**** BEGIN LOGGING AT Thu Sep 13 02:59:57 2007 Sep 13 06:44:00 good morning all Sep 13 07:34:55 morning all Sep 13 07:37:26 03koen 07org.oe.dev * rf01ed979... 10/ (1 conf/distro/angstrom-2007.1.conf): angstrom 2007.1: require sane-srcrevs.inc Sep 13 07:38:16 good morning RP Sep 13 07:38:23 good morning all Sep 13 07:50:46 03koen 07org.oe.dev * r8b47107e... 10/ (4 files in 2 dirs): Sep 13 07:50:46 srcrev files: updates Sep 13 07:50:46 * added new entries to sane-srcrevs.inc and updated old one Sep 13 07:50:46 * added insane-srcrevs.inc that sets every package to AUTOREV Sep 13 07:50:46 * added moko-autorev.inc that sets openmoko*.bb to AUTOREV Sep 13 07:51:28 there we go Sep 13 07:55:13 when cross-compiling with oe, where are located the header files ? Sep 13 07:56:34 morning Sep 13 07:56:46 ~curse flickr for crap called yahooid Sep 13 07:56:46 May you be reincarnated as a Windows XP administrator, flickr for crap called yahooid ! Sep 13 07:56:54 hey hrw Sep 13 07:57:08 nud: ${STAGING_INCDIR} Sep 13 07:57:26 koen: are all the header files in that particular directory ? Sep 13 07:57:50 not all, but the ones you need are :) Sep 13 07:57:59 ok :-) Sep 13 07:58:07 opal build system is such a crap... Sep 13 07:58:36 opal, openh323, pwlib, all crap Sep 13 07:58:50 yeah :-( Sep 13 07:59:57 I've heard someone whas properly autotooling all that Sep 13 08:02:33 hmmm ... i just noticed in my build log that the ipk build of ruby for slugosle fails to build the socket library with the following error: "Fatal: --enable-ipv6 is specified, and your OS seems to support IPv6 feature. But your getaddrinfo() and getnameinfo() are appeared to be broken. Sorry,you cannot compile IPv6 socket classes with broken these functions. You can try --enable-wide-getaddrinfo." Sep 13 08:04:02 do you think it's wise to use a build dir of another package in a bitbake file ? Sep 13 08:04:42 what would be the easiest way to "fix" this ipk package build .. as i need the socket interface regardless of wether or not it supports ipv6 Sep 13 08:12:27 SRCREV = "1" pushed Sep 13 08:13:21 koen: I tweaked the .inc files to use ?= Sep 13 08:13:42 I had a reason for not doing that :) Sep 13 08:13:56 koen: hmm. What was the reason? Sep 13 08:14:11 making it harder for people to shoot themselves in the foot Sep 13 08:14:37 koen: OE is about allowing people to customise things. I think we have to give them that option Sep 13 08:14:40 koen: but if I want to have recent eds-dbus I would have to edit OE confs instead of local.conf Sep 13 08:14:56 It should be changable from local.conf Sep 13 08:15:12 RP: you can customize things either way Sep 13 08:15:59 just do DISTRO_EXTRA_RDEPENDS = "AUTOREV" in local.conf Sep 13 08:16:02 * koen coughs Sep 13 08:17:50 RP: moko-autorev is useless when using ?= Sep 13 08:18:24 koen: why? Sep 13 08:18:44 koen: no way to get the equivalent of the $PREFIX (where dependencies are already installed) ? Sep 13 08:19:00 because angstrom already includes sane-srcrevs.inc Sep 13 08:19:18 koen: So you need to include it before you include angstrom... Sep 13 08:19:44 that is counterintuitive Sep 13 08:20:18 * RP shrugs. Anyone touching the distro config will work it out and local.conf will just work Sep 13 08:20:43 why is local.conf so sacred that I have to butcher distro.confs? Sep 13 08:21:05 Isn't a distro supposed to set policy? Sep 13 08:21:14 koen: because local.conf is the only place when user can change some stuff Sep 13 08:21:24 that's not true Sep 13 08:21:30 koen: Distro sets default policy. The user should be free to override whatever they like Sep 13 08:21:44 and if you *need* to change things, you're not a user anymore, but a developers Sep 13 08:21:45 koen: changing distro files is not user way Sep 13 08:21:58 using angstrom-2007.1 on davinci-dvevm, when i `bitbake bootstrap-image`, i get a bunch of "unsatisfied recommendation for ..." (where, ... includes things like kernel-module-vfat, kernel-module-nfs, pointercal, alsa-states, ncurses-terminfo, and more), followed by: Sep 13 08:22:02 ERROR: Cannot satisfy the following dependencies for task-base: Sep 13 08:22:03 RP: Isn't that up to the distro to decide? Sep 13 08:22:05 task-boot update-modules angstrom-version ipkg-collateral sysvinit-pidof update-rc.d update-rc.d update-rc.d update-rc.d update-rc.d update-rc.d update-rc.d update-rc.d prism-firmware spectrum-fw update-rc.d Sep 13 08:22:21 bernard_: wrong PACKAGE_EXTRA_ARCHS? Sep 13 08:22:22 (sorry to interrupt conversation ...) Sep 13 08:22:22 koen: what if application developer use OE to build it? Sep 13 08:22:41 koen: then he select DISTRO/MACHINE and want to use AUTOREV for own app Sep 13 08:22:58 then he can set AUTOREV in his recipe Sep 13 08:23:04 koen: If you think everything should be locked down and unchangable by the user please go and start your own distro. It is not a good policy for the "default" OE distro which is meant to allow cusotmisation Sep 13 08:23:22 koen: It is also not a good policy for versatile generic include files Sep 13 08:23:23 RP: I *have* my own distro, it's called angstrom Sep 13 08:23:30 koen: in build/conf/local.conf? there's nothing set. it's essentially a copy of OE.dev/conf/local.conf.sample. Sep 13 08:23:32 koen: Angstom is not yours Sep 13 08:24:21 * bernard_ suspects he needs to clean out everything and rebuild, but was hoping for a shortcut (that wouldn't take 12 hours). Sep 13 08:24:22 RP: that's your opinion Sep 13 08:25:06 bernard_: "bitbake update-modules angstrom-version ipkg-collateral sysvinit update-rc.d prism-firmware spectrum-fw" and look does it start do_package tasks for them Sep 13 08:25:51 bernard_: and give info if this is a reason Sep 13 08:26:25 koen: Its a blummin important distinction. If you think you "own" angstrom then its not the OE community distro that half of us think it is... Sep 13 08:27:08 hrw: it does do_package_write, do_build and do_qa_staging, but not do_package Sep 13 08:27:54 RP: I want angstrom to be hard to break, you say I can't do that Sep 13 08:28:13 bernard_: in that build it does it? Sep 13 08:29:35 koen: I think its a waste of time and is just going to force people to make much nastier hacks Sep 13 08:30:01 koen: I also don't want this "policy" of yours/angstroms spilling onto generic conf files Sep 13 08:30:44 03rpurdie 07org.oe.dev * rd6b23e81... 