**** BEGIN LOGGING AT Sun Jan 01 02:59:57 2006 Jan 01 03:47:20 i just created #566 Jan 01 03:47:38 setpwc needs to be updated Jan 01 05:10:55 hey Jan 01 05:11:16 morning zecke Jan 01 05:12:59 zecke: looks like we are the only sober ones :) Jan 01 05:15:57 heh Jan 01 05:16:01 morning zecke, lrg Jan 01 05:16:15 HI !!! Jan 01 05:16:36 morning pb_ gremlin[it] Jan 01 05:17:20 anyone have a hangover from last night ? Jan 01 05:25:41 hangover ? Jan 01 05:26:05 lrg|home: hehe Jan 01 05:26:06 gremlin[it]: it means not feeling well after drinking too much alcohol Jan 01 05:26:25 ~lart Stella Artois for my headache Jan 01 05:26:26 * ibot drops a truckload of VAXen on Stella Artois for my headache Jan 01 05:27:41 hahahah :) ... mhh no i just drink a bottle of wine and few cups of white wine after dinner ;) Jan 01 05:35:00 morning Jan 01 08:12:23 03mickeyl 07org.oe.dev * r186f778d... 10/packages/regina-rexx/ (3 files in 2 dirs): add regina-rexx(-native), a cross-platform REXX interpreter Jan 01 08:14:12 happy new year Jan 01 08:14:30 hail mickeyl Jan 01 08:14:36 happy new year to you Jan 01 08:15:14 Happy New Year! Jan 01 08:16:33 hi all Jan 01 08:16:37 hi florian Jan 01 08:17:15 happy new year all Jan 01 08:17:29 mickeyl: Thanks for restarting monotone... but i suppose i do something wrong. Jan 01 08:17:39 hi mreimer - same to you! Jan 01 08:17:48 thanks florian Jan 01 08:18:42 03pH5 07org.oe.dev * r9dd1ab5b... 10/packages/gaim/ (files/gtk-deprecated-2.0.0.patch gaim_cvs.bb): gaim-cvs: allow gtk-deprecated Jan 01 08:21:16 mickeyl: What does monotone want to tell me: "read from fd 6 (peer monotone.vanille.de) failed, disconnecting"? Jan 01 08:22:44 florian: no idea actually. i would need to look into the code to really tell you :) most of the times, this is just an internal server problem - perhaps it has a limited number of concurrent connections or so - simply retrying should work most of the times Jan 01 08:27:52 mickeyl: i get exactly the same message from ewi too... Jan 01 08:28:22 oh oh !! monotone: multiple update candidates Jan 01 08:28:24 hmm... firewalled? Jan 01 08:29:47 i try just now to sync and it connect but i gor a double update definition warning ;( Jan 01 08:30:03 mickeyl: unlikely... nilsf is too busy to play with the firewall ;) Jan 01 08:30:34 hehe Jan 01 08:30:42 which mt version are you using? Jan 01 08:30:49 the server runs 0.24 iirc Jan 01 08:31:38 0.24, but i wonder if i checked in anything sine the update. Jan 01 08:32:26 I remember i converted the database quite some days ago, pulling works and i have an external key now an tell it mt. Jan 01 08:35:39 ah Jan 01 08:36:28 iirc to ensure we don't get the problematic changes from nslu in again, we shut down all keys and readded them on demand Jan 01 08:37:03 and koen rebuilt the db Jan 01 08:37:10 so it would be better to start with one of his snaps Jan 01 08:37:25 instead of your own db which is likely to have the problematic changesets Jan 01 08:38:54 that would suck... i checked in a huge ile of changes. Jan 01 08:40:57 how can i check if my db includes these problematic bits? Jan 01 08:40:58 yeah, i'm afraid there is no other way without compromising the database again though. may koen has a better idea, he's the MT wizard Jan 01 08:44:15 * florian forgets about oe Jan 01 08:56:06 happy new year Jan 01 09:07:22 chouimat: thanks - same to you. Jan 01 11:42:49 happ new year everybody Jan 01 11:43:05 happy new year everybody Jan 01 12:16:07 to all of OE community, i whish you a happy new year full of joy and good things. Keep going on your wonderfull work, you are all doing a great job. Jan 01 12:16:50 thanks alan. to you too! Jan 01 12:17:07 thanks mickeyl ^^ Jan 01 12:19:25 mickeyl: too bad OZ 3.5.4 couldn't be release when you wanted. Let's hope it will be possible soon. Jan 01 12:48:46 florian: what type of changes did you apply? have you moved files around or removed stuff? Jan 01 12:49:45 florian: otherwise you should be able to export via "montone diff" as long as your local db is still in a working condition Jan 01 12:55:23 reenoo_: my local db is okay... but i removed stuff, added stuff and even worse... i have changes in my repository we won't want to have upstream... Jan 01 12:56:02 i could sdo several diffs, but i don't know if monotone is able to deal with new files in diffs. Jan 01 12:56:37 new files aren't problematic Jan 01 12:57:02 neither are renames and deletions but they aren't displayed as part of the diff Jan 01 12:57:16 so you can't import them with patch Jan 01 12:58:25 reenoo_: then this might be a solution... but i'd be more lucky if: Jan 01 12:59:10 - someone would be able to tell me what this error message really means, even mickeyl just guessed. Jan 01 12:59:30 - i'd be able to check if my db actually contains these broken bits. Jan 01 13:01:18 well. your key is not allowed push access on the server unless you've started with a fresh db and informed koen about that. Jan 01 13:01:36 the server will just disconnect you Jan 01 13:01:53 and yes, the error message could use improvement Jan 01 13:02:20 reenoo_: hum well... you are the first who knows this. noone told me. Jan 01 13:02:53 Koen sent mail about that to oe@ after the latest monotone blowup Jan 01 13:03:43 reenoo_: i asked koen what to do, i installed a new monotone and he told me some command to do some database conversion but not more. Jan 01 13:04:10 well. I, too, had to ask on IRC about the exact upgrade, etc. procedure. Jan 01 13:04:48 btw. would i have these broken bits if i don't even have the nslu branch? Jan 01 13:05:17 ciao all Jan 01 13:05:25 hi pigi Jan 01 13:05:38 ciao Pigi Jan 01 13:05:48 florian: dunno. sorry Jan 01 13:05:57 hi pb_, reenoo_ did you started well the year ? Jan 01 13:07:57 florian: looking at the symptoms I would say, yes, somehow changes from the nslu branch sneaked into .dev and thereby caused the breakage. but koen never reported the actual cause to oe@ so I may be wrong. Jan 01 13:08:58 hey Pigi Jan 01 13:09:15 Pigi: overslept quite a bit this morning, but otherwise, yeah, I can't complain, thanks. Jan 01 13:09:22 hey florian! happy new year ! Jan 01 13:09:29 reenoo_, fine ! Jan 01 13:09:59 Pigi: and a happy new year btw :) Jan 01 13:10:10 reenoo_, thx :) Jan 01 13:10:21 Pigi: thanks! same to you! Jan 01 13:10:56 pb_ I have a question around the ipkg. Is is "legal" to have multiple instance of a package in "packages" file ? Jan 01 13:11:06 Pigi: yeah, it is Jan 01 13:11:55 gosh! what's giving me trouble recently is that the downgrade option stopped to work from when I have changed the constrain function. Jan 01 13:12:28 oh dear Jan 01 13:12:33 basically what's happening is that if there are multiple instance of a package ( for the same arch ) I take always the newest. Jan 01 13:13:16 then, when asked to downgrade, the function is, more or less, unable to locate this version. Jan 01 13:13:39 ah, right. Jan 01 13:14:02 that doesn't sound good. even if Packages only contains one version, the user might supply an .ipk on the command line, and that metadata is effectively merged into ipkg's view of Packages. Jan 01 13:14:22 right! Jan 01 13:14:26 so, "ipkg --force-downgrade install busybox_0.01.ipk" would stop working, which would lose. Jan 01 13:15:50 I know. I have an option or two when the user suppy an ipk "by hand". If the version is lower than the one in Packages, I can easly stop the cycle, Jan 01 13:16:04 fwiw, there have been reports on supplying .ipks on the command not working at all recently Jan 01 13:16:33 http://bugs.openembedded.org/show_bug.cgi?id=510 Jan 01 13:16:42 but if I have in packages a version 0.0.1 and 0.0.3 , and the user supply the 0.0.2 I can get the right version. Jan 01 13:17:23 reenoo_, right. Jan 01 13:17:51 I can also "force" ipkg to choice the one supplied by hand, by using a flag, or something like this. Jan 01 13:18:09 yeah Jan 01 13:18:18 but I think that ipkg has been pushed really hard to the limits..... Jan 01 13:18:22 if the user specifies a particular .ipk on the commandline, that's the one that ipkg needs to install. Jan 01 13:18:46 right. I don't see a need for an additional flag Jan 01 13:19:02 heh. More easy to say than to implement ( in all that hashes in the code :) ) Jan 01 13:19:03 let's move to rpm *flame coming on left and right* :) :) :) : ) Jan 01 13:19:23 gremlin[it], feel free to rewrite a working rpm ;) Jan 01 13:19:49 Pigi: well. think of the filename as a flag Jan 01 13:20:36 stat() the file and set flag = 1 if the file exists or something Jan 01 13:21:08 reenoo_, I know. The real problem is that ipkg __could__ install either ".ipk" and feeds packages from a single command line. Jan 01 13:21:26 like " ipkg install foo.0.0.0.ipk bar " Jan 01 13:21:53 obviously foo.0.0.0.ipk is provided by hands, and bar is from feeds. Jan 01 13:21:57 it's a per package property, yeah Jan 01 13:22:03 yeah Jan 01 13:35:32 I know it is OT, but someone of you have never played with synce ? Jan 01 13:51:17 bbl Jan 01 13:51:23 * Pigi is away: I'm busy Jan 01 13:58:29 03koen 07org.oe.dev * r2eeaab44... 10/packages/gpe-what/gpe-what_0.41.bb: gpe-what: update to 0.41 Jan 01 13:58:33 03koen 07org.oe.dev * r4c96dc4e... 10/packages/gpe-bluetooth/gpe-bluetooth_0.48.bb: gpe-bluetooth: update to 0.48 Jan 01 14:06:25 03koen 07org.oe.dev * r0240e129... 