00:00.15 | rwhitby-away | oh. ok - I'll check back in about 10 hours (after work) |
00:03.06 | *** join/#oe Crofton (~balister@66-207-66-26.black.dmt.ntelos.net) |
00:03.43 | Cwiiis | pb_: Thanks for the configure.in tip, almost fixed now :) |
00:12.57 | *** join/#oe micropal (ident@139.80-202-82.nextgentel.com) |
00:19.37 | Crofton | How do you get kernel modules in the fs image? |
00:31.43 | Ken|JLimeX | kergoth, you left yet? |
00:33.56 | Cwiiis | hmm... How do I specify that the host gcc get used to compile something in a Makefile.am - Something like 'CC = $(HOSTCC)' I'd guess? |
00:35.55 | Cwiiis | oh no, nevermind, that's a silly suggestion |
00:37.20 | treke | ooh libvncclient has been released |
00:39.29 | treke | I should try updating keypeblle to use it |
00:39.56 | treke | that could conceivably fix a ton of performance issues |
00:46.41 | *** join/#oe kergoth`laptop (~kergoth@ip71.churchill-park.dfw.ygnition.net) |
00:46.48 | treke | hey k |
00:46.51 | kergoth`laptop | hey |
00:47.54 | woglinde | kergoth try a icp-vortex |
00:49.22 | kergoth`laptop | heh, pricey |
00:49.29 | woglinde | kergoth *g* |
00:53.06 | woglinde | *sigh* a freind oy my made a trip for 78 days |
00:53.24 | woglinde | 8 countrys around the world |
00:54.59 | *** join/#oe kergoth_ (~kergoth@li11-226.members.linode.com) |
00:55.07 | *** join/#oe kergoth__ (~kergoth@li11-226.members.linode.com) |
00:55.20 | kergoth`laptop | hrm |
00:56.05 | *** join/#oe kergoth_ (~kergoth@li11-226.members.linode.com) |
00:57.38 | *** part/#oe kergoth`laptop (~kergoth@ip71.churchill-park.dfw.ygnition.net) |
01:02.03 | *** join/#oe kergoth_ (~kergoth@proxysb01.ext.ti.com) |
01:03.43 | Ken|JLimeX | Aaaarghhhh |
01:04.23 | kergoth | heh, no luck? |
01:05.00 | Ken|JLimeX | Guess :P , i've tried it all dynamic, static, glibc, uclibc, -O |
01:05.05 | Ken|JLimeX | but no cigar |
01:05.17 | Ken|JLimeX | i got a feeling that its building for SH but the wrong one :) |
01:07.03 | Ken|JLimeX | trying to set arch to sh3-linux |
01:07.12 | Ken|JLimeX | but is OE just checking for keyword sh3? |
01:07.51 | kergoth | ? |
01:08.14 | kergoth | when TARGET_OS is linux, and TARGET_ARCH is sh3, it passes --host=sh3-linux to ./configure's |
01:08.22 | kergoth | (or --target, depending on the package in question) |
01:08.53 | reenoo_ | 'night all |
01:10.14 | Ken|JLimeX | hmm, well im lost for the reason its not working then. Everything looks fine |
01:11.16 | kergoth | Ken|JLimeX: did you try changing the optimizations? |
01:12.20 | Ken|JLimeX | yeah |
01:12.25 | Ken|JLimeX | i set it to "-O" |
01:12.33 | Ken|JLimeX | FULL_OPTIMIZATION = "-O" |
01:12.40 | kergoth | adn still nothing? |
01:12.40 | Ken|JLimeX | O as in Operah |
01:12.40 | kergoth | weird |
01:13.16 | Ken|JLimeX | getting exact error as always, file says its an hitachi elf 32bit |
01:13.20 | Ken|JLimeX | but cannot run it |
01:17.07 | *** join/#oe noclouds (~mhfan@61.191.205.89) |
01:18.09 | Crofton | How do you get kernel modules in the fs image? |
01:28.08 | Crofton | ping |
01:32.07 | *** join/#oe aloph1 (~aloph@cnet-cable-189-84-61.canbrasnet.com.br) |
01:32.13 | Cwiiis | Anyone know why using the program 'more' on recent gpe-images as user says permission denied on /dev/console? (when you give it permission, it complains no such device :p) |
01:33.22 | *** join/#oe noclouds (~mhfan@61.191.205.89) |
01:33.58 | Cwiiis | runs fine as root |
01:36.24 | *** part/#oe dougt (~dougt@adsl-63-198-216-19.dsl.snfc21.pacbell.net) |
01:36.37 | *** join/#oe Anst (~anst@198.36.32.9) |
01:39.14 | Cwiiis | damnit, you spend ages getting a package to build and then it just crashes on you anyway... |
01:56.24 | *** join/#oe rwhitby-away (~rwhitby@rwhitby.nslu2-linux) |
02:04.02 | *** join/#oe offroadgeek (~root@68.184.192.44.charter-stl.com) |
02:04.20 | *** join/#oe noclouds (~mhfan@61.191.205.89) |
02:09.02 | *** join/#oe raster (~raster@vagw.valinux.co.jp) |
02:26.34 | treke | raster |
02:39.20 | Cwiiis | heh, I take it back, I like autotools... It's far more tedious fixing broken custom scripts for OE than broken autotool'd scripts :p |
02:39.36 | woglinde | cwiisss *g* |
02:40.01 | woglinde | mostly the programmers use autotools wrong |
02:41.24 | kergoth | i've said it before and i'll say it again. autotools may suck, but at least it sucks in known predicatable ways that can be worked around. |
02:43.24 | woglinde | hm 4 a.m. |
02:43.51 | raster | treke: trekeman! |
02:44.12 | raster | mmm |
02:44.16 | raster | predictable sucking |
02:44.41 | kergoth | hehe |
02:46.51 | woglinde | bah uclibc is lame |
02:47.05 | woglinde | even with debugging on |
02:47.09 | woglinde | lcars: can't resolve symbol '' |
02:48.07 | raster | speaking of sucking |
02:48.19 | raster | man does kdrive suck on c830's |
02:53.44 | woglinde | hm I will ask andersee tomorrow |
02:53.53 | woglinde | hm or |
02:54.15 | andersee | woglinde: eh? |
02:58.02 | Cwiiis | raster: Slow? |
02:58.53 | Cwiiis | kergoth: Heh, right... I should learn to use them properly at some point - It doesn't look quite as complex as I first thought |
03:00.12 | raster | Cwiiis: try moving a window :) |
03:00.17 | raster | (opaque) |
03:00.18 | raster | :) |
03:00.31 | raster | u can WATCH it blit the pixels |
03:00.35 | raster | as ithas no accel |
03:00.54 | raster | and now it needs to do across-bus to offboard vidoe chipset read/write cycles |
03:01.18 | *** join/#oe _drak0_ (~rob@ip68-98-148-202.dc.dc.cox.net) |
03:01.54 | woglinde | good night |
03:15.44 | Cwiiis | raster: Nasty... |
03:16.42 | raster | jah |
03:24.14 | *** join/#oe offroadgeek (~michael@68.184.192.44.charter-stl.com) |
03:30.17 | treke|home | raster: yeah that's nasty |
03:31.29 | raster | yeah |
03:31.30 | raster | very nasty |
03:31.51 | treke|home | the scary part is it's improved significantly |
03:38.27 | Cwiiis | argh, bloody make! There is no unterminated variable reference! |
03:40.53 | kergoth | heh |
03:41.25 | Cwiiis | Maybe I'm just tired... Can anyone see an unterminated variable reference here? Apparently it's at line 25, it happens just after all DGENOBJS finish compiling - http://www.pastebin.com/246099 (line 26 on this) |
03:41.29 | treke|home | wonder how well this libvncclient works |
03:46.20 | Cwiiis | I've looked over and over and I can't find it, assuming there is one :/ |
03:46.31 | *** join/#oe hedgehair (~hedgehair@user-0c6sdm3.cable.mindspring.com) |
03:48.26 | Cwiiis | oh God, I'm so dumb... found it (in Makefile.conf...) |
03:48.34 | Cwiiis | ${STAGING_LIBDIR)... Nice :p |
03:51.59 | kergoth | hehe |
03:59.20 | Cwiiis | heh, make completely arbitarily leaves a file out unless I specify instructions for it specifically... nice... |
04:12.30 | hedgehair | where do i put the source of the prog i want to cross compile? /stuff/build ? |
04:17.14 | kergoth | what? |
04:17.26 | Cwiiis | Is there an OE variable like CC, but for g++? |
04:17.32 | kergoth | Cwiiis: read bitbake.conf. |
04:17.36 | Cwiiis | thanks |
04:21.18 | _drak0_ | anyone try to compile apache lately? i think it's broken |
04:21.26 | _drak0_ | | cp: cannot stat `/build/find_apu.m4': No such file or directory |
04:25.25 | *** join/#oe micropal_ (ident@139.80-202-82.nextgentel.com) |
04:26.52 | Cwiiis | I'm being punished for cursing autotools... |
04:27.18 | treke|home | not possible |
04:27.32 | treke|home | if you could be punished for cursing autotools, kergoth would be in deep shit |
04:27.42 | kergoth | no kidding |
04:28.24 | treke|home | he's called it things I can't repeat in polite company |
04:33.25 | Cwiiis | http://www.pastebin.com/246113 - Can anyone point out why the code there throws the error at the bottom? I'm afraid my C++ skills leave something to be desired... I guess there's a missing cast or something? |
04:34.25 | Cwiiis | It gives that error three times, so I'm guessing it might have something to do with the NULL terminator at the end? |
04:34.55 | treke|home | bad |
04:34.59 | treke|home | give more info |
04:35.02 | treke|home | like line numbers |
04:35.04 | treke|home | and full files |
04:35.29 | treke|home | it looks like it's related to line 6 |
04:36.31 | Cwiiis | It fails on the last line of the struct - I can paste the full file? It's just a bit bug, I was hoping it'd be obvious :p hehe |
04:38.09 | Cwiiis | ah, it's probably the calls to atoi... |
04:38.21 | Cwiiis | yes |
04:46.40 | Cwiiis | atoi doesn't throw an exception does it? :/ |
04:46.48 | Cwiiis | I can't find any info anywhere... |
04:47.26 | treke|home | doubt it |
04:47.53 | Cwiiis | feh, this isn't even C++ code anyway... |
04:48.00 | treke|home | seems like it would cause problems with c apps calling it |
04:49.54 | Cwiiis | treke|home: Do you think I'm wrong on the error perhaps? |
04:50.30 | treke|home | beats me |
04:50.45 | treke|home | there wasnt nearly enough information to do more than guess |
04:53.20 | Cwiiis | I was right :) |
04:53.38 | Cwiiis | Just created a wrapper _atoi |
05:12.03 | *** join/#oe qip|Virus|da (~Thomas@pD95DBF65.dip.t-dialin.net) |
05:13.48 | *** join/#oe offroadgeek (~michael@68.184.192.44.charter-stl.com) |
05:17.53 | Cwiiis | "NOTE: package dgen-sdl-1.18-r0: task do_build: completed" :) |
05:25.56 | [g2] | kergoth, remember when vapier was using all the non-usual paths for building OpenSlug several weeks ago, including symlinks and stuff ? |
05:28.19 | kergoth | [g2]: yeah |
05:29.14 | [g2] | I'm gonna update the OE/BB procedures for new Beta users to build, things are slightly dated. |
05:29.56 | [g2] | Do you want me to recommend only the standard paths/slots or has all that stuff been fixed and users can try stuff like vaiper did |
05:32.30 | kergoth | you cannot build inside of a path that contains symlinks. |
05:32.36 | kergoth | thats the restriction. |
05:32.46 | kergoth | due to an autoconf/oe bug |
05:32.55 | kergoth | other than that, do whatever you want |
05:33.22 | [g2] | ok sounds good |
05:34.14 | [g2] | thx |
05:34.17 | kergoth | np |
05:36.15 | *** join/#oe offroadgeek (~michael@68.184.192.44.charter-stl.com) |
05:41.31 | *** join/#oe zedkatuf (~dfutak@82-32-58-45.cable.ubr08.azte.blueyonder.co.uk) |
05:44.34 | *** join/#oe Shdwdrgn (Picard@sourpuss.net) |
06:00.06 | CIA-8 | 03cwiiis 07 * r1.3098.1.2 10openembedded/packages/neon/ (neon_0.24.7.bb files/no-func-checks.patch): Add neon, some web-related library thingy that encompass (not added yet) uses. |
06:00.12 | CIA-8 | 03cwiiis 07 * r1.3098.1.1 10openembedded/packages/dgen/ (4 files in 3 dirs): Add two versions of dgen - one known working on collie, the other most recent (seems to segfault on startup). Needs testing. |
06:00.54 | Cwiiis | the 'working' one doesn't seem to work for me :( |
06:01.05 | Cwiiis | It runs and stuff, but roms don't load properly I guess... |
06:01.31 | Cwiiis | I compiled it a long time ago for qpe/sharprom and it was fine then... |
06:04.26 | Cwiiis | Think I'll sleep now... |
06:13.46 | *** join/#oe ibot (ibot@apt.bot.TimRiker.active.supporter.pdpc) |
06:13.46 | *** topic/#oe is OpenEmbedded Developer Lounge | http://bugs.openembedded.org/ (l/p guest/guest) | This is not a distribution support channel | The OE mailing list is back, but everyone needs to resubscribe. | We have a bitbake-dev mailing list for discussions on the core. If you're interested in BitBake-Ng, please subscribe. | LCA 2005 Embedded Miniconf - http://www.openembedded.org/miniconf |
