**** BEGIN LOGGING AT Mon Jun 06 23:59:56 2005 Jun 06 23:59:56 which is correct Jun 07 00:00:33 x11 takes X11 and dl, xau only Xau Jun 07 00:00:41 hrw: and "nm staging/arm-linux/lib/libXdmcp.so.6 | grep -i wrap" reveals "00001fac T XdmcpWrap" Jun 07 00:01:09 ok. so add that lib to linking Jun 07 00:02:28 hrw: hmm, do you have a staging/arm-linux/lib/pkgconfig/x11.pc file? I don't. Jun 07 00:02:47 its share/pkgconfig/ Jun 07 00:02:59 not lib/pkgconfig Jun 07 00:03:20 hmm.. we have both... Jun 07 00:03:41 122 files in share/pkgconfig, 35 in lib/pkgconfig Jun 07 00:04:06 hrw: I have a x11.pc in share/pkgconfig, but not in lib/pkgconfig Jun 07 00:04:21 hrw: after I added -l Xdmcp to the x11.pc, my gpe-soundserver compiled Jun 07 00:04:23 me too Jun 07 00:04:39 it should be fixed in package not in x11.pc Jun 07 00:04:49 there is xdcmp.pc Jun 07 00:12:42 <__law__> NOTE: Update cvs://anoncvs@savannah.gnu.org/cvsroot/gcc;method=ext;tag=csl-arm-branch;module=gcc;date=20050603 Jun 07 00:12:42 <__law__> connect to address 199.232.41.3: Connection timed out Jun 07 00:12:58 <__law__> :-( since 2 days Jun 07 00:19:48 good morning all Jun 07 00:20:31 hi koen Arjan raster Jun 07 00:20:36 burrp Jun 07 00:20:43 hrw: booyah Jun 07 00:20:47 _law_: try to grab it from cvs_tarball_stash Jun 07 00:22:16 ok. time to prepare to goto work Jun 07 00:22:36 hi hrw Jun 07 00:22:51 time to boot windows and work on matlab... Jun 07 00:23:09 koen: windows? Jun 07 00:23:24 <__aZ> hi.. anybody knows how to clean all builded packages without cleaning the toolchain? Jun 07 00:23:24 yeah Jun 07 00:24:02 __aZ: bitbake -cclean all pakcages by name except those from toolchain Jun 07 00:24:19 __aZ: next time build toolchain, backup tmp Jun 07 00:25:42 <__aZ> ok.. if I delete only tmp/work will bitbake rebuild the toolchain also or it will be smart enough to see it already built? Jun 07 00:26:14 as soon as you want to build something that uses pkgconfig, your out of luck Jun 07 00:26:29 hrw: it was pkgconfig being evil, not libtool Jun 07 00:27:56 koen: ok good to know Jun 07 00:28:16 __aZ: tmp/work/ is usually not needed after build Jun 07 00:28:41 ~lart libtool and pkgconfig Jun 07 00:28:42 * ibot whacks libtool and pkgconfig with the cluebat Jun 07 00:29:21 __aZ: if you build opie then you can just rm -rf tmp/work and all will build. if gpe then you have to leave all libs in tmp/work Jun 07 00:29:23 about 25% of the packages in OE use pkgconfig Jun 07 00:29:26 maybe more... Jun 07 00:29:58 hmm Jun 07 00:30:57 <__aZ> hrw: 10x but for now i'm still playing with x86.. i'll try and see what happens:) Jun 07 00:31:30 oik I go to work Jun 07 00:31:39 brb in ~15-20 minutes Jun 07 00:34:19 hrw|gone: Hey, is it possible that opie-image broke with yesterdays commit? Jun 07 00:34:22 ERROR: Nothing provides base-files Jun 07 00:34:22 ERROR: dependency base-files (for task-bootstrap) not satisfied Jun 07 00:34:36 that's after bk pull and a removal of /tmp Jun 07 00:41:01 hey bluelightning Jun 07 00:41:07 hi koen Jun 07 00:41:54 morning Jun 07 00:41:54 how's life in NZ? Jun 07 00:42:03 hey XorA Jun 07 00:42:52 koen: a bit chilly... but otherwise ok :) Jun 07 00:43:16 koen: how about sunny NL? Jun 07 00:44:58 foggy at the moment Jun 07 00:45:14 I hope the sun will get through in the afternoon Jun 07 00:46:59 re Jun 07 00:47:24 hi XorA bluelightning Jun 07 00:47:39 hi hrw|work Jun 07 00:47:44 hey hrw|work :) Jun 07 00:49:28 ascent: strange Jun 07 00:49:44 ~seen zecke Jun 07 00:49:46 zecke <~ich@83-169-170-199-dynip.superkabel.de> was last seen on IRC in channel #oe, 8h 38m 17s ago, saying: 'and good nite'. Jun 07 00:50:12 hrw|work: Can you build task-bootstrap yourself? Jun 07 00:50:32 ascent: can try Jun 07 00:50:39 please :) Jun 07 00:50:49 but not from scratch - not powerfull machine Jun 07 00:51:08 eeks! voodoo python messages on my screen Jun 07 00:51:33 * koen cleans the cache Jun 07 00:51:49 gheh :) Jun 07 00:52:09 koen: you do have a powerfull machine, just a p4-3ghz with only 1g ram in here.. Jun 07 00:52:19 logging from non-utf8 system into utf8 one makes strange efects Jun 07 00:53:04 hrw|work: utf8 support is a nightmare :-) Jun 07 00:54:05 XorA: you live in us-ascii region of world... Jun 07 00:54:27 ascent: builds without problem Jun 07 00:54:40 hrw|work: but, with old tmp, right? Jun 07 00:54:49 ascent: yes Jun 07 00:55:06 let's see what koen gets, he's doing the compile too. Jun 07 00:55:21 Otherwise I'm going to tear my oe tree apart, and get a fresh one :) Jun 07 00:55:22 File "/home/koen/OE/openembedded/packages/kbdd/kbdd_cvs.bb:populate_packages", line 208, in ? Jun 07 00:55:25 File "/home/koen/OE/openembedded/packages/kbdd/kbdd_cvs.bb:populate_packages", line 27, in populate_packages Jun 07 00:55:29 File "/home/koen/OE/openembedded/packages/kbdd/kbdd_cvs.bb:populate_packages", line 17, in update_rcd_package Jun 07 00:55:32 TypeError: cannot concatenate 'str' and 'NoneType' objects Jun 07 00:55:34 that's what I get Jun 07 00:56:44 me too Jun 07 00:56:52 koen: I got it in linux-hotplug Jun 07 00:59:08 let's wait for zecke and mickey|zzZZzz to show up Jun 07 01:00:44 hrw|work: not when you guys put those wierd characters on my screen Jun 07 01:01:22 XorA: like those: ±æê³ñ¶ó¼¿? Jun 07 01:02:15 hrw|work: yeah thats +- ae e' | n~ Jun 07 01:03:44 heh Jun 07 01:03:50 10:01 < hrw|work> XorA: like those: ?????? Jun 07 01:04:04 that's what I get :) Jun 07 01:04:16 I got the above error also in linux-hotplug a few minutes ago Jun 07 01:04:53 koen: ;) Jun 07 01:05:03 arjan: how's about task-bootstrap? Jun 07 01:05:35 it occurred in bootstrap-image here Jun 07 01:05:41 ah okay Jun 07 01:05:50 that's about the same :) Jun 07 01:07:25 morning Jun 07 01:07:29 howdy Jun 07 01:07:35 We wachten in spanning op Judith's antwoord... Laat ook maar zitten. Jun 07 01:07:37 evenin Jun 07 01:07:42 morning :) Jun 07 01:08:54 http://www.joemonster.org/i/i/kino76.jpg Jun 07 01:10:14 koen: I got that in atd Jun 07 01:15:09 Can somebody look at classes/update-rc.d.bbclass. pb's last change breaks this module. He changed updatercd_postrm into updatercd_prerm. But updatercd_postrm is used in line 38. Jun 07 01:18:26 do13: do you want BK write access ? :-) Jun 07 01:19:01 morning all Jun 07 01:19:20 hi RP, ljp Jun 07 01:22:18 do13: I changed this, this fixed my atd package compilation, no I'll try this with dbus Jun 07 01:23:09 RP: morning Jun 07 01:23:30 morning all Jun 07 01:24:38 dbus build fine as well, now I've "bk push"ed. Jun 07 01:25:19 koen: can you try again? Jun 07 01:27:32 do13: I posted the latest status of that copypage problem to arm-linux-kernel. Ideas welcome as it appears to be insane :-/ Jun 07 01:29:08 RP: I read your mail. But I haven't looked in detail. Jun 07 01:30:22 schurig: Now bitbake busybox works. Jun 07 01:32:22 schurig: that seems to fix it Jun 07 01:33:06 I got WRT54GS Jun 07 01:33:25 hrw|work: have you installed an OE baked firmwarew yet? Jun 07 01:33:36 koen: I dont unpack it yet Jun 07 01:34:17 probably will look into it tomorrow - today evening I have meeting with Spanish OZ user Jun 07 01:37:02 now time to check which rev it is Jun 07 01:40:10 v1.1 - supported by openwrt Jun 07 01:43:33 NOTE: [10:34] package task-bootstrap-1.0-r23: task do_build: completed Jun 07 01:46:09 hrw|work: what did you do to get it compiling? Jun 07 01:48:29 morning Jun 07 01:49:08 Arjan: bk pull schurig fix for update-rc.d Jun 07 01:50:39 hrw|work: from which tree should I get it? and with which command? Jun 07 01:50:42 hrw|work: it was do13's fix, I was just the committer Jun 07 01:51:01 Arjan: it's in the developer tree, I don't how how long it takes to propagate to the anonymous tree Jun 07 01:51:17 Arjan: so, it's in bk://oe-devel@oe-devel.bkbits.net/openembedded/ Jun 07 01:51:28 Arjan: and the command is the usual "bk pull" Jun 07 01:52:45 hi mickeyl Jun 07 01:52:57 schurig: is there a way to pull it as anonymous from the tree? I don't have an account Jun 07 01:53:02 good morning folks Jun 07 01:53:23 Arjan: no, you need to wait until the fix propagates to the anonymous tree Jun 07 01:53:33 morning mickeyl Jun 07 01:53:36 ok, thanks Jun 07 01:53:41 hi RP, hrw Jun 07 01:53:48 morsche schurig Jun 07 01:54:45 Arjan: You can pull from it anonymously - jsut run bk parent bk://oe-devel.bkbits.net/openembedded/ then bk pull Jun 07 01:54:51 Arjan: or you do "bk edit classes/update-rc.d.oeclass" and change the postrm's there into prerm's Jun 07 01:54:56 mickeyl: morsche Jun 07 01:55:01 Just don't clone from there Jun 07 01:55:15 RP: does this work? Jun 07 01:55:16 RP: I thought it would ask for a password then Jun 07 01:55:27 RP: except when you have the right RSA key Jun 07 01:55:38 schurig: If you leave the oe-devel@in it will Jun 07 01:56:05 hey mickeyl Jun 07 01:58:59 schurig: thanks! that fixed it! Jun 07 01:59:12 schurig: did you push that fix yet? i've nothing to pull here Jun 07 01:59:24 mickeyl: I've pushed it Jun 07 01:59:33 hmm Jun 07 01:59:47 mickeyl: do "bk citool" Jun 07 02:00:08 mickeyl: what's your last one? Jun 07 02:00:56 ChangeSet@1.3539, 2005-06-07 10:24:03+02:00, schurig@mnz66.mn-solutions.de Jun 07 02:00:59 that looks right Jun 07 02:01:42 hmm ok, guess i pulled on a different xterm Jun 07 02:01:49 t'works now *shrug* Jun 07 02:06:30 d'oh Jun 07 02:06:33 C3100 announced Jun 07 02:06:44 i thought they'd come later with the press release Jun 07 02:06:45 finally? Jun 07 02:06:59 128mb flash Jun 07 02:07:07 I was hoping for bluetooth or wifi Jun 07 02:07:42 url? Jun 07 02:07:54 that ridge around the keyboard looks uncomfortable Jun 07 02:08:34 http://ezaurus.com/lineup/sl/compare_spec.html Jun 07 02:09:01 http://www.sharp.co.jp/corporate/news/050607-a.html Jun 07 02:09:24 heh.. c3000 with 128M flash ;( Jun 07 02:09:52 sharp isn't in to the whole "innovation" thing Jun 07 02:10:08 c700 + more ram + more storage Jun 07 02:10:34 and a different colour Jun 07 02:11:12 I bet they still use that 30 year old "framebuffer" thing Jun 07 02:11:36 koen: yes, and the 60 year old keyboard thing as well Jun 07 02:11:53 ouch Jun 07 02:11:55 better than the 90 year old x11 Jun 07 02:11:57 koen: maybe even a 70 year old loudspeaker? Jun 07 02:12:13 old != bad Jun 07 02:12:37 tell that to my 20 MB harddrive Jun 07 02:12:52 :D Jun 07 02:12:58 ljp: oh, you can use that as a paper weight, or to train your bizeps Jun 07 02:13:34 nope. DOS with a compressed drive! Jun 07 02:13:34 ljp: hmm, paperweight might not work. You're living on the other side of the world, maybe it has negative weight there :-) Jun 07 02:15:37 wow, it took Debian just 15 years to release Sarge :-) Jun 07 02:16:29 openwrt use own buildsystem ;( Jun 07 02:16:44 hrw|work: not very long if you put your fingers in it :-) Jun 07 02:16:46 debian released, apple switched to x86, time for opie to drop qt/e :) Jun 07 02:16:59 koen: good said Jun 07 02:17:25 koen: :) Jun 07 02:17:27 koen: X11+Qt/X11 > Qt/E Jun 07 02:17:39 this is not the new device i've been expecting Jun 07 02:17:45 i think there will come another announcement soon Jun 07 02:17:50 schurig: I probably can build openwrt using OE - but moving all their devs to OE.. Jun 07 02:18:04 mickeyl: hmm, get me customers for 100.000 MNCIs and we can halve the price :-) Jun 07 02:18:17 btw, OE powered guadec too Jun 07 02:18:22 they used meshcubes :) Jun 07 02:18:52 mickeyl: then you have WIFI+Bluetooth+GSM/GPRS ... and a nice paper weight as well :-) Jun 07 02:19:02 schurig: ah bummer, i have just two or three potential ones ;) Jun 07 02:19:06 a bit far from 100.000 :) Jun 07 02:19:14 koen: nope. time for gpe to drop gtk :) Jun 07 02:19:22 only a few order of magnitudes Jun 07 02:19:34 mickeyl: What kind of device are you expecting out of interest? Jun 07 02:19:42 ljp: yeah, lets switch to fltk so we can use dillo again! Jun 07 02:19:48 RP: A C3000 thingy with wlan and/or BT Jun 07 02:20:03 mickeyl: and bigger screen Jun 07 02:20:16 mickeyl: wlan would be nice. Would bring it more level with the ipaqs... Jun 07 02:20:17 yeah, the transreflective display from the 6000 Jun 07 02:20:21 * koen will go for a nokia770 Jun 07 02:20:25 * mickeyl too Jun 07 02:20:41 with a bt keyboard Jun 07 02:20:44 I probably not - 350 eur is too high for me Jun 07 02:20:49 * RP will like the 770 more when the wlan actually works... Jun 07 02:20:51 E99 :) Jun 07 02:20:53 ~change 350 eur to pln Jun 07 02:20:56 350.00 Euro (EUR) makes 1,420.22 Zloty (PLN) (from http://www.xe.com/) Jun 07 02:20:57 hopefully Jun 07 02:21:15 hrw|work: 350 is a steal for the device specs Jun 07 02:21:35 if only they used sd instead of that $#@%^^#$%^ rs-mmc Jun 07 02:21:37 mickeyl: when it got into Poland it will have 550eur atleast Jun 07 02:22:28 mickeyl: and when opie will run on it? :) Jun 07 02:23:14 I dont want to run qvga sized gpe apps on it Jun 07 02:23:34 hrw|work: we have native gpe apps for it Jun 07 02:23:38 using hildon Jun 07 02:23:51 hrw|work: http://oss.kernelconcepts.de/maemo/ Jun 07 02:24:35 add OE already support the device :) Jun 07 02:24:38 i dont see how they can call that 770 a tablet Jun 07 02:24:41 s/add/and/ Jun 07 02:25:14 koen: I know but currently gpe has the same problem as many qpe apps.. qvga sized Jun 07 02:25:24 hrw|work: that depends on the state of Opie/X11 Jun 07 02:26:39 hrw|work: yeah, florian has worked on that for the last months, the pim stuff should be reasonable now (in cvs) Jun 07 02:26:50 the games are still but ugly in vga Jun 07 02:27:17 s/but/butt/ Jun 07 02:27:47 * koen reads the glade tutorial Jun 07 02:28:59 does qt have some form of glade? I need to poke at qt apps for work soon Jun 07 02:29:32 designer + uic Jun 07 02:29:47 thanks Jun 07 02:29:51 * koen fires up google Jun 07 02:30:01 although the designer in qt 2.x is very limited Jun 07 02:30:09 in 3.x it's ok and in 4.x it will rock Jun 07 02:31:36 hi mickeyl Jun 07 02:31:44 hmmmm Jun 07 02:32:25 I have to invent some excuses, it seems the apps are in qt/e Jun 07 02:32:35 good morning pb_ Jun 07 02:32:43 hey pb_ Jun 07 02:32:54 hi pb mallum Jun 07 02:33:48 hey hrw|work + all Jun 07 02:35:11 * RP hides Jun 07 02:36:56 where did RP go? Jun 07 02:37:42 can't see him Jun 07 02:39:22 3 Jun 07 02:39:57 so.. SL-C3100 Jun 07 02:40:18 It would have to happen on the week I get my 3000 :/ Jun 07 02:40:54 don't be disappointed, there's no significant difference Jun 07 02:41:09 that bit of more flash doesn't really count Jun 07 02:41:10 colour + flash Jun 07 02:41:34 That keyboard ridge looks painful and enough to put me off... Jun 07 02:41:46 RP: that's what I thought Jun 07 02:41:51 hey RP is back again! Jun 07 02:41:56 RP: I hadn't noticed that before Jun 07 02:42:33 koen: I'm not here really. You're just confused :) Jun 07 02:42:51 ah, ok Jun 07 02:44:27 http://www.formorer.de/blogs/uploads/phd-sarge.png Jun 07 02:45:31 hehe :) Jun 07 02:48:26 hrw|work: :-D Jun 07 02:57:21 hey Jun 07 02:57:30 hey zecke Jun 07 02:57:57 hi zecke Jun 07 02:59:01 hi zecke Jun 07 04:07:25 flex does not build - it doesn't even try Jun 07 04:07:51 probably something with the flex.inc thing - flex_2.5.31.bb just includes flex.inc Jun 07 04:09:00 jacques: could you check if it downloaded ok? Jun 07 04:09:51 koen, ok checking... Jun 07 04:10:32 -rw-r--r-- 1 fontenot users 681792 Apr 1 2003 ../../sources/flex-2.5.31.tar.bz2 Jun 07 04:10:32 -rw-r--r-- 1 fontenot users 32 May 30 16:33 ../../sources/flex-2.5.31.tar.bz2.md5 Jun 07 04:10:36 looks like it did Jun 07 04:10:37 the flex problems we were seeing were caused by a stupid sf mirror Jun 07 04:10:46 could you rm those? Jun 07 04:10:47 oh those are old problems Jun 07 04:10:58 this is new - bitbake flex does nothing Jun 07 04:11:01 and I hope it are old problems Jun 07 04:11:10 it think it has been built Jun 07 04:11:20 bitbake flex -c clean ; bitbake flex Jun 07 04:11:46 trying... Jun 07 04:13:07 koen, you are right - sorry - I hadn't seen that particular failure mode before Jun 07 04:13:28 I mean, I saw the problems with flex not downloading many times, but this was different Jun 07 04:13:46 bitbake's error reporting is.... different... Jun 07 04:14:05 oK something really si wrong Jun 07 04:14:08 it's not packaging Jun 07 04:14:21 after the build I have no flex ipks Jun 07 04:14:41 ~hail mickeyl for the bitbake shell Jun 07 04:14:42 which explains what it did before - it did nothing because it was already built, and I was looking for ipks Jun 07 04:14:42 * ibot bows down to mickeyl for the bitbake shell and chants, "I'M NOT WORTHY!!" Jun 07 04:15:02 ~lart mplayer Jun 07 04:15:02 * ibot teaches mplayer the basics, including how to RTM Jun 07 