irclog2html for #openembedded on 20030910

00:00.32chouimatI liked Final Fantasy Tactics on the psx
00:01.12_Psychoeverquest is addictive its not because its a game, its because its a social thing with a chat
00:01.14_Psychoyou know like irc
00:01.53W8TVIcool...
00:01.57W8TVIupgrade done...
00:02.33chouimatlol http://geocities.com/pupitetris/Code/Freeciv-Worms/screenshot.png
00:02.38kergothi should allocate some more space to ccache
00:02.44kergothshit keeps falling out of the ccache
00:02.48kergothheh
00:02.59gb2treke: i love RPGs, except for combat
00:03.00chouimathttp://geocities.com/pupitetris/Code/Freeciv-Worms/tileset.png
00:03.06gb2turn based combat gets very boring to me
00:06.45chouimatgb2: need a rpg/strategy/adventure type and I will be happy
00:06.59kergothyeah, thats what i'm looking for too atm
00:07.01gb2i loved metroid prime and zelda: windwaker
00:07.12gb2that's the type of game i like
00:07.31gb2realtime adventure/rpg game
00:07.56gb2i play LARPs at times too, i can't stand tabletop RPGs though.
00:09.50chouimati think I will install this tonight: http://sc2.sourceforge.net/index.html
00:10.13kergoth./libtool: s%^.*/%%: No such file or directory
00:10.13kergoth./libtool: -e: command not found
00:10.20gb2kergoth: cool
00:10.22kergothand libtool comes back with a vengeance
00:10.35kergothibot: spell vengeance
00:10.35'vengeance' may be spelled correctly
00:16.09W8TVIThe system is going down for reboot NOW!
00:16.12W8TVI:)
00:20.44trekegb2: Some RPGs I enjoy
00:20.57trekeNWN I did have a little bit of fun playing
00:21.26gb2the nutballs on another irc network are totally addicted to nwn
00:21.51gb2i think i like games where i actually feel in control and it isn't a purely linear game.
00:22.04gb2ofcourse.. I liked Eternal Darkness which is almost completely linear
00:22.37chouimathttp://ars.userfriendly.org/cartoons/?id=20030907&mode=classic
00:29.23*** join/#openembedded bipolar (~bflong@ctt186159.ceinetworks.com)
00:29.48kergothmake[1]: Leaving directory `/home/kergoth/coding/projects/user/oe/packages/tmp/base/apache-2.0.47-r0/httpd-2.0.47'
00:29.51kergothNOTE: package apache-2.0.47-r0: task do_compile succeeded
00:29.55kergothwoo hoo
00:33.52kergothkergoth@direwolf
00:33.52kergoth~/coding/projects/user/oe/packages$ ls -l tmp/deploy/tar/apache-2.0.47-r0.tar.gz
00:33.55kergoth-rw-r--r--    1 kergoth  kergoth   3260854 Sep  9 19:33 tmp/deploy/tar/apache-2.0.47-r0.tar.gz
00:33.58kergothheh!
00:33.58kergothneed to trim that sucker
00:40.23_Psycho3 megs for apache for OE device ? sound not bad
00:43.07*** part/#openembedded bipolar (~bflong@ctt186159.ceinetworks.com)
00:44.06*** join/#openembedded bipolar (~bflong@ctt186159.ceinetworks.com)
00:49.43ljpneeds tiny-apache
00:50.30W8TVIheh
00:53.40kergoth`bbl_Psycho: thats full, including man pages, docs, etc. completely untouched
00:53.45kergoth`bbl_Psycho: we'll shrink it down quite a bit
00:56.17_Psychoah indeed
01:03.38*** join/#openembedded LV|off (~lordvan@62.218.218.165)
01:27.43*** join/#openembedded bipolar (~bflong@ctt186159.ceinetworks.com)
01:32.14chouimatnight
01:56.12*** part/#openembedded slowhog (~henryjen@ip67-95-198-3.z198-95-67.customer.algx.net)
01:58.10chouimatkergoth: do you still have the url of the disassembler script?
01:58.21kergothhmmm
01:58.29kergothi dont think so, but i might have it on my desktop
02:01.33chouimatkergoth: ok I found it on my desktop backup ...
02:16.55gb2disassembler script?
02:24.43bipolarkergoth: working late tonight? ;)
02:25.08kergoth?
02:25.16kergothi'm not at work, no.. am i working? yeah
02:25.17*** join/#openembedded CIA (CIA@cia.utility.freenode)
02:25.18kergothdoing OE hacking
02:25.36kergothhrmph
02:25.42kergoththis metadata passing thing is pissing me off
02:25.46kergothit broke the ncurses build
02:27.10W8TVIcurses!
02:28.22*** join/#openembedded stigger (stigger@p024-n1-gos.nlc.net.au)
02:28.52muledo we support shared semaphores?
02:29.10mulewhatever that means
02:30.27*** join/#openembedded Bala (~Bala@hsdbsk142-165-117-29.sasknet.sk.ca)
02:30.27*** join/#openembedded LV|off (~lordvan@62.218.218.165) [NETSPLIT VICTIM]
02:30.27*** mode/#openembedded [+bbb *!*@12-227-17-69.client.attbi.com *!*@12-227-27-36.client.attbi.com] *!*@*12-227-27-36.client.attbi.com] by jordan.freenode.net
02:38.45kergothgrr
02:49.53*** join/#openembedded __stigger (stigger@p024-n1-gos.nlc.net.au)
02:53.09W8TVImewyn: what about?
02:53.31mewynhttp://news.yahoo.com/news?tmpl=story2&cid=501&u=/ap/20030909/ap_on_en_mu/downloading_music_11&printer=1
02:58.08kergothheh
03:00.19*** join/#openembedded _munky (~munky@ool-18ba3eed.dyn.optonline.net)
03:00.22_munkyhey
03:00.32_munkyanyone here?
03:01.16bipolarhmmm... new vendor is sending me huge .doc files. Must enlighten this one...
03:01.40_munkyhey do you know anything about OZ?
03:01.55mewynbipolar: with the Clue-by-four?
03:02.32_munkyif someone knows anything about getting the pocketop keyboard to work in OZ please let me know!!
03:03.28bipolarmewyn: I'm a bit more gentle then that. a nice email telling them that emailing .doc files is bad... mmmkay?
03:03.52bipolar_munky: I've never used that kbrd. Sorry, I don't even know where to start.
03:04.07_munkywell it corrupts my system
03:04.11bipolareek
03:04.25kergothno, IRK does.
03:04.29kergothinstall it from the commandline
03:04.32_munkywhen i try to install the package, it locks up and then when i press reset it just blinks the prompt
03:04.37kergothbipolar: i wouldve used the clue-by-four myself
03:04.48_munkyk
03:04.56_munkyill try kergoth
03:05.10bipolarkergoth: I usualy deal with informed people. I rarely get .doc files. Besides, it's a chance to mention openoffice. ;)
03:05.30mewynbipolar: awww... but so much bloodshed can be had with a clue-by-four
03:05.37bipolarhehehe...
03:05.49kergothheheh
03:05.52bipolarbut this is the person getting me a T1 for $650 a month. :)
03:06.04bipolarI don't want to kill her....
03:06.06bipolaryet...
03:09.30*** join/#openembedded LV|off (~lordvan@62.218.218.165)
03:13.43mewynkergoth: http://www.xmradio.com/million/refer/
03:14.27kergothheh, cool
03:14.50W8TVIits too bad xm is crapy around here
03:15.59*** join/#openembedded ChanServ (ChanServ@services.)
03:15.59*** mode/#openembedded [+o ChanServ] by zelazny.freenode.net
03:16.09kergoththis sucks
03:16.23*** join/#openembedded SuKoShi (~sukoshi@dyn-81-166-74-34.ppp.tiscali.fr)
03:16.24treke|homewho sucks?
03:16.37kergothheh
03:17.05kergothif i use make -e, it blows up teh build. if i pass on the cmdline, it fails due to parsing. if i pull it in from another makefile, its the wrong priority.
03:17.05treke|homeW8TVI: When you donwload an mp3 you dont have to listen to people make it clear why they can't find any other job
03:17.11kergothif i do nothing, the vars dont get into the build
03:17.23W8TVII know! lets start a radio music stealing ring!
03:17.26treke|homekergoth: Answer: You don't package that app
03:17.26mewynXM? crappy?
03:17.34W8TVItrees
03:17.51kergothif i use make -e and set MAKEFLAGS="", it works for the general case, but if you _need_ a var to override everything through all submakes, you cant since you set makeflags to ""
03:18.00kergothwhich is what i just ran into with ncurses
03:18.06W8TVIrm -rf /riaa
03:19.54mewynW8TVI: ah.  just burn them down.  that'll make the ents happy ;)
03:19.56W8TVIor even better:
03:20.15W8TVIecho fuck you > /dev/riaa
03:21.17mewyncat /dev/random > riaa
03:22.41W8TVIriaa > /dev/null
03:23.54ljpthe difference between radio play and downloading an mp3 is that the radio station pays either BMI, or ASCAP per play
03:24.13ljpso there are royalties
03:24.20W8TVIits no differnt
03:24.25ljpyes it is
03:24.28W8TVIsome one still buys the CD
03:24.32W8TVIthen shares it
03:25.13ljp'haring' is really distributing it
03:25.22ljp'sharing', even
03:25.32treke|homemusicians ought to go out and get real jobs
03:25.43kergothdamnit damnit damnit.
03:25.49W8TVIthe riaa should go out and get read jobs
03:25.50ljpmost musicians do have day jobsd
03:26.08W8TVII'd rather pay $1 a song to the artist
03:26.09ljponly a small percentage actually are millionaires
03:26.22treke|homeljp: they all suck
03:26.26W8TVIthan pay $20 for a cd that hos only 1 song on it I like
03:26.34ljpahh, well. what about the recording studio? and the record company
03:26.37kergothfuck!
03:26.38kergothgrr
03:26.50treke|homeljp: The artist is free to pay them
03:26.50W8TVIit costs $1 to make a cd
03:26.51ljpdont forget the graphic artists
03:27.08ljpever rented a recording studio?
03:27.10kergothi need to call make -e, with an empty MAKEFLAGS to prevent -e getting passed into submakes, and need to pass a var on the commandline as well
03:27.12W8TVI$1 per cd
03:27.15kergothhow the fuck am i going to do that.
03:27.15ljpits expensive
03:27.32treke|homeljp: Same as any other business.
03:27.35ljpit costs more than one dollar to produce a music album
03:28.30W8TVIbut the charge $20 to for each CD that only costs $1 per cd to make over the run of cds
03:28.56W8TVIthey sell 1mil cds to stores
03:29.05ljpthats not including the cost of the production of the music. thats just the industrial cost of pressing cd's
03:29.16W8TVIthey charge the store $17 for the cd the store sells it for $20
03:30.19W8TVIthe cheapest CD I have ever seen costs $5
03:30.28W8TVIhas 10 songs on it
03:30.56W8TVIthe store pays $2.50 to $3 for it...
03:31.02treke|homeThing is, those cds that were all 20 bucks last time I went to a store are all 13 bucks now. Meaning Someone figured out that that 7 bucks wasnt really necessary to make a nice profit
03:31.53ljpI'm not saying they are overpriced
03:32.00ljparent
03:32.09ljpthey are overpriced, sure
03:32.18W8TVIyou cant tell me that it costs more to make the $20 CD than it does for the $5 CD
03:32.37kergothargh!
03:33.27W8TVIthe cos that stores buy the music charge 2x what it costs to make the CD
03:33.44W8TVIsome times more than 2x
03:34.01W8TVII have worked in retail for over 4 years
03:34.20treke|homemmh. pizza and scotch
03:34.23W8TVII ahve seen what it costs stores for things
03:34.26treke|homeThis is gonna be a good night
03:34.29kergothGODDAMN YOU GMAKE
03:34.40W8TVIgmake--
03:34.47W8TVIkergoth++
03:35.50W8TVIlike I'd have to manualy reset the essid every time I put the card in to get it to work
03:36.06W8TVIkismet even works better with this new card
03:36.06*** join/#openembedded BleedingE (whocares@saus04.usc.es)
03:37.08*** join/#openembedded bipolar (~bflong@ctt186159.ceinetworks.com)
03:38.25ljpyou have no idea how much recording studios are
03:39.11ljpeasily $50,000
03:39.26kergothevery time i think i have a solution
03:39.28kergothit fails miserably
03:39.29kergothwtf
03:40.15*** join/#openembedded W8TVI|z (~zic@3639246105.mi.dial.hexcom.net)
03:40.16gb2still battling gcc?
03:40.21kergothno, make
03:40.24ljpdrink a rum, and forget about it
03:40.55kergothi need to set -e to pick up the vars from the env, MAKEFLAGS="" to prevent it from passing -e down into the submakes, but at the same time there are a few vars I _need_ to pass into the submakes
03:41.12kergothso i thought maybe i could set MAKEFLAGS to just $(MAKEOVERRIDES) so it gets the vars but not the flags, but it doesnt seem t owork
03:41.54kergothmake recursion sucks
03:42.00W8TVI|zdang this desktop....
03:44.30ljpbeep
03:45.13*** join/#openembedded treke|z (~root@lsanca2-ar29-4-41-064-167.lsanca2.elnk.dsl.genuity.net)
03:45.57*** join/#openembedded bipolar (~bflong@ctt186159.ceinetworks.com)
03:46.29W8TVI|zfreep
03:46.57*** join/#openembedded treke|z (~root@lsanca2-ar29-4-41-064-167.lsanca2.elnk.dsl.genuity.net)
03:47.34treke|zso what happens when rootwarn on ibot reaches max count>
03:53.44ljpall the power goes out
03:53.56ljppeople wont have water
03:54.03ljpso you'd better stock up
03:54.19kergothdamnit
03:54.25kergotheither we pass everything onthe cmdline or nothing
03:54.27kergoththat sucks.
03:54.36ljpall or nothing, eh?
03:54.58kergothyeah, i was hoping to use -e for most vars that can go in at toplevel but dont need to recurse all the way through
03:55.08kergothbut there doesnt seem to be a way to do it
03:55.37*** join/#openembedded W8TVI (~w8tvi@3639246105.mi.dial.hexcom.net)
03:56.06kergothdamnit damnit damnit.
03:56.16kergothwe'll hit the commandline length limit if we stick to that
03:56.17kergoth:(
03:56.22kergothhmm
03:56.36ljpso.. just increase that
03:56.58ljpcompressed commandline
03:57.09kergothhehe
03:57.10ljpgzip the commandline
03:57.27kergothnon-ascii cmdline
03:57.31kergothyou sick bastard
03:57.32ljpsomeow I dont think that woulod actually work
03:57.48ljpgzsh!
03:58.23*** join/#openembedded CIA (CIA@cia.utility.freenode)
03:59.38W8TVIhrm...
03:59.42W8TVI4 mirc users
03:59.53W8TVIthey should use xchat
04:00.43kergothgrr!
04:00.45kergothwhyt he hell
04:00.59kergothif i make -e CC=blah, it doesnt pick up that cmdline set of CC
04:01.02kergothit does if i remove -e
04:01.52kergoth~/coding/projects/user/oe/oe-dev/tmp$ THAT=env make -e THIS="that stuff" THAT=cmd 2>&1|grep 'THAT is'
04:01.55kergothMakefile:8: THAT is cmd
04:02.00kergothtest case shows it obeying cmdline above all
04:04.09*** join/#openembedded treke|z_ (~root@lsanca2-ar29-4-41-064-167.lsanca2.elnk.dsl.genuity.net)
04:06.48*** join/#openembedded stigger (stigger@p024-n1-gos.nlc.net.au)
04:08.58*** join/#openembedded CIA (CIA@cia.utility.freenode)
04:14.11*** join/#openembedded treke|laptop (~ggilbert@lsanca2-ar29-4-41-064-167.lsanca2.elnk.dsl.genuity.net)
04:14.22*** join/#openembedded TimRiker (~timr@proxyle02.ext.ti.com)
04:14.22*** mode/#openembedded [+bbb *!*@12-227-17-69.client.attbi.com *!*@*12-227-27-36.client.attbi.com *!*@12-227-27-36.client.attbi.com]] by tolkien.freenode.net
04:21.02treke|laptopkergoth: maybe we should subdivide content by package
04:21.14kergothtreke|laptop: ?
