**** BEGIN LOGGING AT Thu Mar 30 10:59:56 2006 Mar 30 11:00:34 zecke, btw there is oss emu code in linux/sound/core/oss. Mar 30 11:01:06 RDEPENDS_hostap-modules-cs = "hostap-modules (${PV}) hostap-conf" Mar 30 11:01:13 it should be ok for ipkg? Mar 30 11:01:15 do13, This looks like the pin is not signalling the codec. I wonder if it needs to be enabled anywhere ? Mar 30 11:02:22 hrw|work: (>= ${PV}) iirc Mar 30 11:02:44 zecke: pcm_oss.c? Mar 30 11:03:04 do13, I'm checking datasheet.... Mar 30 11:03:39 koen: no Mar 30 11:03:49 lrg: thx. I'll try to read the GPIO status Mar 30 11:03:52 koen: hostap-modules-cs 0.4.7 need hostap-modules 0.4.7 Mar 30 11:04:04 koen: I want to force same ver Mar 30 11:11:30 do13, I'm being called away now. I think your idea is correct to check the level of the pin. If the pins is signalling then it's a bug with the driver. be back in 1 hour. Mar 30 11:12:07 lrg: ok Mar 30 11:16:42 man x is huge. i wish i could just use kdrive Mar 30 11:17:35 TheMasterMind1: "x".islower() -> True Mar 30 11:17:56 it might be huge, but it is not capital Mar 30 11:18:15 TheMasterMind1: you of course can use Xfbdev :) Mar 30 11:19:12 how can one check who is responsible for a certain package, lets say for instance opie-bluetoothmanager Mar 30 11:19:41 zecke: we could organize something to get E3s to interested people Mar 30 11:19:56 HuntrCkr: MAINTAINER field Mar 30 11:20:03 HuntrCkr: grep MAINTAINER opie-bluetoothmanager*.bb Mar 30 11:20:32 bitbake -i ... parse peek opie-bluetoothmanager MAINTAINER Mar 30 11:20:53 koen: dont start me on new toys :-) Mar 30 11:20:57 zecke: that might be poluted by distros Mar 30 11:20:57 koen: my local argos sells them for 30 quid Mar 30 11:21:05 mithro: are you familiar with pickle and unpickle? Mar 30 11:21:12 ade|desk: you can only get them in the UK Mar 30 11:21:23 ade|desk: how much euro are 30 quid? Mar 30 11:21:33 ~change 30 gbp to eur Mar 30 11:21:38 30.00 Pound (GBP) makes 43.2025 Euro (EUR) (from http://www.xe.com/) Mar 30 11:21:39 ~change 30 gbp to eur Mar 30 11:21:42 lol Mar 30 11:21:55 http://wiki.earth.li/E3_Buying_one Mar 30 11:22:20 zecke: yes Mar 30 11:22:31 * XorA has 20 quid coupons for Argos :-) Mar 30 11:22:46 me too Mar 30 11:22:49 what do you want to know? Mar 30 11:23:05 mithro: can I dump more than one object? Mar 30 11:23:13 mithro: and in which order can I load them? Mar 30 11:23:39 zecke: just put them in a list and dump the list Mar 30 11:23:51 note you can't pickle a few things Mar 30 11:24:16 mithro: the documentation is not saying that :} Mar 30 11:24:25 mithro: I will put them in a tuple... Mar 30 11:24:36 03hrw 07org.oe.oz354x * r9487dee9... 10/packages/hostap/ (hostap-utils_0.4.0.bb hostap-utils.inc hostap-utils_0.4.7.bb): added hostap-utils 0.4.7 Mar 30 11:24:40 03hrw 07org.oe.oz354x * r24db71a5... 10/packages/hostap/ (8 files in 3 dirs): hostap-daemon: added 0.4.8 and unified patches for 0.4.4/0.4.8 Mar 30 11:24:43 tuple, list little difference :P Mar 30 11:24:44 03hrw 07org.oe.oz354x * rea6e3bbd... 10/packages/hostap/ (42 files in 9 dirs): Mar 30 11:24:44 hostap-modules: added 0.4.7 as default Mar 30 11:24:44 - hostap-modules-cs now RDEPEND on hostap-conf Mar 30 11:24:44 - its easier to update only hostap-conf when new card is added Mar 30 11:25:00 mithro: right Mar 30 11:25:26 zecke: what are you pickleing out of intrest? Mar 30 11:25:54 mithro: bitbake is pickling Mar 30 11:26:10 mithro: I pickle two more structures Mar 30 11:26:31 mithro: and mickeyl showed my assumption didn't hold true (the assumption I could pickle any amount of data) Mar 30 11:26:41 any amount of objects... Mar 30 11:26:49 any amount of top level objects Mar 30 11:27:07 any amount of not related top level objects... whatever *grin* Mar 30 11:27:16 ? Mar 30 11:27:21 crap, I've broken oe :/ Mar 30 11:28:16 mndctrl: fix it :) Mar 30 11:29:09 I'm trying ;) going with the 'monotone update -rjan2006prebreakage'... Mar 30 11:29:15 lets see if it makes a difference ;) Mar 30 11:29:27 mndctrl: ah you have broken your version of OE? Mar 30 11:29:41 mndctrl: using BitBake trunk by chanche and cache is not working for you? Mar 30 11:30:13 mithro: OT: Did you ever mix PyGame and WxWidgets? Mar 30 11:30:15 I guess... bitbake whatever fails dramatically... Mar 30 11:30:19 ;) Mar 30 11:30:32 mndctrl: use pastebin to post your error Mar 30 11:30:50 I will if the -rjan2006pre thing doesent help ;) Mar 30 11:30:59 lol Mar 30 11:31:35 While waiting for it I feel no guilt over surfing the web and dreaming the day away ;) Mar 30 11:31:49 crap it finished :/ hehe... Mar 30 11:32:35 hmmm, now I get: ERROR while parsing /path/to/every/.bbfie ? Mar 30 11:32:41 clean cache? Mar 30 11:32:54 mndctrl: too old OE + too new bitbake Mar 30 11:33:09 mndctrl: take the offer, or don't take it.. Mar 30 11:33:17 mndctrl: -rjan2006pre OE need old bitbake (r329 for example) Mar 30 11:33:24 zecke: btw why dont you put a version in the cache file? Mar 30 11:33:26 mndctrl: use pastebin instead of wrongly summarising errors Mar 30 11:33:55 mithro: the cache is broken regardles of the version :) Mar 30 11:33:57 zecke: hehe ;) okok... I'm sorry... but cleaning cache helps for now.. I'll use pastebin next time... Mar 30 11:34:20 mndctrl: reading short anouncements on bitbake-dev and oe mailinglist helps as well Mar 30 11:34:40 mithro: the only proper solution would be to go through the legals departments Mar 30 11:34:52 zecke: sounds like me getting confused ;) Mar 30 11:35:06 mithro: bitbake should have a way to spread news, let the people print that, hand it over to the legal department Mar 30 11:35:13 mithro: and they will read it to you :) Mar 30 11:36:22 mndctrl: do you use BitBake trunk? Mar 30 11:38:17 yeah finally this battery is running flat Mar 30 11:38:44 zecke: ?? Mar 30 11:38:51 mithro: ah you missed that Mar 30 11:39:10 mithro: I was at my grand parents until yesterday. Now I want to get some work done for a deadline tomorrow Mar 30 11:39:43 mithro: I drive to the 'office', just to crash one device, while the tool to reset it is at home, I go back to my car, the car isn't starting anymore Mar 30 11:39:58 mithro: now I'm waiting for the unit to turn off, so I can restart it Mar 30 11:40:28 - legal deparments? Mar 30 11:40:37 mithro: ah okay Mar 30 11:40:54 mithro: well, I wonder how to distribute 'NEWS' Mar 30 11:41:19 mithro: mailing to mailing lists is not enough, I do not expect any user to say thank you, or to recognize bitbake being ueberfast Mar 30 11:41:41 mithro: but I expect them to get some basic informations, I thought mailinglists could be enough - they aren't Mar 30 11:42:17 mithro: I'm considering writing an anouncement framework... Mar 