04:15:21 47minutes of encoding and segfault... Jun 07 04:15:36 hrw|work: it does that on occasion :-) Jun 07 04:15:45 koen, flex is not making a package Jun 07 04:15:58 hrw|work: for real instability you should try the amd64 version Jun 07 04:16:01 ok. another attempt. if that one also fail I will try 1.0pre7 Jun 07 04:16:20 XorA: donate me amd64 hardware so I will try ;) Jun 07 04:16:59 hrw|work: I am running 1.0pre7 and stability is ok at the moment Jun 07 04:17:12 jacques: it looks that way Jun 07 04:19:05 koen, it's even more broken - it's building an x86 binary instead of armeb Jun 07 04:19:10 first bug found in sarge cd/dvd images.. Jun 07 04:20:07 hrw|work: already? Jun 07 04:20:30 you'd think it would be bugless after 3 years Jun 07 04:21:05 entry in /etc/apt/sources.list is for security fixes in testing not in stable Jun 07 04:21:54 flex.inc inherit native... Jun 07 04:23:42 Im fixing it now Jun 07 04:23:45 hm, alsa-tools seems broken too. "no rule to make target `install`'.. Jun 07 04:23:56 hrw|work, thanks Jun 07 04:24:17 BTW, procps fails on package step - it tries to do install --strip Jun 07 04:27:30 mickey|meeting: bitbake interactive does not look at changed PR Jun 07 04:27:48 hi lardman Jun 07 04:28:00 hi hrw|work Jun 07 04:31:05 What provides: libasound2 (>= 1.0.8) Jun 07 04:31:06 ? Jun 07 04:31:27 alsa-lib Jun 07 04:35:08 ah thanks hrw|work, you're right Jun 07 04:42:11 flex fails to build for arm Jun 07 04:42:17 | tables.c: In function `yytbl_write16': Jun 07 04:42:17 | tables.c:313: internal compiler error: in arm_print_operand, at config/arm/arm.c:11410 Jun 07 04:43:17 I can push change which will fix flex to be non-native but it will anyway fail (atleast for arm) Jun 07 04:43:55 http://home.hrw.one.pl/oe/diffs/0607_flex.diff is whole change Jun 07 04:45:02 the change is OK Jun 07 04:45:19 and will push it Jun 07 04:45:22 add a BROKEN=1 and your done :) Jun 07 04:46:08 BROKEN_arm="1" Jun 07 04:50:05 damn, it used to work Jun 07 04:50:34 what about armeb? Jun 07 04:50:49 dont have armeb build here Jun 07 04:52:14 CIA-4: pushed flex Jun 07 04:52:33 hmm wait a sec Jun 07 04:53:01 it's been a long time, but I think I recall I had to build my own flex anyway - the included version was buggy (even when it built) Jun 07 04:53:14 I'll have to check my notes Jun 07 05:00:28 03koen 07 * r1.3539.2.1 10openembedded/packages/gkdial/gkdial_1.8.1.bb: gkdial_1.8.1.bb: point SRC_URI to the hh.org sourcemirror. ~lart disappearing sources Jun 07 05:07:38 no, it looks liek flex did used to work on armeb Jun 07 05:07:54 it was gawk that had issues Jun 07 05:08:44 hrw|work: add a bugtracker entry for everything you spot w/ bb -i please Jun 07 05:12:53 ok Jun 07 05:13:00 thanx Jun 07 05:13:57 sf not responding Jun 07 05:15:25 ok, responding extremely slowly Jun 07 05:31:45 How can I say that I prefer the _cvs version of the 2.6 kernel in my local.conf? Jun 07 05:35:55 ascent: imho you it is done in machine conf file Jun 07 05:36:09 I see. Jun 07 05:36:49 PREFERRED_PROVIDER_virtual/kernel = "handhelds-pxa-2.6" Jun 07 05:37:21 That's a directory, not a .bb file Jun 07 05:37:22 hmm Jun 07 05:37:35 which makes sense Jun 07 05:37:36 I think that is not only thing what is needed for 2.6 kernel Jun 07 05:37:44 a PROVIDER is not a .bb file Jun 07 05:37:48 thats a package name, not a dir Jun 07 05:37:50 But to what kernel does it default. Jun 07 05:37:52 a .bb file PROVIDES something, that's true Jun 07 05:38:37 the default is the latest version Jun 07 05:38:45 if not specified otherwise in local.conf or a machine conf Jun 07 05:39:21 mickeyl: so, the _cvs? Jun 07 05:39:40 if there is a _cvs one, then yes Jun 07 05:39:50 But I now get NOTE: package handhelds-pxa-2.6-2.6.11-hh2: failed Jun 07 05:39:57 while there is a _cvs.bb file Jun 07 05:40:27 well, then perhaps it's specified somewhere Jun 07 05:40:34 hmmz.. Jun 07 05:40:56 you read the .conf ? Jun 07 05:41:43 Yes. Jun 07 05:41:55 I pasted the PROVIDER line from it, a few lines back Jun 07 05:43:41 ascent: PREFERRED_VERSION_handhelds-pxa-2.6 = "2.6.10-hh0+cvs${CVSDATE}" Jun 07 05:45:50 2.6.11 ? :) Jun 07 05:45:53 lets see Jun 07 05:54:02 FYI, flex 2.5.31 builds native on armeb - running testsuite now Jun 07 05:56:12 ascent: no, 2.6.10 Jun 07 05:56:18 * jacques wonders if these are supposed to be in /usr/lib/gcc/armeb-linux/3.4.4/include: Jun 07 05:56:26 lrwxrwxrwx 1 root root 35 Jun 7 03:28 MozillaFirefox -> root/usr/lib/MozillaFirefox/include Jun 07 05:56:32 lrwxrwxrwx 1 root root 28 Jun 7 03:28 mozilla -> root/usr/lib/mozilla/include Jun 07 05:56:46 pb_: why 2.6.10? That's old.. Jun 07 05:56:59 haven't you heard? gcc depends on mozilla nowadays Jun 07 05:57:23 koen, d`oh! Jun 07 05:57:32 like windows and IE Jun 07 05:57:36 makes sense Jun 07 05:57:58 hmm flex testsuite isn't doing so well :-\ Jun 07 05:59:09 ah, OK Jun 07 05:59:31 the flex bison tests are failing, which makes sense considering bison is giving "broken pipe" error Jun 07 05:59:45 all the other tests are succeeding so far Jun 07 06:15:04 afternoon Jun 07 06:16:07 ascent: because that's what is in the .bb file. Jun 07 06:16:39 jacques: heh, that mozilla stuff does seem a bit bogus. I wonder where that is coming from. Jun 07 06:17:54 pb_, I can't even imagine :-) Jun 07 06:18:30 last flex test is taking forever Jun 07 06:18:41 but there is definitely something wrong with bison Jun 07 06:23:18 yeah all flex tests passed other than the three bison ones (because bison aborted) Jun 07 06:33:39 03mickeyl * r241 10bitbake/lib/bb/shell.py: build in a try/except block and intercept faulty builds Jun 07 06:34:13 mickeyl: r241 bitbake will not exit when error occure? Jun 07 06:40:30 hrw|work: bitbake shell shouldn't exit on build error Jun 07 06:40:43 great Jun 07 06:43:04 let every user fear our 1337 z3ck3 d0c5: http://ewi546.ewi.utwente.nl/tmp/d/keys.html Jun 07 06:43:49 a.. I had to look at it and improve look/code Jun 07 06:46:05 sorting it by alfabet would be near Jun 07 06:46:09 neat* Jun 07 06:46:35 and xhtml 1.0 strict Jun 07 06:49:37 03koen 07 * r1.3543 10openembedded/conf/documentation.conf: documentation.conf: typo-- Jun 07 06:49:50 and a working "see also" Jun 07 06:54:35 ibot: botmail for zecke: could you write a small doc on how to use 'groups' and 'see also' in documentation.conf? Jun 07 07:01:04 ~lart docs.python.org Jun 07 07:01:04 * ibot changes docs.python.org's permissions to 0777 and tells the world Jun 07 07:01:11 its so fscking slow... Jun 07 07:01:34 03mickeyl 07 * r1.3539.3.1 10openembedded/packages/wireless-tools/wireless-tools_28-pre6.bb: wireless tools: package libiw (again) Jun 07 07:06:24 heh.. python is crazy... Jun 07 07:11:30 bitdoc need cleaning... Jun 07 07:12:46 hi Jun 07 07:12:53 hrw|work: what is bitdoc? Jun 07 07:13:23 hrw|work: use /usr/share/doc/python then :-) Jun 07 07:13:37 schurig: a parser for documentation.conf that generates http://ewi546.ewi.utwente.nl/tmp/d/keys.html Jun 07 07:13:47 i still don't like using real attributes instead of special formatted comments like in *.conf Jun 07 07:13:54 looks like waste of memory for me Jun 07 07:14:17 koen: someone should do array.sort() one they keywords Jun 07 07:14:29 15:46 < koen> sorting it by alfabet would be near Jun 07 07:14:45 that's what I was thinking Jun 07 07:14:55 according to the comments in the code it should do it Jun 07 07:16:11 mickeyl: can you add me as developer to bitbake's SVN? My berlios name is holgerschurig. Then I'll add the sort Jun 07 07:17:22 if it will be in php then I will add sort.. python is scare... Jun 07 07:17:41 schurig: i don't know if i can do that since i'm not the manager of the project. let me check Jun 07 07:17:50 schurig: no go. only kergoth can add you Jun 07 07:18:11 that code needs cleaning Jun 07 07:18:22 * koen hands hrw a broom Jun 07 07:18:56 hrw|work: php is for basic programmers :-) Jun 07 07:19:21 schurig: after php is perl Jun 07 07:19:38 hrw|work: you're a fan of the obfuscated coding contests? Jun 07 07:19:47 hrw|work: why not program in s/390 assembler? Jun 07 07:19:55 no s/390 here ;( Jun 07 07:20:04 hrw|work: apt-get hercules Jun 07 07:20:13 thats not the same.. Jun 07 07:20:26 E: Invalid operation hercules Jun 07 07:20:37 hrw|work: apt-get install hercules Jun 07 07:20:49 koen: :-) Jun 07 