**** BEGIN LOGGING AT Tue Oct 14 02:59:57 2008 Oct 14 04:18:33 #splashtop Oct 14 06:58:27 morning Oct 14 06:58:38 morning. is there a way when i build a image that a certain package gets build again? like DEPENDS or RDEPENDS# Oct 14 06:59:01 when building minimal image and adding DEPENDS toolchain-sdk the sdk does not get packed Oct 14 06:59:02 bitbake -crebuild foo Oct 14 06:59:05 i assume? Oct 14 06:59:12 morning Oct 14 06:59:40 gilligan not really :) Oct 14 07:05:24 --force? :) Oct 14 07:56:11 bonjour Oct 14 07:56:33 Morning Oct 14 07:56:48 Any one going to CELF? Oct 14 08:11:59 morning Oct 14 08:12:03 morning Oct 14 08:12:06 Esben: s/CELF/ELC you mean? Oct 14 08:12:19 yes Oct 14 08:12:25 6-7 november Oct 14 08:12:46 -> *ibot* seen woglinde Oct 14 08:12:51 I was wondering if a number of OE developers are going to stay at a particular hotel. Oct 14 08:12:58 argh, sorry :) Oct 14 08:13:00 if so, I would like to choose the same hotel :-) Oct 14 08:13:28 ~seen wogline Oct 14 08:13:29 i haven't seen 'wogline', hrw Oct 14 08:13:32 ~seen woglinde Oct 14 08:13:33 woglinde was last seen on IRC in channel #oe, 12h 58m 44s ago, saying: 'bye till later'. Oct 14 08:13:36 is better way mmp Oct 14 08:15:46 hrw: I know, I just didn't want to leave clutter on channel :) Oct 14 08:16:08 so /msg'ing ibot directly was the first choice Oct 14 08:16:41 good morning Oct 14 08:41:11 morning Oct 14 08:41:21 moin henning ;) Oct 14 08:44:51 hi woglinde, hi thebohemian Oct 14 08:50:32 03  07org.openembedded.dev * r6271edc9f0 10org.openembedded.dev.git/contrib/mtn2git/manglenames.py: mtn2git: add current name and email mapping script Oct 14 09:30:59 I am building jpeg-6b with EXTRA_OECONF="--enable-shared", during compile it is saying .." ./libtool: Command not found". libtool is already compiled.can anybody suggest what could be the problem .. Oct 14 09:36:01 nakul_: another library that has problems with libtool2 Oct 14 09:36:45 03  07org.openembedded.dev * ree956ebf7a 10org.openembedded.dev.git/packages/xorg-driver/xf86-video-omapfb_git.bb: xf86 video omapfb: bump SRCREV for YV12 support Oct 14 09:38:23 thebohemian: I think it should generate 'libtool' @ configure time. & that libtool is expected when it says ./libtool .. but libtool is not being generated.. Oct 14 09:39:08 i can successfully get openswan compiling & installing but don't know how to have kernel modules... any tips ? Oct 14 09:41:28 nakul_: look at the makefiles and find out which variable contains ./libtool Oct 14 09:41:45 LIBTOOL Oct 14 09:42:00 then add "VARIABLE=${TARGET_SYS}-libtool" to the oe_runmake call in your recipe Oct 14 09:42:52 if there is no oe_runmake you may extend EXTRA_OEMAKE Oct 14 09:43:15 nakul_: I assume that there is a ${TARGET_SYS}-libtool in the source dir Oct 14 09:44:10 I am not getting ..plz explain Oct 14 09:44:11 03  07org.openembedded.stable * r06ab83ea93 10org.openembedded.dev.git/packages/ (8 files in 5 dirs): Oct 14 09:44:11 opie-mediaplayer1: Add 1.2.3 versions of mediaplayer1 and plugins Oct 14 09:44:11 Applied to .dev in: 03740c225501e66bf2a7218536868643b53812ce Oct 14 09:44:11 03  07org.openembedded.stable * r6fdae89396 10org.openembedded.dev.git/packages/portabase/metakit_2.4.9.3.bb: Oct 14 09:44:14 metakit: fix fetch path & library install/packaging Oct 14 09:44:16 Applied to .dev in: f2ebdb1826372581987a96a662d30d97b544c438 & Oct 14 09:44:18 c9cfaab10b1cc5547009c788d31c5e670d194bc9 Oct 14 09:44:20 03  07org.openembedded.stable * rda8db62f8c 10org.openembedded.dev.git/: Oct 14 09:44:22 merge of '0fa016c65abfeab8646d405ea9db417c1c179b26' Oct 14 09:44:26 and '68085549871330fb9fdd18af4cb53d3634c2bd36' Oct 14 09:44:28 03  07org.openembedded.stable * r6f7b777f96 10org.openembedded.dev.git/packages/ (6 files in 3 dirs): Oct 14 09:44:31 beecrypt: merge duplicate recipes Oct 14 09:44:33 * Merge do_stage from portabase/beecrypt into beecrypt Oct 14 09:44:35 * Delete portabase/beecrypt (patches no longer seem to be necessary) Oct 14 09:44:37 Applied to .dev in: 88676b03209e4a8147b6f9ab4d6f5c66d0132128 Oct 14 09:44:39 03  07org.openembedded.stable * rdbadae189c 10org.openembedded.dev.git/conf/distro/angstrom-2007.1.conf: angstrom 2007: bump to r19 Oct 14 10:09:44 hi mickeyl Oct 14 10:10:00 nakul_: sorry, pidgin crashed Oct 14 10:10:08 (and I didn't noticed) Oct 14 10:10:08 moin hrw Oct 14 10:10:13 hey hrw Oct 14 10:10:21 hrw: how are things? Oct 14 10:10:44 mickeyl: tomorrow last day with OH Oct 14 10:11:19 mickeyl: I am discussing with next company - Poky still rather then OE. Oct 14 10:11:20 hrw: i see. too bad. i hope you'll find something equally productive Oct 14 10:11:28 oh, sounds good Oct 14 10:11:32 what's good for poky is good for OE Oct 14 10:11:35 imo Oct 14 10:11:50 there are lot of things in poky worth adding to OE Oct 14 10:11:51 [if you will commit again ;)] Oct 14 10:12:11 mickeyl: I am waiting for git so next week will start merging Oct 14 10:12:16 super Oct 14 10:12:49 do we support user branches on git.oe.org? Oct 14 10:13:58 as maemo changes are big, also other ones etc Oct 14 10:16:40 ~curse git.freedesktop.org Oct 14 10:16:40 May you be reincarnated as a Windows XP administrator, git.freedesktop.org ! Oct 14 10:30:27 mickeyl: good morning Oct 14 10:30:48 well, afternoon I suppose Oct 14 10:31:43 pb__: morning Oct 14 10:31:51 ya, woke up pretty early (8 something) today Oct 14 10:32:00 pb__, should pb@reciva.com be used as your git id? Oct 14 10:32:05 guess i still didn't arrive in my timezone ;) Oct 14 10:32:06 yo shoragan Oct 14 10:32:08 mickeyl: heh, that's not like you Oct 14 10:32:10 k Oct 14 10:32:15 shoragan: I would prefer philb@gnu.org Oct 14 10:32:52 pb__, ok, changed Oct 14 10:32:56 thanks Oct 14 10:33:21 florian_: good morning Oct 14 10:33:51 re Oct 14 10:34:46 shoragan: which one I used in a list? Oct 14 10:34:50 ok, i should now have nearly all email addresses correct Oct 14 10:35:10 "Marcin Juszkiewicz" : (["Marcin", "Marcin Juszkiewicz ", "Marcin Juzkiewicz ", "hrw@openedhand.com", "hrw@openembedded.org", "local!szczepan", "hrw.one.pl!szczepan", "hrw.one.pl!hrw", "local!hrw"], "openembedded@hrw.one.pl") Oct 14 10:35:42 the list is now in MTN contrib/mtn2git/manglenames.py Oct 14 10:37:17 shoragan: please change main address to hrw@openembedded.org - my monotone databse is broken Oct 14 10:37:23 hi pb Oct 14 10:37:39 pb I fixes the e2fsprogs problem Oct 14 10:37:44 fixed Oct 14 10:40:54 hrw, so hrw@oe.org is a valid address? Oct 14 10:42:30 it was last time Oct 14 10:43:10 Hello. I receive an error if I try to build Angstrom distribution from "org.openembedded.stable" for HTC Blueangel. I have found that this bug is reported at "http://bugs.openembedded.net/show_bug.cgi?id=4465" already. How is possible that there is a successful build for Blueangel (http://www.angstrom-distribution.org/unstable/autobuild/htcblueangel/) while this bug is remaining? Oct 14 10:47:11 drasar stable vs. unstable~ Oct 14 10:48:03 hrw, did you get my test mail? Oct 14 10:50:05 shoragan: @oe.org was always slow... Oct 14 10:50:12 k Oct 14 10:50:21 shoragan: and I do notknow for how long I will keep @hrw.one.pl and @haerwu.biz Oct 14 10:50:44 woglinde: Yes, but "2007.12-r18" that is built by autobuild on that link is just in stable, right? Oct 14 10:51:43 drasar hm where is it failig for you? Oct 14 10:55:58 woglinde: While compiling "openmoko-dialer2". It is the same problem as I mentioned Oct 14 10:57:18 drasar hm you could try to update the openmoko-dialer2 src rev and try again Oct 14 10:57:39 woglinde: "openmoko-dialer2" is compiled while building "bitbake x11-image" Oct 14 10:59:00 drasar in the angstroem stuff is base and console image Oct 14 10:59:07 no xstuff and no openmoko stuff Oct 14 10:59:12 thats the answer Oct 14 10:59:58 woglinde: How can I do it? I have proceeded according to "http://www.angstrom-distribution.org/building-%C3%A5ngstr%C3%B6m" Oct 14 11:00:50 woglinde: Yes I know, but I need x11-image Oct 14 11:01:24 drasar edit conf/distro/include/sane-srcrevs.inc and try the version which is in .dev Oct 14 11:01:32 I am to lunch now Oct 14 11:04:40 woglinde: I modified the x11-image.bb, and removed the openmoko-dialer2 (replace it with empty item) as is mentioned at "http://bugs.openembedded.net/show_bug.cgi?id=4465#c3" Oct 14 11:19:29 03  07org.openembedded.dev * reea3d090db 10org.openembedded.dev.git/contrib/angstrom/build-feeds.sh: angstrom feed builder: cut down on excessive build attempt by building packages only for needed archs Oct 14 11:19:32 03  07org.openembedded.dev * rc6cb0f213a 10org.openembedded.dev.git/contrib/mtn2git/manglenames.py: mtn2git: fix some more email adresses Oct 14 11:19:34 03  07org.openembedded.dev * r2752c1ab1a 10org.openembedded.dev.git/packages/gdal/gdal_1.5.2.bb: gdal: update to 1.5.2 Oct 14 11:19:35 03  07org.openembedded.dev * r26841634a7 10org.openembedded.dev.git/: Oct 14 11:19:35 merge of '2de185389e4f5cf5e749eb1645efe691e9fad4f8' Oct 14 11:19:36 and 'e6fa426a909123cbd9d0a3fdf92a16a53c9722e9' Oct 14 11:19:38 03  07org.openembedded.dev * rf8e1f4186f 10org.openembedded.dev.git/contrib/mtn2git/manglenames.py: mtn2git: fix some more adresses Oct 14 11:19:46 03  07org.openembedded.dev * rc1f4d97cbf 10org.openembedded.dev.git/: Oct 14 11:19:46 merge of '8be7406d131a307bb409ca173b6b0bb55c092bc6' Oct 14 11:19:47 and 'd86d6413edacca9f55683204f6ec67098e2de319' Oct 14 11:19:49 03  07org.openembedded.dev * r3ce0dd713d 10org.openembedded.dev.git/: Oct 14 11:19:51 merge of '3a2d3525d6188443c195ea295aa262c5e3dccb74' Oct 14 11:19:53 and '40fbbc76fa2b3612027b83b97b4e703d113eaa3c' Oct 14 11:23:27 hrw, did you receive the mail? Oct 14 11:25:42 03  07org.openembedded.dev * r6e3f91bcca 10org.openembedded.dev.git/contrib/mtn2git/manglenames.py: mtn2git: update Phil Blundell's address Oct 14 11:33:30 is there a patch for the "error: 'struct sk_buff' has no member named 'nh'" when you compile openswan for kernel ? (2.6.22) Oct 14 11:37:27 gilligan_, you there? Oct 14 11:37:45 shoragan: got reply? Oct 14 11:38:03 hrw, not yet Oct 14 11:41:48 re Oct 14 11:42:33 woglinde: So I have tried "openmoko-dialer2-1_0.1.0+svnr4343-r6" from dev and received the same error Oct 14 11:44:36 03  07org.openembedded.dev * r97e075cc1a 10org.openembedded.dev.git/packages/corsair/ (corsair_0.2.6.bb files/corsair-opie.patch): corsair: fix compile for qte/qte-mt selection Oct 14 11:44:53 woglinde: But removing this package from "x11-image.bb" should be enough for me ;) Oct 14 11:50:23 I wonder why x11-image depends on openmoko-dialer2 at all Oct 14 11:52:31 hrw yes I wonder too Oct 14 11:53:19 hrw but for the htc it would make sense Oct 14 11:53:57 no Oct 14 11:54:02 hm Oct 14 11:54:08 it should be x11-phone-image for it Oct 14 11:54:15 is the dialer this much shitty? Oct 14 11:54:38 dialer adds gsmd, openmoko libs etc Oct 14 11:56:48 tracking x11 git versions suxx Oct 14 11:57:54 03  07org.openembedded.dev * r313883bd4d 10org.openembedded.dev.git/contrib/mtn2git/manglenames.py: mtn2git: add Jordan Crouse's email Oct 14 11:57:57 03  07org.openembedded.dev * rcd529ac20d 10org.openembedded.dev.git/: Oct 14 11:57:57 merge of '4814b69151af2e49760bb9a3837c26ca674a7636' Oct 14 11:57:57 and '6d598225ef381f53c7c567f433cb8223d13f098e' Oct 14 12:02:07 guys, is DEPENDS also valid for the do_unpack stage? I use rpm2cpio to extract the contents of a .src.rpm, I added a dependency on rpm2cpio-native and on cpio-native, but it seems that the unpack stage of my recipe is triggered before those two dependencies are built Oct 14 12:02:17 hrw, ok, got it Oct 14 12:02:29 shoragan: moo Oct 14 12:02:35 zecke, yoo Oct 14 12:02:47 shoragan: when should we begin? Oct 14 12:03:05 zecke, i'm doing a dryrun right now Oct 14 12:03:09 RP: have you seen the pyflakes result? Oct 14 12:03:25 shoragan: ah you applied the existing patches to the maillist? Oct 14 12:03:30 yes Oct 14 12:04:07 zecke, the list is now in mtn Oct 14 12:05:08 out of my reach then :) Oct 14 12:05:37 oh Oct 14 12:05:49 shoragan: so a gitosis setup is the only missing bit? Oct 14 12:06:02 maybe, no idea Oct 14 12:06:22 i've got my own gitosis setup Oct 14 12:07:42 shoragan: currently it is named gittrial :) Oct 14 12:08:03 shoragan: I wonder if we should have a precommit hook to forbid merging of old and new history :) Oct 14 12:08:31 zecke, can you do that? Oct 14 12:09:00 shoragan: merge? or forbid? Oct 14 12:09:24 forbit it in a precommit hook Oct 14 12:09:32 would be nice to have Oct 14 12:09:39 shoragan: I assume we could... Oct 14 12:10:08 zecke: RP is on conference, but I forwarded that request to him Oct 14 12:10:54 hi Cliff Oct 14 12:11:24 hrw: thanks, tell him this can wait :) Oct 14 12:12:24 ok, fast-importing now, then i'll repack and push to my gitosis so you can have a look Oct 14 12:12:29 zecke: he probably will add it to long list of things to do Oct 14 12:14:21 there is one thing which I will miss after working for OH Oct 14 12:14:27 Poky autobuilder Oct 14 12:14:33 hrw: the long list of things to do? Oct 14 12:15:07 hrw: running a buildbot setup is not too difficult, I can have a all red waterfall of OE :) Oct 14 12:15:37 zecke: I have buildbot config on one of my home machines - will need to setup it on fastest one Oct 14 12:16:08 zecke: [14:15:58] hrw: ouch, that pyflakes looks ugly Oct 14 12:17:26 hrw: the cool thing about the poky buildbot is, if things are red you fix them :) Oct 14 12:17:44 * zecke continues to hack on prelink... Oct 14 12:18:25 zecke: if I will get customer for OE builds I will check them Oct 14 12:18:40 zecke, what do you want to try with prelinking/ Oct 14 12:19:46 shoragan: prelink on building the rootfs (I think I know why SuSE is not enabling prelink) Oct 14 12:19:56 shoragan: the other is to use michael meeks -Bdirect emulation Oct 14 12:20:41 what's -Bdirect? Oct 14 12:21:01 re Oct 14 12:21:35 shoragan: it is a sun linker option, michael meeks implemented something like this for GNU... I would google now as well Oct 14 12:22:52 http://sourceware.org/ml/binutils/2006-06/msg00418.html alternative Oct 14 12:23:14 http://sourceware.org/ml/binutils/2005-10/msg00436.html meeks foo Oct 14 12:29:39 zecke: good morning Oct 14 12:32:27 pb__: morning master, did you play with any of prelink, or binutils patches to speed up resolving of symbols? Oct 14 12:33:54 zecke: I did some stuff with prelink a few years ago, but nothing recently. I don't think I've ever looked at the binutils hashing stuff in any detail. Oct 14 12:34:48 The Meeks -Bdirect thing is, if I understand it correctly, probably not a good idea in general because it will change the symbol binding semantics Oct 14 12:34:48 pb__: prelink is a huge win, sadly it changes the binaries and needs to be rerun after upgrades... Oct 14 12:35:34 zecke: yah, indeed. I did think at one point about doing build-time "pre-prelinking" but I don't think I ever got anywhere with that. Oct 14 12:35:35 pb__: yes, ulrich is quite straight on this topic. the gnu hash looks like a good idea Oct 14 12:36:06 in principle there's no reason why you couldn't run prelink offline, so long as you can predict with reasonable accuracy what binaries will be installed. Oct 14 12:36:07 pb__: ah... build time... I will start with prelinking my rootfs Oct 14 12:36:25 obviously that doesn't work in the general case with a desktop type system, but in an embedded environment you generally have rather tighter control over versions. Oct 14 12:36:38 pb__: right, as long as you use the same cache and your address space is big enough you can do it at buildtime for everything Oct 14 12:36:52 in particular, if your whole rootfs is a single readonly cramfs (or whatever) you should be able to prelink the whole lot offline with no problems. Oct 14 12:37:23 if you are allowing people to do piecemeal upgrades of individual packages then things are slightly trickier, but the linker is smart enough that it will disregard prelink information when the binaries don't match. Oct 14 12:37:44 so, in the worst case, users with "inconsistent" systems will just not see any prelink benefit until they have everything back in sync. Oct 14 12:38:48 pb__: right, maybe I should just add a post inst prelink target to opkg... I'm batlling in emulating ld-linux.so to find and resolve libraries to have a cross prelink Oct 14 12:40:18 zecke, could we instead just link each lib to a unique address at build time? Oct 14 12:40:47 shoragan: heh, this is what we used to do in 1992. Oct 14 12:41:02 pb__, so why is it a bad idea? Oct 14 12:41:38 because it requires you to centrally coordinate all your library load addresses to make sure they don't overlap. Oct 14 12:41:52 this is how a.out shared libraries worked in days before elf Oct 14 12:41:53 we have more libraries now ... Oct 14 12:42:02 shoragan: eeek, I don't