30 11:42:32 mithro: bitbake would get news, and use OSD, printers, blinken lights etc Mar 30 11:42:45 mithro: to signalise that some information must be read... Mar 30 11:43:03 mithro: and then a little game is played to 'test' if the information has been read... Mar 30 11:43:07 *frustrated* Mar 30 11:44:37 mithro: well and I think the Legal Department is the only institution that is actually reading Mar 30 11:45:53 zecke: since I don't know what bitbake trunk is, I guess not... Mar 30 11:46:03 mndctrl: which version of bitbake do you use? Mar 30 11:47:06 zecke: 1.3.3 ? Mar 30 11:47:38 mndctrl: okay *nevermind* Mar 30 11:47:39 revision 422 or something... I just did a svn today... Mar 30 11:47:45 zecke: ok, hehe ;) Mar 30 11:47:50 mndctrl: well, not nevermind Mar 30 11:47:57 mndctrl: svn info | grep -i trunk Mar 30 11:48:23 URL: svn://svn.berlios.de/bitbake/trunk/bitbake Mar 30 11:48:45 thats the only output... Mar 30 11:49:10 mndctrl: https://lists.berlios.de/pipermail/bitbake-dev/2006-March/000212.html Mar 30 11:49:15 mndctrl: or how to waste my time :} Mar 30 11:50:40 aha, are there a way do downgrade my bitbake then? Mar 30 11:51:37 mndctrl: svn up -r418 Mar 30 11:51:58 do13: tnx ;) Mar 30 11:58:49 is there dry run functionality (-n) in the interactive shell (-i) Mar 30 11:59:39 no Mar 30 12:13:55 hehe, this whole day went while struggeling with getting oe and bitbake working properly together... :/ But I learned a bit more about the workings of the two ;) Mar 30 12:23:06 * chouimat is back. Mar 30 12:23:32 morning Mar 30 12:24:38 chouimat: morning ;) Mar 30 12:37:02 Ciao all Mar 30 12:37:44 morning all Mar 30 12:38:06 hey Richard Mar 30 12:38:06 morning Mar 30 12:38:33 hey Pigi Mar 30 12:38:52 hey koen. How things are going ? Mar 30 12:39:10 Pigi: pretty good. How's life in italy? Mar 30 12:40:06 Fine thanks. I'm having a really hard period at work, and that is good, and bad. But a part for not having more spare time, everyhing it's OK.... Mar 30 12:40:26 when stuff is put into staging and get gets stuffed in to staging/bin is it supposed to go into staging/foo_arch-linux/bin or can it just be left as is ? Mar 30 12:41:25 ade|desk: Its supposed to get into foo_arch-linux Mar 30 12:42:33 ah ok, for an autotooled native package whats the best way to push it into foo_arch-linux ? Mar 30 12:44:21 its only binaries wtf Mar 30 12:44:27 btw even Mar 30 12:45:06 ade|desk: Here, the binaries make it into i686-linux so it appears to work in at least some cases... Mar 30 12:46:10 RP: it has worked for most packages alas not for this one :( Mar 30 12:47:26 ade|desk: Sounds like its autotooling isn't quite right :-/ Mar 30 12:47:38 looks like bash, dbus-1 fail a little and put stuff in staging/etc , perl5 puts things in random places too Mar 30 12:47:49 RP: indeed Mar 30 12:48:16 We do need to try and track down the offenders and fix the packages... Mar 30 12:48:23 quilt is another Mar 30 12:48:45 RP: some sort of LD_PRELOAD sandbox before do_install() Mar 30 12:49:58 ade|desk: The main problem looks like /etc Mar 30 12:50:15 ~seen pb_? Mar 30 12:50:20 pb_ is currently on #gpe (1d 21h 20m 45s) #oe (1d 21h 20m 45s). Has said a total of 61 messages. Is idling for 4h 58m 36s, last said: 'bbl'. Mar 30 12:50:47 local, share are perl problems Mar 30 12:50:50 var is dbus Mar 30 12:51:19 cd .. Mar 30 12:51:21 lol Mar 30 12:51:51 ade|desk: Placing etc, var, info and man there is by design looking at native.bbclass Mar 30 12:52:08 "Architecture independent paths" Mar 30 12:52:48 does native.bbclass need to learn about a tidy staging ? Mar 30 12:54:29 OE has to learn about a tidy staging :) Mar 30 12:57:20 ade|desk: It could be argued the current behaviour is correct. I don't like it though... Mar 30 12:57:43 RP: yeah its correct, yet wrong ;) Mar 30 13:04:34 hi pigi Mar 30 13:05:43 can somebody give me a fast internet Mar 30 13:05:57 mithro: here you go.. Mar 30 13:05:58 * koen hands mithro some dark fiber Mar 30 13:06:20 * mndctrl hands mithro some light fiber instead ;) Mar 30 13:06:43 mithro: go to a shop and buy the internet, its so much easier :-) Mar 30 13:08:34 I guess you would need a storage-facility also... Mar 30 13:09:37 XorA: i'll sell you the complete internet dvd set Mar 30 13:10:21 TheMasterMind1: wicked, can I have the pr0n edition? Mar 30 13:10:38 sure Mar 30 13:10:48 i have a bridge i want to sell you too Mar 30 13:11:01 TheMasterMind1: nah, we sold ours to the yanks Mar 30 13:11:34 | fc-case.c:(.text+0x1e5): undefined reference to `__ctype_b' Mar 30 13:11:39 any idea what would cause that? Mar 30 13:12:12 TheMasterMind1: hacking with glibc internals Redhat stylee? Mar 30 13:12:57 i'm running gentoo.. its occuring on the fontconfig-native compile Mar 30 13:12:59 hi pb__ ! Mar 30 13:13:31 TheMasterMind1: I run on gentoo and havent seen that one, but I cannot currently check as my house is in powercut Mar 30 13:13:42 Pigi: hey, thanks for getting that bug taken care of Mar 30 13:13:53 TheMasterMind1, np. Mar 30 13:14:07 well i haven't updated glibc in a while (cuz its takes forever to compile) so maybe'll ill emerge and try again Mar 30 13:14:22 RP, http://dev.aimao.org/poky/ Mar 30 13:14:29 TheMasterMind1, fwiw I've released 0.99.160 ( not too much fixes tough ) Mar 30 13:15:05 pb__ do you remember when we discussed about the removal of a package before upgrading ? ( a.k.a. the infamous bug ? ) Mar 30 13:15:28 Pigi: yea i saw CIA report it earlier tonight. going to add .160 to OE because both .154 and .159 were having offline root issues Mar 30 13:17:13 Genesis: Neat :) Mar 30 13:17:47 you suggested me to use the "remove_data_files_and_list" function which is good, but do you think it could be usefull to remove also "maintainer_scripts_except_postrm" and "postrm" just to be sure that every single file of previous package has been removed ? Mar 30 13:18:09 TheMasterMind1, ok. I think I will add a couple more of fixes in few time. Mar 30 13:20:57 TheMasterMind1: Im on 2006.0 Mar 30 13:21:12 TheMasterMind1: on an X86_64 Mar 30 13:22:12 Genesis: Could you also dump your shell environment to a file please? Mar 30 13:22:59 yeap Mar 30 13:23:45 XorA: i'm on 2004.something or other Mar 30 13:24:22 http://dev.aimao.org/poky/shellenv.txt Mar 30 13:27:08 Genesis: I think the problem is coming from libXsettings Mar 30 13:27:46 i think too Mar 30 13:28:22 same error on my home and work machines Mar 30 13:29:52 Genesis: There is a reference to /usr/X11R6/lib in build/tmp/staging/arm-linux/lib/libXsettings-client.la which is where I suspec the problem comes from Mar 30 13:31:26 Genesis: As a test, try removing that reference from that file and contiuing the build Mar 30 13:31:40 