07:21:26 koen: maybe your apt has only Super Cow power, but my has Brainread power :-) Jun 07 07:23:25 let's combine those :) Jun 07 07:26:18 mickeyl: I agree with you on the comments issue. I also wish it was being done differently... Jun 07 07:26:47 RP: perhaps we can convince zecke to overthink that Jun 07 07:29:44 morning Jun 07 07:33:35 good morning CosmicPenguin Jun 07 07:34:34 mickeyl: The argument was a new parser would be needed. I can't see if being difficult to write something as we don't need to use the main bitbake parser for something simple like this... Jun 07 07:34:41 morning CosmicPenguin Jun 07 07:37:13 RP: if only bitdoc would read in documentation.conf, but no other .oe/.oeclass/.bb/*.conf file, then it doesn't matter Jun 07 07:37:37 that's right Jun 07 07:53:26 hmmm... ipkg appears to have a 100 char limit on filename lengths Jun 07 07:53:36 and it doesn't do bounds checking very well Jun 07 07:53:46 CosmicPenguin: bugzilla it :) Jun 07 07:54:18 More debug is needed - but until then, I'll do a work around Jun 07 07:54:23 stupid gcov - be more accomodating! Jun 07 07:59:30 schurig: I thought the idea was to document alongside the places the variables are defined/used? Jun 07 08:01:08 RP: this hasn't happened yet, e.g. a grep -rl '\[doc\]' * only shows documentation.conf Jun 07 08:15:06 schurig: yet :-/ Jun 07 08:16:09 hello Jun 07 08:16:30 pb_: Have you looked at the pxa copypage fix rmk posted to arm-linux-kernel ? Jun 07 08:17:31 fsck.. zaurus discharged ;( Jun 07 08:17:42 I'm in trouble building handhelds-sa-2.4.19-rmk6-pxa1-hh40, I get an "build/tmp/work/handhelds-sa-2.4.19-rmk6-pxa1-hh37.4-r5/kernel/include/linux/kernel.h:10:20: stdarg.h: No such file or directory" Jun 07 08:18:34 pb_: I suspect a function declared as attribute naked and with a hardcoded restore of the pc at the end of it (in asm) is being inlined with unpredictable results... Jun 07 08:21:47 cu Jun 07 08:22:30 hmm, both bison 1.875 and 2.0 build and pass all their tests native on armeb Jun 07 08:25:46 RP: I saw his note, but I haven't really looked at it. What leads you to think tht this function is being inlined? Jun 07 08:26:16 pb_: The results of objdump -d... Jun 07 08:28:49 RP: well, does it help if you mark it attribute((noinline))? Jun 07 08:31:00 pb_: In fact its messed up in all cases. This should never really appear in a dissassembly: Jun 07 08:31:02 144: e89da830 ldmia sp, {r4, r5, fp, sp, pc} Jun 07 08:31:04 148: e89d6830 ldmia sp, {r4, r5, fp, sp, lr} Jun 07 08:31:06 144: e89da830 ldmia sp, {r4, r5, fp, sp, pc} Jun 07 08:31:08 148: e89d6830 ldmia sp, {r4, r5, fp, sp, lr} Jun 07 08:31:30 pb_: I'll mark it as noinline and see what it does... Jun 07 08:33:03 nuts - a '-' isn't valid in a package name for pkg_postinst_ ? Jun 07 08:37:29 RP: well, there are some cases where that might be legitimate, but I agree that it does look dubious. Jun 07 08:41:13 CoreDump|home: WFM, although I'm not on an up-to-date BB tree ATM Jun 07 08:41:16 err Jun 07 08:41:28 CosmicPenguin Jun 07 08:45:59 heh Jun 07 08:46:04 I haven't updated in a while Jun 07 09:19:49 sigh - guess I have to update Jun 07 09:23:01 CosmicPenguin: you need to upgrade your bitbake Jun 07 09:23:09 just did Jun 07 09:23:19 or just patch the copy you have; it's a one line change in ConfHandler.py Jun 07 09:23:21 ah, heh Jun 07 09:23:22 hi pb Jun 07 09:23:34 err.... Jun 07 09:23:36 hm when will oe move from bk? Jun 07 09:23:55 I read 1st july is the deadline Jun 07 09:24:15 pb_: I'm at revision 241 - ConfHandler.py didn't seem to change from the previous revision that I had Jun 07 09:26:13 hi woglinde Jun 07 09:26:17 CosmicPenguin: oh, hm, that's a bit odd Jun 07 09:26:36 presumably this is the busybox parse problem you're running into, or something like that? Jun 07 09:27:00 pb_: no - I'm trying to do a pkg_postinst_ for a package with a dash in the name Jun 07 09:27:10 CosmicPenguin: yah, that's what busybox does Jun 07 09:27:26 can bitbake parse that .bb file ok? Jun 07 09:27:31 kind of like in modutils_2.4.27.bb - if I bbread that, it only shows the pkg_postinst_modutils() not the others Jun 07 09:27:35 yeah, the parse is cool Jun 07 09:27:40 it just doesn't seem to show up in the metadata Jun 07 09:27:46 ah, I see. that must be a different problem. Jun 07 09:28:02 let me inspect the busybox one with bbread Jun 07 09:29:11 oh, sorry, busybox is INITSCRIPT, not postinst. Jun 07 09:29:17 heh Jun 07 09:29:27 yah, I see the same effect that you do with modutils. Jun 07 09:29:45 pkg_postinst_modutils-depmod is nowhere. Jun 07 09:30:19 FILES_modutils-depmod is okay, though, so maybe it's something weird about functions. Jun 07 09:30:26 yeah Jun 07 09:30:40 I see the same thing Jun 07 09:40:54 I'll take a look at it a bit later on. Jun 07 09:40:55 Good evening. Jun 07 09:41:09 gotta figure out why my test rig is destroying these boards first. Jun 07 09:42:36 pb_: I'll toss around and see if I can figure it out Jun 07 09:42:41 righto Jun 07 09:42:43 doubtful though - I'm not familiar with this code Jun 07 09:43:33 quit Jun 07 09:45:20 CosmicPenguin: the place to start would be BBHandler.py, and the regex that it uses to find function definitions. Jun 07 09:45:45 might want to compare that to the regex for standard variables, which is in ConfHandler.py Jun 07 09:46:35 thats a hell of a regex Jun 07 09:52:37 heh, yeah Jun 07 09:53:17 one of the reasons our parser is not the fastest is that it's made up entirely of these hairy regular expressions. Jun 07 10:18:30 pb_: well, its probably not in the __func_start_regexp__, \- is there, plain as day Jun 07 10:36:46 koen: ping Jun 07 10:36:53 pong Jun 07 10:37:05 koen: I plan to update documentation.conf Jun 07 10:37:12 ok Jun 07 10:37:21 I'll update the docs after that Jun 07 10:37:21 koen: but see the commit message of my bitdoc checkin on how a docu could look like Jun 07 10:37:43 koen: I might not make it tonight (I've to write a protocol) Jun 07 10:38:52 me can not get how innovative M$ is Jun 07 10:39:20 we've now two articles on heise.de that praise M$ for having 'soon' a way to update all M$ products Jun 07 10:39:34 just like you do Windows Updates today Jun 07 10:40:25 riiight, thats what i always wanted, my apps calling home to MS Jun 07 10:41:57 it would rock if I could debian with just one tool Jun 07 10:42:13 instead of invoking each update tool independenly... Jun 07 10:42:23 hehe Jun 07 10:42:23 kergoth: If you use M$, you trusted them... Jun 07 11:17:04 mickeyl: oh Qt4 RC1 Jun 07 11:47:47 anyone here a suse expert? Jun 07 11:48:07 CosmicPenguin: what is that? Jun 07 11:48:12 ~suse Jun 07 11:48:14 methinks suse is the name of Bill Gates' mistress Jun 07 11:48:44 heh Jun 07 11:49:27 I'm just trying to figure out if anybody has dumped the contents of this liveCD on the web so I don't have to go through all that crap Jun 07 11:49:38 Protocol: I've made the potentiometer have a resistence of ~0 Ohm that made the Transistor smell Jun 07 11:49:56 somehow my values are totally unexpected afterwards Jun 07 12:04:39 I just got bitbake from svn and downloaded openembedded-exported-2005-05-04-1003.tar.bz2. I edited the config file (MACHINE=collie,DISTRO=familiar), but I get "ERROR: nothing provides nano" Jun 07 12:05:03 i.e. no matter what I build it's not found Jun 07 12:05:19 any ideas? Jun 07 12:05:29 well, do you have the package? Jun 07 12:05:36 openembedded/packages/nano Jun 07 12:06:12 yes Jun 07 12:06:30 there's 2 bb files for 2 versions Jun 07 12:06:32 dmag: sounds like your BBPATH and/or BBFILES are set incorrectly. Jun 07 12:06:34 and set BBPATH to that too? Jun 07 12:06:38 and BBFILES in the conf? Jun 07 12:06:41 :) Jun 07 12:06:45 kergoth: :) Jun 07 12:06:52 hmm, looking.. Jun 07 12:07:37 config file has BBFILES = "/s/openembedded/packages/*/*.bb", env says BBPATH=/s/build:/s/openembedded Jun 07 12:07:43 hmm, should that be /s/bitbake? Jun 07 12:07:52 no Jun 07 12:09:29 ok, what to check next? Jun 07 12:10:18 what configuration file contains that BBFILES line? Jun 07 12:10:27 local.conf Jun 07 12:10:32 full path. Jun 07 12:10:41 it's in /s/conf/local.conf Jun 07 12:10:43 depends on where you put it :) Jun 07 12:10:52 everything is in /s, no symlinks Jun 07 12:10:58 thats not right. Jun 07 12:10:58 it belongs inside build/ Jun 07 