**** BEGIN LOGGING AT Mon Jan 09 02:59:57 2006 Jan 09 03:00:49 bizarre the glibc no longer compiles Jan 09 03:01:07 arm-audit patch fails, but the actual version of glibc hasnt changed for ages Jan 09 03:15:36 XorA: glibc cvs made nasty changes which broke old version before a certain cvs date :-( Jan 09 03:16:07 RP: erp, means glibc-2.3.5+cvs20050627 is b0rked then Jan 09 03:16:26 XorA: I thought I'd checked in a patch which fixed it? Jan 09 03:16:42 RP: before or after 9am this morning? Jan 09 03:16:54 "fixed" as in removed all the files that mysteriously appeared Jan 09 03:16:58 before Jan 09 03:17:04 so evidently not Jan 09 03:17:26 The tricky appeared to be to remove all the files in the bits directory... Jan 09 03:17:41 RP: this is the patch arm-audit failing to apply Jan 09 03:17:52 RP: the file is there but is obviously not one patch was for Jan 09 03:18:21 There was a post to oe@ from koen with a link to correct tarball Jan 09 03:18:39 which file does it fail on? Jan 09 03:19:20 RP: gimme a sec, I lost error off top of screen Jan 09 03:21:28 RP: http://pastebin.com/497572 Jan 09 03:23:24 XorA: That does look rather serious. Better find the archive copy of cvs from ewi :-/ Jan 09 03:23:38 and we'll have to talk treke into mirroring it Jan 09 03:23:47 hi all, :) Jan 09 03:24:24 x86 can build konqueror? Jan 09 03:24:54 does OE can build konqueror for x86? Jan 09 03:26:23 Hmm. python has a module with an RDEPENDS on libgdbm3 which doesn't exist in OE :-/ Jan 09 03:26:44 RP: found the problem, for some reason its not getting cvs from 20050627 its getting toadys Jan 09 03:26:56 umm can anyone here comment on the status of e-image? Is it still developed? Jan 09 03:26:57 RP: I guess its a problem with CVSDATE/SRCDATE change Jan 09 03:27:11 XorA: You did update bitbake? Jan 09 03:27:30 RP: crapit Jan 09 03:27:49 ~lart me Jan 09 03:27:49 * ibot installs a bad bootloader on xora and turns xora into a brick Jan 09 03:27:52 * RP mutters something about reading mailing list posts and topics ;-) Jan 09 03:28:24 The trouble is if you can do it, everyone is going to :-/ Jan 09 03:28:35 RP: mailing list confused me, as at one point someone said bitbake doesnt need to change at others they say it does Jan 09 03:29:03 but this morning I should have thought clearer Jan 09 03:29:52 ~lart me some more Jan 09 03:29:52 * ibot squishes me some more like a bug Jan 09 03:29:55 Hi all! I am about to bitbake gpe-image with oe but get an error on the gettext-native package: /bin/sh ../lib/csharpcomp.sh -O -g -o GNU.Gettext.dll /home/DP/dev/oe/tmp/work/i686-linux/gettext-native-0.14.1-r3/gettext-0.14.1/gettext-runtime/intl-csharp/intl.cs Jan 09 03:29:55 | mscorlib: No such library Jan 09 03:30:36 is this a fault of the package or my system? Jan 09 03:30:44 XorA: bitbake needs to change for several reasons mainly that we could work around the CVSDATE one but not the DEPENDS one Jan 09 03:31:12 RP: yeah, I read the stuff, and now I think I got the mails backwards in my mind Jan 09 03:32:07 * XorA goes back to reading 101 Jan 09 03:32:46 it is a pity it doesnt die completely when doing my mistake Jan 09 03:33:04 that would have woken me up, but cant go back and break old release versions of bitbake :-) Jan 09 03:33:24 We could have OE check the revision of bitbake... Jan 09 03:34:44 RP: building nicely now, sorry for the moronicness Jan 09 03:34:59 anybody plz help me....can i use kernel from www.linux.org with some patch for my hx4700 instead downloading cvs from hh? Jan 09 03:35:10 XorA: What are you building out of interest? Jan 09 03:35:15 RP: gpe-image Jan 09 03:35:35 XorA: It won't work. I'm working on a fix... Jan 09 03:35:47 Well, fixes Jan 09 03:35:53 RP: cool, I was just trying out all the new stuff to give it some testing Jan 09 03:36:14 XorA: ok. It will get a reasonable way into it :) Jan 09 03:36:19 RP: theres been no actual app updates I want anyway Jan 09 03:36:43 I'd not recommend app updates given the current PV issues :-/ Jan 09 03:36:55 but this is all planned I guess :) Jan 09 03:37:08 RP: you havent seen the mess my c7x0 is in already :-) Jan 09 03:37:23 well, mine too :) Jan 09 03:37:52 RP: I accidently shut it down after an ipkg upgrade and forgot to sync disks, lost all my ipkg status files Jan 09 03:38:02 RP: so it doesnt remeber half the stuff thats installed Jan 09 03:38:08 You will notice the meta/packages and conf/machine files look a lot cleaner now they've been stripped of all the DEPENDS uglyness... Jan 09 03:39:06 <[lala]> RP: since you removed bootstrap_extra_depends, i was wondering if i can still add packages to bootstrap-image? Jan 09 03:40:11 [lala]: Yes, just use RDEPENDS Jan 09 03:40:44 [lala]: bitbake should figure out the DEPENDS for itself Jan 09 03:41:18 RP: we got the white screen again. first boot after flash... second boot fine Jan 09 03:41:19 <[lala]> RP: well, i certainly hope so ;-) Jan 09 03:42:27 greentux: At what point does the white screen appear? Jan 09 03:42:36 Do you see the kernel messages or not? Jan 09 03:43:34 <[lala]> RP: (on behalf of greentux): you flash the kernel, then after flashing the new kernel is booting and there is the white screen Jan 09 03:43:40 <[lala]> just some secondes after boot Jan 09 03:44:00 [lala]: So you see some of the kernel messages, then it white screens? Jan 09 03:44:08 <[lala]> RP: exactly Jan 09 03:44:32 <[lala]> RP: but doesn't happen all the time and a lot less since rc7 Jan 09 03:44:45 [lala]: Can you tell what line the kernel prints before it white screens? Jan 09 03:45:15 <[lala]> boah ... have to try reading that next time :) kinda difficult Jan 09 03:45:32 [lala]: Nothing changed in the -rc7 kernel that would affect this so it looks like some timing problem Jan 09 03:45:45 [lala]: The trouble is working out what the timing issue is... Jan 09 03:46:03 Not helped by the fact I've only ever seen this twice on my device... Jan 09 03:46:12 and they might have been other problems... Jan 09 03:46:38 <[lala]> i see, try to catch the line next time Jan 09 03:48:15 [lala]: thanks - it may give some clue. It could be a bad interaction with the bootloader if its around framebuffer init time... Jan 09 03:51:35 morning lrg|home Jan 09 03:51:51 morning greentux, all Jan 09 03:52:27 lrg|home: headset recieved? Jan 09 03:52:54 greentux: yes, got it this morning thanks Jan 09 03:53:26 hi lrg|home Jan 09 03:53:35 hey RP Jan 09 03:53:44 lrg|home: we ad a voip call, but only one direction.... the mixer overloads me :) Jan 09 03:54:19 greentux: the current mixer overloads me as well. I'm working on reducing it's size Jan 09 04:23:24 morning Jan 09 04:24:09 morning zecke Jan 09 04:24:24 yiha ! NOTE: package nano-1.3.9: completed Jan 09 04:26:15 feel like my first C++ hello world... except that it took more time to compile... Jan 09 04:26:28 hehe Jan 09 04:27:28 hail zecke Jan 09 04:27:36 did you fail your exam yet? Jan 09 04:27:48 pb__: not yet it is at 13h Jan 09 04:27:55 so 30 more minutes Jan 09 04:28:12 aha Jan 09 04:28:35 I might be lucky and pass this time Jan 09 04:28:55 ~hail zecke's drugs Jan 09 04:28:58 * ibot bows down to zecke's drugs and chants, "I'M NOT WORTHY!!" Jan 09 04:29:08 ^^ Jan 09 04:29:18 zecke: good luck Jan 09 04:29:49 zecke: let's hope so Jan 09 04:30:13 zecke: good luck! Jan 09 04:30:54 sneak in a PDA and we all try and answer the questions Jan 09 04:31:44 hi Jan 09 04:31:47 hi hrw Jan 09 04:31:58 hey hrhr Jan 09 04:32:03 how to get bitbake/OE going? (both up-to-date) Jan 09 04:32:16 mmm... Jan 09 04:32:23 http://pastebin.ca/36248 Jan 09 04:33:14 hrw: This is my fault - I've broken bitbake... Jan 09 04:33:17 hrw: Light an effigy of Tux, dance around clockwise for 3 minutes while humming at least two ABBA songs, and plug christmas lights into a firewire socket.. Jan 09 04:33:18 brb Jan 09 04:33:30 hrw: Basically, don't use the shell :-/ Jan 09 04:33:34 re Jan 09 04:33:50 RP: nasty you Jan 09 04:34:14 lrg: ping, ping, ping ;) Jan 09 04:34:23 alan|laptop: hi Jan 09 04:34:28 anyway looks like I have ewi 100% idle ;) Jan 09 04:34:34 RP: please tell lrg (if you chat with him) to send me his snail mail address Jan 09 04:34:45 hrw: Its also broken in more subtle ways. As soon as I have opie and gpe builds working again, I'll give provide some updates for bitbake... Jan 09 04:34:47 RP: I want to send the SL 6000er as soon as possible to him Jan 09 04:34:55 hi zecke Jan 09 04:34:58 lrg: hey Jan 09 04:35:04 zecke: tell lrg|home :) Jan 09 04:35:16 hrw: or downgrade bin/bitbake Jan 09 04:35:42 lrg|home: you would like to have a SL6000er for audio and touchscreen testing/implementing? Jan 09 04:35:45 I can live with no shell now Jan 09 04:36:05 zecke: yes please :) Jan 09 04:36:27 lrg|home: send your snail mail address to zecke@handhelds.org please Jan 09 04:36:40 lrg|home: I will ship it tomorrow then (wanted to ship it today...) Jan 09 04:36:54 lrg|home: any preferred way of sending this package? Jan 09 04:37:16 zecke: thanks, no prefs on delivery. whatever is easiest Jan 09 04:39:54 morning all Jan 09 04:40:00 morning Dirk Jan 09 04:40:03 hi dirk Jan 09 04:40:09 lrg|home: I will send the device tomorrow then Jan 09 04:40:14 hi Liam, Richard Jan 09 04:40:18 zecke: thanks Jan 09 04:41:21 "Auf den Kampf Toreror, hast du auch keine Lust, bist du voller Frust..." Jan 09 04:41:27 * zecke heads to the other building... Jan 09 04:43:45 RP: please can you test this on corgi and spitz http://www.do13.de/oz/patches/corgi-bl-generic-r0.patch Jan 09 04:47:26 do13_: Will do in the next kernel build (I need to upgrade to recent git soon anyway). Looks sane though. Jan 09 04:47:29 heya folks, got a build failure on libtool-native due to "configure already having been run". the do_configure stage runs... configure, and then, rather unusually, ./config.status --recheck re-running configure bombs out with a "source directory already configured; run "make distclean" there first error. this is libtool-native-1.5.6 _and_ 1.5.10 Jan 09 04:47:36 do13_: Does that work on tosa? Jan 09 04:48:29 RP: basically yes. I'll build a new image and test with opie. Jan 09 04:48:42 i have been successfully building for the past... two-three days, i did a monotone pull last night and went "uh-ohhh" as approx 100 files whizzed past my eyes (oh - and a monotone merge - v0.24 - bombs out with an internal error) ... Jan 09 04:49:26 i noted down everything i did to setup OE on my mandriva 2006. I can easily write a how-to now. The problem is everything is alerady explained in GettingStarted. If i write a how-to, it will be a real step by step for newbies like me... I'm not sure wether i should release this doc or not, as you may see more non dev users coming here to ask silly questions (like i do each day...). What do you guys think ? Jan 09 04:49:36 clues anyone? Jan 09 04:50:14 lkcl: current tree has many revolutions. keep the one from thursday Jan 09 04:51:27 lrg|home: I have a prob with pendown detection. The driver detects pendown events without touching the screen. ts_configure is finished after a couple of seconds. Jan 09 04:52:38 do13: in accelerated or polling mode ? Jan 09 04:53:03 lrg|home: polling mode Jan 09 04:54:45 do13: hmmm. not sure about that one. :( I'll get a better picture when my tosa arrives later in the week, you could try altering some of the module params to