ok Mar 30 13:31:54 If that works, we can at least blame a particular package :) Mar 30 13:33:33 that seems to work :O Mar 30 13:33:52 arf i should really learn more about find/grep Mar 30 13:34:01 my syntax was wrong Mar 30 13:34:25 i also try to configure beagle for the multifile searsh Mar 30 13:34:47 Genesis: The real question now is why did it put that there in the first place... Mar 30 13:35:12 yeap Mar 30 13:36:15 i should see on the old version first Mar 30 13:37:42 Genesis: Looking at tmp/work/libxsettings-client/Makefile, it has references to /usr/X11R6 in X_LIBS and C_LIBX Mar 30 13:37:59 X_CFLAGS even Mar 30 13:51:55 Genesis: I think this will fix it properly: http://www.rpsys.net/openzaurus/temp/libxsettings_fix-r0.patch Mar 30 13:56:00 noticed , ive a look after that after Mar 30 14:03:32 i'll have more time for OE if i succeed to port it for my board :) Mar 30 14:10:06 zecke: is your car working again now? Mar 30 14:10:13 Pigi: yeah, I remember that Mar 30 14:10:29 pb_ you suggested me to use the "remove_data_files_and_list" function which is good, but do you think it could be usefull to remove also "maintainer_scripts_except_postrm" and "postrm" just to be sure that every single file of previous package has been removed ? Mar 30 14:10:40 pb__: I guess not. It is not using Windows Mobile for Cars yet Mar 30 14:10:47 pb__: so it lacks the autorepair functionality Mar 30 14:16:12 grrrr :( Mar 30 14:17:16 the maintainer for most of the opie packages is the opie mailing list (opie@handhelds.org)... if i try and send an email to that address, it comes back with (you are not allowed to post to this mailing list), so how exactly do you contact them? Mar 30 14:17:43 HuntrCkr: subscribe to opie-devel :) Mar 30 14:17:56 HuntrCkr: alternatively if you want to file a bug use opie-bugs.oszine.de Mar 30 14:18:14 HuntrCkr: or just ask your question here, or in #opie Mar 30 14:18:22 thanks zecke... finally, a useful answer :) Mar 30 14:18:50 HuntrCkr: don't be too shy :) Mar 30 14:31:21 mickey|zzZZzz: bastard :) Mar 30 14:44:11 hi Mar 30 14:44:20 hey CoreDump|home Mar 30 14:48:09 03ade 07org.oe.dev * rfb0a7d68... 10/packages/qemu/qemu-qop-nogfx-native_svn.bb: qemu: add a differing native qemu that works with gcc4, arm support only atm Mar 30 15:03:03 RP, : is there a trick to pass the crosshairs calibration screen in qemu ? Mar 30 15:03:19 my initrd seems to fork fine :) Mar 30 15:07:49 Genesis_: The trick is to disable it Mar 30 15:08:06 Genesis_: We need to find a nice way to do this :-/ Mar 30 15:08:24 yeap Mar 30 15:08:36 add pointeur emulation in qemu ? Mar 30 15:09:06 stilet Mar 30 15:09:11 Genesis_: The board we're meulating doesn't have a touchscreen so this is tricky. Mouse emulation is already there Mar 30 15:09:22 yeap Mar 30 15:09:28 with good eyes Mar 30 15:09:36 we can see a little pixel :p Mar 30 15:10:31 but more than mouse emulation , stylet emulation will see the cross only when we 'll click on the right button for example Mar 30 15:15:02 sorry Mar 30 15:41:38 hi guys Mar 30 15:41:45 ey mickeyl Mar 30 15:42:02 hi mickeyl Mar 30 15:42:07 yo mickeyl Mar 30 15:42:24 * NAiL can't build TCL suddenly. http://pastebin.com/630886 Mar 30 15:42:42 I've looked at the configure sources, but I can't figure out what's wrong. Looks like a normal case-statement to me. Mar 30 15:43:39 NAiL: I had the same problem Mar 30 15:43:55 NAiL: I assumed the EABI stuff broke it Mar 30 15:44:05 apparently it's a general problem Mar 30 15:44:44 hmm Mar 30 15:45:22 works for me Mar 30 15:45:33 koen: What distro are you running? Mar 30 15:45:43 debian on the host Mar 30 15:45:48 angstrom-2006.9 in OE Mar 30 15:46:00 debian testing iirc Mar 30 15:46:13 I'm running ubuntu dapper Mar 30 15:46:16 ie, testing Mar 30 15:46:26 And I think it suddenly broke a while back Mar 30 15:47:01 * NAiL suspects it's the build environment that is broken. Because suddenly it won't work on the sources in SVN either. Mar 30 15:47:29 iirc mickeyl uses mandriva Mar 30 15:47:44 righto Mar 30 15:51:29 03mickeyl 07org.oe.dev * r570c957f... 10/packages/boost/boost_1.33.0.bb: boost 1.33.0: make it compile w/ linux-gnueabi Mar 30 15:51:34 03mickeyl 07org.oe.dev * ra2b816a5... 10/packages/qt/ (11 files in 4 dirs): qt-x11-free: upgrade to 3.3.6 and 4.1.1 respectively Mar 30 15:51:38 03mickeyl 07org.oe.dev * r028fac41... 10/packages/python/ (12 files in 3 dirs): python: upgrade to the latest and greatest version: 2.4.3 Mar 30 16:09:12 03mwester 07org.oe.dev * r1f027d0a... 10/conf/distro/unslung.conf: Unslung: conf/distro/unslung.conf updated to Unslung 6.8-beta Mar 30 16:11:23 later all Mar 30 16:19:15 bonne soiree Mar 30 16:23:35 hmm now what was the command to revert a revision Mar 30 16:23:46 rm -rf Mar 30 16:23:56 monotone disapprove Mar 30 16:24:17 which in this case will create an extra head which you need to merge Mar 30 16:24:32 I usually get lazy and just fix it by hand and commit Mar 30 16:24:40 koen: thanks Mar 30 16:24:45 revert as in undo the changes locally, or a commit? Mar 30 16:24:56 NAiL: a commit Mar 30 16:25:01 ah Mar 30 16:25:08 * NAiL just re-read Mar 30 16:25:35 ~lart tcl for whatever makes it fail to build Mar 30 16:25:36 * ibot puts on a hockey mask and jumps out at tcl for whatever makes it fail to build Mar 30 16:44:34 03coredump 07org.oe.dev * r4089b02c... 10/packages/gpe-session-scripts/ (2 files in 2 dirs): disapproval of revision 'e00cbe4865a1571655d38a2fe99892f1ec2abb83' Mar 30 16:50:57 * france_ is away: Away Mar 30 17:12:42 03coredump 07org.oe.dev * rd0b03a40... 10/packages/altboot/ (altboot_0.0.0.bb files/altboot.func): altboot: Be more verbose about when all messages are coming from the new rootfs Mar 30 17:53:14 oooooh Mar 30 17:53:16 http://xaos.tenmei-pl.org/oz/bootsplash-landscape.png Mar 30 17:54:16 purty Mar 30 17:54:38 hehe Mar 30 17:54:50 nice Mar 30 18:02:53 RP: have you finally implemented parallel builds? Mar 30 18:03:01 RP: It can't be too hard Mar 30 18:05:55 koen: sweet! Mar 30 18:09:30 koen: I think you know it already, caching if bitbake is fixed again Mar 30 18:09:38 zecke: cool! Mar 30 18:09:43 koen: it would be nice if you could confirm that building glibc is way faster Mar 30 18:11:12 RP: hi Mar 30 18:11:12 hi all... Mar 30 18:11:21 hey cedric_ Mar 30 18:13:07 wow debian will break evolution for everyone :) Mar 30 18:14:17 indeed Mar 30 18:14:22 thanks to glice Mar 30 18:14:37 ? Mar 30 18:16:10 re Mar 30 18:16:24 wb Mar 30 18:16:36 network outage at work Mar 30 18:16:51 hrw: no Audi for you :) Mar 30 18:17:57 zecke: you make it sound like evolution wasn't broken to start with Mar 30 18:18:13 zecke: debian breaks due to the new glice allocator in glib Mar 