04:21.17treke|laptopkergoth: So we'd have content/apache/apache-2.0.47.oe
04:21.23kergothah
04:21.33kergothdoestn really matter, that wont affect build. just hte repos structure
04:21.38treke|laptopyup
04:21.50treke|laptopI can just see the directory structure getting messe
04:22.31treke|laptopno glibc-2.2.4? We going to switch to gcc 3 for the base system?
04:22.44kergothyep
04:22.48treke|laptopgood
04:25.01treke|laptopwe using the debian dosfstools patch for a specific reason?
04:25.17kergothprobably not
04:25.22kergothwe should ditch debian patches unless we need them
04:25.27kergothin other words, unless they touch code
04:25.31kergothwe dont need debian/
04:25.31treke|laptopyup. That's what I'm thinking
04:25.38kergothso we need to look at each debian patch
04:25.44kergothcheck if it touches anything outside debian/
04:27.21treke|laptopcool. Get to play "find the upstream"
04:27.56treke|laptopIt's a bad sign when something is last marked as maintained by an yggdrasil employee
04:28.11kergothhehe
04:31.36kergothshould write the source uploader class to automatically populate mirrors
04:31.56treke|laptopjust let me know, I've got space and bandwidth
04:33.23gb2hm.  wonder what i should buy for my birthday
04:34.14W8TVIone of my shirt-tail cousins died of cancer 2 days ago... I am sad for her husband that will no longer have her around, but I am happy for her that she no longer suffers.
04:34.42kergothbrb
04:34.46W8TVIhttp://links.net/
04:35.07W8TVImy own words... after reading one of the posts there
04:38.53treke|laptopdoes oemake not autosatisfy deps if you don't have it build everything?
04:39.08kergothit doesnt know how to _locate_ an oe
04:39.13treke|laptopah
04:39.17kergothyou must specify every .oe that it will know about
04:39.25kergothso it cant satisfy a dep if it doesnt know where to get the metadata
04:39.34kergothright now "what to build" is the same as "what can be built"
04:39.37kergothits all one list of oe files
04:39.38kergothnot split
04:39.46kergothcheck doc/TODO
04:39.51treke|laptopif I build something, it will know it's been built?
04:39.57kergothyeah
04:40.01kergothtmp/stamps/
04:40.06treke|laptopok. Wanted to be sure
04:40.46treke|laptop<PROTECTED>
04:40.49treke|laptop<PROTECTED>
04:40.53treke|laptop<PROTECTED>
04:40.55treke|laptop* kergoth wonders where that came from
04:40.58treke|laptop<kergoth> brb
04:41.00treke|laptop<W8TVI> http://links.net/
04:41.03treke|laptopωνω Signoff CIA: #openembedded (Read error: 110 (Connection timed out))
04:41.05treke|laptop<W8TVI> my own words... after reading one of the posts there
04:41.08treke|laptop<treke|laptop> does oemake not autosatisfy deps if you don't have it build
04:41.11treke|laptop<PROTECTED>
04:41.13treke|laptop<kergoth> it doesnt know how to _locate_ an oe
04:41.16treke|laptop<treke|laptop> ah
04:41.18treke|laptop<kergoth> you must specify every .oe that it will know about
04:41.21treke|laptopbase/patcher-1.0-r0 is both parent and child of patcher, aborting
04:41.23*** kick/#openembedded [treke|laptop!~kergoth@dsl081-228-056.chi1.dsl.speakeasy.net] by kergoth (control your mouse)
04:41.30*** join/#openembedded treke|laptop (~ggilbert@lsanca2-ar29-4-41-064-167.lsanca2.elnk.dsl.genuity.net)
04:41.37kergothboth parent and child? hmm
04:41.38treke|laptopThe trackpad on this laptop is fucked
04:41.41kergothoh, its recursing
04:41.47kergothcaught an infinate recursion
04:41.50kergothpatcher is depending on itself
04:41.50treke|laptopit doesnt always register left mouse down :)
04:41.52kergothhehe
04:42.07kergothtreke|laptop: .oe files automatically pull in base.oeclass
04:42.17kergothtreke|laptop: and base.oeclass automatically appends patcher to deepnds
04:42.18kergothboom
04:42.18kergothhehe
04:42.24kergothwe worked around it
04:42.27kergothlook at the patcher.oe
04:42.31kergoththere's a _delete
04:42.37treke|laptopbut it does nothing?
04:42.57kergothno
04:42.59kergothits incorrect
04:43.03kergothbase appends 'patcher' to depends
04:43.09kergothpatcher tries to remove base/patcher
04:43.13kergoththey arent the same, so boom
04:43.16treke|laptopoh. I need to delete patcher-1.0?
04:43.19kergothno
04:43.25kergothlast line in the patcher .oe
04:43.30kergothlook at it
04:43.37kergothDEPENDS_prepend_delete="base/patcher "
04:43.38treke|laptopdoh
04:43.39treke|laptopbase/
04:43.47kergothsecond line in bin/classes/base.oeclass
04:43.52kergothDEPENDS_prepend="patcher "
04:43.54kergothmismatch
04:44.05kergothchange the patcher _delete to just 'patcher' not 'base/patcher'
04:44.09kergoththen the problem will go away
04:44.12treke|laptopyup. DOne. It works
04:44.16kergothcool
04:47.17treke|laptopnext problem is oe never seems to go past patcher and onto glibc even though both are in OEFILES and patcher seems to have build successfully
04:47.43kergothglibc depends on kernel-headers
04:47.47kergothyou missed an error
04:47.52kergothunsatisfied dependency
04:48.06kergothan unsatisfied dep disables the build of a package and things that depend on it
04:48.14kergothbut doesnt abort the build, cause there may be things it _can_ build
04:48.35kergothsaw what build?
04:48.43treke|laptopso even if it's already been built, it has to be in OEFILES?
04:49.07treke|laptopkergoth: kernel-headers
04:49.08kergothyes, because the .oe can affect package name, version, and revision
04:49.10kergothwhich are in the stamp
04:49.11treke|laptopok.
04:49.18kergothso without loading the .oe, it cant know the stamp to look for
04:49.24kergothand anything can PROVIDES something
04:49.32treke|laptopthat does make some sense
04:49.34kergothbut it wont know what was built that provides it without loading the .oe
04:49.44kergothunless you were to create a mess of provides symlinks int he stamps are
04:49.47kergothhehe
04:50.05treke|laptopthis is gonna be a mess :)
04:50.19kergoththings will be cleaner once we teach oemake to distinguish between what we request built, and what oe files are available.
04:51.07treke|laptopyou know, my pizza should be here by now
04:51.17kergothheh, pizza guy get lost?
04:51.23kergothi've had that happen. fucker called me from his cell
04:51.28treke|laptopapparently. Ordered it an hour and a half ago
04:51.28kergothnot knowing where he was
04:51.48kergothheh, damn
04:54.07treke|laptopam I correct in concluding I need to set $PATH to include the toolchain?
04:54.10kergothyeah
04:54.14kergothit doesnt do that fancy shit yet
04:54.22kergothsince we have no config system, who cares
04:54.25kergothheh
04:54.51treke|laptopwho needs easy :)
04:55.01treke|laptopthere we go. Now glibc is happier
04:55.14kergothdefault is to install, but not package
04:55.30kergothINHERIT=package_tar oebuild blah blah will caues it to inherit the package_tar class, and make a tarball
04:55.34treke|laptopI'm guessing it can't make ipks yet, so that probably isnt a big deal :p
04:55.35kergothor package_ipk, or package_rpm
04:55.55kergothbut most .oe files dont have all the vars set needed to produce an ipk control file
04:56.01kergothwe need more default var values in oe.conf
04:56.08treke|laptoppackaging is usually the easy part anyways
04:56.18kergothyeah, making the shit build is the real work
04:56.36treke|laptopalthough packaging can be pretty tedious for things like gtk
04:57.01kergothpackaging is easier here. FILES_libc6="/lib/lib*.so*"
04:57.02kergothheh
04:57.06kergothwildcards++
05:00.29W8TVIhttp://linuxworld.com/story/34011.htm
05:01.04treke|laptoplinuxworld.com is the devil
05:01.27W8TVI"The giant sucking sound you're hearing, Darl, is the sound of your customers all contributing little pieces to the open source picture - so they can get better value and get rid of you."
05:01.45kergothhah
05:02.17W8TVILinuxWorld Magazine's Editor Responds to SCO CEO's Open Letter
05:03.23kergothljp: fix the cygwin buildroot build
05:03.25kergothljp: :)
05:04.38ljpI'm using that codewarrior zaurus thing
05:04.50ljpits kinda cool
05:05.46W8TVI"Taking the Geek Out of Linux"? what the world comming to?!
05:08.52W8TVIif they want people to use linux insted of windows, they need to get all the hardware co's to make linux drivers, and make a gui for stupid people
05:09.59ljpthere is a linux gui for stupid people - gnome. smart people use fluxbox
05:10.18kergothwhats kde, somewhere in between?
05:10.22kergoth:P
05:10.33W8TVII like kde
05:10.37treke|laptopno. kde is for germans
05:10.43kergothhaha
05:10.43kergothyeah
05:10.44W8TVIeasy to use but powerful
05:10.47treke|laptopand other eurofreaks
05:11.02W8TVIbut right now I'm using gnome for some reason that I dont rememeber
05:11.48ljpheheh actually, I tend to use kde
05:12.09ljpand switch to fluxbox when it pisses me off
05:12.27ljpI havent tried gnome for along time
05:12.39kergothi'm the other way around. use fluxbox and occasionally try kde to see how its doing
05:12.43ljpgnome would be good.. if it used Qt
05:13.13treke|laptopqt would be cool if it was lgpl
05:13.17ljpits better than SCO support for linux
05:13.41ljprevive harmony, and lgpl it
05:13.43W8TVIture
05:13.46W8TVItrue
05:21.52*** join/#openembedded CIA (CIA@cia.utility.freenode)
05:23.25treke|laptopsomehow /lib/libc.so.6 is sneaking into the build
05:24.14ljpoohh. afraid of a little committment, eh?
05:24.22treke|laptopnope. up to date
05:25.17kergothk bk pull
05:25.28kergothremove tmp/staging/target/lib/libc.so if it exists
05:25.31kergoththen try again
05:25.56*** join/#openembedded LV|off (~lordvan@62.218.218.165)
05:32.53*** join/#openembedded Bala (~Bala@hsdbsk142-165-117-29.sasknet.sk.ca)
05:33.52*** join/#openembedded CIA (CIA@cia.utility.freenode)
05:34.06treke|laptopcool libc6 compiled
05:35.12treke|laptopthis is the first thing I've gotten to build in oe
05:35.59ljpprogress
05:39.02*** join/#openembedded TimRiker (~timr@proxyle01.ext.ti.com)
05:41.09ljpTimRiker: hi
05:41.45*** join/#openembedded stigger (stigger@p080-n1-gos.nlc.net.au)
05:42.51Balaso does everyone on the OE team work for TI? or just Tim and Kergoth?
05:43.28kergothtim doesnt work on oe
05:43.40kergoththe only ti employee that does is myself
05:43.48Balaah well that answers my question then :)
05:43.54kergothheheh
05:44.03treke|laptopcool. dosfstool looks done
05:44.34kergothi love how with inheritence, the .oe files get more and more slim and clean
05:44.41kergothless duplicate implementations
05:45.26ljpthats inconceivable!
05:46.30ljpbleh
05:49.35kergothhehe
05:49.53treke|laptopTimRiker: You're the boss. You dont have to do any work
05:50.06kergothjust sit in meetings, right? ;)
05:50.10ljpya, you can drink beer like my boss
05:50.16treke|laptopand sign  timecards
05:50.39TimRikermeetings meetings.. that's about it.
05:51.09ljpdo you have meetings to discuss other meetings?
05:51.14TimRikergetting some time in on the side now. looking forward to getting a dev board to start hacking on.
05:51.24TimRikerljp: yep. =/
05:51.45ljpoh, thats funny
05:54.55treke|laptopcool. e2fsprogs builds
05:54.58treke|laptopthis shit is nice
05:55.04treke|laptopTimRiker: give chris a raise
05:59.54ljpya, me too
06:03.51fdaski did up a revamped downloads page on my OZ site mirror... can ppl check it out and lemme know if they like the interface for selecting files?
06:03.53fdaskhttp://ffff.ca/oz_website/content/download
06:04.28fdasknot done yet as far as pulling up the right files...  the box to select things at the top is the main bit i'm interested in feedback on
06:05.09*** join/#openembedded cyn_ (~cyn@adsl-18-208-15.mem.bellsouth.net)
06:05.39ljpi like it
06:06.11kergothi like it too. simple, clean.
06:06.59fdaskneed a bit of info on the directory structure for releases...   right now, i notice everything kinda jumbled...  like 5000d/5500 images in the same dir.   is this gonna get cleaned up?
06:07.10fdasklike maybe a REL_VER/MODEL/  type structure?
06:07.17kergothyeah, we need to restructure the site
06:07.24kergothi had a plan for that, but never got around to doing it
06:07.49fdaskcuz once thats in place, putting something like this (the rom selection thing) into place will be a snap.
06:07.53kergoths/site/site directory structure/
06:12.51kergothfdask: make sure you use valid xhtml, or noda will likely kill you
06:12.53kergothhehe
06:13.09kergothhe's a stickler for standards copmliance
06:13.14kergothbleh, cant type
06:13.21fdaskheh thx for the tip.  :)
06:13.49fdaskdoes noda come on freenode?  or any of the other website devs?
06:14.01kergothyeah he does, but not all that often nowadays
06:14.07kergothi atke it he's busy with contracts
06:14.08kergothibot: seen noda
06:14.08noda <~noda@modemcable013.119-201-24.mtl.mc.videotron.ca> was last seen on IRC in channel #openembedded, 1d 12h 33m 41s ago, saying: 'Huh? Why change /etc/network/interfaces?'.
06:15.58*** join/#openembedded CIA (CIA@cia.utility.freenode)
06:17.44kergothW8TVI: is cluebat the distribution menas?
06:17.46kergothmeans
06:18.15W8TVInope, clue-by-four
06:18.30kergothah, excellent.
06:18.42W8TVI<W8TVI> do I have to get the clue-by-four out?  if you keep asking if you can ask a question, I'm going to tell you no.
06:18.48*** join/#openembedded djr|z (pda@68.170.56.122)
06:18.48*** join/#openembedded _Psycho (~Bob@MTL-HSE-ppp165980.qc.sympatico.ca)
06:18.48*** join/#openembedded treke|home (~ggilbert@lsanca2-ar29-4-41-064-167.lsanca2.elnk.dsl.genuity.net)
06:18.48*** join/#openembedded absentia (~scott@ip216-179-64-58.cust.bestweb.net)
06:18.48*** join/#openembedded Shdwdrgn (Picard@sourpuss.net)
06:18.48*** join/#openembedded killefiz (~sven@ssv-niederkruechten.de) [NETSPLIT VICTIM]
06:18.48*** mode/#openembedded [+bbb *!*@12-227-17-69.client.attbi.com *!*@*12-227-27-36.client.attbi.com *!*@12-227-27-36.client.attbi.com]] by calvino.freenode.net
06:22.41*** join/#openembedded treke|z (~root@lsanca2-ar29-4-41-064-167.lsanca2.elnk.dsl.genuity.net)
06:26.49W8TVIkergoth: http://knerd.com/~kb8sjt/vlog.txt
06:27.53kergothhehe
06:28.15W8TVI:)
06:28.31W8TVIlike the way I handled it?
06:28.49treke|z<msg treke|home get laudndry 30 minutes
06:29.04kergothtreke|z: you missed.
06:29.59treke|zclose enough
06:30.31treke|zI just need a timestamp so I know when to get it :P
06:30.44kergothhehe. i should try that
06:30.55kergothi always forget it in the dryer and end up with wrinkley clothes
06:31.18treke|zi dont use dryers much :p
06:31.31treke|zmostly hang to dry
06:32.32treke|znot that I have much laundry to do
06:32.41ljpI just dont wash the clothes
06:32.51ljpthat way. they never get wrinkled
06:33.01*** join/#openembedded djr|z (pda@68.170.56.122)
06:33.01*** join/#openembedded _Psycho (~Bob@MTL-HSE-ppp165980.qc.sympatico.ca)
06:33.01*** join/#openembedded treke|home (~ggilbert@lsanca2-ar29-4-41-064-167.lsanca2.elnk.dsl.genuity.net)
06:33.01*** join/#openembedded absentia (~scott@ip216-179-64-58.cust.bestweb.net)
06:33.01*** join/#openembedded Shdwdrgn (Picard@sourpuss.net)
06:33.01*** join/#openembedded killefiz (~sven@ssv-niederkruechten.de) [NETSPLIT VICTIM]
06:33.01*** mode/#openembedded [+bbb *!*@12-227-17-69.client.attbi.com *!*@*12-227-27-36.client.attbi.com *!*@12-227-27-36.client.attbi.com]] by calvino.freenode.net
06:33.18treke|zshirts and pants all go to the cleaners, except in emergencies
06:33.44kergothcleaners.. i should do that. save time
06:33.56W8TVIcool
06:34.03W8TVIkismet runs on my desktop...