make the pen down signal less sensitive in the mean time. Jan 09 04:56:09 lrg|home: OK. Next days I'll send you my patches so you can build an actual kernel. Jan 09 04:57:28 oh, bugger. how to i return to thursday's thingy, then? Jan 09 04:57:41 do13: Thanks, I'm thinking now it may be too sensitive and holding the pda could be causing the spurious pen down events. Try increasing the rpu param to 8 in wm9712.c Jan 09 04:57:58 like with cvs and svn, is it possible to do an "update -D '2 days ago'"? Jan 09 04:58:28 lkcl: read channel topic Jan 09 04:58:45 ack. ta. Jan 09 04:59:06 lrg|home: OK, will check after work Jan 09 04:59:26 channel topic.... channel topic.... dumb stupid unfamiliar irc program.... *thinks*... unjoin #oe and rejoin... Jan 09 04:59:27 brb Jan 09 04:59:57 thanks! Jan 09 05:02:32 oh dear that's a lot of changes back :) Jan 09 05:05:28 RP: Corgi doesn't use the gpio_batfull IRQ? Jan 09 05:07:33 corgi doesn't, husky and shepherd do Jan 09 05:09:39 RP: How do you detect if the battery is full? read_devdata(SHARPSL_STATUS_CHRGFULL) is only called from sharpsl_off_charge_battery Jan 09 05:11:17 do13_: It basically doesn't for corgi and will preiodically always trigger the charger Jan 09 05:11:44 do13_: Reading through the sharp code, that was my understanding of how it worked. I remain slightly sceptical though... Jan 09 05:12:20 I did have a borrowed corgi which this appeared not to harm (the charger IC itself has the final say about charging anyway) Jan 09 05:13:49 oh fer pity's sake. same error. "source directory already configured" Jan 09 05:15:46 RP: strange... Jan 09 05:16:41 do13_: I suspect Sharp made a mess of charging in corgi... Jan 09 05:17:01 Sounds like it.. Jan 09 05:17:08 tn Jan 09 05:17:26 this is _after_ doing a monotone update -rjan2006prebreakage Jan 09 05:18:31 RP: on tosa the charging status is checked in the battery thread. Jan 09 05:20:05 do13_: I honestly can't remeber exactly what the other SL models did in there. Over a long period of time, I was able to remove a lot of that code. Perhaps some of it was removed incorrectly? Jan 09 05:22:18 RP: I'll remove 500 #idefs in sharps code. Maybe is human readable. Jan 09 05:23:23 do13_: I should really go back and reread it now we know what most of it is trying to do and check whether what we ended up with is correct. Part of me never wants to see that code again though Jan 09 05:25:31 RP: I can't image sharp / lineo knows how the charger works. Jan 09 05:36:18 do13_: No. It'd be nice to know if the hardware on corgi is the same as husky/shepherd... Jan 09 05:44:08 RP: Do you have pictures of corgi/husky and shepherd? Then we can compare the charger IC. Jan 09 05:44:41 RP: I read a bit from weekend discussion about CVS_TARBALL_STASH. when you will change it could you add support for file:// locations? On ewi I use koen's DL_DIR as source of archives.. Jan 09 05:47:31 hrw: I'd also like to see that and had file:// mind :) Jan 09 05:52:24 what do you think about new oz website look? I just changed fontsize, rightsidebar Jan 09 05:52:53 "/usr/bin/ld: cannot find -lc"... i had this error on my first computer. I can't remember what lib provides this... anyone ? Jan 09 05:53:53 hrw: clean and sober. Just allright. Jan 09 05:53:56 wtf... I could have passed Jan 09 05:54:10 zecke: but you didn't ? Jan 09 05:55:29 hrw: openzaurus misses a logo or something... Jan 09 05:55:51 alan|laptop: I will know in about a week Jan 09 05:56:01 alan|laptop: oz never had a logo - it used opiezilla Jan 09 05:56:18 zecke: keep finger crossed... Jan 09 05:56:47 hrw: I guess it is the other way around :} Jan 09 05:56:57 hi i did this to create my openslug crossdevelopment system: download Makefile, make update, make openslug-image but after some seconds i always get this error: Jan 09 05:57:14 File "/home/kami/slug/openslug/bitbake/lib/bb/fetch/wget.py", line 66, in md5_sum Jan 09 05:57:14 kami22: use pastebin please Jan 09 05:57:14 return parm['md5sum'] Jan 09 05:57:14 KeyError: 'md5sum' Jan 09 05:57:15 make[1]: *** [image] Fehler 1 Jan 09 05:57:24 okay sorry Jan 09 05:57:26 kami22: you may lart me, then do a svn up on bitbake Jan 09 05:57:28 zecke: truely speaking: I dont know how it was. it was long before I ever heard that zaurus and opie exist Jan 09 05:57:57 how can i do svn up on bitbake? Jan 09 05:58:08 hrw: I think kergoth, bipolar, harlekin just used the OZ logo... Jan 09 05:58:15 kami22: where do you have your bitbake tree? Jan 09 05:58:35 zecke: now opiezilla is recognized rather as opie logo then as oz one Jan 09 05:58:49 # /slug/bitbake Jan 09 05:59:22 kami22: here is a quick guide Jan 09 05:59:32 kami22: cd /slug/bitbake && svn up Jan 09 06:02:19 alan|laptop: I think I should have passed because I spotted a integer overflow Jan 09 06:02:57 The use of PKG_ in OE totally breaks the RDEPENDS code. Thinking of openobex which has PACKAGES = "${PN} ..." yet ends up generating a package called libopenobex-${SOMVERSION}. gpe-beam then uses RDEPENDS = "libopenobex-1.0-1" Jan 09 06:03:19 RP: :( Jan 09 06:03:26 lrg|home: did you send an email yet? Jan 09 06:04:21 lrg|home: ah the spam filter ;) Jan 09 06:04:30 Any ideas how to solve this would be welcome. I'm losing the will to live :-( Jan 09 06:04:36 zecke: yep Jan 09 06:06:04 lrg|home: oh you already mailed me :} Jan 09 06:06:10 stupid spam emails Jan 09 06:06:15 *sorry* Jan 09 06:06:25 zecke: np Jan 09 06:07:22 http://home.hrw.one.pl/wordpress/ - someone can reach it? Jan 09 06:07:49 yes Jan 09 06:07:53 hrw: i can Jan 09 06:08:02 hrw: lacks some images etc.. Jan 09 06:08:02 what do you think about intro text on top? Jan 09 06:08:55 hrw: stylesheet points to http://localhost Jan 09 06:09:06 ah.. Jan 09 06:09:33 now should work Jan 09 06:10:21 works Jan 09 06:10:51 intro need rewrite but it atleast give something - will add link to about inside Jan 09 06:10:59 this was again a complete rebuild of the entire distribution from scratch. Jan 09 06:11:16 we always build the distro from scratch? Jan 09 06:11:19 I dont know who wrote 'about openzaurus' Jan 09 06:12:09 RP: well you would have to apply the same vodoo tactics some bbclass'es apply.. :} Jan 09 06:12:24 Actually, this is totally insane. If someone implements a nondebian.bbclass , libopenobex breaks Jan 09 06:12:56 I don't know what the solution is but currently, it isn't right... Jan 09 06:13:33 zecke: We'd be putting policy into bitbake then :-( Jan 09 06:13:54 RP: right :( Jan 09 06:13:57 RP: isnt that RDEPENDS generated by shlibs code anyway and isnt needed? Jan 09 06:15:02 * RP hugs XorA. I like this solution :) Jan 09 06:15:16 * alan|laptop bbl Jan 09 06:15:53 It just means we hide the problem though :-/ Jan 09 06:21:04 zecke: sorry i am no back from lunch Jan 09 06:23:16 kami22: don't worry, svn up and you will be fine Jan 09 06:23:33 if i do svn up i always get only this : /slug$ cd bitbake && svn up Jan 09 06:23:33 Überspringe '.' Jan 09 06:23:57 kami22: svn status is saying what? Jan 09 06:24:19 kami22: alternatively go to /slug/bitbake/lib/bb/fetch/wget.py Jan 09 06:24:32 kami22: go to line 66 Jan 09 06:25:02 svn status Jan 09 06:25:02 svn: '.' ist keine Arbeitskopie Jan 09 06:25:58 OZ devels: feel free to add to 'developers' link category on openzaurus.org Jan 09 06:26:34 NOTE: package meta-opie-1.0-r32: task do_build: completed Jan 09 06:26:37 yes :) Jan 09 06:27:59 kami22: replace your lib/bb/fetch/wget.py with http://svn.berlios.de/viewcvs/*checkout*/bitbake/trunk/bitbake/lib/bb/fetch/wget.py?rev=332 Jan 09 06:28:25 yes but there are more errors Jan 09 06:28:41 wait a second i will post them on the internet? Jan 09 06:28:47 ~pastebin Jan 09 06:28:53 extra, extra, read all about it, pastebin is a place to paste your stuff without flooding the channel - try http://pastebin.com/, or http://pastebin.ca Jan 09 06:30:36 this is my error : http://pastebin.com/497699 Jan 09 06:31:35 kami22: well... do what I said ;) Jan 09 06:31:42 kami22: long backtrace, one error Jan 09 06:33:29 lrg|home: would 5-7 days for package delivery be okay? Jan 09 06:35:59 okay it works thanks Jan 09 06:36:12 will post here if i get any new errors :) Jan 09 06:36:21 kami22: your welcome to Jan 09 06:36:34 thanks :) Jan 09 06:38:06 zecke: yep, thats fine Jan 09 06:41:22 zecke: wb Jan 09 06:44:19 cross your fingers! Jan 09 06:45:02 right :-) Jan 09 06:46:33 someone tried xfce-image on Z or other pda? Jan 09 06:49:46 RP: oz 3.5.4.1 will get 2.6.15 for cxx00 and for c7x0? Jan 09 06:50:32 hrw: I expect so given the advances we've made Jan 09 06:50:38 ok. Jan 09 06:50:52 RP: I'm just adding 'openzaurus 3.5.4 release plans' post to openzaurus.org Jan 09 06:51:14 i've had to comment out the "./config.status --recheck" from the Makefiles in libtool-native _half-way through_ the bitbake build. Jan 09 06:51:29 RP: you can look at it in wp-admin Jan 09 06:51:40 hrw: I'll probably switch to 2.6.15 as the stable version and 2.6.15+2.6.16-git as the unstable soon Jan 09 06:51:59 this "gets it through" Jan 09 06:52:11 hrw: You saw my page about 2.6 keymaps? Hope that's ok. I might add some hints about alsa audio to it as well Jan 09 06:52:13 RP: 2.6.15 works ok on my 760 Jan 09 06:52:19 RP: not yet Jan 09 06:52:40 RP: give url Jan 09 06:52:44 I haven't tried it yet but I don't see why it won't work as well as 2.6.15-rc7 :) Jan 09 06:53:14 hrw: http://openzaurus.org/wordpress/howto/c7x0-and-cxx00-26-kernel-information/ Jan 09 06:54:46 RP: nice Jan 09 06:55:09 hrw: Did you add pageup/pagedown keys? Jan 09 06:55:26 RP: add a post with info that such howto is added - category 'changes' is going to right sidebar automatically Jan 09 06:55:37 RP: dont remember now Jan 09 06:56:20 RP: http://openzaurus.org/wordpress/2006/01/09/openzaurus-354-release-plans/ Jan 09 06:57:11 feel free to edit Jan 09 06:59:00 cu Jan 09 07:11:21 hi, is anyone experimented with c3000 initrd image makeing here please ? Jan 09 07:56:50 glurp.... "/usr/bin/ld: cannot find -lc Jan 09 07:56:50 "... what the hell do i miss ? Jan 09 07:57:04 libc, apparently Jan 09 08:01:32 I've just sent a mail explaining the latest RDEPENDS problem to oe@. If anyone feels like helping, please let me know... Jan 09 08:05:53 RP: woooo my head aches now Jan 09 08:10:48 RP: you define the default OE kernel settings? Jan 09 08:11:01 greentux: I guess I do, yes Jan 09 08:11:08 RP: earlier kernel had TKIP CCMP etc set, recent not... why? Jan 09 08:12:32 greentux: I think that's accidental. Perhaps the config options changed format or something... Jan 09 08:12:52 greentux: Its certainly not something I did deliberately Jan 09 08:13:11 RP: ok NP, but how it can become default again? Jan 09 08:13:18 RP: bugzilla? Jan 09 08:13:30 greentux: I'll fix locally and it will appear in the next kernel Jan 09 08:13:45 morning Jan 09 08:13:53 RP: tnx Jan 09 08:22:46 XorA: You don't fancy fixing debian.bbclass then? :) Jan 09 08:26:55 RP: I was trying to think about the problem and my brain got stuck in a loop where a needs b and b needs a Jan 09 08:27:31 RP: but on paper you solution sounds sensible, it avoid b needing to know a Jan 09 08:27:39 pb_: thanks. :) Jan 09 08:27:57 XorA: I've stuck bitbake in that loop a few times recently :) Jan 09 08:28:22 RP: hopefully someone cleverer than me cade code that solution for you Jan 09 08:29:35 XorA: You