10/ (5 files in 3 dirs): srcrev: Change default to '1', make assignments in .inc files weak assignment and update generator code, set SRCREV to the old default insane-srcrevs.inc Sep 13 08:30:44 first I get yelled at for creating angstrom- and now I get yelled at for not making angstrom- files Sep 13 08:31:16 koen: Its this underlying idea of forcing users to use certain settings I disagree with Sep 13 08:31:26 koen: It is absolutely pointless Sep 13 08:31:26 hrw: nope. that bitbake command never called do_package. should i blow away the stamps for those packages and bitbake -c package them? Sep 13 08:31:45 koen: All it does is force users to hack the distro or other config files Sep 13 08:31:58 bernard_: do_package_write_ipk are called? Sep 13 08:32:06 koen: It won't actually stop them changing anything,.. Sep 13 08:32:14 RP: hmm, right Sep 13 08:32:33 maybe I overestimate the common sense of users Sep 13 08:32:35 hrw: nope. do_package_write is though, not do_package_write_ipk Sep 13 08:33:01 So does koen own Angstrom or not? The answer to that determines whether SlugOS and MokoSlug survive as distros ... Sep 13 08:33:02 koen: It also makes changes difficult for anyone with a genuine need Sep 13 08:33:19 bernard_: other packages were generated? Sep 13 08:33:31 rwhitby: I'm very worried about what koen said above and I think others will be as well :/ Sep 13 08:34:15 maybe we need a new distro called 'Micro' who's scope of decision making is a bit wider than an Angstrom ... Sep 13 08:34:16 hrw: yep. about 2960 of them, but just not those ones it seems. Sep 13 08:35:13 RP: the angstrom core team was suggested a few weeks back, but no action was taken, so I guess people don't want that team Sep 13 08:35:33 koen: We're waiting on Mickey on that as you yourself has pointed out Sep 13 08:35:48 I think if one person thinks they own Angstrom solely, then the idea of a team is not going to succeed anyway ... Sep 13 08:35:58 rwhitby: quite :/ Sep 13 08:36:17 But I hope it was just said in jest above ... Sep 13 08:36:43 it was Sep 13 08:37:10 the "start your own distro" remark was nonsense Sep 13 08:37:43 since RP demonstrated that even if you start your own distro OE will take control over it Sep 13 08:37:44 anyone fancy a pint ? Sep 13 08:38:33 hrw: i think i've got enough to guide me through. cheers for pointing me in the right direction. Sep 13 08:38:37 I think the main point is that if a certain distro is always touted as being the "reference distro for OE", then it should be a product of a community of people, not just one point of view. Sep 13 08:39:56 koen: I have no intention of "taking control" of openmoko or any other specific distros. As rwhitby said, angstrom has been held up as a reference distro which is why "OE" is expecting to have some influence over it. It is meant to set an example Sep 13 08:41:30 ade|desk: yes please :) Sep 13 08:41:43 Coopers Pale Ale please Sep 13 08:42:26 A pint of black sheep ale for me please Sep 13 08:42:33 one guinness Sep 13 08:42:59 * ade|desk pops down the pub .. i'll see what i can get with 32p and 1 shirt button and randomly a dried pea Sep 13 08:43:40 * ade|desk is a little worried by the pea Sep 13 08:43:41 * hrw can accept good Czech lager Sep 13 08:44:44 koen: i'd advise not having the guinness in my local, not a good example sadly Sep 13 08:45:37 when are folks nearing Cambs I will certainly buy you all a drink :) Sep 13 08:46:42 * XorA catches train to Cambs :-) Sep 13 08:46:48 hrw: i thought lager and good are mutually exclusive Sep 13 08:47:03 come on down XorA :) Sep 13 08:47:15 although Cambs scares me Sep 13 08:47:28 ade|desk: maybe in UK Sep 13 08:47:48 nothing beats a good real ale Sep 13 08:47:48 hehe Sep 13 08:48:00 ade|desk: but there is possibility to buy Polish beer in UK now ;) Sep 13 08:48:20 deuchars? Sep 13 08:48:59 your due a ....... Sep 13 08:49:19 bleh. this may be something to do with my problem... Sep 13 08:49:20 NOTE: package linux-davinci-2.6.x+git20070913-r1: task do_package: started Sep 13 08:49:20 sh: arm-angstrom-linux-gnueabi-depmod-: command not found Sep 13 08:49:41 XorA: can I talk you into bringing a few tins of BRU into OEDEM? Sep 13 08:50:15 koen: Ill see how luggage weight goes, would bottles be OK? Sep 13 08:50:18 damn the hh16 kernel is a big download 56 mb ! Sep 13 08:50:26 XorA: sure! Sep 13 08:50:55 We should invent a Bru + Leffe cocktail for mickey :-D Sep 13 08:51:12 as in the orange sticky sweet sparkle juice ? Sep 13 08:51:19 ade|desk: yes Sep 13 08:51:22 wow Sep 13 08:51:59 i still think a cheese and chocolate drink is the way forard, pleases both sexes Sep 13 08:52:08 forward even Sep 13 08:52:58 * XorA likes neither runny cheese or chocolate Sep 13 08:55:12 XorA: it's a concept, not a viable product sadly. men and women will never be able to agree without it though Sep 13 08:55:34 ade|desk: sounds like a girls only drink to me Sep 13 08:56:14 man and cheese, maybe it's a southern thing Sep 13 09:05:24 cbrake: ping Sep 13 09:16:51 ade|desk: probably ;-) Sep 13 09:16:56 morning zecke Sep 13 09:17:38 moin Sep 13 09:39:49 !oebug 2928 Sep 13 09:40:29 yo Laibsch Sep 13 09:41:00 hi XorA, good morning everyone. Sep 13 09:42:40 hi Laibsch Sep 13 09:43:44 gm Sep 13 09:47:37 how can one propose a patched / new bb file for inclusion ? Sep 13 09:47:44 via bugtracker Sep 13 09:53:31 * nud wonders why the hell AC_CHECK_FILE is not supposed to work when cross-compiling Sep 13 09:54:36 because it checks /usr/, which is on your host Sep 13 10:25:22 03koen 07org.oe.dev * rb37835d2... 