10/packages/libeventdb/libeventdb_0.18.bb: libeventdb: add 0.18 Jan 01 14:24:12 is some one running a 2.6.14 kernel on a StrongARM? Jan 01 14:25:23 mhhh wait Jan 01 14:26:14 nope last cvs is 2.6.13 :( Jan 01 14:27:02 where u got 2.6.14 ? Jan 01 14:33:36 zecke ... Jan 01 14:36:10 gremlin[it]: vanilla tree. My shark will not boot off nfs Jan 01 14:36:38 ah ok Jan 01 14:39:15 zecke: still? that's very sad. Jan 01 14:39:22 can you see it trying to open any files if you run tcpdump? Jan 01 14:39:51 pb_: yes I see a lot of NFS calls (mount, get, getattr...) Jan 01 14:40:00 what's the last thing it tries to open? Jan 01 14:40:06 pb_: but then it stops ;) Jan 01 14:40:35 pb_: don't remember but the last thing I see are retransmission's and timeout of packages Jan 01 14:40:58 oh dear Jan 01 14:40:59 pb_: but I still can ping the machine, scroll the boot log.. Jan 01 14:41:03 retransmissions and timeouts can't be good Jan 01 14:41:09 maybe you have some fragment-related problem Jan 01 14:41:30 pb_: a more important issue Jan 01 14:42:01 pb_: AC_CHECK_PROG_VER is unknown when I do the usual autotools.bbclass do_configure on glibc Jan 01 14:42:17 pb_: and it says I should use m4_allow_pattern to make it known... Jan 01 14:42:27 pb_: I'm about to patch the 'configure' script... Jan 01 14:42:45 good night Jan 01 14:42:53 mm, interesting Jan 01 14:43:09 glibc's aclocal.m4 in cvs contains that macro, but there's no clue as to where it comes from Jan 01 14:43:54 maybe seeting acpaths = " -I." helps.. let me try this Jan 01 14:44:26 I doubt that will help. autotools.bbclass blows away aclocal.m4 Jan 01 14:44:40 you could copy the macro out of there and patch it into acinclude.m4, though Jan 01 14:45:07 pb_: ah thanks Jan 01 14:45:26 this explains why my manual autoreconf worked Jan 01 14:45:52 yeah Jan 01 14:46:02 autotools.bbclass removes that file because it is a common hideout for bogus libtool macros Jan 01 14:46:17 (and, in theory, it can be regenerated by "aclocal" so contains no valuable data) Jan 01 14:46:54 but unfortunately it's quite common for packages to treat aclocal.m4 as a source file rather than a generated one, and add their local macros there by hand. Jan 01 14:47:00 ~lart Mr. Drepper Jan 01 14:47:00 * ibot farts in Mr. Drepper's general direction Jan 01 14:47:08 lol Jan 01 14:47:19 crumbs, fighting talk Jan 01 14:48:04 ah yeah, the cvs history for aclocal.m4 reveals the culprit Jan 01 14:48:15 revision 1.11 Jan 01 14:48:15 date: 1998/04/02 17:40:54; author: drepper; state: Exp; lines: +63 -33 Jan 01 14:48:15 (AC_PROG_CC_LOCAL): Remove superfluous tests. Jan 01 14:48:15 (AC_PROG_CHECK_VER): New macro. Jan 01 14:48:45 lol he was even responsible for it Jan 01 14:49:03 yeah Jan 01 14:49:14 nearly eight years ago, though, so I guess one should cut him some slack Jan 01 14:51:20 thanks master Jan 01 14:51:30 I might have a glibc building at the end of this year Jan 01 14:51:43 heh Jan 01 14:51:50 no time to lose, then Jan 01 15:02:31 pb_: what place is preferable over configure.in? creating a m4 dir? Jan 01 15:03:03 zecke: just make aczecke.m4 in the top level Jan 01 15:03:09 or acinclude.m4, or anything other than aclocal.m4 Jan 01 15:03:21 pb_: so ac*.m4 Jan 01 15:03:54 yeah, might even be *.m4 Jan 01 15:04:30 if you pick something other than acinclude.m4, you might need to add "-I ." to the acpath. Jan 01 15:06:14 if he picks a different filename, -I . will only help if he does an m4_include() to pull the file in Jan 01 15:06:38 hm, really? I thought aclocal searched all the dirs in the path to find its macros. Jan 01 15:07:10 I don't recall ever having needed to use m4_include() before. Jan 01 15:07:28 happy new year kergoth, by the way Jan 01 15:08:00 it does search those dirs, but it doesnt load every .m4 laying around Jan 01 15:08:17 unless he has an AC_REQUIRE()? Jan 01 15:08:22 heh Jan 01 15:08:24 likewise Jan 01 15:08:34 my resolution is to get and keep a job that i dont despise Jan 01 15:08:34 autoheader: warning: missing template: ASM_GLOBAL_DIRECTIVE Jan 01 15:08:35 autoheader: Use AC_DEFINE([ASM_GLOBAL_DIRECTIVE], [], [Description]) Jan 01 15:08:47 fun fun fun Jan 01 15:09:13 kergoth: heh. what's your current employment situation? Jan 01 15:09:22 presumably either unemployed or in a job that you despise, given that resolution Jan 01 15:09:38 desperatley seeking a jb he doesn't despise Jan 01 15:10:07 * Crofton needs to resolve to improve his typing Jan 01 15:10:31 Crofton: yeah, I figured that much out Jan 01 15:14:38 pb_: I'm facing the problem that