can use PKG_${PN} to lookup what an entry needs to change to. The problem is that the appropriate PKG_ namespace might not be available if you're working between packages Jan 09 08:30:33 For instance, the packaging of gpe-beam needs to lookup PKG_openobex to correctly translate the RDEPENDS line Jan 09 08:31:40 pb__: Have you any thoughts on this? Jan 09 08:35:35 pb_: in fact it was glibc-static-devel, even if i had glibc-devel... Jan 09 08:51:36 * france is away: Away Jan 09 08:53:32 tailor stopped on ewi, I had to restart it Jan 09 08:53:51 we have git history until the 21st of september now Jan 09 08:54:58 zecke: Neat. I'll have to find time to experiement with it and setup this repository... Jan 09 08:55:13 I'm looking forward Jan 09 08:55:20 * RP will manage that when he fixes OE :-/ Jan 09 08:55:37 RP: good luck, tough job though Jan 09 08:56:46 RP: just a thought, does this mean all RPDEPENDS must be built as the debian code doesnt know the name of the output package until it is build Jan 09 08:57:07 XorA: It does :-( Jan 09 08:57:54 I'm hoping we don't have debianisms in many RDEPENDS and I can ignore this issue until after we have working metadata... Jan 09 08:58:26 XorA: Basically, debain.bbclass will have to set BUILD_ALL_DEPS = "1" :-/ Jan 09 08:59:11 RP: yeah, that sounds okay to me Jan 09 08:59:57 RP: there cant be many cases where you RDEPEND on libXXX as normally you need the headers/link stuff for it before using it, and linux people dont tend to do general plugins much Jan 09 09:01:32 XorA: Yes, its not as bad as it sounds. But having BUILD_ALL_DEPS set all the time isn't a very nice thing to insist on :-/ Jan 09 09:03:59 NOTE: package gpe-image-1.0-r19: task do_rootfs: started Jan 09 09:04:50 NOTE: package gpe-image-1.0-r19: task do_rootfs: failed Jan 09 09:04:56 Wasn't to be :) Jan 09 09:05:00 Close though Jan 09 09:05:04 RP: you will get there! Jan 09 09:05:37 hrw|gone: ? Jan 09 09:05:58 serial_cs: RequestIO: No more items Jan 09 09:06:02 any idea? Jan 09 09:06:50 Noooooooo. Its just failing on one package. fontconfig-utils (aka, libfontconfig-utils) Jan 09 09:07:46 rhaaaa ! dam router ! Jan 09 09:08:53 Should fonts be RDEPENDing on fontconfig-utils? Jan 09 09:10:49 hi can anyone help me i am trying to compile php for openslug i started this command : bb php Jan 09 09:11:14 and i always get this error : checking for iconv... no Jan 09 09:11:14 checking for libiconv... no Jan 09 09:11:14 configure: error: iconv not found, in order to build xmlrpc you need the iconv library Jan 09 09:11:21 may i request once more an option in bitbake to re-try the compilation if (and only if) network acess fails while retrieving sources ? Jan 09 09:20:54 RP: the postinst calls fc-cache is I remeber rightly Jan 09 09:22:39 RP: but thats been commented out in ttf.inc now so maybe they dont anymore Jan 09 09:23:29 can someone help me with my libiconv problem? Jan 09 09:23:42 zecke: hey Jan 09 09:23:54 XorA: The question is what should correctly depend on fontconfig-utils (and how it should do so given the naming issues) Jan 09 09:24:34 Currently, task-mythfront and ttf-bitstream-vera do... Jan 09 09:25:34 task-mythfront should probably be allowed to in some form but debian.bbclass prevents this. I'm not so sure ttf-bitstream-vera should be doing this... Jan 09 09:27:22 RP: the fonts in ttf-fonts do not, hrw is maintainer and has specifically removed those lines Jan 09 09:28:25 * mithro ponders Jan 09 09:28:38 should I try and catch the 7:15 train tommorrow morning Jan 09 09:32:35 <[lala]> RP: i had now two times the white screen after switching vt (from gpe to console) Jan 09 09:32:51 <[lala]> using raw keyboard Jan 09 09:34:11 RP: what do we need to talk me into mirroring? Jan 09 09:36:25 holy treke! Jan 09 09:38:11 treke: glibc broke their cvs so the commonly used libc cvs 20050627 (or whatever the date is) is corrupt if pulled from them. If oemirrors had a good copy on, that would be useful Jan 09 09:39:07 treke: http://ewi546.ewi.utwente.nl/OE/source/libc_sources.redhat.com__20050627.tar.gz Jan 09 09:39:11 ah Jan 09 09:39:35 the mirror script hasnt been running since my drive has long run out of space, and I've got no clue what is safe to delete :/ Jan 09 09:40:05 treke: I can imagine it being difficult to keep track of :-/ Jan 09 09:40:30 so there's a 100gb of source, and I imagine most of it is unneeded Jan 09 09:40:44 treke: I know I have archives here going back "quite a while" and keep meaning to try and clear out but don't want to lose anything valuable... Jan 09 09:41:27 treke: At release times for familiar/openzaurus, we really need to generate a list of the sources the releases images used... Jan 09 09:42:06 there's got to be some sane way of doing this Jan 09 09:42:35 [lala]: That's interesting to know. Looks like its something is wrong with the mode changes in pxafb... Jan 09 09:42:42 Something in the setvar function... Jan 09 09:43:18 treke: I think someone added a sourcespkg to oe that kept some kind of record of the sources used to upload with the images Jan 09 09:43:25 cant the final release build be done with an empty sources dir, but CVS_TARBALL_STASH and the like set to a source archive, then the sources directory will end up with sources only in the release, but wont be slowed by downloading them all again Jan 09 09:43:28 (as a .bbclass) Jan 09 09:44:15 maybe I can do some log munging and figure out what source hasnt ever been requested from oesources.org Jan 09 09:44:49 good idea Jan 09 09:44:53 treke: With some of the cvs dates not being locked down in OE, that's only going to work for some packages Jan 09 09:45:35 RP: yeah, but it should knock out the the ancient CVS stuff Jan 09 09:45:44 which is the only big problem Jan 09 09:45:49 hi can anyone help me i am trying to compile php for openslug i started this command : bb php Jan 09 09:46:09 can someone help me with my libiconv problem? Jan 09 09:46:48 having 20 copies of modutils isn't too tough. having 300 copies of opie cvs is more so :) Jan 09 09:47:32 treke: Most things should be ok - it'll be the x libs and tslib you'll see a lot of hits for Jan 09 10:05:06 Why does the subpackage metadata live in $WORKDIR rather than somewhere in staging or deploy (with the packages it belongs to)? Jan 09 10:17:45 ok. someone should be heading to where oesources.org is shortly Jan 09 10:38:04 pb_: feel like pulling 608 rev's using darcs? Jan 09 10:38:13 zecke: always Jan 09 10:38:19 let me rush to my command prompt Jan 09 10:38:40 * pb__ darcs pull Jan 09 10:39:14 heh, does it really need to give me the "EXTREME CARE" thing every time? Jan 09 10:39:53 pb__: I could place viagra advert there as well? Jan 09 10:40:04 pb__: or spread my hate word on GPE? Jan 09 10:40:08 whatever you prefer Jan 09 10:40:16 darcs: getCurrentDirectory: resource exhausted (Too many open files) Jan 09 10:40:16 both of those sound good Jan 09 10:40:35 pb__: darcs pull's the motd Jan 09 10:41:36 right Jan 09 10:41:46 it seems to be thinking at the moment. must be my pesky slow network connection again Jan 09 10:42:01 pb__: well it failed applying 608 patches here :( Jan 09 10:42:09 :-( Jan 09 10:43:35 good evening Jan 09 10:43:56 ah, here we go Jan 09 10:43:58 875 patches for me Jan 09 10:44:07 I'll say "a" Jan 09 10:44:31 pb__: I think no one used darcs on a project as big as OE Jan 09 10:44:36 doh Jan 09 10:44:38 pb__: it could apply one patch a time Jan 09 10:45:06 yeah. I wonder what would happen if I just ask it to apply some random selection of patches and not the others. Jan 09 10:46:20 crumbs, now it's using 550MB of ram Jan 09 10:47:17 that sucks Jan 09 10:47:31 pb__: http://bugs.darcs.net/issue48 sounds nice as well Jan 09 10:47:55 ade|desk: yah; just another 100MB and it would be using nearly as much as evolution. Jan 09 10:48:10 hehe Jan 09 10:48:19 or older bb Jan 09 10:48:46 * pb__ waits for firefox to page back in Jan 09 10:49:26 I guess I need to get a better computer if I'm going to use darcs. Jan 09 10:50:00 zecke: mm, yeah, that issue does sound hilarious Jan 09 10:50:13 "This is a result of the new improved and more efficient code for applying Jan 09 10:50:13 patches." Jan 09 10:50:20 eleet. I wonder what other improvements they have planned. Jan 09 10:50:36 pb__: well. Monotone might be slow. but the used crypto code is secure! Jan 09 10:51:03 indeedy Jan 09 10:51:28 I've been cursing evolution and its memory usage. Keeps locking this machine up :-/ Jan 09 10:52:10 zecke: in rev 332, bin/bitbake has the new buildAllDeps parameter, but in lib/bb/shell.py it's called with just one parameter. Jan 09 10:52:13 zecke: true, but the crypto wouldn't help much in that case. it would just mean that darcs took longer to apply mickeyl's patches against my home directory. Jan 09 10:52:58 RP: that does sound bad. I don't think I've seen any lockups as such; it just seems to sit there consuming 650MB or so. Jan 09 10:53:07 pb__: right, it should have a False/True there ;) Jan 09 10:53:30 zecke: tell that to ph5 Jan 09 10:53:30 RP: maybe try evolution-win32 on a decent OS? Jan 09 10:53:47 dear tab completion, you suck! Jan 09 10:53:49 pb__: In my case the memory used can start to eat into swap and the system near enough locks up Jan 09 10:54:24 so yeah Jan 09 10:54:27 it was iptables Jan 09 10:54:57 zecke, pH5: It should really do a getVar(BUILD_ALL_DEPS) Jan 09 10:55:32 RP: pH5 : or that Jan 09 10:56:29 zecke: darcs is still thinking. I guess I need a faster cpu as well as more ram. Jan 09 10:56:39 9976 pb 25 0 570m 480m 1996 R 97.2 47.5 10:49.48 darcs Jan 09 10:56:45 pb__: maybe a darcs co-processor Jan 09 10:57:16 good idea Jan 09 10:57:24 I wonder what it can be doing that takes so long Jan 09 10:57:37 diff -- /home/pb/owned\n+++ /home/pb/owned Jan 09 10:57:49 ah yes Jan 09 10:58:00 drat Jan 09 10:58:11 * pb__ 0wn3d Jan 09 10:58:37 * RP -> food Jan 09 11:06:22 pb__: holding 'y' and then pressing 'd' after 100 patches is a nice workaround... Jan 09 11:13:29 pb__: it took 398 patches here ;) Jan 09 11:16:29 I dont feel quite so dumb any more Jan 09 11:17:00 treke: ? Jan 09 11:17:12 The problem was with a openvpn/shorewall problem. Shorewall wasnt properly coming up since the vpn interface wasn't active Jan 09 11:17:23 but the vpn doesnt come on until much later in the boot process Jan 09 11:17:52 This is just the first time I'd rebooted since installing shorewall :) Jan 09 11:17:56 heh Jan 09 11:23:51 re Jan 09 11:24:35 zecke: seems that it worked Jan 09 11:25:18 the latest change I have now is: Jan 09 11:25:19 Sun Nov 13 20:45:47 GMT 2005 koen@openembedded.org Jan 09 11:25:19 * [project @ 81893f6385397851b77ac4dab982e8f0ae5a9198] Jan 09 11:25:19 qof: add qof 0.6.0-pre1 Jan 09 11:25:49 pb__: just two more month Jan 09 11:27:52 last RP changes allow to find interesting bugs in metadata.. Jan 09 11:28:37 zecke: right Jan 09 11:29:52 RP: ping Jan 09 11:29:59 ERROR: /home/hrw/devel/oe/org.openembedded/packages/opie-security/opie-security_cvs.bb depends on itself (eventually) Jan 09 11:30:02 ERROR: upwards chain is: opie-image (opie-image) -> meta-opie (meta-opie) -> opie-securityplugin-blueping (opie-securityplugin-blueping) -> opie-security (opie-security) -> opie-securityplugin-pin (opie-securityplugin-pin) Jan 09 11:30:18 RP: opie-security RRECOMMENDS opie-securityplugin-pin