06:33.05 | *** join/#oe DN (DN@cpe-24-175-124-238.houston.res.rr.com) |
06:33.10 | *** part/#oe DN (DN@cpe-24-175-124-238.houston.res.rr.com) |
06:45.14 | *** join/#oe ibot (ibot@apt.bot.TimRiker.active.supporter.pdpc) |
06:45.14 | *** topic/#oe is OpenEmbedded Developer Lounge | http://bugs.openembedded.org/ (l/p guest/guest) | This is not a distribution support channel | The OE mailing list is back, but everyone needs to resubscribe. | We have a bitbake-dev mailing list for discussions on the core. If you're interested in BitBake-Ng, please subscribe. | LCA 2005 Embedded Miniconf - http://www.openembedded.org/miniconf |
06:58.11 | *** join/#oe noclouds (~mhfan@61.191.205.89) |
07:09.35 | *** join/#oe Timelord (~TL@64-147-65-91.access-3.intertex.net) |
07:34.00 | *** join/#oe ljp (~lpotter@203.94.178.46) |
07:48.02 | *** join/#oe cedric_ (~cedric@195.94.59.29) |
08:11.05 | *** join/#oe gnufan (~mhfan@61.191.206.153) |
08:29.02 | *** join/#oe Virusmaster (~Thomas@p5085C449.dip0.t-ipconnect.de) |
08:41.44 | hrw|work | morning all |
08:53.05 | *** join/#oe aloph1_ (~aloph@cnet-cable-160-104-148.canbrasnet.com.br) |
08:55.09 | koen | 'morning all |
08:55.20 | hrw|work | hi koen |
08:55.27 | koen | hey hrw|work |
08:55.34 | hrw|work | koen: familiar 0.8.1 will be released? |
08:56.17 | koen | somewhere in the next few hours |
08:57.08 | hrw|work | great |
08:59.57 | hrw|work | hardfloat? |
09:00.22 | koen | yes |
09:00.31 | koen | we need compat with debian-arm |
09:01.05 | hrw|work | ah |
09:01.28 | hrw|work | during oz 3.3.6-pre1 times I sometimes used pacakges from debian |
09:02.16 | koen | once OE has more packages we might break compat |
09:02.50 | hrw|work | argh.. my toolchain is older then current so even gcc-cross-initial is going to rebuidl ;( |
09:11.31 | *** join/#oe mallum (~mallum@host81-155-189-183.range81-155.btcentralplus.com) |
09:14.03 | *** join/#oe noclouds (~mhfan@61.191.206.153) |
09:24.53 | hrw|work | ok. bought keyboard. |
09:25.24 | *** join/#oe ade|desk (~adavey@213-232-83-137.dsl.prodigynet.co.uk) |
09:25.37 | ade|desk | morning all |
09:26.13 | hrw|work | hi ade |
09:26.50 | ade|desk | anything exciting happend while i've been ill ? |
09:31.00 | *** join/#oe RockClimber (~Philip@spud.chem.ox.ac.uk) |
09:35.39 | RockClimber | Are any updates to .bb considered too petty? I mean things like improving the package description? |
09:38.43 | hrw|work | I think that better descriptions etc are good |
09:41.16 | RockClimber | I'm trying to build gnokii, and I've found it depends on X11. This is because gnokii can have an X11 front end, but it would be very useful for opie people to have it without the front end? What is the solution? Can 2 .bb packages exist, gnokii-x11 and gnokii? |
09:41.54 | *** join/#oe Dair|work (~dair@ip23.127.adsl.wplus.ru) |
09:43.27 | hrw|work | RockClimber: there is gammu in OE |
09:44.35 | Dair|work | again about bitbake: I have two different programs in two different dirs with different .pro-files. But they should be in one package (ipk). How should I write .bb for this? |
09:45.06 | hrw|work | Dair|work: iirc there are aps like that in OE |
09:45.37 | hrw|work | basically you have to write own do_configure() and do_compile() |
09:45.54 | RockClimber | hrw|work, what about gammu? |
09:46.26 | hrw|work | RockClimber: gammu is console util to manipulate phones |
09:47.17 | Dair|work | hrw|work: right, I already came this long :-) But sorry I don;t understand your first phrase ("iirc ..." ) |
09:47.28 | hrw|work | ~iirc |
09:47.29 | ibot | well, iirc is "if I recall correctly" |
09:47.44 | Dair|work | oh, sorry |
09:48.03 | koen | or you could make gnokii.bb make a gnokii.ipk and gnokii-frontend.ipk |
09:48.16 | hrw|work | gnokii.ipkg and xgnokii.ipk |
09:49.57 | *** join/#oe pb_ (~pb@2002:3e03:42cc:2:20a:5eff:fe00:391e) |
09:56.17 | Dair|work | hrw|work: excuse me, don't you recall just one place where could it be? ;-) I'm really confused what to 'find' or 'grep' ;-) |
09:57.38 | hrw|work | cumulus/cumulus_1.2.bb |
09:58.13 | *** join/#oe micropal (ident@139.80-202-82.nextgentel.com) |
09:58.15 | Dair|work | thanks again |
09:58.16 | hrw|work | Dair|work: btw - if it is your app then consider rewrite to have one .pro file or split to two packages |
09:59.45 | Dair|work | hrw|work: well, I have the power to do that... But first I'd like to force BB to fit to our project not vice versa ;-) |
10:00.59 | hrw|work | ;) |
10:01.30 | hrw|work | iirc even trolltech styleguide recommends one .pro for whole project |
10:02.17 | Dair|work | it depends of what you call "project" then ;-) |
10:02.48 | hrw|work | for me a project for me project == package |
10:05.40 | hrw|work | Dair|work: anyway - as long as you will maintain .bb for it I will be quiet about how it is done |
10:06.02 | RockClimber | how do you go about learning how to write a .bb file. I've read the manual, read a few .bb files still strugling. |
10:06.42 | Dair|work | RockClimber: I AM struggling. Here with big help from hrw :-) |
10:06.56 | hrw|work | RockClimber: 1. touch PN_PV.bb 2. put SRC_URI = "source-link" 3. build |
10:07.19 | hrw|work | RockClimber: if it has configure then add "inherit autotools" |
10:07.59 | hrw|work | if it has .pro file and is qt./e app then "inherit palmtop" - if works then "inherit opie" instead and add APPTYPE, APPNAME, APPDESKTOP |
10:08.08 | Dair|work | hrw|work: :-)))) my boss wants complete explanation of what and how's done on our corp wiki ;-) |
10:11.18 | hrw|work | so write it |
10:12.15 | Dair|work | hrw|work: I will as soon as I build whole svn-tree. I'm half on my way doing it ;-) |
10:12.52 | hrw|work | ;) |
10:13.58 | *** join/#oe cedricbompart (~ced@adsl-soho-1-c18-p123.vtx.ch) |
10:13.58 | *** join/#oe techsoup (~Chipoclud@AMarseille-151-1-37-57.w82-122.abo.wanadoo.fr) |
10:21.50 | *** join/#oe guillermo (~chatzilla@dsl-082-083-145-044.arcor-ip.net) |
10:23.08 | hrw|work | omg... perl do_package() step is HUGE |
10:23.26 | hrw|work | I vote for separate dir in feed for it ;) |
10:25.49 | RockClimber | Building gnokii I get "| /usr/lib/libgtk.so: could not read symbols: File in wrong format". It appears to be trying to link with my libgtk (i386) not an arm lib. Any hints on how to fix this - where do I start looking? |
10:26.44 | hrw|work | RockClimber: configure.ac or configure.in - fix searching for libs |
10:26.56 | hrw|work | or Makefile.ac/Makefile.in |
10:28.11 | *** join/#oe andersee (~andersee@codepoet.org) |
10:29.20 | RockClimber | hrw|work, if I modify the Makefile int tmp/work/gnokii.... and re-run bitbake will it continue using my modified file or overwrite it? |
10:30.35 | hrw|work | modify, create patch, add patch to SRC_URI, rebuild |
10:32.56 | *** join/#oe minchia (~asd@156.54.249.54) |
10:36.02 | RockClimber | I don't see libgtk in the oe packages, but ipkgfind shows a package exists |
10:36.49 | minchia | hello, anyone can help me to choose a CF ethernet for Zaurus? |
10:39.15 | *** join/#oe AMDPenguin (~nobody@aus-ext-proxy01.amd.com) |
10:40.15 | *** join/#oe mallum (~mallum@host81-155-189-183.range81-155.btcentralplus.com) |
10:43.30 | hrw|work | minchia: wired one? |
10:45.28 | minchia | yeah .. ethernet RJ45 |
10:49.08 | hrw|work | dont know any |
10:49.24 | RockClimber | Socket Low Power is one example |
10:52.56 | minchia | is one example of what? |
10:53.26 | pb_ | of cf ethernet cards |
10:54.03 | minchia | yeah .. but i'm looking for a model ... with support chipset |
10:54.18 | pb_ | eh? he just told you a model. |
10:54.28 | pb_ | <RockClimber> Socket Low Power is one example |
10:54.30 | pb_ | there it is. |
10:54.49 | RockClimber | In fact I even pasted a link in #openzaurus |
10:55.02 | pb_ | heh |
10:55.18 | minchia | i'm sorry |
10:55.19 | pb_ | oh well, what can you do. |
10:57.45 | pb_ | hm, what's going on here? |
10:57.46 | pb_ | cp: cannot create regular file `/home/pb/oe/build-h3900/tmp/staging/arm-linux/qt2/lib/fonts/helvetica_80_75_t10.qpf': Permission denied |
10:58.28 | RockClimber | you haven't run out of disk space - funny but I managed to |
10:58.43 | pb_ | heh, no, I have 14G free. |
11:01.20 | pb_ | oh well, maybe I can avoid building qte at all. |
11:02.04 | ade|desk | pb_: best choice |
11:02.17 | pb_ | yah |
11:02.36 | pb_ | I was trying to build frozen-bubble, but I had accidentally left libsdl-qpe selected as my preferred sdl provider. |
11:02.45 | ade|desk | oops |
11:05.06 | hrw|work | pb_: I just built it with qte without problem |
11:06.12 | pb_ | hrw|work: I guess you are as one with qte. |
11:06.38 | pb_ | I've never managed to quite attain the necessary karma for a successful build of anything qte-related. |
11:07.56 | pb_ | ah, here we go |
11:07.58 | pb_ | building perl now |
11:08.00 | hrw|work | ;)) |
11:08.17 | ade|desk | pb_: ah your anti TT campaining is coming back to haunt ;) |
11:08.26 | pb_ | ade|desk: yeah, it seems. |
11:22.52 | *** join/#oe freyther (~freyther@pD9E7E47E.dip0.t-ipconnect.de) |
11:24.46 | hrw|work | hi zecke |
11:25.18 | rwhitby | pb_: ping |
11:26.42 | rwhitby | pb_: I have another error like yesterday, and I need assistance from someone who knows how things should be done when rootfs_ipk.bbclass runs the postinst scripts after everything in IPKG_INSTALL gets installed in offline root mode |
11:27.41 | rwhitby | instead of the postinst scripts being installed in .../tmp/rootfs/usr/lib/ipkg/..., they are being installed in .../tmp/rootfs/home/slug/openslug/tmp/staging/i686-linux/lib/ipkg/... |
11:28.38 | rwhitby | which means that when rootfs_ipk.bbclass looks in .../tmp/rootfs/usr/lib/ipkg/... to find and run them, they are not there, and important things like the /sbin/init symlink don't get created (which makes it hard for the system to boot). |
11:29.04 | pb_ | ah, that's no good |
11:29.11 | pb_ | I wonder why the scripts are ending up in staging. |
11:29.28 | pb_ | did this start happening with ipkg-native 0.99.144? |
11:29.56 | rwhitby | I expect it's another case where there was a hard-coded /usr/lib in there, and kergoth changed it to some variable which is not what he thought it was in that place (as was the case for gcc-cross yesterday). |
11:30.17 | pb_ | yeah, could be |
11:30.28 | pb_ | anyway, tmp/rootfs/usr/lib/ipkg is the correct destination for those scripts. |
11:30.38 | rwhitby | this all started happening in the last two or three days. We were ramping up for an OpenSlug beta binary firmware release, and are now dead in the water :-( |
11:31.07 | rwhitby | ok, can you tell me what .bbclass file writes out the postinst scripts? |
11:31.52 | rwhitby | or is it when ipkg-cl is run to install all the packages that the postinst scripts get unpacked into the wrong place? |
11:34.08 | rwhitby | the call to ipkg-cl in real_do_rootfs looks ok: ipkg-cl -f /home/slug/openslug/tmp/work/openslug-image-1.0-r4/temp/ipkg.conf -o /home/slug/openslug/tmp/rootfs inst\ |
11:34.09 | rwhitby | all base-files base-passwd ... |
11:34.15 | hrw|work | maybe it's "lists_dir" option? |
11:34.44 | rwhitby | where is that set? |
11:37.21 | pb_ | rwhitby: right, ipkg-cl is responsible for unpacking the scripts |
11:37.28 | pb_ | they aren't installed by any .bb or .bbclass |