config.h lacks results of tests that were ran Jan 01 15:15:05 pb_: at the bottom of config.log I have a 'confdefs.h' section with all the #define's Jan 01 15:15:16 pb_: I wonder where I should start looking at? Jan 01 15:15:39 zecke: config.h.in should probably be your first port of call Jan 01 15:16:08 pb_: config.h is generated by autoheader? Jan 01 15:18:21 config.h.in is, config.h is generated by configure Jan 01 15:18:25 iirc Jan 01 15:18:29 * kergoth gets food Jan 01 15:18:36 well enough glibc toying for today... Jan 01 15:19:40 If I need some betatesters for my new fixes, what can be the right procedure ? Jan 01 16:26:39 Pigi: ask koen ;) Jan 01 16:26:49 zecke, ;) Jan 01 16:27:23 I will send a mail in oe@ or familiar-dev@ and see what's happen :) Jan 01 16:34:38 Time to sleep. Jan 01 16:34:47 Nite all! Jan 01 16:41:03 good nite Jan 01 18:27:31 03jbowler 07org.oe.dev * rb17ad878... 10/ (3 files in 2 dirs): nslu2-kernel: bug fix kernel release 2.6.15.5 **** BEGIN LOGGING AT Mon Jan 02 11:55:02 2006 Jan 02 11:55:58 I have a problem creating an initrd with OE. Jan 02 11:56:19 I have to use "mount -o loop ...", which is only possible as superuser. Jan 02 11:56:25 Is there a way arroung this ? Jan 02 11:56:40 s/ung/und/ Jan 02 11:56:59 hi Jan 02 11:57:50 <_law_|iBook> hi hrw Jan 02 12:00:04 Do we have pakets as example, which do not use any autoconfig, qmake, ...m just use a preconfigured Makefile ? Jan 02 12:01:03 uv1: do_compile() { oe_runmake } ones? Jan 02 12:02:12 uv1: acpid Jan 02 12:02:35 uv1: Dont know. I have to move existing SW packages without auto stuff into bitbake files. Jan 02 12:02:46 hrw: checking acpid Jan 02 12:03:52 btw - acpid looks like need update-rc.d inherited Jan 02 12:04:23 hrw: OK looks like worth a try ... Jan 02 12:04:40 hrw: Can you also help with ... Jan 02 12:04:45 I have to use "mount -o loop ...", which is only possible as superuser. Jan 02 12:04:56 Is there a way arround this ? Jan 02 12:05:13 uv1: I dont know any Jan 02 12:05:57 hrw: I want to build an initrd image during build process and need to loop mount this image. Jan 02 12:06:28 hrw: As we work as non root users, this is a serious problem. Jan 02 12:07:29 uv1: you cannot get this done before creating initrd? Jan 02 12:08:37 Why do you need to loop mount it? Jan 02 12:09:04 hrw: Dont know what you mean. "dd if=/dev/zero of=initrd bs=1M count=4;mke2fs initrd;mount -o loop initrd /mnt" and then copy all needed stuff into the image. Jan 02 12:09:30 That required "mount", which is not possible as a non superuser (at least on my Debian and Suse system) Jan 02 12:10:05 whats wrong with genext2fs? Jan 02 12:10:36 exactly.. if there is a tool to generate ext2 images then why not use it? Jan 02 12:10:47 OE create ext2 images with it Jan 02 12:10:55 CosmicPenguin: to be able to copy linuxrc, thing like /bin/echo, /bin/sleep, modules ... Jan 02 12:11:05 again - whats wrong with genext2fs? Jan 02 12:11:22 CosmicPenguin: What does genext2fs do ? Jan 02 12:11:27 guess? Jan 02 12:12:19 CosmicPenguin: Any docs available (or at least example packages) ? Jan 02 12:12:47 Well, OE uses it, when you specify the FSTYPE to be ext2 Jan 02 12:13:15 and google is very useful Jan 02 12:13:26 http://genext2fs.sourceforge.net/ Jan 02 12:14:17 uv1, What is genext2fs you ask? Simply, it generates an ext2 filesystem as a normal (i.e. non-root) user. It doesn't require you to mount the image file to copy files on it. It doesn't even require you to be the superuser to make device nodes or set group/user ids. Jan 02 12:15:09 Thanks for your comments. That'll help ;_)) Jan 02 12:15:19 Different question. Jan 02 12:16:30 "cvs co" does require a dir structure "cvs/abc-x.y.z/abc". I have packages, which do not have the "abc" subdir. Jan 02 12:16:41 Any hints to get arround this ? Jan 02 12:17:03 module=. Jan 02 12:17:13 we had such packages in OE Jan 02 12:17:14 n Jan 02 12:17:35 or something like that Jan 02 12:18:06 zecke: n? "module=." does work ? does not ? Jan 02 12:18:16 uv1: look at clucene/*.bb Jan 02 12:20:37 hrw: Thanks, that enough stuff to do tomorrow. I possibly also have to extend bb rpm classes again. My collegues do strange things with rpm ... Jan 02 12:20:49 Thanks a lot and see you tomorrow ... Jan 02 12:25:47 03rw 07org.oe.dev * r0c952375... 