30 18:18:25 koen: right :) Mar 30 18:18:50 well not even the evil KDE think tank could have plotted that better Mar 30 18:19:03 :) Mar 30 18:19:07 one should put a SUGGESTS Kontact into the evolution package Mar 30 18:19:12 ouch Mar 30 18:19:56 koen: glice is not enabled on Ubuntu? Mar 30 18:20:29 zecke: I suspect the ubuntu people just use a matching evo build Mar 30 18:20:31 hi cedric_ Mar 30 18:24:38 zecke: It will happen the moment I get enough free time :) Mar 30 18:24:59 zecke: bitbake is also slow again Mar 30 18:25:22 4m45.909s/1m2.738s Mar 30 18:25:31 uncached/cached Mar 30 18:25:47 koen: ewi is busy Mar 30 18:26:17 RP: hehe take a day off :) Mar 30 18:26:24 Its just started running quite fast for me. I think I have disk IO problems on my server :-/ Mar 30 18:26:52 RP: as Schilling would say, try to use a supported OS Mar 30 18:26:58 RP: like solaris 2.x Mar 30 18:27:41 cu Mar 30 18:27:49 cu marcin Mar 30 18:28:09 zecke: in a minute you'll tell me Macs rule as well ;-) Mar 30 18:28:35 RP: Darwin/OSX have awesome tools Mar 30 18:28:40 hi cedric_ Mar 30 18:28:44 RP: but disk io seems poor like on any laptop Mar 30 18:29:00 threading sucks on osx Mar 30 18:29:06 as will all microkernels Mar 30 18:29:16 zecke: use lots of memory as a disk cache Mar 30 18:29:25 s/will/with Mar 30 18:29:36 my laptop flies with the old bitbake - haven't tried the new version on it :) Mar 30 18:29:41 I have tried running dope in qemu Mar 30 18:30:13 oh apropos qemu Mar 30 18:30:27 RP, could you give me some hints how to go on after having build an image for qemuarm? Mar 30 18:30:35 i.e. how to start that beast? Mar 30 18:30:48 mickeyl: Its actually on the OE wiki :) Mar 30 18:30:52 oops Mar 30 18:30:57 * mickeyl looks Mar 30 18:31:00 vim src/kernel/qgfxraster...cp Mar 30 18:31:01 mickeyl: MACHINE=qemu-arm :) Mar 30 18:31:06 vim src/kernel/qcolor.h Mar 30 18:31:09 when compiling fontconfig-native OE is passing the font config directory in my cross staging area as a include argument to gcc, causing the compile to fail later on at linking Mar 30 18:32:11 mickeyl: http://oe.handhelds.org/cgi-bin/moin.cgi/UsefulConfigurations which links to a webpage of mine Mar 30 18:32:22 aaah, right. thanks Mar 30 18:33:17 how does autotools know where to look for include dirs and where not to Mar 30 18:33:53 RP: " qemu are know" is a 'n' missing there? Mar 30 18:34:50 zecke: thanks, fixed :) Mar 30 18:35:15 TheMasterMind1: sorry, i have no idea. i don't even remember who our autofoo wizard was Mar 30 18:35:20 was that kergoth by chance? Mar 30 18:35:43 mickeyl: We really need to have some patches for that machine to disable the touchscreen and enable the mouse for both opie and gpe Mar 30 18:36:11 mickeyl: yes kergoth and pb_ are GNU utils experts Mar 30 18:36:16 RP: right. Mar 30 18:36:24 hmmm Mar 30 18:36:34 any autoconf gurus handy? Mar 30 18:36:37 TheMasterMind1, I think it's something in auto* switches. Mar 30 18:36:39 RP: can qemu do PS2 mouse input? Mar 30 18:36:53 hi Mar 30 18:36:57 zecke: It already does, and it works under kdrive nuder gpe Mar 30 18:37:05 zecke: but we hide the cursor :) Mar 30 18:37:07 TheMasterMind1, And yes, aclocal *IS* auto* :) Mar 30 18:37:17 hrm Mar 30 18:37:19 that's a bug Mar 30 18:37:24 RP: ah we need to set QWS_MOUSEROTO and compile the cursor into QtE Mar 30 18:37:30 gpe should only disable the cursus when it finds a touchscreen Mar 30 18:37:48 anyone her, who could help with some pkg-config/bitbake oddities? Mar 30 18:37:54 koen: Its a bug then. It should also only run xtscal if it finds a touchscreen really... Mar 30 18:38:00 but i have no switches.. i'm just calling autotools_do_install Mar 30 18:38:04 whats the command again to update the package list? Mar 30 18:38:14 RP: heh, indeed Mar 30 18:38:22 Ken|JLime: bitbake package-index Mar 30 18:38:24 mickeyl, zecke: We also have a problem with colours under opie. The framebuffer is BGR instead of RGB Mar 30 18:38:26 ah, thx man Mar 30 18:38:26 RP: Opie shows a Calibration when a QCalibratedMouseHandler is used Mar 30 18:38:27 TheMasterMind1, Well, I was talking about autotools, not how they're used in OE tree... Mar 30 18:39:12 liek why "pkg-config --variable=foodir package" prepends the staging dir? Mar 30 18:39:18 RP: that sounds challenging enough Mar 30 18:39:21 TheMasterMind1, I believe if OE has its own auto* then it should look in their install directories by default. Mar 30 18:39:34 RP: do you plan to implement the PXA270 fb for qemu? Mar 30 18:39:44 RP: I want to toy with overlays :) Mar 30 18:39:52 wow, the locale glibc stuff was marvelous, the time of first boot dropped from 25 min well under 5 in fam083 & h6300. Mar 30 18:40:02 zecke: I'd like to. See my time problem again :} Mar 30 18:40:34 zecke: Will QWS_MOUSEROTO mean the call to QCalibratedMouseHandler isn't used? Mar 30 18:40:37 RP: time, time, time... Mar 30 18:40:43 lamikr: excellent Mar 30 18:40:57 RP: It smells like TPanel is set as mouse proto Mar 30 18:41:22 RP: which will select a QCalibratedMouseHandler :) Mar 30 18:41:49 RP: hmm I need to get my paper into a 'releasable' state until tomorrow Mar 30 18:41:51 zecke: I don't know what I'm doing with qte ;-) Mar 30 18:42:33 RP: wait until I have time :} Mar 30 18:42:48 zecke i had same problem on a ep93xx arch ... :( Mar 30 18:42:58 gremlin[it]: what problem? Mar 30 18:43:11 no calibration is performed ;( Mar 30 18:43:17 gremlin[it]: check TT's QWS documentation Mar 30 18:43:28 gremlin[it]: well, we have a different issue with qemu Mar 30 18:43:49 ops ok sorry Mar 30 18:44:08 gremlin[it]: export QWS_MOUSE, QWS_MOUSE_PROTO or QWS_MOUSEPROTO to TPanel Mar 30 18:44:12 gremlin[it]: and tslib will be used Mar 30 18:44:33 cya later :) Mar 30 18:44:42 koen: bitbake -b foo -cclean Mar 30 18:44:52 koen: cleans all dependencies? Mar 30 18:45:01 zecke: what the hell is git doing? Mar 30 18:45:22 (except for using 100% cpu and loads of IO) Mar 30 18:45:30 koen: tailor Mar 30 18:45:58 screen 21611.pts-1.bitbake Mar 30 18:46:07 you can CTRL+C it if you want to Mar 30 18:46:31 naa, I'm just waiting for some idle time to do another bitbake test Mar 30 18:46:33 zecke nope, no tslib support for rustom ts interface :) Mar 30 18:48:30 zecke: will you come to LinuxTag this year? Mar 30 18:48:49 mickeyl: I asked yesterday if we should try to get a OE workshop done Mar 30 18:49:05 mickeyl: but I think I could get there for a couple of days Mar 30 18:49:47 it's in Wiesbaden this year Mar 30 18:52:40 right :) Mar 30 18:59:34 need to write some more Mar 30 19:03:19 sounds familiar to me Mar 30 19:03:22 *cough* Mar 30 19:10:28 * france_ is back (gone 02:19:29) Mar 30 19:18:49 write more what? Mar 30 19:34:13 all: What is a simple OE conf I can start with and that works? I have tried numerous so far, they all fail in some spot. Mar 30 19:37:36 all: Can someone explain me this error: /home/leon/sandbox/openembedded/build/tmp/cross/lib/gcc/armeb-linux/4.0.2/../../../../armeb-linux/bin/ld: unrecognised emulation mode: armelf_linux Mar 30 19:42:57 anyone here using 2.6.16 on a PXA270? Mar 30 19:45:37 neuromancer: pH5 and RP might Mar 30 19:45:46 koen, thanks Mar 30 19:46:36 koen, well thing is i was wondering if someone is using the EABI, and which CFLAGS to use (failed trying a coupleof different ones ... :( Mar 30 19:46:54 neuromancer: DISTRO=angstrom-2006.9 Mar 30 19:47:05 that will set up an EABI toolchain Mar 30 19:47:10 hmm Mar 30 19:47:24 ~ibot Mar 30 19:47:26 it has been said that ibot is a blootbot written in perl run by TimRiker on his server. logs on http://ibot.rikers.org// , ibot, jbot, apt are all the same process. It uses sqlite, but mysql or other SQL storage is also supported. Mar 30 19:47:28 yeah, that was what i was reading from the conf files Mar 30 19:47:55 sine you need glibc and gcc for EABI it isn't a matter of cflags Mar 30 19:47:58 but stupid as i am i wanted to give gcc 4.0.2 or 4.1 a try Mar 30 19:47:59 s/sine/since/ Mar 30 19:48:10 EABI only works with 4.1 Mar 30 19:48:17 and which flags? Mar 30 19:48:30 (not quite true, since it doesn't have thumb interwork enabled) Mar 30 19:48:46 neuromancer: no cflags, it does eabi by default Mar 30 19:49:00 it does? Mar 30 19:49:08 yes Mar 30 19:49:25 even if i don't define -mabi=aapcs or iwmmxt Mar 30 19:49:26 hmm Mar 30 19:49:27 run the arm-objumb -x on it Mar 30 19:49:45 neat Mar 30 19:49:56 so everything i ought to do is use 4.1? Mar 30 19:50:10 just use that angstrom distro Mar 30 19:50:20 it's quite complicated Mar 30 19:50:26 well .. Mar 30 19:50:42 the angstromdistro isn'tusing the standardgcc Mar 30 19:51:00 'standard'? Mar 30 19:51:20 it's using the only gcc in OE that has eabi support Mar 30 19:51:28 hmm Mar 30 19:51:29 can't get more standard that that Mar 30 19:51:52 well i thought i readsomewhere that they were using somekind of patched gcc ... Mar 30 19:52:03 no we aren't Mar 30 19:52:04 i'll give it a try then ... Mar 30 19:52:09 :) Mar 30 19:52:18 good Mar 30 19:52:30 (yes it's patched, but not patch hell like the csl one) Mar 30 19:52:37 same goes for the glibc Mar 30 19:52:41 sure Mar 30 19:52:42 koen: what would hold me from trying 4.1 for say an IXP4xx platform? Mar 30 19:53:01 koen, any downsides so far? Mar 30 19:53:03 likewise: nothing Mar 30 19:53:03 (except for my newbie-ness in OE) :-/ Mar 30 19:53:18 neuromancer: problems with pthread and lack of testing Mar 30 19:53:49 well testing could be something i could waste my time on ... Mar 30 19:53:54 neuromancer: the company doing the EABI stuff make money supporting their own toolchain, so they don't care much about 'upstream' Mar 30 19:53:59 but define problems with pthread ... Mar 30 19:54:04 s/make/makes/ Mar 30 19:54:17 neuromancer: mickeyl|bbl mentions something with pthread Mar 30 19:54:19 * neuromancer acks. Mar 30 19:54:45 pH5 got it booted into X on his ipaq Mar 30 19:55:08 hmm Mar 30 19:55:27 so you suggest i should get started with angstrom.conf ? Mar 30 19:55:38 or the 2006-9.conf ? Mar 30 19:56:58 neuromancer: ? Mar 30 19:57:27 21:53 < koen> neuromancer: DISTRO=angstrom-2006.9 Mar 30 19:57:44 set that, set your MACHINE and do 'bitbake Mar 30 19:58:55 ok Mar 30 19:59:45 so essentially these 2angstrom files (angstrom.conf and angstrom-2006.9.conf) hold what i needto make my own distro ready for EABI :) Mar 30 19:59:59 yes Mar 30 20:00:15 koen: spawning a build also, MACHINE=ixp4xx, DISTRO=angstrom-2006.9 Mar 30 20:00:54 likewise: cross your fingers and toes, since I suspect armeb wasn't high on CSL's priority list Mar 30 20:02:21 koen: ah yes. well all the others builds I tried failed for me as well :-/ Mar 30 20:02:28 koen, thanks@ Mar 30 20:02:50 likewise: I haven't built for ixp4xx for a while Mar 30 20:03:03 koen: I have been busy filling the bug databases :-) Mar 30 20:03:08 :) Mar 30 20:03:20 * koen should find some time to empty it Mar 30 20:03:44 koen: Don't you guys have bug-squash days on IRC? Mar 30 20:03:51 we have Mar 30 20:03:58 but not recently Mar 30 20:04:00 The threshold for fixing bugs via the bug database is a bit high. Mar 30 20:04:21 Where can i find a usefull setup for running multiple targets (HP jornada 6xx/7xx) for my distro? Mar 30 20:04:47 Ken|JLim1: see angstrom.conf how to include multimachine Mar 30 20:04:53 oki, thx Mar 30 20:04:57 koen: On the wiki there is a page about a .nl meeting in Eindhoven. Do you know anything about this? Mar 30 20:05:52 likewise: that was a while ago (I got sick right before) Mar 30 20:06:05 likewise: it would be a nice idea to organize a new one Mar 30 20:06:10 ~seen gerwinin Mar 30 20:06:13 gerwinin was last seen on IRC in channel #oe, 6d 22h 1m 51s ago, saying: 'My boss bought eight licenses and 1 of them I send back to micro$oft '. Mar 30 20:07:37 If i decide to put a copy of my repository (org.openembedded.dev / build ..) in a different dir, do i need to change anything to my OE.db? Mar 30 20:08:00 Ken|JLim1: no Mar 30 20:08:09 Ken|JLim1: and you can use a buildscript like this: http://ewi546.ewi.utwente.nl/tmp/viewmtn/file.psp?id=ea6e3bbdf7bb7190ff6a5c4b061fc6abb83a6907&path=contrib/buildscripts/openzaurus/build-oz.sh Mar 30 20:26:30 koen: indeed: GCC 4.1.x for ARM fails for BE: /home/leon/sandbox/openembedded/build/tmp/cross/lib/gcc/armeb-linux/4.1.0/../../../../armeb-linux/bin/ld: unrecognised emulation mode: armelf_linux Mar 30 20:26:54 koen: I think I had the same fault with 4.0.2 earlier. Mar 30 20:26:59 hrm Mar 30 20:29:32 koen: and right under: Supported emulations: armelfb_linux armelfb Mar 30 20:29:52 (which is more promising I guess) Mar 30 20:29:59 I guess Mar 30 20:30:11 I don't even know what that error means Mar 30 20:31:36 koen: lennert to the rescue: http://www.kegel.com/crosstool/crosstool-0.42/patches/gcc-3.4.2/gcc-3.4.0-arm-bigendian.patch Mar 30 20:32:13 that patch was in oe, i take it nobody looked at which of the 3.4.x patches, if any, were still needed in 4.x Mar 30 20:32:26 hey kergoth Mar 30 20:32:28 hey Mar 30 20:32:54 hey kergoth Mar 30 20:34:40 hi kergoth, woglinde Mar 30 20:34:51 hi pb Mar 30 20:34:53 hey pb_, kergoth Mar 30 20:35:02 yo stephmw Mar 30 20:35:03 * Stephmw crawls out from beneath his rock Mar 30 20:35:36 heh Mar 30 20:35:58 all I can say is that encrypted rom images stink... Mar 30 20:36:00 bleh. Mar 30 20:36:19 suck Mar 30 20:36:45 my dreams of flashing a consumer device with the kit at work... *sigh* Mar 30 20:37:11 