:00:59 hrw: Yeah Aug 01 12:01:14 And it's a pita to update, too Aug 01 12:01:32 since the cvs server is really slow in responding sometimes.. Aug 01 12:01:56 hrw: Anyway it builds with "20080802" Aug 01 12:02:20 At least everything that is needed for illume builds... Aug 01 12:05:31 alphaone, fine, as long as the rest of us can bitch :) Aug 01 12:05:41 alphaone: let openmoko donate some bandwith to rasterman to get him moved to svn or git Aug 01 12:06:06 hrw: I think there's an experimental git mirror already Aug 01 12:06:15 even mtn would be an improvement .... Aug 01 12:06:17 stefan_schmidt: Knows more Aug 01 12:06:20 :-) Aug 01 12:08:36 got a problem while compiling with bitbake the linux-handhelds package... how do I solve that? Aug 01 12:09:17 while doing "do_configure" in it Aug 01 12:14:57 uclibc/images/htc himalaya// does not exist, you may need to bitbake it seperatly Aug 01 12:15:07 how do I bitbake that seperatly? Aug 01 12:15:24 i'm trying to port webif for openembedded , i wonder if some router use it with openembedded ? i don't find anyone Aug 01 12:16:51 anybody? Aug 01 12:18:54 Crofton: Okay, you can start bitching now ;-) Aug 01 12:19:28 I'm not sure I actually do any builds that use it :) Aug 01 12:19:47 Crofton: Even better :-) Aug 01 12:20:17 I need to get a monitor for my beagle board Aug 01 12:20:50 Crofton: You have nothing DVI capable? Aug 01 12:20:55 03  07org.openembedded.dev * rf517aae780 10OE.dev/packages/openmoko-projects/ (illume/module.illume.src illume_svn.bb): Aug 01 12:20:55 illume_svn.bb: The keyboard is now runtime configurable Aug 01 12:20:55 This unbreaks the build and also enables the internal keyboard again. Aug 01 12:21:05 I do, but it is attached to another computer Aug 01 12:21:13 Ah :-) Aug 01 12:21:14 it is a buisance moving the cable around Aug 01 12:21:26 Yeah, I know Aug 01 12:21:28 and 90% of what I need to do works via console/ssh Aug 01 12:22:37 03  07org.openembedded.dev * rb3122d2565 10OE.dev/conf/distro/include/sane-srcdates.inc: conf/distr/include/sane-srcdates.inc: Update EFL_SRCDATE (unbreaks eet) Aug 01 12:23:06 03  07master * rb3122d2565 10OE.dev/conf/distro/include/sane-srcdates.inc: conf/distr/include/sane-srcdates.inc: Update EFL_SRCDATE (unbreaks eet) Aug 01 12:24:44 03  07master * rf517aae780 10OE.dev/packages/openmoko-projects/ (illume/module.illume.src illume_svn.bb): Aug 01 12:24:44 illume_svn.bb: The keyboard is now runtime configurable Aug 01 12:24:44 This unbreaks the build and also enables the internal keyboard again. Aug 01 12:27:24 * ant|work went to OESF hell and bothered the PDAX' u-boot developer... Aug 01 13:30:11 ant|work, my kexecboot programm is 52K with static linking against klibc on x86. Now only my oe build on the laptop has to finish. Aug 01 13:31:59 thesing: the one with menu or kexec-only? Aug 01 13:32:35 only the menu program. Aug 01 13:32:53 congrats, it's minimal size! Aug 01 13:32:55 kexec is another 100K. Aug 01 13:35:41 yeah, with not-so-much-stripped defconfig should easily fit (just -oabi and -network support). Aug 01 13:35:51 hrw: what do you think? Aug 01 13:38:05 (hrw gave me some preconfigured minimal 2.6.24 defconfigs_c7x0, perhaps has some 2.6.26 ready for the purpose) Aug 01 13:39:57 I do not worked on minimal configs Aug 01 13:42:24 ant|work, it should be easy to modify the collie config for other machines. Aug 01 13:42:37 ok, hrw had some no-pm patches Aug 01 13:42:44 I'll see if I can port Aug 01 13:43:04 (I have it somewhere @home) Aug 01 13:44:10 hrw: kernel size was around 750 kb IIRC? Aug 01 13:46:21 about Aug 01 13:46:24 found it: one 766144b and other 819428b Aug 01 13:46:37 766144 Aug 01 13:46:43 :-] Aug 01 13:47:44 thesing: what's your partition size on collie? 1Mb instead 1,2 like c7x0/akita? Aug 01 13:48:10 approximately Aug 01 13:59:05 collie has exactly 1MB for kernel Aug 01 14:16:11 ah, hrw, and about that u-boot-mkimage-openmoko-native in OE? Another openmoko Easter Egg? Aug 01 14:22:26 no Aug 01 14:27:16 bah, I don't see why other distros should 'require u-boot-openmoko_git.bb' when using uImage (kernel.bbclass) Aug 01 14:28:43 poky use uboot-mkimage-native for it Aug 01 14:28:56 exctly the point... Aug 01 14:30:38 hrw: was this commit http://gitweb.openembedded.net/?p=org.openembedded.dev.git;a=blobdiff;f=classes/kernel.bbclass;h=06ae2466ea0100919ea00aa20628982fb52c6840;hp=f111c9f0921e61f2b18b8260b22ea1b6265e0dce;hb=750021c27fc4f8ccdf4f74211b06e2cfc2c26d25;hpb=fff0ee28a2540a73950871ac4adf2419f69a6aef Aug 01 14:32:33 was this commit.... Aug 01 14:32:42 irssi do not like long urls Aug 01 14:32:58 Works fine for me :-/ Aug 01 14:33:01 http://gitweb.openembedded.net/?p=org.openembedded.dev.git;a=commitdiff;h=750021c27fc4 is same Aug 01 14:33:07 eh, Laibsch pointed this out...too long links Aug 01 14:33:24 piroko: works here too but I do not see if there was something after link Aug 01 14:33:37 ant|work: it is a matter of looking what you paste Aug 01 14:33:45 me too I don't see (java webchat) Aug 01 14:33:48 :-( Aug 01 14:34:02 Huh. Mine just wraps around Aug 01 14:36:43 piroko: I use irssi in screen over ssh session Aug 01 14:36:46 Irssi 0.8.10 (20051211) - http://irssi.org/ Aug 01 14:36:53 quite old irssi Aug 01 14:36:56 hrw: So do I Aug 01 14:37:10 irssi 0.8.12 (20071006 0939) Aug 01 14:37:31 hrw: Two screens sessions actually. One on host, one on server. Screen in screen :D Aug 01 14:37:49 hrw: Hit ctrl-A F Aug 01 14:37:53 (shift F) Aug 01 14:38:37 hrw: And if that doesn't do it, make sure wrap is on. (ctrl-a r) Aug 01 14:39:18 mine is Client: irssi 0.8.10 (20051211 0941). And i see links to Aug 01 14:40:05 ok, guys, the point is not irrsi but mkimage-native ! Aug 01 14:41:15 heh Aug 01 14:41:33 lol. Sorry! Aug 01 14:41:33 pb__: hello Aug 01 14:41:55 piroko: np ;-) Aug 01 14:42:56 ant|work, have you asked to jeremy_laine? Aug 01 14:43:32 no, I just discovered this Aug 01 14:44:18 actually I do mkimage step by hand, but would integrate it like other machines do Aug 01 14:51:36 pb__: hi Aug 01 14:56:47 mickey|zzZZzz: Amethyst has disappeared offline :/ Aug 01 14:57:35 RP: hello Aug 01 14:57:36 mickey|zzZZzz, also dead here Aug 01 14:57:46 hi khem Aug 01 14:57:50 RP: I want to commit http://paste.debian.net/13593/ Aug 01 14:58:21 RP: this fixes the problem with dummy libc.so linking with some shared objects when building final glibc Aug 01 14:58:25 khem: Can we tell gcc-cross-intermediate to look somewhere else for that? Aug 01 14:58:58 you mean not in staging ? Aug 01 14:59:03 khem: yes Aug 01 14:59:45 hi khem Aug 01 14:59:49 RP: hmmm I think that will need more telling to gcc because our staging is our sysroot Aug 01 15:01:46 pb__: RP there is another issue with satging that I am seeing thats specific to toolchain because of intermediate packages. Aug 01 15:02:35 khem: What kind of issue? Aug 01 15:02:53 pb__: I explained why busybox is needing libgcc in !oebug 4464 Aug 01 15:03:19 khem: right, thanks, I saw that Aug 01 15:03:29 RP: when I want to rebuild say glibc-initial Aug 01 15:03:43 khem: and then it trashes staging? Aug 01 15:03:49 on a already staged sysroot Aug 01 15:04:09 khem: Yes, I've worried about this with packaged-staging too Aug 01 15:04:21 the different gccs and glibc all corrupt each other Aug 01 15:04:26 it deinstalls gcc-cross and gcc-cross-intermediate but all the gcc phases share files Aug 01 15:04:45 so indirectly when it removed cross-gcc is also removed the other two Aug 01 15:04:45 it shouldn't deinstall them Aug 01 15:05:01 what deinstalled them? Aug 01 15:05:11 and it does not deinstall gcc-cross-initial because thats required by glibc-initial Aug 01 15:05:41 but the compiler executable the deleted already so glibc-initial rebuild fails due to that Aug 01 15:06:09 RP: it must be bitbakes logic of staging Aug 01 15:06:21 khem: Its possible there are some races in packaged-staging Aug 01 15:06:44 RP: I see, what is best way to look at it. Aug 01 15:07:13 khem: Actually its not a race. Its the face the gcc packages have overlapping contents Aug 01 15:07:17 RP: I think it sees the dependency chain Aug 01 15:07:28 uninstalling gcc-cross breaks gcc-cross-in* Aug 01 15:07:44 I saw this problem with BB_STAMP_POLICY=whitelist Aug 01 15:07:45 The only solution there is to stop the packages overlapping Aug 01 15:07:57 RP: surely if they have overlapping contents you shouldn't have been able to install them both in the first place Aug 01 15:08:19 pb__: I've not been able to enable that yet due to this exact problem Aug 01 15:08:27 oh, right Aug 01 15:08:29 doh Aug 01 15:09:32 RP: I think there should be something like if you dinstall gcc-cross that would mean gcc-cross-intermediate and gcc-cross-initial are also deinstalled Aug 01 15:09:32 well, yeah, if you want to support rebuilding arbitrary bits of the early bootstrap toolchain then the only solution is to deconflict all the files by creating a separate bootstrap staging area. Aug 01 15:10:06 That solution does look attractive Aug 01 15:10:23 personally I think my attitude would be "don't do that, then", but if somebody else wanted to do the deconflicting then I would consider it a splendid idea :-} Aug 01 15:10:31 same is true for glibc and glibc-initial Aug 01 15:11:04 khem: yes, the two problem cases are gcc and glibc. If they were fixed, I'd have bitbake error on finxing conflicting files... Aug 01 15:11:28 * ant|work wonders if the Gentoo solution to nuke intermediate libc could be helpful Aug 01 15:11:39 RP: rpm has something like package a obsolete package b Aug 01 15:11:58 khem: So do .ipks and .debs Aug 01 15:12:06 yeah, dpkg and ipkg (and presumably opkg) have that too, but it doesn't really help in this case Aug 01 15:12:06 a revdep-rebuild for oe would be a dream :D Aug 01 15:12:08 and then when you install package a and package b is already installed then it get removed from database Aug 01 15:12:11 khem: but the way staging works we can't use it Aug 01 15:12:41 valhalla: BB_STAMP_POLICY = "whitelist" is what you refer to Aug 01 15:12:46 RP: I meant this http://dev.gentoo.org/~vapier/eh/?part=1&chap=4 Aug 01 15:13:15 pb__: I have this patch http://paste.debian.net/13593/ to remove dummy libc.so Aug 01 15:13:41 i didn't know about that: interesting Aug 01 15:13:44 thanks Aug 01 15:14:38 khem: well, another benefit to the thing we are discussing here is that it would remove the dummy libc.so problem as a side effect. Aug 01 15:15:02 if the fake libc.so is corralled off in some bootstrapping ghetto where only gcc-cross-intermediate looks, it won't hurt anything