12:11:05 oh Jun 07 12:11:12 (more accurately, "conf/local.conf" must exist in a path in BBPATH) Jun 07 12:11:23 but putting it in build makes most sense, since thats where the output of the build goes. Jun 07 12:12:05 ah. (goes to re-check webpages he got the info from..) Jun 07 12:12:10 could I please get a url that will point me at the list of task* targets? I've lost the place I read it Jun 07 12:12:22 ls openembedded/packages/meta/ Jun 07 12:12:27 tnx Jun 07 12:14:07 yup, that worked. It's doing something... Jun 07 12:33:37 is there any known problem with the libtool package? It won't compile here (failing in running the configure) Jun 07 12:35:52 not yet known Jun 07 12:38:41 Hmmm, I just learned something I didn't realize, that I need to cleanout the old build dir on a build failure (I didn't realize it woulc fail to do a clean). My libtool is still failing, but the real failure is in the staging target, because it can't find a "i686-linux-libtool" file Jun 07 12:40:16 Hm.. with a opie-image, are all the dependencies built as ipkg, or are they built into the image? Jun 07 12:40:58 ascent: yes Jun 07 12:41:17 ascent: you can configure it to do otherwise though Jun 07 12:42:02 yes to as ipkg or yes to as builtin? :) Jun 07 12:42:21 because I see tslib scroll by, but I thought that it was not included in the final image Jun 07 12:42:44 what makes you think it wasnt? Jun 07 12:43:02 I needed to install the built ipkg to the rootfs once I booted from it :) Jun 07 12:43:16 what ipkg? Jun 07 12:43:25 tslib related ipkgs Jun 07 12:43:39 tslib wont go into bootstrap by default. it goes into the gpe and opie images. Jun 07 12:43:54 the commandline ts_print, etc utils wont be Jun 07 12:44:08 I'm building opie-image now. Jun 07 12:44:24 well, I'll see in a few :) Jun 07 12:44:27 (when its done) Jun 07 12:44:42 question: The staging target of my libtool is failing, and I need to verify the value of some variables in a .bb file. I don't know the right method to be compatible with the bk files ... could I edit in a 'echo ${}' line, something like that? Jun 07 12:44:58 bbread /path/to/blah.bb Jun 07 12:45:20 is there a bbread man page/wiki page? Jun 07 12:45:32 I couldn't find it Jun 07 12:45:44 I'm not familiar with wikis, though Jun 07 12:51:06 chuckr: not much docu is present (while we're workin on it) Jun 07 12:51:13 chuckr: but bbread does not much of a docu Jun 07 12:52:36 My problem is, I'm blocked right at the start, I can't get task-bootstrap done. The libtool (package?) is failing to stage, it looks like a misuse of a dirname, but I don't know what's safe to change inthe bik environment, so I can begin to troubleshoot Jun 07 12:53:15 chuckr: the basics things. Check if you versions are current, don't have another bitbake in the way Jun 07 12:53:26 chuckr: check BBPATH, BBFILES... Jun 07 12:53:39 chuckr: use bitbake -DDD and -v to get more output Jun 07 12:53:52 if that is an AMD64 might try linux32 Jun 07 12:54:10 that's why I asked for anything yo have on e what they do, I dislike starting things just randomly hoping I won't break it Jun 07 12:54:42 it's a 686 (pentium4) environment Jun 07 12:55:39 (I'm doing it) Jun 07 12:55:49 chuckr: ok let us start over again Jun 07 12:55:50 chuckr: hi Jun 07 12:56:01 chuckr: what do you want to build? (target, distro) Jun 07 12:56:17 chuckr: what version of bitbake and what version of the openembedded package database do you use? Jun 07 12:57:41 I just refreshed both 60 minutes ago Jun 07 12:58:09 bitbake from svn, openembedded from bk Jun 07 12:58:40 ok, that run of bitbake, with -v -DDD, helped out Jun 07 12:59:23 chuckr: is it libtool or libtool-native thats failing? Jun 07 13:00:12 that's odd, the build dir is totally empty (yes, that would make it fail, wouldnt it?) Jun 07 13:00:28 libtool-native Jun 07 13:01:58 I'm still new at this, give mea little time, but I think I'll be able to find something so drastically wrong Jun 07 13:03:03 are you not running bitbake in the build directory? Jun 07 13:03:45 I am running it from the buil didr, but everything seems to happen in the /usr/tmp/bb dir that I have defined at TMPDIR Jun 07 13:04:21 which is what you'd expect. why would an empty build dir (other than conf/local.conf) be a problem? Jun 07 13:05:43 yes, mine is empty(nic eto konw that's correct) Jun 07 13:12:57 If I want to do a rebuild of a failred package, I need to ask all of the required steps. It looks like I want to clear out the sources (cvs download) file, the tmpdir's work subdir, and I now *think* maybe the stamps in tmpdir/stamps (at least the ones for that package), is that all> Jun 07 13:14:01 heh Jun 07 13:14:05 bitbake -c clean blah Jun 07 13:14:08 or Jun 07 13:14:13 bitbake -c clean -b /path/to/blah.bb Jun 07 13:14:18 thanks Jun 07 13:15:34 every time I run bitbake, it spends 2 minutes running thru all the 2400 packages again, is there any way to shortcut that? Jun 07 13:15:47 not yet :| Jun 07 13:16:05 ok, I wasn't cl;eaning it correctly Jun 07 13:16:10 -b bypasses all deps and acts on that package. if you know that all its deps have already been built, you can use that Jun 07 13:16:16 (this _is_ documented in the wiki) Jun 07 13:16:28 I didn';t realize about the stamps Jun 07 13:17:03 I' Jun 07 13:17:21 I'm back to the REAL failure, I think (ned to look again) Jun 07 13:18:15 hm.. when I bitbake opie-image, and I set distro, I get ipkgs _and_ images.. cant I get only images? Jun 07 13:19:24 you get something for free! What can suck about that? Jun 07 13:19:50 well, it confuses me what's in the image and what needs to be injected into it :) Jun 07 13:20:45 it isnt hard to ask ipkg what packages are installed. learn to use it. Jun 07 13:22:10 During my config run (and I have the failing config.log) it's faling because "source directory already configured; run make distclean". I notice in the .bb file, the do_configure has a line "autotools_do_configure", is there anything that does? Jun 07 13:22:42 chuckr: that generally happens when ./configure is run inside of a path that contains symlinks. Jun 07 13:22:48 set TMPDIR to the _real_ path Jun 07 13:23:01 sorry, it aLREADY IS Jun 07 13:23:09 apparently not. Jun 07 13:24:34 would be easier if it weren't, but it's just /usr/tmp/bb, nothing hinky Jun 07 13:25:08 no links in the environment at all, in fact Jun 07 13:25:17 at least, oe's env Jun 07 13:29:29 Hm, open embedded worked on freebsd, right? How's about netbsd? Jun 07 13:37:30 I just used the syntax "bitbake -c clean -b ", it worked, so what are the other valid targets I can use there? Jun 07 13:37:59 err... ls packages/*/*.bb? Jun 07 13:38:31 I mean, things like "install" or "stage" Jun 07 13:39:42 I just successfully did the clean, if I wanted to rebuild the target that's failing, step by step, those steps would be ... Jun 07 13:40:06 well, there's not reason to go through the steps individually if you just cleaned it Jun 07 13:40:15 just bitbake -b and go Jun 07 13:40:39 it's failing, it looksl ike configure is being called incorrectly I do need to call things with some control Jun 07 13:41:15 the stages are done in order - so if configure fails, it will fail in the configure stage Jun 07 13:41:32 its not wise to type in four commands when one will do Jun 07 13:54:54 I'm thinking that, with all the messing around I have done in the last 2 weeks with other compilers, it;s *my* environment that's causing libtool to fail, but one question: the libtool-native, it shouldn't have something in it that looks like a zaurus target (like armv5te) should it? Jun 07 13:55:58 probably yes Jun 07 13:56:18 but I'm an idiot when it comes to libtool Jun 07 13:57:24 I was contrasting in my mind, libtool versus libtool-native, would the libtool-native have only i686 stuff? Jun 07 13:58:25 the listing is all full of zaurus-type targets, I wanted to check that Jun 07 13:59:48 libtool-native would just be a standard libtool for 686 Jun 07 13:59:57 with all of our hacks to strip out certain paths Jun 07 14:00:33 moment, I just had a lightbulb go off Jun 07 14:02:52 * CosmicPenguin goes back to trying to figure out why dashes in the pkg_postinst_ override don't work Jun 07 14:03:35 CoreDump|home: look at bb.data Jun 07 14:04:20 * CoreDump|home taps CosmicPenguin on the shoulder and points to zecke Jun 