**** BEGIN LOGGING AT Wed Sep 27 02:59:56 2006 Sep 27 06:11:14 Good morning all Sep 27 07:05:05 morning Sep 27 07:06:06 koen: goodmorgen koen Sep 27 07:07:56 What is the smallest image available in OE? (preferably kernel + 1 init binary)? Sep 27 07:11:27 good morning all Sep 27 07:11:39 likewise: don't make me go zen on you :) Sep 27 07:12:37 RDEPENDS_woestenberg-image = "kernel-image busybox" Sep 27 07:12:44 or something like that Sep 27 07:21:32 koen: what will happen to me when you do go zen on me? Sep 27 07:28:16 "the smallest images is the no image at all" Sep 27 07:28:24 hrw|gone: congrats on the new job Sep 27 07:29:31 koen: but will it boot then? :-) Sep 27 07:30:15 morning Sep 27 07:30:51 koen: whats that terminal crgram you us Sep 27 07:31:33 likewise: "the fastest boot is no boot at all" Sep 27 07:31:41 XorA: ? Sep 27 07:31:47 koen: for serial Sep 27 07:31:53 XorA: mostly cu Sep 27 07:31:54 and that should have read program Sep 27 07:32:05 cu thats it, too early for memory banks to work Sep 27 07:32:07 cu supports colours :) Sep 27 07:39:12 my terrier is circling around edin :-) Sep 27 07:39:54 morning Sep 27 07:39:55 throw him a bone Sep 27 07:39:57 XorA: minicom Sep 27 07:39:59 hey hrw|work Sep 27 07:40:51 Crofton: ping Sep 27 07:41:23 XorA: not house trained yet ? Sep 27 07:42:00 Crofton: linux-omap1 does not build with gcc 4.1.1 Sep 27 07:42:01 | arch/arm/plat-omap/built-in.o: In function `omap_mcbsp_free': Sep 27 07:42:01 | mcbsp.c:(.text+0x3fa4): undefined reference to `omap_dsp_release_mem' Sep 27 07:42:01 | arch/arm/plat-omap/built-in.o: In function `omap_mcbsp_request': Sep 27 07:42:01 | mcbsp.c:(.text+0x406c): undefined reference to `omap_dsp_request_mem' Sep 27 07:45:44 x11 building requires host x11 headers :-( Sep 27 07:46:24 XorA: it does ? Sep 27 07:47:08 i don't have the headers on my box and as far as i can tell xlib builds fine Sep 27 07:49:33 diet-x11 makekeys needs them Sep 27 07:49:44 unless there is a libx11-native missing Sep 27 07:57:46 morning all Sep 27 07:57:59 hey RP Sep 27 07:58:41 XorA: i can't see where a native xlib would be called in from Sep 27 07:58:44 morning Sep 27 07:59:38 hey RP Sep 27 08:00:49 ade|desk: the do_compile Sep 27 08:03:47 RP: any idea on howto cleanly solve http://bugs.openembedded.org/show_bug.cgi?id=1440 ? Sep 27 08:06:03 koen: Add it to multimachine.bbclass I guess Sep 27 08:06:19 but where? Sep 27 08:07:06 between ${MULTIMACH_ARCH} and ${TARGET_OS} ? Sep 27 08:07:10 koen: s/${TARGET_OS}/${TARGET_VENDOR}-${TARGET_OS}/ Sep 27 08:08:29 ~lart wolfson proxy Sep 27 08:08:29 * ibot chops wolfson proxy in half with a free Solaris 7 CD Sep 27 08:09:37 http://oe.pastebin.com/795265 Sep 27 08:10:25 pastebin is on the go slow :-( Sep 27 08:10:56 * kergoth likes pastey.net Sep 27 08:11:26 kergoth: !!! what time is it over your way? Sep 27 08:11:35 only 1:30am Sep 27 08:11:39 koen: right, that should work Sep 27 08:11:51 koen: It will force a total rebuild of course... Sep 27 08:11:58 kergoth: you moved lately then ? Sep 27 08:12:18 NOTE: Unpacking /home/koen/OE/downloads/flex-2.5.31.tar.bz2 to /data/build/koen/OE/build/tmp/angstrom/work/i686--linux/flex-native-2.5.31-r2/ Sep 27 08:12:20 not in the past 8-10 months Sep 27 08:12:34 RP: and the '--' is target vender isn't set Sep 27 08:12:40 RP: but I can live with that :) Sep 27 08:12:58 koen: If we want to maintain backwards compatibility, drop the '-' Sep 27 08:13:02 * koen ponders about TARGET_VENDOR ?