**** BEGIN LOGGING AT Tue May 17 23:59:56 2005
May 18 00:18:53 <koen> good morning all
May 18 00:19:58 <koen> 12677 koen      25   0 42064 8632 2984 R 96.5  0.4   4:14.06 python
May 18 00:20:15 <koen> 0.4% memory usage when having parsed 2.3k .bbs
May 18 00:29:12 <Spyro> cool
May 18 00:29:36 <koen> that's 8 mb in case you are wondering 
May 18 00:29:44 <Spyro> v. cool
May 18 00:29:53 <koen> it is about 2 or 3 times slower at parsing in the first run
May 18 00:30:41 <robert> milan vs liverpool .championship final
May 18 00:38:11 <mithro> wtf!? 10M!
May 18 00:38:30 <mithro> yay! I can use OE again :P
May 18 00:39:03 <koen> you could even use it on your pda
May 18 00:39:15 <Spyro> hehe
May 18 00:41:13 <mithro> yeah
May 18 00:41:22 <mithro> native building
May 18 00:41:35 <mithro> koen: who did the work fixing up bitbake?
May 18 00:42:02 <koen> zecke used proti's patches and extended them
May 18 00:50:49 <XorA> morning
May 18 00:50:55 <BigAl> Has anybody tried to build gnu-config-native recently?
May 18 00:50:57 <koen> hey XorA 
May 18 00:51:11 <koen> BigAl: http://bugs.openembedded.org/show_bug.cgi?id=2
May 18 00:51:25 <XorA> koen: those patches ready to roll now then?
May 18 00:51:47 <BigAl> koen: Yup, that'd be it.
May 18 00:54:13 <Spyro> is there any reason the oe packages are held in bitkeeper?
May 18 00:54:46 * XorA curses bitkeeper
May 18 00:55:12 <koen> Spyro: distributed repositories
May 18 00:55:29 <Spyro> koen: its a bar to my contributing back :(
May 18 00:55:52 <koen> Spyro: july 1st it'll be all over
May 18 00:56:13 <koen> then we'll switch to baz, monotone or cogito
May 18 00:57:22 <Spyro> argh!
May 18 00:57:26 * Spyro has none of those
May 18 00:57:33 <Spyro> whats wrong with svn and cvs ?
May 18 00:57:43 <koen> it's not distributes
May 18 00:57:53 <koen> we want the repo's on our computers
May 18 00:57:58 <koen> commit offline and stuff
May 18 00:58:22 <Spyro> :/
May 18 00:58:50 <koen> cogito is the new kernel thing
May 18 01:01:40 <koen> Spyro: it also needs to support file moving and renaming, cvs+svn don't support that 
May 18 01:01:46 <koen> not with losing the history
May 18 01:23:20 <RP> Spyro: Send patches to the oe mailing list
May 18 01:23:40 <RP> morning all btw
May 18 01:24:00 <RP> BigAl: The pcmcia problem crept in between rc4-mm1 and rc4-mm2
May 18 01:26:23 <Spyro> RP: PCMCIA problem ?
May 18 01:27:43 <RP> Spyro: pcmcia is dead on rc4-mm2. See my post to LKML
May 18 01:28:20 <Spyro> eep
May 18 01:28:45 <RP> Spyro: I think some proc entries have disappeared which is stopping cardmgr from playing nicely
May 18 01:29:00 <RP> Spyro: Just for reference, bk does work well for managing the openembedded tree. I have to give it that much....
May 18 01:29:24 <Spyro> from what I hear bk is very good. but I run an open source box.
May 18 01:30:12 * RP shrugs
May 18 01:30:31 <Spyro> RP: yeah, I know, Im nuts, balha blah :)
May 18 01:30:53 <XorA> Spyro: as of 1st July it all becomes irrelevant anyway
May 18 01:31:06 <RP> XorA: indeed
May 18 01:31:07 <Spyro> XorA: indeed
May 18 01:31:13 <Spyro> GMTA :)
May 18 01:36:25 <robert> hi XorA
May 18 01:38:36 <koen> hey greg72 
May 18 01:38:57 <robert> hi koen
May 18 01:39:06 <koen> hey robert 
May 18 01:42:11 <BigAl> RP: Sounds about right, cardmgr was complaining about proc
May 18 01:44:50 <koen> RP: pcmcia-utils not working right yet?
May 18 01:44:50 <robert> yesterday, kergoth told me it wasnt correct to set ASSUME_PROVIDED = "virtual/arm-linux-gcc" and PATH=/usr/local/arm/2.95.3/bin:$PATH when i wanted to use local arm-linux-gcc.
May 18 01:44:52 <robert> who can tell me how to set?
May 18 01:44:58 <koen> I saw your post to lkml
May 18 01:46:20 <RP> koen: I've not tried it yet. I really need to get a working build system running and try that...
May 18 01:47:12 <robert> koen:do you know that?
May 18 01:47:17 <koen> messed up OE tree?
May 18 01:47:20 <koen> robert: no idea, sorry
May 18 01:47:42 <robert> koen:)
May 18 01:47:48 <koen> robert: see the zauruskernels wiki page on openembedded.org
May 18 01:48:16 <robert> i have seen many times ,but i have no idea too.
May 18 01:49:25 <robert> waiting for kergoth.........
May 18 01:50:01 <RP> robert: kergoth is in the process of moving so you may have a long wait :-(
May 18 01:50:14 <robert> bad news!
May 18 01:50:16 <RP> robert: I'm afraid I don't know the answer to your quesiton either
May 18 01:50:27 <XorA> robert: ASSUME... goes in conf/local.conf
May 18 01:50:28 <RP> (other than the zaurus kernels page)
May 18 01:50:40 <BigAl> robert: I've got this in my local.conf: ASSUME_PROVIDED = "virtual/arm-linux-gcc-2.95"
May 18 01:50:51 <XorA> robert: path is set however you set path in your shell
May 18 01:50:58 <hrw|gone> robert: 1. add ASSUME into conf/local.conf
May 18 01:51:12 <robert> haha .hrw
May 18 01:51:13 <hrw|gone> robert: 2. set PATH in shell before you start bitbake
May 18 01:51:26 <robert> hrw:i have set it
May 18 01:52:14 <robert> but ..........build failed 
May 18 01:54:03 <hrw|gone> ~shoot openlogging.org
May 18 01:54:04 * ibot shoots openlogging.org in the foot with a cork gun!
May 18 01:56:41 <hrw|gone> Error: Max pending log exceeded, commit aborted.
May 18 01:56:50 <hrw|gone> This error indicates that the BitKeeper program is unable to contact
May 18 01:56:50 <hrw|gone> www.openlogging.org to send the ChangeSet log[s]. The number of unlogged
May 18 01:56:51 <hrw|gone> changesets have exceeded the allowed threshold of 40 changesets.
May 18 01:56:53 <hrw|gone> fuckers
May 18 01:57:20 <robert> hrw:yesterday, kergoth told me it wasnt correct to set ASSUME_PROVIDED = "virtual/arm-linux-gcc" and PATH=/usr/local/arm/2.95.3/bin:$PATH when i wanted to use local arm-linux-gcc.
May 18 01:57:39 <hrw|gone> robert: and he said what to do?
May 18 01:57:58 <robert> dont told me
May 18 01:58:44 <wrobbie> should I put libstdc++6 or libstdc++ in RDEPENDS? With '6', I get:
May 18 01:58:45 <wrobbie> | Collected errors:
May 18 01:58:45 <wrobbie> | ERROR: Cannot satisfy the following dependencies for task-bootstrap:
May 18 01:58:46 <wrobbie> |        libstdc++
May 18 01:59:07 <XorA> hrw|gone: wouldnt it also require a PREFFERED_VERSION_gcc=2.95.3?
May 18 01:59:54 <hrw|gone> maybe
May 18 02:01:01 <robert> XorA;my board use its own arm-linux-gcc, so the arm-linux-gcc oe build for me isnt the one i need
May 18 02:03:15 <hrw|gone> I switch to R/O copy of oe tree here..
May 18 02:03:54 <CIA-9> 03koen 07 * r1.3332 10openembedded/packages/gnome/goffice_0.0.1.bb: goffice_0.0.1.bb: fix deps
May 18 02:04:16 <koen> damn
May 18 02:04:26 <koen> cache udpating is broken with the new bitbake
May 18 02:04:42 * koen CACHE=""
May 18 02:04:50 <XorA> robert: PREFFERED_VERSION combined with ASSUME_PROVIDED I think will stop oe building one
May 18 02:21:52 <hrw|gone> heh.. robert quit before I get fix for him
May 18 02:23:09 <hrw|gone> robert: you have binutils, gcc-cross, glibc headers/libs/etc own - right?
May 18 02:23:22 <hrw|gone> ASSUME_PROVIDED = "virtual/arm-linux-gcc"
May 18 02:23:24 <hrw|gone> ASSUME_PROVIDED += "binutils-cross"
May 18 02:23:24 <hrw|gone> ASSUME_PROVIDED += "glibc"
May 18 02:23:45 <hrw|gone> robert: but you need to copy glibc headers/libs into tmp/staging/
May 18 02:24:03 <robert> i have only "virtual/arm-linux-gcc"
May 18 02:24:58 <hrw|gone> robert: come to query private (from beggining)
May 18 02:25:45 <robert> in distro.conf 
May 18 02:25:47 <robert> PREFERRED_VERSION_glibc-initial = "2.2.5"
May 18 02:25:49 <robert> PREFERRED_VERSION_glibc = "2.2.5"
May 18 02:26:53 <hrw|work> robert: somewhere in irc client you have open window with "Hrw" in title - look there
May 18 02:53:55 <hue> hi, all. I want to build a uclibc-based mini-distro on x86, how should I start it?
May 18 02:55:38 <hue> any suggestion? or hints.
May 18 02:58:15 <zecke> hey
May 18 02:58:40 <hue> :zecke hi
May 18 02:58:54 <hrw|work> hi zecke 
May 18 02:59:43 * zecke announces free cookies to the one finding my bug with data_dict.py. It must be quite obvious
May 18 03:04:02 <koen> zecke: broken cache?
May 18 03:04:33 <zecke> koen: no
May 18 03:04:48 <koen> it doesn't update the cache over here after editing .bbs
May 18 03:06:47 <CIA-9> 03koen 07 * r1.3333 10openembedded/packages/gnome/goffice_0.0.1.bb: goffice_0.0.1.bb: fix more deps, bitbake -b is bad
May 18 03:06:52 <zecke> koen: then it didn'tdo it before either ;)
May 18 03:07:06 <zecke> koen: you can go to bb/make.py and uncomment two print messages
May 18 03:07:07 <koen> zecke: which patches are needed to get bittest running?
May 18 03:07:14 <zecke> none
May 18 03:07:53 <hrw|work> ok. maybe with my help robert will get his build going
May 18 03:08:12 <koen> zecke:   File "/data/build/oe/bitbake/lib/bb/__init__.py", line 283, in decodeurl
May 18 03:08:12 <koen>     m = re.compile('(?P<type>[^:]*)://((?P<user>.+)@)?(?P<location>[^;]+)(;(?P<parm>.*))?').match(url)
May 18 03:08:16 <koen> TypeError: expected string or buffer
May 18 03:09:13 <koen> that's with lastest svn
May 18 03:09:39 <zecke> koen: clean the cache first
May 18 03:10:28 <koen> where does bittest store it's cache?
May 18 03:12:31 <zecke> koen: oh that is with bittest?
