**** BEGIN LOGGING AT Thu Jul 06 02:59:57 2006 Jul 06 03:23:44 v8jlene: go ahead :) Jul 06 06:16:49 morning Jul 06 06:17:10 JustinP: its 32bit debian testing Jul 06 06:17:27 ~lart CoreDump for broken openssh Jul 06 06:17:27 * ibot declares CoreDump a moron for broken openssh Jul 06 06:18:44 hrw: morning Jul 06 06:19:38 morning... Jul 06 06:19:43 hi tkp Jul 06 06:19:48 tkp: how goes build? Jul 06 06:20:16 hey, I left oe building bootstrap-image overnight Jul 06 06:20:27 It's still on do_populate_staging for glibc Jul 06 06:20:39 7 hours? Jul 06 06:20:44 tkp: you build on pentium1? Jul 06 06:20:56 it's in a vm actually Jul 06 06:21:05 but hosted on a fairly decent pc Jul 06 06:21:36 jkp left his going too... started at the same time Jul 06 06:21:44 and his is in the same place Jul 06 06:21:51 on a brand new Intel mac Jul 06 06:22:00 (in a debian vm) Jul 06 06:22:17 something must be up Jul 06 06:22:53 make has been running for 387 minutes says top Jul 06 06:23:22 never tried to run OE in emulated Jul 06 06:24:02 well, stracing the pid shows lots of activity Jul 06 06:25:35 * v8jlene seems to recall make looping issues in glibc a few weeks back... Jul 06 06:25:49 hmm. I think thats whats happening Jul 06 06:25:57 sh|t :( Jul 06 06:26:23 v8jlene: recall a way round it? Jul 06 06:26:54 tkp: It wasn't me... but I'm looking back at the irc logs now to see what it was about... Jul 06 06:27:09 this is glibc-2.3.5+cvs20050627-r7 Jul 06 06:27:20 tkp: that one works quite ok Jul 06 06:29:11 * tkp wonders how he can find out what it's playing at Jul 06 06:29:20 tkp: Ok, unrelated... it was glibc 2.4 looping due to -isystem changes which were then removed... Jul 06 06:29:37 tkp: look in tmp/work/*glibc*/temp/ dir Jul 06 06:31:34 well I'm not sure what to look for in there, but the directories size is slowly increasing Jul 06 06:32:10 BTW, I have glibc 2.4 for x86 working now. Jul 06 06:32:21 v8jlene: was just going to ask! Jul 06 06:32:22 ahhh - we should try that Jul 06 06:32:38 v8jlene: thanks for the .info link btw Jul 06 06:32:46 just saw it now....looks handy Jul 06 06:33:06 morning all Jul 06 06:33:16 hi dirk Jul 06 06:33:28 hey Marcin Jul 06 06:33:45 hrw: still using wlan-ng ??? Jul 06 06:34:00 There's a few changes needed for glibc 2.4... and I have two fixes... and I'm not 100% sure which is better... just testing the second now (takes a long time!) to make it works as well... Jul 06 06:34:13 do13_: to first suspend Jul 06 06:34:37 v8jlene: has been committed? Jul 06 06:34:45 do13_: on resume khubd OOPS Jul 06 06:35:24 tkp: Nope. Such changes tend to get noticed if they break something that currently works ;) Jul 06 06:35:55 hrw: these are the reasons for starting 2.6 development after a few days with 2.4 Jul 06 06:37:12 understable Jul 06 06:38:22 tkp: First change is to remove the #rm -rf ${S}/bits from glibc_2.4.bb - that looks like it's a stray from the .bb it was derived from and shouldn't be in there anyway. But I want someone using glibc 2.4 on arm to test that before I consider commiting it.. Jul 06 06:39:39 hrw: btw I booted 2.6 opie image on collie Jul 06 06:39:53 and it works? Jul 06 06:39:57 tkp: Second change there's two options. One is to use make 3.81 - I'm still verifying this but it builds at least. Other option is to add libc_cv_386_tls=${libc_cv_386_tls=yes} to the site file (linux/i586-linux) for example. Jul 06 06:40:16 hrw: It has a lot of issues Jul 06 06:40:25 sounds like the first is cleaner v8jlene Jul 06 06:40:26 03hrw 07org.oe.oz354x * rbcfe5d9b... 10/packages/openssh/ (11 files in 5 dirs): Jul 06 06:40:26 openssh: removed RREPLACE/RCONFLIT dropbear, dropped old version Jul 06 06:40:26 - openssh packages does not replace dropbear correctly Jul 06 06:40:26 on image create time I got only openssh-sshd instead of ssh/sshd/scp Jul 06 06:40:51 do13_: it was expected Jul 06 06:40:53 which do you prefer? Jul 06 06:41:23 hrw: first mistake: .dev build the qte-mt libs as default and there are no free space in flash :) Jul 06 06:41:52 do13_: collie future is in root-on-card anyway Jul 06 06:42:06 full ack Jul 06 06:42:14 jkp_: The make change was recomended on another list... it results in lots of TLS errors into the log files... but it seems to be compiling fine.. Jul 06 06:42:32 v8jlene: on another note Jul 06 06:42:33 "A large number of packages will therefore fail to work correctly if you do not do a umask 022 prior to building." Jul 06 06:42:43 this is from your pages - is that still relevant? Jul 06 06:43:07 jkp_: Does the result actually work... apparently it should... I'm waiting for the build to finish... Jul 06 06:43:14 do13_: what do you think about having page in OZ wiki with info about 2.6 status on zauruses? Jul 06 06:43:47 v8jlene: ok - im thinking im going to try a build with the first fix...was there a reason why bits/* was removed in the frist place? Jul 06 06:44:37 hrw: good idea. so we have this info on one page instead of multiple pages Jul 06 06:44:55 jkp_: The claim was that directories were moved around in cvs and some old directories were bein gleft behind on peoples system and preventing the build... or something like that. I'm not really sure... that was before my time here... Jul 06 06:45:14 jkp_: Actually, I have the old checkouts... so it was around the time I started here ;) Jul 06 06:45:29 heh Jul 06 06:45:39 but it sounds irrelevant for 2.4 anyway though no? Jul 06 06:46:07 jkp_: Yeah, pretty sure it was just left there when copied over from a cvs .bb file. Jul 06 06:46:37 do13_: I will write startup Jul 06 06:46:48 hrw: thx Jul 06 06:46:58 jkp_: Currently arm is probably the only thing that used glibc 2.4 and it's not in the main glibc, it's in ports and they copy in lots of bits stuff manually... so this rm shouldn't effect arm (as far as I can tell). Jul 06 06:48:16 am i right in thinking that its only one of these changes i need? Jul 06 06:48:31 ie - i either need to remove the rm command OR i need to move to newer make? Jul 06 06:48:33 do13_: fyi, i just received a SL-6000 and have arm kernel hacking experience; if there's anything i can do to contribute in that capacity, let me know Jul 06 06:51:45 jkp_: http://pastebin.ca/80260 Jul 06 06:52:13 jkp_: That's my notes that I was keeping to paste into bugtracker once I'm happy that I know what I'm talking about... Jul 06 06:52:21 excellent :) Jul 06 06:52:28 ...when it loads :p Jul 06 06:53:08 jkp_: Oh, you must have NPTL for glibc-2.4 - MACHINE="x86" has this comment out, MACHINE="epia" is ok. Jul 06 06:53:24 zwelch: cool. Let me think about it. Jul 06 06:53:43 ok - we're on epia so thats good Jul 06 06:53:58 i just realised what the second fix you are trying to do fixes....thats good you are working on that Jul 06 06:53:59 do13_: one problem that i have to fight is the camera module, for which it appears there only exists a binary module Jul 06 06:54:11 means we should be able yo build native on the AMD64 then Jul 06 06:54:16 (and for 2.4... the agony!) Jul 06 06:54:33 zwelch: sorry, I do not have this module Jul 06 06:55:01 jkp_: Yeah, I'm building on x86_64 (FC4) Jul 06 06:55:11 that's okay, just advising you on one of my areas of kernel interests Jul 06 06:55:17 we resorted to VM's running debian (32bit) Jul 06 06:56:12 zwelch: I can recommend good tool: http://www.idapro.com/ Jul 06 06:57:13 zwelch: I used this to analyse the collie 2.4 sd driver Jul 06 06:58:36 do13, RP, mickeyl: http://wiki.openzaurus.org/KernelInfo26/MachineStatus Jul 06 06:59:30 do13_: thanks for that tip Jul 06 07:01:01 hrw: thx Jul 06 07:01:30 do13_: register and edit Jul 06 07:03:53 do13_: do you know that tosa keyboard had to have japanese keys? Jul 06 07:03:56 jkp_: Looks like make is looping when using 3.81... so I'd suggest going the site file change way and make sure you aren't using 3.81. Jul 06 07:04:25 jkp_: Actually, scratch that... Jul 06 07:04:29 heh Jul 06 07:04:51 jkp_: Looks like I just completed do_stage afterall ;) Jul 06 07:08:05 jkp_: I takes around 1hr from a clean start (rm -fr tmp) for me to get to the end of the libc-2.4 build.. that's with gjc disabled (takes about 4 times as long with that enabled). This is an AMD X2 4400+/2G. Jul 06 07:08:09 cu Jul 06 07:08:37 hrw: Ive never seen a tosa with japanese keys Jul 06 07:09:27 v8jlene: gjc? Jul 06 07:09:52 tkp: Java compiler thingy. Jul 06 07:10:11 zwelch:Is it worth the effort to port the camera driver to 2.6, because of the resolution? Jul 06 07:11:00 is there alternative hardware available? Jul 06 07:11:20 i mean, if we can get the specs, it's not a huge effort Jul 06 07:11:34 if we can't, then it's worth finding hardware for which specs can be obtained Jul 06 07:12:19 v8jlene: gcj is part od gcc right? how do you disable it? we have no need for a java compiler Jul 06 07:12:21 zwelch: use a usb camera? Jul 06 07:12:25 tkp: http://www.openembedded.info/wiki/New_targets#distro - I set LANGAUGES to "c,c++" Jul 06 07:13:52 tkp: Yeah, it's part of gcc. My build with make-3.81 just completed. Jul 06 07:14:16 v8jlene: on your