10/ (1 packages/linux/linux_2.6.21.bb): linux 2.6.21: remove gumstix patch that breaks busybox Sep 13 10:27:29 03hrw 07org.oe.dev * r7370b7cb... 10/ (1 packages/starling/starling_0.2.bb): starling 0.2: fix DEPENDS Sep 13 10:27:34 03hrw 07org.oe.dev * r159bde97... 10/ (1 conf/distro/openmoko.conf): openmoko: use moto-autorev.inc to get OpenMoko apps from SVN always Sep 13 10:39:48 03koen 07org.oe.dev * rf17497ce... 10/ (1 packages/uclibc/elf2flt_svn.bb): elf2flt: add wrapper for uclinux binaries needing funky linkers Sep 13 10:39:52 03koen 07org.oe.dev * rc9f11994... 10/ (1 conf/distro/include/sane-srcrevs.inc): sane-srcrevc: add elf2flt Sep 13 10:39:57 03koen 07org.oe.dev * r0fc94aea... 10/ (1 packages/uclibc/bfin-uclibc_svn.bb): bfin-uclibc: add blackfin fork of uclibc Sep 13 10:40:02 03koen 07org.oe.dev * rab0d9a48... 10/ (1 conf/distro/include/sane-srcrevs.inc): sane-srcrevs: add bfin-uclibc Sep 13 10:58:59 hi i get a error on compiling gcc-cross can any1one help plz? Sep 13 10:59:01 tmp/work/gcc-cross-4.1.2-r6/gcc-4.1.2/build.i686-linux.arm-linux-uclibc/arm-linux-uclibc/libstdc++-v3/include/cstdlib:111: error: '::atof' has not been declared Sep 13 10:59:10 TARGET_OS=linux-uclibc Sep 13 10:59:18 host is debian Sep 13 11:02:37 how does oe resume broken downloads ? Sep 13 11:04:33 remove the files from the sources directory and fetch again? Sep 13 11:07:44 "/usr/bin/env svn update -r ${AUTOREV} failed with signal 1" :( Sep 13 11:15:50 wow the folks at fd.org are a bit naff at systems admin, its not too hard to keep a git feed up but they seem to defy belief. Sep 13 11:19:59 e Sep 13 11:20:04 hups sorry Sep 13 11:46:25 koen: This mail was stuck in the outbox, port 25 is filtered in the .nl? ;) Sep 13 11:47:44 gr, what is poky's svn url? Sep 13 11:48:15 Crofton|home: go to projects.o-hand.com and check Sep 13 11:48:22 thanks Sep 13 11:48:31 o-hand was forgotten Sep 13 11:48:57 zecke_: most dutch isps indeed filter port 25 Sep 13 11:49:20 koen: why? Sep 13 11:49:33 because we are dutch and want to be annoying Sep 13 11:49:51 I use my Uni VPN to send mails but.... Sep 13 11:50:35 zecke_: http://scap.linuxtogo.org/files/032809c7df976acfe029d7fb8c145719.png Sep 13 11:50:56 is that sort of usable? Sep 13 11:51:53 yes Sep 13 11:52:17 the font is a bit bug and the searchbar thingy takes up too much space, but I think it's usable Sep 13 11:53:49 s/bug/big/ Sep 13 11:54:28 zecke_: http://scap.linuxtogo.org/files/028a3a989f0242f801f80f81200628b0.png Sep 13 11:58:36 hi everyone, after my OE crashed for an unknown reason after an OE update. I tried to reinstall everything from scratch but now I get the following error AttributeError: 'pysqlite2.dbapi2.Connection' object has no attribute 'execute' Sep 13 11:58:56 any suggestions why this is happening? Sep 13 11:59:26 the error comes after I try to run bitbake Sep 13 11:59:42 your python-pysqlte2 is too old Sep 13 12:00:12 well I installed the latest Sep 13 12:00:22 at least that is the latest in the repository Sep 13 12:00:29 I use Mepis 6.5 Sep 13 12:05:38 bbl Sep 13 12:06:23 03rwhitby 07org.oe.dev * r367c1221... 10/ (4 files in 2 dirs): *-(src|auto)revs: Added revs for some missing openmoko packages. Sep 13 12:08:54 rwhitby: does AUTOREV works for you? Sep 13 12:09:17 hrw: dunno - I added a known good rev and AUTOREV in different files at the same time. one of them worked,. Sep 13 12:10:18 I still don't agree with openmoko using AUTOREV, and I'm considering making MokoMakefile add known good versions to preferred-openmoko-versions.inc so that openmoko developers are forced to change the srcrev each time they *intentionally* want the community to test a new version. Sep 13 12:11:42 Inflicting Joe Random OpenMoko developer's random SVN checkin on hundreds of community members is simply criminal, and will continue to lead to frustration and developers leaving the community. Sep 13 12:11:49 rwhitby: use DISTRO-angstrom-2007.1 in mokomakefile then :) Sep 13 12:11:57 rwhitby: you'll have to take that up with the OpenMoko maintainer :-D Sep 13 12:12:18 koen: XorA: yes, I'm fully aware of those two options :-) Sep 13 12:12:47 * XorA plays hunt the maintainer Sep 13 12:13:31 XorA: kick him Sep 13 12:13:47 But I figure that DISTRO=angstrom-2007.1 already has an "if (running-under-MokoMakefile-and-not-via-openmoko.conf) then print "use bitbake directly instead"; ;-) Sep 13 12:14:25 For some reason, I had the idea in my head that XorA was going to be the maintainer ... dunno why. Sep 13 12:14:28 to feed both words, just needs an if $OPENMOKO_DEV_BUILD then include sane.inc else inslude autorev.inc Sep 13 12:14:40 worlds Sep 13 12:14:46 bloody hell this keyboard is getting worse Sep 13 12:15:19 * XorA posts keyboard to hrw for professional cleaning Sep 13 12:15:43 and lots of polish Sep 13 12:15:58 XorA: I accept only MS4000 keyboards Sep 13 12:16:12 and not warrant getting them back Sep 13 12:16:40 koen: did mickey|zzZZzz specifically request moko-autorev.inc, or was it just put in there to maintain the status quo and mickey|zzZZzz hasn't given input yet on whether it should be there or not? Sep 13 12:16:44 hrw: this is an ms4000 Sep 13 12:16:55 XorA: black? Sep 13 12:17:00 hrw: out of three, 2 are mine, one is Wolfsons, only one doesnt have sticky keys Sep 13 12:17:10 hrw: yeah, black usb only Sep 13 12:17:43 rwhitby: I only added so the openmoko can use it if they want, hrw activated it Sep 13 12:17:44 ms4000 is the only one which can replace keyboard which I use now Sep 13 12:17:58 hrw: did mickey|zzZZzz specifically request moko-autorev.inc, or was it just put in there to maintain the status quo and mickey|zzZZzz hasn't given input yet on whether it should be there or not? Sep 13 12:18:40 koen: would it have the same effect if it was removed and bleeding-edge developers put that line in their local.conf? Sep 13 12:18:47 * XorA wonders if he can sneak ms4000 out of Wolfson and to OEDEM :-) Sep 13 12:19:03 rwhitby: it was not requested iirc but I added it as OM 2007.2 is too much changing to use one ver Sep 13 12:19:15 rwhitby: it should have the same effect Sep 13 12:19:20 for example yesterday