May 18 03:12:39 <zecke> koen: ah right
May 18 03:12:40 <koen> yes
May 18 03:15:03 <CIA-9> 03zecke123 * r170 10bitbake_qa/source_checker/source_checker.py: 
May 18 03:15:04 <CIA-9> bitbake_qa/source_checker/source_checker.py:
May 18 03:15:04 <CIA-9>  -A less intrusive patch to lib/bb/fetch.py was
May 18 03:15:04 <CIA-9>  checked in for bitbake1.2.1.
May 18 03:15:04 <CIA-9>  -switch data and url again
May 18 03:18:01 * koen svn up
May 18 03:19:43 <koen> ok, bittest is running
May 18 03:20:10 <koen> zecke: does it log errors somewhere? 
May 18 03:20:48 <zecke> koen: nope
May 18 03:21:00 <zecke> koen: you could change the output 'handler' though
May 18 03:21:02 <koen> hmmm
May 18 03:21:14 <zecke> koen: it prints them at the end (when all tests ran)
May 18 03:21:59 <koen> I'd it bittest to print it to a file that I can parse and put on the web
May 18 03:22:21 <koen> so we can have green and red cells for packages
May 18 03:22:53 <zecke> koen: just hook a different output formatter to it
May 18 03:23:02 <koen> as a side note: bittest doesn't needs to be serial
May 18 03:23:11 <zecke> koen: but we need someway to overwrite bb.note to output to a file as well
May 18 03:23:34 <zecke> koen: sure it doesn't
May 18 03:24:05 <zecke> koen: but it is hard to do it async, each test should get its own ${WORKDIR}
May 18 03:24:25 <ade|desk> morning all
May 18 03:24:38 <koen> hey ade|desk 
May 18 03:24:49 * zecke starts bitbake programming on a zaurus
May 18 03:25:40 * zecke wonders where to put that doc test utility
May 18 03:27:31 <RP> Is there a known problem with diet-x11?
May 18 03:27:34 <koen> a finished, moving on to b
May 18 03:27:48 <koen> RP: not that I'm aware off
May 18 03:27:53 <koen> RP: build or runtime?
May 18 03:30:07 <RP> koen: build time - library link error
May 18 03:30:26 <koen> RP: it worked yesterday
May 18 03:30:34 <koen> did a familiar and oz build from scratch
May 18 03:30:55 <RP> koen: maybe its a parallel_make problem :-/
May 18 03:31:10 <koen> -j4 over here
May 18 03:31:19 <RP> koen: same here :-/
May 18 03:31:26 <RP> koen: which machine?
May 18 03:31:43 <koen> 312:28 < miromiro> Do sombody work with eCos system ?
May 18 03:31:43 <koen> 12:29 < miromiro> nobody ?
May 18 03:31:51 <koen> that's *one* minute
May 18 03:32:34 <RP> koen: This build is using MACHINE=ipaq-pxa-2.6
May 18 03:32:59 <koen> hmm
May 18 03:33:04 <koen> I tried h3900 and c7x0
May 18 03:33:09 <RP> koen: a build with MACHINE=c7x0 worked for gpe-image :-/
May 18 03:33:57 <koen> oz or fam?
May 18 03:34:04 <RP> koen: oz
May 18 03:34:20 <RP> Is that asking for trouble? :)
May 18 03:35:43 <koen> it's a bit backwards, but it should work
May 18 03:36:11 <koen> since an *ipaq* is not a *zaurus*
May 18 03:36:20 <RP> It hadn't occured to me to change it to be honest
May 18 03:36:37 <koen> it shouldn't matter
May 18 03:36:53 <koen> the only difference is softfloat and the distroname
May 18 03:38:26 <hrw|work> ~lart gnumeric for crashing on save
May 18 03:38:26 * ibot hits gnumeric with an anvil and laughs with a contralto voice ... Haha Ha HA Ha for crashing on save
May 18 03:38:53 <hrw|work> why I have to use oo.org for simple excel editing...
May 18 03:39:24 <RP> http://www.rpsys.net/openzaurus/temp/log.do_compile.3056-1 is the cut down log. The full log is also there
May 18 03:41:01 * koen blaims ccache
May 18 03:41:33 <koen> I disabled ccache months ago for the crap it was giving me
May 18 03:41:49 <RP> It does give some massive speedups :-/
May 18 03:42:10 <koen> I can build a gpe-image from scratch in 3 hours or so
May 18 03:42:15 <RP> Can I disable it for that package to test if its the problem?
May 18 03:42:16 <koen> fast enough for me
May 18 03:42:36 <RP> koen: Its when I'm recompiling the same kernel every 5 minutes that its handy
May 18 03:42:36 <koen> or you can wipe the ccache cache before rebuilding it 
May 18 03:47:36 <RP> koen: Even stranger, cleaning the package and rebuilding worked. No changes to ccache
May 18 03:48:46 * france eyes a new eight processor hammer box....
May 18 03:49:04 <zecke> france: looks like a perfect tinderbox
May 18 03:49:33 <zecke> france: each test case gets two cpus
May 18 03:51:36 <france> heh
May 18 03:51:57 <zecke> france: what are you going to run on it? Solari10 ;)
May 18 03:52:28 <france> heh
May 18 03:52:30 <france> true64
May 18 03:52:31 <koen> slowaris?
May 18 03:53:09 <france> zecke: I have not tried my alpha to arm cross toolchain in a while.
May 18 03:53:25 <france> zecke: I might have an idle quad processor alpha at crl that could be used.
May 18 03:53:41 <zecke> *my eyes are shining*
May 18 03:54:01 <koen> hmmm, alpha, that one is missing in my "collection"
May 18 03:54:23 <france> I think the quad processor has 8M for l2 cache on each cpu.
May 18 03:54:35 <koen> indeed
May 18 03:54:48 <france> and either 8 or 16 GB Dram on the box
May 18 03:54:49 <koen> the only reason I didn't buy one was that it weighed a ton
May 18 03:55:02 <france> yup, that one weights about 145 pounds
May 18 03:55:34 <koen> about 90 kilos
May 18 03:55:43 * koen is not in to imperial measures
May 18 03:55:46 <france> sounds about right
May 18 03:56:08 <france> when I am i the USA, I tend to do as the natives do, but sometimes slip into the usage of meters.
May 18 03:56:43 <france> the us will never go on metric until the sports teams start using it.
May 18 03:57:16 <france> but 90 some odd meters does not sound right for a football field.
May 18 03:57:17 <koen> like the simpsons episode: "we now even run on metric time"
May 18 03:57:23 <france> heh
May 18 04:04:02 <koen> lunch break!
May 18 04:04:11 <koen> bye bye report
May 18 04:04:14 <france> I am off to breakfast
May 18 04:12:00 * france is away: Away
May 18 04:14:03 <zecke> mickey|zzZZzz: I need some help on improving my expression
May 18 04:14:07 <zecke> mickey|zzZZzz: a = None
May 18 04:14:33 <zecke> mickey|zzZZzz: is a None or is the i18n_rev("Verweis") of a None
May 18 04:32:33 <[cc]smart> zecke: python ?
May 18 04:35:42 <hrw|work> [cc]smart: hi
May 18 04:35:48 <[cc]smart> hi :)
May 18 04:36:49 <zecke> [cc]smart: yes
May 18 04:37:01 <zecke> [cc]smart: I'm still confused if I should call it refrences
May 18 04:37:37 <[cc]smart> references ? in what respect ? after your assignment a is defined , but woithout value
May 18 04:38:48 <zecke> [cc]smart: As in "Call by Reference", You hold a reference to a instance
May 18 04:38:54 <zecke> [cc]smart: you pass the reference
May 18 04:39:01 <[cc]smart> There is no instance involved
May 18 04:39:14 <zecke> [cc]smart: that would mean a = None. The reference of a is None
May 18 04:39:40 <[cc]smart> Actually, maybe i'm wrong, but even see no reference involved.
May 18 04:40:21 <[cc]smart> To me its like asking about the first element of an empty array
May 18 04:40:45 <[cc]smart> There is none.
May 18 04:41:02 <koen> there is no spoon
May 18 04:41:32 <CIA-9> 03zecke123 * r171 10bitbake_qa/doc_checker/ (. __init__.py doc_checker.py): (log message trimmed)
May 18 04:41:32 <CIA-9> bitbake_qa/doc_checker:
May 18 04:41:32 <CIA-9>  A small test case that goes through every key in a concrete
May 18 04:41:32 <CIA-9>  bitbake file (this way uses the bbclasses as well).
May 18 04:41:32 <CIA-9>  And checks if a 'doc' flag is present.
May 18 04:41:33 <CIA-9>  The idea came from Chris and now we've a tool to look at it.
May 18 04:41:35 <CIA-9>  It might be better to directly scan the bbclasses but this approach
May 18 04:41:43 <[cc]smart> The only thing visible to me is, that after your assignment, your app won't break accessing a.
May 18 04:42:28 <[cc]smart> You can now do if a: even if a cannot be assigned any sensible thing
May 18 04:42:59 <koen> zecke: bittest is almost done with "c"
May 18 04:45:16 <CIA-9> 03zecke123 * r172 10bitbake_qa/ (bin/bittest lib/bittest/__init__.py lib/bittest/config.py): (log message trimmed)
May 18 04:45:16 <CIA-9> bitbake_qa/lib/bittest/config.py:
May 18 04:45:16 <CIA-9>  Move the configuration evaluating
May 18 04:45:16 <CIA-9>  into config.py.
May 18 04:45:16 <CIA-9> bitbake_qa/lib/bittest/__init__.py:
May 18 04:45:18 <CIA-9>  Make the method available to people using
May 18 04:45:20 <CIA-9>  this module
May 18 04:45:31 <zecke> koen: hehe. We can do better when we tune some wget options
May 18 04:45:35 <zecke> koen: and kill some mirror
May 18 04:45:58 <zecke> I might have broken bittest now :}
May 18 04:47:28 <CIA-9> 03zecke123 * r173 10bitbake_qa/lib/bittest/testresult.py: 
May 18 04:47:28 <CIA-9> bitbake_qa/lib/bittest/testresult.py:
May 18 04:47:28 <CIA-9>  Do not add None references to the list of test results.
May 18 04:47:28 <CIA-9>  If the to be added item is a list, extend it into the
May 18 04:47:28 <CIA-9>  the testresult list
May 18 04:48:10 <zecke> RP: if you want, you can import your data implementation into trunk
May 18 04:48:18 <zecke> RP: I won't make it default yet though ;)
May 18 04:56:36 <zecke> koen: I've the biggest headache with the missing fakeroot
May 18 04:56:45 <zecke> koen: I remember problems when we switched the last time
May 18 04:56:57 <koen> cups is missing too
May 18 04:57:08 <koen> I'll check the log when it's finished
May 18 04:58:29 <koen> I'll take a look at hooking it up to php tonight
May 18 04:58:58 <zecke> koen: I blame ecn for it
May 18 04:59:03 <zecke> koen: php?
May 18 04:59:14 <koen> a php parser for the log
May 18 04:59:28 <koen> make pretty colored output on the web
May 18 04:59:34 <zecke> koen: spyce
May 18 04:59:46 <koen> I don't know python
May 18 04:59:50 <koen> only php
May 18 05:00:08 <koen> mickeyl use spyce, right?
May 18 05:00:22 <koen> s/use/uses/
May 18 05:00:28 <zecke> koen: yupp
May 18 05:07:58 <wrobbie> does anyone else have a weird delay after 'building provider hash' is done (new bitbake)
May 18 05:08:00 <wrobbie> ?
May 18 05:08:08 <koen> yes
May 18 05:09:44 <wrobbie> is it doing something, or waiting areound for something? 