which RDEPENDS on opie-security Jan 09 11:31:01 RP: sugestions? Jan 09 11:31:31 that's not a metadata bug Jan 09 11:31:45 pb__: that is not - I had other before it Jan 09 11:34:29 I got another one: Jan 09 11:34:36 ERROR: dependency bluez-utils (for opie-securityplugin-blueping) not satisfied Jan 09 11:34:47 but bluez-utils-nodbus got build and provide bluez-utils Jan 09 11:40:26 pb__: you can do cg-update as well if it pleases you?! Jan 09 11:41:27 zecke: it delights me, but my git tree is at home Jan 09 11:41:40 so it will have to wait 30 minutes or so until I, also, am at home Jan 09 11:41:57 pb__: you want to stop working that early? Jan 09 11:42:22 ok. few things left Jan 09 11:43:03 zecke: yeah, I have been here for some hours already Jan 09 11:43:25 zecke, RP: http://bugs.openembedded.org/show_bug.cgi?id=582 Jan 09 11:44:10 pH5: s/0/False please Jan 09 11:44:18 pH5: and why do you want that var not to be expanded? Jan 09 11:46:08 zecke: getVar is called with 1 instead of True in many places. should this variable be expanded? Jan 09 11:47:15 pH5: it should be convertible to boolean, "${SOME_OTHER_VAR}" doesn't look pretty bool Jan 09 11:50:07 zecke: either a package is a task- package that needs all its depends to be built, or it's not. i can't imagine any case where expansion would be useful. Jan 09 11:51:04 opie-sh fixed Jan 09 11:52:20 fsck.. my favorite pizzeria ended delivery ;( Jan 09 11:52:22 pH5: still, there's no reason to inhibit expansion unless it would be actively damaging Jan 09 11:52:37 ok, so I'll s/0/True/ Jan 09 11:52:47 actually, (iirc) you can just omit that argument Jan 09 11:52:53 I'm pretty sure true is the default Jan 09 11:53:10 there is some default... ;) Jan 09 11:53:41 ah, no, false is still the default Jan 09 11:53:58 we did talk about changing it to true, way back when, but I guess we were too cowardly in the end Jan 09 11:54:03 03hrw 07org.oe.dev * r5f32961e... 10/packages/opie-sh/opie-sh.inc: opie-sh: fix RDEPENDS self dependency Jan 09 11:54:19 * pb__ go home now Jan 09 11:55:07 * RP returns Jan 09 11:55:12 RP: wb Jan 09 11:55:37 RP: was it really necessary to kill pb_ for your return? Jan 09 11:55:58 RP: any ideas of my bugs? Jan 09 11:56:08 hrw: Have you any idea whether ttf-bitstream-vera should be depending on fontconfig-utils? Jan 09 11:56:33 hrw: Yes, I have some patches locally which try to fix at least the blueutils problem Jan 09 11:56:53 hrw: But I see others with the patch which is why I haven't pushed anything yet... Jan 09 11:57:20 RP: ttf fonts... basically they need fontconfig-utils for X11 enviroments and opie-ttf-support for opie Jan 09 11:57:40 zecke: hi still there? can you help me with a little problem? i trying to compile php5 for openslug on my cross development system and always get this error :checking for iconv... no Jan 09 11:57:40 checking for libiconv... no Jan 09 11:57:40 configure: error: iconv not found, in order to build xmlrpc you need the iconv library Jan 09 11:57:54 hrw: So something in gpe should be going this rather than the ttf-bitstream file itself? Jan 09 11:58:34 kami22: no idea. It can be a bug in our package database Jan 09 11:58:43 kami22: look at config.log and see if something failed? Jan 09 11:58:48 RP: ttf fonts postinst/postrm should call 'fc-cache' for fontconfig and 'update-ttffontdir' for opie Jan 09 11:59:03 kami22: glibc or uclibc? Jan 09 11:59:19 how can i find this out? Jan 09 11:59:21 hrw: and fontconfig-utils just does this? Jan 09 11:59:38 RP: no - it contain fc-cache Jan 09 11:59:53 hrw: Another thing I could use your insight for - the DEPENDS in meta-opie shouldn't be needed but I don't know enough about opie to be sure removing them is safe... Jan 09 11:59:58 kami22: cd tmp/work/php/?? Jan 09 12:00:03 kami22: and then look in config.log Jan 09 12:00:11 RP: thats why I'm looking at meta-opie now Jan 09 12:00:33 hrw: Right, so that rdepend is totally wrong. Just a question of what should be doing that Jan 09 12:00:42 okay wwait a second Jan 09 12:01:01 sqlite is screwed up Jan 09 12:01:19 hrw: I'll make a diff of my current bitbake patches - they'll mean you get slightly less confusing build messages Jan 09 12:01:26 ok Jan 09 12:03:30 hrw: http://sam.rpsys.net/bitbake-rproviders-r0.patch Jan 09 12:03:45 hrw: Keep in mind this is even more experimental ;-) Jan 09 12:04:11 RP: why sam? Jan 09 12:04:30 hrw: cant find anything like glibc or uclibc in the config.log Jan 09 12:04:50 kami22: did you find 'that' config.log? Jan 09 12:05:02 kami22: iconv failed, so grep for that in the log Jan 09 12:05:15 kami22: You will need to find out why the test failed Jan 09 12:05:32 configure:53616: checking for iconv support Jan 09 12:05:32 configure:53655: result: no Jan 09 12:05:35 zecke: Experiemental bitbake is running on sam so sharing the patches there was easier Jan 09 12:05:36 kami22: as a heavy Qt addict I do not use iconv for converting all kind of stuff.. Jan 09 12:05:55 RP: why did you name sam sam was my initial question Jan 09 12:05:57 * france is back (gone 03:14:21) Jan 09 12:06:37 hrw: http://sam.rpsys.net/oe-temp.patch are my current differences with OE. Most of these changes are known to be wrong but do illustrate where I know the problems are Jan 09 12:07:24 zecke: I like machine with three letter names. sam is a handy three letter name which also happens to have other interesting meanings... Jan 09 12:07:44 zecke: Some of my other machines are tim, max and dax :) Jan 09 12:07:45 but what i can't understand the configure command says without-iconv and why did xmlrpc need iconv? Jan 09 12:08:21 ~lart linux Jan 09 12:08:21 * ibot turns linux into a lifesized tux doll Jan 09 12:08:30 hrw: http://sam.rpsys.net/oe-temp.patch are my current differences with OE. Most of these changes are known to be wrong but do illustrate where I know the problems are Jan 09 12:08:30 ~lart kde^h^honsole Jan 09 12:08:31 * ibot hurls dozens of incontinent, insomniac, hungry kittens with tiny little razor-sharp claws and a wide variety of contagious intestinal parasites at kde^h^honsole Jan 09 12:08:33 kami22: dunno. look at configure for what iconv test it runs and then look at config.log for more information Jan 09 12:08:42 I hate linux Jan 09 12:08:55 hrw: try GNU/Linux instead Jan 09 12:09:03 fscking kernel which is not able to report WHY it die Jan 09 12:09:21 hrw hm Jan 09 12:10:33 I use it, then it die.. and no way to get that info.. I just want to boot again and get something like 'oops.. you 6th usb device tried to get cpu off from mainboard' Jan 09 12:12:19 and then I again have to wait 2 minutes for loading kde sessin ;( Jan 09 12:13:56 hrw is it your desktop pc? Jan 09 12:14:30 woglinde: yes - plain x86 box which run some daemons also for my local needs Jan 09 12:14:45 sorry i cant find the error :( Jan 09 12:15:01 i dont understand what configure is testing Jan 09 12:15:16 kami no config.log? Jan 09 12:15:46 i have got a config.log but i don't understand what i have to look vor Jan 09 12:16:10 RP: could you give me url to previous patch? Jan 09 12:16:17 03hrw 07org.oe.dev * r45134ada... 10/packages/opie-examples-python/ (3 files): Jan 09 12:16:17 opie-examples-python: few fixes Jan 09 12:16:17 - corrected dependency on opie-pyquicklaunchapplet Jan 09 12:16:17 - set PACKAGE_ARCH to all as it contain only Python script Jan 09 12:16:17 - there is only one example - do not package .desktop files for other Jan 09 12:16:37 kami look at the last test should near before the variable statements begins Jan 09 12:16:55 RP: s/svn-native/subversion-native/ Jan 09 12:19:17 kami soemthing like http://pastebin.ca/36279 Jan 09 12:22:32 RP: ERROR: upwards chain is: opie-image (opie-image) -> meta-opie (meta-opie) -> opie-securityplugin-blueping (opie-securityplugin-blueping) -> opie-security (opie-security) -> opie-securityplugin-pin (opie-securityplugin-pin) Jan 09 12:23:01 sqlite has cross-compile bugs Jan 09 12:23:15 RP: any idea how to make it working again? I could remove RDEPENDS opie-securityplugin-pin from opie-security but woule prefer not have to Jan 09 12:23:41 Luke-Jr: works for me (host:x86, target: x86, arm) (host: x86-64, target: arm) Jan 09 12:24:05 hrw: your host and target pointer sizes match Jan 09 12:24:19 I suggest testing the x86_64-built one again Jan 09 12:24:49 sqlite compiles and runs a program to determine its sizeof(char*) Jan 09 12:25:11 does not take it from site/* cache? Jan 09 12:25:27 site/*? Jan 09 12:26:02 zecke: now I update my git tree Jan 09 12:26:04 openembedded/site/* - config.cache for autotools Jan 09 12:26:17 Luke-Jr: NOTE: package sqlite3-3.2.7-r0: task do_build: completed Jan 09 12:26:22 hrw: it doesn't use autotools Jan 09 12:26:24 03jbowler 07org.oe.dev * r85da89c6... 10/packages/ (16 files in 12 dirs): (log message trimmed) Jan 09 12:26:24 slugos: remove/correct references to openslug Jan 09 12:26:24 - apex: contains an absolute build sys path name (commented) Jan 09 12:26:24 - base-files: rename openslug to slugos in 3.0.14 Jan 09 12:26:24 - busybox: rename openslug to slugos in 1.00, 1.01 Jan 09 12:26:25 - netbase: rename openslug to slugos in 4.21 Jan 09 12:26:27 - slugos-init: rename openslug to slugos in 0.10 Jan 09 12:26:28 hrw: compile is ok, but runtime is not Jan 09 12:26:29 03jbowler 07org.oe.dev * r93b8d4a0... 10/conf/ (distro/slugos.conf machine/nslu2.conf): slugos: update floating CVS packages and ipkg to latest versions in conf Jan 09 12:26:35 03jbowler 07org.oe.dev * rd5130c6d... 10/ (586 files in 358 dirs): Jan 09 12:26:35 explicit_merge of '85da89c673cc2faec29d5ea692e6e7638f4c1b5e' Jan 09 12:26:36 and '5f32961e5d0f4132f4513cbeac7fe59dc5ae5315' Jan 09 12:26:38 using ancestor '' Jan 09 12:26:40 to branch 'org.openembedded.dev' Jan 09 12:26:41 Luke-Jr: ah.. ok Jan 09 12:26:42 03jbowler 07org.oe.dev * r77ee5dc3... 