FC4 64bit? Jul 06 07:14:21 tkp: Yep. Jul 06 07:14:27 nice Jul 06 07:16:21 tkp: I haven't actually tested the resultant libraries... it's taken enough time to just get them to build ;) And i'm not really setup for x86. Now I need to work on getting it to build for sh4 which is what I really care about ;) Jul 06 07:17:02 v8jlene: were going to set it running on our FC5 64bit box and test Jul 06 07:19:05 tkp: Were you going to use a newer make or change the site file? Jul 06 07:19:40 v8jlene: were on 3.80 on our box Jul 06 07:19:41 newer make maybe? Jul 06 07:19:43 it sounds easier Jul 06 07:20:12 v8jlene: with the newer make thing, we still need to comment out the removal of bits/* ? Jul 06 07:20:29 jkp_: Yes, the remove of bits/* has to go in either case. Jul 06 07:20:32 ok Jul 06 07:20:39 so its remove bits, and update make Jul 06 07:20:45 and we should be good to go Jul 06 07:20:53 *should* :p Jul 06 07:21:34 good morning all Jul 06 07:21:34 For the newer make I just copied each of the make .bb's to the 3.81 versions, the built make-native before starting the build of anything else. The native packages are put in the path before everything else. Jul 06 07:21:57 ie, so I didn't actually upgrade make on my system, although that should have worked fine as well. Jul 06 07:22:27 koen: morning Jul 06 07:23:43 jkp_: And set all the glibc versions to 2.4 - as in the pastebin entry... Jul 06 07:23:49 yeah, done that bit Jul 06 07:24:10 im testing in a VM so i set it going already with just the bits/* change Jul 06 07:24:19 tkp is going to do native on x86_64 Jul 06 07:26:13 v8jlene: I just literally need to duplicate and rename the two make bb files (rename to 3.81)? Jul 06 07:26:26 jkp_: It may or may not work with just that change on 32-bit. I suspect it'll work but I woudln't bet on it! Jul 06 07:26:27 tkp: Yep. Jul 06 07:26:41 do I need to duplicate and rename the make-3.80 directory too? Jul 06 07:26:54 heh Jul 06 07:26:55 tkp: Yeah, directory as well. Jul 06 07:27:42 tkp: No editing of anything needed. And when you do bitbage make-native it'll pick up the new version. Jul 06 07:28:46 v8jlene: tkp's just failed glibc initial with the picking up the wrong patches issue Jul 06 07:29:04 how did you solve that? its using the CVS patches instead of the 2.4 patches Jul 06 07:29:25 the way i solved that was to hack it amd move the cvs folder of patches, it then fell back on the 2.4 folder Jul 06 07:29:36 but thats wrong - you must have hit this issue no? Jul 06 07:29:48 jkp_: If you have the must current version from monotone it should be fixed? Have you done an "mt pull;mt update" since we fixed that? Jul 06 07:29:56 nope Jul 06 07:30:05 hmmm Jul 06 07:30:39 jkp_: Ah, so you don't have the fix. It was just changing FILESDIR near the top of glibc-2.4. Par tof it says glibc-cvs when it should be glibc-2.4 Jul 06 07:30:51 Near the top of glibc-2.4.bb that is. Jul 06 07:31:38 v8jlene: if we mt-pull / mt-update in just the glibc dir it should work right? Jul 06 07:31:50 want to avoid having to update everything maybe Jul 06 07:31:59 maybe we should just do the whole lot, not sure Jul 06 07:32:33 jkp_: I don't think you can get just bits... Jul 06 07:32:37 ok Jul 06 07:33:12 v8jlene: ours looks like: Jul 06 07:33:13 FILESDIR = "${@os.path.dirname(bb.data.getVar('FILE',d,1))}/glibc-2.4" Jul 06 07:33:19 no nmention of cvs in there Jul 06 07:34:10 tkp: Yep, looks fine. Jul 06 07:34:28 tkp: What error did you get again? Jul 06 07:34:48 v8jlene: http://pastebin.ca/80277 Jul 06 07:36:19 tkp: You might need to clean that before building: bitbake -c clean glibc-initial Jul 06 07:36:36 I had done rm -rf tmp Jul 06 07:37:02 re Jul 06 07:37:11 tkp: Yeah, that would have done it. Jul 06 07:38:14 v8jlene: just ran clean and re bitbaked... same error Jul 06 07:38:21 do13_: run diagnostics on tosa and select 'keyboard check' - some keys and named differently to real keys Jul 06 07:38:26 maybe we really should mtn pull Jul 06 07:38:26 tkp: You din't mess something up with moving directories around trying to fix this before? Jul 06 07:38:46 v8jlene: that was jkp... I never touched them on mine Jul 06 07:40:11 hey hrw|work Jul 06 07:40:34 mine fails in the same place v8jlene Jul 06 07:41:34 Sorry, it was FILESPATH in glibc-initial_2.4.bb that needed to be changed... Jul 06 07:42:18 ah, ok Jul 06 07:43:24 http://www.openembedded.org/viewmtn/revision.psp?id=27e5267a63cc6eabfebf01fa54915c4307784a99 Jul 06 07:43:25 morning Jul 06 07:43:35 hrw|work: will try after work Jul 06 07:43:40 hey koen, XorA Jul 06 07:43:42 v8jlene: patches work now ta :) Jul 06 07:43:54 Hey Dirk Jul 06 07:44:29 hi Jul 06 07:44:55 do you use some test/benchmarking software ? Jul 06 07:47:45 koen: That "rm -rf ${S}/bits" in glibc_2.4.bb breaks anything target not in ports. Could you check that ARMs ok with it gone (no hurry)? I think you'd be the main one currently using glibc 2.4. Jul 06 07:48:29 arm is in ports, so I guess it would break? Jul 06 07:48:37 (as is linuxthreads) Jul 06 07:49:20 koen: you missed a not in there Jul 06 07:49:41 koen: I think things in ports will be fine either way. Anything *not* in ports - like x86 and sh, don't work with that there. Jul 06 07:50:51 feel free to remove it Jul 06 07:51:09 we can alays add do_munge_append_arm = "rm -rf bits" if needed Jul 06 07:51:18 always* Jul 06 07:52:00 koen: Ok, just didn't want to go pulling a working glibc out from under you ;) I guess this qualifies for a PR bump as well? Jul 06 07:53:04 it doesn't alter the resulting package, right? Jul 06 07:53:35 PR should only be needed is the resulting package/binary is (or could be) different Jul 06 07:54:16 s/PR/PR bump/ Jul 06 07:54:46 koen: For x86 and sh it's no change because they currently don't build at all! For arm, I don't know... it's doing all sorts of copying bits related stuff around in the .bb file ;) I don't believe so... Jul 06 07:55:54 jkp_/tkp: I take it's progressing ok now? Jul 06 07:56:05 yep :) Jul 06 07:56:06 aye - im on glibc-intermediate Jul 06 07:56:14 just about to set it going on FC5 Jul 06 07:57:29 Cool. Lets hope things built with it all work ok. I probably should retry with gcc 4 as well now that it seems to be working with gcc 3.x. Jul 06 07:58:25 that would be cool as well Jul 06 07:58:43 would be nice to know if we could use a later compiler Jul 06 07:59:07 It's not a real priority for me since gcc 4.x can't build working kernels for sh4 ;( Jul 06 07:59:35 Time for dinner now, I'll drop back in a few hours and see how things went. Jul 06 07:59:58 nice one v8jlene, will let you know how it goes Jul 06 08:12:16 hi folks Jul 06 08:20:10 hey ade|desk Jul 06 08:48:34 koen: NOTE: package gimp-2.3.10: completed Jul 06 08:50:24 XorA: cool Jul 06 08:54:54 * XorA so needs a terrier Jul 06 08:59:29 XorA: borzoi/spitz are better in overclocking Jul 06 09:00:50 hrw|work: why? different cpu in terrier? Jul 06 09:04:12 isn't a terrier just borzoi +2g diskspace? Jul 06 09:04:25 morning all Jul 06 09:04:53 hey RP Jul 06 09:05:38 hey RP Jul 06 09:07:08 hey Richard Jul 06 09:07:33 yay, cherrypicker test passes Jul 06 09:09:12 * njs propagates it to mainline Jul 06 09:10:16 mickeyl: http://bugs.openembedded.org/show_bug.cgi?id=745 Jul 06 09:10:22 njs: that's good news Jul 06 09:10:32 njs: I suspect that will make some OE people very happy :) Jul 06 09:11:15 RP: that is the hope Jul 06 09:21:58 ~praise njs Jul 06 09:22:01 All hail njs! Jul 06 09:22:11 ~botsnack Jul 06 09:22:11 aw, gee, kergoth Jul 06 09:22:44 kergoth: I thought you used svk? :-) Jul 06 10:04:54 XorA: Meanie on oesf wrote that sharp used newer set of chips which had a problem with OC Jul 06 10:05:03 njs: cool Jul 06 10:05:21 hrw|work: not that it needs overclocking :-) Jul 06 10:06:21 they still need to be 800 MHz and have a vga port Jul 06 10:08:54 reports from pdaXrom are without overclock they can play DivX where X<640 Y<480 Jul 06 10:10:40 anything that you can fit in the sram of the pxa270 would display lightning fast Jul 06 10:11:13 03mickeyl 07org.oe.dev * r02ca6b9a... 10/packages/binutils/ (4 files): binutils 2.15.94.0.1 add matching binutils-cross-sdk, remove other 2.15 ones Jul 06 10:11:15 03xora 07org.oe.dev * r5beb3d8e... 10/packages/gimp/gimp_2.3.10.bb: Jul 06 10:11:15 gimp_2.3.10.bb : new version, DEFAULT_PREFERENCE=-1 at moment as it starts to Jul 06 10:11:15 load but at end of splash screen quits with message gimp-2.3: terminated: Aborted. This is not OOM. Jul 06 10:39:16 hello everyone Jul 06 10:42:22 NOTE: package gettext-0.14.1: completed Jul 06 10:42:24 mickeyl, RP, do13: someone of you tried USB speakers with Zaurus? Jul 06 10:42:41 hrw|work: not me Jul 06 10:43:00 yay: uclibc svn gettext Jul 06 10:43:02 friend bought such beasts so I will check them in august Jul 06 10:43:22 hrw|work: I read that as breasts on first pass :-) Jul 06 10:45:04 XorA: we have proverb for such thinking - something with being hungry ;D Jul 06 10:45:22 * XorA chuckles Jul 06 10:45:30 hrw|work: never checked Jul 06 10:47:34 RP: locomolcd needs reworking. It reenables the display on each backlight change, has wrong contrast settings, .... Jul 06 10:48:26 do13_: Can the backlight and power controls be seperated? Jul 06 10:48:36 do13_: It does seem to be a bit of a mess :-/ Jul 06 10:48:49 do13_: (I just cleaned up its use of the backlight class) Jul 06 10:50:23 RP: yes, It seems the power controls can be seperated Jul 06 10:50:42 do13_: There is an lcd power class as well as a backlight class Jul 06 10:50:51 RP: the comadj value should be passed via platform data Jul 06 10:51:01 do13_: agreed Jul 06 10:59:40 03kristoffer 07org.oe.dev * redd37393... 