May 18 05:09:57 <zecke> wrobbie: yupp
May 18 05:09:59 <koen> it's doing something
May 18 05:10:12 <zecke> wrobbie: if you use trunk, read the commit messages ;)
May 18 05:10:20 <wrobbie> ah, i will
May 18 05:10:33 <zecke> wrobbie: bitbake is like the crazy wombat it does only use some mb of ram
May 18 05:10:40 <koen> zecke: let's integrate 'sl' into it
May 18 05:11:05 <wrobbie> did i read above that cache handling is bit buggy now? (I had to kill a cache file manually)
May 18 05:11:21 <zecke> wrobbie: nope nothing is buggy, the format is different
May 18 05:11:22 <wrobbie> wait - let me read log msg 1st :)
May 18 05:11:42 <CIA-9> 03koen 07 * r1.3334 10openembedded/packages/gpe-conf/gpe-conf_0.1.24.bb: gpe-conf_0.1.24.bb: add it
May 18 05:11:56 <zecke> cya later
May 18 05:23:17 <wrobbie> well, -c clean and rebuild to re-make a package after hacking its .bb doesn't work anymore. I have to kill the package.bb in tmp/cache too, after -c clean
May 18 05:23:48 <koen> wrobbie: same here
May 18 05:25:26 <wrobbie> ok, buglet then, i guess.. is -c clean <package> followed by bitbake <package> the normal way to deal with .bb hacks? -c <other cmd> never works for me..
May 18 05:25:50 <koen> -c <command> -f should force it
May 18 05:26:14 <wrobbie> but what command? depends on what i hacked in bb?
May 18 05:26:14 <koen> I usually use -b package.bb to rebuild stuff
May 18 05:26:35 <wrobbie> oh, i never used that....
May 18 05:46:03 <VoodooZ_Work> hey
May 18 05:46:18 <VoodooZ_Work> Any major infrastructure changes in the last 48 hours?
May 18 05:46:39 <VoodooZ_Work> My build fails at the first package now (quilt-native) with lots of python errors. 
May 18 05:46:48 <VoodooZ_Work> I get this as the last line: AttributeError: 'module' object has no attribute 'createCopy'
May 18 05:50:39 <wrobbie> did you update bitbake?
May 18 05:53:14 <VoodooZ_Work> When was the last version change?
May 18 05:53:34 <wrobbie> today, loads of work done, according to the log
May 18 05:54:12 <VoodooZ_Work> thanks I'll check it out.
May 18 05:54:17 <koen> about one hour ago
May 18 05:54:24 <Ken|JLime> I updated last night (hadnt updated in ages) and i got same amount off errors
May 18 06:05:46 <Ken|JLime> AttributeError: 'module' object has no attribute 'createCopy'
May 18 06:06:13 <koen> Ken|JLime: have done svn up in the last hour?
May 18 06:06:42 <Ken|JLime> I updated (with bk pull) last night and got the same error
May 18 06:06:59 <Ken|JLime> havent updated in like 2 months so it could be anything
May 18 06:07:02 <koen> did you update bitbake too?
May 18 06:07:17 <RP> zecke: My work isn't ready for adding to bitbake yet. Now your code is in, it gives me something to devlope against though. Watch this space :)
May 18 06:07:23 <Ken|JLime> only bk pull; bk -r co -q
May 18 06:07:50 <koen> Ken|JLime: http://oe-devel.bkbits.net:8080/openembedded/cset@1.3330?nav=index.html|ChangeSet@-7d
May 18 06:08:04 <VoodooZ_Work> sorry guys, I had somebody at my desk...
May 18 06:08:08 * RP builds new opie and gpe images for the c7x0 and hx2750
May 18 06:11:49 <chouimat> ugmorning
May 18 06:12:38 <VoodooZ_Work> What is the latest "stable" version of bitbake anyways? I'm using 1.1 right now and the web page has an old one.
May 18 06:13:30 <VoodooZ_Work> actually, I just realised It's 1.2 so I'll start by updating to that...
May 18 06:13:41 <koen> http://developer.berlios.de/project/showfiles.php?group_id=2701
May 18 06:14:05 <koen> VoodooZ_Work:  http://oe-devel.bkbits.net:8080/openembedded/cset@1.3330?nav=index.html|ChangeSet@-7d
May 18 06:14:50 <mallum> koen: is TSLIB_TSDEVICE set right ?
May 18 06:15:11 <koen> mallum: wrong window, but yes
May 18 06:17:31 <mallum> koen: wrong tab you mean 
May 18 06:17:51 <koen> window in irssi :)
May 18 06:18:24 * chouimat is back.
May 18 06:20:30 <Ken|JLime> Updated bitbake and still getting same error
May 18 06:20:44 <VoodooZ_Work> koen: So I guess I'm wasting my time by installing 1.2 then? 
May 18 06:20:52 <koen> pretty much
May 18 06:20:56 <VoodooZ_Work> ok.
May 18 06:21:08 <VoodooZ_Work> So I guess I have to use svn to get it right?
May 18 06:21:14 * VoodooZ_Work checks his notes...
May 18 06:21:16 <koen> yes
May 18 06:21:26 <koen> check the gettingstarted wiki
May 18 06:21:26 <VoodooZ_Work> ok. thanks. I'll figure it out.
May 18 06:22:18 <VoodooZ_Work> yep. I found my note.
May 18 06:22:23 <VoodooZ_Work> svn co svn://svn.berlios.de/bitbake/trunk/bitbake; cd bitbake; ./setup.py \
May 18 06:22:23 <VoodooZ_Work> install --prefix=/usr/local --install-data=/usr/local/share
May 18 06:22:31 <VoodooZ_Work> kergoth gave me that a while ago.
May 18 06:23:12 <Ken|JLime> VoodooZ_Work, i just did that but the problems still remains
May 18 06:23:15 <wrobbie> 'svn update' if you still have that tree..
May 18 06:23:17 <VoodooZ_Work> ok.
May 18 06:23:51 <VoodooZ_Work> yeah, but I think I did the "easy" method last time using the tarball so..
May 18 06:26:46 <VoodooZ_Work> I got 1.73
May 18 06:27:25 <wrobbie> r173, that's correct
May 18 06:27:36 <VoodooZ_Work> ok
May 18 06:28:15 <VoodooZ_Work> Do I really need to run ./setup though? I just extracted the folder in my /opt
May 18 06:29:08 <wrobbie> I'd remove tmp/cache/*, but I'm a oe newbie too..
May 18 06:29:25 <wrobbie> no need for setup, if you set path to it
May 18 06:29:27 <VoodooZ_Work> hehehe. I've been a newbie for way too long! :)
May 18 06:29:35 <VoodooZ_Work> did that.
May 18 06:29:41 <hrw|work> I have bitbake in path when use OE - never installed it
May 18 06:29:47 <VoodooZ_Work> now it says: NOTE: multiple providers are available (glibc, glibc-intermediate);
May 18 06:29:47 <VoodooZ_Work> NOTE: consider defining PREFERRED_PROVIDER_virtual/armeb-linux-libc-for-gcc
May 18 06:29:49 <wrobbie> me too..
May 18 06:30:15 <VoodooZ_Work> Probably my year old config file!
May 18 06:30:27 <wrobbie> oh-oh :)
May 18 06:31:24 <wrobbie> it's non-lethal I think. I have the same thing, and don't know if i have to set it to 'glibc' or 'glibc-initial'. (i'm only building i686)
May 18 06:31:44 <VoodooZ_Work> I'm build openslug for the NSLU2 (ARM)
May 18 06:31:47 <hrw|work> you can ignore it
May 18 06:32:06 <VoodooZ_Work> Well, it doesn't do anything. as if it aborts.
May 18 06:32:10 <wrobbie> hrw: does it save building stuff if i define it?
May 18 06:32:56 <hrw|work> wrobbie: I dont know which one it had to be - on my setup it will be glibc
May 18 06:33:34 <VoodooZ_Work> In my case it suggested i use: PREFERRED_PROVIDER_virtual/armeb-linux-libc-for-gcc for some reason.
May 18 06:34:00 <wrobbie> no, that's the variable you have to set..
May 18 06:34:27 <VoodooZ_Work> yeah and judging from my local.conf file the notation is different.
May 18 06:34:32 <wrobbie> (or not, doesn't matter apparently)
May 18 06:35:31 <VoodooZ_Work> Anyways, It seems all this is causing more trouble so I'll just revert to 1.2 and wait it out I guess.
May 18 06:36:03 <VoodooZ_Work> eventually the nslu2 repo should sync up or somebody else will get the same problem.
May 18 06:36:16 <hrw|work> VoodooZ_Work: bitbake r170 or newer is needed for current OE
May 18 06:36:36 <hrw|work> and vice versa probably - 
May 18 06:36:37 <VoodooZ_Work> Since today/yesterday?
May 18 06:36:57 <VoodooZ_Work> I've been using 1.1 for while and it started failing yesterday
May 18 06:39:28 <hrw|work> 20 hoursmickeyl1.3330apply zecke's patches to prepare for the low memory bitbake. the repository needs bitbake r159 or better now
May 18 06:39:32 <hrw|work> ops
May 18 06:39:37 <hrw|work> 20 hours mickeyl 1.3330 apply zecke's patches to prepare for the low memory bitbake. the repository needs bitbake r159 or better now
May 18 06:39:50 <hrw|work> since 1.3330 OE needs newer bitbake
May 18 06:40:16 <VoodooZ_Work> ok. I just wonder if the NSLU2 guys are not using older stuff though.
May 18 06:40:35 <koen> hrw|work: that was in the  http://oe-devel.bkbits.net:8080/openembedded/cset@1.3330?nav=index.html|ChangeSet@-7d
May 18 06:40:41 <koen> link I posted earlier :)
May 18 06:40:45 <hrw|work> ok
May 18 06:41:12 <mickeyl> morning
May 18 06:41:16 <koen> good morning mickeyl
May 18 06:43:12 <hrw|work> hi mickeyl 
May 18 06:44:08 <CIA-9> 03mickeyl * r174 10bitbake/bin/bitbake: add missing copyright and fix __ignored_dependencies scoping problem
May 18 06:44:47 <wrobbie> they just keep coming.. :)
May 18 06:47:16 <hrw|work> hmm. maybe I will push AUTHORS change 
May 18 07:00:44 <hrw|work> hmm.. I'm writing 'Projects' page on my website and cannot remember where I put hands in..
May 18 07:05:02 <mickeyl> heh
May 18 07:06:04 <hrw|work> do we have policy for ttf fonts?
May 18 07:06:34 <mickeyl> not yet
May 18 07:09:34 <hrw|work> http://www.hrw.one.pl/index.php/en/projects/
May 18 07:12:32 <zecke> re
May 18 07:12:38 <RP> hi zecke
May 18 07:12:57 <Crofton> I need to pull from bk, and svn to get new bb running?
May 18 07:13:04 <RP> Crofton: yes
May 18 07:14:05 <Crofton> rerun setup.py with what options?