10/ (570 files in 349 dirs): Jan 09 12:26:44 explicit_merge of '93b8d4a095a60dfe2b4b6e82f4ce68721a1d4f11' Jan 09 12:26:46 and 'd5130c6d17c31fa91dee51b0074c7e5e8c87717c' Jan 09 12:26:48 using ancestor '' Jan 09 12:26:50 to branch 'org.openembedded.dev' Jan 09 12:26:58 hrw: it fails an assert at runtime Jan 09 12:27:01 Luke-Jr: do you compile for amd64? or on amd64? Jan 09 12:27:08 amd64 host, arm target Jan 09 12:27:14 see sqlite/main.mk Jan 09 12:27:17 config.h rules Jan 09 12:27:21 Why would it run at runtime? Jan 09 12:27:24 Thats stupid Jan 09 12:27:24 Luke-Jr: ah.. Jan 09 12:27:34 CosmicPenguin: sqlite is stupid and stores pointers in an int format Jan 09 12:27:36 Luke-Jr: well check configure.log? Jan 09 12:27:45 CosmicPenguin: indeed, running is the last thing you would expect at runtime Jan 09 12:27:48 Luke-Jr: and we apply a special sizeof patch to autoconf Jan 09 12:27:49 zecke: there is no log Jan 09 12:27:56 zecke: again, it's not autoconf Jan 09 12:28:00 CosmicPenguin: I guess people get confused by the name or something Jan 09 12:28:04 Luke-Jr: then it is not sqlite3? Jan 09 12:28:11 zecke: 2 Jan 09 12:28:12 speaking about sqlite... new versions are available Jan 09 12:28:24 hrw: I need to update the packages Jan 09 12:28:29 * zecke needs to get his shark running Jan 09 12:28:40 pb_: heh Jan 09 12:29:07 hrw: Did you get the url? Jan 09 12:29:11 CosmicPenguin: hey Jan 09 12:29:20 lrg|home: hows it going? Jan 09 12:29:35 RP: yes - applied but opie-security loops Jan 09 12:29:46 CosmicPenguin: not bad. thanks. I have a db1500 in front of me. Jan 09 12:29:52 RP: ERROR: upwards chain is: opie-image (opie-image) -> meta-opie (meta-opie) -> opie-securityplugin-blueping (opie-securityplugin-blueping) -> opie-security (opie-security) -> opie-securityplugin-pin (opie-securityplugin-pin) Jan 09 12:30:00 maybe I should reduce the size of the MTU Jan 09 12:30:16 hrw: I'm thinking about this... Jan 09 12:30:26 RP: I use 'bitbake -nv opie-image' for this Jan 09 12:31:28 CosmicPenguin: I've made a start to the pb1500 (old alchemy) platform layer during Christams, but I've been caught up lately with some support and other tasks Jan 09 12:31:47 hrw: Good idea. I'm been running actual builds as they highlight more issues but that's good for just testing the dependencies Jan 09 12:31:55 CosmicPenguin: I'm hoping to have something going by the end of January Jan 09 12:32:11 zecke: hm, this can't be good: Jan 09 12:32:20 error: Failed connect to ewi546.ewi.utwente.nl:80; No such file or directory (curl_result = 7, http_code = 0, sha1 = 46c37676b71ae663d916ae2e3b5bb27e19f8c518) Jan 09 12:32:25 lrg|home: cool. I'm not sure I'll be able to get to anything soon, but you have my full support if you need it Jan 09 12:32:31 woglinde: so here is the libiconv error : http://pastebin.ca/36281 Jan 09 12:32:32 zecke: I try again Jan 09 12:32:32 pb_: try again Jan 09 12:32:42 pb_: hehe, It might not have written that file yet Jan 09 12:32:43 CosmicPenguin: thanks and appreciated :) Jan 09 12:32:49 lrg|home: unfortunately, I probably can't get you a DB1200 - the last run was completely snapped up by potential customers Jan 09 12:32:51 zecke: oh, heh Jan 09 12:32:52 *I* don't even have one Jan 09 12:32:56 pb_: org.openembedded.dev is a symlink Jan 09 12:33:28 kami yes there it is -> cannot find -lexpat Jan 09 12:33:31 hrw: I think the answer might be just to remove it from the DEPENDS in meta-opie Jan 09 12:33:38 and now? Jan 09 12:33:54 CosmicPenguin: they must be in short supply, however I did see a db1200 based media player on someones desk lately Jan 09 12:33:56 RP: I will do it Jan 09 12:34:05 kami skipping incompatible /usr/lib/libexpat.so build a libexpat for your target Jan 09 12:34:18 zecke: okay, it seems to be running along happily enough at the moment Jan 09 12:34:23 * pb_ goes looking for an ipaq power brick Jan 09 12:34:31 hrw: I suspect all of those DEPENDS are actually wrong and if our metadata is correct actually unneeded... Jan 09 12:34:49 pb_: try cg-update as well Jan 09 12:34:59 CosmicPenguin: I think they would be annoyed if I re flashed their media player ;) Jan 09 12:35:39 RP: I will work on them - first want to get working meta-opie Jan 09 12:36:25 hrw: I'd try commenting out that whole DEPENDS and see what breaks Jan 09 12:37:04 RP: probably meta-opie and then opie-*image Jan 09 12:37:06 so does someone w/ commit access want to fix sqlite 2? Jan 09 12:37:24 woglinde: but where can i find libexpat.so? Jan 09 12:37:46 echo '#define SQLITE_PTR_SZ 4' >$WORKDIR/sqlite-2.8.16/config.h <-- before make Jan 09 12:37:49 kami22: in STAGING_LIBDIR? Jan 09 12:37:54 kami22 you have to build it Jan 09 12:38:07 Luke-Jr: each arch has SQLITE_PTR_SZ 4? Jan 09 12:38:25 kami22 maybee the deps for libiconv are incorrect Jan 09 12:38:35 hrw: nah, and echo probably isn't the best way, but that gets the idea there Jan 09 12:38:49 hrw: With the DEPENDS commented out, meta-opie should just work - it will just generate a set of packages with the appropriate RDEPENDS Jan 09 12:38:51 hrw: I'm sure those w/ commit access also know the proper way to determine the # Jan 09 12:39:08 okay i now that i have to build but i dont know to what package lexpat ? Jan 09 12:39:18 Luke-Jr: I have r/w but I'm not able ;( Jan 09 12:39:37 RP: ok. now I'm building meta-opie* Jan 09 12:39:38 kami22 maybee expat? Jan 09 12:39:38 hrw: hrm, I don't have a clue unfortunately Jan 09 12:39:48 good idea :) Jan 09 12:39:49 zecke: roger, will try that next Jan 09 12:40:08 RP: solved that opie-security thing Jan 09 12:40:15 Is any issue known that 2.6.14.4 will not work on StrongARM? Jan 09 12:40:29 hrw: Keep in mind that building meta-opie won't actually build all the dependencies now. Jan 09 12:40:43 RP: ok - but it build something Jan 09 12:41:02 We really need to add a "set BUILD_ALL_DEPS" option to the bitbake commandline... Jan 09 12:42:04 as in "BUILD_ALL_DEPS=1 bitbake opie-opie-image"? Jan 09 12:42:19 pb_: image_ipk.bbclass sets that Jan 09 12:42:20 pb_: master I would like to abuse you now! Jan 09 12:42:52 RP: ah. well, s/opie-opie-image/meta-opie-opie-opie/ or whatever. Jan 09 12:42:59 zecke: I am at your disposal Jan 09 12:43:22 pb_: :) (but yes, you're right, we can do that already) Jan 09 12:43:37 RP: should I need to do anything to get sound working on c760? Jan 09 12:43:54 Luke-Jr: unmute something in alsamixer I expect Jan 09 12:44:38 RP: you know the driver has Left/Right as seperate volumes, right? Jan 09 12:45:28 Luke-Jr: yes, I had noticed that when I wrote the codec driver... Jan 09 12:45:31 can someone kick monotone on vanille? Jan 09 12:45:45 RP: purposely? Jan 09 12:45:51 Luke-Jr: I'm planning on merging L+R controls in ASoC soon. Jan 09 12:46:14 Luke-Jr: They are independent atm because they stride 2 registers Jan 09 12:46:18 lrg|home: Why? Jan 09 12:46:28 ah, I thought RP wrote it ;) Jan 09 12:46:37 * Luke-Jr is failing to get any audio out :\ Jan 09 12:46:46 hrw: I think zecke has rooted vanille.de in order to use it for his darcs trial Jan 09 12:46:53 lrg|home: That would be implementing policy in kernel space which I'd really like to avoid Jan 09 12:47:14 lrg|home: It should be userspace that turns them into one volume control Jan 09 12:47:24 ah. ok Jan 09 12:47:34 RP: alsamixer still allows independent L/R volumes control from a stereo mixer Jan 09 12:47:45 RP: it's all done in alsa-lib :) Jan 09 12:47:47 RP: ALSA supports left/right modifications to a single volume control Jan 09 12:48:06 Luke-Jr: I wrote the codec, lrg|home wrote ASoC which is the framework the codec uses Jan 09 12:48:23 RP: it's just I've not implemented it yet in the core - been lazy :) Jan 09 12:48:54 Luke-Jr: all my fault I' afraid - RP is blameless Jan 09 12:49:17 * RP hides from the scary ALSA internals ;-) Jan 09 12:49:27 I'm assuming I'm not torturing your code by trying 'cat /dev/urandom>/dev/dsp'? Jan 09 12:49:41 * lrg|home hides from scary bitbake internals :) Jan 09 12:49:42 that's usually how I test audio ;) Jan 09 12:49:55 03hrw 07org.oe.dev * ra1193f39... 10/packages/ (4 files in 2 dirs): Jan 09 12:49:56 opie-security: removed dependency loop Jan 09 12:49:56 - moved opie-securityplugin-pin from opie-security to meta-opie/task-opie-settings Jan 09 12:49:58 Luke-Jr: no Jan 09 12:50:03 hrm Jan 09 12:50:11 afk Jan 09 12:50:11 * Luke-Jr ponders why he can't get any audio Jan 09 12:50:51 Luke-Jr: what player are you using ? Jan 09 12:50:57 how does one create an initrd again... and how do I get it compiled into the kernel :} Jan 09 12:51:05 Luke-Jr: aplay, madplayer both work ok Jan 09 12:51:06 lrg: cat Jan 09 12:51:16 Luke-Jr: using OSS emulation Jan 09 12:51:32 zecke use "the source(kernel) luke" Jan 09 12:51:42 woglinde: so configure is now not a problem the expat was missing but now i am getting for compiling this error : Jan 09 12:51:51 lrg: yes Jan 09 12:52:03 woglinde: pb_ : http://page.mi.fu-berlin.de/~freyther/shark_pcap Jan 09 12:52:08 Luke-Jr: can you cat audio and then at the same time cat /sys/devices/platform/soc-audio/dpm Jan 09 12:52:21 /home/kami/slug/openslug/tmp/cross/lib/gcc/armeb-linux/3.4.4/../../../../armeb-linux/bin/ld: skipping incompatible /usr/lib/libcrypt.a when searching for -lcrypt Jan 09 12:52:21 /usr/lib/libexpat.so: could not read symbols: File in wrong format Jan 09 12:52:22 collect2: ld returned 1 exit status Jan 09 12:52:25 woglinde: pb_ : could you please look at that dumb, shake your magic ball and tell me what is wrong with my nfs ;) Jan 09 12:52:37 kami22: /usr/lib/libexpat sounds wrong ;) Jan 09 12:53:03 http://pastebin.com/498211 Jan 09 12:53:04 pH5: did you update that patch for bitbake? Jan 09 12:53:09 lrg: http://pastebin.com/498211 Jan 09 12:54:06 aha... unmuting "Output Mixer HiFi Playback Swit" helped Jan 09 12:54:11 whatever that means o.o Jan 09 12:54:33 kami22 your buildsystem still using the libexpat from your system which will be wrong for your target Jan 09 12:54:52 bipolar: you can bypass problem, remove libcre stuff from Makefile.in line 228, and LIBCRE from MakeFile in konqueror-embedded-snapshot-20030705/konq-embed/src Jan 09 12:54:54 Luke-Jr: ah, it means your DAC output is played into the output mixer and is a good thing :) Jan 09 12:54:58 ops Jan 09 12:54:59 kami22 and same as with libexpat for now is libcrypt Jan 09 12:55:07 lrg: DAC output? Jan 09 12:55:19 aquadran: or install libexpad-dev on your system :} Jan 09 12:55:26 Luke-Jr: Playback audio Jan 09 12:55:48 zecke: that not solved problem. Jan 09 12:55:51 lrg: ok, so I'm guessing you'll tie that in with a volume control at some point ;) Jan 09 12:55:56 like w/ Left/Right? Jan 09 12:56:04 right Jan 09 12:56:22 zecke: try turning down the rsize on your nfs mount Jan 09 12:56:25 set it to 1024 or something Jan 09 12:56:28 zecke: http://bugs.openembedded.org/attachment.cgi?id=465 Jan 09 12:56:38 hey zecke Jan 09 12:57:03 Luke-Jr: it depends on the path. If the control is after the mixer then yes, else no. Jan 09 12:57:08 pb_: (no kidding) what is rsize? (I will do man exports) Jan 09 12:57:12 mithro: darcs lost ;) Jan 09 12:57:12 how would I View the Capture part in alsamixer? Jan 09 12:57:21 Luke-Jr: can you alsactl store Jan 09 12:57:24 mithro: it fails with applying +398 patches Jan 09 12:57:24 zecke: read block size. you need to set it on the client, not the server. Jan 09 12:57:26 Luke-Jr: press tab Jan 09 12:57:30 zecke: switching fast between windows and up+enter buttons is not good :) that was old message Jan 09 12:57:34 aha Jan 09 12:57:58 zecke use nfs version 3 with aply ,v3 to the nfs-options Jan 09 12:58:08 Luke-Jr: alsact can be used to save and restore your mixer settings Jan 09 12:58:17 woglinde: but how to change this ?? Jan 09 12:58:21 s/alsact/alsactl Jan 09 12:58:21 lrg: no initscripts? ;) Jan 09 12:58:38 i build expat in cross development Jan 09 12:58:40 kami22 build libcrypt for your target Jan 09 12:58:44 Luke-Jr: not yet, we've only jus recently started audio Jan 09 12:58:59 kami22 dont know where libcrypt is in oe Jan 09 12:59:00 zecke: see Documentation/nfsroot.txt, but ignore the bit that claims the default is 1024. this is all lies; the default is, in fact, 4096. Jan 09 12:59:36 zecke maybee you can try nfs v4 but I didnt test that for sure Jan 09 12:59:58 mhh the 'new' oe have move all the work dir wrom tmp/work to tmp//work ? Jan 09 12:59:59 sorry i think i am to stupid for this so is libexpat install right and is the error in lcrypt? Jan 09 13:00:08 can i delete the old one ? Jan 09 13:00:17 zecke: did you mean libexpact on host ? Jan 09 13:00:58 aquadran: sorry I meant libpcre-dev ;) Jan 09 13:01:00 lrg: any ideas on mic? cat /dev/dsp does nothing Jan 09 13:01:43 Luke-Jr: I suggest installing aplay and arecord. This will allow better control of the audio capture/playback