10/conf/local.conf.sample: local.conf.sample: explain "versioned distros". replace reference to familiar-0.8.3 with familiar-unstable. Jan 02 12:30:04 koen|away: ping Jan 02 12:31:44 That reminds me - I still want to hack genext2fs to setup the filesystem in a way that pleases extlinux Jan 02 12:33:55 someone use suspend2 on desktop? Jan 02 12:57:02 anyone ever seen libtool breaking bluetooth-nodbus ? Jan 02 12:57:14 | ccache arm-linux-gcc -march=armv4 -mtune=xscale -I/local/pkg/oe/collie/tmp/staging/arm-linux/include -I/usr/include -I/local/pkg/oe/collie/tmp/staging/arm-linux/include -fexpensive-optimizations -fomit-frame-pointer -frename-registers -O2 -Wl,-rpath-link -Wl,/local/pkg/oe/collie/tmp/staging/arm-linux/lib -Wl,-O1 -o .libs/hcitool hcitool.o oui.o -L/local/pkg/oe/collie/tmp/staging/arm-linux/lib -L/usr/lib /local/pkg/oe/collie/tmp/stagi Jan 02 12:57:14 | arm-linux-gcc: /local/pkg/oe/collie/tmp/staging/arm-linux/lib/.libs/libbluetooth.so: Datei oder Verzeichnis nicht gefunden Jan 02 12:57:51 ya. naturlich Jan 02 12:58:01 how to fix that? Jan 02 12:58:05 and what's the cause? Jan 02 12:58:11 I dont understand error message Jan 02 12:58:15 * chouimat is back Jan 02 12:58:22 LC_ALL=C is best for OE Jan 02 12:58:23 Datei oder Verzeichnis nicht gefunden = file not found Jan 02 12:58:41 ok Jan 02 12:58:45 never had it Jan 02 12:58:49 hmm bummer Jan 02 12:59:11 so much for changing CVSDATE to something more recent Jan 02 12:59:16 * mickeyl turns it back again Jan 02 12:59:23 mickeyl: check does you build libusb before bluez or not.. Jan 02 12:59:43 no, libusb is not built Jan 02 13:00:36 libtool somehow misinterprets STAGING_LIBDIR as the point to look for a library Jan 02 13:00:45 strange... Jan 02 13:02:20 zecke: how I use icecream with oe? Jan 02 13:02:21 the only floating CVS packages are ipkg-utils* Jan 02 13:02:34 chouimat: INHERIT += "icecream" Jan 02 13:02:39 funny.. I'm watching 'bad girls' in TV and there are lot of 'hardware' shops Jan 02 13:03:10 zecke: ok thanks Jan 02 13:03:21 !!! Jan 02 13:03:22 gremlin[it]: Dunno? Jan 02 13:04:28 chouimat: there might be some more options Jan 02 13:06:15 Ciao all Jan 02 13:06:32 hi pigi Jan 02 13:06:40 hi pb_ ! Jan 02 13:07:24 did you discovered any new bug^H^H^H feature recently ;) ? Jan 02 13:07:49 hehe Jan 02 13:07:52 no, not recently :-) Jan 02 13:08:00 good. :) Jan 02 13:08:31 did you have any further thoughts about this Replaces: thing? Jan 02 13:08:50 as far as I could tell with my testcase, Replaces: (without Conflicts:) has no effect at all Jan 02 13:09:02 No, just looked trough the code, and it doesn't seems to be wrong. Jan 02 13:09:22 Apologize for my bad understanding, this means it works or it's broken ? Jan 02 13:09:31 It's broken Jan 02 13:09:43 good ( well bad... ) Jan 02 13:10:14 I'm reading your mail just now..... Jan 02 13:10:57 The theory is that Replaces on its own should allow a package to overwrite files belonging to another package without causing an error. (It also needs to make sure that the file lists are updated to reflect the fact that the old package no longer owns those files.) Jan 02 13:11:48 see... http://www.us.debian.org/doc/debian-policy/ch-relationships.html#s-replaces Jan 02 13:11:57 but what should happen to the "replaced" package ? Should it be removed or just be left there ? Jan 02 13:12:43 Just left there. Jan 02 13:12:59 yeah. Just reading now. Jan 02 13:12:59 It should only be removed if the new package also says that it Conflicts with the old one. Jan 02 13:13:29 I should find a couple of packages in opie environment to test the baheviour. Jan 02 13:13:41 good plan Jan 02 13:13:49 zecke: I'm tired to wait after the openslug compile :) Jan 02 13:13:51 That's rather stange, as I'm sure I did the tests.... Jan 02 13:14:21 yeah, it is strange Jan 02 13:16:10 I'll now be known as pb_|kitchen for a short while. bbiab. Jan 02 13:16:16 pb_ where ( a part from oe ) can I get those two packages ( xserver-common and gep-scripts ) afayk ? Jan 02 13:16:25 gosh, too late :) Jan 02 13:19:12 chouimat: you're actually compiling something? Jan 02 13:19:17 * CosmicPenguin falls over Jan 02 13:21:49 Pigi: http://familiar.handhelds.org/snapshots/v0.8.3/feed/ for the new packages. and http://familiar.handhelds.org/releases/v0.8.2/feed/x11/ for the old gpe-session package Jan 02 13:22:29 reenoo_, thx, as those in ipkgfind aren't replacers.... Jan 02 13:23:12 Pigi: well. you may still have to add the Replaces: field Jan 02 13:23:49 Pigi: to xserver-common*.ipk Jan 02 13:24:03 sure, but we need to be sure that those from feed ( and oe too ) are good, before I open again those ( not so sympathic ) part of code Jan 02 13:28:11 like I said, xserver-common probably doesn't have the Replaces: field because that doesn't result in the desired behaviour right now. you will need to manually add it for testing purposes. Jan 02 13:28:20 bye Jan 02 13:28:29 reenoo_, ok. Jan 02 13:29:05 yeah, what reenoo said