May 18 07:14:27 <zecke> VoodooZ_Work: use tags/bitbake-1.2.1 instead of trunk/bitbake
May 18 07:15:53 <zecke> koen: for your reporting tool, you can easily hook up a different TestOutput
May 18 07:17:03 <koen> ok
May 18 07:17:15 <koen> will try when someone plugs in the power
May 18 07:18:48 <zecke> hehe
May 18 07:18:57 <RP> Crofton: Perrsonally, I don't install bitbake - I just run it from some directory
May 18 07:19:04 * zecke checks options for remote compiling
May 18 07:19:08 * zecke too
May 18 07:19:51 <Crofton> grin, I just built the rpm and installed
May 18 07:19:52 <Crofton> it
May 18 07:20:18 <Crofton> on the last machine I used, we had two people using seperate oe trees with common bb
May 18 07:21:56 <koen> Crofton: same here
May 18 07:23:26 <koen> one tree for work stuff and one personal OE tree
May 18 07:23:35 <koen> with a common bitbake
May 18 07:24:06 <Crofton> we used to do two builds at once, poor machine :)
May 18 07:24:30 * CosmicPenguin fires up the new and improved bitbake
May 18 07:24:48 <Crofton> hmm, raining in Itlay
May 18 07:24:50 <Crofton> er Italy
May 18 07:25:10 <CIA-9> 03zecke123 * r175 10bitbake_qa/conf/bitbake.conf: 
May 18 07:25:10 <CIA-9> bitbake_qa/conf/bitbake.conf:
May 18 07:25:10 <CIA-9>  For running the tests decrease the timeout from 900 seconds
May 18 07:25:10 <CIA-9>  for dns, connect and read to 40 seconds. If that leads to false
May 18 07:25:10 <CIA-9>  positives when fetching source we can easily increse the timeout
May 18 07:25:40 <koen> zecke: flase negatives
May 18 07:25:45 <koen> ehm
May 18 07:25:48 <koen> false even
May 18 07:26:05 <zecke> koen: depends on the view of the test
May 18 07:26:13 <koen> that's true
May 18 07:26:19 <zecke> koen: does it succeed if it doesn't find source?
May 18 07:26:30 <koen> no idea
May 18 07:26:31 <zecke> koen: but you're right admitingly I meant false negatives
May 18 07:28:12 <RP> What's the status of the jlime patches? They look sane enough - should I push them?
May 18 07:28:21 <koen> go ahead
May 18 07:28:53 <koen> looks like we can add sh3 to our new distro
May 18 07:29:23 <CIA-9> 03zecke123 * r176 10bitbake_qa/conf/bitbake.conf: 
May 18 07:29:23 <CIA-9> bitbake_qa/conf/bitbake.conf:
May 18 07:29:23 <CIA-9>  Reduce the number of tries for the same reason. Failing
May 18 07:29:23 <CIA-9>  three times on a non existant sourcefile should be enough.
May 18 07:33:25 <zecke> cya later
May 18 07:40:03 <Crofton> I suppose I should expect bb to be really slow handling bb files the first time after the upgrade?
May 18 07:46:01 <RP> Crofton: The slowness is one of the sideaffects of the decreased memory consumption
May 18 07:46:35 <Crofton> would using psyco help?
May 18 07:50:18 <CosmicPenguin> Come on, just buy the opteron
May 18 07:50:20 <CosmicPenguin> you know you want to
May 18 07:50:50 <Crofton> heh
May 18 07:50:57 <Crofton> yes, but I am cheap
May 18 07:51:02 <hrw|work> CosmicPenguin: you sell them in promotion? cheap opterons with extra coffee?
May 18 07:51:12 <Crofton> and it is much faster on the second go round
May 18 07:51:17 <CosmicPenguin> hrw|work: I wish
May 18 07:55:42 <CoreDump|home> hi
May 18 07:57:02 <CIA-9> 03RP 07 * r1.3329.1.3 10openembedded/conf/distro/jlime.conf: Add jlime distro (patch from Kristoffer Ericson)
May 18 07:57:08 <CIA-9> 03RP 07 * r1.3329.1.2 10openembedded/ (4 files in 3 dirs): Add jornada6xx machine (patch from Kristoffer Ericson)
May 18 07:57:12 <CIA-9> 03RP 07 * r1.3329.1.1 10openembedded/packages/linux/ (linux-jlime_2.6.11.bb linux-jlime-2.6.11/defconfig_jlime): Kernel for jornada6xx and jlime distro from Kristoffer Ericson
May 18 07:57:42 <RP> Crofton: I expect we'll end up making bitbake more efficient. Its just a question of how long that will take
May 18 08:00:43 <CP|Laptop> Crofton: if you are bored, here are some performance hints I found on the web:
May 18 08:00:47 <CP|Laptop> http://wiki.python.org/moin/PythonSpeed/PerformanceTips
May 18 08:01:15 <Crofton> unfortunately I am not bored :(
May 18 08:02:47 <CP|Laptop> You will be when you stare at python parsing all day... :)
May 18 08:05:16 <Crofton> heh, trying to figure out how to get usbmon going right now
May 18 08:05:25 <Crofton> seem to be missing debug_fs ...
May 18 08:18:40 <hrw|work> I want to patch pcmcia-cs to support 2 partitions on cf cards - any objections?
May 18 08:19:27 <wrobbie> who knows how to solve my dependency problem:
May 18 08:19:29 <wrobbie> | ERROR: Cannot satisfy the following dependencies for task-bootstrap:
May 18 08:19:29 <wrobbie> |        libstdc++
May 18 08:19:43 <hrw|work> wrobbie: where you got it?
May 18 08:19:58 <wrobbie> "but it worked yesterday!" :)
May 18 08:20:11 <wrobbie> in making my image
May 18 08:20:37 <wrobbie> I have 'libstdc++6' in RDEPENDS
May 18 08:20:48 <CoreDump|home> hrw|work: sounds great. Any chance to have an automatic detection which partitions are available instead of two hard-coded partitions?
May 18 08:21:09 <wrobbie> dunno who's asking for 'libstdc++', without '6'
May 18 08:21:22 <hrw|work> CoreDump|home: probably can be done by extra script
May 18 08:21:35 <hrw|work> CoreDump|home: I have to look into it in next days
May 18 08:22:27 <CoreDump|home> :)
May 18 08:23:27 <wrobbie> .. if there is some way to trace a dependency, i could investigate further
May 18 08:29:14 <hrw|work> ~shoot dsteen
May 18 08:29:16 * ibot shoots dsteen in the head with a glue gun!
May 18 08:29:36 <mickeyl> ?
May 18 08:29:40 <mickeyl> what did he do?
May 18 08:30:11 <hrw|work> he had to write 'root on sd' howto
May 18 08:30:57 * CoreDump|home is working on a root-on-sd solution ATM
May 18 08:31:44 <hrw|work> why sf.net is so fscking slow...
May 18 08:32:21 <CoreDump|home> +mhentges@chroot01:../OpenEmbedded/tmp/rootfs >du
May 18 08:32:22 <CoreDump|home> 73M     .
May 18 08:32:24 <CoreDump|home> sigh
May 18 08:32:39 <hrw|work> good size for collie
May 18 08:33:32 <CoreDump|home> heh
May 18 08:33:51 <CoreDump|home> actually, this is going to be an ext2 image to boot off of my SD card
May 18 08:34:00 <CoreDump|home> which is only 64MB heh
May 18 08:34:11 <hrw|work> my collie has 64M sd too
May 18 08:34:24 <hrw|work> but I prefer moving /opt/ to it
May 18 08:34:49 <CoreDump|home> ~lart termcap for eating 6MB
May 18 08:34:50 * ibot explains, ever so gently, that if termcap doesn't give the channel more information, they can't help for eating 6MB
May 18 08:35:28 <hrw|work> what for you need full termcap?
May 18 08:35:31 <CP|Laptop> that was stupid, ibot
May 18 08:35:33 <CoreDump|home> cardfs currently weighs in w/ 62MB
May 18 08:35:49 <CoreDump|home> ah, something depends on it. mc for one
May 18 08:36:11 <hrw|work> ok. sent mail to dsteen with question..
May 18 08:36:35 <mickeyl> koen: to prevent longish discussions, i wouldn't call it "Trolltech crap"
May 18 08:36:43 <CoreDump|home> hehe
May 18 08:36:47 <mickeyl> but that's me as a diplomatic person
May 18 08:36:48 <mickeyl> heh
May 18 08:36:59 <mickeyl> (zug)
May 18 08:37:08 <hrw|work> I wrote it simple: Trolltech sucks (on my website)
May 18 08:38:16 <hrw|work> ~seen offroadgeek
May 18 08:38:19 <ibot> offroadgeek <~offroadge@offroadgeek.sustaining.supporter.pdpc> was last seen on IRC in channel #openzaurus, 3d 22h 12m 39s ago, saying: '4gb... I don't know'.
May 18 08:41:19 <koen> mickeyl: well I did respond to someone who clearly lacks reading skills, so I guess not many people will actually read it
May 18 08:41:31 <koen> let alone understand it
May 18 08:41:56 <mickeyl> fair enough
May 18 08:41:59 <mickeyl> Qt/X11 is quite nice
May 18 08:42:00 <mickeyl> :)
May 18 08:42:06 <mickeyl> at least 3.x
May 18 08:42:07 <mickeyl> heh
May 18 08:42:18 <koen> I like gtk and it's modelview
May 18 08:42:27 <koen> or whatever they call it nowadays
May 18 08:42:54 <hrw|work> gtk... I dont like its so called 'filerequester' which is crap
May 18 08:43:06 <koen> it's crap
May 18 08:44:36 <molivier> does oe required a new version of bitbake ?
May 18 08:44:50 <mickeyl> yes
May 18 08:44:54 <mickeyl> read the Changelogs now and then
May 18 08:45:19 <koen> http://oe-devel.bkbits.net:8080/openembedded/cset@1.3330?nav=index.html|ChangeSet@-7d
May 18 08:45:24 <molivier> mickeyl, ok thanks 
May 18 08:45:31 <koen> http://oe-devel.bkbits.net:8080/openembedded/cset@1.3330?nav=index.html|ChangeSet@-7d
May 18 08:45:37 <koen> <- read that link
May 18 08:45:55 <koen> s/"<-"/"^"/
May 18 08:46:47 <CP|Laptop> hehe
May 18 08:47:54 <hrw|work> ok. I go home and then back to the cinema
May 18 08:48:28 <hrw|work> back as when I go from home to cinema I walk near place where I work
May 18 08:49:57 <molivier> sorry , how i apply all this patchs ?
May 18 08:50:05 <wrobbie> off to bed, nightie night
May 18 08:50:13 <hrw|work> molivier: bk pull?
May 18 08:52:16 <molivier> since i have make bk pull this morning i can't build my apps anymore - got many python errors
May 18 08:53:05 <hrw|work> molivier: update bitbake also
May 18 08:53:33 <CP|Laptop> when python isn't the largest memory user on my system
May 18 08:53:34 <CP|Laptop> thats good
May 18 08:53:41 <CP|Laptop> very very good
May 18 08:54:02 <koen> 'svn up'
May 18 08:59:51 <CosmicPenguin> oops - p4 fetch is broken
May 18 08:59:55 <CosmicPenguin> heh
May 18 09:00:01 * CosmicPenguin smacks self across head
May 18 09:00:02 <CosmicPenguin> idiot
May 18 09:00:21 <hrw|sw> cu
May 18 09:00:41 <RP> hrw|sw: enjoy the film!
May 18 09:00:51 <hrw|sw> RP: I will
May 18 09:01:04 <hrw|sw> RP: every saga has the end
May 18 09:01:58 <koen> "I just found out the department is getting moved"
May 18 09:02:11 <koen> that's my boss explaining why my buildbox went offline
May 18 09:02:29 <chouimat> koen: nice
May 18 09:02:41 <molivier> koen, thanks for all - it's work fine now
May 18 09:02:52 <koen> molivier: good to hear
May 18 09:03:14 <koen> chouimat: I knew it was getting moved months ago, but the "when" was an unknown
May 18 09:04:40 <molivier> time to code !
May 18 09:06:06 <CosmicPenguin> great - spilled coffee on myself
May 18 09:06:20 * koen stumbles across a tree containing .oe files 
May 18 09:08:22 <CosmicPenguin> mmm.... disk space
May 18 09:15:56 * RP still has a buildroot somewhere :)
May 18 09:17:46 * koen downloads bitkeeper
May 18 09:17:50 <CosmicPenguin> I've got tux buildroots at home
May 18 09:18:52 <offroadgeek> are any of the oe devs on and have a second?