else :-) Aug 01 15:15:24 pb__: yes Aug 01 15:17:09 pb__: with current satging and a way to remove the intermediate packages as I was suggesting will reproduce same toolchain bit everytime Aug 01 15:17:45 khem: The problem is that with the way staging currently works, conflicting packages will not help you Aug 01 15:18:43 RP: may be we can enhance bitbake to add a new keywork which has this semantics Aug 01 15:19:07 You could have something at the bitbake level that says "if X is removed, Y and Z must be too" but that is a really nasty piece of logic to put into an already very probelmatic piece of code Aug 01 15:20:39 khem: If I were to look at the problem I'd seriously consider a new staging area and take advantage of gcc's sysroot and differing build-sysroot options Aug 01 15:21:27 mickeyl: You know amethyst appears to be having problems? Aug 01 15:22:41 RP: we could use --with-build-sysroot :) Aug 01 15:22:55 but that wont work with older gcc Aug 01 15:22:56 khem: thats what I mean Aug 01 15:23:20 khem: I'd accept letting the older gcc's break things and doing something sane with newer ones Aug 01 15:23:58 If the breakage bothered people we can tell them to use a more recent gcc... Aug 01 15:23:58 RP: I would think about the idea. But it will break I think even gcc 4.1 Aug 01 15:24:15 khem: hmm, that would be bad :/ Aug 01 15:24:32 I thought you meant 3.3.x as in "older gcc" Aug 01 15:24:45 I thought the build-sysroot option had been around longer Aug 01 15:26:26 RP: hmmm excavation tells me that with-build-sysroot was implemented some 3 years back so gcc 4.1 might work. Aug 01 15:26:41 I'm sure it was even older than that Aug 01 15:26:57 but I could be wrong Aug 01 15:26:58 RP: pb__ for now I will commit the dummy libc.so patch Aug 01 15:27:18 khem: ok, but it is just a bandaid Aug 01 15:27:53 RP: in CSL branch yes it was proposed upstream in 2005 Aug 01 15:28:22 RP: Got it Aug 01 15:28:57 RP: yes. no idea what's going on. traceroute seems to indicate it's something on the way Aug 01 15:29:16 i hope it will vanish alone. until then we might think pushing to one of our backups Aug 01 15:30:29 mickeyl: ok, I'm not too worried yet, I just wanted to make sure you knew! Aug 01 15:33:05 khem: ok Aug 01 15:33:24 mickeyl: the patch that khem is about to commit should fix your ISO8859-1.so issue Aug 01 15:34:36 excellent Aug 01 15:43:14 have a nice weekend guys Aug 01 15:43:41 hrw|gone: You too! Aug 01 15:46:54 What's up with amethyst? Aug 01 15:48:46 what mirrors do we have except ewi? Aug 01 15:50:25 no idea Aug 01 15:50:27 just sent mail about it Aug 01 15:50:35 [to oe-devel and oe-users] Aug 01 15:52:06 we have mirrors? ;) Aug 01 15:52:41 sure Aug 01 15:52:47 for the database at least Aug 01 15:52:52 not for anything else Aug 01 15:53:52 ah right, wolfson mirrors the db Aug 01 15:59:40 hmm Aug 01 15:59:43 seems to be alive again Aug 01 16:00:58 morning Aug 01 16:02:25 g'day kergoth Aug 01 16:08:07 something wrong with monotone server today ? Aug 01 16:08:14 I dont seem to get connection Aug 01 16:08:48 mtn: connecting to monotone.openembedded.org Aug 01 16:08:48 mtn: network error: failed to connect: Connection refused Aug 01 16:09:55 i've just received a mail that says that it is down Aug 01 16:10:17 on openembedded-users Aug 01 16:10:30 * khem cant commit. Leaves for jail Aug 01 16:13:02 khem: commit to mirrors Aug 01 16:13:29 unfortuneately this demonstraits what we lose with git :-( Aug 01 16:15:04 XorA: well, if nobody knows about the capability, it isn't much of a loss Aug 01 16:16:34 hello :) Aug 01 16:16:44 can anyone tell me what provides 'makedevs'? Aug 01 16:17:00 i had to rebuild my machine and can't seem to find the package that provides this.. Aug 01 16:17:27 busybox probably, or perhaps a seperate makedevs package, iirc Aug 01 16:17:55 hmm...kergoth`work there is no makedevs package in gentoo Aug 01 16:18:03 perhaps i can't reuse my old tmp directory Aug 01 16:18:09 kgilmer: er, this isn't the gentoo channel Aug 01 16:18:11 perhaps it's provided as a native package Aug 01 16:18:11 you didnt say anything about gentoo Aug 01 16:18:15 what makes you think this is the place to ask? Aug 01 16:18:15 there is a makedevs package in oe :-} Aug 01 16:18:40 pb__ :) Aug 01 16:18:41 * kergoth`work plays around with the qemu gdb stub Aug 01 16:19:07 i'll try a fresh tmp dir and see what happens. thx! Aug 01 16:19:29 okay, good luck Aug 01 16:21:39 btw who uses makedev ? MAKEDEV seems the debian one Aug 01 16:22:14 also, for those that have any interest in Jalimo and java in OE we did some benchmarking here: http://bugblogger.com/java-vms-compared-160/ Aug 01 16:23:18 ant|work, makedevs is called in do_rootfs() Aug 01 16:23:43 no, I mean 'makedev' seems a relict of 2005 Aug 01 16:23:53 http://gitweb.openembedded.net/?p=org.openembedded.dev.git;a=history;f=packages/makedev;hb=HEAD Aug 01 16:24:12 ah, these long url's... Aug 01 16:24:20 damn, jamvm's startup time is fast Aug 01 16:24:45 ant|work: there's still a place for static device creation Aug 01 16:25:18 and you need .man for embedded ? Aug 01 16:25:20 ant|work: nb, makedevs != makedev Aug 01 16:25:57 pb__: who's in the same family as mknod ? Aug 01 16:26:16 neither, really Aug 01 16:26:34 makedevs is the program that reads