Jan 02 13:29:10 reenoo_, just to be sure: xserver-common replace gpe-session-scripts, right ? Jan 02 13:29:26 just ipkg-unbuild xserver-common, edit CONTROL/control to add "Replaces: gpe-session-scripts", then ipkg-build it again Jan 02 13:29:37 sure. Jan 02 13:36:24 RP: no more prominent place? Jan 02 13:36:46 zecke: Have you any suggestions? Jan 02 13:37:18 RP: no Jan 02 13:38:13 zecke: mentioning it in the changelog is fine for now. We can always rethink it if OH makes any more significant code contributions Jan 02 13:39:22 right Jan 02 14:05:09 ive not being able to build for a few weeks now, always fails at gcc-cross-4.0.2-r1 with oninline.h:121: error: cannot convert 'const __ctype_touplow_t*' to 'const int*' in assignment Jan 02 14:05:10 heh Jan 02 14:05:22 bitbake -i doesn't warn you when the critical variables aren't set Jan 02 14:06:18 note to self: /unpart != /quit Jan 02 14:06:20 I guess I had gotten used to that - because i couldn't figure out for the life of me why binutils was failing Jan 02 14:11:21 pb_, ping Jan 02 14:11:30 hi all Jan 02 14:11:56 Pigi: at your service Jan 02 14:12:49 heh. I have just flashed a gpe image, but on this image ( 20060102 ) both xserver-common and gpe-sessio-script are installed. Jan 02 14:13:01 Is this correct or I'm missing something ? Jan 02 14:13:43 (sometimes I feel really dumb ) heh Jan 02 14:16:16 Pigi: hm, that is rather odd Jan 02 14:16:19 I suspect it would have been better to flash a 0.8.2.... Jan 02 14:16:24 yeah Jan 02 14:16:35 oh, right, yeah, I understand Jan 02 14:16:43 the 0.8.3 versions of those packages don't conflict. Jan 02 14:17:13 the problem is that xserver-common is a new package in 0.8.3, which includes some of the files that were previously (0.8.2) in gpe-session-scripts Jan 02 14:17:32 so, the case we were worried about was an 0.8.2 user typing "ipkg upgrade" Jan 02 14:17:44 I see. So I should flash a 0.8.2 .... downloading .... Jan 02 14:18:41 right Jan 02 14:18:48 ~lart vim for displaying some (not all) prolog files with perl syntax highlighting Jan 02 14:18:49 * ibot smacks vim up side the head with a clue-by-4 for displaying some (not all) prolog files with perl syntax highlighting Jan 02 14:28:43 hi - Jan 02 14:31:38 has anybody hit this problem? My build stopped at the final do_rootfs stage: Jan 02 14:31:41 | + ipkg-cl -f /home/jsun/build/oe/build/tmp/work/bootstrap-image-1.0-r0/temp/ipkg.conf -o /home/jsun/build/oe/build/tmp/rootfs install task-bootstrap Jan 02 14:32:08 apparently there is no package for task-bootstrap built. Jan 02 14:32:18 the direction is empty Jan 02 14:32:37 rebuilding task-bootstrap was successful. Jan 02 14:32:59 I am not sure what is *supposed* to happen here. Jan 02 14:33:05 check your DISTRO. Jan 02 14:33:33 should there be a task-bootstrap package? Jan 02 14:34:05 yeah Jan 02 14:34:16 DISTRO = "familiar-0.8.3" Jan 02 14:34:33 sorry for being dumb -what should I check? Jan 02 14:34:34 do you have a familiar-0.8.3.conf in conf/distro/? Jan 02 14:34:43 hold on - Jan 02 14:35:50 I used the snapshot from dec 27 - there is familiar.conf and familiar-unstable.conf under conf/distro Jan 02 14:36:25 then you will have to set DISTRO to "familiar-unstable" or "familiar". Jan 02 14:36:30 at this point I *just* want to be build *an image*. Is there any other distro known to work better? Jan 02 14:36:37 OK, I can try that too. Jan 02 14:37:01 which distro is considered more stable? Jan 02 14:37:33 not sure. I think "familiar-unstable" is probably the most likely to work. Jan 02 14:38:02 thanks - other than familiar distro, is there any one (ones) considered stable? Jan 02 14:38:17 I'm sure there are, but I don't know which ones. Jan 02 14:38:51 (this might be dumb question too) can one build an image without a distro? Jan 02 14:38:58 jsun_: are you subscribed to oe@handhelds.org? if not I'd recommend you do that now and read the archives at gmane.org Jan 02 14:39:12 reenoo_, thanks. Jan 02 14:39:16 I will do that right now. Jan 02 14:40:06 no, you can't build an image without a distro Jan 02 14:44:25 hmm - unless I am seriously careless, I don't seem to find the mailing list subscription link on the oe wiki site. Jan 02 14:44:28 any pointer? Jan 02 14:45:25 Have a look on handhelds.org - its address is oe@handhelds.org Jan 02 14:45:27 ok - found it. Jan 02 14:45:42 it is mentioned in getting started page. Jan 02 14:57:30 I wonder if this replace thing has never worked. Jan 02 15:01:15 pb_ I think I found the problem. Jan 02 15:11:14 there is a little problem with matchbox bb file. It tries to retrive a source file from http:// while it should be