May 18 09:20:02 <CoreDump|home> shoot
May 18 09:20:11 <RP> offroadgeek: I'd just ask your question
May 18 09:20:27 <koen> http://oe-devel.bkbits.net:8080/openembedded/cset@1.3330?nav=index.html|ChangeSet@-7d
May 18 09:20:33 <koen> pre-emptive paste
May 18 09:20:34 <RP> koen: :)
May 18 09:20:44 <RP> we need to update the topic
May 18 09:21:02 <koen> s/we need/kergoth needs/
May 18 09:21:04 <CosmicPenguin> Who else has oper priviliages - treke?  
May 18 09:21:30 <RP> I don't know...
May 18 09:21:48 <offroadgeek> RP:  did you get an answer?
May 18 09:22:11 <RP> offroadgeek: An answer to what?
May 18 09:22:50 <offroadgeek> RP:  oh.. I misread your line... I thought you had "just asked your question" ;)
May 18 09:23:20 <offroadgeek> no, I have a specific question for one of the oe devs... kinda an offer
May 18 09:23:37 <RP> offroadgeek: We don't know what your question is yet... ;-)
May 18 09:24:35 <offroadgeek> ok... I have a quad-zeon box just sitting racked in my basement turned off, and I'm wondering if any of them want it for a remote build environment?
May 18 09:24:49 * koen looks at zecke 
May 18 09:25:26 <CosmicPenguin> Tinderbox
May 18 09:25:32 <CosmicPenguin> who has rack space?
May 18 09:26:02 <RP> offroadgeek: I'd ask zecke about his tinder box ideas
May 18 09:26:24 <CosmicPenguin> mmm.... tinderbox
May 18 09:26:35 <offroadgeek> zecke:  let's here about your tinderbox ideas ;)
May 18 09:26:56 <koen> I have (although it's being moved now) a server for tinderbox builds, it's just awaiting a working install :)
May 18 09:27:05 <offroadgeek> I actually have three servers just sitting there.. but that one is the best of them
May 18 09:28:52 <CosmicPenguin> Now, all we need is somebody with mad python skilz to put together a tinderbox script and we're there, man
May 18 09:29:39 <koen> isn't tb in perl?
May 18 09:31:22 <CosmicPenguin> koen: yeah, but the client script can be in anything, I think
May 18 09:31:47 <CosmicPenguin> all it needs is mail support - and I figure if you do a python script, you can avoid having the extra perl process
May 18 09:32:06 <CosmicPenguin> I might be wrong, though
May 18 09:32:23 <koen> it's all the same to me
May 18 10:34:00 * chouimat is away: stuff and people are boring in kde-devel
May 18 10:45:06 <treke> bleh
May 18 10:45:16 <treke> so it'll be 900 bucks getting the car repaired
May 18 10:57:31 <zecke> re
May 18 10:57:34 <zecke> _alwin_: hey
May 18 10:57:46 <_alwin_> hi zecke 
May 18 10:58:50 <zecke> CosmicPenguin: right a tinderbox
May 18 10:59:32 <CosmicPenguin> zecke: tell offroadgeek - he was offering hardware
May 18 11:01:06 <zecke> CosmicPenguin: CRL might have a Quad Box for us as well
May 18 11:01:30 <Spyro> Mmmmmm. Quad...
May 18 11:03:17 <CosmicPenguin> With the bitbake changes, and a parallel make, we can probably do four or five builds at one time
May 18 11:04:19 <treke> now that memory consumtion is down I can run one
May 18 11:04:20 <treke> maybe two
May 18 11:07:48 <CosmicPenguin> I have three gigs on my Opteron, and I could do three, but it would struggle
May 18 11:07:54 <CosmicPenguin> I bet I can do five or six now
May 18 11:08:01 <CosmicPenguin> Not that I have that many things to compile
May 18 11:08:15 <treke> zecke_: almost got kdepim built ;p
May 18 11:08:19 <zecke_> hehe
May 18 11:08:22 <zecke_> now update
May 18 11:08:46 <treke> hehe once I get the debian package to build I can update
May 18 11:09:30 <Crofton> ~nag zecke_ 
May 18 11:10:33 <treke> zecke_: why can't you people be more like gnome :p
May 18 11:10:55 <zecke_> treke: like not being able to save the history in a gnome-terminal :}
May 18 11:11:02 <Crofton> bitbake is smaller than gnome now
May 18 11:11:10 <zecke_> treke: we don't want to be dropped from Slackware
May 18 11:11:16 <treke> zecke_: making it easier to upgrade just one thing :p
May 18 11:11:27 <treke> I migh get to port ktnef to osx
May 18 11:11:51 <zecke_> treke: hehe not that I ever understood why ktnef is important for mail :}
May 18 11:12:05 <treke> outlook sometimes sends them out :)
May 18 11:12:32 <treke> and our mac users like being able to get ot their attachments :p
May 18 11:12:40 <zecke_> treke: on the one hand KDE is more integrated code and gui wise than gnome
May 18 11:13:13 <zecke_> treke: but yes with non stable libraries it is bad :}
May 18 11:14:37 <CosmicPenguin> Somebody should add a bitbake version, and bail out in the base.bbclass in OE if the versions don't match
May 18 11:14:47 <CosmicPenguin> s/version/version variable/
May 18 11:15:04 <kergoth> there is a version variable, but nobody ever updates it.
May 18 11:15:11 <kergoth> bb.__version__ or something like that
May 18 11:15:23 <kergoth> the bitbake binary dumps the version when you pass it -v, iirc
May 18 11:16:34 <CosmicPenguin> setup.py
May 18 11:17:12 <kergoth> that has it too, but you cant easily get to that, and that isnt guaranteed to exist in an installed bitbake
May 18 11:17:14 <kergoth> heh
May 18 11:17:28 <roh> hm... seems the opensimpad 0.9.0rc1_opie is crashing at some points...
May 18 11:19:23 * zecke_ hides
May 18 11:20:04 <zecke_> kergoth: is __version__ something python spefic? a number or a str?
May 18 11:20:22 <kergoth> python specific?
May 18 11:20:47 <zecke_> kergoth: hehe sorry. Something to define by convention?
May 18 11:20:54 <zecke_> roh: hehe what is rc1?
May 18 11:21:09 <zecke_> roh: could you've restored your simpad
May 18 11:21:11 <kergoth> oh, dunno
May 18 11:22:07 <Crofton> so when are you headed for MN kergoth 
May 18 11:22:27 <kergoth> in a day or two.  need to finish cleaning up the apartment and the car, and take the car in for its 9000 mile service before the drive
May 18 11:22:30 <roh> zecke_: yep.. i have
May 18 11:22:50 <roh> zecke_: i now own a custom-built passive jtag ;)
May 18 11:23:01 <roh> zecke_: http://phoenix.hyte.de/jtag_simpad/
May 18 11:24:02 <roh> zecke_: and i flashed the 0.9.0rc1_opie_XXL.jffs2 filesystem after the bootloader-updater (both from the opensimpad-forum) to test-drive it.. but it crashed fairly often
May 18 11:24:12 <zecke_> kergoth: do not disturb could we put a bitbake-1.2.1 tarball somewhere?
May 18 11:24:48 <zecke_> roh: crashing is uncommon - as I would describe it
May 18 11:25:13 <zecke_> actually my SIMpad didn't crash for a long term (only at CeBit while using wellenreiter)
May 18 11:25:39 <treke> zecke_: stil no good :(
May 18 11:25:45 <roh> i suspend and resume (to get the noise away), open a browser, fullscreen it, go to heise.de and scroll around a bit on the startpage with the touchscreen.. then it freezes
May 18 11:26:03 <zecke_> roh: that is a 2.4 kernel?
May 18 11:26:27 <zecke_> roh: could you check if it froze (is ssh accessible/pingable)
May 18 11:26:32 <roh> stays running (on console) but the screen is frozen and even the cursor keys on the side are dead.. stopping and running opie again (via initscript) does not help
May 18 11:26:56 <zecke_> roh: Keys are always dead until you set QWS_KEYBOARD="Buttons"
May 18 11:27:03 <roh> 2.4.25-vrs2-pxa1-jpm1
May 18 11:27:12 <treke> zecke_: it still deletes shit like crazy
May 18 11:27:14 <roh> zecke_: the keys work before the freeze
May 18 11:27:27 <zecke_> roh: weird
May 18 11:27:38 <zecke_> treke: damn (KitchenSync that is?)
May 18 11:27:50 * chouimat is back.
May 18 11:28:16 <treke> zecke_: yup
May 18 11:28:36 <zecke_> treke: which resources do you use?
May 18 11:28:42 <treke> remote and local
May 18 11:30:29 <treke> zecke_:  both sides are set up to sync the local file against a file on a server
May 18 11:30:36 <treke> over fish
May 18 11:31:33 <zecke_> treke: happen to have a debug build?
May 18 11:31:40 <treke> nope, but I can rebuild it 
May 18 11:31:53 <treke> lemme try this
May 18 11:32:10 <treke> maybe if I go a->b instead of a->c and b->c
May 18 11:33:30 <zecke_> treke: oh wait you use the same resource/file more than once?
May 18 11:33:54 <treke> zecke_: my work machine syncs with a file on my web server
May 18 11:34:04 <treke> zecke_: and then I was trying to sync that with my home machine
May 18 11:34:14 <treke> so that I could keep them both in sync with each other
May 18 11:37:48 <CIA-9> 03zecke123 * r177 10bitbake/bin/bitbake: 
May 18 11:37:49 <CIA-9> bitbake/bin/bitbake:
May 18 11:37:49 <CIA-9>  Update the version to 1.2.1 (now a string)
May 18 11:38:37 <kergoth> zecke_: you changed the bitbake binary?  __version__ in bin/bitbake is the version of the binary.  __version__ in lib/bb/__init__.py is the version of the bitbake python module.
May 18 11:38:41 <kergoth> heh
May 18 11:38:50 <kergoth> guess it makes sense for them to match though :)
May 18 11:39:26 <CIA-9> 03kergoth * r178 10bitbake/ (lib/bb/__init__.py setup.py): Change the bb.__version__ to 1.2.1, and adjust setup.py to actually import the bb module to obtain the version it uses.
May 18 11:39:50 <zecke_> damn
May 18 11:40:21 <CIA-9> 03kergoth 07bitbake-1.2.1 * r179 10/bitbake/: Tag bitbake 1.2.1.
May 18 11:40:28 <treke> zecke_: I'll try two local connectors
May 18 11:41:20 <kergoth> zecke_: should be ready to do a ./setup.py sdist and upload the tarball to berlios for release
May 18 11:41:38 <zecke_> kergoth: I'm not allowed to 
May 18 11:41:44 <kergoth> hmm
May 18 11:41:54 <zecke_> I can do a tarball now (thanks to your setup.py info)
May 18 11:42:40 <kergoth> hmm, what ht ehell
May 18 11:42:52 <kergoth> the bitbake-1.2.1 tag is showing the old __version__ in setup.py
May 18 11:43:16 <zecke_> kergoth: I did svn copy was that wrong? :}
May 18 11:43:30 <kergoth> bah, you did the svn copy before changing the versions :P
May 18 11:43:47 <zecke_> kergoth: Yeah yesterday :}
May 18 11:49:13 <koen> hmmm, this uniprocessor box is an order of magnitude slower 
May 18 11:49:19 * koen gets some coffee
May 18 11:55:40 <CIA-9> 03kergoth 07bitbake-1.2.1 * r180 10/ (bitbake/ lib/bb/__init__.py setup.py): Undo that last extranious svn cp, and pull over the change to trunk which sets bb.__version__ to 1.2.1 and makes setup.py import bb to get to that.