device-table.txt and populates /dev from it by calling mknod(2) Aug 01 16:26:46 ok, this one Aug 01 16:26:52 I'm not entirely sure what the makedev package does, it might just contain the MAKEDEV script Aug 01 16:27:03 but, in any case, that is not the package at issue here Aug 01 16:27:24 well when I installed the makedev ebuild the binary that resulted was MAKEDEV. Aug 01 16:27:26 I didn't understand your comment about man pages at all Aug 01 16:27:27 (fyi) Aug 01 16:27:40 kgilmer: ok, makes sense Aug 01 16:27:45 install -m 0644 MAKEDEV.man ${D}${mandir}/man8/MAKEDEV.8 Aug 01 16:28:13 do we ship man now? Aug 01 16:28:15 we don't just leave things out, ant|work. we make more granular packaging. Aug 01 16:28:33 ant|work: yes, in the -doc package, like we always have done Aug 01 16:28:38 ok, I need to read and learn Aug 01 16:29:03 not too much bug & patch activity going on yet Aug 01 16:29:28 well, heh, if bugzilla and mtn are out of order then that isn't too much of a surprise Aug 01 16:29:41 well, I just saw that mail Aug 01 16:29:47 but things are working for me just fine Aug 01 16:29:56 ah, maybe they are back online now Aug 01 16:30:11 bugzilla was broken as recently as about an hour ago, and khem seemed to be having trouble with mtn more recently than that Aug 01 16:30:48 should the repo really point to monotone.oe.org instead of .net? Aug 01 16:32:40 "mtn pull" doing fine here Aug 01 16:33:07 Looks like everything is back to normal Aug 01 16:39:54 XorA: Why should we not be able to use mirrors with git? Aug 01 16:43:13 Laibsch: You can have read mirrors but not multiple push targets. Aug 01 16:43:14 Laibsch, git doesn't support multiple equal heads Aug 01 16:43:27 mtn and hg do Aug 01 16:44:28 * NAiL closes one bug Aug 01 16:48:18 * Laibsch applauds NAiL Aug 01 16:48:34 heh, no changes needed ;) Aug 01 16:48:54 The bug had a .bb for iscsi-target, which I happened to add a couple of weeks ago Aug 01 16:49:00 broonie, jilles: I just asked about this in #git and it sounded like it should be possible to keep on going as we did before Aug 01 16:49:12 XorA: ^^ Aug 01 16:49:36 quoting from #git: (18:47:44) Mikachu: just do a pull from the other server, and if it fails, reset --hard and mail someone Aug 01 16:49:43 of course, just need to set up the right hooks or cron jobs Aug 01 16:50:32 Laibsch: mtn and hg can do that semi automagically. Aug 01 16:58:58 bbl Aug 01 16:59:36 Laibsch: Another round of applause, please :-P Aug 01 17:00:06 * * OE Bug has been RESOLVED by nail(AT)nslu2-linux.org Aug 01 17:00:08 * * speex_1.2-beta3.bb (r0) does not install libspeexdsp Aug 01 17:00:10 * * http://bugs.openembedded.net/show_bug.cgi?id= Aug 01 17:01:12 funky... no bug ID Aug 01 17:10:54 NAiL: stop faking bug solved :-) Aug 01 17:11:55 haha :-P Aug 01 17:12:22 Both of those bugs were already fixed, no work necessary. There are probably more of those... Aug 01 17:12:27 03  07org.openembedded.dev * r8cf27e8076 10OE.dev/packages/linux/ (9 files in 2 dirs): linux omap2 git: update serial hang patches for beagleboard Aug 01 17:12:33 03  07master * r8cf27e8076 10OE.dev/packages/linux/ (9 files in 2 dirs): linux omap2 git: update serial hang patches for beagleboard Aug 01 18:16:09 nail Aug 01 18:16:19 NAiL: Thanks for combing through the list Aug 01 18:16:32 it *does* help, don't mind XorA|gone ;-) Aug 01 18:17:30 broonie: yes, it looks like the support for that kind of thing in mtn and hg is better. From XorA's earlier remark I had gotten the impression that we would not be capable of having mirrors to push and pull going into the future with git Aug 01 18:20:17 Depends on the branching structure, of course. Aug 01 18:20:50 If we wish to use no branches, then it is true that git will have troubles because mirrors are not easy. Aug 01 18:22:45 sometimes life is funny. This morning I wondered if there is anything useful out there to use my pcmcia slot. And now I was happy to be able to use my cf->pcmia adapter for cf cards. Aug 01 18:26:03 * mwester has a Sony memory-stick PCMCIA adaptor, which probably will remain useless (I wish I knew how proprietary the MS devices were *before* I bought them). Aug 01 18:26:05 * * OE Bug 2473 has been RESOLVED (FIXED) by philb(AT)gnu.org Aug 01 18:26:07 * * libgpevtype: wrong depends, fix inside Aug 01 18:26:09 * * http://bugs.openembedded.net/show_bug.cgi?id=2473 Aug 01 18:31:05 * * OE Bug 4189 has been RESOLVED (INVALID) by thommycheck(AT)gmx.de Aug 01 18:31:08 * * Errors about CF FAT Filesystem on boot Aug 01 18:31:09 * * http://bugs.openembedded.net/show_bug.cgi?id=4189 Aug 01 18:36:14 Laibsch, did you find some time to try the collie stuff? Aug 01 18:53:05 * * OE Bug 4456 has been RESOLVED (FIXED) by philb(AT)gnu.org Aug 01 18:53:07 * * DNS Cache Poisoning Bug Aug 01 18:53:09 * * http://bugs.openembedded.net/show_bug.cgi?id=4456 Aug 01 19:26:41 pb_: the patch in 2140 is not applicable? Aug 01 19:26:58 I assume you are philb@ Aug 01 19:29:20 Laibsch: seemingly not, for the reasons given in comment #9 Aug 01 19:32:03 plus, additionally, it doesn't look like it will apply anymore because the commented-out parts of insane.bbclass have now been deleted altogether Aug 01 19:34:54 OK Aug 01 19:35:04 setting dev- keyword for the patch Aug 01 19:42:20 Laibsch: I