06:34.14W8TVIeven though I have no wireless card in it
06:34.16W8TVIlol
06:37.44*** join/#openembedded treke|z_ (~root@lsanca2-ar29-4-41-064-167.lsanca2.elnk.dsl.genuity.net)
06:47.13fdaskEtherkillers?
06:48.29W8TVIyep
06:48.55fdaskwhat are those
06:49.12W8TVIkergoth: his problem was with not having the dir that kismet wanted to write to a real folder
06:49.15W8TVIFATAL: Dump file error: Unable to open dump file Kismet-Sep-10-2003-1.dump
06:49.24kergothhehe
06:50.09treke|homehmmm
06:50.13treke|homemplayer was really hurting
06:51.23ljphmm.. how is dhcp handled in oz?
06:51.37treke|homewith udhcp
06:51.38W8TVIfdask: http://www.fiftythree.org/etherkiller/
06:52.13ljphmm
06:52.28ljpwhat package?
06:53.06ljpbuxybox?
06:53.12ljperr busybox
06:54.33*** join/#openembedded lsmith (~vandal@p213.54.51.34.tisdip.tiscali.de)
06:55.45ljphmm
06:56.01kergothyeah busybox includes udhcpc
06:56.45treke|homeoh sweet. sash is working on reverse engineering the w100 stuff
06:56.52kergothyeah
06:56.56kergoththats what i told you and mallum earlier
06:57.04kergothforgot it was sash doing it though
06:57.16kergothdoes he have that on a site somewhere or something?
06:57.24treke|homezaurus forums
06:57.42kergothbleh. i'll never understand the desire to use forums
06:58.06treke|homenntp is your friend
06:58.13kergothas is mutt.
06:58.28treke|homethere is a patch for nntp support in mutt
06:58.38kergothis there? coool
06:58.41treke|homeyeah
06:58.52treke|homehabit keeps me from ditching tin for it though
07:02.39ljpthats funny. theres a screenie of busybox on the web site
07:09.38W8TVInight guys
07:22.10*** join/#openembedded LordVan (~lordvan@62.218.218.165)
07:25.47*** join/#openembedded cmartin_ (~Chris@gw.krypnet.de)
07:27.29ljpbleh
07:41.18*** join/#openembedded stief (~stief@213-136-25-234.adsl.bit.nl)
07:41.19*** join/#openembedded lordvan_ (~lordvan@62.218.218.165)
07:58.54*** join/#openembedded LordVan (~lordvan@62.218.218.165)
08:17.43Twiungood morning
08:28.07*** join/#openembedded SuKoShi (~sukoshi@dyn-81-166-74-34.ppp.tiscali.fr)
08:30.07*** join/#openembedded lsmith (~vandal@port-212-202-39-108.reverse.qsc.de)
09:19.50*** join/#openembedded CAMCAM (~camrein@AMontsouris-108-1-4-70.w193-252.abo.wanadoo.fr)
09:26.26*** join/#openembedded LordVan (~lordvan@62.218.218.165)
09:37.24mickeylmorning Twiun
09:37.27mickeylhi rest
09:37.29mickeyl;)
09:42.58*** join/#openembedded LordVan (~lordvan@62.218.218.165)
11:24.38*** join/#openembedded djr|z (pda@68.170.56.122)
11:40.24*** join/#openembedded bisho (~bisho@136.Red-213-97-191.pooles.rima-tde.net)
12:00.01cmartinmickeyl: have any idea why console messages are still being displayed on the FB even though Opie already runs?
12:00.35cmartinmickeyl: i can't make why you don't see it
12:00.55mickeyldidn't read my botmail?
12:01.10mickeylcmartin: do /msg ibot botmail read
12:01.20cmartinmickeyl: no i wasn't informed of a botmail
12:01.28mickeylstrange
12:01.30mickeylanyway
12:01.32cmartinbotmail read
12:01.39mickeylno not here, /msg ibot
12:02.09mickeylcmartin: you must recompile the kernel for it to work
12:02.16mickeylcmartin: because i changed the kernel command line
12:02.20cmartinmickeyl: i had rebuilt my kernel, did you change something yesterday?
12:02.24cmartink
12:02.32cmartinmind reading
12:03.23mickeylno, i didn't
12:03.25mickeylhmm
12:03.56mickeylcmartin: do a 'cat /proc/cmdline
12:04.28cmartinmickeyl: what did you change in the cmd line - i think bk left mine in place since i had edited it before
12:04.52cmartinmickeyl: can i do a bk unedit on spec. files?
12:07.07mickeylwhat is a spec. file?
12:07.40mickeylah "specific"
12:07.44mickeylbk unedit works, yeah
12:08.49mickeylhah - no wonder then :)
12:08.49cmartinso a "bk unedit packages/kernel-simpad/Makefile" would remove my changes and allow to grab the version from bitkeeper?
12:08.55mickeylcmartin: yeah
12:09.01mickeylcmartin: as soon as you didn't have commited it
12:09.12mickeyls/as soon/as long
12:10.47cmartinbk pull says nothing to pull
12:11.26prhbah. netbase cant be found in ftp://ftp.debian.org/debian/dists/potato/main/source/base/ which is fine... it's not there...
12:12.33cmartin. /proc/cmdline says:
12:12.36cmartinmtdparts=sa1100:512k(boot),1m(kernel),-(root) console=tty1 root=1f02 noinitrd jffs2_orphaned_inodes=delete rootfstype=jffs2 M
12:12.47mickeylthat seems ok
12:12.53cmartini don't know what the M on the is doiing there
12:12.55mickeylnow what is the contents of /etc/init.d/banner ?
12:13.37cmartin# cat /etc/init.d/banner
12:13.39cmartin#!/bin/sh
12:13.40cmartin#switch to vt2
12:13.42cmartinchvt 2
12:13.43cmartin#disable non-critical output messages
12:13.45cmartinecho "2" >/proc/sys/kernel/printk
12:13.46cmartin#turn off cursor and show a logo
12:13.48cmartinecho -ne '\033[?25l' >/dev/vc/2
12:13.49cmartinzcat /etc/bootlogo.gz >/dev/fb/0 2>/dev/null
12:13.51cmartin#clear bottom if booting for the first time
12:13.52cmartinif test ! -e /etc/.configured; then
12:13.54cmartin<PROTECTED>
12:13.55cmartin<PROTECTED>
12:13.57cmartin<PROTECTED>
12:13.59cmartin<PROTECTED>
12:14.00cmartin<PROTECTED>
12:14.02cmartin<PROTECTED>
12:14.15mickeyllooks ok
12:14.21mickeylno idea why it doesn't work - it should
12:15.00mickeylread for your self... it should work, shouldn't it?
12:16.13cmartini noticed something that may have to do with the same thing:
12:17.24cmartinafter the bootlogo comes up and the reconfiguring of packages start - USB messages from the kernel garble the reconfiguring part and moves it down a row
12:18.25prhmickeyl: seems potato has been relegated in the debian archive... and netbase is fetched from a non-existant location now... how should I fetch it?
12:18.41mickeylprh: no idea.
12:18.47mickeylprh: look for a mirror
12:19.22prhhmm - why are we using such an old netbase anyway?
12:19.35mickeylcmartin: that's definitly wrong. chvt2 changes to vt2 and the echo says don't dump kernel messages to the active console...
12:19.41mickeylprh: because noone has upgraded it yet
12:20.06mickeylprh: you can't just update something without thinking 'bout the consequences, e.g. in this case
12:20.24mickeylnewer netbase change /etc/interfaces format which would break all userspaces tools relying on that
12:20.39mickeylprh: so generally spoken, you have to think a lot about the consequences before updating anything
12:21.07prhthe changes weren't that dramatic though were they?
12:21.42prhof course that could help explain why I can't get my wcf11 working under OZ or the sharp rom
12:21.51mickeylweren't ?
12:21.58mickeylwe didn't update the package
12:22.15prh"newer netbase change /etc/interfaces format"
12:22.27mickeyl"which would break"
12:22.30mickeylfuture tense
12:22.43prhI'm lost :O)
12:23.07mickeylnewer netbase packages are around, but we didn't update because of said reasons.
12:23.37prhok, gotcha. bah, I'm still hoping on that day that I manage to do an OZ build :O)
12:24.08fdaskprh: how long ya been at it?
12:24.26cmartinmickeyl: k, after reboot the cursor is no longer there, but kernel messages are still being displayed to the FB
12:24.28*** join/#openembedded mithro (~tim@dsl1-83.gw1.adl1.airnet.com.au)
12:24.46prhfdask: since early August.
12:25.11fdask:o
12:25.14prhfdask: I try few a few hours every few days but there's always things that fail I end up giving up
12:25.42fdaskim sure you've learned a lot in that time though, right?
12:25.50prhnot really
12:26.19fdasktook me about 6 days of fiddling to get my first image built.  flashed it this morning.  
12:26.21prhdoesn't help that I'm probably not meant to be using bk because I've been fiddling a little with arch/tla
12:26.35prhfdask: what machine is that for?
12:26.40fdask5500
12:27.05prhcool
12:27.14fdaskprh: what machine u trying to compile for?
12:27.25prhI'm currently working on making Debian work nicely from SD card on a 750
12:28.23*** join/#openembedded SuKoShi (~sukoshi@dyn-81-166-74-34.ppp.tiscali.fr)
12:28.30prhI've made an init script that will pivot to a debian install on SD and run the real init there (of failing that on ROM) so if there is a debian install on a plugged in SD card... debian boots nicely. just need to usb and pcmcia stuff working nicely
12:28.44prhthe init should work on any Z I guess
12:29.01Twiunprh: does SD on the 7x0 series work faster than on the old zaurii?
12:29.22prhno idea
12:29.37prhwant me to try bonnie on it?
12:30.04Twiunbonnie?
12:32.13mithrohello
12:32.14cmartinTwiun: why should SD work faster - SD is in itself slow (CF for example is much faster)
12:34.59mickeylcmartin: you might want to make sure, you're on vt2 with Opie (see /etc/init.d/opie and cat stuff to /dev/tty2) and that /proc/sys/kernel/printk really contains 2. might also experiment with echoing "1" or "0" to .../printk
12:40.12Twiuncmartin: I wasn't sure if they'd enabled the additional datalines
12:44.42cmartinmickeyl: here's the important part of /etc/init.d/opie:
12:44.45cmartinecho "Starting Opie..." >/dev/tty1
12:44.47cmartin<PROTECTED>
12:44.49cmartin<PROTECTED>
12:45.47cmartin# cat /proc/sys/kernel/printk
12:45.48cmartin7       4       1       7
12:48.37cmartinmickeyl: is that correct that "chvt 1" is executed after starting Opie? - and printk doesn't seem to contain "2" either
12:53.10*** join/#openembedded noda (~noda@modemcable013.119-201-24.mtl.mc.videotron.ca)
12:53.11nodaMoo!
12:56.17prhTwiun: disc speed tester
12:57.17prhcmartin: on machines with full sd support there are 4 data lines rather than 1 (which is the mmc standard) so SD cards can be potentially 4 times faster than mmc... if the hardware has the 4 data lines... the old Z's dont... there are rumours that the newer ones do
12:58.07nodaprh: The actual flash writing procedure is the bottleneck though, isn't it? So sending data over 4 lines wouldn't make it any faster....
12:58.44cmartinprh: k thanx for that clarification
12:59.47prhnp
13:00.07cmartinmickeyl: it seems the "echo "2" >/proc/sys/kernel/printk" in /etc/init.d/banner  isn't successful - after running it manually the messages are gone from the FB
13:00.11prhnoda: no the mmc line data rate is much slower than newer flash
13:00.25prhnoda: and data reads on flash are much faster than writes...
13:00.35nodaprh: I'm not talking about the line data rate. I'm talking about -- ah :)
13:01.00nodaOh, I am talking about it.
13:01.02nodaAh, I'm completely wrong :)
13:01.06nodaThanks for clearing that up :)
13:01.14prhnp :O)
13:07.57cmartinmickeyl: i think the problem is that /proc is mounted after banner is executed - look:
13:08.42cmartin# ls -al /etc/rcS.d/
13:08.44cmartindrwxr-xr-x    1 root     root            0 Jan  1  1970 .
13:08.46cmartindrwxr-xr-x    1 root     root            0 Jan  1  1970 ..
13:08.47cmartinlrwxrwxrwx    1 root     root           16 Jan  1  1970 S01banner -> ../init.d/banner
13:08.48cmartinlrwxrwxrwx    1 root     root           17 Jan  1  1970 S01devices -> ../init.d/devices
13:08.50cmartinlrwxrwxrwx    1 root     root           22 Jan  1  1970 S10checkroot.sh -> ../init.d/checkroot.sh;.... in checkroot /proc is mounted
13:10.07cmartinmickeyl: are you still there? - i feel like i'm talking to a wall
13:11.12Twiunhmmm
13:11.29Twiuncmartin: have you tried changing the order?
13:11.50cmartinTwiun: no but i will do that now
13:15.35*** join/#openembedded stief (~stief@213-136-25-234.adsl.bit.nl)
13:16.11mickeylmany tasks here
13:16.20mickeylcan't watch the windows all the time .)
13:16.22mickeyllet me see
13:17.01mickeylcmartin: your analysis seems good
13:18.05mickeylcmartin: fixing that might be easy
13:18.17cmartinTwiun, mickeyl: moving S01banner to S15banner didn't help
13:18.55mickeylhmmm
13:18.58mickeylit shoudl
13:21.21mickeylcmartin: try mounting /proc in banner
13:21.33*** join/#openembedded bobbyd (~none@217.204.66.43)
13:21.34bobbydhi
13:21.55bobbydI'd like to apologise for my ranting last night. it had been a long day and I was in a foul mood.
13:22.24chouimatmorning
13:22.45bobbydam I right in thinking that there is no ALSA driver for the zaurus in the buildroot? I grepped for alsa but just found files for the ipaq
13:23.06*** join/#openembedded DanWilliams (~DanWillia@proxyle02.ext.ti.com)
13:26.31Twiunmickeyl: seen the new HP Ipaq H2210/H2215 preview on TomsHardware?
13:26.47mickeylTwiun: not yet
13:27.39mickeylTwiun: have an URL handy?
13:27.51cmartinmickeyl: tried moving banner to S37 even and mounting /proc within it - no success - /proc/sys/kernel/printk still contains 7 and not 2 after reboot
13:27.52mickeylbobbyd: no ALSA driver for Zaurus
13:28.05bobbydmickeyl: ok great, I'll set about writing one then ;)
13:28.16mickeylbobbyd: now that would be damn cool!
13:28.18mickeyl:)
13:28.29Twiunmickeyl: for all the faults of the MS OS, I quite like the look/layout of the screen apps
13:28.29mickeylcmartin: odd
13:28.33Twiunmickeyl: sure, one sec
13:28.43bobbydmickeyl: shouldnt be too difficult. do you have a list of the zaurus' chips? including the sound chip?
13:29.00Twiunmickeyl: http://www.tomshardware.com/newsletter/vol3/36/ipaq.html
13:29.01mickeylbobbyd: well - its not that easy as you think
13:29.13mickeylbobbyd: iirc it is a sharp custom chip
13:29.20bobbydmickeyl: you mean finding the info on the chip?
13:29.21bobbydahh ok
13:29.33bobbydso the specs arent open?
13:29.47mickeylbobbyd: specs are available, but in japanese
13:29.50mickeylTwiun: thanks
13:29.51bobbydanyway, I'll have a look when I get home tonight, I'm sure I'll be able to sort it out
13:29.59mickeylbobbyd: good. for which machine, btw?
13:30.02bobbydmickeyl: that's fine, I have a japanese friend
13:30.19bobbydmickeyl: well, I have a c700, so for that for a start
13:31.17mickeylbobbyd: best seems to dive through the embedix patches and look how lineo did it. iirc the name of the chipset is written in the comments
13:31.29mickeylTwiun: dull OS, sexy hardware.