dialer fails, todays works Sep 13 12:19:37 hrw: that's exactly the reason why it should be removed. Sep 13 12:19:38 does dialer 2 still restart gsmd at will? Sep 13 12:19:49 koen: its not dialer but libmokogsmd Sep 13 12:20:04 koen: I reported bug on it in OM bugtracker Sep 13 12:20:11 the person who realises that today's dialer works should be the one who updates the fixed SRCREV, it should not happen simply as a result of someone doing an svn checkin Sep 13 12:20:15 it makes debugging my a780 modem quite hard Sep 13 12:20:32 koen: so do not use any mokogsmd app Sep 13 12:20:36 you mean "untested svn checking that breaks horribly" Sep 13 12:20:38 then the community sees a monotonically increasing stable functionality, not an up and down instability, Sep 13 12:20:57 rwhitby: ok, then remove that line Sep 13 12:21:13 and the mistic OM/OE maintainer will handle it Sep 13 12:22:15 hrw: time for your crystal ball Sep 13 12:22:19 I think the draw of new working functionity (i.e. someone sees a commit comment and says "hey, I'd like that") will mean that OM community devs with OE commit access will bump the revs at the right times. Sep 13 12:22:34 (hopefully after they have tested it) Sep 13 12:23:18 and it means that the person who breaks it by bumping the SRCREV is immediately answerable due to them having OE commit access - i.e. they are not one step removed from accountability for their actions Sep 13 12:24:16 (BTW, I think this conversation applies to all OE distros that have more than one user, which is why I am making these points here rather than in #openmoko) Sep 13 12:24:27 rwhitby: ok then. today2 2963, dialer2 2964, contacts2 335 please Sep 13 12:24:47 I like the idea of floggings Sep 13 12:25:42 hrw: are you asking me to add those to preferred-openmoko-versions.inc, or to sane-srcrevs.inc ? Sep 13 12:26:28 rwhitby: I asked for a policy on updating sane-srcrevs.inc a few days ago, no usefull replies so far Sep 13 12:26:40 rwhitby: those ones works on my neo Sep 13 12:26:41 * rwhitby puts his 2yr daughter in mickey|zzZZzz's room, to wake him up. Sep 13 12:26:52 Crofton|home: I think in OEDEM we can go to KitKatKlub to get floggings :-) Sep 13 12:27:04 ok Sep 13 12:27:18 that might encourage this sort of thing though Sep 13 12:27:45 http://en.wikipedia.org/wiki/Kitkatclub ? Sep 13 12:27:59 koen: I would think a sensible policy would be that the specific package builds and runs, and worked with other packages in at least one distro. I don't think a sane-srcrevs bump can be expected to guarantee operation on more than one distro. Sep 13 12:28:23 Crofton|home: yup Sep 13 12:31:05 * rwhitby wonders what a bunch of male OE developers would get up to in the KitKatClub ... Sep 13 12:33:15 rwhitby: I noticed a pattern with breakages and who did the svn checkin Sep 13 12:33:45 so for bumping revs for some recipes need to have extensive testing Sep 13 12:34:48 koen: well I think it's better to blame the process than the person (who may not even know that OE is using their SVN checkins directly). No software process should release an artifact to hundreds of customers directly from trunk (unless development is mandated on branches not trunk) Sep 13 12:35:17 that person knows about OE Sep 13 12:35:45 he knows what is on the table and doesn't filter his commits Sep 13 12:36:28 koen: maybe you should default to dated svn snapshot packages instead of "latest commits" Sep 13 12:36:38 I still think it is important to focus on process Sep 13 12:37:07 * Crofton|home promises never to use the six sigma wrt to OE though :) Sep 13 12:37:30 koen: everywhere I work there is always a layer of isolation between one developers direct svn checkins, and another developers build. Sep 13 12:37:48 rwhitby: how it should be :) Sep 13 12:38:01 rwhitby: I tried to do that for the openmoko kernel Sep 13 12:38:20 suddenly things like "we'll fork" came up Sep 13 12:38:27 Crofton|home: I've been part of CMM level 5 evaluations, and have worked in level 4 organisations :-) Sep 13 12:38:36 * Crofton|home pukes Sep 13 12:38:50 yeah, I did too. Sep 13 12:39:21 there is a lot of good in that sort of thing, sadly the people in charge of implentation tend to be incompetent Sep 13 12:39:36 the good thing here is that most of us are "competent" Sep 13 12:40:05 koen: it should be the case that an individual developer can do a checkin at any time, without directly affecting another developer. that either means that development is on branches and release from trunk, or development on trunk and release from branches, or development on trunk and build from tags. Sep 13 12:40:41 Crofton|home: there is lots of good - in my experience the bad happens when the CMM level is the goal, rather than business results from improved quality is the goal. Sep 13 12:41:53 koen: in our case here, it means that no OE recipe should use a floating svn rev. It should always be a fixed rev, and either the original developer bumps it (saying "I assert this is a new good version") or a downstream developer using it bumps the rev (saying "I've tested this new version and it looks good to me") Sep 13 12:42:44 I think choosing a rev is up to the distro Sep 13 12:43:05 ok, that's the second of those two alternatives Sep 13 12:43:27 (the distro maintainer(s) being the downstream developer I referred to) Sep 13 12:43:29 woowoo, my ADSL is cut Sep 13 12:44:18 anyone know if I can diff betweem branches in mtn? Sep 13 12:45:23 Crofton|home: you can, just diff between the heads of both branches Sep 13 12:45:30 good :) Sep 13 12:47:30 132 files changed, 71095 insertions(+), 36 deletions(-) Sep 13 12:47:32 morning Sep 13 12:47:38 hey chouimat Sep 13 13:01:12 re Sep 13 13:01:35 just added a serial console to my storecenter unit :> time for an OE image Sep 13 14:09:18 Anyone have a hint how to resolve an edje_cc segnmentation fault when building emotion? http://rafb.net/p/lH6XGn29.html Sep 13 14:27:30 CM: it's edje-cc native, you could gdb it on your host Sep 13 14:28:34 ok, thanks