11:37.47 | rwhitby | so the arguments to ipkg-cl look ok - could it have a bad path compiled into it? |
11:37.52 | pb_ | yeah, that's possible |
11:38.02 | rwhitby | where would I look for that? |
11:38.13 | rwhitby | ipkg-native? |
11:38.16 | pb_ | yeah |
11:38.40 | rwhitby | ipkg.h: #define IPKG_STATE_DIR_PREFIX LIBDIR"/ipkg" |
11:38.54 | pb_ | that looks like the one |
11:38.57 | pb_ | what's LIBDIR at that point? |
11:39.00 | rwhitby | Makefile: ALL_CFLAGS = -g -O -Wall -DHOST_CPU_STR=\"i686\" -DBUILD_CPU=i686 -DLIBDIR=\"/home/slug/openslug/tmp/staging/i686-linux/lib\" |
11:39.00 | rwhitby | l |
11:39.04 | pb_ | hah |
11:39.14 | rwhitby | bzzt! next contestant please! |
11:40.19 | rwhitby | native.bbclass: |
11:40.41 | pb_ | #define IPKG_STATE_DIR_PREFIX "usr/lib/ipkg" |
11:40.47 | pb_ | that's what was in ipkg 0.99.140 |
11:40.52 | pb_ | so, I think Pigi broke this in 0.99.144. |
11:41.07 | rwhitby | base_prefix = "${exec_prefix}" |
11:41.07 | rwhitby | prefix = "${STAGING_DIR}" |
11:41.07 | rwhitby | exec_prefix = "${STAGING_DIR}/${BUILD_ARCH}-${BUILD_OS}" |
11:41.07 | rwhitby | # Base paths |
11:41.08 | rwhitby | base_bindir = "${base_prefix}/bin" |
11:41.08 | rwhitby | base_sbindir = "${base_prefix}/bin" |
11:41.10 | rwhitby | base_libdir = "${base_prefix}/lib" |
11:41.11 | pb_ | try rolling back your ipkg-native and see if matters improve. |
11:41.12 | rwhitby | # Architecture independent paths |
11:41.14 | rwhitby | sysconfdir = "${prefix}/etc" |
11:41.17 | rwhitby | sharedstatedir = "${prefix}/com" |
11:41.18 | rwhitby | localstatedir = "${prefix}/var" |
11:41.20 | rwhitby | infodir = "${datadir}/info" |
11:41.22 | rwhitby | mandir = "${datadir}/man" |
11:41.24 | rwhitby | docdir = "${datadir}/doc" |
11:41.26 | rwhitby | servicedir = "${prefix}/srv" |
11:41.28 | rwhitby | # Architecture dependent paths |
11:41.30 | rwhitby | bindir = "${exec_prefix}/bin" |
11:41.32 | rwhitby | sbindir = "${exec_prefix}/bin" |
11:41.34 | rwhitby | libexecdir = "${exec_prefix}/libexec" |
11:41.36 | rwhitby | libdir = "${exec_prefix}/lib" |
11:41.38 | rwhitby | oops - sorry! |
11:41.40 | rwhitby | note in there that prefix is set to STAGING, and then that is used for libdir |
11:41.41 | *** join/#oe cbrake_away (~cbrake@a22c174.neo.rr.com) |
11:41.50 | pb_ | yeah, that's no good |
11:41.54 | rwhitby | (I meant to cut that down to three lines, but accidentally hit return) |
11:43.48 | rwhitby | I wonder why OpenSlug is the first to find these things? Are all the other distros not compiling anything at the moment? |
11:46.38 | rwhitby | the problem was caused by kergoth in native.bbclass rev 1.14 and later |
11:47.16 | rwhitby | s/1.14/1.4/ |
11:47.26 | pb_ | well, it looks like an unfortunate combination of kergoth and pigi. |
11:48.02 | hrw|work | pb_: ipkg >.140 have option "lists_dir" for "usr/lib/ipkg" thing |
11:48.24 | pb_ | prior to 0.99.144, ipkg didn't care about LIBDIR by the looks of things. |
11:48.24 | pb_ | so, kergoth might have reasonably thought that his change was safe. |
11:48.24 | pb_ | hrw|work: that's for the lists specifically, not for the other files. |
11:48.38 | hrw|work | aha |
11:48.49 | rwhitby | should I patch ipkg.h back to #define IPKG_STATE_DIR_PREFIX "usr/lib/ipkg" ? |
11:49.05 | pb_ | it's there so that you can put the lists in ramfs, while still keeping the status file and other metadata in flash. |
11:49.06 | rwhitby | (I think that is what I will do locally until kergoth looks at it) |
11:49.14 | pb_ | yes, please do |
11:49.19 | pb_ | I think that is the right solution. |
11:49.33 | pb_ | ipkg.h is a generated file now, so you probably need to patch ipkg.h.in. |
11:52.07 | rwhitby | ah - so Pigi's change was in the ipkg CVS, not in OE ? |
12:04.09 | pb_ | right |
12:05.26 | rwhitby | pb_: can you let kergoth know if you see him before I do ? This must be affecting every single #oe distro in a major way - I cannot understand why no-one else has seen it. |
12:05.58 | *** join/#oe reenoo_ (~r@pD9528F1F.dip.t-dialin.net) |
12:07.05 | reenoo_ | morning |
12:13.21 | pb_ | rwhitby: ipkg 0.99.144 has only been in the repository for about 14 hours. |
12:13.26 | pb_ | I guess you were just the first person to notice. |
12:13.42 | pb_ | and yeah, Pigi's change was in the ipkg CVS tree |
12:13.49 | pb_ | hi reenoo_ |
12:13.57 | reenoo_ | hey pb_ |
12:15.21 | hrw|work | hi Rene |
12:15.35 | rwhitby | pb_: so if we set preferred version on ipkg to 0.99.143 we will be sweet? |
12:15.56 | pb_ | I think the one before 0.99.144 was actually 0.99.140, but yes |
12:16.05 | pb_ | oh, and you need ipkg-native, not ipkg itself |
12:16.09 | rwhitby | right |
12:21.34 | *** join/#oe mallum (~mallum@host81-155-189-183.range81-155.btcentralplus.com) |
12:32.26 | RockClimber | how do you make a patch file in the write format if you have configure.in configure.old? |
12:32.41 | pb_ | diff -u configure.old configure.in |
12:42.47 | CIA-8 | 03reenoo 07 * r1.3100.1.1 10openembedded/packages/ (6 files in 5 dirs): openvpn, ppp-dsl, prism54, shorewall, snes232: more kernel-module-* RDEPENDS/RRECOMMENDS fixes. |
12:42.47 | reenoo_ | ~lart br1 |
12:42.48 | RockClimber | earlier I asked about building gnokii without X support. It was suggested that maybe gnokii and xgnokii could build to different packages. But having tried this, I've hit the problem that gnokii links with libXpm and therefore requires X11. If you remove this you lose xgnokii. |
12:43.12 | reenoo_ | ~botsnack |
12:43.12 | ibot | thanks, reenoo_ |
12:43.41 | koen | RockClimber: in that case to .bbs are a better solution, but use a .inc to share the common stuff, like srcuri and license |
12:43.53 | RockClimber | Would it be acceptable to produce gnokii.bb and xgnokii.bb which produce separate packages that conflict (ie you can install one or the other) |
12:46.02 | koen | something like that |
12:46.02 | reenoo_ | pb_: I think I've fixed all of that kind now |
12:46.56 | reenoo_ | (think as in grep doesn't reveal any other ones) |
12:48.10 | pb_ | reenoo_: excellent |
12:49.29 | *** part/#oe rwhitby-treo (~rwhitby@rwhitby.nslu2-linux) |
12:49.36 | RockClimber | once you have oe installed, is it possible to manually compile source code using the cross compiler. what paths need to be set. I think this would be useful for helping me learn how configure and makefile work. |
12:57.20 | Crofton | do you know what dir the cross tools are in? |
12:57.43 | koen | $TMPDIR/cross iirc |
12:57.53 | RockClimber | how to I pass extra instuctions to a configure? |
12:58.35 | Crofton | yeah that sounds about right |
12:58.47 | reenoo_ | RockClimber: from a .bb simply set EXTRA_OECONF = "your configure options here" |
12:59.01 | Crofton | ./configure --target=something CC=arm-linux-gcc |
12:59.09 | Crofton | or something along those lines |
12:59.15 | Crofton | google |
12:59.35 | RockClimber | I meant from the bb file, so reenoo_ had the answer |
13:00.00 | Crofton | ah |
13:00.31 | RockClimber | until I worked out how to patch configure.in I wanted to compile gnokii --withou-x |
13:00.45 | Crofton | I'll be in the same boat you are i a bit, nce I start working |
13:00.50 | reenoo_ | RockClimber: the .bb needs to "inherit autotools" for that to work though |
13:01.32 | *** join/#oe [g2] (~g2@g2.nslu2-linux) |
13:03.21 | RockClimber | it does |
13:03.53 | CIA-8 | 03hrw 07 * r1.3110.1.1 10openembedded/packages/kdepimpi/kdepimpi_2.0.12.bb: kdepimpi updated to 2.0.12 |
13:04.53 | cedric_ | how can I detect the amount of RAM available in the kernel? |
13:06.03 | cedric_ | for example the C750/760/860 have 64M and the C700/SL5600 have 32M |
13:08.33 | hrw|work | cedric_: look how RP did it |
13:09.05 | hrw|work | hm.. but I dont know does he did c700 as 32M device or as pxa250 device.. |
13:09.58 | koen | c700 is in the c7x0 config |
13:10.12 | koen | the configs only know about flash size |
13:20.02 | *** join/#oe mrz80 (1000@caledonia.cns.ufl.edu) |
13:23.54 | *** join/#oe TyggerBob (~myname@64.26.144.98) |
13:27.16 | *** join/#oe stevenh (~xev@65.167.23.2) |
13:38.22 | [g2] | ~logs |
13:38.23 | ibot | apt/ibot/jbot/purl all log to http://ibot.rikers.org/<channelname>/ where channelname is html encoded ie: %23debian | lines that start with a space are not shown | some channels have stats at http://ibot.rikers.org/stats/<channelname>.html.gz |
13:46.39 | *** join/#oe XorA (~dp@81-178-218-37.dsl.pipex.com) |
13:46.52 | XorA | afternoon |
13:47.02 | Crofton | morning :) |
13:47.14 | [g2] | pb_ I'm looking to catch the discussion you had with rwhitby in here 2 hours ago, the logs at rikers aren't current enough yet is there another source ? |
13:47.23 | pb_ | I don't think so |
13:48.12 | pb_ | I can paste it to you in a query if you want. |
13:48.26 | [g2] | that'd be great |
13:48.33 | [g2] | thx |
13:50.39 | *** join/#oe McGregor (mcgregor@helios.et.put.poznan.pl) |
13:52.37 | pb_ | done |
13:55.51 | hrw|work | hi McGregor |
13:56.29 | McGregor | hi hrw |
13:56.56 | McGregor | hrw:tommorow High Noon ? |
13:58.05 | XorA | ok so never roll the time back when running Xvnc :-) |
14:01.11 | hrw|work | McGregor: 12:00 gmt+1 |
14:02.34 | cedricbompart | koen: I was talking about a C method which going to give me the memory amount |
14:14.19 | *** join/#oe J_ (~john@cpc4-mfld2-5-1-cust63.nott.cable.ntl.com) |
14:16.31 | *** join/#oe woglinde (woglinde@pD9E7EADA.dip.t-dialin.net) |
14:23.08 | *** join/#oe koen (~koen@212.41.157.237) |
14:23.21 | woglinde | morning koen |
14:24.14 | koen|train | hey woglinde |
14:24.23 | woglinde | hihi |
14:24.41 | woglinde | koehn where are ypu going? |
14:24.48 | emte | train as in choochoo? |
14:24.57 | koen|train | to my parents, and yes, as in choochoo |
14:24.59 | *** join/#oe chouimat (~dieu@r2351064.cidc.net) |
14:25.02 | emte | lol |
14:25.09 | woglinde | whats choochoo? |
14:25.31 | emte | wasnt sure if you had taken up a triathelon or something |
14:26.08 | emte | big metal machine that pulls cars along a track and makes a "choo choo" sound |
14:26.12 | koen|train | my parents live closer to brussels than I do, so I get to sleep an hour longer tomorrow |
14:26.14 | emte | :) |
14:26.32 | hrw|work | koen|train: nice bonus |
14:26.33 | emte | they must like the sprouts :P |
14:26.46 | emte | sorry bad pun |
14:26.49 | woglinde | koen yeah one hour can make big different |
14:27.58 | koen|train | Apeldoorn - Brussels is 'only' 4.5 hours |
14:39.54 | *** join/#oe RockClimber (~Philip@spud.chem.ox.ac.uk) |
14:40.05 | woglinde | hi rockclimber |
14:40.42 | RockClimber | reenoo_, what was that line in .bb to pass commands to configure "something_OECONF=" I think? |
14:40.58 | koen|train | EXTRA_OECONF |
14:45.04 | reenoo_ | RockClimber: like koen|train said, but best use grep on other .bbs for reference. |
14:45.44 | reenoo_ | RockClimber: if you remember OECONF you'd use something like "cd packages; grep OECONF */*bb | less" |
14:52.53 | *** join/#oe cougar_ (~cougar@ASt-Lambert-102-2-1-215.w193-252.abo.wanadoo.fr) |
14:54.42 | *** join/#oe qfh (~qfh@ppp-82-135-3-138.mnet-online.de) |
14:56.30 | *** join/#oe qfh (~qfh@ppp-82-135-3-138.mnet-online.de) |