May 18 11:55:48 <kergoth> there, thats better
May 18 11:55:50 <kergoth> heh
May 18 11:56:09 <zecke_> :)
May 18 11:56:36 <zecke_> I know your busy, but could we create a tar ball
May 18 11:56:41 <koen> one hundred and eighty!!
May 18 11:58:03 <CosmicPenguin> sold
May 18 11:58:20 <kergoth> zecke_: write a human comprensible changelog for the release.
May 18 11:58:30 <kergoth> s/human/user/
May 18 11:58:33 <koen> "It's better, use it"
May 18 11:58:41 <kergoth> hehe
May 18 11:58:54 <zecke_> :}
May 18 11:59:13 <zecke_> kergoth: Yeah technical writing is obviously not (yet) my strongest part
May 18 11:59:28 <CosmicPenguin> thats a common programmer affliction
May 18 11:59:29 <zecke_> kergoth: I wanted to comment on the API differences
May 18 11:59:47 <kergoth> right, API differences are good to note, but now we need a summary that a user can understand
May 18 12:00:15 <zecke_> hehe, ok "Nothing changed, just use it"
May 18 12:00:35 <koen> "uses far less RAM now, use it"
May 18 12:00:47 <zecke_> koen: no 1.2.1 does not include that part
May 18 12:00:51 <koen> damn
May 18 12:00:53 * zecke_ hopes that it doesn't
May 18 12:00:58 <kergoth> 1.2.1 is the groundwork only
May 18 12:01:05 <kergoth> i only pulled from trunk the changes i mad,e not the rst :)
May 18 12:01:43 <koen> zecke_: could you incorporate a "calculating foo, please wait" message after the provider hash has been built?
May 18 12:02:01 <zecke_> koen: no. I prefer to fix the shortcoming
May 18 12:02:08 <koen> even better
May 18 12:02:20 <CosmicPenguin> "If this pause pisses you off, bother zecke"
May 18 12:02:27 <kergoth> zecke_: current oe repo requires recent trunk, or 1.2.1?
May 18 12:02:37 <kergoth> "Incremental release to lay the groundwork necessary for upcoming substantial performance and memory usage improvements." <-how's that for the 1.2.1 notes?
May 18 12:02:49 <koen> oooh
May 18 12:02:53 <koen> buzzwords
May 18 12:02:56 <CosmicPenguin> ship it!
May 18 12:02:56 <kergoth> hehe
May 18 12:03:05 <zecke_> kergoth: yupp put that in, and create an API.changes
May 18 12:04:35 <koen> treke: did you change the default email for bugzilla yet?
May 18 12:04:44 <kergoth> zecke_: https://developer.berlios.de/project/showfiles.php?group_id=2701
May 18 12:04:48 <kergoth> zecke_: doing the freshmeat release now
May 18 12:05:13 <treke> koen: no I'll do it right now
May 18 12:05:41 <CosmicPenguin> Careful - this is starting to act like a real project... :)
May 18 12:06:20 <kergoth> hehe
May 18 12:06:22 <zecke_> are you going to mail oe@handhelds.org to inform about the new requirement?
May 18 12:06:39 <kergoth> you can
May 18 12:06:44 <koen> I'll do it
May 18 12:06:46 <zecke_> ok
May 18 12:07:09 <koen> trunk or 1.2.1?
May 18 12:07:15 <zecke_> 1.2.1
May 18 12:07:28 <kergoth> k, freshmeat release submitted
May 18 12:07:55 <kergoth> hey [g2] 
May 18 12:08:07 <[g2]> kergoth, HEY!!
May 18 12:08:11 * [g2] hugs kergoth 
May 18 12:08:18 <zecke_> thanks kergoth
May 18 12:08:28 <kergoth> np
May 18 12:08:37 <kergoth> actually moving forward with the move has broken my rut
May 18 12:08:41 <kergoth> feel like coding or something
May 18 12:08:49 <[g2]> excellent
May 18 12:08:50 <kergoth> guess i should finish cleaning so i can leave instead though :)
May 18 12:09:05 <zecke_> yeah get off that fscking computer ;)
May 18 12:09:13 <[g2]> you know you can't keep a good man down for long :)
May 18 12:09:36 <RP> I have a user "getting excited" becuase he's written a perl script that can apply my patches based on the linux-openzaurus .bb file...
May 18 12:09:53 <zecke_> hehe
May 18 12:10:03 <zecke_> bitbake oz.bb -cpatch
May 18 12:10:08 <CosmicPenguin> whatever works, I guess
May 18 12:10:13 * kergoth starts a clone of the current oe tree so he can build himself a new dev env for colinux, and heads into the kitchen to make sure he can actually see the countertops
May 18 12:10:37 <[g2]> kergoth, I'm one commmand short of having my first knoppix livecd, just add soure and meta data and then build openslug/unslung
May 18 12:10:48 <RP> I just can't quite believe some people....
May 18 12:10:50 <[g2]> ok... then maybe some testing :)
May 18 12:11:30 <koen> mail sent
May 18 12:11:48 <zecke_> [g2]: As you might have read, we've plans to setup a complete QA infrastructure for our packages
May 18 12:12:13 <zecke_> [g2]: I think once we've it, it will be a pleasure to build unslung configs as well
May 18 12:14:38 <[g2]> zecke_, THX, openslug first please, and I'm out-of-the-loop on that.  Is there a place to read more ?
May 18 12:15:34 <zecke_> [g2]: not yet
May 18 12:15:50 <[g2]> teaser! :)
May 18 12:16:21 <zecke_> ;)
May 18 12:17:28 <[g2]> zecke_, so are you going to tell me more or make me sweat ?
May 18 12:17:59 <zecke_> [g2]: We've now bitbake_qa that can run tests under different configurations
May 18 12:18:19 <zecke_> [g2]: a tinder client could now regulary run any or all of the tests
May 18 12:18:58 <zecke_> [g2]: it could go as long as having a base toolchain on a machine and run bitbake moo on the newyl checked in files
May 18 12:20:24 <[g2]> zecke_, sounds interesting... I'm going to be setting something up for OpenSlug and all its packages
May 18 12:21:25 <[g2]> this knoppix disk (and the follow-on with a snapshot of the metadata and open tarballs) would allow a user to just build on any x86 box that had some memory and disk space
May 18 12:28:06 <kergoth> fyi, handhelds.org cant host us anymore
May 18 12:28:11 <kergoth> as of the 31st of htis month
May 18 12:28:59 <RP> kergoth: can't host what exactly?
May 18 12:30:56 <koen> I haven't seen a mail on admin@ about that
May 18 12:31:29 <RP> I'm not on admin@ :-/
May 18 12:31:39 <koen> I am
May 18 12:31:46 <koen> (but not an admin)
May 18 12:32:11 * CosmicPenguin tries to gather motivation to update xfce
May 18 12:32:13 <RP> Are any other hh.org services affected?
May 18 12:38:34 <zecke_> mickeyl: ping
May 18 12:38:49 <zecke_> type(bb.data.init()) dict
May 18 12:40:41 <kergoth> RP: openembedded.org.
May 18 12:40:54 <kergoth> RP: they could host a handhelds.org subdomain for us, but not our domain
May 18 12:41:12 <kergoth> should probably move it to berlios or sf for the time benig
May 18 12:41:13 <kergoth> being
May 18 12:41:34 <kergoth> koen: france emailed me
May 18 12:41:47 <koen> aha
May 18 12:42:13 <koen> hah
May 18 12:42:49 <koen> the pmamobilecafe dweep is reporting me the the relevant authorities because I accused him of spamming 
May 18 12:44:29 <RP> kergoth: I guess that's not too bad. I was expecting something much worse...
May 18 12:53:28 <zecke_> WTF
May 18 12:53:44 <Crofton> ?
May 18 12:56:00 <zecke_> somehow a old data.pyc comes back and strikes me...
May 18 12:56:23 <zecke_> even if I delete it, it is created again (from the wrong source)
May 18 12:56:43 <Crofton> I hate it when stuff like that happens
May 18 12:58:05 <koen> treke: would it be possible to write a bugzilla trigger that sends a msg to #oe?
May 18 12:58:35 <koen> s/possible/easy/
May 18 13:03:28 <RP> BugBot? :)
May 18 13:10:34 <zecke_> mickeyl: can you confirm that bb.data.getVar is slower than data.getVar slower than getVar?
May 18 13:11:39 <zecke_> kergoth: what regexp is COMPATIBLE_HOST? is MACHINE included?
May 18 13:12:51 <zecke_> ah ok found
May 18 13:12:54 <zecke_> HOST_SYS only
May 18 13:14:50 <zecke_> kergoth: mind if I add DISTRO and MACHINE as well?
May 18 13:15:43 <zecke_> kergoth: or should I add COMPATIBLE_DISTRO, COMPATIBLE_MACHINE?
May 18 13:16:56 <kergoth> i dont really have a preference either way.  pb_ added that, see what he thinks
May 18 13:17:18 <zecke_> okay :)
May 18 13:18:44 <kergoth> but i'm glad you're planning on doing that, not having one for MACHINE has been bugging me for a while now :)
May 18 13:19:37 <zecke_> kergoth: I hate that bbfiles use anonymous python functions to raise the SkipPackage exception
May 18 13:37:39 <Twiun> yo
May 18 13:39:55 <zecke_> pb_: Could we make COMPATIBLE_HOST a bit more generic
May 18 13:40:07 <zecke_> pb_: there is need to have it available for MACHINE and DISTRO as well
May 18 13:40:28 <zecke_> pb_: examples are unslung/opensl* and the latest OZ ATI Core modules
May 18 13:41:53 <pb_> zecke_: can you explain a bit more about that?
May 18 13:42:24 <pb_> are you saying that the aticore modules need to be different for each MACHINE?
May 18 13:45:43 <zecke_> pb_: some binary only modules only apply to a specefic machine
May 18 13:46:06 <zecke_> pb_: some packages are distribution specefic
May 18 13:46:37 <pb_> distribution specific in what way?
May 18 13:46:44 <treke> abi
May 18 13:46:45 <treke> ?
May 18 13:46:57 <koen> CONFFILES_nylon = /etc/blah
May 18 13:47:14 <treke> aticore has two different source tarballs
May 18 13:47:15 <koen> SOMEVAR_familiar = foo
May 18 13:47:23 <treke> one for softfloat and one for hardflaot
May 18 13:47:25 <koen> treke: only one works
May 18 13:47:36 <koen> aticore..tar.bz2 doesn't exist
May 18 13:47:46 <treke> koen: it's under hard iirc
May 18 13:47:52 <treke> ask mithro :p
May 18 13:47:55 <pb_> treke: abi differences wouldn't be sufficient to merit a COMPATIBLE_HOST kind of thing.
May 18 13:48:09 <zecke_> pb_: look at the unslung binary only packages
May 18 13:48:39 <zecke_> pb_: or even kernels if you build hh2.4.0-test0 it fails with a defconfig-collie missing
May 18 13:48:45 <koen> treke: could ne, but the .bb is broken
May 18 13:48:47 <pb_> treke: COMPATIBLE_HOST exists purely to stop a "world" build from choking on a package that is hopelessly uncompilable on a particular architecture (i.e. has x86 assembler in the .c files).