10/packages/linux/ (3 files in 2 dirs): Jul 06 10:59:40 linux/linux-jlime-sh3-2.6.17/LinuxSH-2.6.17.patch: Updated patch Jul 06 10:59:40 * Updated patch from CVS to fix build problems. Jul 06 10:59:40 linux/linux-jlime-sh3-2.6.17/defconfig_jlime: Updated .config Jul 06 10:59:40 * Updated .config to reflect kernel changes. Jul 06 10:59:41 linux/linux-jlime-sh3_2.6.17.bb: Changes to reflect updates Jul 06 11:03:09 why do i always type task-boostrap Jul 06 11:03:33 * ade|desk slaps self Jul 06 11:28:22 is bootstrap-image supposed to give you a final rootfs? Jul 06 11:29:08 tkp: yes Jul 06 11:29:16 tkp: check tmp/deploy/images Jul 06 11:29:23 hmm. odd. It said it completed Jul 06 11:29:45 there is only tmp/depoly/ipk Jul 06 11:29:50 no images directory Jul 06 11:30:21 do you have IMAGE_FSTYPES defined? Jul 06 11:31:08 yes, set to ext2 Jul 06 11:31:20 IMAGE_FSTYPES = "ext2" Jul 06 11:31:45 perhaps todo with: BBMASK="(task-bootstrap-unionroot)" ? Jul 06 11:35:21 that should be harmless Jul 06 11:38:12 what command creates the packages listing for ipkg Jul 06 11:38:53 bitbake package-index Jul 06 11:39:01 cheers Jul 06 11:39:03 or ipkg-make-index -p Packages -m . Jul 06 11:39:18 '.' being the dir where the ipkgs are Jul 06 11:43:06 hmm can't find task-bootstrap when building bootstrap-image, yet it is there in deploy/ipk/ Jul 06 11:56:00 opinions? Jul 06 11:56:02 RP, mickeyl, XorA, koen: http://pastebin.ca/80437 Jul 06 11:59:18 hrw|work: http://pastebin.ca/80442 Jul 06 11:59:20 03kristoffer 07org.oe.dev * rb21ed980... 10/packages/linux/ (9 files in 3 dirs): Jul 06 11:59:20 packages/linux/linux-jlime-sh3-2.6.16.20: Dropping packages Jul 06 11:59:20 * Dropping jlime-sh3-2.6.16.20 kernel due to 2.6.17 being Jul 06 11:59:20 availabel. Dropping all patches and .bb file Jul 06 11:59:20 packages/linux/linux-jlime-sh3-2.6.17/unexpected-int-fix: Add patch Jul 06 11:59:21 * Adding patch to silence unexpected interrupts. Jul 06 12:00:12 look ok Jul 06 12:02:50 koen: I was thinking about adding OE notice Jul 06 12:03:01 ah yes Jul 06 12:03:04 and an OE notice Jul 06 12:05:38 03kristoffer 07org.oe.dev * r9fcf7591... 10/packages/linux/linux-jlime-sh3-2.6.17/unexpected-int-fix.patch: Jul 06 12:05:38 linux/linux-jlime-sh3-2.6.17/unexpected-int-fix.patch : ported to 2.6.17 Jul 06 12:05:38 * Updated patch to reflect 2.6.17 Jul 06 12:07:01 koen: Release work would not be possible without OpenEmbedded buildsystem which took most of work from our shoulders. Jul 06 12:10:12 'without the OpenEmbedded buildsystem' Jul 06 12:10:39 and 'most of the work' Jul 06 12:10:54 * koen suspects polish doesn't use 'the' and 'a' a lot, like russian Jul 06 12:11:18 koen: the/a/an does not exist in our language group Jul 06 12:12:08 like I thought Jul 06 12:15:42 so... anyone with amd64 here? Jul 06 12:16:00 ixoth: us Jul 06 12:16:09 yep. Jul 06 12:16:14 am I the only one who has qemu segfaults during glibc binary locale generation? Jul 06 12:16:15 I have access to such one Jul 06 12:16:18 well, em64t. Jul 06 12:16:39 NAbyss: em64t from rev E (iirc) == amd64 Jul 06 12:17:37 hrw|work: ah, I thought the instruction set was slightly different? Jul 06 12:18:09 NAbyss: first em64t lacked some amd64 instructions. then intel added some and finally in one rev they added last one Jul 06 12:20:15 so, nobody on amd64 uses binary locale generation? Jul 06 12:22:07 ixoth: if you will wait 2h then I can look Jul 06 12:22:37 yes, please, if you can... Jul 06 12:23:09 ~oe.db Jul 06 12:23:16 methinks oe.db is http://www.openembedded.org/snapshots/OE.db.bz2 Jul 06 12:23:24 ixoth: how do enable/disable binary local generation? Jul 06 12:23:32 my goal is to make an ugly hack to make it work under amd64, but there's no use in doing it if I am the only one affected :)) Jul 06 12:23:46 cbrake: ENABLE_BINARY_GENERATION=1 Jul 06 12:23:49 iirc Jul 06 12:24:05 cbrake: ENABLE_BINARY_LOCALE_GENERATION = "1" Jul 06 12:24:17 ENABLE_BINARY_LOCALE_GENERATION = "1" to be exact Jul 06 12:24:26 oh, hrw|work is faster than me Jul 06 12:24:36 * cbrake checks ... Jul 06 12:25:01 cbrake: if you check then I will do not - more speed for ya Jul 06 12:28:21 hrw|work: looks like binary local generation is off by default -- ok, I'll turn it on and kick off a build ... Jul 06 12:28:55 cbrake: its off in .dev Jul 06 12:29:28 hrw|work: ok Jul 06 12:31:38 ixoth: ok, should have some info in an hour or so ... Jul 06 12:32:13 cbrake: great... btw, you use system qemu or qemu-native from OE? Jul 06 12:32:24 and what's your distro, just to know Jul 06 12:32:46 ixoth: starting using OE qemu as I don't have it installed natively Jul 06 12:32:58 ixoth: gentoo amd64. what are you using? Jul 06 12:33:17 the same Jul 06 12:33:24 ixoth: :-) Jul 06 12:33:49 in fact, i tried using qemu from portage with same results (segfaults) Jul 06 12:35:32 03florian 07org.oe.dev * rc858dcb5... 10/packages/glib-2.0/glib-2.0_2.12.0.bb: glib: add 2.12.0 Jul 06 12:35:36 03florian 07org.oe.dev * r7fea829f... 10/packages/pango/ (pango-1.13.2/opentype-makefile.patch pango_1.13.2.bb): pango: add 1.13.2 Jul 06 12:35:40 03florian 07org.oe.dev * r1331b7fc... 10/packages/cairo/cairo_1.2.0.bb: cairo: add 1.2.0 Jul 06 12:35:44 03florian 07org.oe.dev * r3c1e67c8... 10/packages/gtk+/ (17 files in 2 dirs): gtk+: Add version 2.10.0. Jul 06 12:35:48 03florian 07org.oe.dev * r6de86d14... 10/packages/gtk+/gtk+_2.10.0.bb: gtk+: Lower preference of 2.6.10 because of missing patches and missing test. Jul 06 12:41:10 if a header files is placed in staging by the libc that includes a define that is not valid what is the best way to remove it? do the staging bit and then patch the header afterwards, how do you this do_stage_append ?? Jul 06 12:42:22 patch it before staging Jul 06 12:44:37 well its needed for uclibc-initial to build uclibc, but after uclibc is built it very much needs to not be there afterwards Jul 06 12:45:08 grr meeting , bbiab Jul 06 12:47:28 RP: finally got time to test audio, all the issues I had are fixed apart from speaker which I cant test in office :-) Jul 06 12:52:20 XorA: you'll probably like the next commit Jul 06 12:52:25 http://ewi546.ewi.utwente.nl/tmp/hrw/3.5.4.1-test/final/fixed2/ - what do you think? Jul 06 12:53:09 hrw|work: I'd change 'archive to install' to 'install kit' or something Jul 06 12:53:19 hrw|work: IMO too wide Jul 06 12:53:33 otherwise you end up spending weeks reading forum posts like 'it doesn't flash the tarball' Jul 06 12:54:03 ixoth: thats true - have to reorg it Jul 06 12:55:03 hrw|work: and s/procedd/proceed/ Jul 06 12:55:30 s/support/supports/ Jul 06 12:55:36 03koen 07org.oe.dev * r3b10fe49... 10/packages/gtk+/gtk+_2.10.0.bb: gtk+ 2.10: depend on cups to enable the printing plugin Jul 06 12:55:46 koen: heh heh Jul 06 12:55:53 koen: one day I shall buy ink for my printer Jul 06 12:56:11 XorA: and one day I shall actually build gtk 2.10 to see if it works Jul 06 12:59:28 now it looks too packed.. Jul 06 13:01:15 what do you think of current ver? Jul 06 13:03:36 looks pretty good Jul 06 13:06:20 hmm.. link to tarballs with rootfs would be usefull probably too.. Jul 06 13:12:37 what are the difference between BOOTSTRAP_EXTRA_DEPENDS and BOOTSTRAP_EXTRA_RDEPENDS? Jul 06 13:12:41 do I understand correctly that BOOTSTRAP_EXTRA_DEPENDS are built before image Jul 06 13:12:47 and BOOTSTRAP_EXTRA_RDEPENDS is the stuff that gets built into image? Jul 06 13:12:53 or I am wrong? Jul 06 13:18:31 kergoth: ping Jul 06 13:24:08 http://openzaurus.org/wordpress/2006/07/06/openzaurus-3541-release/ Jul 06 13:24:55 We Made It ;) Jul 06 13:25:06 hrw|work: congrats! Jul 06 13:25:32 ixoth: DEPENDS are build time dependencies and RDEPENDS are runtime dependencies Jul 06 13:25:50 ixoth: (slightly wrong) RDEPENDS end up in the rootfs Jul 06 13:26:29 koen: so if I list package in RDEPENDS but not in DEPENDS, will it work? Jul 06 13:26:43 ~hail hrw|work Jul 06 13:26:44 * ibot bows down to hrw|work and chants, "I'M NOT WORTHY!!" Jul 06 13:26:55 ixoth: with a recent OE and bitbake, yes Jul 06 13:27:58 * XorA needs a faster rm Jul 06 13:28:09 03kristoffer 07org.oe.dev * r4c2c6d22... 