want to use a.out anymore... Oct 14 12:42:05 I'm happy with ELF Oct 14 12:42:35 one of the main drivers for the linux binary format change from a.out to elf was that the old shared library system just didn't scale. Oct 14 12:43:10 yay! a.out! Oct 14 12:43:21 * raster pulls out the sarcasm Oct 14 12:43:24 libc4 :) Oct 14 12:43:24 in theory it would be possible to do something like that at the DISTRO level, but it would still be a nightmare. remember that libraries can change in size from one version to the next, so you might need to keep giving them completely different address ranges. Oct 14 12:43:43 all in all, it is a pretty bad idea. Oct 14 12:43:43 heh Oct 14 12:43:51 http://rafb.net/p/Bln8J762.html do you guys think this is a good idea? Oct 14 12:43:59 prelink basically gives you that benefit but without the administrative nightmare. Oct 14 12:44:14 prelink would be good for readonly images Oct 14 12:44:18 pb__: lies... libraries only get smaller and faster :) Oct 14 12:44:22 as they will never get updates Oct 14 12:44:44 zecke: ah yes, I forgot you do not use gnome Oct 14 12:45:04 I guess a.out would be fine for kde :-) Oct 14 12:45:29 pb__: actually I have a gnome-terminal, pidgin and such things running Oct 14 12:45:44 * pb__ recoils from the screen in shock Oct 14 12:45:52 pb__: but I think version numbers in GNOME world just say how much more space and memmory your require :) Oct 14 12:46:11 pb__: I thought gtk+-2.0 means you neeed now twice as much resources :) Oct 14 12:46:12 zecke: right, it is part of a unified marketing message with intel and the computer vendors. Oct 14 12:46:29 big corporates and GNOME :) Oct 14 12:46:36 the idea is that the cpu model number, ram and disk size will all be synchronised with the gnome release number. Oct 14 12:46:59 i thought the deal was exclusively made with ms regarding that? Oct 14 12:47:15 or like, gnome being a copycat :) Oct 14 12:47:46 * pb__ stabs evolution Oct 14 12:47:58 why did they implement this stupid feature of turning the preview window off whenever it crashes? Oct 14 12:48:16 pb__: I have to use outlook, do you think that's a steup up on the ladder? :) Oct 14 12:48:18 wouldn't be so bad if they had made it crash less often at the same time, I suppose Oct 14 12:48:26 pb__: :) Oct 14 12:48:32 BlindMan: heh. well, sometimes I do wonder. Oct 14 12:48:58 zecke: regarding your patch, I find myself wondering whether the whole PE business is somewhat misconceived. Oct 14 12:49:32 PE seems like another thing that is basically just a matter of distro policy, and I kind of feel it belongs in the distro config files. Oct 14 12:50:19 i.e. if the 4ng5tr0m people want to bump PE on mythtv or something, they can set PE_mythtv = "500" in their distro configuration without bothering other people. Oct 14 12:51:10 pb__: I think we really have more than one epoch. One is the classic use case for epoch 1.0 > 1.0-beta1, the other is some kind of policy change in the distro... Oct 14 12:51:20 the only case where a global, cross-distro PE thing would seem to make sense is where the upstream version number really did go backwards, and that happens very rarely indeed. Oct 14 12:51:37 99% of the time, PE is just used as a way to recover from some packaging screw-up that has affected some or other distro. Oct 14 12:51:52 pb__: e.g. if I recompile everything with new flags I want everything to be recompiled... Oct 14 12:52:02 right, indeed Oct 14 12:52:08 beware of libx11 1.1.9x Oct 14 12:52:13 or if you recompile against a different set of shared libraries, or some such Oct 14 12:52:23 even then, arguably that is more a PR change than PE. Oct 14 12:52:25 zecke: I though 1.0 > 1.0-beta1 is solved by PV="0.9+1.0-beta" :) Oct 14 12:52:34 thebohemian: yes Oct 14 12:52:44 pb__: I just saw that the packagemanager actually has no idea about PE Oct 14 12:52:53 pb__: it is just us prepending a number?! Oct 14 12:53:12 or actually the package managers knows %d: Oct 14 12:53:13 zecke: ipkg certainly did use to understand epochs, and I imagine opkg does too. Oct 14 12:53:20 it would be pretty weird if the openmoko folks had removed that feature Oct 14 12:53:32 pb__: sorry, my confusion Oct 14 12:54:10 pb__: so in package manager land we have 1x epoch support, in OE land we have two cases where we would like to change the epoch... Oct 14 12:54:29 pb__: and what is your thought on mixing these two things? Oct 14 12:55:15 zecke: I think my position is that epoch changes should be reserved solely for recovering from version number mistakes, e.g. accidentally releasing a package with a version number that is out of sequence and now sorts higher than you wanted. Oct 14 12:55:32 whether that happened because of a "1.0-beta3" kind of thing, or because upstream actually changed their mind about versioning, is more or less irrelevant. Oct 14 12:55:52 the other case, where you are just recompiling with different flags, seems like it should be handled either by a PR bump, or by an architecture change. Oct 14 12:56:14 or fix ipkg/opkg to get 1.0~beta3 < 1.0 as it is in dpkg? Oct 14 12:56:28 and then we can use PV="1.0~beta3" Oct 14 12:56:42 both opkg and ipkg understand architectures as having a range of attractiveness, so you can have libkde3_1.0.0_arm+zecke.ipk built with your own eleet options and ipkg will understand it is better than the regular arm version. Oct 14 12:57:02 it will, I think, even upgrade from one to the other automatically without the version number having to change. Oct 14 12:57:13 pb__: okay, architecture change... so you would put build config the archnames Oct 14 12:57:30 pb__: armv4t+zecke's-magic om-gta02+zecke's-magic? Oct 14 12:57:33 zecke: depends on the situation, either in the architecture name or just as a component of PR. Oct 14 12:57:52 pb__: changing the PR globally is rather difficult Oct 14 12:58:06 arguably the biggest problem in this space is that DISTROs have no way to influence PR without fiddling every individual .bb file. Oct 14 12:58:17 I think that is probably my one biggest complaint about the oe metadata structure. Oct 14 12:58:23 pb__: e.g. the only non obvious thing would be to have an "auto increment pr" Oct 14 12:59:04 yeah, or build it up from a set of components. PR = "${DISTRO_PR}.${FILE_PR}" or some such. Oct 14 12:59:44 pb__: the pro's of abusing PE, one single line to change for a global upgrade, easy upgrade path... but I understand your point Oct 14 12:59:59 the problem with PE is that once you use it, you can't really go back. Oct 14 13:00:01 pb__: yes, I have proposed something like this for PV Oct 14 13:00:17 PR gets effectively reset on every PV bump, so it is much more transient in that sense. Oct 14 13:00:39 pb__: you are right Oct 14 13:03:46 looks like I will have a go at poking PRs... Oct 14 13:03:55 ok, very good Oct 14 13:08:14 http://git.sicherheitsschwankung.de/?p=jan/oetest.git;a=summary Oct 14 13:08:19 this is my dryrun Oct 14 13:08:36 everyone please check if it looks sane :) Oct 14 13:08:51 send a mail :) Oct 14 13:09:23 :p Oct 14 13:11:04 shoragan: dreambox was not updated for 18month? that does not look right :) Oct 14 13:11:35 where should i pull to get the most recent stuff? Oct 14 13:11:47 shoragan: what is your current glob? Oct 14 13:11:55 no idea Oct 14 13:11:59 i just use mtn pull :) Oct 14 13:12:07 shoragan: mtn pull server org.openembedded.