07 14:04:22 zecke: you already fixed it? Jun 07 14:04:43 gah - I hate updating the OE tree Jun 07 14:04:50 * CosmicPenguin should merge some of his crap back up Jun 07 14:06:07 CosmicPenguin: no not fixed, but look there Jun 07 14:06:23 CosmicPenguin: Show me an example line? Jun 07 14:07:27 zecke: showdata on modutils_2.4.27.bb - notice that pkg_postinst_modutils-modinfo() doesn't show up in the metadata Jun 07 14:07:56 I have a very detailed listing of the failure. I *think* it's doing a full configure, I can't quite see where it eneded, but it's long enough, and the Makefile exists, and the failure at the end is because its already runa configure, it doesn't need to run it again Jun 07 14:08:56 yeah, that's it, it's trying to rerun an already successful configure Jun 07 14:09:35 chuckr: BBINCLUDELOGS = "1" Jun 07 14:09:42 chuckr: and you have log files of the build Jun 07 14:09:45 what does that do? Jun 07 14:09:55 I have full logs as it is Jun 07 14:10:00 ah okay Jun 07 14:10:15 it's very definitely incorrectly duplicatinghte configure call Jun 07 14:10:31 I just need to figure out why Jun 07 14:27:02 couold someone pelase double check me: in the openembedded/packages/libtool/libtool-native_1.5.10.bb file, inside it, it;'s got several references to the name of the .bb file excepting that those references DON'T have the "native" in it Jun 07 14:27:28 ERROR: log data follows (/home/ascent/tmp/work/alsa-tools-1.0.8-r0/temp/log.do_install.20829) Jun 07 14:27:31 | NOTE: make DESTDIR=/home/ascent/tmp/work/alsa-tools-1.0.8-r0/image install Jun 07 14:27:33 | make: *** No rule to make target `install'. Stop. Jun 07 14:27:36 | FATAL: oe_runmake failed Jun 07 14:27:58 is what I'm seeing correct practice? Jun 07 14:28:34 makes sense to me Jun 07 14:28:51 dropping the "native" is right? Jun 07 14:29:02 yes Jun 07 14:29:04 ok Jun 07 14:29:18 go read up on include Jun 07 14:30:03 I would if I could find something to read on it Jun 07 14:30:54 it's on the front page of openembedded.org Jun 07 14:31:25 it's also under the bitbake page in the wiki Jun 07 14:39:58 bleh Jun 07 14:45:44 is it just me or do all alarm companies suck? Jun 07 14:47:07 making money out of fear of theft? Jun 07 14:47:22 not being able to keep an appointment Jun 07 14:47:49 treke: move to Norway or another scandinavian country and you don't need any ;) Jun 07 14:48:03 but that might not be too practicable Jun 07 14:50:54 ~lart metric system Jun 07 14:50:54 * ibot gives metric system a "free" copy of Windows and then charges double for "Upgrades" Jun 07 14:50:59 amen Jun 07 14:51:18 my car gets 40 rods to the hogshead and that's the way I like it Jun 07 14:51:37 treke: the issue is if you once not used the metric system you're lost Jun 07 14:51:44 106,3 mV -> V Jun 07 14:51:57 I'm soo stupid I use python just to double check... Jun 07 14:52:06 the real issue is that you people put the , in the wrong place Jun 07 14:52:25 hehe Jun 07 14:52:29 s/,/./ Jun 07 14:52:53 well, I couldn't find out what was broken, but I found the problem, and discovered a useable workaround: there was, in the two configure files, a section in each (about 6 lines) that checked ont he age of config.log (grep for "already"), I removed that section from both configures, and it completed withought furhter problems Jun 07 14:53:32 zecke: no, nokia did not respond yet Jun 07 14:53:34 chuckr: I don't know if this is correct but you might just file a bug report Jun 07 14:53:52 * zecke really hopes to get one Jun 07 14:53:57 I know so little about what I[m doing, I feel incorrect about filing a bug on this Jun 07 14:54:02 I will even write Mono and Anyway on the back Jun 07 14:54:27 chuckr: little people understand the merits of libtool, with an issue you document your findings Jun 07 14:54:32 I wouldn;t mind repeating what I did, if anyone else wished to take a look or cared enough to login Jun 07 14:54:33 anyway, mono rocks Jun 07 14:54:56 koen: anyway, mono rocks, as it made me rich Jun 07 14:55:03 :) Jun 07 14:55:17 filthy rich even Jun 07 14:55:18 koen: anyway, mono rocks, because I spent all money I got from SUN on it Jun 07 14:55:59 koen: admitingly both of us are just jealous Jun 07 14:56:08 yeah Jun 07 14:57:44 errr... hello all! Jun 07 14:58:13 I was refered here by lardman on the OESF forums Jun 07 14:58:21 ah Jun 07 14:58:25 usbhost or simpad? Jun 07 14:58:37 koen! Jun 07 14:58:40 Hi! Jun 07 14:58:46 usbhost Jun 07 14:58:47 hello LurkerZ Jun 07 14:59:23 I saw you reading the thread I started Jun 07 14:59:27 earlier today Jun 07 14:59:49 any ideas? Jun 07 15:00:09 my idea was "unload the usb modules", but that wasn't so Jun 07 15:01:12 I've slowly come to believe that USB host is not meant to be working in OZ 3.5.3 yet Jun 07 15:01:46 it should be a matter of unloading and loading modules Jun 07 15:01:52 it clearly says "USB not completely working yet" in the release notes Jun 07 15:02:03 yeah Jun 07 15:02:10 that's why I said "should" :) Jun 07 15:02:16 but I could never get a straight answer on what exactly that meant Jun 07 15:02:42 Oh Jun 07 15:05:49 if I could have one wish fullfilled here, it would be to have the manual section on the parsing of variables rewritten, so that it no longer relies purely on examples to get the idea across Jun 07 15:06:47 the task-bootrap I'm running is chugging along nicely, at great long last. Jun 07 15:06:50 would be nice, but examples are probably the simplest way to do it :) Jun 07 15:07:05 too many odd cases Jun 07 15:07:14 like FOO = ${FOO} Jun 07 15:07:19 too much flexibility Jun 07 15:07:58 hey pb_ Jun 07 15:08:14 hi all Jun 07 15:08:23 treke, I had someone suggest I file a bug report, but I didn't want to (I know so little about what is supposed to occur) but if you say I should, I will, about the small sections of the configure files i was able to cut out, and fix myu problems Jun 07 15:08:44 huh? Jun 07 15:08:57 why on earth would it matter if I say you should put in a bug report Jun 07 15:09:20 zecke knows far more about the bitbake internals than I do Jun 07 15:10:11 there are two configure files in libtool. If yoiu search for the "already" and "pwd" in the same paragraph, you can quite easily locate the sections I'm referring to (I would even supply lines offsets if you wished) and cutting out those sections cleared my libtool propblems Jun 07 15:10:32 I have no plans on looking. Jun 07 15:10:41 ok, no skin off my back Jun 07 15:15:46 can I ask, I wish to check out of the packages a linux tarball, if I pick a particular one, could I please get the bitbake command that would unpack + patch it, ready for me to pick it up? I wish to use it outside of oe Jun 07 15:16:04 bitbake -c patch -b Jun 07 15:16:13 tnx Jun 07 15:18:00 could I ask, I'm pl;aying my own game here, with a build system that uses a bunch of gmake macros, so I can put my own environment together. What might be a good kernel for me to choose (assuming I can cobble together a good config file)? Jun 07 15:18:35 well, that rather depends on your platform. Jun 07 15:18:43 most of the kernels we have in OE are highly machine specific. Jun 07 15:18:47 the zaurus, the C3000 Jun 07 15:19:15 I keep forgetting I[m dealing with folks who aren't as foccussed on the zarus as I am Jun 07 15:19:15 one of the openzaurus-pxa kernels, I guess. look at the appropriate MACHINE.conf file to see which one it selects. Jun 07 15:19:24 chuckr: write your own meta package? Jun 07 15:19:38 you're kidding! Jun 07 15:20:04 chuckr: why would you write gmake macros? Jun 07 15:20:33 well, maybe after I finished getting it to compile, but putting together what I intend, then re-inserting it into oe, seems like guilding the lily Jun 07 15:22:09 because I don't need to, by build system already has them all, and it's about a quarter of the complexity of oe, and I know gmake so well already. If oe were better documented, but as it stands now, I would have to learn by reading sourfces, and I'm not quite good enogh at python to do that Jun 07 15:22:19 nearly, though Jun 07 15:22:36 your idea isn't bad. Jun 07 15:22:59 lol Jun 07 15:22:59 anyway have fun Jun 07 15:23:01 I actually have a great fondness for gmake Jun 07 15:23:12 I copied opie-image.bb to my own file, added ASCENT = "strace alsa-lib alsa-oss alsa-utils hostap-modules