for the pointer Sep 13 14:29:12 probably wrong number of arguments Sep 13 14:29:22 efl tends to segfault on crap like that Sep 13 14:30:42 I'm guessing mickey|zzZZzz knows :) Sep 13 14:31:02 how does one get around a git server not being available ? Sep 13 14:31:33 stab linus in the eye Sep 13 14:31:42 koen: with a spoo Sep 13 14:31:43 n Sep 13 14:32:47 seems a bit harsh as its the one at fd.o Sep 13 14:41:23 has anyone here tried the sdhc patch on zaurus yet? Sep 13 14:45:59 here is the untranslated page, google is a readable translation if anyone can test it http://ebsnap.lkj.jp/zaurus/memo/hiki.cgi?SDHC+%A5%AB%A1%BC%A5%C9%A4%F2%BB%C8%A4%A6 Sep 13 14:46:28 SDHC is in mainline kernel AFAIK Sep 13 14:47:58 the last information i had heard was that sdhc was unsupported, but the info on their said we just need an updated driver in the kernel, do you know of anyone who tried it yet? Sep 13 14:49:02 I own no SDHC cards Sep 13 14:49:09 otherwise it would be tested Sep 13 14:49:24 ah great shoving a tag=[hex] in the hal-info_git.bb , downloading that tag from angstrom sources, able to continue building now, at last Sep 13 14:49:45 * ade|desk shakes a fist at fd.o Sep 13 14:50:04 ade|desk: remove their kidneys with a spoon Sep 13 14:50:12 on a seperate note, which branch in oe should i be pulling from? Sep 13 14:51:26 hello, anyone here able to help be with editing the wiki? Sep 13 14:51:28 XorA: no, a fine scalpel , so i can see them on a Russian ebay-alike site Sep 13 14:51:38 ade|desk: heh heh Sep 13 14:51:52 there's some problem with anonymous editing: i can make only one edit Sep 13 14:51:56 per page Sep 13 14:53:13 i'm trying to fix the link to bbfile on the front page, it currently goes to empty page Bbfile Sep 13 14:54:08 i don't understand how the name of the page can be bbfile but [[bbfile]] creates a link to Bbfile Sep 13 14:57:39 wirelessdreamer: that sdhc patch is in 2.6.22 AFAIK Sep 13 14:58:58 koen: thanks Sep 13 15:00:51 koen: do you know if openprotium is maintained in oe, or do they use an own branch? Sep 13 15:01:09 koen: if not in 2.6.21 even Sep 13 15:02:09 send me a free 8gig card and I will test :-) Sep 13 15:03:34 CM, if you can edit the main page, can you change the link to bbfile to the uri http://www.openembedded.org/wiki/bbfile for now? Sep 13 15:04:13 Jin^eLD: I think they have their stuff in svn, but you'd have to ask hopsnbarly Sep 13 15:04:34 ah, he is doing that! cool Sep 13 15:04:40 thanks Sep 13 15:05:32 I added serial to my box today, so want to play around with it a little more Sep 13 15:08:01 bye Sep 13 15:08:18 tuukkah: I tried to edit http://www.openembedded.org/wiki/DirectoryStructure again to fix another spelling error, but got "This content has been modified by another user, changes cannot be saved." just like you did Sep 13 15:08:36 CM, ok so it's not just me Sep 13 15:09:05 Maybe there's a timeout per page on anonymous edits? Sep 13 15:09:47 * CM goes home from work Sep 13 15:10:01 ok, there probably is Sep 13 15:10:30 so, anyone with an oe account willing to fix the main page link to bbfile? Sep 13 15:32:30 likewise: hello, sorry I did not have my status set correctly this morning, and I missed your ping Sep 13 15:39:26 cbrake: pong. ever worked with openocd and ixp42x? Sep 13 15:39:48 likewise: no, I've only tried it on the PXA270 and the AT91SAM9260 Sep 13 15:40:33 cbrake: got eveything to work for flashing (even with workspace) 16 MB worth of Flash image, except that the second nibble always is zero. Sep 13 15:40:41 likewise: was successful with 1 out of 3 PXA270 boards, and still working on the AT91SAM9260 ... Sep 13 15:41:08 likewise: strange -- I've not run into that one Sep 13 15:43:37 cbrake: what key worked for you? I'm using the Amontec JTAGkey here Sep 13 15:44:01 likewise: I have the Olimex ARM-USB-OCD Sep 13 15:44:48 likewise: seems like once you are connected, the jtag adapter should not make a lot of difference, but who knows Sep 13 15:46:19 cbrake: I'm now performing low-level CFI command to write individual 16-bit words. The problem then occurs on each 16-bit write. 0xeeee becomes 0xe0ee. During block writes (using workspace) only the first 16 bit word is affected, the remaining 16 MB are OK. Sep 13 15:50:12 likewise: does working_area == "workspace"? Sep 13 15:52:30 Hi, I am getting a ton of errors like Sep 13 15:52:39 IOError: [Errno 2] No such file or directory: '/usr/src/oe/tmp/work/armv5te-angstrom-linux-gnueabi/util-linux-2.12r-r12/install/util-linux-fdisk.lock' Sep 13 15:53:04 I assume this is because I have manually removed the work dir earlier Sep 13 15:53:29 Assuming this was about the equivalent to rm_work Sep 13 15:53:49 Is there any way to recover other than restarting from scratch? Sep 13 15:54:39 I can do "bitbake -c rebuild util-linux" and the package will build fine. but there are so many failing that it becomes tedious. Sep 13 15:55:56 nop. clear all tmp and restart Sep 13 15:56:08 its the safest way Sep 13 16:09:33 cbrake: I only know about "working_area". i have set it to SDRAM which I initialized beforehand. Sep 13 16:09:51 tuukkah: Not much interest from the oe people in their wiki it seems :-/ Sep 13 16:10:24 likewise: for the PXA270, I think I have been using internal SRAM as working_area -- have not tried SDRAM yet Sep 13 16:11:04 cbrake: If I kill openocd during the write_binary, the first 16-bits look ok when I subsequently read them. So seemingly, openocd does an extra write of "f0ff" to address 0 afterwards. Sep 13 16:11:50 cbrake: thanks for listening, I have emailed the openocd list, and will continue debugging tomorrow. Sep 13 16:12:16 likewise: good luck. Could also try #openjtag -- I think drath is the openocd author Sep 13 16:12:33 cbrake: he seems silent today. I will ping him just to make sure. Sep 13 17:03:54 ho Sep 13 17:05:33 any bitbake experts in here? i recently upgraded bitbake, and i notice a difference in behaviour, i think, when doing OE builds. before it would make a package for every dependency of