14:56.33 | cougar_ | Hi all, I was wondering to use bitbake to create my own linux distrib. I was asking myself which was the good solution to do a such thing : 1) create an ext3 filesystem and copy it on the target host (likely the jffs2 images for ipaq) or 2) create some packages and an installer on the target host... Ideas ? |
14:57.20 | koen|train | an installer would be the nicest thing |
14:57.29 | koen|train | ext3 the easiest |
15:03.18 | cougar_ | Is there already a bitbake project which generate ext2/ext3 filesystem which I could inspire from? |
15:04.14 | koen|train | all you need is IMG_FSTYPE="ext3" (or something like that) |
15:04.28 | koen|train | you can taak a look at mythfront-image |
15:04.35 | koen|train | take* |
15:04.44 | cougar_ | Thanks, I'll try |
15:05.07 | Rupert-Giles | anyone ever put gentoo on a Z? |
15:05.19 | koen|train | yes |
15:05.28 | koen|train | not me, but I've heard of it |
15:06.27 | Rupert-Giles | hmmm wonder if it's worth a try |
15:06.30 | *** join/#oe qfh (~qfh@ppp-82-135-3-138.mnet-online.de) |
15:06.50 | koen|train | given the ram and cpupower available: it isn't |
15:07.06 | Rupert-Giles | hmm |
15:17.43 | reenoo_ | there's no point in it either |
15:18.02 | reenoo_ | no device specific stuff etc. |
15:23.54 | *** join/#oe micropal_ (ident@139.80-202-82.nextgentel.com) |
15:25.00 | ade|desk | oh shoot, council tax is going up :( |
15:25.57 | Rupert-Giles | haha |
15:26.49 | ade|desk | they are re doing the boundaries too. |
15:27.22 | *** join/#oe tigrux (~tigrux@charybdis-ext.suse.de) |
15:27.59 | *** join/#oe Laibsch (~leggewie@G151b.g.pppool.de) |
15:28.41 | *** join/#oe zwi (~zwi@216.88.131.43) |
15:28.54 | Rupert-Giles | n000 |
15:30.04 | woglinde | coreutils rocks |
15:30.08 | woglinde | *g* |
15:32.17 | ade|desk | ~change 150000 gbp to usd |
15:40.01 | hrw|work | bye |
15:46.05 | RockClimber | I finally got gnokii to build without X11, and install in on my Zaurus and then found out that IrDA doesn't work with my phone a gnokii (it breaks the phone) - how annoying! |
15:57.04 | *** join/#oe zinga (~arnaudb@186.80-203-227.nextgentel.com) |
16:08.13 | *** join/#oe CoreDump|home (~mhentges@hentges.net) |
16:12.06 | cedric_ | I'm palying with frequency scaling under 2.6 |
16:13.30 | cedric_ | I've overclock to 471Mhz like on 2.4 |
16:19.29 | kergoth_ | bleh |
16:19.30 | kergoth_ | morning |
16:19.47 | woglinde | morning kergoth |
16:19.55 | woglinde | sleep well? |
16:20.41 | *** join/#oe ade|desk (~adavey@213-232-83-137.dsl.prodigynet.co.uk) |
16:21.05 | kergoth_ | heh |
16:21.27 | kergoth_ | yeah, then woke up to no power. wonder if i was late on the bill, or if they needed to cut it during my xfer of service from txu to reliant |
16:21.39 | kergoth_ | then lost my work security key, so floundered about in my apartment for like half an hour |
16:21.58 | woglinde | ouch buts friday not monday |
16:22.05 | kergoth_ | could always be worse |
16:23.14 | CosmicPenguin | yeah, it really *could* be monday |
16:23.32 | CosmicPenguin | Or wait, thats what they just said |
16:24.13 | kergoth_ | hehe |
16:24.51 | woglinde | hm no tea here anymore |
16:34.35 | *** join/#oe Ken|JL (~Ken@h237n2fls34o1010.telia.com) |
16:37.53 | *** join/#oe marcansoft (1337@80.26.222.224) |
16:42.06 | *** join/#oe ar_ (~ar@port-ip-213-211-233-162.reverse.mdcc-fun.de) |
16:52.34 | *** join/#oe Ken|JLimeX (~Ken@h237n2fls34o1010.telia.com) |
16:57.40 | *** join/#oe rphillips (~rphillips@rphillips.developer.gentoo) |
17:02.55 | treke | blah |
17:05.21 | kergoth | heh |
17:06.29 | *** join/#oe SirFred (~mteira@85.Red-81-34-170.pooles.rima-tde.net) |
17:06.34 | SirFred | Hi. |
17:12.46 | CosmicPenguin | kergoth: every played with this: http://directory.fsf.org/devel/compilers/acovea.html |
17:13.01 | CosmicPenguin | s/every/ever/ |
17:13.35 | kergoth | nope |
17:14.38 | CosmicPenguin | Looks interesting - I think it would be valuable to hack it so that instead of measuring performance it measured size instead |
17:15.06 | CosmicPenguin | find the set of gcc options that resulted in the smallest running binary |
17:16.47 | *** join/#oe zap (~zap@217.170.93.196) |
17:18.04 | kergoth | nod, looks interestin |
17:18.07 | kergoth | g |
17:21.36 | *** join/#oe Ken|JL (~Ken@h237n2fls34o1010.telia.com) |
17:30.42 | *** join/#oe Crofton (~balister@mprg5213.mprg.ee.vt.edu) |
17:31.19 | *** join/#oe treke (~ggilbert@tinman.treke.net) |
17:34.42 | *** join/#oe dkey (~dkey@L0010P25.dipool.highway.telekom.at) |
17:36.18 | Crofton | | ERROR: Cannot satisfy the following dependencies for task-bootstrap: |
17:36.19 | Crofton | | kernel-module-pcmcia_core kernel-module-omap_cf kernel-module-nls_iso8859-1 kernel-module-nls_cp437 kernel-module-nls_base |
17:36.19 | Crofton | | |
17:36.22 | Crofton | hmm |
17:37.12 | kergoth | Crofton: your BOOTSTRAP_EXTRA_* are such that those modules are depended on, but your kernel build doesnt have those things enabled, so ipks for them werent created. |
17:37.15 | kergoth | Crofton: fix one or the other. |
17:37.38 | Crofton | I added some stuff to the machine file for the modules |
17:37.50 | Crofton | since the machine needs the modules to suuport the OSK HW |
17:37.54 | kergoth | yes, apparently you added dependencies on modules that dont exist. |
17:38.05 | kergoth | adjust your kernel defconfig to build those things as modules |
17:38.25 | Crofton | I am almost 99% certian they are built |
17:38.42 | kergoth | well i'm 100% certain they arent. or there's something else screwed up in your kernel's .bb |
17:38.49 | Crofton | # IDE modules |
17:38.49 | Crofton | BOOTSTRAP_EXTRA_RDEPENDS += "kernel-module-ide-cs kernel-module-ide-disk" |
17:38.49 | Crofton | BOOTSTRAP_EXTRA_RDEPENDS += "kernel-module-ide-core" |
17:38.52 | kergoth | oe's kernel.bbclass automatically does all the module packaging |
17:38.54 | Crofton | Heh |
17:38.55 | kergoth | all you need to do is build them |
17:38.59 | Crofton | hmm |
17:39.24 | Crofton | so the .ko's automatically are "exported" from the kernel build |
17:39.41 | kergoth | do_install installs them |
17:39.48 | kergoth | via make modules_install or whhatever |
17:39.49 | kergoth | into ${D} |
17:39.57 | Crofton | and D is? |
17:40.04 | kergoth | nothing you need to care about. |
17:40.08 | CosmicPenguin | heh - as modules are installed, they get touched and unionfs puts them into the rw partition |
17:40.13 | kergoth | its the rootfs |
17:40.14 | Crofton | well I want to know :) |
17:40.18 | CosmicPenguin | Somebody should address that |
17:40.20 | Crofton | ok |
17:40.22 | kergoth | its where you install things, the base point from which oe packages |
17:40.39 | Crofton | tmp/rootfs |
17:40.42 | Crofton | =D |
17:40.42 | kergoth | install installs everything into a rootfs location. package then breaks that out into subpackages |
17:40.46 | kergoth | no. |
17:40.48 | kergoth | tmp/rootfs is _image creation time_ |
17:40.57 | kergoth | each package has an individual "rootfs" location that it installs to |
17:41.02 | kergoth | which oe then uses to package it. |
17:41.12 | kergoth | ls tmp/work/blah/image/ |
17:41.14 | kergoth | ls tmp/work/blah/install/ |
17:41.40 | Crofton | arghh |
17:41.47 | Crofton | nothing there but dir structure |
17:41.50 | kergoth | it does a make install on the package into image |
17:41.59 | kergoth | then oe moves files from image to install/subpackage to break things out into the subpackages |
17:42.04 | kergoth | _moves_ |
17:42.08 | kergoth | an empty image/ dir is a good thing |
17:42.23 | kergoth | an empty install/ is not :) |
17:43.11 | Crofton | OK , .ko's are in install |
17:43.23 | kergoth | if they're in install, then oe emitted ipks for them |
17:43.23 | Laibsch | kergoth: have you decided yet on whether or not you will automatically move the old wiki over to moin? |
17:43.27 | kergoth | (given package_ipk inherited) |
17:43.50 | kergoth | Laibsch: havent really thought about it. i'm leaning towards not, since there's so much old crufty content in the old wiki that we probably dont want |
17:44.38 | Crofton | it is entirely likely I do not know what I am doing :) |
17:44.48 | kergoth | Crofton: is hte omap_cf.ko in install? |
17:45.20 | kergoth | Crofton: i find it extremely unlikely that thats the case. again, if your kernel build builds the modules, they get stuffed into ipks. its all automatic, unless you're overriding do_install in your kenrel .bb. |
17:45.23 | Crofton | drwxr-sr-x 3 balister users 4096 Feb 25 12:19 kernel-module-omap-cf |
17:45.50 | kergoth | 12:21 < Crofton> | ERROR: Cannot satisfy the following dependencies for task-bootstrap: |
17:45.53 | kergoth | 12:21 < Crofton> | kernel-module-pcmcia_core kernel-module-omap_cf kernel-module-nls_iso8859-1 |
17:45.57 | kergoth | <PROTECTED> |
17:45.57 | kergoth | wheres that coming from? |
17:46.00 | kergoth | omap_cf? |
17:46.00 | kergoth | omap_cf != omap-cf |
17:46.06 | Crofton | ah |
17:46.08 | Laibsch | kergoth: Well, but there still is valuable content as well. Manual conversion will most likely never (!) happen and my guess would be that throwing away bitrot after conversion ought to be rather swift. IMHO it would be a shame to lose all that old info even if some of it was outdated. |
17:46.23 | *** join/#oe keturn (~kevint@c-24-21-141-200.client.comcast.net) |
17:46.53 | kergoth | Laibsch: manual conversion has been happening just fine up till now. someone needs info that isnt int he new wiki, the9~y move it over |
17:46.59 | kergoth | you expect that to not continue? |
17:47.05 | Crofton | - != _ let me follow this insite |
17:47.07 | kergoth | i agree, users are fucking lazy, but.. |
17:47.12 | kergoth | heh |
17:47.47 | Laibsch | so are programmers. I guess without laziness, algorithms might have never been invented |
17:47.49 | Laibsch | ;-) |
17:47.52 | kergoth | yep |
17:47.59 | koen | ~laziness |
17:48.00 | ibot | LAZINESS: The quality that makes you go to great effort to reduce overall energy expenditure. It makes you write labor-saving programs that other people will find useful, and document what you wrote so you don't have to answer so many questions about it. Hence, the first great virtue of a programmer. |
17:48.06 | kergoth | i dunno, it doesnt matter much to me. we need to comb through every page individually no matter what |
17:48.32 | Laibsch | koen: I am perplexed at the profound knowledge of this bot. Who told him all that stuff? |
17:48.41 | kergoth | ~factinfo laziness |
17:48.41 | ibot | laziness -- created by greycat <~wooledg@imadev.eeg.ccf.org> at Wed Jun 26 15:59:40 2002 (975 days); it has been requested 10 times, last by koen, 41s ago. |
17:48.54 | kergoth | all sorts of people :) |
17:49.27 | *** join/#oe mrz80 (1000@caledonia.cns.ufl.edu) |
17:49.28 | Laibsch | I guess the coming about of this bot is the best example of ~laziness in action. ;-) |
17:49.48 | kergoth | indeed |
17:49.53 | kergoth | less typing is a Good Thing(tm) |
17:50.52 | Crofton | the module is called omap_cf.ko |
17:51.07 | Crofton | BOOTSTRAP_EXTRA_RDEPENDS += "kernel-module-omap_cf" |
17:51.20 | kergoth | Crofton: thats wrong. _ isnt a legal seperator in an ipk file |
17:51.26 | kergoth | Crofton: _ is used to seperate name from version and stuff |
17:51.29 | kergoth | Crofton: so oe automatically mangles it |
17:51.31 | Crofton | heh |
17:51.34 | kergoth | Crofton: you want to depend on kernel-module-omap-cf |
17:51.44 | Crofton | how do you get around it with pcmcia_core? |