different certificates for r&d and release... hw based too :( Mar 30 20:46:46 Stephmw: What hardware are you looking at? Mar 30 20:47:17 I copied my current 6xx folders and created new 7xx (changed local.conf) but it looks like it is draging some files from the old 6xx folders.? Mar 30 20:47:24 kergoth: I am now testing the patch... it needs a bit more fuzz than default, so I had to manually do it... Mar 30 20:47:28 likewise, no specific model :) Mar 30 20:47:34 likewise, and def not linux based Mar 30 20:47:39 Ken|JLim1: BBPATH and BBFILES? Mar 30 20:48:00 Stephmw: is that bad? I like QNX as well :-) Mar 30 20:48:21 likewise, naw, it's just not an open source platform (and no, not windows either) Mar 30 20:48:31 I've changed the BBPATH, not BBFILES though. How should i do that? Mar 30 20:48:31 likewise, but I can't be too specific Mar 30 20:49:43 Ah doh Mar 30 20:49:53 your are right Koen, i hadnt changed BBFILES Mar 30 20:49:54 thx Mar 30 20:57:27 Ken|JLim1: I've walked into that soooo many times... Mar 30 21:18:42 is there a way to escape a # Mar 30 21:21:08 in what context? Mar 30 21:21:40 hi reenoo Mar 30 21:22:15 hey pb_ Mar 30 21:22:46 You cannot escape the # Mar 30 21:22:50 it will hunt you down and kill you Mar 30 21:26:08 darn Mar 30 21:26:09 oh well Mar 30 21:26:43 I was just kidding Mar 30 21:26:47 I'm sure you can escape the # Mar 30 21:26:49 heh Mar 30 21:27:24 heh Mar 30 21:29:35 tmm1: what are you actually trying to accomplish? Mar 30 21:29:53 in terms of the bitbake parser, I don't think you can escape a #, but I can't immediately think of any reason you'd want to Mar 30 22:08:27 france_: so I have heard yeah. I love his work on behavior based robotics. ("Planning is just a way of avoiding figuring out what to do next.") Mar 30 22:08:45 err wrong channel Mar 30 22:09:00 reenoo: exactly.. it is ok, I am flexible... Mar 30 23:41:58 monotone is beginning to grow on me. I just wish it was faster Mar 30 23:42:45 They have a cream you can use to get rid of that Mar 31 00:38:44 hrm, first time i attempt to use oe in months and i get a db schema mismatch Mar 31 00:38:47 thats no fun Mar 31 00:41:27 confirm passphrase for key ID [kergoth@handhelds.org]: Mar 31 00:41:27 monotone: fatal: std::exception: Botan: ARC4 cannot accept a key of length 36 Mar 31 00:41:27 monotone: this is almost certainly a bug in monotone. Mar 31 00:41:29 * kergoth sighs Mar 31 00:41:36 methinks i should upgrade to current monotone :) Mar 31 00:55:24 heh Mar 31 00:55:32 where you working now kergoth? Mar 31 00:55:55 TheMasterMind1: for a subcontractor on an army contract, out in phoenix Mar 31 00:55:59 TheMasterMind1: google for Land Warrior Mar 31 00:57:11 nice Mar 31 00:59:34 i dont hate it, and thats about as good as it gets, methinks Mar 31 01:00:17 yea Mar 31 01:00:33 still debating leaving computing Mar 31 01:02:52 really? for what Mar 31 01:05:11 hrm i need to submit all these oe patches to bugzilla Mar 31 01:05:15 should do that soon Mar 31 01:07:15 no idea, just getting sick of doing this for a living. i miss actually enjoying coding as a hobby Mar 31 01:07:44 yep yep i know what you mean Mar 31 01:08:29 problem is, i like money. Mar 31 01:08:31 :P Mar 31 01:08:40 heh Mar 31 01:09:04 start a startup Mar 31 01:09:19 read some paul graham Mar 31 01:10:26 alright i'm out Mar 31 01:11:29 later Mar 31 02:31:18 i'm having a terrible time building native and cross fontconfig and freetype. the builds are referencing each other's libs and include files Mar 31 03:49:10 * france is back (gone 22:49:27) Mar 31 04:07:35 i have a none native .bb file that insists on not using the cross compiler Mar 31 04:11:02 ? Mar 31 04:11:43 all that means is that the source needs to be patched to behave Mar 31 04:12:04 it has some normal variable explicitly set Mar 31 04:12:12 variable(s)* Mar 31 04:18:56 yea just hard tracking it down Mar 31 06:27:25 where can I get a fat libx11 not the diet one Mar 31 06:53:14 Hey, can I move stuff from bitbake.conf to local.conf? e.g. a line like IMAGE_DEPENDS_nbi = "netboot-native" Mar 31 07:26:32 anyone have a full build on X11...ie not the diet? Mar 31 07:29:36 it seems that I can move between bitbake.conf and local.conf Mar 31 07:30:20 greyl: Sorry, I can't answer you question :( Mar 31 07:32:49 its ok, someone suggested using the debian arm libs...who knows maybe itll work...maybe I flash again :) Mar 31 07:33:10 greyl, probably one of eth tablet projects do Mar 31 07:33:24 they wont fit on any small device Mar 31 07:33:41 they say meth tablets work... :) Mar 31 07:33:57 emte what do you mean "they" Mar 31 07:34:15 the full X11 libraries adn server Mar 31 07:34:19 and* Mar 31 07:34:21 hi Mar 31 07:34:58 well I seem to remember this before(3.5.2) I want to compile tcl/tk/...everything else and it complains about missing some functions... Mar 31 07:35:37 hi all Mar 31 07:35:59 greyl: now tcl/tk builds without problems Mar 31 07:36:11 via oe? or by hand native? Mar 31 07:36:42 by the way I had to edit my /usr/lib/libc.so file again...is that expected? Mar 31 07:36:54 oe Mar 31 07:37:10 (I'll crawl into a hole now) Mar 31 07:37:24 (dialup is non functional for developing with oe) Mar 31 07:38:24 * emte shrugges Mar 31 07:38:39 just get a source cache from someone/where on broadband Mar 31 07:39:24 (Im too dumb to dev on my own and too dumb to figure out oe...head sand join) Mar 31 07:40:54 i wasnt aware taht oe is really that hard to use ... initial setup is a bit troubling tho Mar 31 07:41:54 (hmm...I know nothing about it. Im sure it's easy. I don't want to start a flame war. I just want ./configure;make;make install on my Z) Mar 31 07:42:15 roughly taht is how oe works Mar 31 07:42:35 hmmm... enybody else having trouble building tslib? Mar 31 07:42:50 it seems the CVS server is unavailable for it Mar 31 07:43:43 HuntrCkr: one of arm.linux.co.uk machines failed Mar 31 07:46:09 ok... anyway to get the CVS then? Mar 31 07:47:31 look at CVS_TARBALL_STASH and set SRCDATE_tslib to latest available snapshot Mar 31 07:49:28 morning all Mar 31 07:50:23 hi dirk, schurig Mar 31 07:50:44 hey Marcin Mar 31 07:50:51 morning Mar 31 07:51:01 hey XorA Mar 31 07:51:39 anyone wanna give a quick recap on the diff between hardware fp and software fp (in the arm world) Mar 31 07:52:04 greyl: i'm working on a full x11 bulid for OE Mar 31 07:52:07 (debian libx11 is hardware fp...Im guessing oe is software fp) Mar 31 07:52:08 greyl: most arms lack FPU so hardwareFP hit kernel emulator Mar 31 07:52:27 hrw|work: hi! Mar 31 07:52:46 hrw|work: OK... where do i set SRCDATE_tslib Mar 31 07:52:54 mastermind1 - praise be to al^H^H whoever....that'd be awesome! Mar 31 07:52:55 greyl: I wrote such