hostap-utils prism3-support tslib" Jun 07 15:23:20 and puth ${ASCENT} behind DEPENDS and IPKG_INSTALL, but it hangs on alsa Jun 07 15:23:27 | ERROR: Cannot satisfy the following dependencies for alsa-utils: Jun 07 15:23:27 | ${ALSA_PACKAGES} Jun 07 15:23:30 what did I forget :) Jun 07 15:23:36 I like python, but the size ofhte oe codebase means that it would take ma a good while to learn it all Jun 07 15:23:55 heh, oe exists because gmake based systems didn't come close to meeting our needs. the maintainance load is far too high, and it doesnt scale nearly well enough Jun 07 15:24:19 otoh, perhaps taking an oe project and bending it my way, that might be a workable idea Jun 07 15:24:27 ascent: that looks like just general bogosity in alsa-utils.bb Jun 07 15:24:29 me Jun 07 15:24:44 means dropping all my nice work so far Jun 07 15:25:00 pb_: oh. hm.. Jun 07 15:25:25 I want alsa to be in my image :) Jun 07 15:25:28 I don't know where ALSA_PACKAGES is meant to be defined, but it isn't in alsa-utils.bb itself and I don't think it is in any standard .conf file. Jun 07 15:25:51 I like chattering about it, and might be boring you. Seriously, if I am, don;'t reply to me, it's the best way to shut me up Jun 07 15:25:52 I recommend you check the bitkeeper history to find out where that line came from, and send hate-mail to whoever added it. Jun 07 15:26:00 pb_: lol :) Jun 07 15:26:15 Maybe I should set ALSA_PACKAGES myself, to "" ? :) Jun 07 15:26:28 could be in a collections package Jun 07 15:26:47 make sure something isn't being included that isn't in the path Jun 07 15:26:48 possibly, but alsa-utils.bb wouldn't be able to see it in that case. Jun 07 15:26:51 it aint in opie-collections.inc Jun 07 15:26:58 variables aren't shared across packages. Jun 07 15:26:59 (at least, it's included, and still gives the error) Jun 07 15:27:18 how big will my tmp become while task-bootrap is b uildng? Jun 07 15:27:27 pretty big :) Jun 07 15:28:10 how many gigs to the nearest, oh, 5? Jun 07 15:28:25 something like that Jun 07 15:28:42 ok, Ihave enough Jun 07 15:28:44 that much? Jun 07 15:28:46 let's see. Jun 07 15:29:09 after it's finshed, what can I trim? Jun 07 15:29:21 pb_: do you think there is request to clean each package in the workdir after beeing built? Jun 07 15:29:28 chuckr: anything Jun 07 15:29:38 zecke: that will break stuff with pkgconfig Jun 07 15:29:41 I know there isn't Jun 07 15:29:44 or libtool Jun 07 15:29:52 one of the two is being a pain in OE Jun 07 15:29:52 koen: these packages are broken then Jun 07 15:29:53 zecke: yah, people ask for that feature every so often. It would be a neat thing to do. Jun 07 15:30:04 koen: pkgconfig should install stuff to the pkg-site Jun 07 15:30:14 I think hrw had some kind of class file to do it automatically. Jun 07 15:30:20 rm_work Jun 07 15:30:28 yeah Jun 07 15:30:37 which breaks a gpe build at the moment Jun 07 15:30:46 or a !opie build Jun 07 15:32:06 koen: pkg config should be easy to fix? Jun 07 15:32:06 zecke is right, those packages are just broken. Jun 07 15:32:20 koen: I dunno how to create 'right' .la files though Jun 07 15:32:49 zecke: lock kergoth in a room with sed and some food Jun 07 15:33:19 I'm curious, oe having 3 different scm's, doesn't that bother anyone? Jun 07 15:33:26 3? Jun 07 15:33:36 bk and svn are 2 in my book Jun 07 15:33:38 cvs, subversion,and bk Jun 07 15:33:43 cvs? Jun 07 15:33:58 oe has one scm. bitbake has one scm. their needs are quite different. Jun 07 15:34:01 well, it's a scm Jun 07 15:34:04 [ascent@192 packages]$ du -sm ~/tmp Jun 07 15:34:04 4307 /home/ascent/tmp Jun 07 15:34:18 that's after opie-image, which is much larger Jun 07 15:34:53 I wasn't proposing any one in particular, just wishing that there were fewer tools needed Jun 07 15:35:02 What is this deal with everybody getting all worked up about scms? They're different tools, they useful, and you should know how to use them all regardless Jun 07 15:35:41 CosmicPenguin: Jun 07 15:35:53 it's not everybody, I was just suggesting that 3 scms mbiht be one or two too many Jun 07 15:35:55 ~lart zealots Jun 07 15:35:59 there arent 3. Jun 07 15:36:03 get your facts straight Jun 07 15:36:11 bitbake is stored in svn, oe is stored in bk. Jun 07 15:36:25 ok Jun 07 15:36:35 once again, their needs are different. Jun 07 15:36:42 there was in the past a cvs mirror for the free zealots Jun 07 15:36:54 right, that was just so people could get sources without using bk Jun 07 15:36:56 how do you mean that? Jun 07 15:36:57 mirror only Jun 07 15:37:12 amazing how much people were working on svn Jun 07 15:37:28 "no I have patches for svn" Jun 07 15:37:41 I used to dislike svn, it was very fragile, but it looks like it outgrew that Jun 07 15:37:59 it is mature Jun 07 15:38:46 oh, and why not either fold bitbake into bk, or openembedded into svn (I hope I don't have that backwards) Jun 07 15:39:03 we folded out bitbake Jun 07 15:39:13 we want a distributed scm for OE Jun 07 15:39:16 what part of "they have different needs" do you not understand? Jun 07 15:39:20 bitbake didn't need that Jun 07 15:39:22 how many fucking times do i have to repeat this before it sinks in? Jun 07 15:39:28 kergoth: join #oe-bofh ;) Jun 07 15:39:29 actually, that was my question Jun 07 15:40:07 bitbake was put in svn because it didn't need a distributed scm and to stop the whiny "bk is evil" bitches Jun 07 15:40:56 for OE we want to work offline *with* complete history Jun 07 15:41:04 svn doesn't do that Jun 07 15:41:15 I'm not much of a bk person, so I'm learning here Jun 07 15:41:29 I'm not against, it, merely ignorant of it Jun 07 15:41:33 we also need to easily do star merges, pushes and pulls between independent "offline" or online repositories Jun 07 15:41:41 there are distributed scms that dont do that part well Jun 07 15:41:45 well, "distributed" Jun 07 15:41:55 ~lart kergoth for arguing when he could be looking at scms Jun 07 15:41:55 * ibot hauls kergoth up by the scruff of the neck and spanks him until he waddles for arguing when he could be looking at scms Jun 07 15:42:00 hehe Jun 07 15:42:09 you be productive Jun 07 15:42:12 we will abuse chuck Jun 07 15:42:34 distributed larting Jun 07 15:42:45 ~dlart Jun 07 15:43:14 damn Jun 07 15:43:30 'night all Jun 07 15:43:42 nigh t Jun 07 15:43:55 | Collected errors: Jun 07 15:43:55 | Cannot find package tslib. Jun 07 15:43:55 | Check the spelling or perhaps run 'ipkg update' Jun 07 15:43:57 koen|sleep: http://gtkperf.sourceforge.net/index.php?page=main Jun 07 15:44:04 while bitbake tslib works.. Jun 07 15:44:11 but putting it in my own image seems to fail Jun 07 15:44:21 libts Jun 07 15:44:26 debian style naming Jun 07 15:44:32 oh :) Jun 07 15:44:36 thats.. new :) Jun 07 15:44:54 zecke: http://projects.o-hand.com/xresponse/ Jun 07 15:44:55 * koen|sleep zzzz Jun 07 15:45:02 gnight Jun 07 15:45:09 koen|sleep: have seen that as well Jun 07 15:45:44 somehow it does not work with Opie/QtE Jun 07 15:45:46 that crazy mallum Jun 07 15:45:53 always thinkin up something new Jun 07 15:46:09 CosmicPenguin: :) Jun 07 15:46:17 that crazy mallum, up late too... Jun 07 15:46:23 CosmicPenguin: :D Jun 07 15:46:49 | Cannot find package tslib. Jun 07 15:46:49 | Check the spelling or perhaps run 'ipkg update' Jun 07 15:46:50 hmz. Jun 07 15:48:22 ascent: kergoth already told you what the problem is Jun 07 15:48:31 one item: someone suggested I use BBINCLUDELOGS=1, I assume it means it drops log files, where, named what? Jun 07 15:48:40 er koen Jun 07 15:49:05 treke: he suggested tslib, but tslib fails too Jun 07 15:49:29 or am I missing some point here :) Jun 07 15:49:34 read what he said again Jun 07 15:49:46 eh Jun 07 15:49:47 hang on :) Jun 07 15:49:51 *coffee* Jun 07 15:50:13 OT: someone familiar with tex and \minipage? Jun 07 15:50:31 only the first.. never used minipage. sorry Jun 07 15:51:10 hm.. "ERROR: Nothing provides libts" Jun 07 15:51:22 ascent: actually I've ~10 pictures and I want two side by side... Jun 07 15:51:27 ascent: DEPENDS = "tslib" Jun 07 15:51:36 ascent: RDEPENDS = "libts" Jun 07 15:51:50 hum.. that's .. odd :) let's try. Jun 07 15:52:15 ascent: but if you build Qt/E, libts will be automatically built and installed Jun 07 15:52:23 ascent: YOu want to build tslib Jun 07 15:52:33 root@spitz:~# modprobe