17:51.48 | kergoth | what do you mean? |
17:51.56 | kergoth | kernel-module-pcmcia-core |
17:51.57 | kergoth | ta da |
17:52.10 | CosmicPenguin | Anyone wanna fight passionately for chosing sablevm over jamvm? |
17:52.11 | Crofton | so I depend on omap-cf |
17:52.16 | kergoth | 12:36 < kergoth> Crofton: you want to depend on kernel-module-omap-cf |
17:52.26 | reenoo_ | CosmicPenguin: huh? |
17:52.43 | koen | CosmicPenguin: jamvm segfault for me |
17:52.49 | Crofton | This begs the questioin: how does oe knw the module is really omap_cf? |
17:53.03 | CosmicPenguin | koen: I'll call that a good enough reason - sablevm it is |
17:53.18 | Laibsch | kergoth: Coming back to it not mattering much to you. The bot says that documenting is a virtue. So while you personally may not need that docu I hope you see the value of it. Combing through stuff to throw stuff away is easy IMHO. combing through stuff to throw stuff away and hand-convert the rest is much more pain, I'd say. To the computer doing the conversion it won't matter much if you later throw away half of it. |
17:53.19 | reenoo_ | koen: well.. so does sablevm |
17:53.34 | koen | D'oh |
17:53.57 | reenoo_ | CosmicPenguin: the difference is that jamvm crashes _every_ time while sablevm crashes _occasionally_ only |
17:54.24 | reenoo_ | and that's only for AWT/Swing apps anyway |
17:54.52 | reenoo_ | non-gui apps should just work |
17:55.53 | reenoo_ | (unless they rely on on buggy Sun code from older JDK releases) |
17:55.56 | CosmicPenguin | Well, I'm gonna give it a shot on a x86 |
17:56.00 | CosmicPenguin | maybe I'll have better luck |
17:56.08 | reenoo_ | heh |
17:56.15 | reenoo_ | that's a different story then |
17:56.44 | reenoo_ | sablevm even supports Eclipse on x86 |
17:58.18 | reenoo_ | CosmicPenguin: let me know about your findings. might help tracking down why things crash on ARM |
17:59.07 | kergoth | Crofton: yes. the package contains omap_cf.ko. just hte package name is different. |
17:59.24 | CosmicPenguin | reenoo_: I'll let you know . I've never even seen the inside of a JVM before ,so I'll be researching for a while |
17:59.34 | kergoth | Laibsch: nod, thats a valid point. automatic conversion is probably best then, but we need to get volunteers to comb through the content |
17:59.38 | Crofton | Heh |
17:59.43 | reenoo_ | CosmicPenguin: cool |
18:00.10 | Crofton | I think it is going better now, I missed the nls modules whe I s/_/-_ |
18:00.28 | kergoth | oh wow, i found an annoying oe/bb "bug" |
18:00.32 | kergoth | the gcc packaging does: |
18:00.33 | kergoth | +OLD_INHIBIT_PACKAGE_STRIP := "${INHIBIT_PACKAGE_STRIP}" |
18:00.33 | kergoth | +INHIBIT_PACKAGE_STRIP = "1" |
18:00.57 | kergoth | and then decides whether or not to strip the target packages (libgcc, etc) based on OLD_INHIBIT_PACKAGE_STRIP |
18:01.06 | kergoth | because the automatic stripping doesnt work, it runs the x86 strip (cross package) |
18:01.08 | kergoth | however |
18:01.14 | kergoth | if INHIBIT_PACKAGE_STRIP isnt set at _all_ |
18:01.23 | kergoth | := doesnt actually expand, because epxansion of non set variables returnst he same value |
18:01.45 | kergoth | so OLD_INHIBIT_PACKAGE_STRIP := "${INHIBIT_PACKAGE_STRIP}" only results in it being set to "${INHIBIT_PACKAGE_STRIP}" |
18:01.52 | kergoth | which later expands to 1 |
18:02.08 | kergoth | logic flaw in the .inc, it assumes that the variable has a default value that it can back up |
18:02.12 | reenoo_ | Crofton: kernel.bbclass translates all module names to valid package names the same way. |
18:03.11 | Crofton | OK |
18:03.15 | Laibsch | kergoth: I was just about to propose doing the combing from time to time. Maybe it would be a good idea to have all old pages in some kind of sub-directory so they are not part of the current wiki right, but they are there in the correct form. they could then be moved one at a time w/o having to do the really tedious part of wiki code conversion. As I said, I'll go through one page or the other from time to time if I do not have to convert wiki code. |
18:03.34 | kergoth | Laibsch: thats a good idea if we can manage it |
18:03.44 | kergoth | (the subdirectory thing) |
18:04.54 | Laibsch | I guess what you'd have to do is prepend oldwiki/ or some such to the created file names. Not sure if that is already implemented in the converter. |
18:04.58 | Crofton | Am I doing the "write thing" by specifying kernel modules in the machine.conf file? |
18:05.03 | CosmicPenguin | nutts |
18:05.15 | CosmicPenguin | I should have told andersee about the SHLIB thing... :) |
18:05.16 | CosmicPenguin | oops... :)( |
18:05.44 | kergoth | Crofton: generally yeah, things that hardware specific belong to the machine |
18:05.55 | kergoth | Crofton: though arguably, depending on what they are, some bits might belong to the distro |
18:05.57 | *** join/#oe mrz80 (1000@caledonia.cns.ufl.edu) |
18:06.31 | Crofton | yeah |
18:06.47 | Crofton | In my case I am using the CF feature |
18:06.50 | kergoth | modules for a piece of functionality which isnt device specific, thats probably a distro decision |
18:07.27 | Crofton | My thoughts are the machine file should support the hw preent on the board |
18:07.33 | Crofton | present |
18:15.37 | Laibsch | ~seen cbrake |
18:15.39 | ibot | cbrake is currently on #nslu2-linux #oe #elinux. Has said a total of 4 messages. Is idling for 2h 45m 4s |
18:16.33 | cbrake | Laibsch: I'm back |
18:17.40 | Laibsch | cbrake: Good. Just wanted to test your nick which is good to have so we can discuss things related to the users manual. |
18:20.40 | *** join/#oe Ken|JLimeX (~Ken@h237n2fls34o1010.telia.com) |
18:23.37 | koen | ~lart emte |
18:24.06 | koen | imlib 1.2.0 != imlib2 CVS HEAD |
18:26.00 | kergoth | heh |
18:27.46 | CIA-8 | 03kergoth 07 * r1.3113 10openembedded/packages/gcc/ (8 files): Unbork the gcc-cross ipk file creation bits per the recent target path changes. |
18:29.05 | kergoth | [g2]-lunch, VoodooZ_lunch: there. |
18:29.17 | VoodooZ_Work | thanks |
18:29.18 | Crofton | task-bootstrap is still looking for the modules with the _ in the name |
18:29.30 | kergoth | Crofton: you didnt rebuild task-bootstrap after changing the bootstrap dep variables. |
18:29.36 | kergoth | Crofton: warned you about that :) |
18:29.54 | Crofton | I just tried a bitbake task-bootstrap -c clean |
18:30.12 | Crofton | and did the same to the kernel |
18:30.35 | *** join/#oe Tygger_Bob (~myname@CPE00045a265f64-CM000f9f7c58fa.cpe.net.cable.rogers.com) |
18:32.42 | kergoth | Crofton: bbread /path/to/task-bootstrap.bb|grep RDEPENDS |
18:34.59 | Crofton | bother _ still present there ... |
18:36.11 | Crofton | thanks |
18:36.20 | Crofton | I must be blind |
18:36.45 | Crofton | bbread shows you the actual build script? |
18:39.12 | kergoth | bbread shows the metadata oe parses, exactly as it will be used |
18:39.31 | kergoth | so omap_cf must be getting added by someone in the .confs or .bbclasses |
18:39.41 | Crofton | heh that is what th manual said :) |
18:39.51 | kergoth | heh :) |
18:39.55 | Crofton | I was missing some in the nls modules actually |
18:40.20 | Crofton | I was hunting through all the places I though I might have put _'s |
18:42.10 | Crofton | if I do a bitbake task-bootstrap -c build |
18:42.19 | Crofton | Will that force a rebuild? |
18:43.05 | kergoth | no. |
18:43.11 | Crofton | heh |
18:43.12 | kergoth | that runs the build task |
18:43.22 | Crofton | clean the buld then? |
18:43.25 | kergoth | which has already been run, and oe knows that |
18:43.35 | Crofton | oe knows to much |
18:43.38 | kergoth | ls tmp/stamps/ |
18:43.42 | Crofton | I does not read my mind |
18:43.50 | kergoth | yes, clean and then build it |
18:44.08 | Crofton | or wack a stamp? |
18:44.46 | Crofton | although I had some issues with not being able to overwrite files during those attempts |
18:45.01 | kergoth | removing a stamp doesnt remove the files in tmp/work/ |
18:45.06 | kergoth | its sources will still be sitting there |
18:45.07 | Crofton | yeah |
18:45.09 | kergoth | things already built |
18:45.12 | Crofton | that seems to cause problems |
18:45.14 | kergoth | unless you, i.e. do a make distclean manually.. |
18:45.31 | kergoth | rarely a good idea unless you have a good reason for it |
18:45.36 | Crofton | but -c clean will do the right things |
18:45.41 | kergoth | like working on a patch inside of tmp/work/, so dont want the source tree wiped |
18:45.53 | kergoth | clean just wipes the stamps and its dir in tmp/work/ |
18:45.58 | CIA-8 | 03koen 07 * r1.3115 10openembedded/packages/kbdd/kbdd_cvs.bb: |
18:45.58 | CIA-8 | kbdd_cvs.bb: |
18:45.58 | CIA-8 | <PROTECTED> |
18:46.00 | CIA-8 | 03koen 07 * r1.3114 10openembedded/packages/imlib2/imlib2_1.2.0.bb: fix imlib_1.2.0 src_uri imlib 1.2.0 != imlib2 CVS HEAD |
18:46.04 | CIA-8 | 03koen 07 * r1.3113 10openembedded/conf/distro/familiar-0.8.2.conf: add familiar-0.8.2.conf |
18:46.09 | Crofton | OK I think I have beaten that into my this morning |
18:46.56 | Crofton | so task-bootstrap does the module stuff? |
18:47.11 | kergoth | task-bootstrap is the only package bootstrap-image installs. |
18:47.20 | kergoth | so its its responsibility to depend on everything you expect to be in a bootstrap image |
18:49.44 | Crofton | gah, just unpacked tar file in my oe dir |
18:49.58 | *** join/#oe nofpu (~matthew@82-45-183-10.cable.ubr06.dals.blueyonder.co.uk) |
18:50.00 | kergoth | bk sfiles -x shows files that arent under bk's control |
18:50.05 | kergoth | a little creative shell scripting is your friend |
18:50.08 | kergoth | heh |
18:50.16 | Crofton | actually one up from there |
18:50.20 | Crofton | fixbale |
18:50.38 | Crofton | excellent, we have modules |
18:52.35 | Crofton | next add, pcmcia stuff :) |
19:01.48 | Laibsch | kergoth: I have just talked with cbrake who is also helping with the users manual. He said he will try to install a moin wiki to one of his machines this weekend probably. We could do the conversion there but would need the phpwiki dump from you at some point in time. Pages could then be moved individually or thrown away but the code conversion would be automatic. What do you say? |
19:02.50 | *** join/#oe andersee (~andersee@codepoet.org) |
19:03.41 | kergoth | Laibsch: sounds good to me |
19:07.22 | CIA-8 | 03kergoth 07 * r1.3112.1.2 10openembedded/packages/ipkg/ (3 files in 2 dirs): Apply patch to ipkg 0.99.144 to add a --with-ipkgdir configure option, then alter ipkg-native 0.99.144 to back up the target libdir before native is inherited, and pass that via the new configure option. |
19:09.09 | Crofton | What's the difference between DEPENDS and RDEPNDS? |
19:09.15 | kergoth | build time and run time. |
19:09.20 | kergoth | DEPENDS is oe packages. .bb files. |
19:09.24 | kergoth | RDEPENDS is runtime. ipk files. |
19:09.36 | Crofton | one bb file can make severl ipk files? |
19:09.41 | koen | yes |
19:09.47 | koen | see tslib or sysvinit |
19:10.04 | kergoth | Crofton: every package does by default, actually. |
19:10.19 | kergoth | Crofton: we default to a split of packagename, packagename-dev (headers, etc), and packagename-doc (man pages, /usr/share/doc files) |
19:10.46 | Crofton | so I have to mention all the required BB files in the DEPEnDS variable |