a recap recently into the arm kernel mailing list Mar 31 07:53:02 HuntrCkr: local.conf Mar 31 07:53:02 greyl: use the mailing list archive Mar 31 07:53:08 will do thanks Mar 31 07:53:18 greyl: OE is build system not distro Mar 31 07:54:19 hrw|work: are you using gcc 4.1 (arm-none-linux-gnueabi-gcc) / binutils 2.16.91.0.6 ? Mar 31 07:54:31 (head bowed humbly .... just hoping for fat libx11) Mar 31 07:54:47 hrw|work: when I use this toolchain with my 2.6.16.1 kernel, sometimes error messages of missing symbols are wrong, they refer to the wrong .o/.c file Mar 31 07:55:42 TheMasterMind1: I need X11 (and Java+GTKAWT) really soon (in one, two months) Mar 31 07:56:09 schurig: I use .oz354x with openzaurus-3.5.4.1 currently Mar 31 07:56:46 hrw|work: I'm not familiar with the various zauri thingies ... I've mostly copied the settings from the conf/distro/angstrom*.conf file Mar 31 07:57:08 schurig: koen and ph5 are local EABI hackers Mar 31 07:57:09 hrw|work: hmmm.... would you believe there isn't a snapshot for tslib? Mar 31 07:57:18 HuntrCkr: shit happens Mar 31 07:57:19 hrw|work: i.e. I don't know which zauri distro uses what toolchain Mar 31 07:57:23 hrw|work: ahh, good to know Mar 31 07:57:36 mastermind1 - you meant tonight right? Mar 31 07:57:39 schurig: ah. oz user gcc 3.4.x, familiar user 3.4.x Mar 31 07:57:43 hrw|work: lol... yes it does Mar 31 07:57:49 schurig: angstrom use gcc 4.1/eabi Mar 31 07:58:21 hrw|work: yeah, I have two devices, the older one is stull using 3.4.3, and kernel 2.4.21 Mar 31 07:58:43 2.4.21.. last 2.4/arm.. Mar 31 07:58:48 hrw|work: I'm now porting this one to kernel 2.6.16.1 and thought that I could use the newest toolchain glibc as well. So far I have success, Mar 31 07:59:06 schurig: http://www.hrw.one.pl/2006/03/29/wireless-extension-v19-on-collie2418/ Mar 31 07:59:26 hrw|work: i.e. the 4.1 EABI compiled kernel can successfully nfs-root-mount an old filesystem image which was compiled with a non-EABI 3.4.3 Mar 31 07:59:33 schurig: you need glibc 2.4 (and some extra patches which are in OE) to use EABI Mar 31 07:59:56 koen: how are the images looking on boot now? Mar 31 08:00:25 koen: usespace EABI comes later :-) Mar 31 08:00:28 schurig: Those errors are due to arm mapping symbols and also occur under gcc 3.4 Mar 31 08:00:29 XorA: they boot into X, but stuff like 'busybox touch' errors out Mar 31 08:01:03 koen: now thats funky, something as complex as X works, but touch command flakes :-) Mar 31 08:01:07 morning all btw :) Mar 31 08:01:08 koen: for now, I have CONFIG_OABI_COMPAT set Mar 31 08:01:17 hey Richard Mar 31 08:01:28 schurig, greyl: i've been working on it for a few days now. but i just got all the deps compiled and xorg-server is compiling as we speak. i have no idea if it will work though. there are atleast 60 different ipks and i need to figure out which ones are actually required Mar 31 08:01:54 TheMasterMind1: yeah, X is a beast! :-) Mar 31 08:02:00 XorA: you underestimate the complexity of busybox =P Mar 31 08:02:26 TheMasterMind1: maybe I play with GTK in Framebuffer (via DirectFB) and use then GNU Classpath with it's awt-peer, not sure Mar 31 08:02:55 RP: what do you mean with "mapping symbols"? Mar 31 08:03:04 schurig: DirectFB is nice on pda, I didnt get as far as gtk on directfb though Mar 31 08:03:28 XorA: that's because it didn't work til a few weeks ago Mar 31 08:03:48 the debian people are using it for their installer and have fixed it Mar 31 08:04:04 I wish I had a memory...I delt with fat x11 in 3.5.2...murphys law I deleted that from my harddisk this last weekend...3.5.4's out, I have 3.5.3 too, why keep 3.5.2... :( Mar 31 08:04:35 koen: gpe-directfb, qt->directfb, and gtk->directfb would make some nice PDA stacks Mar 31 08:04:48 s/gpe/x11/ Mar 31 08:04:57 RP: for example, in drivers/misc/mn-sysctl.c I put a call to a function which isn't in the vmlinux because of a wrong .config file. After a "make" the toolchain claims "drivers/built-in.o: In function `mn_sysctl_handler':mn_kbd.c:(.text+0x24d7c): undefined reference to `mn_get_lcd_brightness'". And the error is certainly not in mn_kbc.c ! Mar 31 08:05:36 schurig: http://mail-index.netbsd.org/port-arm/2005/06/19/0000.html Mar 31 08:07:35 that explains a lot Mar 31 08:08:03 schurig: I'm just saying it was a problem with gcc 3.x as well and it doesn't mean the resulting object files are wrong, just that it can't read the symbols back in them properly Mar 31 08:08:04 * koen retracts some curses directed at cirrus Mar 31 08:08:34 hi Mar 31 08:08:41 RP: I noticed that the .o files are correct, after all I have few problems with my 4.1-EABI compiled kernel Mar 31 08:08:43 hey obergix[work] Mar 31 08:08:59 RP: do you mean I should patch binutils with http://lists.gnu.org/archive/html/bug-binutils/2004-07/msg00041.html ? Mar 31 08:09:03 if i inherit autotools it should run autoconf for me correct? what would cause it not to Mar 31 08:09:26 RP: it's actually only cosmetical, I know my kernel good enougth to find the real culprit of missing symbols, even without this patch :-) Mar 31 08:09:46 schurig: Me too :) Mar 31 08:09:55 hi obergix[work] Mar 31 08:10:21 TheMasterMind1: you need a configure.in or configure.ac in the unpacked source? Mar 31 08:10:30 schurig: I'm not sure that patch is correct - I don't think anyone has written a patch that solves the problem in a way everyone agrees on Mar 31 08:10:43 TheMasterMind1: details in /classes/autotools.bbclass, near line 80 Mar 31 08:10:48 schurig: yea there're there. log.do_configure shows a ./configure: not found Mar 31 08:10:54 k will look Mar 31 08:11:09 schurig: for every EABI problem you encounter, you should poke pbrook or drow on #debian-arm, they have patches usually aren't in upstream (but present in the csl-arm madness) Mar 31 08:11:19 TheMasterMind1: autoreconf will create the configure for you, that means, you don't need a ./configure file if you use inherit autotools Mar 31 08:11:55 koen: hehe, the problem is to find out if it IS a EABI or gcc 4.1 problem Mar 31 08:12:31 schurig: found it. S was set wrong so it wasn't finding configure or .in/.ac =P Mar 31 08:12:48 schurig: Asking if they have a patch wouldn't hurt, regardless :) Mar 31 08:12:59 RP: real 0m40.960s for a cached run of 'bitbake foo' Mar 31 08:13:29 koen: Yay :) Mar 31 08:13:48 RP: I have a strange difference between my 3.4.3 compiled kernel 2.4.15.6 for RT3000 and my 4.1 compiled 2.6.16.1 for MNCI. They both have the same 32 MB of Strataflash, even the same chips. Mar 31 08:13:53 and 2m30-ish for uncached Mar 31 08:14:29 schurig: note that gcc-41/glibc2.4 don't have the 'real' EABI yet Mar 31 08:14:37 RP: however, on RT3000 it works flawlessly, even