13:31.47mickeylTwiun: however I won't buy another PDA without WLAN integrated.
13:32.09Twiunmickeyl: dull perhaps, but the simplicity of the look seems more appropriate for a small screen.... PalmOS is still the pinnacle for me
13:32.41Twiunmickeyl: I find the Opie/Qtopia desktop-isms innapropriate
13:32.52Twiunibot: spell inappropriate
13:32.53'inappropriate' may be spelled correctly
13:34.29*** join/#openembedded qip|Virus|da (~Virusmast@p5088FDBA.dip.t-dialin.net)
13:34.37mithrospecially the taskbar - couldn't stand not know what was running on my pocketpc
13:35.07mickeylTwiun: you're right in that.
13:35.43mithroTwiun: which desktop-isms don't you like?
13:37.13Twiunmithro: the top buttons and the taskbar mainly
13:37.25mickeylTwiun: i like versatile controls (read widgets) though, IMO the problem is just the chunky default look of the Qt widgets. I think with a dedicated lightweight Opie widget theme, we would rule!
13:37.26Twiunmithro: I'm not too hot on the tabs on the root desktop either :)
13:38.37Twiunmickeyl: agreed, the versatility is important
13:40.06*** join/#openembedded SuKoShi (~sukoshi@dyn-81-166-74-34.ppp.tiscali.fr)
13:40.21Twiunsomething I find nice as an architectural feature is to define Actions for a program - not necessarily tie that action with particular widgets though
13:40.43Twiunso the close button's action could be operated through a different visual mechanism
13:41.14Twiunor removed completely and have the app close/hibernate after being in the background for a period of time
13:41.41Twiunhere we go for some PalmOS-isms :)
13:41.44mickeylTwiun: you have good ideas. Help us with Opie 2.x. - there we will decouple the launcher, taskbar and inputmethods.
13:42.42Twiunmickeyl: kergoth and I brainstormed a while back about the pluggable launcher - so my ideas do come through. I'll try to find time to help with opie 2.x
13:43.13chouimatmickeyl: opie 2.x will on qt3?
13:43.26bobbydwhen I put my zaurus in standby mode, what actually happens? what's draining the battery?
13:44.37bobbydI'm wondering if there's still work to be done on prolonging the battery life in standby mode as the battery doesnt seem to last very long
13:45.40*** join/#openembedded LV|work (~lordvan@62.218.218.165)
13:45.50yeiazeldoesn't seem to last very long ?
13:46.04mickeylchouimat: definitly. we will start developing it on top of qt3. most likely, bevor we release, TT will release qt4
13:46.49chouimatmickeyl: I heard that the snapshot will beging very soon
13:46.52yeiazelbobbyd: I've been able to use my Z without charging the battery 3 weeks
13:47.30yeiazel+"during"
13:47.50Twiunmickeyl: TT have given a lead time for Qt4?
13:48.50mickeylchouimat,Twiun: we're anticipating a solid beta release in Q1/2004. Qt4 will rock for embedded systems. Early in-house tests have reported a 33% drawing speedup
13:49.23mickeylby that time, OZ will have migrated to uclibc and g++3 giving even more speedup
13:50.03chouimatmickeyl: i was in n7y when tt give a talk about qt4
13:50.49*** join/#openembedded prpplague (~billybob@12.148.134.14)
13:51.43mickeylchouimat: n7y ?
13:51.55Twiunmickeyl: Nice! Namespaces support?
13:52.06chouimatmickeyl: nove hrady the kde conference
13:52.25mickeylchouimat: ah so you got it straight from the horses mouth :)
13:52.54chouimatmickeyl: from mattias ettrich, he was doing the speech
13:54.27mickeylcool
13:54.29mickeylhow is he?
13:54.35mickeylgeeky or "normal" ? ;)
13:55.19chouimatmickeyl: somewhat normal geek :) he spoke to me about two or three time that week.
13:56.04bobbydyeiazel: well using the OZ version for the c700 I only get about 4-6 hours in standby mode
13:56.41bobbydyeiazel: are you loggin out each time you shut dowm, or just using standby mode?
13:58.34yeiazeljust standby
13:58.39yeiazelwith a 5000d
13:59.11bobbydyeiazel: ok, there's probably something that hasnt been done yet for the 'collie' build, is the c700 the collie?
13:59.26yeiazelibot: collie
13:59.27i guess collie is sharp sl-5500 and sl-5000d, or a dog
13:59.42bobbydibot: corgie
13:59.43bobbyd: i haven't a clue
13:59.50bobbydibot: c700
13:59.51hmm... c700 is http://zaurus.kruss.com/files/zaurus2003/, or http://www.dynamism.com/zaurus/, or http://shop.conics.net/pda/zaurus-sl-c700/images
14:00.00bobbydhehe :)
14:00.06mickeylibot: corgi
14:00.08methinks corgi is sharp sl-c700, or a dog
14:00.15yeiazelPoodle (5600/b500), Corgi (c700), Shepherd (c750), and Husky (c760)
14:00.19yeiazelhttp://www.openzaurus.org/oz_website/content/news
14:00.21yeiazel:)
14:00.38*** join/#openembedded bipolar_ (~bflong@dhcp-103-630-1.msns.sm.ptd.net)
14:00.40yeiazeland, of course, Collie (5500/5000d)
14:00.50mickeylactually, a lot of stuff labelled corgi in kernel is really for corgi+shepherd+husky
14:01.01bobbydah ok
14:01.41bobbydalso it takes my corgi about 10-30 seconds to go into standby mode. I suppose all this stuff will be getting fixed. I'll build the buildroot and test that.
14:02.15mickeylbobbyd: already improved
14:02.19mickeylbobbyd: 5 seconds now
14:02.22bobbydmickeyl: excellent
14:02.35bipolarbobbyd: I've got the same prob with my 5500.
14:02.36bobbydgotta work, bye ;)
14:02.46mickeylbobbyd: test usb for us, please. we fixed that (hopefully), but noone tested yet.
14:03.05mulemickeyl: also fixed for poodle?
14:03.12mickeylbipolar: that's known. kergoth b0rke hotplug for collie
14:03.34bipolardoes oe build a system image compareable with buildroot yet?
14:03.34mickeylbipolar: currently, too get a working collie build, you have to downgrade 3 packages to OZ3.2-versions
14:05.09stiggerhi all whats the ope-quicklauncher in the wox feed
14:05.57mickeylopie-quicklauncher is a part of the Opie 1.0.2 fast-launch scheme.
14:06.01chouimathmmm this release is internaly know as Syphilis
14:06.10mickeyl*lol*
14:06.17mickeylmule: i think so
14:06.48chouimatmickeyl: I'm wondering if it's the disease  that the secretary didn't get yet ...
14:07.23stiggerso any idea why after I ipkg-upgrade I cannot start any programs from the launcher
14:10.52mickeylstigger: you must upgrade _all_ opie packages or it won't work. and you need the quicklauncher.
14:11.10*** join/#openembedded pb_ (~pb@dsl-62-3-66-201.zen.co.uk)
14:11.15mickeylstigger: i recommend flashing the complete image. upgrading probably doesn't work anyway
14:12.49stiggeryeah I'm downloading it now thought I maight get away with it
14:13.07mickeylstigger: please report problems here. I consider that snapshot to be
14:13.25mickeyla release candidate for OZ 3.2.6
14:13.41stiggerok
14:13.46mickeyl|bblthanx
14:14.04cmartinmickeyl: BTW: in my opie image with opie-quicklauncher it isn't version 1.0.2, but rather 1.0.1-1
14:14.15stiggerit's midnight here so don't expect anything soon
14:15.52mickeylcmartin: yes. we changed versioning for better upgrading, because we had problems upgrading from 1.0.2-snapshot to 1.0.2. so we just call it 1.0.1.x before incrementing.
14:16.08mickeylit will be called 1.0.2 when 1.0.2 is released.
14:16.08cmartink
14:18.14mickeylcmartin: btw., how do you like the fast app launching?
14:18.34stiggercool the icons bounce now
14:18.39mickeylhehe
14:18.55pb_hi chaps
14:18.57mickeylstigger: like that? devels are disputing about that.
14:19.01mickeylhi pb_
14:19.17mickeylpb_: how's your gpe-integration progressing?
14:19.43stiggerI don't think it looks as professional but it looks cool
14:19.52pb_mickeyl: I think libgpewidget is correctly integrated and should build out of the box.  I need to tweak the makefiles in the other applications to add the appropriate pkg-config stuff.
14:20.30mickeylpb_: sounds good.
14:20.32pb_I'll see if I can find some time later today to work on that.
14:21.02mickeylpb_: if you got that, it's time to do a bootmenu
14:21.42pb_yeah, that'd be good.
14:21.58pb_oh, and matchbox didn't compile last time I tried.  maybe someone could prod mallum about that.
14:23.33mulemickeyl: ok thanks
14:24.31cmartinmickeyl: k, added 'echo "2" >/proc/sys/kernel/printk" to /etc/init.d/opie and moved banner to S25 (to keep the USB messages from modutils from garbeling the banner)
14:25.27cmartinmickeyl: i don't know if it is good idea to put that in /etc/init.d/opie, but since we're using our own version of that for the Simpad it probably doesn't matter
14:25.52mickeylcmartin: yes for adding that to init.d/opie
14:26.02mickeylcmartin: as for moving banner to S25 i don't like that
14:26.11mickeylcmartin: seems very intrusive to the whole boot process
14:26.20mickeylcmartin: we need to discuss that
14:26.27mickeylwith the others
14:26.35cmartinmickeyl: well i just did it on my simpad, not in buildroot
14:27.10absentiare
14:33.12*** join/#openembedded icefox (~ben@xltpat1.symbol.com)
14:35.17*** join/#openembedded SuKoShi (~sukoshi@dyn-81-166-74-34.ppp.tiscali.fr)
14:37.16*** join/#openembedded stigger (stigger@p080-n1-gos.nlc.net.au)
14:59.50CIArepository=03buildroot user=11cmartin@nb005
14:59.52CIA07ChangeSet@1.1314 fixed kernel messages on opie fb on the simpad
15:17.15mickeyibot: seen andersee
15:17.15andersee <~andersee@codepoet.org> was last seen on IRC in channel #elinux, 6h 12m 37s ago, saying: 'fontenot: well, it compiled for me...'.
15:17.25mickeyhrmpf
15:31.38kergothyo
15:31.58absentiay0
15:32.58Twiunhey
15:36.28trekemornin
15:37.15kergothtreke: send me a pub key and i'll give you OE push access
15:39.12trekehas to be ssh v1?
15:39.35kergothyeah
15:39.38kergothstupid bkbits
15:40.23trekesent. Think I have dosfstools, e2fsprogs, and something else
15:44.37koyotehrm
15:44.48koyotefirst thing I get is a recursive dependency
15:48.07koyoteafk a bit
15:48.58kergothk, pushed the silly patcher recursive dep fix
15:50.10*** join/#openembedded TimRiker (~timr@proxyle02.ext.ti.com)
15:51.26kergothheh, i didnt notice before
15:51.33kergothbut commit -d works alot better in OE
15:51.38kergothsince the filenames are all different
15:51.43kergothdont need the path to know whats what
15:56.59treke~lart people who use SDL for games on the zaurus
15:57.05kergothhehe
15:59.08*** join/#openembedded Zippo_ (~thomas@p508C371A.dip.t-dialin.net)
15:59.17Twiuntoo right! use the console like everyone else!
16:00.56Twiunkergoth: what's the expected behaviour if running ./bin/oebuild without arguments?
16:02.02kergothTwiun: depends on context.
16:02.13kergothTwiun: if there's an oe file in the current dir, it builds that using the 'build' task
16:02.17kergothfor convenience
16:02.25kergothif there isnt, it dumps usage information
16:03.55Twiunhmmm
16:04.10TwiunI was running it from the oe root dir....
16:04.16Twiunit definately _wasn't_ happy
16:04.42TwiunNOTE: package oe-1.0_20030910-r0: task do_install started
16:04.42Twiuninstall: cannot stat `/usr/share/oe/bin/oemake': No such file or directory
16:04.43kergothoe root dir contains a .oe
16:04.59kergothhmm thats odd, your D var didnt get set. note the absolute path
16:05.58kergothdid the other tasks complete fine?
16:06.12Twiunyup, the ones before that did
16:06.18kergothstrange.
16:06.30kergothworks here
16:06.31kergothhmm
16:08.41Twiuncomplains about no SRC_URI, significant?
16:08.55kergothharmless, just means theres nothing for it to download
16:09.25kergothhmm, should patch libtoolize to accept an alternate path to the libtool share files
16:11.30Twiuncould it be related to the version of python? How can I tell which version I've got? (-h -v --version don't give me that info)
16:14.47kergothyou need very recent
16:14.52kergoth2.2+
16:16.57Twiunyeah, python-2.2.2-6mdk installed
16:18.10kergothhmmm
16:18.16kergothgo into oe.oe
16:18.19kergothdo_install function
16:18.22kergothecho D is ${D}
16:18.26kergothhey mickey
16:18.41muleis there any filespace in oz that is reccommended for use for multiple read writes?
16:18.50mickeyhi kergoth
16:20.10TwiunD is /home/twiun/oe/tmp/base/oe-1.0_20030910-r0/image
16:20.18kergothweird
16:20.25kergothyyp.. echo one of the install commands
16:20.30Twiunok
16:20.40Twiun... just stick echo in front?
16:21.13kergoththat function is shell :P
16:21.26kergothblah () { } is a shell function executed at task time
16:21.37kergothpython blah () { } is a python function " " " "
16:21.46Twiunahhhh
16:22.03chouimathttp://www.wired.com/wired/archive/11.10/fileshare.html
16:23.07Twiunchouimat: You're our userfriendly expert... what's the best cluebat clip?
16:25.12chouimatTwiun: don't know
16:25.17*** join/#openembedded sieve_work (~klu5ter@12.148.134.14)
16:26.06chouimatTwiun: I'm trying to eat my sardines without my cat knowing it :)
16:26.22kergothgood luck
16:26.47chouimatkergoth: excatly ... he had a whole can to eat so ...
16:27.02kergothweird, libtool just blew up again
16:27.12kergothit referenced a SED variable without defining it, thereby exploding horribly
16:27.15kergothodd.
16:27.15chouimatshit he finished his before mine ...
16:29.28mickeyhmm
16:29.35mickeyhaving problems with libxine
16:29.44mickeyusing 3.3.1-uclibc-toolchain i get the following:
16:29.47mickeyduring conftest
16:29.54mickeysr/local/arm/3.3.1-uclibc/lib/gcc-lib/arm-linux/3.3.1/../../../../arm-linux/bin/ld: conftest: hidden symbol `__udivsi3' in /usr/lo
16:29.54mickeycal/arm/3.3.1-uclibc/lib/gcc-lib/arm-linux/3.3.1/libgcc.a(_udivsi3.oS) is referenced by DSO
16:29.54mickeycollect2: ld returned 1 exit status
16:30.05mickeynow what is that?
16:30.14mickeyand why is it an error
16:30.30kergothheh, sounds like an andersee question
16:30.39mickeyright. where is he when I need him
16:30.39mickey;)
16:31.22Twiunkergoth: I'll get back to oe when I get home...
16:31.26kergothk
16:31.37kergothhmm, filename parsing bug
16:31.45kergothlibtool-1.4.3-r1.oe breaks badly
16:31.53kergothwhich sucks, since we want ot keep old versions of things around
16:32.31pb_mickey: inspect stuff like libc.so for relocs against __udivsi3.
16:34.00mickeypb_: hmm... that would be uClibC.so
16:34.18mickeypb_: thanks but i probably cry for help to andersee ;)
16:34.25pb_heh, right
16:42.02kergothanyone here really good with make?
16:44.00pb_(deathly silence ensues)
16:44.04kergothhehe
16:45.09*** join/#openembedded LV|work (~lordvan@62.218.218.165)
16:50.09*** join/#openembedded collord (~collord@dt.collord.net)
16:59.22mickeynah!
16:59.24mickey;)
16:59.31trekemickey: http://treke.net/downloads/PythonDocs.pdb <- complete python manual
17:00.24mickey.pdb?
17:00.47trekeplucker doc. Readable in opie-reader
17:00.49mickeyaah this is the thing you've converted in the last days ?