19:11.03 | Crofton | and mention all the required ipk in the RDEPENDS |
19:11.06 | kergoth | right |
19:11.21 | Crofton | so for a bb file I need to mention it "twice" |
19:11.29 | kergoth | so if you specify just hte packagename in rdepends, you automatically dont get the -dev and -doc stuff, which is whats usually expected |
19:11.35 | Crofton | example abse-fiiles and base-files |
19:11.38 | kergoth | Crofton: we have automatic runtime package dependnecies for a lot of things. |
19:11.51 | kergoth | Crofton: so i.e. if your package DEPENDS on a library, and links to it, that gets in rdepensd automatically |
19:11.56 | kergoth | same for kernel module interdependencies |
19:12.01 | kergoth | but yeah, for what you're doing, thats right |
19:12.03 | kergoth | add to both |
19:12.29 | Crofton | I am getting ready to ad pcmcia support, guess I should read pcmcia.bb :) |
19:13.42 | CosmicPenguin | The native java compiler stuff doesn't work in oe, am i correct? |
19:13.47 | kergoth | 2.6, your kernel provides the modules, userspace is pcmcia-cs |
19:13.52 | kergoth | bleh |
19:14.32 | Crofton | yes looking through that now |
19:14.41 | reenoo_ | CosmicPenguin: define "native compiler stuff" |
19:15.00 | CosmicPenguin | reenoo_: I'm looking at the jolden benchmarks |
19:15.22 | CosmicPenguin | reenoo_: and I see this in the Makefile |
19:15.24 | CosmicPenguin | $(DEST)/%.class: %.java |
19:15.24 | CosmicPenguin | <PROTECTED> |
19:15.33 | CosmicPenguin | Where JAVAC=javac |
19:16.19 | CosmicPenguin | Since I am proud to be 100% java clueless, I don't know if this mysterious javac exists, where it exists, or if it is even needed |
19:16.19 | reenoo_ | you can use jikes-native for that |
19:16.26 | CosmicPenguin | tasty |
19:16.50 | reenoo_ | but you will need to provide it with a classpath |
19:17.01 | CosmicPenguin | sablevm-classpath will work for that? |
19:18.27 | reenoo_ | it should. have a look at the jikes-classpath or jikes-sablevm debian packages |
19:18.32 | VoodooZ_Work | CosmicPenguin: Have you heard of any changes with the AMD64 problem with ipkg? |
19:18.51 | CosmicPenguin | VoodooZ_Work: no, I haven't |
19:18.55 | VoodooZ_Work | cool. thanks. |
19:18.56 | CosmicPenguin | VoodooZ_Work: I've been meaning to try it |
19:19.17 | VoodooZ_Work | yeah, I remember somebody saying there was a new version. |
19:19.28 | VoodooZ_Work | I'm building right now so we'll see in 25 minutes' |
19:33.12 | *** join/#oe Crofton (~balister@mprg5213.mprg.ee.vt.edu) |
19:35.41 | kergoth | ~nslookup cvs1.sourceforge.net |
19:40.47 | Crofton | spelling is everything .... |
19:45.35 | Laibsch | kergoth: I just came across http://www.openembedded.org/contribution.php. I think you might want to specifically solicit contribution via documentation more actively. |
19:48.18 | VoodooZ_Work | CosmicPenguin: Seems to work. ipkg built ok under my opteron. |
19:48.46 | *** join/#oe Pigi (~NoOne@host3-70.pool21345.interbusiness.it) |
19:48.54 | Pigi | Ciao all |
19:49.37 | CosmicPenguin | VoodooZ_Work: excellent news |
19:49.53 | CosmicPenguin | ~praise the ipkg devs |
19:49.55 | ibot | All hail the ipkg devs! |
19:50.01 | CosmicPenguin | for the moment... :) |
19:50.12 | CosmicPenguin | Now, if we can get rid of /usr/lib/ipkg/info, I would be a happy man.. .:) |
19:50.13 | *** part/#oe ade|desk (~adavey@213-232-83-137.dsl.prodigynet.co.uk) |
19:50.14 | VoodooZ_Work | yipee! |
19:50.31 | VoodooZ_Work | 30 minues build is great! |
19:50.47 | Pigi | CosmicPenguin, what's the problem with /usr/lib/ipkg/info ? |
19:51.08 | CosmicPenguin | Pigi: heh - I don't need it on a read-only filesystem |
19:51.35 | Pigi | CosmicPenguin, so you would it to be relocable ? |
19:51.36 | CosmicPenguin | Pigi: except for the posinst - kergoth and I have discussed it would be nice to have the option to install a minimal set of files |
19:52.20 | Pigi | but the contents of info dir is vital for ipkg. |
19:52.23 | kergoth | Pigi: in a static environment, there's no need for anything but postinst. postinst for first boot setup for teh ones that didnt run at image creation time. |
19:52.26 | kergoth | how so? |
19:52.42 | CosmicPenguin | The question is, how much of the info dir is vital for 'ipkg configure' ? |
19:52.44 | kergoth | status seems to have everything you'd need for a static environment |
19:53.31 | *** join/#oe jonathang (~jon@cable-62-205-87-57.upc.chello.be) |
19:54.03 | Pigi | kergoth, that could also be, if nothing should be removed or added. But if you plan some mods, a lot of check are done with the contents of info files. |
19:54.21 | Pigi | And btw also a lot of check for what is installed depends on that files. |
19:54.43 | CosmicPenguin | Pigi: the point is that its a read only filesystem - nothing will be added ore removed |
19:55.04 | *** join/#oe Tygger-Bob (~myname@64.26.144.98) |
19:55.20 | Pigi | It could also be possible to do some tests. |
19:55.31 | kergoth | CosmicPenguin: heh, you know, if we screw with union / |
19:55.40 | kergoth | CosmicPenguin: the ipkg status files will overlap between dests |
19:56.24 | Pigi | What kind of options from ipkg are then needed or used in a readonly filesystem ? |
19:56.38 | kergoth | course, if / isnt actually an ipkg dest, but only the explicit compnents of hte union are, then ig uessi t isnt a concern |
19:57.22 | kergoth | s/ig uessi t/i guess it/ |
20:01.52 | CosmicPenguin | damn - sablevm-classpath needs all the graphical stuff |
20:01.53 | CosmicPenguin | suck |
20:02.58 | CosmicPenguin | kergoth: yeah, it would be ideal if they didn't try to write status on a read only filesystem, but unionfs works in a pinch |
20:03.04 | CosmicPenguin | And is probably a cleaner solution in the end |
20:04.36 | Pigi | kergoth, how is it possible to know what's the name of a package in oe ? |
20:05.06 | Pigi | I mean, I would like to build alsa, but if I do bitbake alsa I get a !nothing provide alsa" |
20:05.29 | Pigi | but the alsa di is there and contains some bb file |
20:05.46 | Pigi | s/di/dir |
20:05.59 | kergoth | Pigi: ? |
20:06.08 | kergoth | the directory names are irrelevent. |
20:06.14 | kergoth | read the bitbake users manual. |
20:09.24 | CosmicPenguin | Holy crap - a java benchmark who's source is 72MB |
20:09.33 | kergoth | eep |
20:22.18 | Laibsch | Pigi: I think your build system is incorrectly configured. Have set PATH and especially BBPATH as specified in http://www.openembedded.org/cgi-bin/moin.cgi/GettingStarted |
20:22.45 | Laibsch | Pigi: I meant: Have you set PATH and especially BBPATH as specified in http://www.openembedded.org/cgi-bin/moin.cgi/GettingStarted ? |
20:23.07 | Pigi | Laibsch, yes, Absolutely. I have already build some 10/15 images. |
20:23.48 | Pigi | Laibsch, why that question ? Did you managed to build alsa ? |
20:24.15 | Laibsch | No, but I once had that problem of nothing provides $BLA |
20:24.32 | kergoth | you clearly havent read the documentation. |
20:24.45 | Pigi | eheheh. Thx for your effort, but as kergoth said I should have do read the whole manual. |
20:24.45 | kergoth | it tells you exactly what doing a 'bitbake alsa' means |
20:24.49 | kergoth | and why it says it isnt provided |
20:24.59 | Pigi | yes. I know now. :) |
20:25.00 | kergoth | bitbake takes _providers_ on the commandline |
20:25.04 | kergoth | PROVIDES. |
20:25.14 | kergoth | by default a package PROVDIES its packagename, packagename-version, and packagename-version-revision. |
20:25.20 | kergoth | there is no .bb with a packagename of alsa. |
20:25.27 | kergoth | hence, there's no provider of 'alsa' |
20:25.29 | kergoth | simple enough. |
20:25.39 | Laibsch | That is what I figured by looking into alsa dir. ;-) |
20:25.40 | Pigi | sure, and now I also know what bitbake -s is for :) |
20:25.54 | Laibsch | Wow, learning time! |
20:26.22 | Crofton | bb needs a spell checker ... |
20:26.22 | Pigi | If you want to try it, be sure to have at least a 2099 linus of history in konsole :) |
20:26.29 | Pigi | s/linus/lines |
20:26.38 | kergoth | | tee log is your friend |
20:26.38 | kergoth | :) |
20:26.38 | Laibsch | I have NO intention of buidling alsa. |
20:27.38 | CosmicPenguin | ahh - come on. it will be fun... |
20:27.43 | Pigi | kergoth, yes, also :) |
20:28.11 | Pigi | CosmicPenguin, it's some day now I'm trying to build it. It's kinda fun :) |
20:28.21 | Crofton | When I change BOOSTRAP_EXTRA_RDEPENDS, do I need to clean task-bootstrap before rebuiling bootstrap-image |
20:28.31 | Crofton | to inculde the new package? |
20:28.40 | Pigi | Especially from when the alsa guys changed the configure :) |
20:28.55 | CosmicPenguin | The alsa guys are aways changing the configure |
20:29.02 | CosmicPenguin | thats sort of like their trademark |
20:29.02 | Crofton | yes |
20:29.08 | Crofton | :) |
20:29.26 | *** join/#oe kergoth (~kergoth@proxysb02.ext.ti.com) |
20:29.31 | Pigi | Yes, but in 1.0.8 every value derive from test.c compilation and execution :) |
20:29.44 | Pigi | so if you are in a cross environment..... |
20:29.55 | CosmicPenguin | Trust me, I know... |
20:30.23 | CosmicPenguin | In fact, I compile on a x86_64, so I get the pleasure of dealing with both -native and cross compiled autotools |
20:30.46 | Pigi | eheeh. Btw there is some chance from 0.9.6 that jamey already ported in old 0.7.2 time |
20:31.13 | *** join/#oe kergoth_ (~kergoth@li11-226.members.linode.com) |
20:31.51 | *** join/#oe Zero_Chaos (zero@pa-robinson1b-106.pit.adelphia.net) |
20:47.00 | *** join/#oe Crofton (~balister@mprg5213.mprg.ee.vt.edu) |
20:48.41 | *** join/#oe micropal (ident@139.80-202-82.nextgentel.com) |
20:49.25 | CIA-8 | 03kergoth 07 * r1.3118 10openembedded/packages/ipkg/ (ipkg_0.99.144.bb ipkg-0.99.144/paths.patch): Fix the recent ipkg-native paths fix. One of my changes to ipkg.h was inadvertantly lost. |
20:50.29 | kergoth | VoodooZ_Work: thats what you need. |
20:50.48 | VoodooZ_Work | cool. |
20:50.53 | VoodooZ_Work | I'm resolving |
20:52.57 | kergoth | Crofton: about to change a few things in openomap |
20:53.05 | Crofton | ok |
20:53.05 | kergoth | Crofton: adding udev to bootstrap |
20:53.13 | kergoth | Crofton: and setting SERIAL_CONSOLE by default |
20:53.17 | Crofton | ok |
20:53.17 | kergoth | that fine with you? |
20:53.41 | Crofton | I am using linux instead of uclinux |
20:54.23 | Crofton | an dI have modules in omap5912.osk |
20:54.34 | Crofton | rest of my stuff is in local.conf |
20:54.52 | Crofton | I put the use_vt and serial console in the machine file |
20:55.04 | kergoth | ahh right, that works |
20:55.27 | kergoth | i should probably add some more omap machines |
20:55.34 | kergoth | omap1710h3, omap2420h4 |
20:55.41 | Crofton | I am building image with gdb-server noew then I will boot, I'm not sure if I have the serial console right |
20:55.50 | Crofton | We only have 5912 :( |
20:55.54 | kergoth | SERIAL_CONSOLE = "115200 ttyS0" would be fine |
20:56.02 | kergoth | assuming you arent using devfs (kernel doestn mount it by default) |
20:56.04 | kergoth | and 115200 baud rate |
20:56.12 | Crofton | baud rate is fine |
20:56.19 | Crofton | not using devfs |
20:56.29 | Crofton | I need to figure out if I should be :) |
20:56.35 | kergoth | no need |
20:56.38 | kergoth | devfs is deprecated in 2.6 |
20:56.41 | kergoth | udev is the right way |