10/conf/distro/ (jlime-donkey.conf preferred-opie-cvs-versions.inc): Jul 06 13:28:09 conf/distro/jlime-donkey.conf : changed opie SRCDATE = 20060627 Jul 06 13:28:09 conf/distro/preferred-opie-cvs-versions.inc : set =? instead of = Jul 06 13:28:13 03kristoffer 07org.oe.dev * r2abc517b... 10/packages/linux/ (2 files in 2 dirs): Jul 06 13:28:13 linux/linux-jlime-sh3-2.6.17/defconfig_jlime : Changes Jul 06 13:28:13 * Added software stack/encryption/mac for wifi Jul 06 13:28:15 linux/linux-jlime-sh3_2.6.17.bb: Added patch to file Jul 06 13:28:17 XorA: keep build on separate partition Jul 06 13:28:17 * Added unexpected-ide-fix.patch to .bb file Jul 06 13:28:43 hrw|work: if Id know I was going to be doing quite so much OE work when I partitioned this drive :-) Jul 06 13:29:10 koen: and if I list package in DEPENDS but not in RDEPENDS - it won't be in rootfs? Jul 06 13:29:20 XorA: and which FS do you use? Jul 06 13:29:43 hrw|work: "the Netherlands" Jul 06 13:29:52 ixoth: correct Jul 06 13:29:57 koen: ok Jul 06 13:30:13 koen: thanks! Jul 06 13:30:15 koen: 'in the Germany' too? Jul 06 13:30:35 hrw|work: no, that's fine Jul 06 13:30:44 .nl just has a weird name Jul 06 13:33:40 hrw|work:Thanks for all the work :) Jul 06 13:39:35 re Jul 06 13:41:01 v8jlene: ping? Jul 06 13:43:30 ixoth: ext3 Jul 06 13:44:21 XorA: i prefer ReiserFS, it deletes files faster :) Jul 06 13:45:01 ixoth: I treid it on amd64, it just made garbage instead of a filesystem :-) Jul 06 13:45:29 hmmm :) perhaps we have different amd64's :) Jul 06 13:46:09 ixoth: this was 2 years ago, my x86 FS was fine, Reiser had many amd64 problems Jul 06 13:46:35 NS2531972056 will be on linuxdevices Jul 06 13:47:32 XorA: fortunately, I bought myself amd64 only a few moths ago :) Jul 06 13:48:24 ixoth: I need to upgrade mine Jul 06 13:50:06 XorA: i've got myself Shuttle ST20G5 barebone, it is very silent Jul 06 13:50:42 My previous machine was a dual athlon which sounded like vacuum cleaner :) Jul 06 13:51:37 heh Jul 06 13:51:45 * koen is sitting behind a dual 2800+ Jul 06 13:51:59 ixoth: qemu patch failed :-( http://pastebin.ca/80505 Jul 06 13:52:37 tkp: What's up? Jul 06 13:52:42 seems like that patch needs a maxdate or updating Jul 06 13:52:52 hey v8jlene Jul 06 13:53:03 ixoth: Ive got a shuttle SN85G and it overheats :-( Jul 06 13:53:05 my FC5 build failed :( Jul 06 13:53:47 somewhere round here: http://rafb.net/paste/results/go5OX719.html Jul 06 13:53:54 tkp: Damn. You guys were also trying a build on something else? How'd that go? Jul 06 13:54:15 yeah, that was on our Debian 32bit vm Jul 06 13:54:18 it finished Jul 06 13:54:32 /usr/include/gnu/stubs.h:7:27: error: gnu/stubs-32.h: No such file or directory Jul 06 13:54:34 although for some reason I had no rootfs at the end of it Jul 06 13:54:40 looks like something is broken on the host Jul 06 13:54:42 (I had run bootstrap-image) Jul 06 13:55:03 cbrake: i had other problems with qemu-native-0.7.0 Jul 06 13:55:19 cbrake: it just segfaulted during binary locale generation Jul 06 13:55:33 is 0.8.1 in OE already? Jul 06 13:55:49 koen: hmm.. I'll look into it Jul 06 13:56:18 cbrake: can you set PREFERRED_PROVIDER_qemu-native="qemu-native" and continue? Jul 06 13:56:21 tkp: Missing glibc-devel? Jul 06 13:57:27 XorA: no, my barebone has some weird cooling system which keeps CPU about 50 degrees centigrade during load Jul 06 13:57:38 IMAGE_FSTYPES = "ext2" - is that right, or would it need to be ext2.gz Jul 06 13:57:48 v8jlene: it seems so! Jul 06 13:58:04 hi Jul 06 13:58:12 hrw|work: what's wring w/ openssh? Jul 06 13:58:22 tkp: FYI, my build with gcc 4.1.1, binutils 2.17.50.0.1 and glibc 2.4 worked fine as well.. Jul 06 13:58:38 gcc 4.1.1... good stuff Jul 06 13:59:20 ixoth: SN85G is notorieous for overheat, wish that had been all over the web before I bought it Jul 06 13:59:21 CoreDump|home: it does not fully replace dropbear Jul 06 13:59:36 CoreDump|home: I got images with openssh-sshd only instead of dropbear Jul 06 13:59:55 oops Jul 06 14:00:02 tkp: I actually meant to test 4.2-20060513 but accidently left the gcc selection totally commented out... 4.1.1 must be the default it builds... Jul 06 14:00:17 ixoth: ok, building ... Jul 06 14:00:57 v8jlene: ok, just installed glibc-devel.... gonna set it going again on the FC5 Jul 06 14:01:46 tkp: I use my own image .bb file - the image .bb are in pacakges/meta - so you can have a look at them to see what they are doing/meant to be doing. Jul 06 14:01:48 http://heanet.dl.sourceforge.net/sourceforge/dejavu/dejavu-ttf-2.7-1.tar.gz doesnt exist Jul 06 14:02:03 you think I need to clean out and start agin, or just carry on from where it failed? Jul 06 14:02:28 tkp: I'd just clean whatever package it failed on and then start it again. Jul 06 14:02:35 right Jul 06 14:05:19 koen, who maintains oesources again :) Jul 06 14:05:21 keep forgetting Jul 06 14:05:27 treke Jul 06 14:05:53 oesources is complimentary, not mandatory Jul 06 14:06:01 http://heanet.dl.sourceforge.net/sourceforge/dejavu/dejavu-ttf-2.7-1.tar.gz doesnt exist Jul 06 14:06:10 so fix the .bb Jul 06 14:06:11 Kristoffer: hmm.. I pushed -1? Jul 06 14:06:39 koen, doh Jul 06 14:06:43 hrw|work, ? Jul 06 14:06:43 commited Jul 06 14:06:51 Kristoffer: its my recipe Jul 06 14:06:59 ~blame me Jul 06 14:06:59 aha Jul 06 14:07:01 * ibot blames hrw|work (and Canada) for all the evil in the world Jul 06 14:07:30 I take it no one has any issue with my mutiple device tables thing that I sent to mailing list and so can push it? Jul 06 14:07:41 * REdOG slaps ibot Jul 06 14:07:55 v8jlene, havent looked at it, but hasnt affected my builds Jul 06 14:08:41 v8jlene: if no-one responded, just push it Jul 06 14:08:48 v8jlene: makes life much easier Jul 06 14:09:00 Kristoffer: pushed fix Jul 06 14:09:05 hrw|work, thx man Jul 06 14:09:25 what's up with qemu-qop ? Jul 06 14:09:44 03hrw 07org.oe.dev * r83caa66d... 10/packages/ttf-fonts/ttf-dejavu_2.7.bb: ttf-dejavu: fixed SRC_URI Jul 06 14:09:58 tkp: This is what bootstap-image looks like for me: http://pastebin.ca/80519 Jul 06 14:10:12 ade|desk: no idea -- its what OE decided to build. Jul 06 14:10:48 v8jlene: that's prety much exactly what it looked like for me, only no tmp/deploy/images dir at the end of it Jul 06 14:12:20 it's running again, so we'll see what happens this time Jul 06 14:12:25 ixoth: more patching failures: http://pastebin.ca/80521 Jul 06 14:12:43 of, and FC5 seems to be going ok now I installed glibc-devel Jul 06 14:13:36 tkp: Run bootstrap-image with -D on the command line and see if gives any clue Jul 06 14:13:48 v8jlene, any reason to change to glibc 2.4? Jul 06 14:13:49 cbrake: can you try to use qemu-0.7.0 (that's what is used in org.oe.oz354x) Jul 06 14:13:54 v8jlene, and did it build well for you? Jul 06 14:14:05 ixoth: ok Jul 06 14:14:14 v8jlene: will do if it fails to give me a fs image again this time Jul 06 14:14:37 Kristoffer: I've got it building for x86, not for sh yet - I'll get to that. Jul 06 14:15:57 Kristoffer: The main issue is that if you try and checkout any of the cvs glibc release they no longer work... if you have an old cvs checkout they work fine. Jul 06 14:18:43 cbrake: for qemu-qop just remove the trunk_no_docs patch , it is nolonger needed as build_docs = "no" id default Jul 06 14:18:49 s/id/is Jul 06 14:19:13 ixoth: ok, building Jul 06 14:19:21 morning Jul 06 14:19:26 ade|desk: ok, I'll test that after I run a build w/ 0.7.0 Jul 06 14:21:13 03lenehan 07org.oe.dev * r4059b35d... 10/classes/image_ipk.bbclass: (log message trimmed) Jul 06 14:21:13 classes/image_ipk: Add support for multiple device tables. The original code Jul 06 14:21:13 support a single value as IMAGE_DEVICE_TABLE which needed to be an absolute Jul 06 14:21:13 value. If that is definied it'll be used as is and this change will have no Jul 06 14:21:13 effect. If is is not definied it'll check for IMAGE_DEVICE_TABLES which can Jul 06 14:21:13 be a list of device images tables which are searched for relative to the Jul 06 14:21:17 BBPATH and it'll apply each one in turn. If neither of these are definied Jul 06 14:21:23 oh crap , if you have texi2html AND pod2man then it defaults to building the docs Jul 06 14:21:53 man alive the qemu build system is strange Jul 06 14:22:31 why oh why did i ever push qemu-qop into OE Jul 06 14:24:59 03lenehan 07org.oe.dev * r86bbc3c1... 