* Oct 14 13:12:15 shoragan: and also get the latest nslu2 bits Oct 14 13:12:20 ok Oct 14 13:12:23 which servers? Oct 14 13:12:58 shoragan: for any OE mirror and monotone.nslu2-linux.org for org.nslu2* stuff Oct 14 13:15:21 ah good our glibc builds itself with --hash-style=both Oct 14 13:15:21 zecke, will update those branches Oct 14 13:15:27 cool :) Oct 14 13:15:31 very good Oct 14 13:17:06 zecke: by the way, if you are interested in improving performance on openmoko, you might also consider turning off OABI compatibility in the kernel. that has, or did have last time I checked, a measurable impact on system call latency. Oct 14 13:17:18 if you're in a pure eabi world nowadays then you shouldn't need that. Oct 14 13:17:36 gta01 isnt pure EABI :-( Oct 14 13:17:46 oh, and you should compile your kernel without CONFIG_FRAME_POINTER, but you knew that. Oct 14 13:18:00 XorA: which bits use oabi? Oct 14 13:18:00 * mwester was actually wondering about that the other day, as SlugOS will move to EABI shortly. That's unfortunate... :( Oct 14 13:18:08 pb__: stupid binary only gps driver Oct 14 13:18:35 XorA: oh, that is a shame. what form does the binary come in, a shared library or something? Oct 14 13:18:46 pb__: binary Oct 14 13:18:54 Gaah! gllin strikes again! Oct 14 13:19:00 pb__: executable Oct 14 13:19:06 pb__: and their attempts to get eabi version failed as it does not wanted to work at all Oct 14 13:19:17 XorA: ah, so does that mean you ship your entire userspace as oabi? Oct 14 13:19:20 XorA: tried Andrew schwartz tool? Oct 14 13:19:20 pb__, we could disable oabi compat in the gta02 config Oct 14 13:19:28 or do you just have a little oabi world for the gps and everything else runs as eabi? Oct 14 13:19:41 if it's just one binary then it should be easy enough to wrap it with LD_PRELOAD to make it run against an eabi kernel. Oct 14 13:19:45 pb__: no, minimal bits of glibc and maybe other libraries Oct 14 13:20:14 The defconfigs for gta01 and 02 will differ anyway; and never will boot correctly if interchanged -- I think we could limit OABI support to the 01 Oct 14 13:20:19 pb__: not sure if its in a chroot, or using ld.so tricks Oct 14 13:20:40 XorA: ah, I see. well, if it has a special glibc anyway then you could easily enough make that use the eabi syscall interface and turn off oabi in the kernel. Oct 14 13:20:45 mwester: yeah, I forgot we officially dropped the one kernel Oct 14 13:20:53 Currently uses ld.so tricks, but might be able to do a chroot Oct 14 13:21:16 mwester: I didn't understand your comment about unfortunate for slugos. surely if you are moving to eabi soon then that's good, right? Oct 14 13:21:21 mwester, the ld.so tricks save a lot of space Oct 14 13:21:28 compared to a chroot Oct 14 13:21:52 heh, I guess I should just look on my gta01 rather than asking questions :-} Oct 14 13:21:54 Yes, we're eager to go to EABI -- I was just hoping we could keep OABI support as well without performance penalty. Oct 14 13:22:16 zecke, will nslu2 also switch to git with us? Oct 14 13:22:42 shoragan: I'm coming to the conclusion that we'll end up having to do rootfs on SD for GTA01 anyway, it's pretty tight. Oct 14 13:22:54 maybe Oct 14 13:23:07 shoragan: I think they will Oct 14 13:23:08 mwester: ah, I see. well, it's not so much a performance penalty, more lack of a performance improvement. Oct 14 13:23:21 nslu2-linux is ready to switch. Oct 14 13:23:28 zecke, it is important that the use the same conversion Oct 14 13:23:34 we could just finish making glling work under schwartz and forget about OABI on gta01 as well Oct 14 13:23:41 mwester: eabi with oabi-compat is about the same as pure oabi, just not as fast as pure eabi. Oct 14 13:24:36 pb__: Thanks for the clarification -- one of the goals is performance, so I think we may have to consider a clean break, and just not support the old binaries. I'll toss the question out to the NSLU2 mailing list... Oct 14 13:25:09 mwester I fixed the efsprogs problem Oct 14 13:25:19 hi diego Oct 14 13:25:32 shoragan: I think they will just switch as well and use our result Oct 14 13:25:37 mwester: the issue, for what it's worth, is that oabi encodes the syscall number in the swi instruction. if you have oabi compat enabled then you have to inspect every swi to find out whether it is oabi or not. Oct 14 13:25:41 ok Oct 14 13:26:01 mwester: I guess you could actually do something with personality() to mark pure eabi tasks as such and avoid doing the lookup for them. that might be a good way to mitigate the loss. Oct 14 13:26:54 or, if you were prepared to lose the ability to mix and match within a single process, you could just set a flag based on whether you had ever seen an eabi syscall and assume that all syscalls will be eabi if yes. Oct 14 13:27:10 that would still allow you to run legacy binaries without any real impact on new ones. Oct 14 13:28:42 sorry for offtopic hm does anyone knows how the filter in scandir works or has a example? Oct 14 13:28:48 <_diego_> hi Oct 14 13:29:24 <_diego_> /mnt/sdd1/oe2/tmp/work/i486-angstrom-linux/e2fspr ogs-1.38-r13/e2fsprogs-1.38/e2fsck -> good, i was going to ask about that :D Oct 14 13:29:45 zecke, from monotone.nslu2-linux.org i didn't get any revisions Oct 14 13:29:51 <_diego_> (some problem in cut&paste ... :( ) Oct 14 13:29:58 _diego_ i fixed it yesterday for eal Oct 14 13:31:37 shoragan: what glob did you use? Oct 14 13:31:52 mtn pull monotone.nslu2-linux.org org.nslu2* Oct 14 13:31:52 mtn: doing anonymous pull; use -kKEYNAME if you need authentication Oct 14 13:31:52 mtn: connecting to monotone.nslu2-linux.org Oct 14 13:31:52 mtn: finding items to synchronize: Oct 14 13:31:52 mtn: certificates | keys | revisions Oct 14 13:31:52 mtn: 1024 | 5 | 340 Oct 14 13:31:54 mtn: bytes in | bytes out | certs in | revs in Oct 14 13:31:56 mtn: 1.1 k | 1.3 k | 0/0 | 0/0 Oct 14 13:32:52 http://monotone.nslu2-linux.org/viewmtn/revision.psp?id=18cd7bbf2fd04ca0f1937cc96beabf526c2c6d47 makes sense Oct 14 13:34:20 so they didn't do any developement for 5 months?! Oct 14 13:34:48 shoragan: in this branch yes, but it only contains their Master Makefile and such Oct 14 13:35:00 ah Oct 14 13:35:09 so should i pull with a different glob? Oct 14 13:35:32 shoragan: you have converted some revs from them. but e.g. their bitbake branch is not converted? Oct 14 13:35:43 * mwester wonders where rwhity is... Oct 14 13:35:49 er, rwhitby even. Oct 14 13:36:11 sleeping? :) Oct 14 13:36:21 zecke, i've converted everything i've got in my mtn db Oct 14 13:36:44 shoragan: okay, and org.openembedded.dreambox has rev's from yesterday? Oct 14 13:37:02 currently i'm updateing with the less restrictive glob on org.openembedded.* Oct 14 13:37:02 The nslu2 bitbake branch has historically been local only, I can't imagine that it would change with git, unless git doesn't allow that. Oct 14 13:37:21 shoragan: there is one branch you did not convert at all but this is due multiple heads :) Oct 14 13:37:37 zecke, which one? Oct 14 13:37:57 mwester: yes, my idea would be to convert everything, let you guys move the stuff you want to nslu2-linux.org and we can remove the branches from the OE servers Oct 14 13:38:07 shoragan: the entity branch Oct 14 13:38:39 FWIW, the need for a bitbake branch is to be able to tag releases including the bitbake versions used for said release -- so if bitbake is ever added to the main git repo, we certainly would be happy to use that (there are no local mods to bitbake) Oct 14 13:38:53 zecke, so should i try to migrate that one, too? Oct 14 13:39:17 shoragan: don't bother :) Oct 14 13:39:29 k :) Oct 14 13:40:00 mwester: okay, I think currently we only make sure that everything is converted from mtn to git, we can fix policies afterwards :) Oct 14 13:40:12 That seems wise to me. Oct 14 13:41:01 zecke, who can switch mtn to readonly? Oct 14 13:41:12 mickeyl: ping :) Oct 14 13:41:17 shoragan: do we want to switch to ro now? Oct 14 13:41:40 nah, you announced wednesday Oct 14 13:41:59 zecke: Do it! :-) Oct 14 13:42:19 shoragan: I announced tuesday readonly Oct 14 13:42:25 ah Oct 14 13:42:39 * mwester checks his calendar ;) Oct 14 13:42:50 Tuesday: Oct 14 13:42:50 - Set monotone readonly 16h german time Oct 14 13:42:50 - merge the multiple heads (if they exist) Oct 14 13:42:50 - Try to add the revs from now to tuesday to git Oct 14 13:42:51 shoragan: and I don't remember what I did 10 minutes ago, how should I remember what I wrote in taipei ages ago? Oct 14 13:42:52 :) Oct 14 13:43:13 wtf is german time? it is not well defined, let me declare it is 16h Oct 14 13:43:25 Hehe Oct 14 13:43:54 zecke: let me just add that one commit that breaks builds for everyone ;-) Oct 14 13:46:39 zecke: Btw your patch fixes the branch cache issue for me. Oct 14 13:49:46 shoragan: http://git.sicherheitsschwankung.de/?p=jan/oetest.git;a=commit;h=5f3bd83a121d59e95e426be6dea862bc8ed1fa22 - this probably was made by Phil Oct 14 13:50:00 alphaone: try to break it Oct 14 13:51:10 it should be ro now Oct 14 13:51:19 who is trini@kernel.crashing.org? Oct 14 13:51:32 Tom Rini (Tim Rini)? Oct 14 13:51:46 http://www.google.de/search?q=Tom+Rini&ie=utf-8&oe=utf-8&aq=t&rls=com.ubuntu:de:unofficial&client=firefox-a Oct 14 13:51:47 found my first commit :) Oct 14 13:52:08 http://git.sicherheitsschwankung.de/?p=jan/oetest.git;a=commitdiff;h=c2a8a34bbf3656607d8a222f2cdcff1f1b212f51;hp=e9fc69562fcd8c405ca4279027a0c155c5874638 - 4.5 year ago Oct 14 13:52:14 lot of time Oct 14 13:52:15 ah found him Oct 14 13:52:20 hrw: cool :-) Oct 14 13:53:08 that week of trying to get openzaurus 3.5.0 booted on collie.. Oct 14 13:55:30 pure fun Oct 14 13:57:27 pb__: http://git.sicherheitsschwankung.de/?p=jan/oetest.git;a=commit;h=5f3bd83a121d59e95e426be6dea862bc8ed1fa22 is yours? Oct 14 14:00:05 mickeyl: RP: I created a git user account and will move the gitosis config there Oct 14 14:00:20 hrw, thanks, fixed the bug Oct 14 14:02:12 zecke: Oct 14 14:02:13 d5287ff..7ebedb5 org.openembedded.documentation -> org.openembedded.documentation Oct 14 14:02:13 b074e06..a17df2d org.openembedded.dreambox -> org.openembedded.dreambox Oct 14 14:02:13 2c1e901..1aa68c6 org.openembedded.stable -> org.openembedded.stable Oct 14 14:02:20 should be current now Oct 14 14:04:25 cool Oct 14 14:04:36 shoragan: did you send your ssh key yet? Oct 14 14:04:44 please verify though Oct 14 14:04:49 zecke, no ;) Oct 14 14:05:11 zecke: Did I send mine yet? Oct 14 14:05:36 alphaone: not in the last 5 minutes, who are you anyway? I'm z nice to meet to you. :) Oct 14 14:05:49 zecke: are you copying keys from git trial? Oct 14 14:05:55 I sent it, but I haven't heard back so I'm not sure if it made it anywhere Oct 14 14:06:07 zecke: Hehe, I sent it some months ago Oct 14 14:07:18 zecke: Send them to you? Oct 14 14:07:19 XorA: hehe, I will Oct 14 14:08:16 zecke, so you switched to readonly? Oct 14 14:08:25 shoragan: try :) Oct 14 14:08:34 i just noticed :) Oct 14 14:09:11 mtn-0.40: warning: error: denied 'nslu2-linux@nslu2-linux.org' write permission for 'org.openembedded.dev' excluding '' Oct 14 14:09:20 mtn-0.40: warning: error: denied 'jluebbe@openembedded.org' write permission for 'org.openembedded.dev' excluding '' Oct 14 14:09:24 I see everything :) Oct 14 14:09:46 hmm, I better send in my key Oct 14 14:09:59 I just turned off the nslu2 sync. Oct 14 14:10:02 and re-read dummmies guide to git Oct 14 14:10:21 Er, quick question -- how do I make a mtn repo ro? Oct 14 14:10:34 mwester: no idea about the proper way Oct 14 14:10:44 mwester: in ~/.mtn/monotone.rc you can define who has write access Oct 14 14:10:55 mwester: -- are comments in lua and I put in return false :) Oct 14 14:11:05 That'll work. Oct 14 14:11:05 mwester: drop all public keys? Oct 14 14:11:18 hrw: that's a little extreme :-D Oct 14 14:11:19 hmmm, better make sure broonie shuts down wolfson mirror as well Oct 14 14:12:39 brb ~one hour or such Oct 14 14:12:50 and then I will work on the gitosis setup Oct 14 14:14:28 heh - rwhitby already had a monotonerc.readonly file ready to be put in place. :) Oct 14 14:18:15 hrw: hm, no idea. Oct 14 14:18:22 where was oe's metadata hosted before bitkeeper? Oct 14 14:19:56 * pb__ too senile to remember that far back Oct 14 14:20:39 pb__: there were pre-bk times? Oct 14 14:20:49 I think so Oct 14 14:20:54 old man mickeyl might remember Oct 14 14:21:06 pb__: when I started with OE it was bitkeeper Oct 14 14:21:10 I thought OE had started on bitkeeper Oct 14 14:21:20 OZ was in sourceforge svn wasnt it Oct 14 14:21:27 s/svn/cvs Oct 14 14:21:35 oh yeah, even older :-) Oct 14 14:27:00 oh, I remember, mickeyl had some... er, thing for non-bk users to commit with. Oct 14 14:27:14 http://openembedded.bkbits.net:8080/oe.deprecated/?PAGE=cset&REV=40a3f5f7gm439VgcT-LmpsixEKhp4Q Oct 14 14:27:25 that must be why I am pb_@uni-frankfurt.de :-} Oct 14 14:28:42 it looks like http://openembedded.bkbits.net:8080/oe.deprecated/?PAGE=cset&REV=405d0fd0_q59bLm0kNigU3FEtX9ZHA was my actual first patch according to the bk history, though I don't really remember it. Oct 14 14:34:38 ah, right... Oct 14 14:35:45 btw - Shawn Rutledge sounds familiar but I do not remember why - someone? Oct 14 14:36:33 is DEPENDS also valid for do_unpack? (i.e. is it ensured that do_unpack will run after all deps have been processed) Oct 14 14:38:16 Shawn - Openmoko IIRC? Oct 14 14:40:17 ah.. yes Oct 14 14:54:06 Hey Oct 14 16:08:42 g'day kergoth Oct 14 16:08:56 heh Oct 14 16:11:39 hey Oct 14 16:12:26 okay, the conversion should be ok now Oct 14 16:12:34 i've fixed some broken names Oct 14 16:23:05 zecke, jetzt bist du dran :) Oct 14 16:23:18 ich habe noch ein paar bugs gefixed und neu exportiert Oct 14 16:23:35 der aktuelle zustand ist im oetest repo auf sicherheitsschwankung.de Oct 14 16:24:22 shoragan: ich sprechen Deutsch nicht Oct 14 16:24:27 hi kergoth Oct 14 16:24:48 hrw, sorry, my brain hurts Oct 14 16:25:12 kergoth: was OE stored in other SCM then bitkeeper in old times? Oct 14 16:25:46 pretty sure it was cvs or something before it went to bk, there's probably some old bk commits with funky commit messages from the conversion... but my memory may be failing me again Oct 14 16:25:50 that was quite a few years ago Oct 14 16:25:56 :) Oct 14 16:26:24 * kergoth thinks Oct 14 16:26:34 kergoth: we have whole bitkeeper history imported into git now Oct 14 16:26:43 i heard, that's excellent Oct 14 16:26:43 thanks to shoragan's work Oct 14 16:26:57 kergoth: and we wondered was there something before bk times Oct 14 16:27:09 My beer addled memory says that there was only ever bk Oct 14 16:27:28 http://sourceforge.net/projects/openzaurus/ Oct 14 16:27:33 CosmicPenguin: thats why I ask kergoth as he is oldest Oct 14 16:27:48 or at least one of oldest Oct 14 16:27:54 could ask schurig, he and i both acted against the initial repo Oct 14 16:27:57 at the same time Oct 14 16:28:06 he pulled in the initial portage code, and i hacked it into being functional Oct 14 16:28:19 http://openzaurus.cvs.sourceforge.net/viewvc/openzaurus/buildroot-oz/ Oct 14 16:28:33 Crofton|work: buildroot-oz was before OE Oct 14 16:28:43 oh wow, the buildroot-oz, that brings back memories Oct 14 16:28:48 haha, last update 6 years ago Oct 14 16:28:50 Crofton|work: OE became because buildroot was crap Oct 14 16:29:01 this is where I get searching for openzaurus on sf.net Oct 14 16:29:06 pretty sure buildroot-oz was the first buildroot to ever get make menuconfig Oct 14 16:29:13 this is why we should be clean on OE Oct 14 16:29:14 * hrw plaed with builtroot-oz for 1-2 days before found OE Oct 14 16:29:31 refuses undocumented bbclass for example Oct 14 16:29:50 kergoth: yeah - thats right Oct 14 16:29:58 kergoth: before that, it was just the monolithic makefile for tuxscreen Oct 14 16:30:00 heh.. I was such n00b then... Oct 14 16:30:08 http://sourceforge.net/projects/openembedded/ Oct 14 16:30:11 course buildroot-oz isn't as clean as the new ones are, since the new versions of make let you define rules in a variable Oct 14 16:30:14 hrw: I don't speak german as well Oct 14 16:30:33 zecke : do you plan to write new architecture document for bitbake ? Oct 14 16:30:39 shoragan: which bugs? Oct 14 16:30:41 Genesis: no Oct 14 16:31:18 I totally can't believe the tuxscreen website is still up Oct 14 16:31:28 oki , i wonder if coming at fosdem is very useful for OE developemetn Oct 14 16:31:30 hehe, tim rikers is still there Oct 14 16:31:54 shoragan: what did you fix? Oct 14 16:31:56 Genesis: it is useful to meet there Oct 14 16:32:01 yeah, from the copy/paste of the tuxscreen buildroot main makefile, thast why his copyright is there Oct 14 16:32:04 eh Oct 14 16:32:07 fosdem is a good place to meet many OE people Oct 14 16:32:12 6 years 8 months ago was the initial commit of hte oz buildroot Oct 14 16:32:12 wild Oct 14 16:33:04 hmmm, http://www.fosdem.org/2009/node/150 Oct 14 16:33:06 Crofton|work: and drink beer Oct 14 16:33:20 http://tuxscreen.cvs.sourceforge.net/viewvc/tuxscreen/buildroot-tux/Makefile?revision=1.1.1.1&view=markup Oct 14 16:33:32 There is the ancestor of all things OE Oct 14 16:33:38 um beer Oct 14 16:33:59 * kergoth chuckles Oct 14 16:34:00 * Sup3rkiddo removes beer from his highlight list Oct 14 16:34:17 urg, a Makfile ... Oct 14 16:34:29 kergoth: :) Oct 14 16:34:31 * kergoth wonders if the original wiki pages from the oe brainstorming on irc are still around Oct 14 16:34:35 shoragan: what bugs did you fix? Oct 14 16:34:40 this year I am printing a list of beers and sticking to the low alcohol ones Oct 14 16:34:47 "Initial import of my nifty buildsystem" :-) Oct 14 16:35:10 zecke, the regex did not match all authors, so some mtn ids were leaking through Oct 14 16:35:34 ahhh... memories Oct 14 16:35:39 * CosmicPenguin wipes away a tear Oct 14 16:38:22 shoragan: thanks so much :) Oct 14 16:39:47 zecke, i think we can even use this dryrun Oct 14 16:39:59 i don't have any more changes pending Oct 14 16:45:01 someone remember how to make Scheduler in buildbot which will be possible to start from web client? Oct 14 16:45:05 <[X]Spoty> Crofton|work hhi Oct 14 16:45:33 <[X]Spoty> I removed vlc from my box, and then tried to install the compiled one with OE : opkg install vlc_0.9.2-r1_armv7a.ip but receive Oct 14 16:45:39 <[X]Spoty> * Package vlc md5sum mismatch. Either the opkg or the package index are corrupt. Try 'opkg update'. Oct 14 16:45:43 <[X]Spoty> any ideas ? Oct 14 16:49:47 try opkg update? Oct 14 16:49:50 bye Oct 14 16:50:15 or remove /var/lib/opkg/lists (or /usr/lib/opkg/lists/) contents before 'opkg install' Oct 14 16:51:16 wow I already locked me out of gitosis :) Oct 14 16:54:40 <[X]Spoty> opkg update, then remove /usr/lib/opkg/lists... but the same Oct 14 16:54:41 <[X]Spoty> oot@myhome:/packages# opkg install vlc_0.9.2-r1_armv7a.ipk Oct 14 16:54:41 <[X]Spoty> Multiple packages (vlc and vlc) providing same name marked HOLD or PREFER. Using latest. Oct 14 16:57:33 anyone know how to have ipsec in the kernel 2.6.22 for arm ? (using openembedded) ? pliiiize... i tried all patchs available on openswan site and i always got errors (never same error...) Oct 14 16:59:58 skrew: er, you shouldn't need openswan anymore. recent kernels have ipsec built in. Oct 14 17:00:22 2.6.22 are recent kernel ? Oct 14 17:00:25 just enable CONFIG_INET_AH, CONFIG_INET_ESP and so on Oct 14 17:00:27 recent enough, yes Oct 14 17:00:36 thanks i try now Oct 14 17:01:14 shoragan: you have setup gitosis? is there something special I have to do to the gitosis.conf? Oct 14 17:01:18 <[X]Spoty> hrw|gone Oct 14 17:01:21 <[X]Spoty> any ideas ? Oct 14 17:01:38 zecke, something special? Oct 14 17:01:58 shoragan: I'm in the member list of gitosis-admin but can not push/clone anymore :) Oct 14 17:02:51 zecke, gitosis-admin is a group like all others Oct 14 17:03:21 you need to have your name in a group that has the oe repo under writable Oct 14 17:05:27 shoragan: sure, that is with the theory, how to debug gitosis-serve? Oct 14 17:05:52 hmm, no idea Oct 14 17:06:13 i've got to go now Oct 14 17:10:53 flood Oct 14 17:10:56 DEBUG:gitosis.access.haveAccess:Access check for 'zecke' as 'readonly' on 'repositories/gitosis-admin.git'... Oct 14 17:10:59 DEBUG:gitosis.access.haveAccess:Stripping .git suffix from 'repositories/gitosis-admin.git', new value 'repositories/gitosis-admin' Oct 14 17:11:03 DEBUG:gitosis.group.getMembership:found 'zecke' in 'gitosis-admin' Oct 14 17:11:05 DEBUG:gitosis.group.getMembership:found 'zecke' in 'oedev-developers' Oct 14 17:11:08 ERROR:gitosis.serve.main:Repository read access denied Oct 14 17:11:10 fun Oct 14 17:11:15 zecke: bad luck Oct 14 17:30:28 ah it tries to be clever Oct 14 17:46:38 skrew: you will still need an ike/keying-daemon though. Oct 14 17:58:21 Kensan: ok ... but i got all keys/config for openswan .. so i'm blocked now :( Oct 14 18:18:17 how far is the git migration? Oct 14 18:18:58 We're on the mtn is read-only step Oct 14 18:19:10 ok thanks Oct 14 18:41:04 are there any packages depending on qt4? Oct 14 18:43:27 hm Oct 14 18:43:30 why you aksing? Oct 14 18:44:16 to learn how to use :) my problem is that I built qt4 and my image seems ok Oct 14 18:44:46 however I can not build my own applications due to missing(or incomplete) qmakespecs Oct 14 18:47:27 yca hm Oct 14 18:47:42 could please pastebin your recipe? Oct 14 18:47:47 I could take short look Oct 14 18:48:49 I don't have a recipe, I'm trying to build my app outside of my oe tree Oct 14 18:49:04 ah okay Oct 14 18:49:06 sorry Oct 14 18:49:09 I'm using qmake-qt4 provided by oe in staging dir Oct 14 18:49:37 hm and what is the problem? you dont have a good app.pro? Oct 14 18:51:23 when I invoke qmake it complains that it can not find good makespecs Oct 14 18:51:35 jepp Oct 14 18:51:43 you need a app.pro file Oct 14 18:52:12 hm and qmake do not has the paths Oct 14 18:52:15 skrew: for test purposes you could do manually keyed sa's via setkey Oct 14 18:52:36 no, I have an app.pro file, besides I build my app dozens time before Oct 14 18:53:00 only this time I'm trying to use qt4 provided by oe Oct 14 18:53:07 yca you could look at org.openembedded.dev/classes/qmake2.bbclass Oct 14 18:53:21 but I suggest write a recpie Oct 14 18:53:23 its easier Oct 14 18:53:30 then do all the stuff by hand Oct 14 18:54:24 yca and look at org.openembedded.dev/classes/qmake_base.bbclass Oct 14 18:55:43 skrew: try the ipsec-tools package instead of openswan and use in-kernel ipsec. Oct 14 18:56:19 ok I see it now, I should have used linux-oe-g++ not linux-oe-g++ Oct 14 18:56:46 ok, I'm going to write recipe for this. thanks for the help Oct 14 18:57:25 yca :) Oct 14 18:58:16 now you see that I'm confused, second one must be linux-arm-g++ Oct 14 18:58:32 hm its only a symlink Oct 14 19:02:07 he florian Oct 14 19:03:35 i can't find the bitbake files for PhoneME Oct 14 19:03:46 echelon???? Oct 14 19:03:52 are you using stable or dev? Oct 14 19:03:58 stable Oct 14 19:04:00 phoneme is only in dev Oct 14 19:04:05 ah, thanks Oct 14 19:04:40 re Oct 14 19:04:43 hms now it is clear while debian is not failing on e2fsprogs an MKINSTALLDIR Oct 14 19:04:47 they don use it Oct 14 19:04:55 and use theire install routine Oct 14 19:04:56 hahah Oct 14 19:05:09 hello, what are .macro pld reg ? Oct 14 19:05:19 has anyone tried to use midpath on their device? Oct 14 19:05:22 guntoo assembler? Oct 14 19:05:25 theses are macros in an arm assembly file... Oct 14 19:05:26 echelon yes Oct 14 19:05:33 midpath runs fine on u386 Oct 14 19:05:35 woglinde: what vm did you try? Oct 14 19:05:36 but where do i find infos on how they works Oct 14 19:05:42 cacao Oct 14 19:05:47 and java-se Oct 14 19:05:51 ah, i guess i should try that Oct 14 19:05:57 opera mini is running Oct 14 19:06:01 nice! Oct 14 19:06:05 that's what i wanted to use lol Oct 14 19:06:09 hehe Oct 14 19:06:18 java-se is in dev? :\ Oct 14 19:06:25 echelon nope Oct 14 19:06:36 only cacao and jamvm Oct 14 19:06:42 it's for the mplayer patch...they don't want #ifdef HAVE_ARMV5TE Oct 14 19:06:43 with classpath Oct 14 19:06:55 gnutoo why this? Oct 14 19:06:55 ah, gotcha Oct 14 19:07:20 woglinde, because 1)it's ugly 2) HAVE_ARMV5TE is not defined Oct 14 19:07:31 GNUtoo hehe this you didnt say Oct 14 19:08:29 s/not defined/not defined when compiling the assembly file Oct 14 19:12:32 woglinde: how do you bind the keys? Oct 14 19:14:35 echelon mouse and keyboard Oct 14 19:15:19 i mean, how do you set the keys when using midp apps Oct 14 19:15:56 i don't have a working mouse on my device Oct 14 19:16:08 hm what a device you have? Oct 14 19:16:15 touchscreen is working too Oct 14 19:16:21 http://openzipit.org/ Oct 14 19:16:25 the z2 Oct 14 19:16:27 virtual keyboard I do not know Oct 14 19:16:33 just try it Oct 14 19:17:24 moin Oct 14 19:17:30 guten morgen Oct 14 19:17:37 re zecke Oct 14 19:18:46 ~moin Oct 14 19:18:47 extra, extra, read all about it, moin is North German for everything you would say for salutatory, or http://moinmoin.wikiwikiweb.de/MoinMoinWiki Oct 14 19:19:15 woglinde: I don't have your ssh key Oct 14 19:19:47 zecke I know Oct 14 19:19:57 will mail you one Oct 14 19:20:01 in the near future Oct 14 19:24:20 wtf Oct 14 19:24:39 i'm bitbaking classpath, and it's downloading Python Oct 14 19:24:51 why does gnu classpath need Python? Oct 14 19:26:49 hm than there is dependency of the dependency in classpath to python Oct 14 19:41:24 shoragan: I miss your ssh key :) Oct 14 19:48:52 http://code.google.com/p/stablebox/ looks interesting Oct 14 19:53:21 hoi kergoth Oct 14 19:54:09 like busybox has churned that much since Oct 14 20:08:50 cp -fpPR /tmp/ZipitZ2/tmp-zipit2-angstrom/work/i686-linux/cacao-initial-0.98-r0/temp-staging//tmp/ZipitZ2/tmp-zipit2-angstrom/staging/i686-linux/lib/* /tmp/ZipitZ2/tmp-zipit2-angstrom/staging/i686-linux/lib Oct 14 20:08:54 cp: cannot stat `/tmp/ZipitZ2/tmp-zipit2-angstrom/work/i686-linux/cacao-initial-0.98-r0/temp-staging//tmp/ZipitZ2/tmp-zipit2-angstrom/staging/i686-linux/lib/*': No such file or directory Oct 14 20:09:21 what do i do? Oct 14 20:09:48 ERROR: '/tmp/ZipitZ2/meta/packages/cacao/cacao-initial_0.98.bb' failed Oct 14 20:10:00 aeh Oct 14 20:10:07 the dir seems falss Oct 14 20:10:15 /tmp/ZipitZ2/tmp-zipit2-angstrom/work/i686-linux/cacao-initial-0.98-r0/temp-staging//tmp/ZipitZ2/tmp-zipit2-angstrom/staging/i686-linux/lib/* Oct 14 20:10:20 cant be a right dir Oct 14 20:13:51 hi ant Oct 14 20:14:04 hi Oct 14 20:14:08 ahrgh Oct 14 20:14:12 last gift from mtn Oct 14 20:14:14 mtn: 5f2600b1acc33de2cbb3e1f80dfa05badd4beec1 koen@openembedded.org 2008-10-14T11:55:02 Oct 14 20:14:16 mtn: choose one with 'mtn update -r' Oct 14 20:14:18 mtn: error: multiple update candidates remain after selection Oct 14 20:14:41 wasn't it read only? Oct 14 20:16:25 * ant__ wanted to keep the last OE.mtn for the posterity... Oct 14 20:17:14 woglinde: how would i fix it? Oct 14 20:17:22 brb Oct 14 20:22:04 wanted to keep the last OE.mtn for the posterity... Oct 14 20:22:07 args Oct 14 20:22:15 echelon do you now have .dev? Oct 14 20:22:33 nope, stable Oct 14 20:23:10 then I can not help you Oct 14 20:23:20 :\ Oct 14 20:23:23 because I do not have the time to mess with stable Oct 14 20:23:24 very well Oct 14 20:23:25 re lorn Oct 14 20:23:27 re zecke Oct 14 20:24:54 hm I wonder where the qt extended image for the greenphone is Oct 14 20:26:28 CosmicPenguin: yea, the fork was relatively recent. not a bad concept in general though, lots of distros end up doing the same thing anyway Oct 14 20:30:08 i guess stable isn't so stable Oct 14 20:31:01 what i can use for syslogd ? syslog-ng don't work (missing libglib dep) Oct 14 20:32:52 * zecke hates runsv Oct 14 20:33:27 zecke, how goes the git work? Oct 14 20:34:09 * zecke is getting annoyed with runsv as it does not honor common init practices Oct 14 20:34:28 zecke: runit? Oct 14 20:34:37 Crofton|work: relax, it´s not Wednesday yet :) Oct 14 20:35:05 he has 3.5 hours :) Oct 14 20:35:06 Jay7: no idea, I'm on a gprs line... I have no fun to read any man page Oct 14 20:35:31 as I remember runsv is subpart of runit :) Oct 14 20:35:37 although I think zecke is defining his own time Oct 14 20:36:14 and yes, it's does not honor init practices :) Oct 14 20:36:25 * zecke just deleted the service Oct 14 20:39:57 hihi Oct 14 20:39:59 use minit Oct 14 20:40:04 * zecke kicks runsv Oct 14 20:41:22 * Jay7 have some ideas about right and fast init scheme.. Oct 14 20:41:40 Jay7 hehe Oct 14 20:42:07 may be we make some prototype under FreeBSD Oct 14 20:43:57 launchd from MacOS X is interesting thing Oct 14 20:44:07 * zecke likes the freebsd shutdown Oct 14 20:44:27 SIGTERM, sleep, SIGKILL, flush buffers, flush buffers, shut off Oct 14 20:45:50 okay Oct 14 20:46:22 Jay7 I find the init system from sun intressting Oct 14 20:46:35 * mwester likes Linksys' modified busybox "halt" command (for the NSLU2) -- "halt" does nothing at all except turn off a gpio that kills the power. Oct 14 20:46:35 1) everything converted 2) I have commit access 3) gitweb is there 4) wiki updated 5) git-daemon serves Oct 14 20:46:43 I think I will go to bed now Oct 14 20:47:30 zecke: thanks! I'm glad to see this happening. Oct 14 20:47:58 hm Oct 14 20:48:07 should I fast send my key to zecke Oct 14 20:48:42 sleep well zecke Oct 14 20:49:22 git fast-send Oct 14 20:49:47 cbrake: You could test if git clone works for you... I will let the dust settle though and not send an announcement Oct 14 20:50:05 zecke: will do Oct 14 20:50:33 there is one hook i would like to add... Oct 14 20:53:14 zecke wait Oct 14 20:53:19 I will send it now Oct 14 20:53:22 zecke: what am I doing wrong? warning: remote HEAD refers to nonexistent ref, unable to checkout. Oct 14 20:53:45 ant__: ? Oct 14 20:53:55 git clone git://git.openembedded.net/org.openembedded.dev Oct 14 20:54:00 cd /oe/org.openembedded.dev git pull --rebase Oct 14 20:54:34 the tiral and the final result have nothing in common Oct 14 20:55:08 http://rafb.net/p/hrQiPp86.html Oct 14 20:55:10 ant__: you can help and write some howto use git-rebase to go from trial to final system Oct 14 20:55:53 fun is I was inkitially following 4ngstr=m howto Oct 14 20:55:57 zecke send Oct 14 20:56:04 woglinde: sent you mean? Oct 14 20:57:19 zecke, du muss HEAD auf dem git-server anpassen Oct 14 20:57:26 http://groups.google.com/group/github/browse_thread/thread/eb3046a7b4c8bec7 Oct 14 20:57:33 und ich würde das repo nicht so wie einen branch nennen Oct 14 20:57:50 arg, german again :) Oct 14 20:58:01 kein Problem ;-) Oct 14 20:58:02 shoragan: just names :) Oct 14 20:58:43 because from now on there will be many org.oe.dev branches (one per repo) Oct 14 20:59:15 shoragan: HEADS, ack Oct 14 20:59:38 shoragan: repo name... hail-zecke would really sound better, good thinking! Oct 14 21:00:06 or just call it openembedded.git Oct 14 21:00:27 shoragan: I assume we will rename the branches as well Oct 14 21:00:32 actually koen suggested git clone git://git.openembedded.net/org.openembedded.dev.git org.openembedded.stable Oct 14 21:00:42 the mtn-naming scheme doesn't make much sense anymore Oct 14 21:01:19 ant__, doesn't make sense, you always get all branches Oct 14 21:01:20 seems that all progressive peoples going to git :) Oct 14 21:01:37 should I do? :) Oct 14 21:03:25 jay7 yes Oct 14 21:03:37 zecke, HEAD is still wrong Oct 14 21:03:39 is there any manual? :) Oct 14 21:03:44 jay7 sure Oct 14 21:03:46 google Oct 14 21:03:52 shoragan: I'm on GPRS... Oct 14 21:03:56 hehe :) Oct 14 21:04:01 jay7 but its nearly the same as mtn Oct 14 21:04:09 clone pull update Oct 14 21:04:28 ant__, is that your commit? Oct 14 21:04:31 http://git.openembedded.net/?p=openembedded.git;a=commit;h=309c832c2c21ceb97fd589ebce8327d0d2bba731 Oct 14 21:04:35 ?karma shoragan Oct 14 21:04:47 karma? Oct 14 21:04:51 %karma shoragan Oct 14 21:05:20 shabble_: he, no, was committed by thesing Oct 14 21:05:25 * Jay7 is going to migrate to git Oct 14 21:05:29 shoragan: _: he, no, was committed by thesing Oct 14 21:05:33 shoragan: karma points... I would give you karma for fixing mtn2git and remove some because you annoy me :) Oct 14 21:06:05 shoragan: org.openembedded.dev is a good default? Oct 14 21:06:11 ant__, i mean if you were the author Oct 14 21:06:17 zecke, yes, i think so Oct 14 21:06:24 shoragan: me and Jay7 Oct 14 21:06:42 kergoth: yah, stablebox does look interesting. it always takes me about a week to psych myself up to install any new busybox version, and about a week afterwards to calm down. Oct 14 21:06:58 shoragan: oh you left unknown.openembedded.org? I thought you want invalid? Oct 14 21:07:06 not good for my heart, anything that reduces the adrenaline shock must be good news Oct 14 21:07:08 busybox is just so central to a functional system,i think everyone is paranoid about upgrading it, like upgrading the libc Oct 14 21:07:14 right Oct 14 21:07:15 zecke, there were only 3 or 4 left, so i didn't change them Oct 14 21:07:31 zecke: shoragan: use the committer name Oct 14 21:07:35 ~help Oct 14 21:07:37 it's even worse than libc in that busybox does tend much more to random churn. Oct 14 21:07:59 ant__, no, i'll need some mapping for the author Oct 14 21:08:08 I keep finding that some previously-standard feature that we really depend on is now optional and buried five levels down in cryptic config options. Oct 14 21:08:38 ~shoragan++ for fixing the firt mtn2git bug with git-reset Oct 14 21:09:05 ~shoragan++ for finding out that git-fast-import got improved Oct 14 21:09:14 ~shoragan++ for tackling the task Oct 14 21:09:16 ~karma shoragan Oct 14 21:09:16 shoragan has neutral karma Oct 14 21:09:23 wtf Oct 14 21:09:29 maybe someone can fix the missing email adresses in http://git.openembedded.net/?p=openembedded.git;a=blob;f=contrib/mtn2git/manglenames.py;hb=HEAD Oct 14 21:09:48 ibot: shoragan++ Oct 14 21:09:49 ~karma shoragan Oct 14 21:09:49 shoragan has karma of 1 Oct 14 21:09:52 ibot: shoragan++ Oct 14 21:09:53 even if not i'm rather happy Oct 14 21:10:05 hm Oct 14 21:10:30 I've fixed my item already on wiki Oct 14 21:10:48 seem I've did this too late :) Oct 14 21:10:52 ah, maybe i'll take a look tomorrow Oct 14 21:10:53 :) Oct 14 21:10:59 or not ;) Oct 14 21:11:09 good night anyway Oct 14 21:11:13 shoragan: reuse 'ant' for Jay / Yuri Oct 14 21:11:15 s/"jay4mail"/"Yuri Bushmelev"/ :) Oct 14 21:11:18 shoragan: wait :) Oct 14 21:11:33 it all depends on if zecke announces that git is r/w Oct 14 21:11:34 shoragan: HEAD is fixed, repo renamed, everything looks fine right? Oct 14 21:12:15 zecke, yes, just some minor missing mappings... Oct 14 21:12:25 shoragan: I'm already there (Andrea A.) Oct 14 21:12:26 i don't mind either way Oct 14 21:13:32 shoragan: well, people were late... I think it is okay Oct 14 21:13:37 zecke: oops fatal: The remote end hung up unexpectedly Oct 14 21:13:39 ok Oct 14 21:16:05 ok, checkout works :) Oct 14 21:16:47 good night, zecke, everyone Oct 14 21:18:02 nitr shoragan Oct 14 21:19:16 Jay7: mizar ~ # emerge --unmerge monotone Oct 14 21:19:23 :-) Oct 14 21:19:52 ant__: 'apt-get remove' now :) Oct 14 21:20:22 removing is quicker in Gentoo :-D Oct 14 21:20:41 :) Oct 14 21:21:37 Jay7: do you think we would ever need some indexing service like 'beagle' on Zaurus? It is about ext3 extended attributes Oct 14 21:22:19 we have no enough RAM :) Oct 14 21:22:55 129 packets transmitted, 91 received, 29% packet loss Oct 14 21:22:56 :( Oct 14 21:23:12 my ISP have problems.. Oct 14 21:23:13 03Cliff Brake  07org.openembedded.dev * r2660ba617f 10org.openembedded.dev.git/conf/machine/cm-x270.conf: cm-x270.conf: test edit in new git repo Oct 14 21:24:02 Jay7: ok, I'll do some minimal defconfigs changes then (+CONFIG_EXT3_FS=y +CONFIG_JBD=y) Oct 14 21:24:41 ant__: I think it's enough for beginning :) Oct 14 21:24:43 hehe Oct 14 21:25:01 and about usb? Not sure is possible on c7x0... Oct 14 21:25:20 (booting from usb) Oct 14 21:25:23 hm.. Oct 14 21:25:37 but other models in theory... Oct 14 21:25:55 it can be possible in theory on C1000/C3xxx Oct 14 21:26:29 can you check size difference with usb subsystem enabled? Oct 14 21:26:42 he..could be huge... Oct 14 21:26:58 I'll do some tests Oct 14 21:26:59 hm.. Oct 14 21:28:04 I know no one who asks about booting zaurus from usb Oct 14 21:28:24 not only Zaurus, but the name is koen ;-) Oct 14 21:28:33 only when usb flash is soldered in :) Oct 14 21:29:08 ok Oct 14 21:29:23 we can provide some option in machine/local.conf Oct 14 21:29:37 like ENABLE_USB_BOOTING=yes :) Oct 14 21:30:02 for peoples who real wants this Oct 14 21:30:13 well, with the actual sdhc 8gb..usb keys are less attractive for handelds Oct 14 21:30:14 but kernel size should be checked.. Oct 14 21:30:35 yep Oct 14 21:35:04 * cbrake ponders the historic moment -- OE has switched to git :-) Oct 14 21:35:43 vbrake *g* Oct 14 21:35:46 ups Oct 14 21:35:48 cbrake Oct 14 21:36:40 how i can find in what "packages" i can find a command ? EG i want real awk, not the busybox one, how i can compilate real awk Oct 14 21:37:14 skrew: find -name "*awk*" Oct 14 21:37:29 skrew: bitbake gawk Oct 14 21:37:47 i have make gawk .. i can just ln -s awk gawk ? Oct 14 21:38:35 ok yes i can ... :) thanks Oct 14 21:38:50 skrew: good. some of the packages use update-alternatives to do this type of linking Oct 14 21:39:11 skrew: so that could probably be added in this case if someone really wanted it Oct 14 21:39:19 ok thanks Oct 14 22:08:36 zecke: Anonymous clone seems not working Oct 14 22:11:09 ok, I have installed git :) Oct 14 22:11:24 where to get repository? :) Oct 14 22:11:38 http://wiki.openembedded.net/index.php/Git Oct 14 22:12:55 fatal: The remote end hung up unexpectedly Oct 14 22:13:53 git clone git://git.openembedded.net/org.openembedded.dev Oct 14 22:14:32 shoragan: HEAD is fixed, repo renamed, everything looks fine right? Oct 14 22:14:36 ?? Oct 14 22:14:57 ant__ ??? Oct 14 22:15:12 one hour ago I could pull Oct 14 22:15:19 i could clone both types Oct 14 22:15:19 now ...hung up Oct 14 22:16:07 atal: Unable to look up (port 9418) (Name or service not known) Oct 14 22:16:35 fatal: The remote end hung up unexpectedly Oct 14 22:16:43 hm Oct 14 22:16:46 I have same error Oct 14 22:16:52 too old git? Oct 14 22:16:56 with git clone Oct 14 22:17:02 dont know Oct 14 22:17:12 git version 1.5.6.5 Oct 14 22:17:13 git version 1.5.5.5.GIT Oct 14 22:17:45 zecke is at sleep right now Oct 14 22:17:57 ok Oct 14 22:17:59 [ebuild R ] dev-util/git-1.5.6.4 Oct 14 22:19:47 Jay7: reading here http://git.or.cz/gitwiki/GitFaq Oct 14 22:21:11 kensan ? Oct 14 22:23:06 pb ? Oct 14 22:24:56 thanks all .. i finally got openswan working without patching kernel :) Oct 14 22:25:38 skrew hm Oct 14 22:25:52 I found strongswan a little bit better Oct 14 22:26:45 woglinde: maybe ... but openswan are on the "requirement" list... so i use openswan ... :) Oct 14 22:27:34 * Jay7 -> sleep Oct 14 22:28:20 requirement ? Oct 14 22:28:33 bye Jay Oct 14 22:29:11 woglinde: yep i help a friend who develop a card for his society Oct 14 22:42:02 good nite Oct 14 22:42:12 nite Oct 14 22:46:11 ok, for the memory, after the renaming the correct command seems to be: $ git clone git://git.openembedded.net/openembedded.git Oct 14 22:48:34 the e2fsprogs mkinstalldirs.patch not only breaks slugos but also openmoko builds Oct 14 22:50:57 actually, the old behaviour is obtained by: git clone git://git.openembedded.net/openembedded.git /oe/org.openembedded.dev Oct 14 22:51:50 mwester: thx for looking after the nslu2 repo during the switch Oct 14 23:03:16 np :) I'm looking forward to git. Oct 14 23:12:50 good night Oct 15 00:18:49 mtn: error: I/O failure while talking to peer monotone.openembedded.org, disconnecting Oct 15 00:19:06 what do i do after this? Oct 15 00:19:17 make: *** [org.openembedded.dev] Error 1 Oct 15 00:21:25 Switch to git Oct 15 00:21:26 sec Oct 15 00:23:05 what? Oct 15 00:24:44 http://wiki.openembedded.net/index.php/Git Oct 15 00:24:49 Are you using a Makefile for OE? Oct 15 00:24:58 If so, run the command to update the Makefile Oct 15 00:25:27 I don't use the Makefile, so I don't know what the command is off the top of my head, but I know it will update things for you :) Oct 15 00:25:34 a make file Oct 15 00:26:00 ok Oct 15 00:26:02 try 'make update' Oct 15 00:26:16 Some quick googling suggests that will update the Makefile you're using Oct 15 00:26:19 then try your other target again Oct 15 00:26:34 no rule :\ Oct 15 00:27:32 * Tartarus digs a bit more Oct 15 00:28:10 Don't have the email I was thinking of around anymore Oct 15 00:28:13 http://rafb.net/p/7n3YI696.html Oct 15 00:28:15 But, where'd you get the Makefile? Oct 15 00:28:41 http://linux.zipitwireless.com/wiki/OpenEmbedded Oct 15 00:29:02 Ahh Oct 15 00:29:34 You're possibly out of luck for a little bit Oct 15 00:29:39 OE is changing from monotone to git right now Oct 15 00:29:48 And I guess the monotone server is down now, rather than read-only Oct 15 00:29:54 I guess. Oct 15 00:29:55 :\ Oct 15 00:34:00 k, got it working now Oct 15 01:15:36 monotone raised my cpu temp up 180 degrees F Oct 15 01:48:00 Tartarus: OE has *changed* from monotone to git :-) Oct 15 01:59:45 rwhitby: but mtn is still available, yeah? Oct 15 02:00:04 echelon: read-only, yes. it will never be updated, and will at some point go away. Oct 15 02:00:10 ok Oct 15 02:00:37 error: branch org.openembedded.dev has multiple heads Oct 15 02:00:39 ^_- Oct 15 02:00:58 Hmm - multiple heads on a read-only mtn server is not good. Oct 15 02:01:02 which server? Oct 15 02:01:23 monotone.openembedded.org Oct 15 02:01:48 you should report that on the oe-devel mailing list then. Oct 15 02:01:53 (so zecke can fix it) Oct 15 02:02:29 but first be sure that the problem is on the server, and not something in your local checkout Oct 15 02:06:29 http://rafb.net/p/mEQLE010.html Oct 15 02:06:34 that's my makefile Oct 15 02:06:43 after it failed, i tried to manually do it myself Oct 15 02:07:13 @mtn --db=OE.mtn pull monotone.openembedded.org org.openembedded.dev Oct 15 02:07:21 that one just completed Oct 15 02:07:28 @mtn --db=OE.mtn checkout --branch=org.openembedded.dev Oct 15 02:07:33 the problem occurs when i run that Oct 15 02:09:03 and the org.openembedded.dev directory is empty before you do all this? Oct 15 02:09:16 and you've never committed anything to monotone locally? Oct 15 02:09:30 i don't have any local org.openembedded.dev directory :( Oct 15 02:14:30 mtn: 4a6941ceebdb9a7f82d8a8d2314c79e1801c068e jluebbe@openembedded.org 2008-10-14T12:33:15 Oct 15 02:14:36 mtn: 5f2600b1acc33de2cbb3e1f80dfa05badd4beec1 koen@openembedded.org 2008-10-14T11:55:02 Oct 15 02:14:42 i'm supposed to pick one? Oct 15 02:14:47 what's the id? Oct 15 02:16:04 i added.. -r4a6941ceebdb9a7f82d8a8d2314c79e1801c068e Oct 15 02:16:57 worked :) Oct 15 02:20:28 ok, you should post that to the mailing list Oct 15 02:23:25 it just told you to pick one Oct 15 02:23:51 i'll notify the people on zipit **** ENDING LOGGING AT Wed Oct 15 02:59:57 2008