20:56.42 | *** join/#oe ar_ (~ar@port-ip-213-211-233-162.reverse.mdcc-fun.de) |
20:56.53 | Crofton | heh |
20:57.05 | Crofton | I can only keep up with so much stuff at my age |
20:57.18 | Crofton | do you want my machine diff? |
20:57.41 | kergoth | do you have a version that works with the release kernel? (per the modules you depend on, etc) |
20:57.48 | Crofton | no |
20:57.56 | kergoth | make one, and i'll put it in upstream oe |
20:58.20 | Crofton | what is the best way to get it to you |
20:58.31 | kergoth | email. oe mailing list, or email me directly |
20:58.39 | kergoth | ml is usually best, since if one of us is busy, someonke else can do it |
20:58.48 | kergoth | oe@handhelds.org |
20:58.54 | Crofton | my diff may not work :) |
20:59.00 | Crofton | yeah I signed up last night |
20:59.23 | Crofton | SERIAL_CONSOLE = "ttyS0 115200,8n1" |
20:59.30 | Crofton | does that have a chance of working? |
20:59.37 | kergoth | no. |
20:59.41 | kergoth | SERIAL_CONSOLE is for sysvinit |
20:59.42 | Crofton | heh |
20:59.44 | kergoth | getty commandline |
21:00.05 | kergoth | you want: |
21:00.06 | *** join/#oe Anst (~anst@198.36.32.9) |
21:00.13 | kergoth | SERIAL_CONSOLE = "115200 ttyS0" |
21:00.38 | kergoth | and if your kernel obeys it: |
21:00.43 | woglinde | re |
21:00.50 | kergoth | there s avariable for the console on the kernel commandline as well |
21:00.52 | kergoth | i forget what it is |
21:01.01 | Crofton | yeah I have that in uboot |
21:01.07 | Crofton | kernel messages come out |
21:01.11 | kergoth | cool |
21:01.22 | kergoth | then just set SERIAL_CONSOLE the way i recommended, and thatll work fine |
21:01.29 | kergoth | i should add u-boot to oe |
21:01.35 | woglinde | ~seen andersee |
21:01.37 | ibot | andersee is currently on #uclibc (1h 58m 47s) #oe (1h 58m 47s) #elinux (1h 58m 47s). Has said a total of 49 messages. Is idling for 22m 1s |
21:01.43 | woglinde | kergith yes |
21:01.48 | Crofton | That would be nice so you have mkimage |
21:01.57 | kergoth | yeah |
21:02.05 | Crofton | I need to figure out how to get it to make kernel images for flashing also |
21:02.05 | kergoth | the u-boot build could stage the mkimage |
21:02.11 | Crofton | yes |
21:02.25 | andersee | woglinde: g'day |
21:03.29 | Crofton | I am going to try and boot on new rootfs soon |
21:03.33 | [g2] | andersee, hey! |
21:03.38 | Crofton | then I will send 5912 diffs |
21:03.58 | andersee | [g2]: hey |
21:03.59 | Crofton | in openomap.conf I have targetOS=linux |
21:04.13 | andersee | [g2]: hows the nslu2 stuff these days? |
21:04.19 | Crofton | serious stuff is on 5912.conf |
21:04.21 | [g2] | Rock'n ! |
21:04.55 | [g2] | OpenSlug (the 2.6.9 based generic server appliance is about to go Beta) |
21:05.04 | andersee | [g2]: cool |
21:05.16 | [g2] | a couple guys are running uClibc on it |
21:05.26 | andersee | [g2]: any closed source bits needed? |
21:05.50 | [g2] | just a pull from Intel if you want to used the onboard ixp NPE |
21:06.10 | andersee | whats an NPE? |
21:06.13 | kergoth | Crofton: cool |
21:06.16 | kergoth | andersee: network processor |
21:06.24 | andersee | i.e. ethernet? |
21:06.26 | kergoth | yep |
21:06.35 | [g2] | however, we resolved the issue with usbnet and you can boot a usb2.0-network adapter out-of-the-box (build) too and be Intel Free |
21:06.49 | [g2] | it's an IXP420 processor |
21:06.53 | andersee | cept for the chip of course |
21:06.55 | andersee | :-) |
21:07.16 | andersee | [g2]: remind me again, how fast is it? |
21:07.18 | [g2] | If you run the USB2.0 adapter you just don't use the coprocessor |
21:07.19 | CIA-8 | 03kergoth 07 * r1.3119 10openembedded/conf/ (3 files in 2 dirs): Add SERIAL_CONSOLE to the OMAP MACHINEs, and include udev in bootstrap for openomap. |
21:07.27 | [g2] | 266 Xscale BE |
21:07.46 | [g2] | 32 MB ram 8MB flash to usb2.0 ports |
21:07.55 | [g2] | We're replacing the Bootloader too |
21:08.06 | andersee | nice, with what? |
21:08.20 | [g2] | the coolest bootloader in the world! |
21:08.23 | [g2] | APEX |
21:08.34 | woglinde | andersee hm which board is arm bigendian? |
21:08.42 | [g2] | beewoole wrote it |
21:08.52 | [g2] | beewoolie that is |
21:08.56 | andersee | woglinde: Linksys nslu2 network attached storage box |
21:09.14 | andersee | [g2]: hmm, url? |
21:09.19 | [g2] | it's an ARM specific bootloader *really* small < 16K right now |
21:09.49 | andersee | [g2]: features match i.e. blob, uboot, etc? |
21:09.50 | [g2] | APEX boots both the 2.4 and 2.6 kernels from flash |
21:09.52 | kergoth | andersee: to translate, beewoolie is marc singer. |
21:09.54 | kergoth | heh |
21:10.02 | CIA-8 | 03kergoth 07 * r1.3120 10openembedded/conf/machine/ (omap1510inn.conf omap2420h4.conf omap1710h3.conf): Add MACHINE files for OMAP 1510 Innovator, OMAP 2420 H4, and OMAP 1710 H3. |
21:10.06 | VoodooZ_Work | [g2]: where do I sign? |
21:10.35 | *** part/#oe Zero_Chaos (zero@pa-robinson1b-106.pit.adelphia.net) |
21:10.39 | [g2] | VoodooZ_Work, sign ? |
21:10.52 | VoodooZ_Work | kidding. I meant to jump on the APEX bandwagon. |
21:10.57 | [g2] | nod. |
21:11.19 | VoodooZ_Work | Looks mighty good. the smaller and faster the better :) |
21:11.27 | kergoth | Crofton: heh, i should probably consolodate the common omap machine bits into an omap.conf. same stuff is in too many places |
21:11.31 | [g2] | beewoolie is looking at putting ext2 support in right now |
21:11.40 | VoodooZ_Work | It sitll takes a minimum of 256K flash though right? |
21:12.01 | [g2] | it's 16K right now so it fits in 1 (128K) eraseblock |
21:12.26 | VoodooZ_Work | oops. sorry 128k. |
21:12.38 | CIA-8 | 03kergoth 07 * r1.3121 10openembedded/conf/machine/omap2420h4.conf: Add armv6te to the list of ipkg architectures for the OMAP 2420 H4 machine. |
21:12.45 | [g2] | andersee, http://wiki.buici.com/twiki/bin/view/Main/ApexBootloader#APEX_Boot_Loader |
21:13.05 | kergoth | hrm, need to teach oe's toolchain stuff to support building for hard vfp, for arm |
21:13.17 | kergoth | probably need some patches from the codesourcery cvs branch |
21:13.56 | [g2] | andersee we'll probably be setting up some build / executions tests for OpenSlug |
21:14.09 | [g2] | I'd like to include uClibc if possible |
21:14.24 | [g2] | I can talk to you about it in uclibc |
21:15.05 | andersee | [g2]: k |
21:16.07 | andersee | kergoth: btw, the other day I was missing an installed libgcc_s and libstd++. I had to pluck them out of the gcc build tree and install manually. |
21:16.12 | kergoth | andersee: yep |
21:16.15 | kergoth | andersee: fixed that today |
21:16.25 | andersee | kergoth: oh, good |
21:16.39 | kergoth | andersee: its due to those coming out of our gcc-cross build, not gcc. so ${libdir} is out in CROSS_DIR< not the target lib path. so it wasnt packaged. |
21:16.52 | kergoth | fixed by backing up the target libdir before inheriting cross, and using that for the packaging |
21:16.55 | kergoth | heh |
21:17.26 | kergoth | seems that way, i just did a bootstrap image and booted it on my omap board successfully |
21:17.34 | kergoth | and g2 and voodooz and using fresh builds on the nslu |
21:18.06 | VoodooZ_Work | works great now. |
21:19.10 | andersee | kergoth: oh, another thing. I had to manually unpack all my ipks, install my own /dev, etc.. I presume there an automagical mechanism to do that to produce a usable rootfs? |
21:19.21 | kergoth | andersee: the -image targets in packages/meta/ |
21:19.31 | *** join/#oe edbarrett (~edbarrett@adsl-66-72-62-238.dsl.klmzmi.ameritech.net) |
21:19.34 | kergoth | bootstrap-image, gpe-image, opie-image, what have you |
21:19.58 | [g2] | openslug-image :) |
21:20.24 | andersee | kergoth: I ran bitbake opie-image, but I think it wanted me to have a machine or distro set |
21:20.37 | kergoth | those arent required |
21:20.40 | kergoth | i rarely set them, myself |
21:20.57 | kergoth | you shouldve ended up with image files in tmp/deploy/images/ |
21:21.03 | andersee | hmm |
21:21.23 | kergoth | IMAGE_FSTYPES is a space seperated list of filesystem types, you can specify them that way |
21:21.29 | kergoth | i generally do "jffs2 tar" myself |
21:21.32 | kergoth | tar for nfs |
21:24.51 | *** join/#oe _CoreDump|home (~mhentges@hentges.net) |
21:25.45 | CosmicPenguin | damn sablevm-crosspaths |
21:25.58 | CosmicPenguin | Nobody else on the planet uses Xtst, I swear... |
21:26.28 | kergoth | heh |
21:26.33 | kergoth | whats Xtst? |
21:27.06 | CosmicPenguin | libXtst |
21:27.19 | CosmicPenguin | X Window System event recording and testing |
21:27.19 | CosmicPenguin | extension library development files |
21:28.20 | *** join/#oe drw (~drw@c-67-172-218-120.client.comcast.net) |
21:33.09 | [g2] | andersee, btw kergoth put hotplug-ng in the OE repo a week or so ago and OpenSlug is running with hotplug-ng |
21:33.48 | VoodooZ_Work | we're cutting edge baby! :) |
21:34.03 | CosmicPenguin | don't fall off |
21:34.37 | [g2] | CosmicPenguin, we did that yesterday :) |
21:34.43 | kergoth | hehe |
21:34.44 | [g2] | fall off that is :) |
21:34.48 | VoodooZ_Work | hehe |
21:35.15 | VoodooZ_Work | I'm a lot less tense now that it works too. |
21:35.53 | woglinde | voodoo go for it |
21:36.22 | VoodooZ_Work | heheh as if! I can't even compile a single .c file without asking kergoth for help! :) |
21:37.44 | kergoth | hm, need a sane default tslib.sh and ts.conf for some of these omap boards |
21:38.13 | woglinde | voodoo oehm??? |
21:38.27 | woglinde | kergoth try and error? |
21:38.36 | kergoth | ? |
21:39.28 | kergoth | i maintain tslib, and these configuration files are so itll work using the ts driver on this board that i wrote. dont exactly need trial & error to get that right. |
21:39.32 | kergoth | heh |
21:40.13 | [g2] | kergoth, do_install overrides the default mechanism correct? Can I do_install_postpend or prepend to add just a little bit after or before the default install method ? |
21:40.59 | [g2] | or maybe it's do_install_append |
21:41.00 | kergoth | [g2]: yes. |
21:41.08 | kergoth | append or prepend |
21:41.13 | CIA-8 | 03kergoth 07 * r1.3122 10openembedded/packages/linux-input/linux-input_cvs.bb: linux-input shouldnt be inheriting module-base. There's no need for it, and it was breaking the build. |
21:41.14 | kergoth | or just do something like: |
21:41.15 | kergoth | do_install () { |
21:41.16 | [g2] | excellent. Cool |
21:41.19 | kergoth | <PROTECTED> |
21:41.22 | kergoth | <PROTECTED> |
21:41.24 | kergoth | <PROTECTED> |
21:41.25 | kergoth | } |
21:41.28 | CosmicPenguin | reenoo_: any reason why we're not installing the classpath stuff into staging? |
21:41.35 | CosmicPenguin | s/classpath/sablevm-classpath/ |
21:41.37 | [g2] | kergoth, THX again |
21:41.41 | kergoth | np |
21:41.52 | kergoth | pretty sure that one was in the bitbake users manual. |
21:41.55 | kergoth | reread it if you havent recently |
21:42.17 | Crofton | For kernels, what are the valid KERNEL_IMAGETYPE settings? |
21:43.02 | Crofton | I need to make the vnlinux file |
21:43.20 | Crofton | rather than the image |
21:43.29 | CIA-8 | 03kergoth 07 * r1.3123 10openembedded/ (2 files in 2 dirs): Fix sysvinit to obey the base_* target path variables, and kill the old sysvinit 2.85. |
21:44.06 | kergoth | Crofton: read kernel.bbclass |
21:44.12 | Crofton | thanks |
21:44.22 | kergoth | you'll probably want to do: |