10/packages/linux/ (6 files in 4 dirs): Jul 06 14:24:59 linux/linux-titan-sh4: Upgrade the titan kernel from 2.6.16 to 2.6.17 now Jul 06 14:24:59 that the sh tree has been upgraded to 2.6.17. Jul 06 14:27:00 how do I do an update with monotone? Jul 06 14:27:16 or do I need to do a complete co each time? Jul 06 14:27:42 REdOG: mtn pull ; mtn update Jul 06 14:28:18 REdOG: http://venge.net/monotone/monotone.html Jul 06 14:28:20 monotone: misuse: working copy directory required but not found Jul 06 14:29:18 03lenehan 07org.oe.dev * r1580665f... 10/packages/glibc/glibc_2.4.bb: Jul 06 14:29:18 glibc 2.4: Remove the workaround from CVS changes. This isn't a CVS version Jul 06 14:29:18 so the workaround shouldn't be needed. Also this work around breaks glibc Jul 06 14:29:18 2.4 for all targets that are no in ports (x86, sh, sparc, powerpc). Jul 06 14:29:59 hey pH5 Jul 06 14:30:21 hi koen Jul 06 14:30:22 pH5: I tried booting from cf, but I only succeed in loading a kernel, not mounting a rootfs :( Jul 06 14:30:32 koen: mtn diff is the diff between working and local db, what is the cmd for changes in local db vs remote db ? Jul 06 14:31:16 koen: microdrive or flash? Jul 06 14:31:52 ade|desk: there isn't any Jul 06 14:32:03 pH5: 1gb cf (flash) card Jul 06 14:32:17 koen: bugger Jul 06 14:32:22 ade|desk: monotone doesn't think in 'local vs remote' Jul 06 14:32:39 * REdOG is not liking monotone so far Jul 06 14:32:42 hm. the microdrive needed some seconds to spin up, but I guess that won't be the case with flash Jul 06 14:33:24 REdOG: that's your problem Jul 06 14:33:32 koen: thanks for the reminder Jul 06 14:33:41 REdOG: direct your complaints to the monotone devs or /dev/null Jul 06 14:33:45 we aren't interested Jul 06 14:34:06 i think that i have commited stuff in my local that shouldn't go upstream a while ago, but i haven't pushed, nor am i going to,... but there is stuff in there that should go , i just can't remember what the hell the stuff is/was Jul 06 14:34:09 s/complaints/opinions Jul 06 14:34:27 is there a command line ogg player in OE thats not mplayer? Jul 06 14:35:01 does madplay play ogg these days ? Jul 06 14:35:17 ade|desk: one in OE doesnt seem to Jul 06 14:35:28 ade|desk: I have two local copies checked out (into different directories) from the local database. I can update the "master" and the diff with my "working" copy. Jul 06 14:36:33 i shall have to have two dbs and two working copies and see what happens Jul 06 14:36:53 one for the weekend me thinks Jul 06 14:37:11 XorA: oggplay? Jul 06 14:37:20 ade|desk: I can't think of why you'd want two db's though? Jul 06 14:37:23 XorA: iirc that was part of the libogg package Jul 06 14:37:49 damn it... : http://pastebin.ca/80549 Jul 06 14:38:10 (on the 32 bit vm) Jul 06 14:39:07 03lenehan 07org.oe.dev * ra574e430... 10/packages/ssmtp/ (ssmtp-2.61/dont-strip.patch ssmtp_2.61.bb): Jul 06 14:39:07 ssmtp 2.61: Dont use -s to install so it doesn't attemp to use the host Jul 06 14:39:07 strip during do_install. OE will take care of running the target strip Jul 06 14:39:07 during packaging. Jul 06 14:40:03 ixoth: look familiar: http://pastebin.ca/80552 Jul 06 14:40:20 ixoth: trying 0.8.0 next ... Jul 06 14:40:36 libjava... anything todo with LANGAUGES = "c,c++" perhaps? Jul 06 14:41:02 cbrake: yes, that's it... Jul 06 14:41:15 cbrake: exactly like this Jul 06 14:41:19 shouldn't that tell it to not use libgcj? Jul 06 14:41:39 cbrake: I'm outta to write a bugreport Jul 06 14:41:51 03lenehan 07org.oe.dev * rc6de9197... 10/files/ (device_table_add-sci.txt device_table_add-sm.txt): Jul 06 14:41:51 files/device_table_add: Add two device tables fragments. One adds the Jul 06 14:41:51 SCI/SCIF serial interfaces (as found on SH* platforms) and ones adds the SM Jul 06 14:41:51 flash driver (TI flash block mapping driver) devices. This can be used with Jul 06 14:41:51 the new IMAGE_DEVICE_TABLES to add them to the standard minimal table. Jul 06 14:42:33 v8jlene: db for clean OE and one that is my polluted db : i can then commit all my crap in the working copy into the dirty db : fresh checkout of the dirty OE as the working copy of the clean db, then mtn diff it to see what i want to keep Jul 06 14:44:01 tkp: Yeah, that should disable it. Jul 06 14:44:21 hi Jul 06 14:44:33 hrw|work: What's wrong with Opie on C7x0? Jul 06 14:44:35 v8jlene: hmm, odd that it's trying to use it Jul 06 14:45:06 hey mickeyl Jul 06 14:45:12 mickeyl: it probably was build error - I replaced 3541 package with rc4 one and it is working now Jul 06 14:45:27 hrw|work: *phew* ok :) Jul 06 14:46:08 mickeyl: hrw|work: like opie matters ;) Jul 06 14:46:09 tkp: I'll let mine run through to the end of bootstrap-image and see what happens... Jul 06 14:46:19 ade|desk: you would be surprised how many users it has Jul 06 14:46:29 although it doesn't get updated Jul 06 14:46:30 tkp: Although I'm all done for tongiht... much too late again ;) Jul 06 14:46:30 (more or less) Jul 06 14:46:33 mickeyl: indeed i would Jul 06 14:46:50 v8jlene: fair play Jul 06 14:47:02 * mickeyl checks his accounts Jul 06 14:47:19 ~change 13 KRO to EUR Jul 06 14:47:26 13.00 Norway Krone (NOK) makes 1.63562 Euro (EUR) (from http://www.xe.com/) Jul 06 14:47:34 ~change 19 KRO to EUR Jul 06 14:47:37 v8jlene: could you paste your dist and local confs? Jul 06 14:47:42 19.00 Norway Krone (NOK) makes 2.39052 Euro (EUR) (from http://www.xe.com/) Jul 06 14:48:17 mickeyl: did you see the monotone pluck addition? Jul 06 14:48:28 koen: no, what's that? Jul 06 14:48:35 mickeyl: cherry picker :) Jul 06 14:48:55 cbrake: http://bugs.openembedded.org/show_bug.cgi?id=1144 Jul 06 14:49:00 koen: woot!? Jul 06 14:49:08 mickeyl: http://lists.gnu.org/archive/html/monotone-devel/2006-07/msg00033.html Jul 06 14:49:29 * mickeyl drops a sigh of relief Jul 06 14:49:38 that brings us nearer to a OE-Core Jul 06 14:49:57 tkp: http://twibble.org/dist/oe/ Jul 06 14:49:57 OE-core? Jul 06 14:50:12 v8jlene: thanks Jul 06 14:50:17 \away Jul 06 14:50:21 my codename for a magic place where only reviewed changes can go in Jul 06 14:50:37 thinking Q/A Jul 06 14:50:57 * koen practices the secret handshake Jul 06 14:51:03 hehe Jul 06 14:52:52 koen: the one with 16 shakes? Jul 06 14:53:11 yes Jul 06 14:54:53 as long as left nipples are not involved :-) Jul 06 14:54:58 beyond 7 and i can't remember .. thats me out of QA :) Jul 06 14:55:13 XorA: left bigger thean right ? Jul 06 14:55:33 cbrake: if you can, you can also report your info about org.oe.dev qemu failures to the same bug Jul 06 14:55:38 XorA: or it is which 'left' Jul 06 14:55:54 ade|desk: old masons jokes :-) Jul 06 14:55:56 cbrake: and also try the proposed patch :) Jul 06 14:56:17 ah now that handshake i do know Jul 06 14:57:06 who had problems with x86 stuff? Jul 06 14:57:08 i can build task-bootstrap without problems here Jul 06 14:57:09 * XorA runs and hides from ade|desk Jul 06 14:57:10 DISTRO=generic MACHINE=epia Jul 06 14:57:44 mickeyl: we did Jul 06 14:58:41 mickeyl: you built straight from a fresh checkout? no modifications? Jul 06 14:59:07 yes Jul 06 14:59:20 i have a site.conf file here, but that doesn't contain anything substantial Jul 06 14:59:28 mickeyl: if a prefered task-bootstrap is set, then it builds, but then doing bootstrap-image .. foobar Jul 06 14:59:30 let me get you the build log Jul 06 14:59:38 mickeyl: what's your host machine? Jul 06 14:59:40 i didn't build bootstrap-image, just task-bootstrap Jul 06 14:59:48 tkp: dual xeon Jul 06 14:59:50 32bit Jul 06 14:59:58 distro? Jul 06 15:00:03 Mandriva 2007.0 cooker Jul 06 15:00:13 interesting Jul 06 15:00:24 i had one error in binutils Jul 06 15:00:29 which i fixed by adding a patch Jul 06 15:00:31 could you try bootstrap-image Jul 06 15:00:56 and, when was your repo checkout from? Jul 06 15:01:00 mickeyl: I would not really recommend using cherrypicking to do a review-before-integrate workflow, because you'd be subverting monotone's own history keeping mechanisms... Jul 06 15:01:25 mickeyl: OTOH, if you can't convince people to submit changes as single units on branches, maybe it's what you'd have to do :-) Jul 06 15:02:11 tjp: from yesterday. I'm updating it atm. will attempt to build bootstrap-image then Jul 06 15:03:17 njs: Actually, I wasn't interested in cherrypicking that much, but it looks like some people are really needing it. When we have it anyway, I could imagine a .reviewed branch where only reviewed changes come in via cherrypicking. Jul 06 15:03:37 isn't that what it's for? picking individual changes out of a stack?= Jul 06 15:03:53 njs: we already have very granular commits, but monotone isn't helpfull if you want to comment 14 out of 20 changes to another branch Jul 06 15:03:58 XorA: | make[5]: *** No rule to make target `../../dist/lib/libnkjar_s.a', needed by `libnecko.a'. Stop. Jul 06 15:04:04 XorA: thats minimo Jul 06 15:04:18 hrw|work: crap, well thats further than I get anyway Jul 06 15:04:28 XorA: stopped on xpcom? Jul 06 15:04:35 njs: propagating and disapproving 6 changes seems a bit backwards to me Jul 06 15:04:38 hrw|work: no, nsinstall Jul 06 15:05:01 hrw|work: it tries to use -m32 with arm compiler Jul 06 15:05:28 hrw|work: cvsdate today? Jul 06 15:06:08 XorA: no - old one Jul 06 15:06:32 hrw|work: ah, I know for x86 todays cvsdate builds Jul 06 15:07:21 XorA: started today build Jul 06 15:07:22 hrw|work: and runs Jul 06 15:09:42 goodbye everyone Jul 06 15:10:01 cbrake: test my patch from bugzilla... Jul 06 15:11:20 hrw|work: The message looks good to me Jul 06 15:11:35 XorA: You'll note Liam has sent a new patch :) Jul 06 15:12:39 hrw|work: can you respond to http://www.oesf.org/forums/index.php?showtopic=20306&hl= ? Jul 06 15:12:52 a guy asking whether WPA works and what hardware he needs Jul 06 15:13:14 RP: ASoC would make a perfect infrastructure for a phone audio subsystem, wouldn't it? Jul 06 15:13:36 mickeyl: Correct. I suspect people are using it for that already ;-) Jul 06 15:13:47 *nod* :D Jul 06 15:14:38 mickeyl: Its nearly working perfectly for us now