usb_ohci_pxa27x Jun 07 15:52:33 Using /lib/modules/2.4.20/kernel/drivers/usb/usb_ohci_pxa27x.o Jun 07 15:52:33 insmod: init_module: usb_ohci_pxa27x: Operation not permitted Jun 07 15:52:33 modprobe: failed to load module usb_ohci_pxa27x Jun 07 15:52:39 ascent: and it will be split (debian style) into tslib-dev, tslib-doc Jun 07 15:52:40 probably. I want ts_calibrate to be in my image Jun 07 15:52:43 LurkerZ: dmesg Jun 07 15:52:45 (ooops) Jun 07 15:53:03 ascent: that depends on your needs Jun 07 15:53:10 LurkerZ: ok so it is Off Topic here? Jun 07 15:53:22 LurkerZ: or is this progress (2.4.20 sounds old though) Jun 07 15:53:32 zecke: well, I just took the opie-image file, and tried to get ts_calibrate inside the image Jun 07 15:53:43 that's the main purpose of my struggle :) Jun 07 15:53:51 (And of course add a few others as well) Jun 07 15:53:55 ascent: yeah the RDEPENDS += " ts_calibrate " Jun 07 15:54:14 ascent: DEPENDS are - as I understand them - build dependencies Jun 07 15:54:18 Ok, well, dmesg says nothing Jun 07 15:54:26 and rdeps, recursive deps? Jun 07 15:54:31 runtime Jun 07 15:54:33 ah Jun 07 15:54:38 nothing about usb_ohci Jun 07 15:54:38 ascent: Remote, Runtime, Rrrr Jun 07 15:54:42 :) Jun 07 15:54:46 prrrrrrrrrr :) Jun 07 15:54:47 LurkerZ: then happy debugging Jun 07 15:55:29 I was going to ask if there were any Zaurus C1000/C3000 users here before I posted that Jun 07 15:56:13 cause I am not even sure if USB host is supposed to be working yet with OpenZaurus 3.5.3 (Opie 1.2) Jun 07 15:56:15 LurkerZ: I'm sadly not Jun 07 15:57:28 And, sadly, I cannot debug ;-) Jun 07 15:58:22 LurkerZ: why? Jun 07 15:58:35 LurkerZ: did sharp implement USB OTG? Jun 07 15:58:44 Yes Jun 07 15:59:12 I apologize for asking in this channel Jun 07 15:59:30 I did not realize that this is intended for developers Jun 07 15:59:40 until just now Jun 07 16:00:03 I am just a simple end-user :\ Jun 07 16:01:05 It is lardman's fault, though; he pointed me here :P Jun 07 16:02:58 I would be eternally grateful if anyone who thinks they can help me could pop on over to #openzaurus Jun 07 16:04:22 LurkerZ: do not apologize Jun 07 16:04:34 LurkerZ: most of us simply can't anything Jun 07 16:04:58 isn't that the truth Jun 07 16:05:27 Do any of the other handhelds to which OE has been ported to have USB OTG? Jun 07 16:06:08 CosmicPenguin: obvioulsy 'say' was missing but both sentences are true Jun 07 16:06:17 * zecke is tired... and heads to bed Jun 07 16:07:39 bye zecke Jun 07 16:07:41 zecke: thanks for the help Jun 07 17:11:57 finally,the next problem, in linux-jline ... any reported problems with htat? I had an error with bad CC options (made little sense) Jun 07 17:16:36 you know, I just looked at the .bb file for linux-jlime, I just can't figure out why it was selected; the comments say it's for the HP Jornada, and that's nowhere near my listed target Jun 07 18:32:43 as far as targets so in the bitbake command "bitbake -c $(target)", what comes after patch (I need them up to compile) Jun 07 18:32:48 what timezone is germany generally in? Jun 07 18:32:54 +3 +4? Jun 07 18:35:06 I'm sorry, the bitbake docs call them "tasks", but didn't name them, what coems after patch? Jun 07 18:35:48 the tasks arent in the bitbake docs, because bitbake doesnt define them. Jun 07 18:35:52 OE does, in base.bbclass. Jun 07 18:35:58 hi kergoth Jun 07 18:40:55 mithro: +1 during winter, +2 during summer Jun 07 18:41:25 kergoth, I'm trying to assemble (for experimentation) two sets of kernel sources (a 2.4 and a 2.6), using a "bitbake -c " command, which task would you say I ought to do to prepare the sources for myself, patch or, uh, configure? Jun 07 18:41:55 "prepare the sources" means what? Jun 07 18:42:29 I want to be able to visit my tmpdir, and be able to glom the config files and the source files Jun 07 18:43:15 I see the new machine/wrt54.conf and wrt-image... those for WRT54G routers? Jun 07 18:43:26 I'm not going to hold this against you, obviously, because I['m not totally sure what I;m doing, on my own side of the project Jun 07 18:44:03 so I guess I don't really know, kergoth, what the configure target does, so ZI don't know if I require it Jun 07 18:44:22 if you aren't sure, forget it Jun 07 18:44:39 (how could you be sure?) Jun 07 18:46:48 Luke-Jr: most likely :) Jun 07 18:47:18 treke: fun; any idea if they include the Linksys apps? Jun 07 18:47:34 offroadgeek: hey Jun 07 18:47:50 Luke-Jr: I would highly doubt it Jun 07 18:47:56 hey luke-jr Jun 07 18:48:02 treke: why? :| Jun 07 18:48:13 license reasons. Jun 07 18:48:25 last I checked, they were GPL Jun 07 18:48:38 then add them in Jun 07 18:48:44 or at least the majority of forks use them Jun 07 18:48:50 that doesnt mean jack Jun 07 18:49:07 treke: usually, it means it's ok Jun 07 18:49:09 no Jun 07 18:49:14 it means no one really cares Jun 07 18:49:17 there is a difference Jun 07 18:49:33 In most cases I've seen, it would mean it's ok... Jun 07 18:49:44 not sure on this particular case Jun 07 18:50:07 it just means you aren't likely to get sued Jun 07 18:50:32 not that it's ok Jun 07 18:53:25 Luke-Jr: the wrt support in oe is pending making the kernel build sanely from oe. the kernel build is an absolute nightmare, referencing header files from outside the krenel tree, headers which have no stated copyright or license information in their headers or associated files. Jun 07 18:53:39 obi: thanks Jun 07 18:53:55 way to go linksys Jun 07 18:54:07 kergoth: if such headers reference GPL-only exports, they'd need to be GPL, no? Jun 07 18:54:09 well, way to go whomever linksys paid to do this Jun 07 18:54:14 CyberTAN? Jun 07 18:56:41 Luke-Jr: they dont. the kernel files reference bits in them, not the other way around. i'm not clear on the situation in that case. but then again, their code would end up linked into a kernel image, so would be infected. regardless, its an ugly, messy situation Jun 07 18:57:16 kergoth: IIRC, GPL code can only reference compatible license libs/code, also Jun 07 18:57:35 and I don't think there's exceptions in Linux for that Jun 07 18:57:52 what part of "so would be infected" did you not get? Jun 07 18:58:01 I was agreeing Jun 07 18:58:05 heh Jun 07 18:58:18 sounds like it just boils down to "messy" Jun 07 18:58:22 anyway, not having a stated license is crap Jun 07 18:58:26 and legally messy Jun 07 18:58:55 do I really disagree with you so often that it seems necessary to assume a statement is meant to disprove you? =p Jun 07 18:59:13 habit :P Jun 07 18:59:42 boo Jun 07 18:59:49 kergoth: kergie! Jun 07 19:00:19 hmm... even if the license was prohibitive, wouldn't it be possible to comply by telling oesources not to mirror it? Jun 07 19:00:25 treke|home: al licensing is messy :) Jun 07 19:00:28 and always fetch directly from Linksys Jun 07 19:00:33 no Jun 07 19:00:47 because without the license you dont entirely know if you can build it Jun 07 19:01:04 Copyright does not prohibit usage, only distribution Jun 07 19:24:24 Luke-Jr: you still have to be legally able to distribute the binaries. Jun 07 19:24:35 the license needs to allow that, at a minimum Jun 07 19:24:46 but yes, there are items in oe today that we distribute binaries of only, and are excluded from the mirror Jun 07 19:30:58 erm Jun 07 19:31:00 well no Jun 07 19:48:13 apparently that flex patch was necessary for arm/armeb compilation Jun 07 19:48:40 adding it back in (well, reverting to the flex bb's from a few weeks ago, including the patch) causes flex to build for armeb again Jun 07 19:49:08 so can we get it back in please? Jun 07 21:56:33 kergoth: you're assuming someone will be distributing the binaries Jun 07 21:56:39 kergoth: that need not be the case Jun 07 21:57:14 if the distro is going to require that every user build it themselves, sure Jun 07 21:58:36 * Luke-Jr nods Jun 07 21:58:48 there is also the unlikely chance the user is out of the control of copyright laws :) Jun 07 21:58:58 maybe Russia? Jun 07 22:20:44 I'm trying to use bitbake to build nano (as a test). When building quilt-native (the first thing I suppose), it tries to download some random file (such as bluez-utils-2.15.tar.gz) and then quilt-native fails to build (of course). Any idea what I'm doing wrong? **** ENDING LOGGING AT Tue Jun 07 23:59:56 2005