17:01.23trekeyup
17:01.26mickeycool. add a package to buildroot for that :)
17:01.50kergoths/add a package to buildroot/add an oe file/
17:01.56kergoth~lart mickey
17:02.14mickeyhow 'come i knew you would say that :)
17:02.24mickeybut you're right. we have to start the migration
17:02.58kergothtreke: how are you liking it so far?
17:03.08kergothstill has lots of needs, but ..
17:03.30trekekergoth: only have three things done
17:03.52gb2a "quickstart" doc for OE would be nice
17:04.02kergothyeah, we need more basic docs
17:04.08kergotha buildroot transition guide
17:04.42gb2can OE build an installable package of itself easily?
17:04.47kergothyes
17:04.49kergoththats what oe.oe is
17:04.56gb2it looks like it should be able to, but i couldn't figure out how to make it emit a package
17:04.57kergothi use it to build rpms for build machines
17:05.04kergothINHERIT=package_rpm oebuild oe.oe
17:05.11kergothits something the config system would normally decide
17:05.15kergothbut since we dont have one yet..
17:05.26kergothpackage_ipk and package_tar are also valid
17:07.25gb2ok, cool.
17:11.36gb2-rw-r--r--    1 gbritton users       46970 Sep 10 13:10 oe-1.0_20030910-r0.i686.rpm
17:11.37gb2yay
17:12.01gb2i needed to do: INHERIT=package_rpm OEDIR=$PWD bin/oebuild oe.oe
17:12.18kergothyeah
17:12.23kergoththats odd though
17:12.27kergothit shoudlve set OEDIR automatically if unset
17:13.21gb2i guessed it should have too, figured it tried to get $OEDIR/conf/oe.conf, if it fails, searches for oe.conf and sets OEDIR appropriately
17:13.41gb2(atleast.. guessed that from reading comments in oe.conf and the docs)
17:14.06trekekergoth: The build process has been dirt easy, although some things seem strange
17:14.48kergothgb2: yeah, thats correct. actually it should try to find the .conf anywhere in OEPATH
17:15.00kergothgb2: oedir and oedir/bin get stuffed in there by default
17:15.09kergothi'll kill oedir completely one of these days
17:15.21kergothso you can use OEPATH exclusively to control where it looks for things
17:15.57kergothits ${PN}-${PV}-${PR}
17:16.03gb2so the general way to do a project is, put a bunch of .oe files in a dir with a conf/local.conf and do: OEDIR=<that dir> oemake
17:16.03gb2?
17:16.04kergothFILESDIR is set automatically by a python function atm
17:16.06trekeah
17:16.10kergothgb2: yup
17:16.32kergothgb2: note that a local.conf isnt required, it defaults to ARCH=BUILD_ARCH, OS=BUILD_OS, BUILD_ARCH and BUILD_OS set by uname
17:16.39kergothoverride whatever you li ke
17:16.56gb2most of the stuff I do is cross, so i figured that stuff would go in there, yeah.
17:17.57kergothyep
17:34.17gb2if you want the pid, you want to fork() and do exec ...
17:36.51mickeyha!
17:37.15*** join/#openembedded SuKoShi (~sukoshi@dyn-81-166-74-34.ppp.tiscali.fr)
17:37.18mickeylinuxtoday.com: LinuxDevices: Linux 2.6: A Breakthrough for Embedded Systems
17:42.00gb2because of CONFIG_EMBEDDED ?
17:44.04mickeynah
17:44.08mickeybecause of the new subsystems
17:44.11gb2it'd be nice if uclibc built against 2.6.x kernel headers
17:44.13gb2it requires 2.4.x
17:44.14mickeyor rather hierachy
17:44.21gb2ah.
17:45.45gb2nice summary article
17:56.47mickeyibot: seen Harlekin
17:56.47harlekin <~max@pD9545F6E.dip.t-dialin.net> was last seen on IRC in channel #opie, 11d 6h 1m 47s ago, saying: 'back in 10 days'.
17:56.59mickeylol
18:04.12*** join/#openembedded pb_ (~pb@2002:5160:45ef:0:2e0:7dff:fe74:8b87)
18:13.47kergoth`foodugh, ate too much
18:13.52kergoth`foodneed a nap
18:13.55pb_hah
18:15.08*** join/#openembedded BBMatthew (~mas01r@arm.ecs.soton.ac.uk)
18:15.21BBMatthewello people
18:15.27mickey'lo
18:16.26kergothhey BBMatthew
18:16.37kergothso i found #devtools
18:16.45kergoththey didnt know the solution to my make metadata passing problem either
18:17.06trekepb_: How goes the gpe packaging?
18:17.13BBMatthewgot the SL5600 buildroot up and running (finally) but it bitches on the make image with libopenobex1 saying its missing some dependencies, namely opie-login opie-aqpkg and qpf-bitstream-vera
18:17.25kergothBBMatthew: remove one of the veras
18:17.26BBMatthewhey kergoth
18:17.29kergothdunno bout the others
18:17.43BBMatthewdo I need libopenobex?
18:17.51kergothits not libopenobex
18:17.59kergothipkg is misreporting the dependency
18:18.08kergothlibopenobex depends on something which depends on something etc etc which depends on those things
18:18.39BBMatthewany way of doing a cleanup?
18:18.41kergothchouimat: this must be some definition of "fun" i wasnt previously aware of
18:18.48BBMatthewshort of rebuilding everything :)
18:18.56kergothheh
18:18.57gb2kergoth: maybe he has a big distcc cluster
18:18.58kergothuse ccache
18:19.01kergoththen it doesnt matter much
18:19.06chouimatkergoth: using one of my customer network as extra compile box ;)
18:19.07kergothgb2: possible..
18:19.08BBMatthewI do use ccache
18:19.10kergothnice
18:19.15kergothBBMatthew: then who cares?
18:19.16mickeyBBMatthew: update your sources. opie-login and opie-aqpkg didn't package since yesterday's cvs.
18:19.29kergoththe only time you're wasting is extracting an dobtaining sources. the build is miniml
18:19.29trekemickey: What broke it in CVS?
18:19.43mickeytreke: dan forgot to adapt aqpkg
18:19.44BBMatthewfair enough - is there a shortcut to do that
18:19.46chouimatkergoth: since he don't want to pay more than 250cad/month for the maintenance .... I'm sucking some cpu time
18:19.53trekemickey: ah. Is BRANCH usable again?
18:19.55mickeytreke: i forgot that opie-login doesn't need to be quicklaunched
18:20.05mickeytreke: pretty much.
18:20.12trekefinally
18:20.14mickeytreke: conversion to the new launcher is nearly finished. minor issues
18:20.17trekebeen wanting to do a new image
18:20.26trekebut i didnt really want to debug the new launcher :)
18:20.27chouimatwow kdebase in 5 minutes!!!!
18:20.42mickeytreke: ha. sounds familiar to me... i did it anyway though :p
18:20.49BBMatthewchouimat: what are you building on?
18:20.58chouimatBBMatthew: pc
18:21.02*** join/#openembedded prpplague (~billybob@12.148.134.14)
18:21.04pb_treke: working on it now
18:21.15pb_harlekin still slacking off?
18:21.26trekestill on vacation as far as I know
18:21.41mickeyibot: seen Harlekin
18:21.42harlekin <~max@pD9545F6E.dip.t-dialin.net> was last seen on IRC in channel #opie, 11d 6h 26m 42s ago, saying: 'back in 10 days'.
18:22.59BBMatthewwhats the command to buildroot to update its sources? :) I have got the others documented
18:23.08trekemrproper
18:23.13mickeynaw
18:23.16BBMatthewoh yes should be able to submit a patch for xine lib soon
18:23.22mickeybest do it manually
18:23.38mickeyibot: ozscripts
18:23.39hmm... ozscripts is function ozclean() { rm -rf $stamps/.$1.*; make SUBDIRS=packages/$1 clean-sources; make SUBDIRS=packages/$1 clean-sourcetrees; }, or function ozmake() { make SUBDIRS=packages/$1; }, or function oznew() { ozclean $1 && ozmake $1; }
18:23.42BBMatthewwhap the stamps and the folder in source
18:25.31bipolaribot: seen zecke?
18:25.31zecke <ich@trns-d9b8655c.pool.mediaWays.net> was last seen on IRC in channel #opie, 5d 11m 51s ago, saying: 'bipolar: but pb_ seems to know the problem?!'.
18:25.45mickeyhah, that's a nice quote
18:25.47mickey:D
18:26.04bipolar:)
18:26.09pb_heh.  I wish he didn't sound so surprised.
18:26.13bipolarlol
18:26.47bipolarpb_: what are you working on, btw. IIRC you're a kernel hacker, or am I wrong?
18:27.12mickeypb_ is mr.multi-talent
18:27.46bipolarheh
18:28.39BBMatthewhow do I get buildroot to update the sources? sorry to be dense but the only options I know are nuclear ones
18:28.43bipolarWell... I'm just happy that I found out that the kde libs in 3.1.3 are compatible with kitchensync. So I can create a package for debian sid that doesn't try to squish other packages files.
18:28.45gb2mickey: made any progress on making uclibc build for you?
18:28.50mickeyBBMatthew: oznew <package>
18:28.52trekeBBMatthew: make SUBDIRS=packages/foo mproper
18:28.57trekeBBMatthew: make SUBDIRS=packages/foo
18:29.14trekepb is also to guy to blame when debian's glibc blows up
18:29.23bipolarcan oe build a complete flashable image or does it need more work?
18:29.33kergothhow many times do i have to say this
18:29.37mickeygb2: yeah - got a 3.3.1-uclibc toolchain to build. next problem is that this thing has a problem with libxine, but i hope andersee knows what to do...
18:29.39kergothIT DOESNT HAVE JACK FOR APPLICATIONS IN THE BUILD
18:29.45bipolarheh...
18:29.49kergothdo we bulid an image without 90% of what you need to boot?
18:29.51kergothfuck no, we dont.
18:29.53pb_bipolar: well, right at this minute I'm working on busybox.
18:29.55bipolarkergoth: I'll remember that. ;)
18:30.01kergothits hard to boot without 'init'.
18:30.04bipolarlol
18:30.07mickeyhard but not impossible
18:30.07bipolarthat it is
18:30.07mickey:D
18:30.09bipolarlol
18:30.10gb2mickey: standalone build of the toolchain? or did you build it within oe or buildroot?
18:30.16trekeinit is overrated
18:30.20kergothmickey: if you say init=/usr/bin/python i'll shoot you
18:30.23kergoth:)
18:30.25mickeyhahaha
18:30.26mickey:D
18:30.27bipolarpb_: hahahah
18:30.35mickeygb2: uclibc-toolchain
18:30.35trekeplease say it
18:30.40mickeyinit=/usr/env python
18:30.42bipolarer... that wasn't ment for you pb_ :p
18:30.47mickey./usr/bin/env even
18:31.05BBMatthewcheers
18:31.18pb_mickey: heh, that's one of nelson's fantasies as well.
18:31.34trekeoh, so nelson is a freak also?
18:31.35pb_he dreams of a system where the only compiled binary is /usr/bin/python.
18:31.42mickeyhehe
18:31.56mickeythat's not my dream...
18:32.02gb2i suppose that's better than init=/usr/bin/emacs
18:32.07mickeymy dream is a system where the only compiled interpreter is /usr/bin/python :D
18:32.46trekepb_: does he also dream of a system that runs like ass on a texas summer day?
18:33.08*** join/#openembedded imk (dj@68.170.56.122)
18:34.11BBMatthewbloody sparcstation!!!
18:35.29gb2mickey: uclibc's own toolchain builder?
18:35.46mickeygb2: yes
18:35.49gb2ok
18:36.10mickeygb2: worked like a charm... for the very first time, actually
18:36.33gb2that's almost always worked for me
18:36.55mickeywe probably have a complete uclibc based system by the time of OZ 3.4
18:37.09mickeyor OpenWhatver 3.4 ;)
18:37.15mickeyWhatever even
18:37.16bipolarhaha
18:37.29trekemaybe we'll get lucky and qpe wont work on it
18:37.31gb2OpenPileOfStuff
18:37.37kergothso, how are we going to handle the releases?
18:37.38mickeytreke: hehe
18:37.45kergothshoudl we manage them simultaniously?
18:37.49kergothor completely independent versioning
18:38.30mickeykergoth: although i'd like the idea of simultaneous versioning it is not practical. E.g. OpenSIMpad is really 0.8.1, not 3.3.x ;)
18:38.51kergothyeah
18:38.54trekekoyote: That kind of requires having things in a decent state
18:39.04kergothcourse the problem is, how do you define the feed that matches a given release
18:39.09kergothbleh
18:39.20trekekergoth: Tags
18:39.24koyoteI've seen some distros tie everything to one release number and have /alphaware trees under it
18:39.27kergothwonder if official/{branch}/{device}/ is sufficient
18:39.36kergothtreke: for the build, its easier than that
18:39.44kergothtreke: just put the OEFILES you used somewhere
18:39.54kergothsince we keep old versions around
18:40.00koyoteofficial/{stable/alphaware}/{device} ?
18:40.13kergothbranch is stable/testing/unstable
18:40.21koyoteah, cool
18:40.33kergothpossibly experimental as well, for really testing stuff
18:40.35mickeywell... we also need to distinguish gcc2.x and gcc3.x - and maybe soon gcc3.x-uclibc
18:40.38kergothi.e. new device work
18:40.38koyotethat seems reasonable. not all devices will be in stable.
18:40.51bipolarAre we going to have completely seperate feeds for each device? Could we instead have feeds for binary compatible devices and then have task packages to handle device specific info and setup? Is this my ass talking? :p
18:40.58kergothmickey: well, gcc2.x vs 3.x in qte/opie space. the base will be gcc3 regardless
18:41.03kergothbipolar: its your ass talking
18:41.07mickeykergoth: right
18:41.08koyotethough you may want to add pplaceholder directories. like in stable/shepherd have a text file that says use unstable.
18:41.10kergothwe dont fully exercise ipkg Architecture
18:41.16kergothif we did, we could have more commonality
18:41.17bipolarkergoth: THank you.. I'll take care of that right away...
18:41.22mickeyhehe
18:41.28trekeAnother question is, how many feeds to we need?
18:41.31mickeyno really, unifying feeds as much as possible is needed
18:41.47mickeyit's a waste of time, bandwidth and packagers @ the moment
18:41.50kergothotherwise we have 3 per device, not accounting gcc2/3 and uclibc/glibc
18:42.04kergoth12 feeds per device seems rather excessive.
18:42.13gb2what's the kernel issue with gcc3.x?
18:42.17kergoth?
18:42.19gb2is it actually gcc building things wrong?
18:42.21kergothsome kernels build with it, others dont.
18:42.27trekezarusu, ipaqsa, ipaqsa, and simpad
18:42.29gb2or is it the code needs some fixing?
18:42.37kergothgb2: you have to patch the kernel to get it to build with gcc3.
18:42.44trekezarusu, ipaqsa, zauruspxa, and simpad
18:42.45koyoteI'd really like to see a source feed for onboard building. it may be lidicrously time consuming for rebuilding opie, but for smaller apps, it gives people a good way to play with software.
18:42.45bipolarhmmm... what if we just had a feed for B/C and one that works the best? DOes ulibc work on all targets?
18:42.52gb2ok, so it's a kernel bug, not a gcc bug
18:42.52kergothkoyote: you dont need a feed.
18:43.02kergothkoyote: just install the OE ipk and build our oe metadata repos.
18:43.05koyote(of course, you have to include the source for opie, too. so that ludicrous people can compile)
18:43.11kergothour source tools are _independent_ of build box
18:43.22kergothunlike nearly every other distro build tool
18:43.29koyotehrm.
18:43.51*** join/#openembedded alephxx (~aleph@adsl-110-19.38-151.net24.it)
18:43.53koyotethere has to be a centralized place to get the sources, that's all
18:43.56koyote:)
18:44.04kergothno, theres a centralized place to get the metadata.
18:44.06mickeythis is not feasible
18:44.14kergoththe sources could come from any number of OE source mirrors
18:44.15mickeyyou'll never have such a centralized place
18:44.35mickeymetadata are enough + some heuristic logic for autobuilds
18:45.00*** part/#openembedded alephxx (~aleph@adsl-110-19.38-151.net24.it)
18:45.02kergothwe do need to cache sources on mirrors though, so that could be considered semicentral
18:45.13kergothbut its important to note that our metadata doesnt reside in the same location
18:45.38koyotehard to do a wget and archive a state that way.