my target. not it only makes a package for the target. is there a way to tell bitbake to make a package for everything it compiles? Sep 13 17:23:31 CM, indeed... Sep 13 17:25:12 tuukkah: Strange, since the default answer seem to be "see the wiki for documentation" Sep 13 17:25:51 indeed^2 Sep 13 17:26:51 actually rwhitby asked me to do the edits. i suppose i need to wait until he wakes up Sep 13 17:26:58 anyone know where the size arg gets passed to mount during boot? Sep 13 17:27:12 Crofton|work: what size arg? Sep 13 17:27:52 koen, I am starting to work on why gumstix hangs at boot with Angstrom Sep 13 17:27:55 Crofton|work: if you are talking about ramfs, I removed the bogus kernel patch Sep 13 17:28:10 yeah Sep 13 17:28:23 I need to see where the optin comes in Sep 13 17:28:34 and see about using full tmpfs, not embedded one Sep 13 17:29:35 koen: What's up with the oe wiki. Why only one edit per page? Do all changes require some kind of accept or something? Sep 13 17:29:48 CM: I suspect it's the caching mechanism Sep 13 17:29:51 try now Sep 13 17:29:56 *nod* Sep 13 17:30:01 I lowered the cache liftetime from 6 hours to 1 minute Sep 13 17:31:47 I could update the page that was locked 2.5 h ago, I'll try changing it again Sep 13 17:32:24 http://rafb.net/p/fIQagV72.html Sep 13 17:32:46 ah ramfs is the bad option ... Sep 13 17:34:02 koen: Thanks, that worked Sep 13 17:34:43 koen: But the "Edit" page I got the first time was still the old text from before my first change, but after going back to the original page and then Edit again, I got the updated one Sep 13 17:34:50 tuukkah: Works better now :-) Sep 13 17:36:04 CM: good :) Sep 13 17:36:15 let's see Sep 13 17:38:01 koen, could you also rename the page bbfile as "Recipe format" since that's how it's called on the front page and we can't make links to pages that start with a lowercase letter? Sep 13 17:39:14 wowsers, ADSL2+ rocks Sep 13 17:39:38 XorA|gone: Fiber too ;) Sep 13 17:40:07 * XorA|gone has gone from 224/512 if I stood on one leg while balancing an apple on a pin on my nose to 1000/4600 Sep 13 17:40:37 ~seen hopsnbarny Sep 13 17:40:38 i haven't seen 'hopsnbarny', Jin^eLD Sep 13 17:40:46 what was his nick again? Sep 13 17:41:04 ~seen hopsnbarley Sep 13 17:41:05 hopsnbarley was last seen on IRC in channel #oe, 2d 1h 54m 57s ago, saying: 'hi likewise!'. Sep 13 17:41:29 aah, thanks Sep 13 17:41:31 :) Sep 13 17:41:31 Crofton : Austin style entertainment -> http://www.radisson.com/hoteldirectory/pictures/hoteldetailimages.jsp?imageUrl=%2Frad%2Fimages%2Fhotels%2FTXAUSTDT%2Fserv_ent_450.jpg&caption=Austin+Style+Entertainment Sep 13 17:59:52 any bitbake experts in here? i recently upgraded bitbake, and i notice a difference in behaviour, i think, when doing OE builds. before it would make a package for every dependency of my target. not it only makes a package for the target. is there a way to tell bitbake to make a package for everything it compiles? Sep 13 18:21:55 tony_: bitbake now has task dependencies, so if A (build-)depends on B, B only has to be built up to where it populates the staging area. Sep 13 18:22:48 afaik there is (unfortunately) no option (yet) to change to something like the old behaviour where each recipe is built from start to finish. Sep 13 18:23:47 pH5, thanks... too bad, that breaks some things I was doing... I guess I will have to explicitly build each package now Sep 13 18:23:57 to build packages, you have to RDEPEND on the recipes. Sep 13 18:24:20 for example create a task recipe that has RDEPENDs on all the packages you need to build. Sep 13 18:24:31 ok, thanks, i will check that out Sep 13 18:24:37 s/packages/recipes/, s/build/package/ Sep 13 18:24:58 btw, what does the R in REDEPEND stand for? Sep 13 18:25:02 runtime Sep 13 18:25:06 ah, cool Sep 13 18:25:06 runtime dependency Sep 13 18:25:07 runtime Sep 13 18:25:10 versus build time dependency Sep 13 18:25:36 thanks, guys, that helps a lot :-) Sep 13 18:26:06 is it possible to use differend defconfig files for the same distribution? e.g. when i have DISTRO_TYPE="debug" and release? Sep 13 18:26:12 or do i have to build 2 distributions? Sep 13 18:28:14 nik0n: look at angstrom-2007.1.conf, and how it sets up DEBUG_APPS Sep 13 18:28:22 you can use the same mechanism for any variable Sep 13 18:28:44 okay Sep 13 18:36:57 ok, I am almost back to where I was a week ago .... Sep 13 19:24:59 er, how do I bring up the network interface in angstrom> Sep 13 19:25:16 basically, run dhcp? Sep 13 19:25:46 ifconfig eth0 192.168.0.1 up ? Sep 13 19:26:12 How can I have a aptch applied conditional on the svn rev number of the package? Sep 13 19:26:28 I sort of want to run dhcp ... Sep 13 19:36:02 Crofton|work: any dhcp client should bring the i/f up, yes Sep 13 19:37:04 I think I do not have the correct module ... Sep 13 19:38:20 hmm Sep 13 19:42:28 zecke @ planet.nl ? Sep 13 19:42:43 zecke: are you eating frikandels? Sep 13 19:45:46 hehe, actually I have not yet managed to get fastfood Sep 13 19:45:57 * zecke is sitting in son en bruegel Sep 13 19:56:57 * likewise is 5 km to the south Sep 13 20:15:45 03crofton 07org.oe.dev * r1ae5e50f... 10/ (3 files in 3 dirs): Sep 13 20:15:45 linux : Put all defconfig information in the gumstix-verdex defconfig and Sep 13 20:15:45 remove defconfig patch form patch list. Sep 13 20:15:54 03crofton 07org.oe.dev * r37de267a... 10/ (1 packages/linux/linux-omap.inc): Sep 13 20:15:54 linux-omap.inc : Depend on uboot-utils for mkimage. Do not depend on u-boot. Sep 13 20:15:54 If you need u-boot, build it manually. Sep 13 21:22:03 re Sep 13 21:23:56 RP: ping Sep 13 21:25:56 psokolovsky: ping Sep 13 21:27:32 hrw: pong Sep 13 21:28:17 hrw: pong Sep 13 21:29:51 RP: I need to enable few more drivers in Zaurus kernels so I am thinking about enabling all pcmcia/usb drivers available (into modules). Sep 13 21:30:08 psokolovsky: will you add 2.6.22 into defconfigman? Sep 13 21:30:51 hrw: yep, sure. eventually. after hh.org upgrade. and that so far blocked by work on