May 18 13:49:15 <treke> koen: feel free to find another way to do that :p
May 18 13:49:26 <treke> since softfloat isn't in overrides :p
May 18 13:49:28 <kergoth> kernels would be a good example of a truly machine specific package. it might _build_, but it sure as hell isnt going to work
May 18 13:49:28 <pb_> If the package compiles on your host, it is COMPATIBLE in that sense, irrespective of whether the output is actually useful for anything.
May 18 13:49:38 <kergoth> course preferences _should_ ensure that they arent attempted
May 18 13:49:40 <kergoth> heh
May 18 13:49:44 <kergoth> heh
May 18 13:49:45 <zecke_> koen: the problem with ATI Core is that TARGET_FPU is not set at all
May 18 13:50:35 <zecke_> pb_: ok kernels for iPAQs do not compile with a Zaurus MACHINE set
May 18 13:50:45 <pb_> kernels aren't part of world 
May 18 13:51:04 <treke> chkhinge
May 18 13:51:08 <pb_> you'd only get a particular kernel built if you asked for it, in which case you deserve what you get.
May 18 13:51:21 <zecke_> pb_: not with bitbake_qa :}
May 18 13:51:23 <treke> it only makes sense in the context of a zaurus build
May 18 13:51:46 <pb_> treke: sure, but we don't care whether it makes sense, only whether it can be compiled or not.
May 18 13:52:30 <kergoth> what about binary only packages, like the sharp-binary-only stuff?  that isnt compiled at all
May 18 13:52:51 <pb_> if it isn't compiled at all, the chances are the package will build fine on any host.
May 18 13:52:59 <pb_> so, by definition, COMPATIBLE with all hosts.
May 18 13:53:00 <pb_> heh
May 18 13:53:11 <schurig> kergoth: Hi Ho!
May 18 13:53:17 <kergoth> schurig: hey
May 18 13:53:45 <pb_> packages like that should be setting their PACKAGE_ARCH to make sure that they aren't actually _installable_ on the wrong host, but that's another issue.
May 18 13:53:56 <schurig> kergoth: BTW, next week we'll get our second baby
May 18 13:54:05 <schurig> (and we still need good names)
May 18 13:54:18 <kergoth> congrats.. and good luck :)
May 18 13:54:19 <woglinde> schurig holger
May 18 13:54:29 <schurig> woglinde: that's my name :-)
May 18 13:54:36 <zecke_> schurig: but a good one
May 18 13:54:44 <woglinde> schurig holger jr.
May 18 13:54:59 <zecke_> schurig: anekin
May 18 13:54:59 <schurig> woglinde: Holger II Schurig :-)
May 18 13:55:15 <schurig> zecke_: nah, he went to the dark side
May 18 13:55:26 <zecke_> schurig: he returned though
May 18 13:55:55 <zecke_> pb_: ok but isn't SkipPackage raised even if I do bitbake -b that_file.bb?
May 18 13:56:47 <pb_> yeah
May 18 13:57:22 <zecke_> pb_: so If I build a hh.org kernel for collie, I will get an error file not found(defconfig-collie)
May 18 13:57:31 <pb_> indeed.
May 18 13:57:33 <pb_> don't do that, then.
May 18 13:57:35 <zecke_> pb_: so at that point with my set machine I'm not able to build did
May 18 13:58:01 <zecke_> pb_: how should bitbake_qa know that it shouldn't touch that package if no exception is thrown?
May 18 13:58:04 <gb2> anyone know if it's possible to trick out the linker to statically link in a .so as if it were just another object?
May 18 13:58:23 <pb_> zecke_: I have no idea what bitbake_qa even is.
May 18 13:58:39 <pb_> let alone how it should proceed in any given set of circumstances.
May 18 13:58:42 <koen> treke: mickeyl implemented the OE variant of the C ? operator 
May 18 13:58:51 <treke> dont know
May 18 13:58:52 <koen> treke: that could be used for TARGETFPU stuff
May 18 13:58:55 <treke> did he?
May 18 13:59:04 <koen> he did
May 18 13:59:12 <zecke_> pb_: hehe a small tool that goes through all BBFILES and runs some tests on it
May 18 13:59:17 <zecke_> pb_: with varying configuration
May 18 13:59:34 <kergoth> zecke_: perhaps it shouldnt touch packages that are excluded from world.
May 18 13:59:47 * kergoth shrugs
May 18 14:00:06 <zecke_> kergoth: I would prefer to not have a hide for QA option
May 18 14:02:34 <pb_> kergoth: trouble is, an awful lot of packages are excluded from world, in the expectation that bitbake will chase dependencies.
May 18 14:02:34 <pb_> I guess bitbake_qa doesn't do that.
May 18 14:02:46 * kergoth nods
May 18 14:03:09 <pb_> for example, anything that PROVIDES a virtual is excluded, which is how the kernels get left out.
May 18 14:03:27 <zecke_> pb_: yes bitbake_qa just plainly iterates over all found files
May 18 14:03:36 <zecke_> pb_: and for QA I think that makes sense
May 18 14:04:20 <pb_> yeah
May 18 14:04:58 <zecke_> pb_: I try to find a better argument for COMPATIBLE_HOST, _MACHINE, _DISTRO
May 18 14:05:10 <pb_> I guess I see your point of view, but I'm kind of loath to add a new "this package is only compatible with <...>" policy diktat variable.
May 18 14:05:19 <pb_> is it only the kernels that have this problem?
May 18 14:05:29 <zecke_> pb_: regardless of justification some packages already exclude themselves with anonymous python functions
May 18 14:05:30 <pb_> if so, they could just raise SkipPackage by hand.
May 18 14:05:42 <kergoth> bleh.  so much cleaning to do, and so much me not wanting to do it
May 18 14:05:43 <pb_> yeah, like those other packages already do
May 18 14:06:23 <zecke_> pb_: hehe, excuse meI'm the OO guy, if more than n packages do the same trick I want to include some common method ;)
May 18 14:07:19 <pb_> zecke_: that's a valid desire, and generally one that I share, but in this case it is outweighed by not wanting to codify bad practice.
May 18 14:07:24 <zecke_> it is not too urgent though
May 18 14:07:39 <zecke_> okay
May 18 14:08:04 <pb_> if the kernel, and a few other packages, have bits of ad-hoc python in them, we can just write that off as "oh, yeah, those crazy packages do weird stuff".
May 18 14:08:27 <zecke_> inherit weird_package
May 18 14:08:38 <zecke_> WEIRD_COMPTABIBLE_HOST = "oz"
May 18 14:08:50 <pb_> if we add a new COMPATIBLE_HOST type mechanism, people will assume that the mechanism is a legitimate thing they can validly use.
May 18 14:09:18 <zecke_> right and it isn't, virtual/* is the preferred approach
May 18 14:09:55 <treke> the bit that annoys me about skip package is that it makes it hard to mirror all of our source
May 18 14:10:08 <treke> since I can't get the SRC_URIs without setting the MACHINE flag
May 18 14:10:18 <treke> and I can't realyl tell what MACHINE should be
May 18 14:10:46 <treke> some random bb just decides it doesnt want to be parsed and tells me to piss off
May 18 14:11:41 <pb_> right
May 18 14:12:04 <zecke_> hmm I've some problems with overrides
May 18 14:12:09 <pb_> it's not very satisfactory.  I don't really remember why I did it that way in the first place.
May 18 14:12:24 <pb_> might just have been a temporary lapse of reason on my part, or something.
May 18 14:12:34 <treke> pb_: it was probably just the simplest way to do it at that point
May 18 14:12:35 <zecke_> if I've PREFERRED_VERSION_gpe-should = ""
May 18 14:12:49 <zecke_> should it go away when I call bb.data.update_data?
May 18 14:13:10 <pb_> heh.  actually, I have no idea what bitbake will do in that situation.
May 18 14:13:29 <pb_> a variable whose value is the empty string is not quite the same as one that isn't defined at all.
May 18 14:13:35 <treke> pb_: perhaps it could be somehow be modified to raise an error after the parser is done?
May 18 14:13:35 <zecke_> it leaves PREFERRED_VERSION_gpe-shield as key
May 18 14:13:49 <zecke_> pb_: sorry it has some value :}
May 18 14:14:10 <pb_> oh, right, I misunderstood what you were asking.
May 18 14:14:34 <pb_> no, I don't think update_data() deletes variables that have participated in overrides handling.
May 18 14:15:01 <pb_> treke: yeah, seems reasonable.  we could probably figure out a way to have it blow up between do_patch() and do_configure() or something.
May 18 14:15:34 <pb_> or even after parsing but before the build starts.
May 18 14:16:07 <treke> before the build makes the most sense I think
May 18 14:16:59 <Ken|JLime> I've cleaned out my entire oe dir and updated bitbake and still get the same errors
May 18 14:17:00 <treke> that way the behavior is the same as it is now, with the exception that you can still parse the bb if appropriate
May 18 14:17:10 <Ken|JLime> last error being AttributeError: 'module' object has no attribute 'createCopy'
May 18 14:19:06 <CosmicPenguin> woo-hoo!  Attribute error!
May 18 14:20:07 * CosmicPenguin goes to lunch
May 18 14:20:20 <Ken|JLime> :P
May 18 14:20:36 <Ken|JLime> seriously though, ive updated bitbake and still got same error
May 18 14:20:49 <Ken|JLime> its a clean oe folder just downloaded
May 18 14:21:42 <koen> Ken|JLime: bitbake 1.2.1 or svn r180?
May 18 14:22:42 <Ken|JLime> getting BitBake Build Tool Core version 1.1, bitbake version 1.2
May 18 14:22:42 <Ken|JLime>  with bitbake --version
May 18 14:23:08 <koen> you need 1.2.1
May 18 14:23:34 <koen> http://download.berlios.de/bitbake/bitbake-1.2.1.tar.gz
May 18 14:24:54 <Ken|JLime> Hmm, i deleted my bitbake folder and downloaded again, must have been something i did wrong before
May 18 14:25:22 <Ken|JLime> getting correct version now, so it most likely works. Thx koen
May 18 14:25:32 <koen> np
May 18 14:58:27 * zecke_ runs
May 18 14:58:53 <CIA-9> 03ich 07 * r1.3336 10openembedded/conf/ (documentation.conf bitbake.conf): (log message trimmed)
May 18 14:58:53 <CIA-9> OpenEmbedded Configuration Description:
May 18 14:58:53 <CIA-9>  Initial and totally incomplete attempt to document
May 18 14:58:53 <CIA-9> variables OpenEmbedded offers to configure the build. Documentation
May 18 14:58:53 <CIA-9> will be saved as 'doc' flag in the variable. We can easily extract
May 18 14:58:53 <CIA-9> it and create proper description out of these keys.
May 18 14:58:55 <CIA-9> Comment/Add your favorite key in conf/documentation.conf and we're
May 18 14:59:06 <zecke_> damn wtf
May 18 14:59:11 <pb_> heh
May 18 14:59:17 <zecke_> sorry I've removed BK_USER from my bashrc
May 18 14:59:51 <zecke_> I started to run due my bad technical writing skills
May 18 15:09:07 <CIA-9> 03zecke123 * r181 10bitbake/lib/bb/fetch.py: 
May 18 15:09:07 <CIA-9> bitbake/lib/bb/fetch.py
May 18 15:09:07 <CIA-9>  Try to partially apply
May 18 15:09:07 <CIA-9> http://wiki.python.org/moin/PythonSpeed/PerformanceTips?highlight=%28performance%29#dots avoid
May 18 15:09:07 <CIA-9> finding data in bb
May 18 15:18:23 <CIA-9> 03koen 07 * r1.3338 10openembedded/conf/documentation.conf: documentation.conf: some more doc stuff
May 18 15:18:39 <zecke_> koen: thank you
May 18 15:18:53 <koen> zecke_: np
May 18 15:18:55 <zecke_> mickeyl: any: how do I check if a file is present with python?