18:45.44kergoth?
18:45.46kergothnot really.
18:46.56koyotei suppose one could download according to the metadata
18:47.03kergothuh, thats what OE and buildroot does.
18:47.07kergothheh
18:47.31kergothfetch -> unpack -> patch -> compile -> stage -> install -> package
18:47.53kergoth(with a liberal sprinkling of other tasks like tmake .pro creation, ipk creation, etc)
18:48.52koyoteI think it may be hung
18:52.38*** join/#openembedded slowhog (~henryjen@ip67-95-198-3.z198-95-67.customer.algx.net)
18:53.14kergothi was reminded when i had to run a ./configure for apache on the ipaq cluster
18:53.21kergothchrist did that take a long time, and that wasnt even compiling
18:53.22kergoth:)
18:53.23gb2heh.. native OZ builds
18:54.08*** join/#openembedded gruemel|laptop (~Jojo@pD9E50077.dip.t-dialin.net)
18:55.25bipolarthat was a long, long time ago
18:55.47Twiunrehi
18:56.11pb_kergoth: yeah, ./configure is one workload that arm handles particularly poorly.
18:57.27pb_in fact, any shell script is likely to be bad news.
18:58.11Twiunwhy's that?
18:58.29pb_lots of forks.
18:58.42pb_any context switch on strongarm will toast your cache.
19:00.37Twiunah
19:09.51koyoteokay, the Z is hung
19:17.52*** join/#openembedded andersee (~andersee@codepoet.org)
19:24.26*** part/#openembedded BBMatthew (~mas01r@arm.ecs.soton.ac.uk)
19:24.50*** join/#openembedded BBMatthew (~mas01r@arm.ecs.soton.ac.uk)
19:34.16treke~lart gcc 3.3.1
19:34.31trekethis std::time_put thing is getting annoying
19:39.34*** join/#openembedded stigger (stigger@p076-n1-gos.nlc.net.au)
19:39.58*** join/#openembedded W8TVI|z (~zic@3639246389.mi.dial.hexcom.net)
19:54.43*** join/#openembedded bach (~ciaolinux@MSTU83.CIMS.NYU.EDU)
19:58.18*** join/#openembedded Neo|Work (~neotron@proxyle02.ext.ti.com)
20:00.13*** join/#openembedded CAMCAM (~camrein@AMontsouris-108-1-4-70.w193-252.abo.wanadoo.fr)
20:01.03kergothbleh.
20:01.13gb2nice.. birthday cake at work
20:01.15gb2with ice cream
20:01.16gb2mmm
20:01.55koyoteokay
20:01.58koyotethe z won't flash
20:02.10fdask:o
20:02.13kergothkoyote: how wonderfully descriptive of you1
20:02.13fdaskno way!
20:02.21koyoteit just powers off when I select 'yes'
20:02.22kergoth~lart koyote
20:02.34koyotemmmmmmm, solaris
20:02.48kergothhehe
20:03.32koyoteyep, it goes through the slection process
20:03.38koyotethen powers off when I press enter
20:03.49kergothis your AC plugged in?
20:03.52kergothand stuff
20:03.56koyoteI think i've done something bad. or the last flash i tried ot install did
20:04.12koyoteyeah, AC. battery door and switch, all good
20:04.26kergothhmm
20:04.33koyoteI tries the rest option, too
20:04.36koyotesame thing
20:04.47koyoteand a normal power on doesn't work.
20:04.55kergothstrange, almost like the kernel it uses to do the flash got corrupted
20:04.57kergoth(in mtd1)
20:05.00kergothwhat device is this again?
20:05.06koyote750
20:05.18gb2koyote: do a DM flash restore?
20:05.21kergothdid you make a full nand backup at any point?
20:05.26kergothexactly
20:05.42koyoteactually, I hadn't found out that I was supposed to
20:05.54kergothhehe, you dont _have_ to
20:05.59kergothbut its nice in case you mtd1 gets hosed
20:06.09kergothmine was, but thats cause i was flashing manually and fucking with the flash process in updater.sh
20:09.02koyoteso, where do I get the appropriate files for the 750?
20:09.08gb2google?
20:09.15koyoteI found instructions here:
20:09.20*** join/#openembedded SuKoSh| (~sukoshi@dyn-195-242-111-101.ppp.tiscali.fr)
20:09.26koyotehttp://pocketworkstation.org/files/recover/README-flash-recover.txt
20:09.47koyoteI don't speak japanese well enough to navigate the relevant google results
20:10.28gb2not seen that procedure
20:10.51koyoteI'd love verification if anyone has :)
20:12.46gb2http://downloads.conics.net/pda/zaurus-sl-c700/service-menus/
20:13.25gb2http://downloads.conics.net/pda/zaurus-sl-c700/service-menus/original-backups/SL-C750-nand-backup-jp.zip
20:13.28gb2you want this
20:14.40cyn_koyote: use babelfish.altavista.com to translate japanese pages you can't understand
20:14.48gb2http://externe.net/zaurus/forum/viewtopic.php?t=388
20:14.52gb2follow these instructions
20:15.22gb2that c750 zip file should have SYSTC750.DBK in it, put it on a CF card and do the NAND Flash Restore from the DM menu
20:18.33*** join/#openembedded gruemel|laptop (~Jojo@pD9E50077.dip.t-dialin.net)
20:26.09kergothbleh.
20:26.10kergothbleh, i say.
20:26.40prpplaguekergoth: bleh bleh
20:27.42kergothprpplague: in your free time (ha ha) do me a favor and loook at the new OE build tools
20:28.25prpplaguekergoth: i'm on weekend duty this weekend, i'll have a look then
20:28.48kergothk
20:29.08kergothneed more eyes on it so we can clean up any remaining issues
20:33.13koyotelet's hope it works.
20:33.56koyotego little green bar.
20:38.19koyotenow to find the real stock kernel image
20:38.43*** join/#openembedded leffe (~leffe@as5-6-1.h.s.bonet.se)
20:43.22koyotelooks good.
20:47.37*** join/#openembedded lsmith (~vandal@p213.54.51.34.tisdip.tiscali.de)
20:47.59*** join/#openembedded _munky (~munky@ool-18ba3eed.dyn.optonline.net)
20:48.13*** part/#openembedded BBMatthew (~mas01r@arm.ecs.soton.ac.uk)
21:00.38kergothtreke: i'm going to split oemake's knowledge of what oe files exist vs what we want to build
21:00.55kergothtoday
21:01.25kergothso you could do something like OEFILES=$HOME/packages/content/*.oe, leave that laone, and just oemake glibc
21:01.29kergothor what have you
21:01.56kergoththatd improve the user interface big time
21:03.29gb2kergoth: so "oemake glibc" would build glibc, and everything glibc needs to build if it isn't already staged?
21:03.59kergothright
21:04.09kergothassuming all its deps are satisfied by something in OEFILES
21:04.29kergothotherwise that package and anything that depends on it will be removed from the build execution digraph and warn you
21:04.42kergothcool thing is, it makes it _alot_ easier for upstream users of OE
21:04.43kergothfor example
21:04.46gb2*nod*
21:04.47kergothsay you maintain tslib
21:04.54kergothyou create tslib.oe in your repos for making rpms and shit from it
21:05.00kergothoebuild<cr> to build it
21:05.05kergothoemake<cr> to build it and all deps
21:05.10kergothheh
21:06.12koyotewell, the initrd on zug sucks rocks
21:06.23kergothinitrd on zug?
21:06.40koyotezaurususersgroup.org initrd for the 750.
21:06.44kergothah
21:06.49mickeyl|Ole*cough*
21:06.52koyoteit's a variation on the stock image.
21:06.56kergothtell em to use buildroot
21:07.03kergothhey mickeyl
21:07.06mickeylhiho
21:07.11koyoteit doesn't even have proper character support.
21:14.33ljpAs The SCO Turns....
21:16.02mewynhey.  no cussing in here
21:22.34kergothk sysvinit's in OE
21:22.56kergothneed to figure out how to express devicetable like permissions changes when it creates a tarball in the package_tar oeclass
21:23.08killefizkergoth: how far from creating working images is oe?
21:23.20kergothkillefiz: a lot of packages still need to be converted over
21:23.31kergothI mean, i just converted init
21:23.34kergothheh
21:24.19killefizok - i'll keep on trying to build an ipaq image with the old buildroot then
21:24.21*** join/#openembedded cmartin (~Chris@p5087C067.dip0.t-ipconnect.de)
21:25.05kergothkillefiz: thatd be best
21:25.16kergothThere wont be another OZ release before the transition
21:25.23kergoththere will, however, be a set of new experimentals
21:25.56killefiza 2.4.2x based image with sd-support would rock
21:26.05killefiz(for the zaurus of course)
21:26.07chouimathmmm boring night in perspective
21:26.15kergoththats pending me having time to finish the driver
21:26.19kergothwhich isnt looking to be likely for a while yet
21:26.20bipolarkillefiz: I 2nd that....
21:26.28killefizkergoth: :(
21:26.30killefizbad news
21:26.41chouimatkergoth: any plan for a 2.6.0-test5-mm1 kernel for the collie?
21:26.49kergothTI has me working on OE atm, so thats my focus on the side too, i'm in buildsystem mode
21:26.52kergothchouimat: yes.
21:27.00kergothchouimat: i already have a 2.5.x tree somewhere that was semifunctional
21:27.02kergothjust never got around to finishing it
21:27.07kergoththat was before i did the 2.4.21 work
21:27.07killefizouhh .. that is better news then
21:27.18kergothso just pull that forward, and look at some device model bits, and we're set
21:27.23chouimatkergoth: ok ... back to kolab install mode
21:28.01killefiznoone who has ever tried one of the 2.4.18 sharp images is likely to try/use oz again before a new kernel
21:28.10kergothyep
21:28.26kergothi'm fine with less users in the short term
21:28.30kergothgiven the transition, thats actually best
21:28.52killefizso we would need someone who can work on the sd driver
21:29.09mewynhttp://newsobserver.com/news/story/2855036p-2634600c.html
21:29.29kergothwell more specifically, you need someone with the nda's sharp locomo doc and kernel experience, and tiem to give to the project
21:29.40kergoths/nda's/nda'd/
21:30.16killefizoh - i got the impression that there would be code that would allow development withouth being bound by an nda
21:30.26kergothno
21:30.29kergoththe core mmc code is open
21:30.30ljpnda scmnda
21:30.34chouimatkergoth: the problem is I not sure if I  have enough time for it
21:30.45kergoththe Z propietary chip involved we had to get the specs under nda
21:31.16chouimatkergoth: only need the chapter 6?
21:31.36kergothyeah, the chapter effy translated
21:31.53kergoththats sufficient, just need to use that to get the Z mmc slot driver interfacing with locomo's spi
21:32.15chouimatkergoth: will check this weekend if I have enough time in the next few weeks
21:32.16killefizno chance to use the 2.4.18 binary with 2.4.2x?
21:32.18ljpdamn it, a just got a sd card, I was transfering stuff to it, and the sharp 3.1 rom froze, and now fdisk wont even recogize it
21:32.24kergothkillefiz: nope
21:33.08chouimatkergoth: where is the code?
21:33.13killefizhmm - i guess i'll transfer all my pda-stuff to the ipaq then
21:33.33killefizi'm trying to compile an ipaq image for a couple of days now - my impression of the old buildroot is that it's not very stable ...
21:33.48mickeyl?
21:33.56mickeylit is pretty solid here
21:34.26killefizmickeyl: i had to change a lot of things here and there to make it compile
21:34.37mickeylkillefiz: e.g. ?
21:34.56mickeylkillefiz: upstream fixes don't count. that's inherent to a buildsystem like that.
21:35.10mickeylanything else left?
21:35.37chouimatkergoth: do you have the chapter in a more usable format than words?
21:35.45killefizi had to change xine-lib makefiles because it would include stuff from /usr/include instead of from the buildroot include
21:36.02killefiz(in fact i didn't want to compile it at all but opie depends on it)
21:36.03kergothchouimat: nope. could open it in openoffice and save it to something else
21:36.31chouimatkergoth: ok I will reinstall openoffice when II will have finished updating this box
21:36.51mickeylandersee: ping
21:37.04anderseemickeyl: pong
21:37.24mickeylandersee: 1st, congrats on the new uclibc release. vast improvements over .20
21:37.54chouimatmickeyl: faster?
21:37.56mickeylandersee: 2nd, 3.3.1-toolchain builds for me fine, but I get a strange (for me) error while configuring libxine
21:38.06mickeylandersee: config.log says like that:
21:38.18mickeylconfigure:12052: arm-linux-gcc -o conftest -I/usr/local/projects/opensimpad/output/staging/include -I/usr/local/projects/opensimpad/
21:38.18mickeyloutput/staging/usr/X11R6/include -I/usr/local/projects/opensimpad/output/staging/target/include -fexpensive-optimizations -fomit-fra
21:38.18mickeylme-pointer -O2 -march=armv4 -mtune=strongarm1100 -mapcs-32  -L/usr/local/projects/opensimpad/output/staging/target/lib -L/usr/local/
21:38.18mickeylprojects/opensimpad/output/staging/usr/X11R6/lib -Wl,-rpath-link,/usr/local/projects/opensimpad/output/staging/target/lib -Wl,-rpath
21:38.20mickeyl-link,/usr/local/projects/opensimpad/output/staging/usr/X11R6/lib conftest.c -lpthread   >&5
21:38.26mickeylcal/arm/3.3.1-uclibc/lib/gcc-lib/arm-linux/3.3.1/libgcc.a(_udivsi3.oS) is referenced by DSO
21:38.26mickeylcollect2: ld returned 1 exit status
21:38.41mickeylmmh - missing a few chars, but you get the idea in the last line (i hope)
21:38.54mickeylno
21:38.58mickeylreposting the last line
21:39.07mickeylusr/local/arm/3.3.1-uclibc/lib/gcc-lib/arm-linux/3.3.1/../../../../arm-linux/bin/ld: conftest: hidden symbol `__udivsi3' in /usr/lo
21:39.07mickeylcal/arm/3.3.1-uclibc/lib/gcc-lib/arm-linux/3.3.1/libgcc.a(_udivsi3.oS) is referenced by DSO
21:39.09mickeylthat's it
21:39.15mickeylminus the leading /
21:39.34anderseeThe _udivsi3.oS stuff is supposed to be supplied by libgcc_s.so.0.9.9
21:40.02anderseeAre you cross compiling?
21:40.14mickeylyes. that is a arm-linux-3.3.1-uclibc toolchain
21:40.34anderseeIf so, things like "-L/usr/local/" are going to result in very very bad things happening.
21:40.58kergoththats an incomplete line, the rest of the path is on th enext line
21:41.19mickeylandersee: that's -L/usr/local/projects/opensimpad/output/staging/include - which is our staging dir
21:41.58mickeyls/include/bin
21:42.13anderseemickeyl: ok
21:42.17mickeylfunny thing is, that worked with a previous version of the 3.3 toolchain
21:42.30mickeyland uclibc .20
21:42.43anderseemickeyl: you are using "arm-linux-gcc", so this was not built using my build, correct?
21:42.55mickeylandersee: it was built, but I renamed all binaries ;)
21:43.10mickeyli don't like arm-uclibc-gcc ;)
21:43.23anderseehmmk
21:43.26kergothwhy not change the prefix in the toolchain build, strikes me as less hackish
21:43.34anderseekergoth: agreed
21:43.58anderseemickeyl: just remove the "--program-prefix=$(ARCH)-uclibc-" stuff
21:44.15mule<PROTECTED>
21:44.17mickeylwill do.
21:44.19mulewerps
21:45.12anderseemickeyl: anyway, I have seen that sortof problem on and off.
21:45.29mickeylmickeyl: hmm. recommendations about workarounds?
21:45.32ljpanyone here use a Mac?
21:45.47anderseemickeyl: When I built the uClibc buildroot for arm yesterday, everything but gcc_target worked though.
21:46.19anderseemickeyl: The curious thing is why it is linking vs libgcc.a rather than libgcc_s.so.0.9.9
21:48.16mickeylandersee: no idea. the directory $TOOLCHAINROOT/lib/ contains libgcc_s.so.0.9.9
21:48.22anderseemickeyl: Is libgcc_s.so sitting in /usr/local/projects/opensimpad/output/staging/lib
21:48.24andersee?
21:48.31mickeyllemme check...