merging hackndev patchset. Sep 13 21:31:14 psokolovsky: ok Sep 13 21:31:26 hrw: That is ok with me in principle Sep 13 21:31:32 psokolovsky: speaking of merging... do you maintain htcuniversal kernel too? Sep 13 21:33:05 psokolovsky: if u`re the one who broke h2200 kernel, then u`ll die painfully :> Sep 13 21:34:43 re Sep 13 21:35:46 Darth_Wader: lol. throw your spears at koen. the guy can't get that configuration managed is not piece of cake, and bothers to revert maintainer's changes ;-D Sep 13 21:36:12 hrw: err, as you might notice, I kinda (co)maintain entire hh.org tree Sep 13 21:36:36 psokolovsky: I know that Sep 13 21:36:38 i`m not sure whose fault it is, but broken touchscreen AND gpio leds on h2200 makes me angry Sep 13 21:37:12 psokolovsky: in Poky you can find htcuniversal patches being part of linux-rp - thats for 2.6.22 and 2.6.23-rc Sep 13 21:37:12 Darth_Wader: pay for support. make HP pay for support. Sep 13 21:37:53 Darth_Wader: Getting angry with people isn't going to be productive Sep 13 21:38:13 i`d better pay to someone who will teach people in charge NOT to broke already working things :] Sep 13 21:38:54 Darth_Wader: work with koen on that matter. h2200 will remain broken, he said. Sep 13 21:39:02 hmm Sep 13 21:39:31 * Darth_Wader slaps koen with sledgehammer Sep 13 21:40:05 psokolovsky: where can i read about all this? Sep 13 21:40:18 on MLs and above Sep 13 21:40:56 which subjects on MLs should i read? Sep 13 21:41:22 Darth_Wader: you should read everything if you make claims like above Sep 13 21:42:31 hi RP , do you have a sec please :) Sep 13 21:42:55 Marex: a sec, yes Sep 13 21:43:28 RP, well ... I think I found a bug in bitbake, since oe bugtracker appears to be down, can you please check this ? http://cmh.hackndev.com/pastebin/39/ Sep 13 21:43:41 RP, it is NOT a correct fix, but with this is "worksforme" Sep 13 21:44:01 the problem is that bitbake adds "-r 1" to svn parameters everytime Sep 13 21:44:37 just when there is no "rev=something" specified in .bb file, it adds -r 1 automagically and breaks everything Sep 13 21:45:10 yea and errr ... I was warned by psokolovsky ... dont kill me if it's not a bug, spare me :B Sep 13 21:45:19 Marex: that's intended behavior Sep 13 21:45:49 psokolovsky, it's intentionally broken .. I see :) Sep 13 21:45:52 Marex: needed revision, or ${AUTOREV} must be selected explicitly Sep 13 21:46:39 psokolovsky, you should have said that earlier :/ Sep 13 21:46:45 Marex: something like that. I myself have no idea why RP chose -r1 after all, instead of failing out, but that's it Sep 13 21:47:38 Marex: so, I think, just check the stuff once again. it kinda should work. if not, I'll try it tomorrow, and we'll come back to RP then... Sep 13 21:47:53 psokolovsky: question about defconfigman.. Sep 13 21:48:14 psokolovsky, so how should I fix linux-hackndev-2.6 to work with this intentionally broken bitbake ? Sep 13 21:48:29 psokolovsky: to select irda I need to set IRDA=m as PXA_FICP=m is not enough - how does it works for you? Sep 13 21:48:52 Marex: Aha, gotcha! You should learn to tag your stuff! Sep 13 21:49:16 psokolovsky, lucky you Im not going to ukraine any soon ;-E Sep 13 21:49:35 but i have some friends there Sep 13 21:49:39 psokolovsky: beware Sep 13 21:49:42 XD Sep 13 21:49:47 hrw: kernel.subsys.irda ?= $1 Sep 13 21:50:08 ok Sep 13 21:51:40 thx Sep 13 21:52:18 Marex: just rtfm, and you won't need to go here except to drink beer ;-). also maybe some football match? Sep 13 21:55:20 Marex: Its not intentionally broken, the revision of "1" is just a gentle hint that you might want to set a SRCREV for that package Sep 13 21:55:45 Marex: See sane-srcrevs.inc and add an appropriate entry Sep 13 21:55:56 RP, ok, I understand that now, thank you :) Sep 13 21:57:32 Marex: except that local workaround, and the right one is to tag your stuff properly Sep 13 21:58:46 psokolovsky: the more I look at defconfigman the more I do not understand how does it works ;D Sep 13 21:58:46 dcopidl-native is broken too Sep 13 21:59:06 hrw, is it really that surprising? ;) Sep 13 21:59:24 hrw: simple feature structures with trivial constraint propagation ;-) README should have it all. Sep 13 21:59:31 hrw: what's exactly unclear? Sep 13 21:59:52 Marex: all what I am looking for is a way to simplify generating Zaurus configs as now they are more and more not in sync Sep 13 22:01:08 hmmm Sep 13 22:01:34 my gumstix just stopped booting at "Remounting root file system...." Sep 13 22:01:41 no errors upstream .. Sep 13 22:01:42 wait Sep 13 22:01:44 done Sep 13 22:01:47 heh Sep 13 22:01:49 psokolovsky: I want to enable usb hosts, all usb drivers, all pcmcia drivers so looks like I need to create whole base configs from scratch Sep 13 22:01:54 hrw, I fully understand your situation, pxa27x based palms are the same Sep 13 22:01:55 udev on first boot issue Sep 13 22:02:04 hrw: defconfigman can do that. the easiest way is to just start with some existing config. Caveat: it may need a bit of patching for other trees (but it for example work for well for hackndev, but that;s downstream) Sep 13 22:02:22 hrw: NO! Sep 13 22:02:54 psokolovsky: I also do not like current base setups - for example pcmcia debugging by default to yes Sep 13 22:03:32 hrw: caveat, it should be done well, and finegrained, that's why h2200 issue for example. have a look at kernel.subsys.pcmcia.drivers.wifi for example Sep 13 22:04:17 hrw: that's because pcmcia needs debugging here ;-). that's what I mean by "tweaking" Sep 13 22:07:10 hi Sep 13 22:07:17 ewi lost network Sep 13 22:08:27 Crofton: so how goes your gumstix effort? Sep 13 22:08:56 psokolovsky: ok, I will try to get atleast half of config and will show you machine config to review - ok? Sep 13 22:08:56 psokolovsky: which device you suggest to take as base? Sep 13 22:09:18 sakoman, good Sep 13 22:09:35 I have actually loaded Angstrom-minimal onto one Sep 13 22:09:46 I am using the network board Adam sent me Sep 13 22:09:53 Hrww: sure, will be glad to