interpreted XorA|gone's comment to mean that I faked cdbot2's output ;) Aug 01 19:51:06 * * OE Bug 2881 has been RESOLVED (FIXED) by philb(AT)gnu.org Aug 01 19:51:08 * * please change default PATH for bash Aug 01 19:51:10 * * http://bugs.openembedded.net/show_bug.cgi?id=2881 Aug 01 19:54:05 * * OE Bug 4440 has been RESOLVED (FIXED) by philb(AT)gnu.org Aug 01 19:54:07 * * New package: libstroke Aug 01 19:54:10 * * http://bugs.openembedded.net/show_bug.cgi?id=4440 Aug 01 20:08:23 03  07org.openembedded.dev * r9c1207f7ee 10OE.dev/ (2 files in 2 dirs): python-dbus 0.82.4 update to 0.83.0 Aug 01 20:08:24 03  07org.openembedded.dev * rab76ea26f9 10OE.dev/packages/blueprobe/ (blueprobe_0.18.bb files/rx3000.patch): blueprobe: add rx3000 support, patch from roman.moravcik@gmail.com Aug 01 20:08:25 03  07org.openembedded.dev * rf571e0b15f 10OE.dev/: Aug 01 20:08:25 merge of '03e37205ddb308324f3e7b1e882c50510b125716' Aug 01 20:08:26 and '9ed8a9e4b331c4c26542be80abb9775edaf8f5bf' Aug 01 20:08:28 03  07org.openembedded.dev * rde28c8f630 10OE.dev/packages/freesmartphone/frameworkd_git.bb: frameworkd git RDEPENDS += python-subprocess Aug 01 20:08:31 03  07org.openembedded.dev * r58de6d53cf 10OE.dev/packages/python/ (python-dbus_0.82.4.bb python-evas_cvs.bb): python-evas cvs patch has been accepted upstream Aug 01 20:08:34 03  07org.openembedded.dev * r7fd39c4094 10OE.dev/packages/bash/bash_3.2.bb: bash_3.2: apply more patches from upstream Aug 01 20:08:39 03  07org.openembedded.dev * r3a2d86632f 10OE.dev/packages/mesa/ (mesa-dri_6.5.2.bb mesa-dri_7.0.3.bb): mesa-dri: squash bogus dependency on nonexistent package Aug 01 20:08:42 03  07org.openembedded.dev * r1309ff1176 10OE.dev/packages/bash/ (bash_3.0.bb bash_3.2.bb files/default_path.patch): bash: apply patch from clement.hermann@free.fr to set a more sensible default PATH Aug 01 20:08:45 03  07org.openembedded.dev * rbfe7e84a6a 10OE.dev/: Aug 01 20:08:49 merge of 'a061abf4bdeedaccdb10ef3b7372827d261db509' Aug 01 20:08:51 and 'c6e91bf295f3be4e74e18520fdbb8aeb20908d3e' Aug 01 20:08:53 03  07org.openembedded.dev * r04020fe39f 10OE.dev/packages/qt4/ (26 files in 3 dirs): qt4: update qt/e and qt4-x11 to 4.4.1 Aug 01 20:08:58 03  07master * r9c1207f7ee 10OE.dev/ (2 files in 2 dirs): python-dbus 0.82.4 update to 0.83.0 Aug 01 20:09:01 03  07master * r58de6d53cf 10OE.dev/packages/python/ (python-dbus_0.82.4.bb python-evas_cvs.bb): python-evas cvs patch has been accepted upstream Aug 01 20:09:04 03  07master * rf571e0b15f 10OE.dev/: Aug 01 20:09:06 merge of '03e37205ddb308324f3e7b1e882c50510b125716' Aug 01 20:09:08 and '9ed8a9e4b331c4c26542be80abb9775edaf8f5bf' Aug 01 20:09:10 03  07master * rde28c8f630 10OE.dev/packages/freesmartphone/frameworkd_git.bb: frameworkd git RDEPENDS += python-subprocess Aug 01 20:09:15 03  07master * rab76ea26f9 10OE.dev/packages/blueprobe/ (blueprobe_0.18.bb files/rx3000.patch): blueprobe: add rx3000 support, patch from roman.moravcik@gmail.com Aug 01 20:09:18 03  07master * r7fd39c4094 10OE.dev/packages/bash/bash_3.2.bb: bash_3.2: apply more patches from upstream Aug 01 20:09:23 03  07master * r3a2d86632f 10OE.dev/packages/mesa/ (mesa-dri_6.5.2.bb mesa-dri_7.0.3.bb): mesa-dri: squash bogus dependency on nonexistent package Aug 01 20:09:28 03  07master * rbfe7e84a6a 10OE.dev/: Aug 01 20:09:30 merge of 'a061abf4bdeedaccdb10ef3b7372827d261db509' Aug 01 20:09:32 and 'c6e91bf295f3be4e74e18520fdbb8aeb20908d3e' Aug 01 20:09:34 03  07master * r1309ff1176 10OE.dev/packages/bash/ (bash_3.0.bb bash_3.2.bb files/default_path.patch): bash: apply patch from clement.hermann@free.fr to set a more sensible default PATH Aug 01 20:09:39 03  07master * r04020fe39f 10OE.dev/packages/qt4/ (26 files in 3 dirs): qt4: update qt/e and qt4-x11 to 4.4.1 Aug 01 21:39:06 * * OE Bug 4465 has been created by yfan.bugzilla(AT)gmail.com Aug 01 21:39:08 * * Error during bitbake x11-image on Ubuntu 8.04 Aug 01 21:39:10 * * http://bugs.openembedded.net/show_bug.cgi?id=4465 Aug 01 21:44:08 is montone.openembedded.org still having issues ? Aug 01 21:44:18 hm not here Aug 01 21:44:25 I could pull Aug 01 21:44:32 woglinde: I cant do push Aug 01 21:44:49 ah Aug 01 21:44:50 hm Aug 01 21:44:53 not tested yet Aug 01 21:45:06 pull worked Aug 01 21:45:25 yes Aug 01 21:45:36 pull works for me but not push Aug 01 21:47:13 ah now mtn pull stopped working Aug 01 21:47:21 so it must be something latest Aug 01 21:47:37 woglinde: is pull ok for u right now Aug 01 21:50:46 yes I was 10 min ago Aug 01 22:35:10 <_hc> so I am just getting started with OE development, and I was wondering whether other shells beside bash are supported for the dev environment Aug 01 22:35:38 <_hc> I ask, because I am (still) using tcsh, and there is a bug in a script when building gcc-native_3.4.6 Aug 01 22:35:52 <_hc> and I want to know if it is worth me submitting a patch Aug 01 22:39:01 yes it is Aug 01 22:39:07 <_hc> turns out it is a known gcc bug http://mail-index.netbsd.org/pkgsrc-bugs/2006/04/11/0004.html Aug 01 22:39:14 we should not have bashisms in the .bb Aug 01 22:43:02 hi thesing Aug 01 22:44:21 hi again Aug 01 22:46:22 hm I dont have to understand why autotools sometimes copy the newest .m4 and sometimes not Aug 01 22:47:09 <_hc> autotools can be quite a mystery sometimes Aug 01 22:47:15 <_hc> but when it works, its nice... Aug 01 22:47:27 yes Aug 01 22:47:27 <_hc> anyway, gotta run, you'll be seeing more of me around here ;) Aug 01 22:47:47 I am getting more and more a autotool master Aug 01 22:48:29 an Aug 01 22:48:30 even Aug 01 22:51:57 re Aug 01 22:52:09 hi again Aug 01 22:52:24 did you get an anwer from the pdax guy? Aug 01 22:52:24 hi Aug 01 22:52:51 no PM yet... Aug 01 22:53:10 (my first PM got lost...) Aug 01 22:53:27 he says Aug 01 22:54:16 I told him here we have testers for spitz and tosa, besides the 5x00 experiments Aug 01 22:54:36 and told him about initramfs images of you Aug 01 22:54:48 why are you such an uboot fan ? ;) Aug 01 22:55:01 ah, and that we already have good 2.6.26 Aug 01 22:55:11 just for spass Aug 01 22:55:22 uboot rocks Aug 01 22:55:38 my stuff works on desktop now. on collie crashes unfortunately. Aug 01 22:55:50 khem yes Aug 01 22:56:04 khem but to many patches arent included Aug 01 22:56:28 one bootloader that works :) Aug 01 22:56:56 I actually somehow merged this u-boot zaurus in u-boot_git.bb Aug 01 22:57:01 uboot might be cool, but once you have a working bootloader its better to spend time on other stuff imo. Aug 01 22:57:15 but it's unsupported if it's only for 2 machines Aug 01 22:57:40 thesing: I'll need to repartition anyway Aug 01 22:57:50 whats their patch policy? Aug 01 22:57:54 and with u-boot is possible from local.conf Aug 01 22:58:06 like glibc? Aug 01 22:58:39 was GPL IIRC Aug 01 22:58:48 ERROR: log data follows (/disk/d1/oe/build_multi_dev/work/htcuniversal-angstrom-linux-gnueabi/linux-handhelds-2.6-2.6.21-hh20-r20/temp/log.do_configure.16793) Aug 01 22:58:48 | /disk/d1/oe/build_multi_dev/deploy/uclibc/images/htcuniversal// does not exist, you may need to bitbake it separately Aug 01 22:59:00 This loooks bad... Aug 01 22:59:09 * flo_lap builds a glibc image Aug 01 22:59:18 03  07org.openembedded.dev * raa139921a2 10OE.dev/packages/gcc/gcc-cross-intermediate.inc: (log message trimmed) Aug 01 22:59:18 Delete dummy libc.so that was created during glibc-initial step. Aug 01 22:59:18 This library is overwritten by a linker script file called libc.so during Aug 01 22:59:18 complete glibc install so the packaged built after glibc is staged are fine. Aug 01 22:59:18 But the binaries/libraries that build with glibc and depend on libc.so in some Aug 01 22:59:21 form see this dummy libc.so and gets things wrong. Aug 01 22:59:23 This fix deletes the dummy libc.so before gcc-intermediate is staged. Because Aug 01 22:59:25 03  07org.openembedded.dev * rf047d1aabc 10OE.dev/packages/ (26 files in 3 dirs): Aug 01 22:59:28 Move the RRECOMMEND for libgcc from glibc to busybox. Aug 01 22:59:30 Busybox/logger is the place where libgcc is really invoked. Aug 01 22:59:32 03  07org.openembedded.dev * rede8295b83 10OE.dev/packages/expat/ (expat-1.95.7/expat-XMLCALL.patch expat_1.95.7.bb): Aug 01 22:59:35 Consider i386 when making XMLCALL as decl type. Aug 01 22:59:37 This fixes gdb build failures on x86_64 host machines. Aug 01 22:59:41 03  07master * rede8295b83 10OE.dev/packages/expat/ (expat-1.95.7/expat-XMLCALL.patch expat_1.95.7.bb): Aug 01 22:59:44 Consider i386 when making XMLCALL as decl type. Aug 01 22:59:46 This fixes gdb build failures on x86_64 host machines. Aug 01 22:59:50 03  07master * raa139921a2 10OE.dev/packages/gcc/gcc-cross-intermediate.inc: (log message trimmed) Aug 01 22:59:53 Delete dummy libc.so that was created during glibc-initial step. Aug 01 22:59:57 This library is overwritten by a linker script file called libc.so during Aug 01 22:59:59 complete glibc install so the packaged built after glibc is staged are fine. Aug 01 23:00:01 But the binaries/libraries that build with glibc and depend on libc.so in some Aug 01 23:00:03 form see this dummy libc.so and gets things wrong. Aug 01 23:00:05 This fix deletes the dummy libc.so before gcc-intermediate is staged. Because Aug 01 23:00:09 03  07master * rf047d1aabc 10OE.dev/packages/ (26 files in 3 dirs): Aug 01 23:00:11 Move the RRECOMMEND for libgcc from glibc to busybox. Aug 01 23:00:13 Busybox/logger is the place where libgcc is really invoked. Aug 01 23:00:31 flo_lap, another user had the same error earlier this day. Aug 01 23:00:53 only me having issues with gpsd ? http://tinderbox.openembedded.net/public/logs/667783.txt Aug 01 23:01:07 error: invalid Python installation: unable to open /oe/build/tmp/angstrom/staging/arm-angstrom-linux-gnueabi/usr/lib/python2.5/config/Makefile (No such file or directory) Aug 01 23:01:24 ant__, glibc maintainer is known to be quite difficult with patches. I was referring to that. Aug 01 23:01:37 got it Aug 01 23:02:14 ok, let me see who was the last one who touched the hh.o kernel stuff Aug 01 23:02:57 flo_lap: seems like uclibc initramfs Aug 01 23:03:04 (pfalcon) Aug 01 23:04:29 ant__: yeah... it is. that's broken by design... Aug 01 23:04:42 ibot: seen pfalcon? Aug 01 23:04:45 pfalcon was last seen on IRC in channel #handhelds, 401d 3h 6m 2s ago, saying: ';-)'. Aug 01 23:13:40 flo_lap: you could convert it to the initramfs I put into kernel.bbclass Aug 01 23:14:48 thesing: sometimes I really love OE... I just started to think about a sane solution because I need it for a new device :) Aug 01 23:15:35 flo_lap, everything you