21:44.27 | kergoth | KERNEL_IMAGETYPE = "vmlinux" |
21:44.33 | kergoth | KERNEL_OUTPUT = "vmlinux" |
21:44.47 | kergoth | (default kERNEL_OUTPUT is arch/${ARCH}/boot/${KERNEL_IMAGETYPE}, but vmlinux doesnt go there) |
21:46.07 | *** join/#oe treke (~ggilbert@tinman.treke.net) |
21:46.31 | reenoo_ | CosmicPenguin: for use with jikes-native? |
21:46.59 | reenoo_ | CosmicPenguin: staging doesn't sound right for that |
21:47.41 | CosmicPenguin | I just used jikes + the sablevm-classpath to compile jolden |
21:47.55 | CosmicPenguin | I'm not very clear on the concept, but wouldn't that be what you would do for your target? |
21:49.04 | Crofton | heh could leave kimagetyoe blank :) |
21:49.40 | Crofton | linux-omap overides do deploy, so I can process image for dl'ing to omap there I think |
21:50.39 | Crofton | I was hoping the modules would work with the kernel I have loaded, but it was built with gcc-3.3 |
21:50.50 | kergoth | ahh right |
21:51.14 | Crofton | I have the serial console now btw |
21:51.54 | reenoo_ | CosmicPenguin: well.. staging is for target libs etc. if I'm not mistaken. what you're using is jikes compiled for the host. I guess, you'll also need a (sablevm-)classpath-native. |
21:51.58 | kergoth | good |
21:52.16 | [g2] | kergoth the ipkg fix in od-devel worked for me tool |
21:52.17 | reenoo_ | CosmicPenguin: otherwise java code using native libs won't work |
21:52.29 | reenoo_ | brb |
21:52.38 | kergoth | [g2]: good |
21:52.56 | CosmicPenguin | hmmm... I guess I don't understand |
21:53.00 | [g2] | later tonight after that's pulled into nslu2-linux I'll mark that done punchlist in our wiki |
21:53.11 | [g2] | THX for making that happen today |
21:53.18 | kergoth | np |
21:53.50 | [g2] | those were the only outstanding Critical technical issues. |
21:53.56 | CosmicPenguin | I thought that java bytecode was platform independent |
21:54.06 | CosmicPenguin | I guess we need a jikes-cross then |
21:54.30 | [g2] | The only other Critical issue was 3 more Alpha users and there were three ready yesterday but the build repo wasn't in shape |
21:54.42 | [g2] | looks like will get them this weekend |
21:59.07 | reenoo_ | CosmicPenguin: this is not about java bytecode |
21:59.53 | CosmicPenguin | reenoo_: I thought thats what jikes produced |
22:00.02 | reenoo_ | CosmicPenguin: correct |
22:01.02 | CosmicPenguin | reenoo_: and jikes uses the classpath .jars or whatever to produce the bytecode, right? |
22:01.34 | reenoo_ | right. the problem is that sablevm-classpath contains both java bytecode and native libs |
22:01.39 | CosmicPenguin | I see |
22:01.48 | CosmicPenguin | What are the native libs for? |
22:01.59 | CosmicPenguin | And native to what? |
22:02.07 | Pigi | what's that: autoheader: warning: missing template: CONFIG_HAVE_OLD_REQUEST_MODULE |
22:02.11 | Pigi | ? |
22:02.40 | reenoo_ | CosmicPenguin: native as in C code |
22:02.52 | CosmicPenguin | reenoo_: for the target platform? |
22:03.00 | CosmicPenguin | The .so's, right? |
22:03.08 | reenoo_ | CosmicPenguin: when compiled for the target, then yes |
22:03.12 | reenoo_ | CosmicPenguin: right |
22:03.19 | CosmicPenguin | Does jikes use those? |
22:03.27 | reenoo_ | I don't know |
22:03.56 | reenoo_ | I can't think of any particular reason why it would want to do that though |
22:06.06 | reenoo_ | it shouldn't need to actually |
22:12.52 | Pigi | kergoth, in log.do_configure I have a lot of "missing template:". What can be the cause ? |
22:13.20 | Pigi | all the missing template are for the #define needed for alsa-driver |
22:13.34 | Pigi | So it would never been able to compile :( |
22:13.35 | Crofton | hmm you define KERNEL_OUTPUT in the kernel.bbclass, but when you go to install the kernel, you don't use it ... |
22:13.52 | kergoth | Pigi: that indicates there are missing templates. |
22:14.09 | kergoth | Pigi: in other words, there are calls to AC_DEFINE() that dont pass a 'description' argument |
22:14.22 | kergoth | Pigi: either add the descriptions, or call AH_TEMPLATE() on each of them instead |
22:14.47 | *** join/#oe zwi (~zwi@69.25.212.5) |
22:14.52 | Pigi | good. A nice patch to that nasty configure ( from alsa again ) :) |
22:15.14 | Crofton | you guys aer acaring me ... |
22:15.36 | kergoth | autotools? scare people? *gasp* |
22:16.09 | Pigi | the strange thing, kergoth, is that the same configure works in my old crosscompile ambient.... grrrr |
22:16.24 | kergoth | Pigi: those changes are required by recent versions of autotools. |
22:16.30 | kergoth | Pigi: sounds like you were doing builds with old versions or something. |
22:17.17 | Pigi | kergoth, yes, I'm almost sure of that. autoheader -V says 2.59 |
22:17.30 | kergoth | auto* break compatibility fairly often. |
22:17.39 | Pigi | eheheh |
22:19.03 | CosmicPenguin | reenoo_: in libffi, in the install_append(), you should have ${TARGET_SYS} instead of arm-linux... :) |
22:20.06 | reenoo_ | CosmicPenguin: heh. indeed |
22:21.51 | reenoo_ | CosmicPenguin: I guess I missed that after fighting the gcc build system for hours... |
22:22.05 | CosmicPenguin | reenoo_: its understandable... :) |
22:22.22 | CosmicPenguin | reenoo_: getting it to work is the first step - plus, 85% of the people using OE are doing ARM anyway |
22:23.07 | CosmicPenguin | Heck, QT/E has been in there since day one, and we're only now building it with a x86 |
22:23.19 | CosmicPenguin | Or at least, only now finding the bugs |
22:23.35 | treke | god forbid you try opie on non arm |
22:23.49 | CosmicPenguin | treke: it works well |
22:24.01 | CosmicPenguin | and its rill purdy too |
22:24.01 | treke | not if you trust the config scripts :) |
22:24.38 | treke | building for ppc means settign the config to x86 :p |
22:27.26 | treke | the one time I try flashing a new image is when it doesnt boot :p |
22:30.16 | CIA-8 | 03reenoo 07 * r1.3111.1.1 10openembedded/packages/libffi/libffi_2.0+gcc3.4.1.bb: libffi: fix a few hardcoded paths. |
22:30.26 | reenoo_ | CosmicPenguin: thanks for spotting |
22:30.30 | CosmicPenguin | reenoo_: no problem |
22:32.36 | kergoth | whee, electricity was off due to being behind on my bill. |
22:32.37 | kergoth | again. |
22:32.41 | kergoth | man, i suck at bills |
22:32.53 | kergoth | apparently they'll hook it back up tonight |
22:33.03 | woglinde | kergoth dont have a grilfriend for? |
22:33.22 | kergoth | hehe. she doesnt do my finances |
22:33.30 | woglinde | hehe |
22:33.52 | kergoth | guess i could ask her to just yell at me every month |
22:34.10 | woglinde | in german the powercompanies cant switch off the power so easily |
22:34.26 | woglinde | because of human rights and so on |
22:34.38 | kergoth | in minnesota they cant during the winter, cause it could kill your ass |
22:34.41 | kergoth | here they dont seem to care |
22:34.43 | kergoth | hehe |
22:34.46 | woglinde | dont have a pda for that? |
22:34.55 | kergoth | i never carry my pdas |
22:34.57 | kergoth | heh |
22:34.58 | kergoth | :\ |
22:35.00 | kergoth | guess i should |
22:35.02 | kergoth | but even if i did |
22:35.14 | kergoth | neitehr opie or gpe support recurring todos whose deadlines show up in the calander and have alarms |
22:35.17 | woglinde | hm that remeber me someone |
22:35.18 | kergoth | so neither meet my needs |
22:35.42 | woglinde | try palmos? *hide* |
22:35.47 | treke | buy a palm |
22:35.58 | kergoth | i've been seriously consdiering a tungsten |
22:36.23 | treke | or just pay bills the same time every month |
22:36.33 | kergoth | pfft, you and your crazy ideas |
22:37.02 | treke | woglinde: in this part of the world cold isn't nearly as big of a problem |
22:37.13 | treke | so turning off the power is annoying, but not going to kill you |
22:37.22 | kergoth | yeah |
22:37.28 | treke | and it's a damn good way to get people to pay up |
22:37.35 | kergoth | that it is |
22:37.38 | kergoth | i hope my food is still good |
22:37.46 | kergoth | heh |
22:37.47 | treke | should be fine |
22:37.53 | treke | just don't open the refridgerator |
22:37.58 | kergoth | assuming they reconnect it tonight anyway |
22:38.01 | woglinde | hm consider solarpower |
22:38.03 | kergoth | pizza it is. |
22:40.41 | woglinde | hm fuck, why zlib is broken with uclibc |
22:40.49 | woglinde | under my powerpc |
22:41.40 | CosmicPenguin | reenoo_: just ran one of the jolden benchmarks that I compiled with jikes on my target, and it worked fine |
22:41.51 | CosmicPenguin | So it seems to me we can get away with copying the .jars into staging |
22:42.43 | CosmicPenguin | Though technically, they are not platform specific, I don't see any problem copying them into the target libs, because who is going to fire up oE just to build native java apps? |
22:43.51 | Crofton | later all |
22:44.29 | Crofton | kergoth: I'll send diffs when I see the thing boot successfully |
22:44.34 | kergoth | Crofton: k, thanks |
22:49.14 | Pigi | nite all |
22:52.11 | reenoo_ | CosmicPenguin: we might want to run a native VM to run ant for example. we'd need a (sablevm-)classpath-native for that anyway. |
22:52.21 | CosmicPenguin | reenoo_: nod |
22:52.36 | CosmicPenguin | reenoo_: so how do the .so's work? Does sablevm pull them in when it needs them? |
22:53.16 | *** join/#oe jh_ (~jason@69.144.85.206) |
22:53.43 | reenoo_ | you can interface to native code from within java through JNI (Java Native Interface) |
22:53.56 | CosmicPenguin | I see |
22:56.02 | CosmicPenguin | Well, it all seems to be working nicely |
22:56.38 | *** join/#oe Timelord0 (~TL@64-147-65-108.access-3.intertex.net) |
22:59.58 | *** join/#oe jmau (~jmau@pD9E5D3DF.dip.t-dialin.net) |
23:00.04 | woglinde | hrms |
23:00.10 | woglinde | static compile foes |
23:00.15 | woglinde | goes |
23:07.34 | woglinde | steroekit try buildroot from the same site where busybox came from |
23:08.01 | woglinde | args |
23:08.04 | woglinde | wrong channel |
23:12.47 | CIA-8 | 03kergoth 07 * r1.3123.1.1 10openembedded/packages/libiconv/libiconv_1.9.2.bb: uclibc build: Fix libiconv to build its static lib and stage it -- for some reason popt requires a static version for something it builds. |
23:16.12 | koen | 'night all |
23:18.56 | CosmicPenguin | kergoth: any opinions on where platform independent stuff like Java bytecode should live in staging? |
23:19.25 | kergoth | in general, in linux systems, platform independent data files belong in ${datadir} aka /usr/share |
23:19.31 | kergoth | iirc we have a path like that in staging today |
23:19.39 | kergoth | if not, we should :) |
23:20.17 | CosmicPenguin | I see an info, man, and shlib |
23:20.31 | kergoth | thast right.. |
23:20.39 | kergoth | i think STAGING_DATADIR is arch specific right now |
23:20.41 | CosmicPenguin | of course, its quite possible I haven't build anything that would be in stare |
23:20.45 | kergoth | only for .m4 files |
23:20.49 | kergoth | due to cross things |
23:25.22 | CosmicPenguin | kergoth: so you think ${STAGING_DIR}/share would be a good place then? |
23:25.49 | kergoth | thatd work. dunno what you'd call the variable, since we already have a datadir |
23:25.50 | kergoth | heh |
23:27.12 | CosmicPenguin | mabye $STAGING_GLOBAL_DATADIR |
23:27.34 | *** join/#oe Ken|JLimeX (~Ken@h237n2fls34o1010.telia.com) |
23:46.53 | *** join/#oe qfh (~qfh@ppp-82-135-3-138.mnet-online.de) |
23:54.48 | woglinde | n8 |