21:48.36anderseei.e. libgcc_s.so -> libgcc_s.so.0.9.9
21:48.50kergothshouldnt. staging is for interpackage dependency, not toolchain bits
21:49.07mickeylno - not there. just in my toolchain
21:50.40gb2kergoth: if the buildroot builds its own toolchain.. where should it go?
21:50.48gb2staging seems logical
21:50.54mickeylgb2: no.
21:50.57mickeylgb2: into the toolchain path
21:51.09gb2what's the toolchain path?
21:51.21mickeylgb2: make menuconfig --> build configuration
21:51.21kergothgb2: the path is hardcoded, so unless you want to always have the toolchain in your staging area
21:51.28kergothwhich you shouldnt, since staging is by definition temporary
21:51.32gb2erm.. i was meaning in general
21:51.42mickeyl./usr/local/arm/<version>
21:51.51gb2if i want to totally bootstrap it all.
21:51.59ljpsome might say it also belongs in /opt
21:52.05gb2the toolchain needs to be built and installed into the staging area
21:52.07kergothsome might say ljp is insane
21:52.12kergothgb2: no
21:52.15kergothgb2: it doesnt.
21:52.17mickeylsome might say SuSE is a good distribution
21:52.20ljpI was refering to embedix
21:52.32mickeylhah. let us online with embedix ;)
21:52.39mickeyls/online/alone
21:52.42kergothhehe
21:52.43mickeyllol
21:52.50ljpright now, I dont give a hill of beans where it goes, as long as it's in the path
21:53.00ljpand works
21:53.36gb2kergoth: it's a toolchain which is being built purely for the purpose of building stuff in that build system.
21:53.43ljpI am testing that codewarrior thing for zaurus
21:53.50kergothgb2: yes, and?
21:53.59ljpits really weird developing on windows
21:53.59kergothgb2: you're not listening. a toolchain's location cannot be changed.
21:54.13kergothgb2: now, if you plan on not keeping it around after this build, then sure, by all means use staging
21:54.25kergothmyself, i dont build a new toolchain every time i run a build
21:54.39gb2you're adding steps then.
21:54.45kergothhuh?
21:54.52kergothi build a toolchain once.
21:54.52mickeylandersee: hmm... i will rebuild the toolchain once again, just to be sure.
21:54.54kergothinto /usr/local/arm
21:54.59kergothbuildroot uatomatically appends that to the path
21:55.06kergothwhether i built it in this invokation or not
21:55.11kergothwhat extra steps?
21:55.24gb2installing the toolchain
21:55.31kergothbuildroot installs th etoolchain
21:55.35kergothdood, everything needs installing
21:55.40kergothinstalling it to staging is still installing
21:55.44kergothwhere's the difference?
21:55.47kergothoh right, there isnt one.
21:56.58gb2when i do a full bootstrap i generally want to bootstrap the entire thing, and the toolchain isn't going to be used for anything outside the buildroot.
21:57.13gb2it's only used by the build system
21:57.15kergoththen change the path to within buildroot
21:57.18kergoththats your perogative
21:57.20gb2just like say.. staged libraries
21:57.22kergothbut you're the first person to request that
21:57.35kergothyou're still missing it
21:57.38kergothstaging is _TEMPORARY_
21:57.45kergothif you clean buildroot, do you want to lose the toolchain?
21:57.48kergothi wouldnt want to.
21:57.49gb2yes
21:57.55kergoththats foolish.
21:57.58kergothmind bogglingly so
21:57.59gb2it's not like it takes years to build
21:58.07kergothyou're wasting time
21:58.10kergothnot years, but time
21:58.34gb2you could say that you're wasting time re-staging dependencies too.
21:58.56gb2the toolchain really isn't all that different
21:59.28kergothstaging is temporary because versions change
21:59.29kergothapis change
21:59.40kergothtoolchains shouldnt change often at all
21:59.41kergothpackages do.
21:59.46kergoththerein lies the difference
22:01.07kergothyour point is valid though, you could keep staging around, or rebuild everything. we do it this way because its most efficient for the general case
22:02.08gb2depends on your hardware
22:02.13gb2often you endup debugging the toolchain
22:02.22gb2so you somteimes want to totally rebuild the world
22:02.30kergoth"often"
22:02.32kergothnot for the common case
22:02.46kergothsometimes yes. most of the time no.
22:03.33*** join/#openembedded chouimat (~dieu@dsl-157-168.b2b2c.ca)
22:03.34*** mode/#openembedded [+o chouimat] by ChanServ
22:03.36kergothregardless, like i said, the system gives you the flexibility to what you want
22:03.47kergothwe just dont do it as a default, since its less efficient for the common case
22:05.16gb2i wasn't trying to criticize the OZ buildroot
22:05.23gb2i was trying to speak in general
22:06.20mewynhttp://www.infoworld.com/article/03/09/10/HNtorvaldssco_1.html
22:06.31gb2i've gotten attached to the uclibc buildroot
22:07.15ljpyup. buildroot's like a rubber band!
22:07.39ljpIK should start doing something worthwhile
22:07.46gb2which builds a uclibc toolchain then everything else
22:08.29kergothit does just what ours does when its enabled.
22:08.39kergothit puts it in a global area for deployment too, not staging
22:08.39kergothheh
22:08.52chouimatgrrrrrr
22:08.58gb2hrm.
22:09.02ljpwhats an easy app I can compile on windows
22:09.20ljpdont forget to convert that qtopia package, too :D
22:09.27kergothqte 2.3.7 is in
22:09.32kergothwouldnt take much to get qtopia in
22:09.33gb2just happens to be within the buildroot dir by default in uclibc then?
22:09.44ljphehe. I was kidding. I can do that
22:10.03ljpdoes oe build images?
22:10.33kergothi did a quick & dirty image making .ole
22:10.34kergoth.oe
22:10.38kergothbut i'd rather do a new oeimage tool
22:10.46trekekergoth: cool
22:11.05ljpI really need to get dsl at home
22:11.22chouimati should get my static ip back
22:11.32gb2oeimage tool that eats packages some scripted instructions and spits out an image?
22:11.37kergothyeah
22:11.42gb2*nod,nod*
22:11.45kergothjust hand it a configuration and a set of packages
22:12.37kergothi'm thinking i might keep the CONFIG_ style tokens for configuration values
22:12.44kergothbut use PKG_ for package selection
22:13.06ljpmakes sence
22:13.09ljpor sense
22:13.34ljpbut not scents
22:13.43kergothneed to read mconf.c and do somethign similar, but with a seperate piece that links into the OE metadata for package items
22:13.47kergothhmm
22:15.14kergoththatll be a pain in the ass without either 1) c++ oe core lib, or 2) oeread to stuff the metadata into a local sql data cache
22:15.25kergothor both
22:17.55CIArepository=03buildroot user=11mickey@allanon.vanille.de
22:17.56CIA07ChangeSet@1.1315 adapt defconfig-zaurus
22:19.15kergothmickeyl: adapt to what? :)
22:19.51mickeyladapt packages/opie/defconfig-zaurus to quicklauncher :)
22:20.05kergothah
22:20.22kergothCIA encourages good change commenting
22:20.51mickeylindeed... might have to improve my skills ;)
22:21.01kergothhehe. same
22:21.11kergothi still do commit -y"blah" once in a rare while
22:21.11kergothor ""
22:21.14kergoth"fixed"
22:22.24*** join/#openembedded W8TVI|z (~zic@3639246389.mi.dial.hexcom.net)
22:22.32ljpyou should learn German, and make German comments
22:24.08W8TVI|zlol
22:24.52mickeylhah
22:25.00mickeylhehe
22:25.08mickeyljust got a message from the SPND program commitee
22:25.13mickeylmy paper has been accepted
22:25.17kergothSPND?
22:25.31mickeylSNPD
22:25.35mickeyla conference about distributed systems
22:25.41kergothah
22:26.02ljpdont forget to mention Trolltech
22:26.06mickeylit's just laughable because the same paper has been clearly rejected twice for other conferences
22:26.13mickeyland now i got a full accept
22:26.15mickeyloh boy
22:26.30mickeyllike playing bingo
22:26.37ljphehehe
22:30.55kergothhmm
22:32.21*** join/#openembedded bobbyd (~rob@pc2-broo3-3-cust181.renf.cable.ntl.com)
22:32.36kergothhey bobbyd
22:32.43bobbydkergoth: hi
22:33.36bobbyddont know if you saw my rant apology in the logs, but I was in a foul mood yesterday. our external producer 'forgot' about 35 bugs until 2 days before we were supposed to go gold
22:33.37bobbyd:)
22:33.43pb_hi kergoth
22:33.46bobbydso I apologise for my ranting :)
22:34.12W8TVI|zrant on d00d
22:34.16W8TVI|zlol
22:34.37W8TVI|zI hate stupid PHBs too
22:35.22bobbydI cant believe people can be so incompetent!
22:36.00bobbydevery time I join a new company I think it'll change, but it never does :)
22:37.27kergothdidnt see, i usually dont have _time_ to check the logs
22:37.28kergothheh
22:37.29W8TVI|znope, never does
22:38.22W8TVI|zor how about when the boss doesnt like you and trumps up something so they can fire you?
22:38.25kergothbleh
22:38.32kergothW8TVI|z: ugh, been there before
22:38.40kergothdamnit, I cant do some of these conversions complete
22:38.42kergothcompletely
22:38.50kergothuntil we have an if/else/else if/fi block for OE files
22:39.19kergothoh well
22:39.26W8TVI|zkergoth, is fi the close of an if
22:39.45kergothW8TVI|z: we can call it whatever we want. fi, endif. which sounds nicer?
22:40.03W8TVI|zah
22:40.06W8TVI|zendif
22:40.11W8TVI|z:)
22:40.24kergothi'd really rather this was a yacc/lex parser
22:40.25W8TVI|zI'm just use to that
22:40.26kergothnot this python crap
22:40.35W8TVI|zheh
22:41.18kergothpick up k&r
22:41.24W8TVI|z(++,+ whaterver)
22:41.33W8TVI|zk&r?
22:41.43kergoththe authors of "The C programming language"
22:41.49W8TVI|zah
22:41.53kergoththe definitive C book, written by the authors of hte language
22:41.55kergothi lost mine
22:41.56kergothgrumble
22:42.32mickeylW8TVI|z: what do you want to learn C for?
22:42.35W8TVI|z:(
22:42.39pb_mickeyl: hah
22:42.43W8TVI|zlol
22:42.45mickeylno really
22:42.47mickeylhonest question
22:42.49mickeyltell me
22:42.50W8TVI|zso I can get a job
22:43.05mickeylhah, you do everything for money, right? :D
22:43.09W8TVI|zno
22:43.18cyn_money makes my world go round
22:43.26pb_I doubt C is actually that much of a requirement for getting a job nowadays.  It's kind of passe.
22:43.27kergothhrmph
22:43.29W8TVI|zwhat is OZ writen in?
22:43.32kergothheh
22:43.36kergothits a distribution
22:43.37mickeyla variety of languages
22:43.44W8TVI|zlol
22:44.04W8TVI|z(for the most part)
22:44.12bobbydcan I just check, does anyone know if anyone else is working on an alsa driver for the zaurus for open embedded?
22:44.21kergothbobbyd: afaik no one is.
22:44.24kergothbobbyd: and for which device?
22:44.26W8TVI|zbrb
22:44.28kergothbobbyd: zaurus isnt specific enough
22:44.34bobbydkergoth: c700
22:44.38*** join/#openembedded mule_ (~mule@195.137.103.49)
22:44.39mickeylW8TVI|z: the most is hard to tell. the kernel is in C and Assembler, the GUI is mostly in C++, the initscripts are bash
22:44.57kergothmickeyl: _bash_?
22:45.01kergoth~lart mickeyl
22:45.05mickeylkergoth: shell-whatever
22:45.06mickeyl:)
22:45.11kergothsh! /bin/sh!
22:45.35pb_yay bash
22:45.47kergothwonder how small you can make bash
22:45.52kergothreasonably
22:45.57cyn_that depends on how featureless you want it ohhh hehe
22:46.03kergothhehe
22:46.09kergothhmmm
22:46.23pb_mm, dunno.  it'd be interesting to approach it from the other side, see how small you could keep ash while having it support all the bash features.
22:46.24kergothOE can package into any number of output forms
22:46.33kergothwonder how to fit this into the postinst bits
22:46.48kergothi.e. for the tarball creation, we'd want rcX.d links created now, not in a postinst, since there is no postinst
22:47.10trekeone of these days I'm gonna start taking lunch breaks again
22:47.19kergothtreke: same here.
22:47.24pb_maybe have it run the postinst automatically when generating a tarball, so the tarball becomes kind of a frozen snapshot of an installed system.
22:47.24mickeylhehe
22:47.38kergothmy problem is i always get to work later than i'd like
22:47.38pb_treke: good thinking, else you're gonna start getting hungry
22:47.43kergothso i skip lunch so i can leave early
22:47.45kergothhehe
22:47.57trekekergoth: never seems to work like that
22:48.02kergothyeah
22:48.14trekekergoth: I always seem to skip lunch and consider leaving early. Then get stuck working two hours later
22:48.24kergothpb_: well, problem is how to run it when we're crosscompiling
22:48.39kergothpb_: cross our fingers and hope everything obeys the offline root path and doesnt try to execute binaries? :P
22:49.02kergothhehe. guess its just leave it as is for now. no extra bits for some output forms
22:50.15bobbydkergoth: someone (i think mickeyl) said that the docs for the c700 sound chip were almost nonexistant. and that the only real docs were in the origional sharp source. is that correct?
22:50.23kergothbobbyd: correct
22:50.28kergothbobbyd: I can get the service manual under NDA
22:50.35kergothbobbyd: but it'll be in japanese
22:50.39kergothbobbyd: in a scanned pdf no doubt
22:50.42bobbydhmm
22:50.47trekekergoth: you ever seen gcc 3.3.1 bitch about  undefined versioned symbol name std::time_put_w@@GLIBCPP_3.2
22:50.53trekewhen building opie?
22:51.06kergothsounds familiar
22:51.11kergothwhich appp failed?
22:51.18pb_kergoth: chroot?  executing binaries is obviously a problem, but most postinsts don't do that and you could just outlaw it by fiat.
22:51.18mickeylbipolar had that once - no idea what he did to cure it though
22:51.19bobbydI have a japanese friend, dont know how much technical stuff she'd be willing to translate though ;)
22:51.23treketetrix, opierec, opieplayer2, and something else
22:51.26kergothpb_: thats true.
22:51.28kergothhmm
22:51.37kergothpb_: well, at a minimum it'll run ln and the like
22:51.43kergothpb_: which would be problematic
22:52.13pb_kergoth: oh, yeah, duh.
22:52.26kergothtreke: not a bad idea. we can make their drivers in their images Suck Less if they find us a translator for hte odcs
22:52.29kergothheh
22:52.40kergothalthough..
22:52.40kergothhmm
22:53.16trekePlus in Dynamisms case they already have contacts that have done some technical translation work
22:53.31trekeconics may or may not do something similar
22:53.32kergothin other words, not within ipkg's db
22:53.43kergoththen any package just dumps the files there
22:53.44kergothhehe
22:53.58kergothand crosses its fingers that they'll get run on first boot by our reconfiguring step
22:54.09kergoththats actually not a bad solution, really
22:54.10*** join/#openembedded heatxsink (~John@63.146.78.10)
22:54.35pb_yep, I guess that'd work.
22:54.48kergothnot the cleanest, but ipkg would continue to work if we told it to look there..
22:54.51kergothhm
22:55.10kergothwonder who'll be the first user to build a rpm based OZ for their device
22:55.21kergothhehe
22:55.26pb_treke: what's the actual bitching you see about std::time_put_w@@GLIBCPP_3.2?  seems a bit suspicious that it has two @s in the name.
22:55.35kergothone of the devs at ti here used OE's rpm output form along with rpm's relocation support to install the package into their NFS root
22:55.39kergothquite spiffy :)
22:55.50pb_heh, neat.
22:55.56cyn_hehehe
22:55.56trekepb_: One sec
22:55.58pb_gotta hand it to those crazy texans.
22:56.03kergothhehe
22:56.17cyn_no native texans know how to code, they're all imported.
22:57.26trekehes from back east, I think
22:57.42pb_heh.  well, I guess the native texans must have known how to make chips once upon a time.
22:57.57bobbydmaybe potato chips
22:58.02trekepb_: http://ent.brooks.edu/~ggilbert/err.txt
22:58.42pb_hm, libopieplayer2.  does that do anything particularly freaky?