need to do is set INITRAMFS_IMAGE to the desired image (before the inherit kernel line) and change the kernel config to get the initramfs from $S/initramfs_image.cpio.gz (or so) Aug 01 23:16:08 <_hc> ok, back sooner than I thought. I am trying to do my first build with bitbake, and have gotten 175 packages built, but now its dying on qemu-native_0.9.1.bb, it seems to be missing zlib, since I get linker errors about inflate, inflateEnd, etc Aug 01 23:16:09 I use it for collie so it works ;) Aug 01 23:16:13 <_hc> here's the log: http://rafb.net/p/EpdSFw25.html Aug 01 23:16:20 thesing: do we need something sent for upstream glibc Aug 01 23:16:35 <_hc> I installed zlib-native successfully, but qemu still dies in the same place Aug 01 23:17:28 khem, I think we have some glibc patches laying around, haven't we? Aug 01 23:18:17 tharvey: that's pretty good... can we build these images automatically before building the kernel. or owuld we need som kind of kernel-initial? Aug 01 23:18:40 <_hc> hmm, it seems to be trying to build qemu-native before it builds its DEPENDS, I am just now building libsdl-native manually Aug 01 23:19:53 flo_lap, they get build automaticly. kernel:do_compile depends on INITRAMFS:do_rootfs Aug 01 23:20:01 thesing: I havent looked closely what needs upstream submission Aug 01 23:20:58 tharvey: great! Aug 01 23:21:08 flo_lap, you must take care that you have cpio.gz in FSTYPES or add it to the initramfs-image bb (like I did for initramfs-kexec-image) Aug 01 23:23:52 thesing: we might copnsider to add a class that takes care of these initrd image types Aug 01 23:25:34 flo_lap, its the only change to standard images. I'm not sure if this es enough for an own class. Aug 01 23:25:50 hmm...the gpsd issue disappeared after rebuilding python Aug 01 23:26:27 thesing: hmm.. indeed Aug 01 23:27:26 thesing: flo_lap: you need a different defconfig for that mini-kernel Aug 01 23:28:06 so it could be just one .bb with per-machine defconfigs Aug 01 23:29:26 like e.g linux-kexecboot.bb Aug 01 23:30:50 * flo_lap needs to turn his uber cool initramfs update image into a useful thign for more devices Aug 01 23:32:08 flo_lap, what does you initramfs image do? Aug 01 23:33:41 tharvey: If updates the device kernel/fs from a cf card if images are present and a certain button is pressed at boot. Aug 01 23:34:02 s/tharvey/thesing Aug 01 23:34:21 thesing: hmmm dynamic ldconfig patches might make sense I dont know Aug 01 23:34:30 thesing: most of the patches are backports Aug 01 23:35:28 flo_lap, I'm working on a bootmenu application for use in initramfs. It scans all partitions for kernel, display them in a nice grp Aug 01 23:35:44 menu and kexecs the chosen one. Aug 01 23:35:56 thesing: on a different topic I committed a fix which solves libgcc problem in a bit different way. Let me know if you see more problems Aug 01 23:35:57 so we only need kernels in rootfs. Aug 01 23:36:37 thesing: that's a nice idea Aug 01 23:36:52 khem, you mean the "let busybox depend on libgcc" commit? Aug 01 23:36:53 yes: production kernel != initramfs-kexec-kernel Aug 01 23:37:00 thesing: yes Aug 01 23:37:26 thesing: in my case its a little bit nasty because I still have to use the original bootloader which needs a special image format. Aug 01 23:37:46 khem, I will test if BB_STAMP_POLICY works now when my main pc working again. Aug 01 23:37:49 thesing: actually for oabi glibc depends on libgcc not only for pthread_cancel but also cancellable functions however for EABI it only depends if pthread_cancel is used Aug 01 23:37:58 thesing: ok Aug 01 23:38:30 flo_lap, if the bootloader can load a kernel everything should be fine. Aug 01 23:38:42 khem, thanks for digging into the libgcc issue. Aug 01 23:38:53 thesing: it was needed Aug 01 23:39:26 thesing: we also have another bigger problem at hand which is rebuilding toolchain components not from scratch Aug 01 23:39:52 right now with all this bootstrap stuff packages end up shooting each other Aug 01 23:40:00 when rebuilding Aug 01 23:40:23 I alway wondered how this worked with packaged staging. Aug 01 23:40:37 and now it turns out that it doesn't ;) Aug 01 23:40:48 thesing: yes, luckily it does support this... but u-boot and dfu would be muchh better than this custom crap Aug 01 23:41:13 he, sane words Aug 01 23:41:36 flo_lap, I have to use the sharp bootloader. Do I have to say any more? ;) Aug 01 23:41:54 thesing: heh ok :) Aug 01 23:42:00 alien vs. predator Aug 01 23:44:01 hm. my laptop isn't really a compile monster. I takes me nearly the whole day to build a console image. Aug 01 23:44:24 some pcmcia card with additional cpus would be great. Aug 01 23:45:53 * thesing imagines a cluster of a laptop some pdas and some routers. Aug 01 23:46:54 flo_lap, how do you flash from the initramfs? dd? Aug 01 23:47:13 thesing: BTW which recipe should I check to build nandlogical.c against klibc? Aug 01 23:47:39 ant__, check kexec-tools-static Aug 01 23:47:45 ok, thx Aug 01 23:48:27 tharvey: no, I need to use a custom tool similar to nandwrite since the device uses a type of nand hcip that does not guarantee any block to be ok (except of the first) Aug 01 23:49:23 hmm, I put DEPENDS = "mtd-utils" because it was missing some header... Aug 01 23:50:04 ant__, it won't help. klibc