when I reset in the middle of the run, but on the MNCI I get weird Strataflash states, even a JTAG debugger sees only garbage and I have to power-cycle. Mar 31 08:14:43 schurig: it doesn't enable thumbinterworking Mar 31 08:15:12 RP: but how to know if this is a toolchain problem or if it is an error made by me ... Mar 31 08:15:53 schurig: You have a lot of big changes and its hard to say which one is at fault :-/ Mar 31 08:15:58 koen: I think I only need this interworking if i mix thumb and non-thumb? Mar 31 08:16:20 schurig: EABI mandates it (and thus breaking for strongarm) Mar 31 08:16:28 hi, is there a document where the PREFERED_* things are explained ? Mar 31 08:16:37 koen: who cares for StrongARM :-) Mar 31 08:16:37 checking for sys/linker.h... configure: error: cannot run test program while cross compiling Mar 31 08:16:47 remember that ARM ltd has patents on thumb(2) Mar 31 08:17:15 schurig: http://wiki.debian.org/ArmEabiPort is the most comprehensive collection of eabi info Mar 31 08:19:59 koen: ABI also mandated having mapping symbols to enable the possibility of interworking Mar 31 08:21:40 how does one deal with this cannot run test program ./configure error Mar 31 08:22:21 TheMasterMind1: you need to look into config.log and configure to isolate the test Mar 31 08:22:44 TheMasterMind1: then you write what the result of this test should be into the proper /site/... file Mar 31 08:22:55 koen: 0m52.279s for a cached run on my laptop, with a monotone pull running as well :) Mar 31 08:23:09 schurig: k, thanks Mar 31 08:23:28 TheMasterMind1: basically, the configure script tried to run a test program to find something out, e.g. the size of an "unsigned const byte" or whatever. But it cannot run cross-compiled programs on your development pc, therefore the error Mar 31 08:23:55 TheMasterMind1: if you know the result, you can provide this into the site/* file, which ./configure reads to get pre-computed values Mar 31 08:24:57 TheMasterMind1: another option is to find out if you can specify the result with options to configure and put rhem into EXTRA_OECONF. If you can patch configure.ac/configure.in to not make the test in the first case, but this is not good style (the value could be different on MIPS etc) Mar 31 08:25:06 since OZ354: GBytes Transmitted: 142.503 GB Mar 31 08:25:21 gotcha. the check its doing is for sys/linker.h Mar 31 08:25:28 (minus ~3GB for OE stuff) Mar 31 08:25:34 koen: Is that a little higher than normal? :) Mar 31 08:25:53 RP: about 130GB higher :) Mar 31 08:26:19 koen: As long as they don't present us with a bill... ;-) Mar 31 08:26:35 heh Mar 31 08:26:55 it's in the 'lab' section of the uni network Mar 31 08:27:18 they might complain if you do >200GB/day Mar 31 08:27:57 We've a little way to go yet then :) Mar 31 08:41:17 someone has sl-a300 not-needed maybe? Mar 31 08:43:37 ibot: change 12000 jpy to pln Mar 31 08:43:42 12,000.00 Yen (JPY) makes 331.4575 Zloty (PLN) (from http://www.xe.com/) Mar 31 08:54:34 :) Mar 31 08:55:08 ibot, : change all my gold for a lady Mar 31 08:55:10 that doesn't look right Mar 31 08:55:13 :/ Mar 31 08:55:19 priceless. Mar 31 08:57:43 XorA: your blog is unreadable, dim dark green on black ... bleeech! Mar 31 08:57:55 schurig: PCW1512 :-) Mar 31 08:58:27 hi! someone time to explain me some toolchain things, i want to test to compile some packages with another toolchain, only for tests? Mar 31 09:04:49 grma: did you read the wiki ? Mar 31 09:05:50 no which one ? Mar 31 09:07:16 grma: http://oe.handhelds.org/cgi-bin/moin.cgi ... that should help for info about OE Mar 31 09:08:05 is there a project for an OE target wizard ? Mar 31 09:08:47 obergix[work]: i have alread build some oe images and targets, for tests i only want to change the compiler, i want to use a external one only for tests, but on the wiki i didn't find any info for that, am i blind ? Mar 31 09:09:20 Genesis: there are a few, but most of them lack usefull stuff for developers Mar 31 09:09:21 grma: cannot help you then Mar 31 09:09:47 grma, perharps should you look after scratchbox Mar 31 09:09:56 ok thanks Mar 31 09:10:04 grma: see http://planet.maemo.org/ Mar 31 09:10:06 ehm Mar 31 09:10:12 that seems cool to switch between different toolchain Mar 31 09:10:14 grma: http://oe.handhelds.org/cgi-bin/moin.cgi/ZaurusKernels Mar 31 09:10:31 maemo use scratchbox i think Mar 31 09:10:50 Genesis: getting your own toolchain into scratchbox isn't trivial Mar 31 09:10:59 oki Mar 31 09:13:11 build a toolchain is never trivial i think Mar 31 09:13:19 it is with OE =P Mar 31 09:13:26 for sure =) Mar 31 09:13:28 who is Rene Wagner? Mar 31 09:13:40 ~seen ph5 Mar 31 09:13:47 ph5 was last seen on IRC in channel #oe, 7d 10h 42m 45s ago, saying: 'RP: At the same time as := expands (during parsing)'. Mar 31 09:14:02 ~seen pH5 Mar 31 09:14:04 ph5 was last seen on IRC in channel #oe, 7d 10h 43m 2s ago, saying: 'RP: At the same time as := expands (during parsing)'. Mar 31 09:14:13 ah, ibot is case insensitive Mar 31 09:14:29 any one know rene's irc handle? Mar 31 09:14:45 TheMasterMind1: reenoo Mar 31 09:16:09 gracias Mar 31 09:16:51 Stephen Donecker irc handle ? Mar 31 09:19:20 oki :) Mar 31 09:40:50 03rpurdie 07org.oe.dev * r95a9843f... 10/packages/tslib/ (24 files in 16 dirs): Mar 31 09:40:50 tslib: Add 2.6 kernel handling for collie/poodle. Instead of every device using Mar 31 09:40:50 a different ts.conf file for 2.6, combine it all into one file which can be Mar 31 09:40:50 overridden on a per machine basis if needed (all copies were identical anyway). Mar 31 09:40:55 03rpurdie 07org.oe.dev * r51fcd1c5... 10/packages/libxsettings-client/ (libxsettings-client_0.14.bb libxsettings-client_0.16.bb): libxsettings: Include libxt in depends to keep configure happy Mar 31 09:45:07 RP: did you add ts-2.6.conf? Mar 31 09:45:50 koen: That would have been a good idea :) Mar 31 09:48:01 koen: thanks for spotting that Mar 31 09:48:45 I was just finding out it if was really missing or if I needed some more coffee Mar 31 09:48:49 usually it's the latter Mar 31 09:49:01 03rpurdie 07org.oe.dev * r455a5dad... 10/packages/tslib/tslib/ts-2.6.conf: tslib: Add missing file. Mar 31 09:50:34 hmm.. 190 revs to go Mar 31 09:51:08 good morning Mar 31 09:51:18 hi Bernardo Mar 31 09:51:44 hey Bernardo Mar 31 10:03:12 Hi Mar 31 10:03:25 How can i rebuild the Packages List in the deploy dir? Mar 31 10:03:45 bitbake package-inde Mar 31 10:03:46 x Mar 31 10:04:49 ah thank you. Ich package-index a command or is it a .bb file? Mar 31 10:05:07 2nd Mar 31 10:05:24 ah found it thank you Mar 31 10:48:48 The LED patches were jsut sent to Linus :) **** ENDING LOGGING AT Fri Mar 31 10:59:57 2006