= none Sep 27 08:13:12 koen: heh, theres a reason that TARGET_VENDOR includes the - Sep 27 08:13:18 but with will result in i686-pokylinux Sep 27 08:14:02 ah, if TARGET_VENDOR includes a -, that change is wrong :) Sep 27 08:14:26 +STAMP = "${TMPDIR}/stamps/${MULTIMACH_ARCH}${TARGET_VEDNOR}-${TARGET_OS}/${PF}" Sep 27 08:14:27 thats the way its usually done, since the default TARGET_SYS doesnt seperate the vendor with a -, to avoid the -- issue koen just hit Sep 27 08:14:43 heh Sep 27 08:15:25 what about doing TARGET_VENDOR = "-oe" in bitbake.conf? Sep 27 08:15:27 kergoth: Is the dash a prefix or postfix? Sep 27 08:15:51 RP: TARGET_VENDOR = "-angstrom", so prefix Sep 27 08:15:54 koen: It would just break everything in sight. Maybe sometime when we're breaking things... Sep 27 08:16:17 koen: ok, I think you know what to do then :) Sep 27 08:16:22 RP: i never remember if its pre or post, i always, always have to check bitbake.conf when i'm gonna define a vendor. kinda annoying, but *shrug* Sep 27 08:18:50 hrw|work: ewi will get moved tomorrow morning and should be up again a few hours later Sep 27 08:21:05 koen: put it on wireless :-) Sep 27 08:22:42 :) Sep 27 08:23:26 koen: ok Sep 27 08:23:42 03koen 07org.oe.dev * rf4888077... 10/ (1 classes/multimachine.bbclass): multimachine.bbclass: include TARGET_VENDOR, closes #1440 Sep 27 08:28:56 being root on remote machine is weird.. Sep 27 08:29:06 'halt' Sep 27 08:29:15 followed by cries of agony Sep 27 08:30:48 koen: so is EWI dieing for good sometime? Sep 27 08:31:19 XorA: after it's being moved bureacracy sets in Sep 27 08:31:32 so it'll be there for a few weeks more Sep 27 08:32:59 koen: heh heh Sep 27 08:33:09 koen: so have we upset someone? Sep 27 08:33:27 no, the uni ended their ipaq projects afaics Sep 27 08:34:31 ~lart uni Sep 27 08:34:32 * ibot blasts uni to oblivion with a kamehameha wave Sep 27 08:35:03 XorA: I still have a pile of hardware on my desk the seem to have forgotten about Sep 27 08:35:18 koen: quick put that in a drawer Sep 27 08:56:08 ~seen crofton Sep 27 08:56:15 crofton is currently on #casualti (8h 17m 43s) #ol (8h 17m 43s) #oe (8h 17m 43s) #elinux (8h 17m 43s), last said: 'celf must have done their own machine file ...'. Sep 27 08:56:56 hrw|work: are you switching celf to generic or a custom distro? Sep 27 08:57:01 custom Sep 27 08:57:53 and is it going to live in .dev or into /dev/celf/null again? Sep 27 08:57:59 .dev Sep 27 08:58:40 In general, we Sep 27 08:58:40 want code and meta-data results pushed upstream into OE directly. Sep 27 08:58:40 Our automated build system downloads OE every day, so anything that Sep 27 08:58:41 makes its way into OE should be available to us the next day. Sep 27 08:59:33 'downloads OE every day' Sep 27 08:59:40 hey Sep 27 08:59:48 koen: fasist firewalls rox Sep 27 08:59:55 that spell 'huge waste of bandwidth' to me Sep 27 09:02:34 hrw|work: you working with CELF / Tim Bird?? Sep 27 09:03:06 likewise: will Sep 27 09:03:44 hrw|work: Ah ok, good! Sep 27 09:04:21 hey Sep 27 09:04:28 hi zecke Sep 27 09:04:37 hrw|work: that is really nice to hear! Sep 27 09:05:05 florian_kc: moin Sep 27 09:05:21 florian_kc: nils is away this week, right? Sep 27 09:05:22 hi Sep 27 09:05:33 koen: yes, why? Sep 27 09:06:34 florian_kc: So I can wait with asking difficult questions :) Sep 27 09:07:29 koen: :-) Sep 27 09:08:17 koen: he has network access... so maybe he has even more free time this week :-) Sep 27 09:09:44 zecke: tinderbox will be down tomorrow morning Sep 27 09:10:29 florian_kc: enough free time to fill in the kbdd svn repo? Sep 27 09:10:57 koen: maybe Sep 27 09:11:06 just a guess Sep 27 09:12:02 koen: I'm afraid you made a typing mistake in the last patch: TARGET_VEDNOR? Sep 27 09:12:55 ah, right Sep 27 09:13:06 I only checked if the workdir looked sane Sep 27 09:13:32 np, thanks for fixing. Sep 27 09:13:56 Right. It's alive! When building glibc-intermediate, does the error "undefined reference to `__atfct_seterrno'" ring any bells? Sep 27 09:14:24 03koen 07org.oe.dev * r4f6bdeca... 10/ (1 classes/multimachine.bbclass): multimachine.bbclass: fix typo spotted by mr Leemans Sep 27 09:14:36 "guinness.conf" sounds good :-) Sep 27 09:15:26 florian_kc: it's this board: http://www.tyan.com/products/html/thunderk7.html Sep 27 09:15:41 koen: could you delete my home dir please? Sep 27 09:15:54 koen: and make a backup of the tinderbox Sep 27 09:15:55 mr... makes me think about an old pub about fast food ;-) Sep 27 09:15:55 zecke: sure, right now? Sep 27 09:16:38 koen: home dir? yes Sep 27 09:16:45 zecke: does tinderbox use postgres? Sep 27 09:16:55 koen: yes it did Sep 27 09:20:15 zecke: /data/tinderbox1 has been backed up to openemebedded.org:~/backup/tinderbox/ Sep 27 09:29:28 time to backup changes to repo Sep 27 09:31:20 mtn commit -b org.oe.hrw-unsorted-backup ? Sep 27 09:32:26 rather sort changes in .oz354x and push Sep 27 09:33:01 yay, everyone like a pay rise :) Sep 27 09:34:34 03hrw 07org.oe.oz354x * r523d4d2f... 10/ (1 conf/distro/openzaurus-3.5.4.2.conf): oz 3.5.4.2: bump SRCDATE for zaurusd to get poodle support Sep 27 09:34:39 03hrw 07org.oe.oz354x * r432081ea... 10/ (1 packages/zaurusd/zaurusd_svn.bb): zaurusd: OZ 3.5.4.2 use newer snapshot with poodle support Sep 27 09:34:45 03hrw 07org.oe.oz354x * r3e44aeab... 10/ (1 conf/machine/include/poodle-2.6.conf): poodle-2.6: use zaurusd instead of alsa-settings Sep 27 09:34:51 03hrw 07org.oe.oz354x * rbdbd69d2... 10/ (1 packages/linux/linux-openzaurus_2.6.16.bb): linux-oz 2.6.16: remove stuff for tosa/poodle/collie Sep 27 09:35:30 hrw|work: does celf really have a fascist firewall, or where they too lazy to learn mtn? Sep 27 09:36:03 koen: will check when start work for them Sep 27 09:38:41 downloading oe.mtn.bz2 daily = 32GB of traffic per year Sep 27 09:40:14 * koen has a look at xdelta Sep 27 09:40:35 koen: what about rsync? Sep 27 09:41:55 NAbyss: if they can open up an rsync port, they can open up the mtn port Sep 27 09:42:15 koen: True, true Sep 27 09:49:56 good morning Sep 27 09:50:20 hi mickeyl Sep 27 09:50:33 * mickeyl unpacks his brand new shiny x60s Sep 27 09:51:05 hey mickeyl Sep 27 09:51:12 mickeyl: is that the tablet one? Sep 27 09:51:13 Woot. Never mind - hacked around it. openat.c seems to be getting patched badly somewhere along the line. Sep 27 09:51:21 or the one that explodes? Sep 27 09:51:27 koen: no, the x60 is not available as tablet (yet). Sep 27 09:51:36 koen: the x60s is the one with the core duo LV Sep 27 09:51:42 1.2 Kg Sep 27 09:51:48 not powerful, but very light Sep 27 09:51:51 which is what I like