uses its own headers. Aug 01 23:50:28 but I need mtd-utils too... Aug 01 23:50:58 ant__, and you need to compile them separately. Aug 01 23:51:52 actually nandwrite is essential Aug 01 23:52:09 for that update/emergency-image Aug 01 23:52:42 but as you said one could do it from SD Aug 01 23:54:52 thesing: http://bugs.openembedded.net/attachment.cgi?id=9850 Aug 01 23:55:04 it was missing +#include Aug 01 23:55:47 missing = missing header Aug 01 23:55:53 good night Aug 01 23:56:16 nite Aug 01 23:57:02 ant__, klibc should have this header. Aug 01 23:57:59 ant__, I usually compile programs that I port to klibc outside of oe for i386 and modify it until it builds. Aug 02 00:00:32 nite all. Aug 02 00:04:47 ant__: mtd-user.h should come from kernel-headers Aug 02 00:05:33 hmm...it was failing Aug 02 00:06:16 ant__: you should have linux-libc-headers installed I guess Aug 02 00:06:26 http://tinderbox.openembedded.net/public/logs/575588.txt Aug 02 00:08:05 ant__: what kernel headers is it using Aug 02 00:08:43 khem: it was created couple of years ago (kernel 2.6.18 IIRC) Aug 02 00:09:34 I just applied an hack from utx and got it compile with 2.6.24 Aug 02 00:09:41 ant__: All I have here is 2.6.24 Aug 02 00:09:45 but the headers, let me see Aug 02 00:10:01 so I can not tell if 2.6.18 exported this header to userspace or not Aug 02 00:10:51 khem: linux-libc-headers 2.6.23 r3 Aug 02 00:13:29 hmmm it looks like in oe does not install mtd headers Aug 02 00:14:34 khem: against glibc the size is 559977 (547k), urgh Aug 02 00:14:42 static Aug 02 00:15:09 still better glibc itself is above 1.5M Aug 02 00:15:30 he he, without u-boot the space is 1,2Mb Aug 02 00:15:41 for kernel + initramfs image Aug 02 00:15:52 ant__: you should use uclibc Aug 02 00:16:07 too big image (wait) Aug 02 00:16:10 or eglibc once I put the kconfig support in eglibc builds Aug 02 00:16:55 -rw-r--r-- 1 andrea users 124K Jun 2 02:44 Angstrom-initramfs-bootmenu-image-glibc-ipk-2008.1-test-20080601-c7x0.rootfs.cpio.gz Aug 02 00:16:55 -rw-r--r-- 1 andrea users 2.6M Jun 4 08:46 Angstrom-initramfs-bootmenu-image-glibc-ipk-2008.1-test-20080604-c7x0.rootfs.cpio.gz Aug 02 00:17:20 actually was built with uclibc, sorry for the naming Aug 02 00:18:45 khem: the 124k image is the klibc Aug 02 00:18:54 :-] Aug 02 00:19:17 thesing added today 50k of menu's in C Aug 02 00:20:00 I'd like to see if mtd-utils and nandlogical will ever fit... Aug 02 00:24:37 ant__: uclibc will be about double of that Aug 02 00:24:57 but i think uclibc has more applications support Aug 02 00:25:03 khem, it contains busybox Aug 02 00:25:08 no Aug 02 00:25:16 busybox is extra Aug 02 00:25:24 I mean the uclibc image is bloated by busybox Aug 02 00:26:05 but this was in June, now let see how thesing surprises us... Aug 02 00:26:29 ant__: what does busybox has to do with uclibc size Aug 02 00:26:37 image size Aug 02 00:26:53 not uclibc size Aug 02 00:28:11 busybox = 1631012 Aug 02 00:28:13 ah you can strip it out if you do not need it Aug 02 00:28:32 that's the point: thesing rewrote the bootmenu in C Aug 02 00:28:43 uclibc is around 500K Aug 02 00:29:14 you can have something like dash.static Aug 02 00:29:29 as shell and boot into it init=/bin/dash.static Aug 02 00:29:38 ash in our case Aug 02 00:29:48 yeh Aug 02 00:36:16 khem: should I upgrade the headers if I move to 2.6.26 ? Aug 02 00:37:24 03  07org.openembedded.dev * r477f652329 10OE.dev/packages/otpcalc/ (7 files in 2 dirs): otpcalc: add initial version 0.97 based on work from Yuri Bushmelev. Closes 4441. Aug 02 00:37:30 03  07master * r477f652329 10OE.dev/packages/otpcalc/ (7 files in 2 dirs): otpcalc: add initial version 0.97 based on work from Yuri Bushmelev. Closes 4441. Aug 02 00:42:29 gn Aug 02 00:46:45 Laibsch: NAiL sorry my attempt at humour Aug 02 00:47:52 * Laibsch changes batteries in humour detector Aug 02 00:48:00 Laibsch: I am currently under the influence if so much alcohol I can carely stange kick me tomorrow Aug 02 00:48:13 s/stange/stand/ Aug 02 00:48:30 I'll shout right atcha Aug 02 00:48:34 ooops Aug 02 00:48:35 You'll love it ;-) Aug 02 00:48:39 * XorA|gone misunderstands Aug 02 00:49:40 ant__: not necessarily Aug 02 00:57:06 * * OE Bug 4438 has been RESOLVED (FIXED) by Aug 02 00:57:08 * * Bump tangogps to 0.9.2 , consolidate recipes Aug 02 00:57:10 * * http://bugs.openembedded.net/show_bug.cgi?id=4438 Aug 02 01:02:23 03  07org.openembedded.dev * rbb39eaa21c 10OE.dev/packages/tangogps/ (tangogps.inc tangogps_0.9.0.3.bb tangogps_0.9.2.bb): Aug 02 01:02:23 tangogps: update to 0.9.2. Closes 4438. Aug 02 01:02:23 * the fso version should probably be merged back in. leaving as separate package for now. Aug 02 01:02:26 03  07master * rbb39eaa21c 10OE.dev/packages/tangogps/ (tangogps.inc tangogps_0.9.0.3.bb tangogps_0.9.2.bb): Aug 02 01:02:26 tangogps: update to 0.9.2. Closes 4438. Aug 02 01:02:27 * the fso version should probably be merged back in. leaving as separate package for now. Aug 02 01:05:36 XorA|gone: ah - so alcohol is the real reason for the deletions on buildhost ;-) Aug 02 01:12:07 * * OE Bug 4271 has been RESOLVED (FIXED) by Aug 02 01:12:09 * * new package - oneko Aug 02 01:12:10 * * http://bugs.openembedded.net/show_bug.cgi?id=4271 **** ENDING LOGGING AT Sat Aug 02 02:59:57 2008