Jan 09 13:02:04 zecke: i npast i tried libpre-native, but it doesn't work. but i not tried simply install on host OS Jan 09 13:02:09 Luke-Jr: you can then specify the capture rate, format and channels Jan 09 13:02:33 http://openzaurus.org/wordpress/howto/c7x0-and-cxx00-26-kernel-information/ now contains a few snippets of alsa information Jan 09 13:02:49 failed? Jan 09 13:02:54 As this is going to become a FAQ Jan 09 13:02:58 zecke: that package is not used anyway Jan 09 13:03:13 zecke: for compilation Jan 09 13:03:32 lrg: any idea what pkg? Jan 09 13:03:35 zecke: i would like to see the output etc Jan 09 13:03:45 just out of intrest Jan 09 13:03:46 ~hail RP for an audio FAQ Jan 09 13:03:51 I guess L/R at 100% is full volume, isn't it? ;) Jan 09 13:03:56 * ibot bows down to RP for an audio FAQ and chants, "I'M NOT WORTHY!!" Jan 09 13:04:14 Luke-Jr: yes 100% is full Jan 09 13:04:32 Luke-Jr: iirc, alsa-tools ipkg Jan 09 13:06:11 mithro: see bugs.darcs.net for the meaningless report Jan 09 13:06:21 mithro: basicly the haskell runtime refuses to open more files Jan 09 13:06:35 mithro: and there is a one month old security bug in darcs Jan 09 13:06:49 mithro: which would make me feel bad when doing darcs pull Jan 09 13:06:57 lrg: how nice... alsa-tools doesn't build =p Jan 09 13:07:21 Luke-Jr: oh, not good :( Jan 09 13:07:39 lrg: I set something to 'Headset' tho, maybe now Minisip will work Jan 09 13:07:53 lrg: any idea if I can use speakers and headset at once? Jan 09 13:08:20 Luke-Jr: headset will have it's own speakers. Jan 09 13:08:43 Luke-Jr: it's basically what people in call centres wear Jan 09 13:08:45 right... I mean I want to use the internal speaker at the same time as the audio jack Jan 09 13:08:56 is that technically possible? Jan 09 13:09:11 woglinde: sorry cant find anything how to fix the lcrypt problem? Jan 09 13:09:13 Luke-Jr: iirc there should be an option for spk auto, on, off Jan 09 13:09:22 Luke-Jr: set to on Jan 09 13:09:36 lrg: but there's only one audio out stream? Jan 09 13:09:49 so I can't play seperate audio on speaker vs lineout Jan 09 13:09:56 Luke-Jr: it goes to both headset and spk in this case Jan 09 13:10:25 Luke-Jr: ehm no, remember it's only a pda :) Jan 09 13:10:35 nah, it's a handheld Jan 09 13:11:10 kami22 hm libcrypt should be in glibc Jan 09 13:13:38 Luke-Jr: btw, I'm working on a chip that has 2 phy audio channels that will do this. It's about to go on a tier 1 reference board soon. Jan 09 13:13:46 pb_: the pcap output looks a bit better Jan 09 13:14:06 pb_: i will upload a new version if my test here fails Jan 09 13:14:24 yes but glibc is compiled Jan 09 13:14:55 lrg: hrm, is OSS emulation incomplete for parts of the driver? Jan 09 13:15:04 lrg: in particular, audio input and mixer Jan 09 13:15:11 kami22 than run find on your build/tmp/work and staging and look if there is al libcrypt Jan 09 13:15:55 kaim are u building ppp ? Jan 09 13:16:01 Luke-Jr: OSS emulation is not recommended over native alsa since it's mixers are aimed at PC's and not handhelds Jan 09 13:16:18 lrg: no difference =p Jan 09 13:16:30 lrg: the mixer appears to only have a mic volume, tho Jan 09 13:16:57 Luke-Jr: can you see any capture settings ? Jan 09 13:17:47 Luke-Jr: what about the mic capoture switch Jan 09 13:17:47 zecke: righto Jan 09 13:17:57 gremlin[it]: do you mean me? Jan 09 13:18:01 zecke: probably no need to upload the whole thing, just cut and paste the last 20 lines or so of output Jan 09 13:18:06 i am trying to build php5 Jan 09 13:18:18 Luke-Jr: and input select switch Jan 09 13:18:18 kaim22 yess sorry TAB don't complete :) Jan 09 13:18:30 pb_: I have increated my mtu again, set rsize to 1024 Jan 09 13:18:37 pb_: and I see that busybox get loaded Jan 09 13:18:39 zecke: very good Jan 09 13:18:55 zecke: yeah, the important thing is to have mtu > rsize Jan 09 13:18:56 pb_: I will toy some more... but I should start with my paper... Jan 09 13:19:06 lrg: alsamixer shows all that, but OSS-based mixers (GPE) only see mic volume Jan 09 13:19:25 if rsize > mtu, you can run into packet fragmentation problems. Jan 09 13:19:29 i found this : /home/kami/slug/openslug/tmp/staging/armeb-linux/lib/libcrypt.a libcrypt.so .. Jan 09 13:19:39 lrg: what is Input Mux? Jan 09 13:19:40 Luke-Jr: you need to use alsamixer in the meantime Jan 09 13:19:51 Luke-Jr: it's an input switch Jan 09 13:19:58 but what does it do? Jan 09 13:20:01 kami22 i had some problem with pppd cause it look 'statically' for libcrypt on /usr/lib/libcrypt.so ... but on a 64bit system it is on /usr/lib64/ ... but don't know if php5 have same problem ... usually happen when no autotools are used :( Jan 09 13:20:03 switching between mic and line Jan 09 13:20:13 err... there's only one audio jack... Jan 09 13:20:20 Luke-Jr: it sets your input src. Jan 09 13:20:33 there's only one source, tho... O.o Jan 09 13:20:39 Luke-Jr: you can plug a mic or line into your jack Jan 09 13:20:56 is there a difference? ;) Jan 09 13:21:10 but on my system it is /usr/lib not /usr/lib64 :( Jan 09 13:21:15 Luke-Jr: yes, the mic needs power, line in doesn't Jan 09 13:21:58 zecke: o well Jan 09 13:21:58 ops ... i read u are using an amd64 ... Jan 09 13:22:17 lrg: for mic, set Input Mux to Mic, Jack Function to Headset, and what else? Jan 09 13:22:22 Luke-Jr: although, 1 input might not be connected. It depends on the hw design Jan 09 13:22:38 Luke-Jr: Jack func to mic Jan 09 13:22:40 i am using and amd64 opteron but dont got a 64bit system Jan 09 13:23:32 * gremlin[it] ... i'm really one of the few :) ... Jan 09 13:23:47 lrg: err... but I want both in & out? Jan 09 13:24:25 Luke-Jr: are you using a mic or headset ? Jan 09 13:24:42 a headset Jan 09 13:24:54 for a cellphone, but w/ a size adapter Jan 09 13:25:07 Luke-Jr: ok, then jack to headset Jan 09 13:25:40 Luke-Jr: for debug you could enable the mic sidetone switch Jan 09 13:25:54 Luke-Jr: this will let you hear the mic from your spk Jan 09 13:26:13 it made crackling when I had that on Jan 09 13:27:19 Luke-Jr: that might be your mic gain too high Jan 09 13:27:35 turned off Mic Boost and now I hear nothing Jan 09 13:28:19 Is there supposed to be a Mic volume? Jan 09 13:28:22 so this is my error message for compiling php5 : http://pastebin.ca/36293 Jan 09 13:28:28 I only have a Capture flag, no volume Jan 09 13:28:49 Luke-Jr: L/R capture volume Jan 09 13:28:53 lrg|home: Keep in mind c7x0 input hasn't been tested by me (I can't remember if you did on the reference board or not?) Jan 09 13:29:23 RP: ah, it's not tested then. Jan 09 13:29:28 hnn' Jan 09 13:29:39 lrg: so that's not Line-only?\ Jan 09 13:29:40 Maybe I should bring my c760 home and try it :p Jan 09 13:30:03 * RP has no input devices to test it with! Jan 09 13:30:17 * lrg|home has no c760 to test with Jan 09 13:30:28 RP: eh, microphones are cheap Jan 09 13:30:51 * lrg|home is back at the office on Friday and will test on his reference board. Jan 09 13:31:05 Luke-Jr: That may be but I don't have one. Its not been a priority for me to find one... Jan 09 13:31:39 RP: wasn't aware one had to look to find a mic ;) Jan 09 13:31:49 I don't suppose I could help out on this end? Jan 09 13:32:57 Luke-Jr: I'll be able to test the codec code on Friday. Once I'm sure it's ok we should then go about testing the c760 Jan 09 13:33:21 In the meantime, I'll think about obtaining a mic :) Jan 09 13:33:35 * lrg|home will give RP a mic at fosdem Jan 09 13:33:40 mithro: re Jan 09 13:33:44 I probably have something I could use, just need an adapter... Jan 09 13:33:52 mithro: these two issues make darcs unsuited for OE Jan 09 13:34:14 * Luke-Jr thought mics generally had the right size jack for the Z Jan 09 13:34:39 Luke-Jr: they can come in 2.5 and 3.5mm jack varieties Jan 09 13:34:50 lrg|home: thanks :) I will try and find something in the meantime Jan 09 13:35:01 lrg: aren't the 2.5mm ones all headsets? Jan 09 13:35:08 Typically, I have one with a 2.5mm plug.... Jan 09 13:35:25 Luke-Jr: I'm not sure Jan 09 13:35:38 Luke-Jr: I've a hs with a 3.5 Jan 09 13:36:10 Well, as far as Walmart goes, headset = 2.5, mic = 3.5 Jan 09 13:36:32 I picked up a 2.5-to-3.5 adapter at RadioShack Jan 09 13:36:41 funny, the guy who sold it insisted it would not work ;) Jan 09 13:36:47 Luke-Jr: I suppose hs are typically used in phones atm Jan 09 13:36:50 (for a headset function) Jan 09 13:38:19 BTW, IPv6 appears to work on the kernel level Jan 09 13:38:30 just a lack of support in software (gpe-mini-browser seems to be v4 only) Jan 09 13:43:17 The next tricky question: Are valid circular RDEPENDS allowed? They upset ipkg so could be classed as broken packaging... Jan 09 13:45:19 They shouldn't upset ipkg unduly. Circular dependencies ought to be legal, yes. Jan 09 13:45:56 Luke-Jr: that is unfortunate. You might want to file a bug against gpe-mini-browser for that. Jan 09 13:46:11 there isn't really any excuse to be writing v4-only software in this day and age. Jan 09 13:53:12 pb_: hmm everything looks fine now but still no go Jan 09 13:53:22 zecke: suck Jan 09 13:53:29 nothing at all on the screen? Jan 09 13:53:42 pb_: Freeing init... Jan 09 13:53:51 doh Jan 09 13:53:59 pb_: and a blinking cursor, keyboard is working, I can ping the machine Jan 09 13:54:00 does your keyboard work? Jan 09 13:54:02 zecke hm maybe a library problem Jan 09 13:54:19 what happens if you do the printscreen thing, is init running? Jan 09 13:54:25 woglinde: I have tried a debootstrap and one uclibc rootfs from gentoo-netwinder Jan 09 13:54:43 pb_: which printscreen thing? Jan 09 13:54:51 zeck or maybe a console problem Jan 09 13:55:08 zecke: try pressing shift-prtscrn or ctrl-prtscrn or something Jan 09 13:55:15 one of those should display the running processes Jan 09 13:55:54 pb_: a sysrq or how that thing is named you mean? Jan 09 13:56:20 hold on Jan 09 13:56:24 * pb_ goes to find a keyboard to look at Jan 09 13:56:55 ah, maybe I meant scroll lock Jan 09 13:56:56 try that one Jan 09 13:57:04 I don't think it's sysrq. Jan 09 13:57:56 init == running Jan 09 13:58:30 you didn't forget to include floating point support in your kernel or anything like that? Jan 09 13:59:11 am I stupid or such ;) I'm trying a softfloat OE image (if I found one) Jan 09 13:59:18 aha Jan 09 13:59:21 pb_: I was so smart I rm -rf ed my kernel build tree Jan 09 13:59:24 hihi Jan 09 13:59:35 zecke: clever plan, always best to avoid wasting space Jan 09 13:59:51 pb_: but I have the config.gz in the kernel so I could extract it if needed... Jan 09 14:00:06 zecke: might be worth checking your binaries to make sure they really are softfloat Jan 09 14:00:16 so i will go now maybe tomorrow somebody can help me Jan 09 14:00:22 use arm-linux-objdump and scrutinize the output for "wfs" or similar instructions Jan 09 14:00:44 good night all Jan 09 14:01:00 Would there be agreement to moving package meta data such as the shlibs output to ${STAGING_DIR}/pkgmeta ? (call the variable something like STAGING_PKGMETADIR) Jan 09 14:01:38 well, shlibs would end up in $PKGMETADIR/shlibs... Jan 09 14:02:04 RP: what's the problem with the current shlibs location? Jan 09 14:02:25 that data is arch specific, so it needs to stay in a directory with ${HOST_SYS} in the name Jan 09 14:03:01 pb_: In theory you should be able to wipe the directories in tmp after building. Its