help. But please be prepared that it might take some effort to get it boot at first. h4000 is good as the base. Sep 13 22:10:09 I fixed the haning boot by turning on the full tmpfs in the defconfig Sep 13 22:10:20 psokolovsky: I do not plan to compile before I will get config which I will accept Sep 13 22:10:38 ok Sep 13 22:10:59 sakoman, enough for tonight, need to get caught up on stuff at home Sep 13 22:11:21 Glad to hear you are up & running! Got networking? Sep 13 22:12:55 on the old board, not on the new board Sep 13 22:13:15 I need to pull Craig's OE branch and study the diffs Sep 13 22:13:36 anyway gotta run. Sep 13 22:13:39 Is his branch pullable now? Sep 13 22:13:42 May be back on later Sep 13 22:13:50 OK, have a nice evening! Sep 13 22:13:56 He sent Adam an email, but we think the server is down atm Sep 13 22:14:21 which one? just mtn pull`d a couple of minutes ago Sep 13 22:14:33 a private vendor branch Sep 13 22:22:40 [machine.ide] Sep 13 22:22:41 kernel.subsys.ide=y Sep 13 22:22:51 psokolovsky: why I am getting ide=m with it? Sep 13 22:23:23 hrw: something overrides it later Sep 13 22:26:23 psokolovsky: so looks like I need to create empty distro config? Sep 13 22:26:55 psokolovsky: I've finally remembered to reply to that angstrom mail... Sep 13 22:27:15 hrw: I suggest that you start working on machine first. later, we can work on bringing angstrom config to good common shape Sep 13 22:27:21 RP: thanks! Sep 13 22:28:13 psokolovsky: http://pastebin.com/d50c581af for c7x0 Sep 13 22:29:44 hrw: well, yes, seems to be good enough. compiles? Sep 13 22:30:20 psokolovsky: I do not care does it compiles as long as it will not generate proper config for me Sep 13 22:31:45 hrw: compiles as does defconfigman parses it? ;-) next, need to formulate what settings are wrong with current angstrom distro config there Sep 13 22:32:10 Attempt to set feature 'kernel.subsys.sound' to incompatible value 'y' (already 'm' set by 'kernel.subsys.sound') Sep 13 22:32:18 so forget about ALSA in modules? Sep 13 22:32:27 03rwhitby 07org.oe.dev * r2d2fba06... 10/ (9 files in 5 dirs): free42: HP42 RPN calculator, also added to openmoko-feed Sep 13 22:32:46 thx to quicksand for that one Sep 13 22:33:14 hrw: I think copy and paste of the relavent defconfig sections might be easier for the zaurus use... Sep 13 22:33:43 hrw: If we actually got them synced up with the device differences documented it wouldn't be that difficult to maintain... Sep 13 22:33:47 RP: thats what I suspect too Sep 13 22:34:13 hrw: no, just need to resolve this conflict Sep 13 22:34:17 RP: but psokolovsky so often mention defconfigman that I decided to give that tool last chance Sep 13 22:34:26 hrw: :) Sep 13 22:34:30 psokolovsky: what conflict? Sep 13 22:34:33 [machine.sound] Sep 13 22:34:33 kernel.subsys.sound = m Sep 13 22:34:35 kernel.subsys.sound.alsa.soc = m Sep 13 22:35:43 hrw: well, you could give it only 1st chance at this time. and noone said it would be easy. moreover, I'd personally wouldn't push you to use it for zaurus. To do it *right*, there would be lots of organizational effort, and that's probably out of scope now ;-( Sep 13 22:36:16 RP: zaurus.conf + MACHINE.conf + common.conf? Sep 13 22:36:35 hrw: y vs m. try to find what sets it to '= y' and change to '?= y' Sep 13 22:37:19 hrw: Something like that. In theory there are just MACHINE + common Sep 13 22:38:18 psokolovsky: kernel-subsys.conf->[kernel.subsys.sound]->SOUND=$1 is a problem Sep 13 22:38:38 hrw: using ?= might fix it Sep 13 22:40:05 it does Sep 13 22:41:07 RP: right Sep 13 22:41:32 someone know tool which extract common parts from files? Sep 13 22:41:55 hrw: You want an inversediff :) Sep 13 22:42:42 hrw: in the end you'll get the same crap with own idiosyncrasies (if only only better) ;-) Sep 13 22:43:16 psokolovsky: in best situation I will end with two files: common + machine-related Sep 13 22:43:35 psokolovsky: The zaurus machine case is actually really simple Sep 13 22:43:40 hrw: I suggest you look and kernel miniconfigs from ex(?)-busybox maintainer, that funky guy. his crap reminded me of defconfig Sep 13 22:44:16 psokolovsky: There are maybe about 10-20 entries that differ between the machines Sep 13 22:44:17 hrw: you have one now, why bother? defconfigman's aim is to have fine-grained features structure, split among maintainable files ;-) Sep 13 22:44:50 RP: I bet. but lack of structuralism. anyway, I just describe what aims I had with defconfigman. Sep 13 22:45:46 psokolovsky: I make no comment on defconfigman, I've never used it ;-) Sep 13 22:52:20 RP: pxa25x/pxa27x/sa1100 + common + machine Sep 13 23:13:13 RP: conf-c7x0 conf-poodle conf-pxa25x conf-pxa27x conf-qvga conf-tosa conf-vga conf-zaurus Sep 13 23:13:36 RP: where conf-pxa27x is conf-akita-spitz Sep 13 23:14:18 and tosa conf suxx due to lack of its support in 2.6.22 Sep 13 23:27:47 looks like I got it cleaned Sep 13 23:30:00 hrw: I'll be interested to see the result :) Sep 13 23:30:16 but tomorrow now ;-) Sep 13 23:30:20 'night all Sep 13 23:30:29 cu RP Sep 13 23:30:32 is mr brabec present? Sep 14 01:00:34 03mickeyl 07org.oe.dev * r235d3c4d... 10/ (4 files in 4 dirs): Sep 14 01:00:34 libgsmd: add modified start script for fic-gta01 due to problems w/ UART multiplexing Sep 14 01:00:34 See http://bugzilla.openmoko.org/cgi-bin/bugzilla/show_bug.cgi?id=788 for more information Sep 14 01:00:34 This is hopefully just a temporary solution... Sep 14 01:00:39 03rwhitby 07org.oe.dev * r93ee90ab... 10/ (1 conf/distro/openmoko.conf): openmoko.conf: comment out moko-autorev.inc to enable stability for the openmoko community. Change approved by mickeyl. Sep 14 01:01:13 those two commits should be cause for celebration in openmoko land ... Sep 14 01:01:18 ;) Sep 14 01:01:40 ;) Sep 14 01:02:08 and tomorrow (or rather today) I will post another set of bugs into bugzilla Sep 14 01:04:08 cool Sep 14 01:04:59 will go through list from last OHOM meeting Sep 14 01:08:22 bye all **** ENDING LOGGING AT Fri Sep 14 02:59:56 2007