on the Zs - just one bug to find and Liam/XorA are working on that :) Jul 06 15:15:17 RP: excellent. I will try to convince the openezx guys to base the (pretty complicated) audio subsystem on that Jul 06 15:15:36 mickeyl: Do you know which codec it uses? Jul 06 15:15:40 mickeyl: sent Jul 06 15:15:48 hrw|work: thank you Jul 06 15:15:55 np Jul 06 15:15:58 RP: let me check Jul 06 15:16:14 mickeyl: ASoC can do pretty complicated already FWIW :) Jul 06 15:19:27 hmm Jul 06 15:19:28 :/ Jul 06 15:19:42 looks like everything on the A780 is handled by the PCAP2 custom MOTO asic Jul 06 15:19:49 s/everything/everything audio/ Jul 06 15:19:54 mickeyl: did you hear back from harald about how he's enjoying OZ? Jul 06 15:19:57 RP: which bug are we still hunting? Jul 06 15:20:02 RP: I lose track Jul 06 15:20:07 XorA: The external speaker on/off Jul 06 15:20:16 koen: no, not yet. he is in shangai atm. Jul 06 15:20:18 XorA: Input is working Jul 06 15:20:19 RP: ok, thats the bit I cant test in middle of office Jul 06 15:20:28 XorA: :) Jul 06 15:20:33 RP: I noticed the capture mixer had changed Jul 06 15:20:53 mickeyl: Is the PXA used for audio at all? Jul 06 15:21:05 RP: how can i check that easilyß Jul 06 15:21:06 ? Jul 06 15:21:22 mickeyl: 2.4 source? Jul 06 15:21:30 heh Jul 06 15:21:35 no surprise Jul 06 15:21:38 but it's huge :D Jul 06 15:21:56 mickeyl: better quality then crapix? Jul 06 15:21:57 * mickeyl wades through mvista 2.4 Jul 06 15:22:05 hrw|work: no. the same! even worse in some areas :/ Jul 06 15:22:36 mickeyl: It can't be worse ;-) Jul 06 15:22:41 mvista and good code quality are surely two things that never meet Jul 06 15:23:02 mickeyl: If you can point me at the code, I'll take a look Jul 06 15:23:03 ade|desk: I heard the support contract includes a shrink Jul 06 15:23:33 koen: heh Jul 06 15:23:40 http://vanille.de/temp/linux-2.4.20-mvl/ Jul 06 15:23:58 grep for EZX Jul 06 15:24:04 hi mickeyl Jul 06 15:24:09 I hope that traffic_eth0_86400.png on ewi will bump soon Jul 06 15:24:09 yi chouimat, how are things? Jul 06 15:24:14 hrw|work: definitly :) Jul 06 15:24:19 mickeyl: Have you an archive of that somwhere? Jul 06 15:24:21 mickeyl: wget lovers Jul 06 15:24:27 RP: sure, let me tar it up. 1 sec Jul 06 15:24:45 saves me spamming your server with wget :) Jul 06 15:24:55 try http://ewi546.ewi.utwente.nl/rrd/graph/traffic_eth0_7200.png Jul 06 15:24:59 mickeyl: as usual ... lot of paperwork (closing old co and starting new one), and trying to use openwrt buildroot-ng to build a firmware for my wrtsl54gs Jul 06 15:26:23 RP: http://vanille.de/temp/e680-linux.tar.gz Jul 06 15:26:43 RP: def-configs/ezx-780 and def-configs/ezx-680 are the two interesting ones Jul 06 15:27:26 mickeyl: old co is currently registered in quebec ... and I leave in ontario so doing 3 taxes form for it is too much ... and it's a pita because I have to comply to the Canadian rules, the one in quebec and the one in ontario, mostly the same but with some subtle changes ... and it's easy to close and reopen than transfering (a lot cheaper too) Jul 06 15:27:31 tkp: bitbaking bootstrap-image atm. had to set PREFERRED_PROVIDER_task-bootstrap = "task-bootstrap" to stay away from unionroot Jul 06 15:27:36 mickeyl: ezx-audio is using pxa audio dma so ASoC will at least be of some use Jul 06 15:27:51 chouimat: you're running it on your own? Jul 06 15:27:56 RP: ah that's good to know Jul 06 15:28:17 mickeyl: right, I had set BBMASK="(task-bootstrap-unionroot)" to do the same Jul 06 15:28:27 mickeyl: yes ... but for the new one I will have 2 or 3 associates so it will be more fun ;) Jul 06 15:28:43 chouimat: yeah. must be tough to run one alone Jul 06 15:28:53 chouimat: I'm just starting my freelancer career Jul 06 15:28:58 mickeyl: via the nssp port on the pxa which again, ASoC supports Jul 06 15:29:04 RP: yay! Jul 06 15:29:32 mickeyl: do you have to start own business for it? Jul 06 15:30:03 mickeyl: it's fun ... but try to have some*one* with you, for support ... the only one I have is my cat and it's sometimes hard Jul 06 15:30:13 hrw|work: thankfully no. in germany freelancer don't need to have a company. that makes tax much easier. Jul 06 15:30:22 mickeyl: happy you Jul 06 15:30:23 chouimat: *nod* Jul 06 15:31:03 mickeyl: old co is me as freelancer ... new one is me as company ... easier to get $$$ for projects ;) Jul 06 15:31:07 RP: http://www.rpsys.net/openzaurus/patches/alsa/info.html. one typo 1st line: "procesors" missing an "s" and in the last part, it's a SL-6000, not an SL-C6000. Jul 06 15:31:34 chouimat: no kidding. if things run well for a year or so i'll think about a .inc as well. Jul 06 15:32:27 mickeyl: Basically, you need to add a codec driver and a machine support file to ASoC as far as I can see. This PCAP entiity on the SSP port appears to be how to communicate with the codec Jul 06 15:33:10 RP: ok, that sounds doable. there's a good number of examples done yet? Jul 06 15:33:12 mickeyl: is audio in PCAP or attached to PCAP? Jul 06 15:33:23 mickeyl: Fixed, thanks Jul 06 15:34:01 mickeyl: machine support files for mainstone, corgi, spitz and tosa, and about 9 codecs so far Jul 06 15:34:24 koen: on a780 it looks like it's IN the pcap. the e680 has a national semiconductor boomer as well Jul 06 15:34:42 anyone know how I can get monotone to tell me which files have changed in the repository since my checkout? Jul 06 15:35:13 tkp: monotone diff Jul 06 15:35:34 tkp: mt log -rLASTREVBEFOREUPDATE --diffs|diffstat Jul 06 15:35:56 hrw|work: cool, thanks Jul 06 15:36:52 RP: that sounds good. Jul 06 15:37:08 anyone know where I can get help with (what I think is) a tslib problem? Jul 06 15:37:17 mickeyl: fancy pasting your local.conf for comparrison? Jul 06 15:37:22 mickeyl: I ran into the task-bootstrap-unionroot issue recently -- will probably come up more often for new users. Jul 06 15:37:49 mickeyl: is this something that should be added to the distro confs, or add something to local.conf.sample ? Jul 06 15:38:10 bbl lunch time Jul 06 15:38:37 cbrake: i guess distros should define it Jul 06 15:38:44 cbrake: I think we should lower the default preference of unionbootstrap Jul 06 15:39:04 tkp: sure, hold on Jul 06 15:39:16 mickeyl: default preference is already "-1", but for some reason OE still likes it better. Jul 06 15:39:44 mickeyl: I looked at the bb code briefly, and I think default preference only applies to packages w/ the same name? Not sure about this. Jul 06 15:40:09 cbrake: could you please add a bug against bitbake for that? I think it's wrong behaviour. Jul 06 15:40:54 mickeyl: The more I look at the dependency and providers code, the more I feel it totally needs rewriting... Jul 06 15:41:00 tkp: http://pastebin.ca/80580 Jul 06 15:41:05 RP: out of interest are the CMDLINE options to change partition layout the same as for 2.4 kernel? Jul 06 15:41:11 mickeyl: thanks ;) Jul 06 15:41:17 RP: *nod* _ugly_ -- grew historically Jul 06 15:41:41 XorA: yes, I think so. We certainly support cmdline params Jul 06 15:41:47 RP: I had a very difficult time figuring out what the bb provider code was doing without a lot of instrumentation :-( Jul 06 15:42:04 good afternoon all Jul 06 15:42:20 mickeyl: I need to talk to zecke about chopping the main bitbake scipt into lumps... Jul 06 15:42:36 RP: that'd be good. it would help bitbake -i as well Jul 06 15:43:03 cbrake: We need to work on debugging. I think we need debugging domains in future Jul 06 15:43:38 RP: nod Jul 06 15:46:25 RP: how's the pysqlite bitbake effort coming along? Jul 06 15:46:30 RP, mickeyl: http://openzaurus.berlios.de/KernelInfo26/MachineStatus Jul 06 15:46:49 * ade|desk hugs bitbake -i, works most of the time Jul 06 15:47:14 cu all Jul 06 15:47:20 damn it... my hard drive filled up before bitbake finished :( Jul 06 15:47:51 hrw|gone: good site Jul 06 15:48:08 hrw|gone: heh Jul 06 15:48:17 hrw|gone: Pavel Maczek "hack" collie support Jul 06 15:48:19 very funny :D Jul 06 15:48:24 it's Pavel Machek btw. Jul 06 15:48:55 it can't be a hack, because it uses git! Jul 06 15:49:05 and linus uses git, so it can't be bad! Jul 06 15:49:07 * koen hides Jul 06 15:49:19 go back under your rock :D Jul 06 15:49:30 i'm sure git rocks anyway Jul 06 15:50:10 koen: I need to get back to it. I might try and push some of the initial restructuring before I continue with it - I'll talk to zecke Jul 06 15:50:26 RP: that'd be cool. Zecke has some more time when he's back Jul 06 15:50:31 ah well, we have a stable tree for a reason Jul 06 15:51:18 tkp: NOTE: package bootstrap-image-1.0-r0: task do_rootfs : started Jul 06 15:51:18 ... Jul 06 15:51:49 mickeyl: I might just push some of it as I doubt anyone is going to object to adding some structure to bitbake :) Jul 06 15:52:52 RP: that's wise :) Jul 06 15:52:58 tkp: bootstrap-image done Jul 06 15:53:10 tkp: let's see if I can chroot into it Jul 06 15:53:39 mickeyl: I got the "NOTE: package bootstrap-image-1.0-r0: task do_rootfs : started" when I ran it.. but there was no image in deploy Jul 06 15:54:21 tkp: ok, i