more the package files I want around but it makes sense to move both shlibs and the .package Jan 09 14:03:36 I agree it needs HOST_SYS in there Jan 09 14:03:51 RP: did you mean "directories in work"? Jan 09 14:04:05 pb_: I did, yes. sorry Jan 09 14:04:24 ah, hm. last time I looked, shlibs was in staging, not in work. Jan 09 14:04:39 if someone moved it to work then yes, I agree it should be moved back to staging. Jan 09 14:04:47 some of shlibs is in staging, some of it isn't Jan 09 14:05:29 It could be argued the bits in tmp/work are expendable although I think if we plan to support proper cleaning at some point, having it all in staging might work better Jan 09 14:06:00 which bits are currently in work? Jan 09 14:07:38 pb_: The shlibdeps file. pclist files are in staging Jan 09 14:07:56 as for the .package files, yeah, I guess moving them is probably okay. It seems a bit weird to put intra-package data in the staging area, particularly if you write it prior to do_stage, but I suppose it doesn't matter much. Jan 09 14:08:31 RP: the shlibdeps file is transient intra-package data. why do you want that in staging? Jan 09 14:09:08 it's only left around at all so you can eyeball it for debugging; nothing (iirc) uses it after shlibdeps has run. Jan 09 14:10:16 If we wanted to reverse the process of putting the shlibs files into staging, would we need it? Jan 09 14:10:33 sorry, I don't understand the question Jan 09 14:11:00 as in, I'm not sure what you mean by "reverse the process" Jan 09 14:11:30 night all Jan 09 14:11:49 Building a library currently causes a number of files to enter staging. Once staging is under packaging control, we need to be able to reverse that process. Jan 09 14:12:18 Would the shlibdeps file allow us to identify which files to remove? Jan 09 14:12:23 no Jan 09 14:12:43 ok, it can stay in work then Jan 09 14:12:59 the shlibdeps file just contains the dependencies that package_do_shlibs() found for the current package. Jan 09 14:14:27 right, rather than the dependencies a package provides which is what the data in staging records Jan 09 14:15:27 indeed Jan 09 14:17:26 The issue I'm trying to solve currently is the conversion of RDEPENDS into whatever naming scheme is favour of the month in any bbclass like debian.bbclass Jan 09 14:17:42 To do that, we need the information in the .packages files Jan 09 14:18:17 mm, right. Jan 09 14:18:27 I'm not totally convinced that this is a problem worth solving, mind. Jan 09 14:18:55 until and unless someone wants to do package renaming on non-shared-library packages, it seems like an issue that one can just ignore. Jan 09 14:19:47 pb_: Ok, tell me how I add RDEPENDS = "fontconfig-utils" to a .bb and not have ipkg complain about the lack of "fontconfig-utils" as debian.bbclass resulted in libfontconfig-utils Jan 09 14:20:03 that's a bug in fontconfig.bb; it shouldn't be renaming the binary packag. Jan 09 14:20:07 or the binary package Jan 09 14:20:51 ok, and the renaming of kernel-image in the kernel.bbclass Jan 09 14:21:26 I don't know what the issue is there. What's doing the depending there? Jan 09 14:21:56 PROVIDES_kernel-image = "kernel-image-${KERNEL_VERSION}" Jan 09 14:21:56 RDEPENDS_kernel = "kernel-image-${KERNEL_VERSION}" Jan 09 14:22:07 if it's some package that just wants _a_ kernel but doesn't care which one, it should be depending on "virtual-kernel" or some similarly virtual thing. Jan 09 14:22:43 This is a tweak to enable the PARALLEL_MODULES_INSTALL Jan 09 14:22:56 RP: I don't see any problem in what you just pasted. Jan 09 14:23:05 It also sets PKG_kernel-image = "kernel-image-${KERNEL_VERSION}" Jan 09 14:23:14 "kernel" is depending on kernel-image-${KERNEL_VERSION}, which is provided. There's no renaming there. Jan 09 14:23:17 ACKAGES = "kernel kernel-image kernel-dev" Jan 09 14:23:26 but PACKAGES is set to "kernel-image" Jan 09 14:23:36 that's fine Jan 09 14:24:03 No, its not. Its exactly the same issue as above... Jan 09 14:24:23 The PACKAGES != the final package name Jan 09 14:24:31 Why is that a problem? Jan 09 14:24:46 PACKAGES is just an internal thing. Nothing ought to care about it. Jan 09 14:24:48 how does bitbake lookup kernel-image-somerandomnumber? Jan 09 14:25:16 against the Provides:, one would hope, just like ipkg does. Jan 09 14:26:01 It can't. ${KERNEL_VERSION} isn't known at build time Jan 09 14:26:31 That's unfortunate, then. Jan 09 14:27:08 There's nothing actually wrong with what that package is doing though. The results are consistent, and the binary packages should install just fine. Jan 09 14:27:49 The binary packages will be fine. bitbake's metadata is inconsistent though Jan 09 14:28:22 I don't think bitbake's metadata will actually be inconsistent; it's just that you can't statically compute its entire state. Jan 09 14:29:55 You have PACKAGES calling a package one thing and the actual package being called something else with the RDEPENDS not using PACKAGES naming but that of the renamed package Jan 09 14:30:13 indeed. Jan 09 14:30:38 PKG exists to make the mapping between the thing in PACKAGES and the output package names. Jan 09 14:31:08 Yes, and so it should be used to convert the RDEPENDS at package build time as well Jan 09 14:35:37 RP: possibly. I still feel that bitbake ought to be able to cope with this situation on its own, though. It isn't asking much for it to notice that the RDEPENDS string is lexically identical to a provided virtual in the same .bb file, even if it can't fully expand those strings at this time. Jan 09 14:37:01 that reservation aside, I don't have any real objection to the PKG processing on rdepends that you suggest. Jan 09 14:37:21 pb_: As it happens I can "hack" around this, as I've already mentioned on oe@ using PACKAGES_DYNAMIC. It doesn't change the fact this feels very wrong though. Jan 09 14:38:21 and, all that said, I have a feeling that the whole PARALLEL_INSTALL_MODULES thing is a bit misconceived and needs to be redone. So it's possible that this problem would just go away in any case. Jan 09 14:38:58 I'm a bit surprised that PARALLEL_INSTALL was even enabled for openzaurus in the first place. It used to be a familiar only thing. Jan 09 14:39:27 You might find that the easiest fix for your problems is just to turn it off for the time being. Jan 09 14:39:53 Turning off PARALLEL_INSTALL doesn't remove that package naming Jan 09 14:40:11 oh, doesn't it? It almost certainly does remove the motivation for it. Jan 09 14:40:31 it does, but it won't change the naming or the RDEPENDS Jan 09 14:40:45 ah, that sounds like an implementation bug. Jan 09 14:41:10 pb_: As I noted yesterday, you were the one who commited it ;-) Jan 09 14:41:27 indeed Jan 09 14:41:46 it was quite a long time ago, though, and I don't exactly remember what the behaviour was prior to that patch. Jan 09 14:42:06 the intent was certainly that the behaviour before and after would be the same if you didn't have PARALLEL_INSTALL_MODULES set. Jan 09 14:42:32 The kernel-image package didn't exist before the patch and did afterwards Jan 09 14:42:34 and I'm fairly sure that the renaming you're describing was introduced with that patch rather than existing beforehand Jan 09 14:42:53 oh, yeah, that's right Jan 09 14:43:08 previously there was just "kernel" Jan 09 14:43:42 so in that case I guess it isn't really a bug in the patch as such Jan 09 14:44:37 ACKAGES = "kernel kernel-image kernel-dev" Jan 09 14:44:49 http://openembedded.bkbits.net:8080/openembedded/diffs/classes/kernel.bbclass%401.32?nav=index.html|src/.|src/classes|hist/classes/kernel.bbclass Jan 09 14:46:09 yeah Jan 09 14:47:16 I wonder if it'd be safe to just back that patch out. Jan 09 14:48:39 probably not terribly safe, but I guess it's no worse than leaving it in. Jan 09 14:49:19 It depends whether you'd class that as breaking upgrades I guess... Jan 09 14:49:33 and whether you consider that implementation broken... Jan 09 14:50:17 I think it probably would break upgrades, but no more badly than they would be broken anyway next time the kernel version changes. Jan 09 14:50:30 We could get it removed from the release branch, then no release of familiar would have had it in... Jan 09 14:51:36 Mm. It's in -rc1 and the feeds, so it's probably too late for that. But removing it from the release branch mightn't be a bad idea anyway. Jan 09 14:53:02 actually, it looks like 0.8.2 had the patch as well Jan 09 14:53:31 the only version of kernel.ipk in the 0.8.2 feeds is for c7x0 (!) but it does seem to have the kernel-image-.. dependency. Jan 09 14:53:51 ah, too late then :-/ Jan 09 14:53:59 kergoth: good morning Jan 09 14:55:00 RP: well, that doesn't mean it's too late to take it out. I imagine it'd be easy enough to fiddle the metadata to minimize the breakage anyway. Jan 09 14:55:13 hey Jan 09 14:55:18 pb_: I also don't understand the problem with fontconfig - you say its the fault of fontconfig.bb when debian.bbclass is filtering all PACKAGES Jan 09 14:55:21 hi kergoth Jan 09 14:55:35 might need to add a few strategic conflicts/replaces, but I doubt it's any worse than that. Jan 09 14:55:44 RP: well, it might actually be the fault of debian.bbclass. Jan 09 14:56:07 I meant more that it isn't the responsibility of the thing that's doing the rdepending to know what the package will have been renamed to. Jan 09 14:56:52 Ah, right. I agree with that much. The problem is how can they tell? Jan 09 14:57:08 and how do we fix it... Jan 09 14:57:27 What is the correct debian name for fontconfig-utils? Jan 09 14:57:49 probably just that, "fontconfig-utils". Jan 09 14:58:09 so debian.bbclass shouldn't be renaming it... Jan 09 14:58:47 Should fontconfig be libfontconfig? Jan 09 14:59:18 dunno. I think it's "fontconfig" upstream. Jan 09 14:59:28 yeah 2.6.15 git runs on my SHark Jan 09 14:59:38 zecke: including init? Jan 09 14:59:38 pb_: either I forgot floating point emulation... Jan 09 14:59:53 pb_: Enter runlevel: it prints Jan 09 15:00:25 RP: on the other hand, I guess your plan of doing PKG processing in rdepends would make that stuff more robust in the case of putative binary renaming in the future, which mightn't be a bad idea. Jan 09 15:00:45 I can't think of any situation in which you'd want to rename non library packages, but I suppose someone might think of one at some point. Jan 09 15:00:52 pb_: now I only need to get a rootfs running... debootstrap does not cut it... Jan 09 15:00:57 RP: sorry, i havent gotten to checking out that patch. i'm in the middle of getting my shit together for a move down to phoenix. I drive out on thursday, lots of frantic packing and cleaning involved Jan 09 15:00:57 zecke: excellent Jan 09 15:02:38 and irc, it seems Jan 09 15:03:12 practically got lost in my closet. more cleaning than packing involved, i'm sorry to say Jan 09 15:03:15 hehe Jan 09 15:03:34 kergoth: np, I understand you're a "bit" busy :). We've checked it into bitbake and are currently seeing how much it breaks. The answer is a fair bit but I think the patch (and some ammendments are correct in themselves) Jan 09 15:04:02 kergoth: heh Jan 09 15:04:28 kergoth: I don't dare open certain cupboards here :) Jan 09 15:04:57 this afternoon i found the kitchen tabletop again. hadnt seen that in a while. my computer is set up there, lacking a desk..:) Jan 09 15:06:09 