May 18 15:19:01 <zecke_> or should I just use os.stat?
May 18 15:20:00 <zecke_> os.path.exists?
May 18 15:21:36 <koen> http://guessnet.alioth.debian.org/ looks nice
May 18 15:24:59 <treke> os.path.exists probably is the best way unless  you are just trying to open the file
May 18 15:25:17 <treke> then you can just use open and catch the exception
May 18 15:27:25 <koen> 'night all
May 18 15:28:42 <Spyro> any X hackers about?
May 18 15:30:35 <zecke_> mickeyl: how can I shellQuote something with pythong before executing it?
May 18 15:35:30 <Ken|JLime> I'm doing a distro for different archs and i want to let the distro file set the preferred kernel name (and not the machine)
May 18 15:35:51 <Ken|JLime> would "PREFERRED_PROVIDER_virtual/kernel = "jlime-kernel-${TARGET_PREFIX}" accomplish this?
May 18 15:36:12 <Ken|JLime> then i just add bb files for example "jlime-kernel-sh3 (if sh3 was target_arch)
May 18 15:36:34 <pb_> if you want TARGET_ARCH, say that.
May 18 15:36:49 <pb_> but otherwise, yes, that would work.
May 18 15:36:53 <Ken|JLime> ah, thx
May 18 15:37:11 <CIA-9> 03zecke123 * r182 10bitbake/lib/bb/fetch.py: 
May 18 15:37:12 <CIA-9> tbake/lib/bb/fetch.py:
May 18 15:37:12 <CIA-9>  Stop sf.net from bull?#*... us. We will fail when we wanted to
May 18 15:37:12 <CIA-9>  download a file and got something with a different name.
May 18 15:37:12 <CIA-9>  This should solve sf.net related problems as we go immediately to
May 18 15:37:12 <CIA-9>  the next mirror and it should work for all other packages as well.
May 18 15:37:22 <pb_> if you stick with TARGET_PREFIX, you will have to name your kernels "jlime-kernel-sh3-linux-" (complete with trailing hyphen) which is probably not what you want.
May 18 15:37:22 <zecke_> RP: Spyro : sf.net should no longer bother us
May 18 15:37:31 <Spyro> zecke_: you fixed it?
May 18 15:37:41 <Spyro> zecke_: you rock :)
May 18 15:37:55 <zecke_> Spyro: the removal part is dangerous (not more than other bitbake stuff)
May 18 15:37:59 <CosmicPenguin> zecke_: Hey, thats a simple fix
May 18 15:38:16 <CosmicPenguin> I was thinking - sessh, we're going to have to figure out if the file is html or not
May 18 15:38:30 <zecke_> CosmicPenguin: besides os.system() with an unquouted variable... :}
May 18 15:38:32 <RP> zecke_: excellent :)
May 18 15:38:54 <Spyro> is there a procedure for updating bitbake?
May 18 15:38:55 <zecke_> CosmicPenguin: I guess if we want to unpack barrfoo.tar.bz2 put downloaded private.movie
May 18 15:39:00 <zecke_> CosmicPenguin: we should fail anyway
May 18 15:39:04 <zecke_> Spyro: svn up
May 18 15:39:08 <Spyro> or do I just nuke the bitbake/ tree and re-install it ?
May 18 15:39:11 <CosmicPenguin> zecke_: yeah - It will work 
May 18 15:39:18 <Spyro> ah yes, of course :)
May 18 15:42:38 <zecke_> ok cya guys
May 18 15:42:53 <pb_> zecke_: later
May 18 15:44:20 <zecke_> RP: how far is your bitkenstein? (bitbake with 90% of C/C++ code)?
May 18 15:45:06 <zecke_> I just considered the possibility to hook up bitbake-data at one point into the python codebase...
May 18 15:45:12 <zecke_> anyway now I'm really off to bed
May 18 15:49:07 <RP> zecke_: It needs some attention. I haven't had any time to spend on it :-/
May 18 15:49:19 <RP> 'night zecke_
May 18 15:49:29 <CIA-9> 03RP 07 * r1.3335.1.1 10openembedded/packages/linux/linux-openzaurus_2.6.12-rc4-mm2.bb: linux-oz-2.6: Fix pcmcia problem introduced in -mm2
May 18 15:49:42 <zecke_> RP: I seem to be in a hurry, but I'm just tired :)
May 18 15:49:45 <Ken|JLime> what does target_prefix translate to?
May 18 15:54:41 <RP> zecke: The bikenstein just saves and restores data from a C module at the moment. I need to make it handle muliple dicts as I didn't understand that bit of bitbake when I wrote it
May 18 15:55:11 <RP> I also have an idea for flattening the data store and doing a supercow type thing, as I mentioned when we were talking about databases
May 18 15:57:00 <RP> Ahhrg. Someone is trying to use the 2.6 spitz kernel and complaining about the patches not applying :-/
May 18 15:57:08 <zecke_> hehe
May 18 15:57:08 <CosmicPenguin> stupid xfce - too many packages
May 18 15:57:14 <zecke_> cya tomorrow
May 18 15:57:16 <CosmicPenguin> but thank goodness its not Gnome
May 18 15:59:54 <CosmicPenguin> Does anybody even use xfce-4.0.5 any more?
May 18 16:00:07 <CIA-9> 03RP 07 * r1.3340 10openembedded/packages/linux/linux-openzaurus_2.6.12-rc4-mm2.bb: linux-oz-2.6: Hopefully make spitz patches apply
May 18 16:12:39 <RP> Adding this to documentation.conf was not a clever idea: T            = "Tempory directory within the WORKDIR"
May 18 16:13:18 <RP> /me now has a directory within the workdir called "Tempory directory within the 
May 18 16:13:19 <RP> WORKDIR"
May 18 16:14:27 <kergoth> haha
May 18 16:16:28 <CIA-9> 03RP 07 * r1.3341 10openembedded/conf/documentation.conf: Fix documentation error that was corrupting $T. You end up with a work directory called "Tempory directory within the WORKDIR" and whilst that may be true... :)
May 18 16:18:52 <treke> RP: look at the local.conf.sample 
May 18 16:40:25 <RP> treke: What have I done?
May 18 16:41:59 <RP> I'm off to bed. 'night all
May 18 16:46:19 <treke> for an example of how to put a comment in the conf file that wont work
May 18 16:46:45 <Spyro> RP: still here?
May 18 16:46:53 <Spyro> RP: I got X working on the e330
May 18 16:47:04 <Spyro> RP: I disabled pointer updates altogether
May 18 16:47:40 * Spyro pokes RP with a sharp stick
May 18 17:10:33 <[g2]> Ok so I built my first Knoppix LiveCD today which was simply a 3.8.2 Knoppix minus OpenOffice, Thunderbird, and kstars 
May 18 17:11:00 <[g2]> plus some stuff to build Openslug (you know the drill python-psyco etc...)
May 18 17:11:26 <[g2]> I missed SVN on the first go-round, but I'll add it in next
May 18 17:12:28 <[g2]> Then with a bitbake snapshot, OE metadata, and tarballs for unslung/openslug we'll have a complete buildable OpenSlug/Unslung distro
May 18 17:12:44 <[g2]> We were toying with the name Sluggix
May 18 17:16:10 <[g2]> However, less Openslug/Unslung tarballs, it's really OEKnoppix
May 18 17:16:52 <[g2]> Are you guys interested in having something like that avialable to the common user ? Or is OE dev or the highway ?
May 18 17:18:35 * chouimat found out today that answering to "Where do you see you in 10 years?" :" We all know that the job market in the world is changing so the only honest answer I can give you is being able to watch all the Star Wars movies and the Clone Wars cartoons in the right order" is not a good answer for a job interview ... even on the phone
May 18 17:31:16 <[g2]> chouimat, looking for a job ?
May 18 17:31:33 <chouimat> [g2]: since november
May 18 17:31:45 <[g2]> that's a while
May 18 17:31:55 <[g2]> sorry to hear that you are still looking 
May 18 17:32:11 <chouimat> [g2]: it's ok I have contracts to do once in a while
May 18 17:32:25 <[g2]> that's good!
May 18 17:32:35 <[g2]> what kind of work do you do ?
May 18 17:32:44 <chouimat> firewall stuff
May 18 17:33:24 <[g2]> linux firewall ?
May 18 17:33:35 <chouimat> [g2]: the interview was on the phone conference call (stupid imho) and I  got one of the two interviewers laughing 
May 18 17:33:51 <chouimat> linux and freebsd (mopstly freebsd (embedded))
May 18 17:34:11 <[g2]> sounds like the other guys was quite serious
May 18 17:34:16 <[g2]> :(
May 18 17:34:54 <[g2]> chouimat, you may be interested in http://www.gentoo.org/proj/en/base/embedded/gnap-userguide.xml
May 18 17:34:57 <chouimat> [g2]: I actualy prefers getting a not serious enough than a lack of ambition
May 18 17:35:22 <[g2]> so what's your ambition ?
May 18 17:35:40 <chouimat> [g2]: a good and fun job, a new wife and a family 
May 18 17:35:57 <[g2]> and are you old-school freebsd ?
May 18 17:36:30 <chouimat> the one I build is using 4.11 ... 5.3 and 5.4 doesn't seems to work on the 16MB flash I'm using
May 18 17:36:57 <chouimat> this gnap thing is interesting
May 18 17:37:19 <[g2]> iirc it's a hardened uclibc toolchain
May 18 17:38:19 <[g2]> what arches are you familiar with ?
May 18 17:38:46 <chouimat> x86,arm and slowly learning ppc
May 18 17:39:00 <[g2]> which arm ?
May 18 17:39:22 <chouimat> zaurus for now
May 18 17:56:43 <mithro> is there an easy way to say to the kernel only compile a certain part of the kernel?
May 18 18:32:18 <robert_> kergoth:
May 18 20:44:52 <robert_> kergoth:
May 18 21:01:42 <kergoth> robert_: what?
May 18 21:09:12 <subdue> "as the world turns"
May 18 21:09:16 <subdue> (very slowly)
May 18 21:24:33 <_ggilbert> chouimat|Zzzz: xandros "it's greek for bollocks"
May 18 21:41:36 <[g2]> anyone know what kind of NX support is in 2.6.11+ for AMD64 ?
May 18 23:39:15 <robert__> kergoth:
May 18 23:42:21 <kergoth> robert__: what?
May 18 23:42:50 <robert__> kergoth:you told me it isnt correct of setting ASSUME_PROVIDED = "virtual/arm-linux-gcc-2.95" and PATH=/usr/local/arm/2.95.3/bin:$PATH.
May 18 23:42:51 <robert__> could you tell me how to do ? i have no idea.
May 18 23:45:44 <kergoth> that isnt what you told me you set.
May 18 23:45:52 <kergoth> and why are you asking me? i'm not the only one here who can help you
May 18 23:46:03 <kergoth> if you have a question, ask it and someone will respond
May 18 23:46:13 <robert__> ...
May 18 23:50:02 <robert__> sorry to trouble you
May 18 23:56:07 <kergoth> that PATH and ASSUME_PROVIDED is from the _zaurus kernels_ wiki page.  if you arent targeting a zaurus, that isnt correct.
May 18 23:58:02 <jbowler> kergoth: a question, I'm looking at a change to the openslug config of busybox (i.e. a change in packages/busybox/busybox-1.00/openslug).  When I make a change like this I need to bump the PR, but if I do that I bump the PR for all OE builds, even though I only change nslu2 stuff - how do you guys feel about that?
**** ENDING LOGGING AT Wed May 18 23:59:57 2005