22:58.52trekeonly reference I'd seen to it involved linking to -lsupc++ which op2 does
22:59.10trekepb_: tetrix and opierec have the same problems. Neither do anything particularly freaky
22:59.19trekeop2 probably does though
23:00.26pb_I'm fairly sure that std::time_put_w is a libstdc++ symbol.  which is weird, because I thought opie shunned libstdc++.
23:00.38kergothit does, but a few apps use it
23:00.45pb_ah
23:01.15kergothusually doesnt break much on the Z, the toolchains have always used static, not shared libstdc++
23:01.20kergothshould think about changing that
23:01.26pb_oh, that might be the source of the problem.
23:01.35trekehmm. I could be using a bad toolchain
23:01.37kergothnot sure about the chain he's using atm
23:01.42kergothis that the 3.3.1?
23:01.48trekecross-3.3.1pre-oz1.tar.bz2
23:01.49pb_wouldn't surprise me at all if there are bad interactions between versioned symbols and static libs.
23:01.53trekefrom official
23:02.01kergothtreke: is libstdc++ a .so in it?
23:02.05*** join/#openembedded W8TVI (~w8tvi@3639246389.mi.dial.hexcom.net)
23:02.26trekeit's a .a
23:02.59kergothah
23:03.32*** join/#openembedded heatxsink (~John@63.146.78.10)
23:04.00bobbydok I'm off to bed. I intend to write and alsa driver for the c700 so I can have port my sound apps. I don know how succesfull I'll be, but I'm going to give it a shot ;)
23:04.06trekefun
23:04.13cyn_nite
23:04.23cyn_see, he came back
23:04.23kergothhe'll be cursing at lineo before long
23:04.28cyn_they always come back.
23:04.30W8TVIlol
23:04.39heatxsinkhello all
23:04.41trekekergoth: Should I try the 3.3 toolchain instead?
23:04.48kergothtreke: *shrugs*.. worth a shot
23:04.58W8TVIkergoth: should I share the kismet log from last night?
23:04.59trekehehe. Black magic
23:04.59mickeylit's your environment
23:05.01mickeylit works fine here
23:05.05trekemickeyl: Oh?
23:05.11kergothmickeyl: which chain?
23:05.23kergothW8TVI: heheh, why not
23:05.28mickeylcross-2.95.3 for kernel, 3.3.1pre for apps
23:05.34trekehmmm
23:05.41pb_mickeyl: do you have shared libstdc++, or static?
23:05.58mickeylno idea...
23:06.21pb_kergoth: ugh
23:06.28kergothwasnt fun :)
23:06.45heatxsinkanyone know how I can get identified so I can join the #gaim channel?
23:06.57mickeylpb: libstdc++.a
23:07.08heatxsinktreke: a little bit of M$ mentality never hurt anyone eh?
23:07.32trekeheatxsink: You never know. I might have installed one toolchain on top of the other
23:07.38kergothheatxsink: /msg nickserv help
23:07.39heatxsinkright
23:07.42pb_mickeyl: oh, right.  could you nm libstdc++.a | grep time_put_w?
23:07.44trekewhich could cause problems if something old poked through
23:07.50heatxsinkkergoth: I did that
23:08.07heatxsinkkergoth: and got this : The nickname [heatxsink] is not registered
23:08.13kergoth..
23:08.15kergothHELP
23:08.16kergothread the help
23:08.22kergothi didnt say identify
23:08.23kergothi said help
23:08.26kergothheh
23:08.35trekeah yes. oxygen also has the bug
23:08.48mickeylpb:_
23:08.48mickeylusr/local/arm/3.3/arm-linux/lib$ nm libstdc++.a | grep time_put_w
23:08.48mickeyl00000000 B _ZN9__gnu_cxx10time_put_wE
23:08.48mickeyl00000000 B _ZSt10time_put_w@@GLIBCPP_3.2
23:08.48mickeyl<PROTECTED>
23:09.20kergothc++ symbols are always so unpleasant
23:09.26trekeok. Mine is the same. Wish I handt already wiped it out
23:11.40W8TVIhttp://knerd.com/~kb8sjt/vlog.txt
23:11.40W8TVII think
23:11.40W8TVIhmmm.
23:11.42W8TVIcant connect
23:14.07kergothyou know, i dont like the oe naming
23:14.14kergoth${PN}-${PV}-${PR}.oe
23:14.19kergothi think that should be _ not -
23:14.19*** join/#openembedded W8TVI|z (~zic@3639246389.mi.dial.hexcom.net)
23:14.28kergothsince - is used in the version and name
23:14.32kergothmakes it ambiguous
23:14.39W8TVIwerid... stupid keyborad stoped working
23:14.43mickeylthen change it now - not if we have all 300 packages converted ;)
23:14.46kergothhehe
23:15.01W8TVIhttp://knerd.com/~kb8sjt/vlog.txt
23:15.08mickeyli remember your wild staging changes in the old buildroot.... damn... i was pissed at times ;)
23:15.15kergothhehe
23:15.25kergothwell, buidlroot wasnt designed
23:15.28kergothit was hacked
23:15.32kergothso kept running into new things it couldnt handle
23:15.47mickeylyeah
23:15.51kergothspeaking of which, we should adjust staging in oe
23:16.02mickeylha!
23:16.02W8TVIhey kergoth, ever have trouble with your sg20 ignoring all network trafic for about 2 o 3 mins at a time?
23:16.03mickeyl:)
23:16.06kergothhehe
23:16.09W8TVIthen all of the sudden work again?
23:16.20mickeylbed time.
23:16.21kergothnope, havent seen that one
23:16.22mickeylg'night all
23:16.26kergothnight mickeyl
23:16.37W8TVIfor a good laugh guys read this: http://knerd.com/~kb8sjt/vlog.txt
23:18.04W8TVIhmmm
23:18.21W8TVImaybe I'll go and put the drive in the laptop and install RH9 on it...
23:23.19_munkyhey
23:23.28_munkycan someone here help me running X?
23:23.50treke#openzaurus is the place for support questions
23:24.21kergothi dont think he's using oz
23:24.24kergothheh
23:24.24_munkydoes that mean proceed?
23:24.30_munkyso what
23:24.42kergothlook at what channel you're in.
23:24.44_munkyi installed x and icewm
23:24.48kergothsharp/tkc questions belong in #zaurus
23:24.52_munkyi know..come on man just try to give me a hand
23:24.57_munkybut no one answers anything
23:25.02_munkyim running thekompany btw
23:25.14_munkybut honestly OZ is better then all those that i tried
23:25.23_munkyso can u try to help me?
23:25.31kergothi understand your predicament, but that doesnt make it any more on topic.
23:25.42_munkyid really appreciate it
23:25.53_munkyjust two questions
23:26.00kergothno.
23:26.03kergothask them in #zaurus
23:26.10trekemost of us don't use tkc's rom
23:26.12_munkywill you answer them  there?
23:26.20_munkyim running a tkc rom?
23:26.26_munkyim running thekompany
23:26.27kergothtkc is thekompany _munky
23:26.32_munkyk
23:26.36kergothTheKompany.Com
23:27.57kergoth~/coding/projects/user/oe/packages$ ls -l tmp/deploy/tar/|wc -l
23:27.58kergoth<PROTECTED>
23:27.59kergothgetting there.
23:29.02gb2kergoth: can a .oe emit multiple packages?
23:29.18kergothINHERIT="package_tar package_rpm" oemake blahblah
23:29.21kergothits space seperated
23:29.26kergothyou can specify any oeclass in that var
23:29.28kergothto inherit
23:29.43gb2not what i meant
23:29.52kergoth?
23:29.57kergothoh
23:29.58kergothyeah
23:29.59gb2stuff like say.. gcc emitting a libstdc++ package and a gcc package
23:30.10kergothPACKAGES=libc6 libc-dev
23:30.12kergothFILES=
23:30.18kergothFILES_libc6=/lib/*
23:30.26kergothFILES_libc-dev=blahblah
23:30.32gb2ah, they get the -version-revision tacked on to all of them
23:30.33kergothit'll iterate through teh packages
23:30.40kergothyeah
23:30.43kergothor yhou could say
23:30.47kergothPV_libc6=blah
23:30.47gb2is there a way to override the version?
23:30.50gb2ok
23:30.51kergothPV_libc-dev=heh
23:30.51gb2heh
23:30.57kergoththeres an OVERRIDES var
23:30.58kergothread oe.conf
23:31.01trekedamn. kergoth thought of all this shit
23:31.02kergothabout OVERRIDES behavior
23:31.04gb2ok, cool.
23:31.08kergothnah, overrides was schurig's baby
23:31.21kergothpersonally I'd rather have better conditionals
23:31.43kergoth${ARCH} == arm: THIS=that
23:31.45kergothor something
23:31.52kergothto enhance flexibility
23:31.58kergothor just if blocks
23:32.21trekeif {} else if {} else {} would probably be enough
23:32.39kergothyeah
23:32.47kergothoverrides wasnt a bad idea
23:32.48kergothbut..
23:32.55kergothfor example, overrides gets ARCH
23:32.58kergothit doesnt get BUILD_ARCH
23:33.01gb2yeah, conditionals would be very useful
23:33.05kergothso you cant make a decision based on your build machine arch
23:33.17pb_would you want to?
23:33.41gb2sounds like a recipie for problems
23:33.45kergothwell, we need conditionals
23:33.48kergothhere's an example
23:33.55kergothCROSS=${ARCH}-${OS}-
23:34.06kergothbut only if ARCH != BUILD_ARCH or OS != BUILD_OS
23:34.13gb2ah
23:34.15gb2good idea
23:34.26kergothso we need a way to empty CROSS given those conditions
23:34.30kergothwhich we cant do right now
23:34.34kergothso like if i want to do a native build
23:34.40kergothi have to echo CROSS= >> conf/local.conf
23:34.49kergothwhich is tedious at best
23:35.21gb2conditionals based on the build machine can be easily misused though
23:35.30kergothyep, they can
23:35.32pb_fwiw, I think most modern distributions install the toolchain so that i386-linux-gcc is an alias for gcc.
23:35.33kergothbut so can lots of things
23:35.36trekewe need stuff like that for opie though
23:35.37pb_(for example)
23:35.39kergothpb_: yeah, quite true
23:36.01kergothwe also need a way for a .oe to specify that it isnt buildable given a set of conditions
23:36.03gb2redhat has i386-redhat-linux-gcc
23:36.12gb2not i386-linux-gcc
23:36.17kergothebuilds support masking what architectures it supports
23:36.24kergothbut that seems needlessly inflexible
23:36.28kergothso i'm thinking a new task
23:36.36kergoththat returns a value to determine if its buildable
23:36.41kergothdo_canbuild ()
23:36.42kergothor something
23:37.02*** join/#openembedded chouimat (~dieu@dsl-157-168.b2b2c.ca)
23:37.02*** mode/#openembedded [+o chouimat] by ChanServ
23:37.02pb_gb2: yeah, same applies.  you always need to be able to tolerate a vendor vanity string in the middle of the triplet.
23:37.27pb_also -linux vs -linux-gnu variance, to some extent.
23:37.42kergoththing is, we dont have tests. without tests, we rely on the user.  hence the need to override CROSS in certain cases
23:37.47gb2pb: well.. it's technically arch-vendor-os
23:37.56gb2i386-pc-linux isn't uncommon
23:37.59pb_indeed
23:38.00gb2yeah
23:38.04kergothneed ot attempt to keep the complexity and intelligence down
23:38.08kergothof the system proper
23:38.14pb_kergoth: right
23:38.19kergothunless we feel like attempting to replace autotools or gmake
23:38.23kergothwhich is a bit beyond the scope of hte project
23:38.26kergothhehe
23:38.29pb_heh
23:38.35pb_that'd be commendable though
23:38.35gb2very little reliance on the user is a good thing
23:38.43gb2i've been dealing with a bunch of that at work
23:38.47kergothgb2: well, it has sane defaults in most cases
23:38.53gb2trying to make things as idiotproof as possible
23:39.02*** join/#openembedded w8tvi_ (~w8tvi@3639239351.mi.dial.hexcom.net)
23:39.05kergothjust a matter of forcing a user override when given a strange set of conditions
23:39.12kergothsince it lacks the intelligence to divine those conditions
23:39.55*** join/#openembedded noda (~noda@modemcable013.119-201-24.mtl.mc.videotron.ca)
23:39.58nodaMoo!
23:39.59kergothhey noda
23:40.02nodakergoth: Heya :)
23:40.43nodaMy GOD! "Unread: 11; Total: 14" -- you can tell Sobig has expired :)
23:40.44kergothpb_: i've considered it.. the A-A-P project is _alot_ like us with regard to capability and syntax, but acts as a gmake replacement (not autotools)
23:40.49w8tvi_m00
23:41.02nodakergoth: Pet peeve: s/alot/a lot/g
23:41.10w8tvi_I never receved 1 sobig
23:41.13kergothso we certainly _could_ do that, juust a question of whether we want to
23:41.14nodaalot is a verb :)
23:41.31nodaw8tvi_: On Monday I got 550 new messages. On *one* of my email accounts :)
23:41.39kergothnoda: sent out a request for assistance on our download page, fyi. got a few replies, and fdask did a preliminary version
23:42.24nodafdask, eh? That's James W?
23:42.55nodalol, I've been mailing with somebody helping out. http://ffff.ca?
23:43.08nodaibot: seen fdask
23:43.08fdask is currently on #opie #zaurus #openembedded #openzaurus.  Has said a total of 337 messages.  Is idling for 2h 7m 37s
23:43.10chouimathi noodaa
23:43.14nodachouimat: Hola!
23:43.20kergothnoda: hehe, thatd be it
23:43.26nodaHe's nice, I like him :)
23:43.51kergothi warned him to output valid standards compliant xhtml or he'd suffer your wrath
23:43.55kergoth:P
23:44.04chouimathttp://c2.com/cgi/wiki?DiscussionOfCargoCult
23:44.05nodakergoth: lol, he's been mailing me about that ;)
23:44.08kergothhehe
23:44.52W8TVIanyone read my log?
23:46.05W8TVIhttp://knerd.com/~kb8sjt/vlog.txt
23:47.19W8TVIlool
23:47.29*** join/#openembedded ljp (~lpotter@ppp189.adsl238.pacific.net.au)
23:47.49W8TVIlooking at the bounced messages for my one yahoo group:
23:47.50W8TVIRemote host said: 554 <probe-1074048-1079279-bg1dz=sohu.com@returns.groups.yahoo.com>:
23:47.50W8TVISender address rejected: Access denied
23:51.47*** join/#openembedded LV|sleep (~lordvan@62.218.218.165)
23:52.01kergothwhat package does debian store its /usr/include/{linux,asm*} in?
23:52.26pb_libc6-dev
23:52.36kergothk
23:52.58kergotherr, damnit.  that means our libc6-dev needs files from two .oe's
23:53.03kergothhow the hell am i going to work that
23:53.05kergothhehe
23:53.08kergothhm
23:53.53pb_heh.  what are the two .oes in question?
23:54.17kergoththere's a seperate kernel-headers .oe
23:54.24kergothsuppose i could just let glibc pull them down itself
23:54.27kergothbe easier that way
23:55.58pb_ah, right.  the way debian works at the moment is that kernel-headers installs into /usr/src/kernel-headers-N.NN, and glibc repackages those files into libc6-dev during its own build process.
23:56.24kergothah. how wonderfully bound to distribution
23:56.36kergothguess i could do something similar
23:56.48pb_we've just been talking about abandoning the idea of using kernel-headers from the kernel altogether, and just keeping a local copy in the glibc source package.
23:57.06kergoththat strikes me as more sane
23:57.29pb_yeah.  I was dubious at first, but I think I've now been won round.
23:57.29kergothand less painful in packaging
23:57.40kergoths/es/s/
23:58.11gb2pb: redhat does that
23:58.24pb_yeah
23:58.57gb2pb: eliminates the #ifdef __KERNEL__ portions
23:59.05gb2and kernel-only headers are like this:
23:59.06gb2#error Invalid kernel header included in userspace
23:59.20kergothyeah, i kind of like that way of doing it
23:59.31kergothwhich is surprising, given how much i tend to dislike rh distributionisms
23:59.38pb_heh

Generated by irclog2html.pl by Jeff Waugh - find it at freshmeat.net! Modified by Tim Riker to work with blootbot logs, split per channel, etc.