I'm living with parents atm so no chance of that happening! Jan 09 15:06:23 hehe Jan 09 15:07:06 The place is flood wired with CAT5E but I had to do it properly under the floorboards :) Jan 09 15:09:36 pb_: Setting AUTO_LIBNAME_PKGS = "" in fontconfig would be the correct thing to do? Jan 09 15:11:12 RP: no; you want the library packages renamed, just not the one with the binaries in Jan 09 15:14:02 pb_: Well, if fontconfig = fontconfig and fontconfig-utils = fontconfig-utils, there is nothing to rename. I suspect you mean only fontconfig should be renamed then? Jan 09 15:14:19 (and fontconfig-dev) Jan 09 15:15:12 oh, right. sorry, I misunderstood what you meant when you were talking about fontconfig earlier; I thought you were talking about changing the name of the source package. Jan 09 15:15:25 mhh got an error with glibc Jan 09 15:15:27 yes, the library package should be libfontconfig0 or some such Jan 09 15:15:47 | Applying patch arm-audit.patch Jan 09 15:16:45 pb_: That's not what you said before ;-) Jan 09 15:19:02 RP: I guess I was confused before. Jan 09 15:20:40 * RP notices libfontconfig-dev in meta-sdk Jan 09 15:21:08 * RP sobs quietly Jan 09 15:23:08 ah, mm, yeah. now, -dev packages _are_ a good argument for filtering RDEPENDS through PKG. Jan 09 15:23:15 those things do indeed need to be renamed. Jan 09 15:26:54 ok, so we do need a chunk of new code, and all the .packages files... Jan 09 15:27:39 and technically to always follow RDEPENDS so they're built and packaged :-/ Jan 09 15:27:40 * Philippe is away: Time to dream :-) Jan 09 15:28:03 maybe not the actual .package files, but you need something that captures the contents of PKG. Jan 09 15:28:28 it might be better to create a new set of files indexed by PACKAGE, to avoid having to search all the .package files. Jan 09 15:28:57 Yes, I was thinking that Jan 09 15:30:10 re Jan 09 15:30:14 hi CoreDump|home Jan 09 15:30:38 gremlin[it]: update your bitbake and/or read the topic/oe mailing list Jan 09 15:31:57 RP: i update bitbake acouple of hours ago ... i'll read email so :) Jan 09 15:33:15 gremlin[it]: It sounds like your glibc doesn't have a CVSDATE working properly Jan 09 15:33:34 hell, lost my stylus. this is no good. Jan 09 15:39:11 pb_: i sooo hate it when that happens Jan 09 15:39:27 good nite guys Jan 09 15:39:56 n8 zecke Jan 09 15:40:53 'night zecke Jan 09 15:41:06 strangely, I've had more stolen than I've lost :-/ Jan 09 15:41:23 RP: may I ask the status of 2.6.x on Akita? Jan 09 15:43:05 CoreDump|home: Should be working well Jan 09 15:43:23 that's great to hear Jan 09 15:43:45 CoreDump|home: ABout the only thing missing is RNDIS support for usb client and cpufreq support Jan 09 15:43:49 I'm currently compiling -rjan2006prebreakage which uses 2.6 Jan 09 15:44:11 dang, USB clients were kinda funny =) Jan 09 15:45:25 Always wanted to add "Boot from USB" to altboot heh Jan 09 15:47:05 CoreDump|home: Note usbnet works, just not RNDIS (windows) Jan 09 15:47:26 and I've discovered some hacks that in theory make RNDISK work Jan 09 15:47:37 umm sounded like USB clients (like USB sticks etc) didn't work, gues I got that wrong Jan 09 15:48:20 usb host works which means you can plug in usb storage devices fine Jan 09 15:48:30 excellent! Jan 09 15:48:35 screw windows =) Jan 09 16:29:18 cp Jan 09 16:29:41 here tothere Jan 09 16:29:58 crazy proxy Jan 09 16:30:04 you never know what it will do next Jan 09 16:30:49 well you can hope it will connect you with a cute girl Jan 09 16:48:30 http://zaurus.catstar.org/jmqt/index_e.html Jan 09 17:05:06 Does OE have a variable "RSUGGESTS"? Jan 09 17:05:46 yes Jan 09 17:06:02 pb_: Where is that implemented? Jan 09 17:06:16 package_ipk Jan 09 17:06:39 you don't mean RRECOMMENDS? Jan 09 17:06:46 no Jan 09 17:07:32 Suggests: and Recommends: are separate fields with distinct meanings. Jan 09 17:08:03 Yes, I see now. I need to work on my grep skills :-/ Jan 09 18:02:23 Hi, Jan 09 18:02:52 when trying to compile I get an error "NOTE: package glibc-2.3.5+cvs20050627-r1: task do_fetch: failed". What now? Jan 09 18:03:41 glibc version 2.3.5 does not seem to be available from oesources.org Jan 09 18:10:52 OK, tomorrow is another day. Jan 09 18:10:55 L8er Jan 09 18:56:56 NOTE: package gpe-image-1.0: completed Jan 09 18:57:31 opie probably would work if the new code wasn't so zealous at following rdepends... Jan 09 20:52:41 Nah, you're fine for now. I gave IL the space since I just wasn't using it, and its helped make the place nicer. Plus I do like the idea, just apparently can't use the equipment. :( Jan 09 20:52:54 Sorry, wrong window. :( Jan 10 00:11:57 halo, any idea when next release is due? Jan 10 00:12:45 release of what? Jan 10 00:12:53 3.5.4? Jan 10 00:13:00 oz i take it? Jan 10 00:13:24 or something else? Jan 10 00:13:44 sure...but I thought that oz 3.5.4 was related to the oe I'll run on my slug and the familiar I'll run on my 3765 ipaq Jan 10 00:13:52 :) Jan 10 00:14:00 oe is a build system Jan 10 00:14:22 you might get the awnswer you want from the oz channel tho Jan 10 00:14:29 answer* Jan 10 00:15:16 yes...true...sorry...wow feel like the coyote with my butt on fire from the acme diy flame thrower... :) Jan 10 00:15:37 there are some oz devs here but i think they are all asleep Jan 10 00:15:58 what platform are you here for emte? Jan 10 00:16:13 ipaq mostly Jan 10 00:17:00 trying to figure out the current issues with fam 0.8.3 before i try to mangle my version again Jan 10 00:17:19 did we ever get dual boot working ie wince in internal flash and familiar in CF? when I left for Z land 3 years ago it was barely working...now that I bought some used ipaqs I still see it's not perfect(or docd) Jan 10 00:17:41 no idea Jan 10 00:17:49 i dont really have a huge use for wince Jan 10 00:17:50 yeah, I'm waiting for 0.8.3 and oz 3.5.4 thought thier simultaneous delay was related to oe Jan 10 00:18:15 the oe meta is getting another chunk of major chainges Jan 10 00:18:19 changes* Jan 10 00:18:26 I dont know why Im so hung up on dual boot ... just am Jan 10 00:18:50 * emte shrugges Jan 10 00:18:52 meta ... I thought that was the individual packages build instructions... Jan 10 00:19:04 i cant say i find dual booting useful in any capacity Jan 10 00:19:15 it is Jan 10 00:19:23 working Jan 10 00:19:35 just assumed that packages would improve at their own time...hodge podge like Jan 10 00:19:42 sleeping Jan 10 00:20:13 its a try to clean up some poor maintainers parctices Jan 10 00:20:19 practices* Jan 10 00:20:28 as far as i understand it Jan 10 00:20:48 I've been ignoring oe for the last year or so...too complex for my simple brain...has it gotten easier? Jan 10 00:20:58 so that eventually dependancies can be followed properly Jan 10 00:21:06 lol Jan 10 00:21:08 not really Jan 10 00:21:45 its a grand concept but there are too many people like me...not developers, can barely read a make file, just want to compile nethack and play... Jan 10 00:22:04 then your in the wrong place Jan 10 00:22:07 desktop solutions allow even me to compile complex apps...PDAs(linux) should too Jan 10 00:22:20 you would be expecting a SDK releasd from your distro Jan 10 00:22:32 but without me there arent enough users to justify businesses... Jan 10 00:24:42 yeah the hard part is trying to keep up and get enough maintainers to expand what already exists Jan 10 00:25:00 i've been thinking about setting up gcc/zaurus/virtualmhz with the Z emulator to build a fast compile environment Jan 10 00:25:48 a fast compile helps but there is the shear amount of data to compile Jan 10 00:25:49 that's what Im saying emte...we need more "average" people like me playing with it...and it's too difficult for me to compile xscreensavers/vnc/you name it... Jan 10 00:26:28 yes but its an unfort necessity Jan 10 00:26:37 well if I go with virtualmhz, I'll use the latest oz distro, then move to my pc, then install the dev env on pc and native emulated Jan 10 00:26:46 there are too many buildsystems that prohibit using just one type Jan 10 00:27:23 (I got lost, two thought threads: unfort necessity? and too many buildsystems?) Jan 10 00:27:30 has anyone reported that combination will work? Jan 10 00:27:52 yeah say 4 months ago someone told me the virtualmhz would emulate a Z Jan 10 00:28:04 Ill check now...hate spewing crap... Jan 10 00:28:13 oe/bitbake is more or less a wrapper to convert many buildsystem types into a cross compile enviroment Jan 10 00:28:51 autotools, smake, imake, etc Jan 10 00:29:35 yep try www.vitera.com they have a openzaurus section and maybe do ipaqs too Jan 10 00:29:50 (oh ok) Jan 10 00:30:15 they(vitera) say a modern pc can emulate a 200mhz arm proc Jan 10 00:30:31 my dual 1600mhz amd might need upgrading :) or :( Jan 10 00:30:32 seems to be 401 Jan 10 00:30:59 oops www.virtera.com forgot the extra r Jan 10 00:31:01 i suppose emulation helps Jan 10 00:31:25 but if you have the devices there is nothing like running the hardware Jan 10 00:32:04 only if the cpu works well...funny but you could emulate native compiling get big disc space then distcc most processing to the desktop for native cross to emulated native dev...havent had enough rum to figure that one out tonight Jan 10 00:32:31 bitbake doesnt support distcc yet Jan 10 00:32:47 inpart because i dont think gcc supports it for arm Jan 10 00:33:02 arm/pxa Jan 10 00:33:30 i'd abandon oe(oops shouldnt say that here)...that's what Ive done till now...1gb sd card, all native compiling...screw cross compiling its toooo damn hard Jan 10 00:33:55 nope, im pretty sure someone had distcc working on Z Jan 10 00:34:10 yes there is a version avail for Z Jan 10 00:34:24 then i missed what you meant about bitbake Jan 10 00:34:30 but that was the old sharp images was it not? Jan 10 00:34:43 you maybe right, i never got it working... Jan 10 00:35:22 there is a modified gcc which does have it working but thats not mainstream Jan 10 00:35:29 I have a sl-5500 and it's never been "clean" under oz, now oz is migrating to 2.6 and the sd driver for 5500 wont work in 2.6 Jan 10 00:35:44 you sure? Jan 10 00:35:52 i though someone got that working ... Jan 10 00:36:03 thought* Jan 10 00:36:09 someone reverse engineered it but I dont think it got ported to 2.6... I have been sleeping a while though... Jan 10 00:36:46 anyway sleep time for me Jan 10 00:36:51 later... Jan 10 00:37:43 anyone awake? anyone ported klimt to oe? Jan 10 00:38:44 everyone who's not awake snore. Jan 10 01:42:25 morning Jan 10 01:52:09 morning Jan 10 02:04:31 morning all Jan 10 02:13:41 hey RP Jan 10 02:15:57 morning all Jan 10 02:19:26 ~lart transcode for not having rotate Jan 10 02:19:27 * ibot teaches transcode the basics, including how to RTM for not having rotate Jan 10 02:25:41 * RP wonders why his build box has disappeared from the internet :-/ Jan 10 02:26:24 morning Jan 10 02:26:49 DOSed by microsoft because OZ is supier to WinCE Jan 10 02:27:39 * RP notes the router has disappeared as well... Jan 10 02:28:05 RP: does this mean your brain gets a day to cool down? Jan 10 02:28:46 XorA: Yes, because the best copy of the patch to bitbake is on that machine :) Jan 10 02:30:28 RP: ah well, maybe its an ISP problem and it be back soon Jan 10 02:30:41 morning ! Jan 10 02:33:47 morning alan|laptop **** ENDING LOGGING AT Tue Jan 10 02:59:56 2006