can chroot into it fine Jul 06 15:54:40 mickeyl: good stuff. Jul 06 15:54:47 mine should be finished again soon Jul 06 15:54:52 I'll add the PREFERRED_PROVIDER into generic.conf for the time being Jul 06 15:54:55 * tkp touches wood Jul 06 15:55:37 koen: well gtk 2.10 doesnt work :-) Jul 06 15:56:02 gtk 2.8 was bad enough. poky is still using 2.6... Jul 06 15:56:23 I have a fix to make glib 2.10 work on arm btw - must push that... Jul 06 15:56:45 RP: that would be cool, nothing gtk runs here :-) Jul 06 15:57:01 oh glib Jul 06 15:57:42 RP: can't hurt to have gtk 2.10 in OE Jul 06 15:57:56 koen: No, but having it work would be better :) Jul 06 15:57:59 I don't want to end up like Opie, relying on tech from 2001 Jul 06 15:58:16 koen: too late with the cairo problems :-( Jul 06 15:58:29 cairo problems? Jul 06 15:58:34 other than being slow? Jul 06 15:58:54 The slowness is a showstopper :-( Jul 06 15:59:24 * XorA would like a reliable xserver as well Jul 06 15:59:49 a supported reliable xserver Jul 06 16:00:17 I was hoping with all the xorg work someone would find time for xsever-xorg 7.X Jul 06 16:01:36 XorA: didn't amd push xserver drivers and stuff? Jul 06 16:03:00 koen: it's still in the bugzilla. Jul 06 16:03:09 I see Jul 06 16:03:30 03mickeyl 07org.oe.dev * r6957560e... 10/conf/distro/ (4 files): generic.conf: prefer task-bootstrap over unionfs-bootstrap Jul 06 16:03:45 mickeyl: http://bugs.openembedded.org/show_bug.cgi?id=745 Jul 06 16:04:03 mickeyl: I couldn't reply to the hh.org bugtracker since my account has been disabled Jul 06 16:04:23 pH5: ah I must have missed, that, must check it out Jul 06 16:05:07 koen: righto. ok, cool. could you add that with DEFAULT_PREFERENCE = -1 to let us check the packaging problem Jul 06 16:05:17 sure Jul 06 16:05:31 thanks. Jul 06 16:14:42 03koen 07org.oe.dev * rae74551c... 10/packages/busybox/ (22 files in 7 dirs): busybox 1.2.0: add it, closes #745 Jul 06 16:21:10 mickeyl: there you go Jul 06 16:21:57 monotone pull && monotone up: monotone: network error: failed to connect: Connection refused Jul 06 16:22:46 time for me to escape this joint and let Liam debug ASOC Jul 06 16:23:59 * mrz80 is away: Metaphysically I'm still here; physically, I'm elsewhere Jul 06 16:24:44 tkp: try another server? Jul 06 16:25:32 koen: not used to this monotone... generally use svn Jul 06 16:25:55 how would I do that/what other server can I use? Jul 06 16:29:12 monotone pull Jul 06 16:30:12 tkp: it's all in http://www.openembedded.org/wiki/GettingStarted Jul 06 16:30:26 and in http://www.openembedded.org/wiki/MonotonePhraseBook Jul 06 16:31:08 ah I get it Jul 06 16:38:11 03florian 07org.oe.dev * r6daf673d... 10/packages/gpe-timesheet/gpe-timesheet_0.30.bb: gpe-timesheet: Add missing dependency for latest release. Jul 06 16:42:26 koen: ok, packaging breakage fixed. how about the uncommented patches? Jul 06 16:43:02 outcommented even Jul 06 16:43:03 mickeyl: need to be reviewed Jul 06 16:43:08 ok Jul 06 16:43:10 step-by-step Jul 06 16:43:12 * mickeyl commits Jul 06 16:43:23 we need to review all defconfigs Jul 06 16:43:27 options got renamed Jul 06 16:43:33 that's why udhcpc wasn't built Jul 06 16:46:55 heh Jul 06 16:46:56 Symbol: CONFIG_NITPICK [=n] Jul 06 16:49:06 oooh Jul 06 16:49:12 shared libbusybox Jul 06 16:49:18 that would make ipkg a lot smaller Jul 06 16:49:24 yay! Jul 06 16:49:33 ~seen pigi Jul 06 16:49:50 pigi was last seen on IRC in channel #asterisk, 37d 2h 32m 49s ago, saying: 'is there anyone that has successfully got asterisk working with eicon diva pci cards (passive cards ) ?'. Jul 06 16:49:58 cute Jul 06 16:49:59 Symbol: USING_CROSS_COMPILER [=n] Jul 06 16:50:16 oh, it isn't cute Jul 06 16:50:28 it's help for lazy crosstool using people ;) Jul 06 16:50:33 heh Jul 06 16:53:30 oooh Jul 06 16:53:31 03mickeyl 07org.oe.dev * r4fe87eeb... 10/packages/busybox/ (7 files in 7 dirs): Jul 06 16:53:31 busybox 1.2.0 fix packaging due to changed CONFIG_ options Jul 06 16:53:31 WARNING: Need to review ALL defconfigs before setting this version as default Jul 06 16:53:47 I'm so going to solve a lot of angstrom bugs with a new defconfig :) Jul 06 16:53:55 heh Jul 06 16:53:57 excellent Jul 06 16:56:11 ~lart dogs in japa's house that ate his lpi certificate Jul 06 16:56:11 * ibot sends a legion of lawyers after dogs in japa's house that ate his lpi certificate's head Jul 06 17:05:47 bootstrap-image finished... but still no .ext2 file Jul 06 17:06:17 I have IMAGE_FSTYPES = "ext2" in local.conf Jul 06 17:07:18 hang on, did task-bootstrap by mistake Jul 06 17:08:34 mickeyl: are the contents of 'PACKAGES' added to 'RPROVIDES'? Jul 06 17:08:47 dunno offhand Jul 06 17:08:49 but i'd think so Jul 06 17:08:57 check classes/* Jul 06 17:10:34 a quick look indicates that it doesn't Jul 06 17:10:52 I'll try to replicate the bug I was seeing Jul 06 17:10:58 * mickeyl going out tonite Jul 06 17:10:59 l8er Jul 06 17:11:18 NOTE: package busybox-1.2.0: completed Jul 06 17:13:36 finally a 'sleep' that supports arguments Jul 06 17:14:02 koen: fallen in love with busybox again ? Jul 06 17:14:21 ade|desk: loathing it less Jul 06 17:15:07 03koen 07org.oe.dev * rcf1dd943... 10/packages/busybox/busybox-1.2.0/angstrom/defconfig: busybox 1.2.0: add initial defconfig for angstrom Jul 06 17:15:36 252K /data/build/koen/OE/build/tmp/angstrom/deploy/ipk/busybox_1.01-r12_iwmmxt.ipk Jul 06 17:15:36 328K /data/build/koen/OE/build/tmp/angstrom/deploy/ipk/busybox_1.2.0-r0_arm.ipk Jul 06 17:15:38 hrm Jul 06 17:15:47 I guess we'll have to live with that Jul 06 17:16:34 I could leave out the cgi capable webserver, but my inner geek resists too much Jul 06 17:17:58 OMG Jul 06 17:18:08 "1 major gripe: 'ipkg-link add libc6' (linking to SD) froze my Z every time I ran it." Jul 06 17:19:06 koen: lol Jul 06 17:19:28 any ideas why bootstrap-image doest put a kernel in the image? Jul 06 17:19:40 koen: the 252k vs 328k is that ~same defconfig Jul 06 17:19:51 tkp: because the kernel isn't in RDEPENDS? Jul 06 17:19:56 ade|desk: no Jul 06 17:20:07 seems odd if it isnt :/ Jul 06 17:20:13 indeed Jul 06 17:20:14 koen: so there is hope :) Jul 06 17:20:31 70k is a small price to pay to solve lots of bugs Jul 06 17:20:46 too many apps assume a 'full' environment Jul 06 17:23:19 yay bootstrap-image 2.6M for ext2 Jul 06 17:23:58 need to strip out the dropbear and some pointless kernel modules Jul 06 17:24:00 ade|desk: cute, that's inside the initrd size limit for most systems Jul 06 17:24:17 * ade|desk does the uclibc happy dance Jul 06 17:25:14 koen: before the happy happy dance need to test on a target, but thats for tomorrow Jul 06 17:25:37 Having an issue. I used SRCDATE = 20060705 to try something out, now I want to go back to SRCDATE = 20060627 but then I get issues with opie-image cannot satisfying different tasks Jul 06 17:26:15 Ken: that's expected Jul 06 17:26:18 ok - something odd....BOOTSTRAP_EXTRA_DEPENDS which is set per machine has got the kernel in for x86, but not for epoa. not sure why this is, but im not sure where the best place to add it is - in our own distro config? Jul 06 17:26:49 BOOTSTRAP_EXTRA* is too heavily abused Jul 06 17:27:03 koen: heh Jul 06 17:27:20 don't try to make sense of it, just abuse it further ;) Jul 06 17:27:46 tkp: because the epia config is intended for netbooting, where you don't have the kernel inside the rootfs Jul 06 17:28:10 you can add it in your distro config file if you want Jul 06 17:28:24 ok sounds good Jul 06 17:29:05 fingers heavily crossed here Jul 06 17:29:27 grub seems to have trouble recognising the partition as the root partition when it is not hte boot disk Jul 06 17:32:09 woooh Jul 06 17:32:10 2.0M ep93xx/Angstrom-bootstrap-image-test-20060706-ep93xx.rootfs.tar.bz2 Jul 06 17:32:10 03koen 07org.oe.dev * re8fc3871... 10/packages/angstrom/ (angstrom-gpe-image.bb task-angstrom.bb): angstrom: some more tweaks to images and tasks Jul 06 17:38:35 koen: any good way around it? besides increasing SRCDATE? Jul 06 17:38:53 Ken: use gpe Jul 06 17:39:17 koen: gpe doesnt build for me :D gcc gives me hell when trying to compile it Jul 06 17:39:27 any solution that doesnt involve +300 hours? Jul 06 17:39:42 Ken: use a different gcc? Jul 06 17:39:47 tried them all Jul 06 17:39:53 all 3.x and all 4.x Jul 06 17:39:59 have you filed a bug for it? Jul 06 17:39:59 4.1.1 is the one that works best so far Jul 06 17:40:02 yeah Jul 06 17:40:09 not all though, cause there is too many Jul 06 17:40:21 SuperH isnt as well supported as Arm Jul 06 17:41:13 03koen 07org.oe.dev * rc1749e31... 