svn://. Can anyone correct this or should i open a bug report ? Jan 02 15:13:24 Pigi: aha Jan 02 15:14:02 pb_ can you read my just sent mail to discuss it a bit ... Jan 02 15:15:15 alan|home: Which matchbox bb file exactly? Jan 02 15:20:13 Pigi: hm, yeah, that code does look pretty broken Jan 02 15:20:20 heh. Jan 02 15:20:37 Pigi: it seems suspicious that replacee_provides[] is set to pkg->provides, not replacee->provides Jan 02 15:20:47 and likewise for replacee_provides_count Jan 02 15:21:34 It looks like changing those two variables would probably make it work a bit better. Jan 02 15:21:37 right. Jan 02 15:22:01 ~lart jamey Jan 02 15:22:02 * ibot wallops jamey with a main rotation server that needs rehubbing. It won't take long Jan 02 15:22:09 heh Jan 02 15:22:20 ah, harsh lart there Jan 02 15:24:10 Pigi: so, if you fix that, does the testcase work correctly? Jan 02 15:24:31 pb_ just testing. Jan 02 15:27:37 pb_ yes, so it seems Jan 02 15:28:07 basically I have commented out : Jan 02 15:28:09 /* abstract_pkg_t **replacee_provides = pkg->provides; Jan 02 15:28:09 int replacee_provides_count = pkg->provides_count; */ Jan 02 15:28:55 and changed the test : for (j = 0; j < replaces_count; j++) { Jan 02 15:28:55 if (replaces[j] == abstract_replacee) Jan 02 15:28:55 Jan 02 15:29:15 ( so changing the test from replace_provides to replaces. Jan 02 15:29:59 and at least I don't get the "conflicting" stuffs. I haven't installed it by real BTW. Jan 02 15:30:10 RP : mmm... didn't check it yet. Just tried to run a bitbake gpe-image. Translating OZ docs to french for now, so i'll check latter, but if you find out, let me know. Jan 02 15:30:37 hm, I'm not sure that's quite the right thing. Jan 02 15:31:03 abstract_replacee is set to replaces[i], so I fear the test you implemented will always be true Jan 02 15:31:16 I think you want to be testing against replacee->provides[] Jan 02 15:31:54 yeah. you are right. I think I'm too tired ATM :) Jan 02 15:33:48 I need a gdb atm. heh Jan 02 15:39:46 pb_ something like this, then : if (replacee->provides[j] == abstract_replacee) Jan 02 15:41:03 but there should be something wrong yet, as the file owner did not change. Jan 02 15:42:14 I mean, the tests are all right, but "/etc/X11/Xinit.d/30xTs_Calibrate" is still in gpe-session-scripts.list , while, if I have understood correctly, should be in xserver-common.list Jan 02 15:51:16 yeah Jan 02 15:51:30 if the file owners aren't changed, there will be disaster when gpe-session-scripts is uninstalled Jan 02 15:51:38 absolutelly. Jan 02 15:52:57 I will hunt it tomorrow, as now I HAVE to go to sleep. Tomorrow mornig my wake-up call is a 0600 AM :( Jan 02 15:55:29 g'nite all Jan 02 15:55:46 RP: "SRC_URI = "http://projects.o-hand.com/..." in matchbox-wm_0.9.5.bb leads to "svn: Unrecognized URL scheme 'http://svn.o-hand.com/repos/matchbox/trunk/matchbox-window-manager'" error. Jan 02 15:56:23 RP: can you correct this or should i open a bug repport ? Jan 02 15:58:08 mmm... wait a minute... maybe i'm saying something wrong... Jan 02 16:00:46 mmm... i don't understand... it's not writen anywhere to use svn in matchbox-wm_0.9.5.bb... Jan 02 16:01:08 yet i have "NOTE: package matchbox-wm-0.9.5+cvs-20060102: failed" Jan 02 16:02:22 rhaaa... i don't understand... Jan 02 16:09:38 alan|home: I guess you want to be looking at matchbox-wm_svn.bb Jan 02 16:18:08 reenoo_: that is what i did, but then i don't see why svn complains about a wrong url... Jan 02 16:19:27 i see "SRC_URI = "svn://svn.o-hand.com/repos/matchbox/trunk;module=matchbox-window-manager"... does the error come from the proto parameter ? Jan 02 16:23:38 ok... lmook like it does come from the proto. if i remove this part, i have a 404 error. Not good, but better... Jan 02 16:26:18 i really do not understand anything to bb files... Jan 02 16:26:39 they are _much_ over my knowledge... Jan 02 16:35:21 plus i need to sleep, it's late. Good night gentlemen. Jan 02 17:13:50 cool - DISTRO=familiar worked! Jan 02 17:15:26 hmm... i'd suggest to use familiar-unstable instead Jan 02 17:16:41 florian, unfortunately that one failed. :( Jan 02 17:27:19 jsun: oops Jan 02 17:28:31 03pH5 07org.oe.dev * rc1e403c3... 