10/packages/angstrom/task-angstrom.bb: task-angstrom: D'oh, somehow I forgot to add dropbear Jul 06 17:53:27 doh 4.1M Xserver image Jul 06 17:53:44 really need to strip that down Jul 06 17:57:13 hmm libstdc++ is big Jul 06 18:02:26 * mrz80 is back (gone 01:38:28) Jul 06 18:41:24 hi all Jul 06 18:41:47 has the user support channel disapeared? Jul 06 18:44:16 yay Jul 06 18:44:18 http://blogs.gnome.org/view/alexl/2006/07/06/0 Jul 06 18:45:05 Gnome is _eliminating_ a library? Jul 06 18:45:07 say it isn't so! Jul 06 18:45:12 it isn't so Jul 06 18:45:27 it is eliminating CORBA as well Jul 06 18:47:08 well, shut my mouth and call me Susan Jul 06 18:48:25 I need to get a kernel into our dist Jul 06 18:48:46 As I understand it, I need to add BOOTSTRAP_EXTRA_RDEPENDS += "kernel" Jul 06 18:48:59 virtual/kernel Jul 06 18:49:15 *if* you are using task-bootstrap Jul 06 18:49:20 koen: in local.conf? or in our dist.conf? Jul 06 18:49:28 bootsrap-image Jul 06 18:49:29 machine.conf Jul 06 18:49:40 ${MACHINE}.conf even Jul 06 18:49:45 but anywhere is fine Jul 06 18:50:04 so we should create our own machine.conf then (we're using epia atm which doesn't have it defined) Jul 06 18:50:42 yeah, or just add it to local.conf Jul 06 18:50:52 make it work first, worry later :) Jul 06 18:51:00 koen: he ;) Jul 06 18:51:11 local.conf it is then Jul 06 18:51:54 so... BOOTSTRAP_EXTRA_RDEPENDS += "virtual/kernel" ? Jul 06 18:54:38 that should do the trick Jul 06 18:54:58 or grep for "kernel" in conf/machine to see how other machines handle it Jul 06 18:54:59 cool... then maybe we can boot this thing :) Jul 06 18:55:19 x86 has it in BOOTSTRAP_EXTRA_RDEPENDS += .. Jul 06 18:55:44 but they say kernel, not virtual/kernel Jul 06 18:56:33 you could specify your kernel-name directly Jul 06 18:56:57 kernel name? or version? Jul 06 18:57:06 the one mentioned in the PREFERRED_PROVIDER_virtual/kernel line Jul 06 18:57:54 I don't have that line Jul 06 18:58:54 how's the .bb for your kernel called? Jul 06 18:59:12 you can put that name into BOOTSTRAP_EXTRA_RDEPENDS Jul 06 19:00:02 ah yes... PREFERRED_PROVIDER_virtual/kernel = "linux-epia" (in epia.conf) Jul 06 19:04:24 koen: I added that to my local conf and re-ran bootstrap-image... Jul 06 19:04:27 still no kernel Jul 06 19:04:41 do I have to clear some cache to get it to rebuild it with the kernel? Jul 06 19:05:18 rebuild task-bootstrap (iirc) Jul 06 19:05:40 can I rebuild that without having to rebuild the toolchain? Jul 06 19:05:54 or do you mean "bitbake rebuild bootstrap-image"? Jul 06 19:08:35 bitbake -c clean task-bootstrap Jul 06 19:08:49 right.... just worked that out Jul 06 19:09:34 i've flashed and had a quick play with 3.5.4.1 for both gpe and opie.. both seem pretty responsive.. first thing ive noticed is it insists on mounting my microdrive as follows /dev/hda /media/hda as vfat!! its partioned with one ext3 foratted partition however. Jul 06 19:10:13 dan2003: #openzaurus? Jul 06 19:10:36 ahh.. ok Jul 06 19:14:28 grr. still no kernel :( did `bitbake -c clean task-bootstrap`, `bitbake task-bootstrap`, `bitbake bootstrap-image` Jul 06 19:14:49 and I have BOOTSTRAP_EXTRA_RDEPENDS += "virtual/kernel" in local.conf Jul 06 19:16:46 try putting it in epia.conf Jul 06 19:17:07 I just realised that something *after* local.conf might be doing BOOTSTRAP_E_R = "" Jul 06 19:17:21 thereby zapping everything you added in local.conf Jul 06 19:17:25 hmm, righ Jul 06 19:17:26 t Jul 06 19:17:35 (it's a bug) Jul 06 19:19:01 anyway to speed up bitbake -c clean task-bootstrap (does it have to handle every single bitbake file first)? Jul 06 19:19:45 hmm Jul 06 19:19:48 koen: sorry for the barrage of questions Jul 06 19:19:53 is that bug still there? Jul 06 19:20:01 it isn't supposed to clean everything Jul 06 19:20:06 I'd ctrl-c it Jul 06 19:20:34 it doesn't clean everything... but it does NOTE: Handling BitBake files: / (2349/3549) [66 %] first Jul 06 19:20:37 which takes ages Jul 06 19:21:14 I'm guessing it's trying to work out which thing to clean Jul 06 19:21:43 ah, you touched a 'core' files, which triggers a cache flush Jul 06 19:21:48 file* Jul 06 19:22:10 local.conf, DISTRO.conf, MACHINE.conf Jul 06 19:22:11 'core' file... like epia.conf ? Jul 06 19:22:15 right Jul 06 19:22:31 hmm. ERROR: Nothing provides runtime dependency virtual/kernel Jul 06 19:22:41 after adding it to epia.conf Jul 06 19:22:57 even tho that file has: Jul 06 19:22:57 PREFERRED_PROVIDER_virtual/kernel = "linux-epia" Jul 06 19:23:47 i ran in to a similar problem last week Jul 06 19:23:59 I'll try with BOOTSTRAP_EXTRA_RDEPENDS += "linux-epia" Jul 06 19:24:12 that should certainly work Jul 06 19:27:22 also, it seems that IMAGE_FSTYPES in DISTRO.conf overrides the one on local.conf Jul 06 19:27:32 surely it should be the other way round Jul 06 19:27:54 ERROR: Nothing provides runtime dependency linux-epia??! Jul 06 19:27:57 ~change 89 usd to eur Jul 06 19:28:03 89.00 United States Dollar (USD) makes 69.6615 Euro (EUR) (from http://www.xe.com/) Jul 06 19:28:13 handy! Jul 06 19:29:11 weird Jul 06 19:29:28 ~change 1 cad to eur Jul 06 19:29:33 1.00 Dollar (CAD) makes 0.703472 Euro (EUR) (from http://www.xe.com/) Jul 06 19:29:38 ~change 1 cad to usd Jul 06 19:29:43 1.00 Dollar (CAD) makes 0.898687 United States Dollar (USD) (from http://www.xe.com/) Jul 06 19:31:36 ~change 630 usd to eur Jul 06 19:31:41 630.00 United States Dollar (USD) makes 493.076 Euro (EUR) (from http://www.xe.com/) Jul 06 19:33:05 ~change 950 usd to eur Jul 06 19:33:09 950.00 United States Dollar (USD) makes 743.595 Euro (EUR) (from http://www.xe.com/) Jul 06 19:35:33 oh ok, x86.conf (and lot's of others) just do kernel, rather than virtual/kernel or linux-MACHINE Jul 06 19:35:37 seems to have worked Jul 06 19:39:48 oh wait Jul 06 19:40:03 that works because the .ipk Provides: kernel Jul 06 19:40:10 *lightbulb* Jul 06 19:41:06 what's odd tho, is that the file /boot/bzImage is actually a symlink to /boot/bzImage-2.6.12 - which doesn't exist Jul 06 19:42:12 hrm Jul 06 19:42:19 check linux-epia*.bb Jul 06 19:42:32 it might be purging the zImage from the .ipk Jul 06 19:45:36 hmm. not sure I understand that Jul 06 19:46:27 ALLOW_EMPTY_kernel = "1" ? Jul 06 19:49:20 that says it's allowed to generate an empty .ipk for the kernel Jul 06 19:49:51 remove the FILES_kernel = "" Jul 06 19:49:58 and bump PR Jul 06 19:49:58 koen, know we talked about this earlier, but do you have any good suggestion for fixing errors due to turning back the SRCDATE. Jul 06 19:50:28 Kristoffer: if you turn back the srcdate you are basically pointing OE to non-existant software Jul 06 19:51:01 koen, How you mean? OE can only go forward not backwards? Jul 06 19:51:16 opie 1.2.2 didn't exist until very recently Jul 06 19:52:09 Yeah I know. Updating 20060627 -> 20060705 went well, but doing it the other way round created issues where it still wanted 20060705 software Jul 06 19:52:17 koen, no good away round it then? Jul 06 19:53:14 nope Jul 06 19:53:14 Btw we got any Opie core developers around? Jul 06 19:53:18 koen, oki thx Jul 06 19:53:19 #opie Jul 06 19:53:26 koen: bump PR? Jul 06 19:53:40 koen, oki but none that regulary idle here? Jul 06 19:53:49 change PR = "something" to PR = "something +1" in the kernel .bb Jul 06 19:54:17 ok, well it's r0 atm, so I'll make it r1 Jul 06 19:54:22 why is that necessary tho? Jul 06 19:55:33 tkp: you change the build output, so you want to make sure it is rebuilt and existing packages replaced Jul 06 19:56:00 right Jul 06 19:56:25 clean task-bootstrap is not enough then Jul 06 19:57:19 ah ok, not it's rebuilding the kernel Jul 06 19:57:24 now* Jul 06 19:59:30 koen: so, now that I'm altering some of oe's bb files, what happens when I do mtn pull & mtn update? will my changes be lost? Jul 06 20:00:15 monotone will merge it silently (no changes lost) or pop up a merging program like meld or kdiff3 Jul 06 20:00:27 it won't silently revert you changes Jul 06 20:00:33 ok, like svn conflicts Jul 06 20:00:51 perhaps it's better to make my own copies of these files tho Jul 06 20:01:46 damn it, bzImage is still a symlink to a non-existant file Jul 06 20:11:05 :D :D :D - it booted! :D Jul 06 20:12:26 tkp: congrats Jul 06 20:12:42 three days worh of hard work. I'm pleased Jul 06 20:13:05 Now I just need to turn it from a bootable base system, into the system I want it to be Jul 06 20:27:53 mickey|out, got a sec? Jul 06 20:34:29 hi Jul 06 20:42:16 hrw: hi :) Jul 06 20:42:22 hi marcos Jul 06 20:43:18 thejapa, hey Jul 06 20:44:04 Kristoffer: hey. got lucky with your SRCDATE? :) Jul 06 20:44:29 thejapa, yeah works like a charm. And also found some opie packages that were forgotten Jul 06 20:45:16 Kristoffer: like which? Jul 06 20:45:17 nice :) Jul 06 20:45:35 hrw, opie-stumbler and addressbook was misspelled Jul 06 20:45:54 noticed when it draged 20060706 instead of 20060627 Jul 06 20:46:05 Kristoffer: o.. Jul 06 20:46:24 havent moved them over to the other opie-versions file yet though Jul 06 20:48:44 so If I want to add a package to my distro, where would be the best/usual place to do it? Jul 06 20:49:02 tkp: distro or image? Jul 06 20:49:05 tpk package inside image or distro? Jul 06 20:49:18 well, into my final image Jul 06 20:49:32 tkp: add it into IPKG_INSTALL/RDEPENDS of tkp-image.bb Jul 06 20:49:58 so... export IPKG_INSTALL = "task-bootstrap mplayer" for example? Jul 06 20:50:13 yes Jul 06 21:10:09 cu Jul 06 21:18:21 cya Jul 06 21:19:05 mickey|out, ? Jul 06 23:17:27 all is silent **** ENDING LOGGING AT Fri Jul 07 02:59:56 2006