10/packages/tslib/tslib_cvs.bb: tslib_cvs: add support for blueangel, closes #516 Jan 02 19:00:13 hello Jan 02 19:27:29 did anyone stumble across the "bits" directory, which recently appeared in the glibc CVS, and causes configure to fail? it looks like a server problem, as the "cvs content changed" (older checkouts, with the same CVSDATE, do not contain that directory) Jan 02 19:28:18 or somebody tried to move a cvs directory in "the incorrect" (not that there were any correct) way Jan 02 20:25:01 I compiled a distri for epia boards anyone a suggestion for a name ? Jan 02 20:31:04 O-Sepia :P Jan 02 20:31:26 emte good idea :) Jan 02 20:31:39 Let me check if the domain is still free Jan 02 20:31:52 * emte thought it was funny and has an easy pohoto/logo crossover Jan 02 20:32:00 photo( Jan 02 20:32:03 ** Jan 02 20:32:27 Maybe a stupid question but I want to release the sources as well any ideas what the best way is to do that ? Jan 02 20:32:41 that is how OE works ... Jan 02 20:32:54 or do you mean images>? Jan 02 20:33:16 The images I have an idea I just made an image that you can put directly on the cf Jan 02 20:33:38 for release? Jan 02 20:33:48 now i am not sure what your asking Jan 02 20:34:20 I was wondering if I should point people for the sources just at oe or that I should make an iso image for the sources Jan 02 20:34:41 Do you get now a bit what I mean ? Jan 02 20:34:43 just at OE if you want development and maintainance help Jan 02 20:35:06 otherwise people will walk with it and possibly not recontribute Jan 02 20:35:25 Okay good idea , I will release first the pabx and stunserver Jan 02 20:36:00 the version with an xserver still needs some development and I need to choose if I want to go with opie , gpe or something else Jan 02 20:36:05 as for release type stuff you could ask france for a pointerpage from hh.org Jan 02 20:36:20 Okay Jan 02 20:36:38 and whoever admins the OE page as well Jan 02 20:36:48 exposure always helps Jan 02 20:37:01 Emte what is the best desktop environment to bundle you think ? (gpe, opie , or something else) Jan 02 20:37:26 both really ... Jan 02 20:37:37 there are dedicated fans to both opie and gpe Jan 02 20:37:48 gpe sorley needs more developers tho Jan 02 20:37:54 sorley* Jan 02 20:38:02 bah i still cant spell it Jan 02 20:38:09 Yeah I am a bit worried with the mouse support Jan 02 20:38:23 ? Jan 02 20:38:30 Because people mostly don't have the money to buy a touchscreen Jan 02 20:38:55 Do opie and gpe both have support for hid based devices ? Jan 02 20:39:07 some level i bleive Jan 02 20:39:12 belive* Jan 02 20:39:24 how much i am unsure Jan 02 20:40:06 I will check this because that is my biggest worry , as long as that is passed than it will be okay , because the other things I have already pretty much working Jan 02 20:41:50 I am waiting now for my 19 inch mini-itx rack cases to come :) Jan 02 20:42:49 lol Jan 02 21:16:41 emte: e-mail admin@handhelds.org or france@handhelds.org Jan 02 22:33:10 does anybody know how to build libdvbpsi Jan 02 22:59:44 france, i think you mean gerwinin not i Jan 02 23:00:00 gerwinin, e-mail admin@handhelds.org or france@handhelds.org Jan 03 01:15:02 morning Jan 03 01:29:19 hrw: you around? Jan 03 02:11:45 morning all Jan 03 02:13:04 morning RP Jan 03 02:36:43 ~seen lrg Jan 03 02:36:46 ~seen lrg|home Jan 03 02:37:39 lrg was last seen on IRC in channel #oe, 10d 20h 45m 11s ago, saying: 'hey Spyro '. Jan 03 02:37:46 lrg|home was last seen on IRC in channel #oe, 1d 21h 11m 23s ago, saying: '~lart Stella Artois for my headache'. Jan 03 02:38:36 morning Jan 03 02:38:40 df00z: pong Jan 03 02:45:03 Ah Jan 03 02:45:12 Sup? I got your message earlier Jan 03 02:45:19 i tried to build OZ 3.5.4 last night Jan 03 02:45:22 it fails on glibc-cvs Jan 03 02:45:31 | ./bits/endian.h:13:2: #error Machine byte order unknown. Jan 03 02:45:39 NOTE: package glibc-2.3.5+cvs20050627-r1: task do_compile: started Jan 03 02:45:43 ERROR: function do_compile failed Jan 03 02:45:44 df00z: I prepared image for you already Jan 03 02:45:56 df00z: you fscked someting in configuration Jan 03 02:46:00 Right, I'll prolly try it out Jan 03 02:46:27 distro = openzaurus-3.5.4 Jan 03 02:46:31 should it just be "openzaurus" ? Jan 03 02:46:46 Err, whoops Jan 03 02:46:50 machine = collie supposed to be poodle Jan 03 02:47:03 df00z: there's a patch you need for it.. let me dig up the url Jan 03 02:47:09 I had the same issue 2 days ago Jan 03 02:47:13 Ah Jan 03 02:47:19 http://zoo.weinigel.se/n30/downloads/glibc.patch Jan 03 02:47:31 then just: Jan 03 02:47:41 cd org.openembedded.dev/ && patch -p1 < ../glibc.patch Jan 03 02:47:52 thanks Jan 03 02:47:56 np Jan 03 02:48:15 I'm still trying to get the bugger going with 2.6.. it's somewhat of a challenge. Jan 03 02:48:27 What machine? Jan 03 02:48:31 df00z: use openzaurus-unstable or better use branch Jan 03 02:48:57 hrw: kk Jan 03 02:49:38 df00z: c3100.. borzoi Jan 03 02:50:00 mickey|dinner: oz 3.5.4 conf in branch has 3.5.3 in feed_uris? **** ENDING LOGGING AT Tue Jan 03 02:59:59 2006