Sep 27 09:51:56 a powerfull explosive :) Sep 27 09:52:14 heh Sep 27 09:52:17 i hope not Sep 27 09:52:31 mickeyl: I'm building angstrom for my workstation right now :)_ Sep 27 09:52:35 s/-// Sep 27 09:52:44 koen: excellent. I'm not very far from using OE on my main machine as well ;) Sep 27 09:53:00 it probably doesn't suck less, but at least I can fix it ;) Sep 27 09:53:31 I'm anxious to try out the .dev/apt stuff the o-hand are working on Sep 27 09:53:38 s/dev/deb/ Sep 27 09:53:40 ~summon crofton Sep 27 09:53:42 apt takes out 20 clean, identical-looking phones, some extra hands, and pretends to be a telemarketer for a large corporation, so he gets delivered a phonelist containing crofton's coordinates Sep 27 09:54:08 heh, that's a cool summon Sep 27 09:54:13 ~summon mickeyl Sep 27 09:54:14 apt takes out 20 clean, identical-looking phones, some extra hands, and pretends to be a telemarketer for a large corporation, so he gets delivered a phonelist containing mickeyl's coordinates Sep 27 09:55:15 btw, we really need a clear policy on which kernel modules to enable in OE Sep 27 09:55:20 *nod* Sep 27 09:55:32 could you add that to oedem session Sep 27 09:55:38 perhaps the 2nd one Sep 27 09:55:39 ? Sep 27 09:55:49 * mickeyl likes installing new computers because he can come up with a new name Sep 27 09:57:32 mickeyl: done Sep 27 09:57:35 thanks Sep 27 09:57:49 koen: why policy? its matter of machine maintainers rather Sep 27 09:58:33 hrw|work: we have a lot of packages that need a kernel module to function Sep 27 09:59:01 like all the routing/firewall/dns packages Sep 27 09:59:47 i think it's a good think to discuss it -- even if the result is that we don't need a policy Sep 27 09:59:49 hi mickeyl Sep 27 09:59:52 s/think/thing/ Sep 27 09:59:55 morning florian Sep 27 10:00:05 how's the kiddy? Sep 27 10:01:17 Crofton: http://pastebin.ca/183871 is needed to build linux-omap 2.6.x-git for omap5912osk machine Sep 27 10:01:21 hungry... :-) Sep 27 10:01:28 hehe Sep 27 10:01:35 nothing unusual ;) Sep 27 10:02:02 morning all Sep 27 10:02:24 uargh Sep 27 10:02:28 how i hate preinstalled windows Sep 27 10:02:36 like... 1 million taskbar symbols Sep 27 10:02:40 1 million services in background Sep 27 10:02:40 ;) Sep 27 10:02:42 ~lart lenovo Sep 27 10:02:43 * ibot blasts lenovo with a huge firehose then strangles lenovo with it Sep 27 10:02:47 morning lrg Sep 27 10:02:53 hi Liam Sep 27 10:02:59 hey lrg Sep 27 10:03:13 mickeyl: and a million ISP shortcuts Sep 27 10:03:35 mickeyl: he destroyed his first bowl yesterday :-) Sep 27 10:03:37 ~curse git fetcher badly Sep 27 10:03:40 hi lrg Sep 27 10:03:43 May you be reincarnated as a Windows XP administrator, git fetcher badly ! Sep 27 10:06:28 git fetcher is CRAP Sep 27 10:06:53 fetch, pack, pack, archive. next time do fetch, archive Sep 27 10:07:12 Already up-to-date. Sep 27 10:07:12 No changes. Sep 27 10:07:12 NOTE: Creating tarball of git repository Sep 27 10:07:47 ${TARGET_VEDNOR} who needs a spelling checker :-) Sep 27 10:08:09 XorA: you need to use mtn pull/merge/update before reporting bugs Sep 27 10:08:16 ;D Sep 27 10:09:03 I want sources.oe.org service... Sep 27 10:15:48 hrw|work: that's a good idea... Sep 27 10:17:38 3rd rebuild of kernel.. Sep 27 10:17:52 ARGH... git fetcher again.. Sep 27 10:20:19 building git kernels can be split into two equal time parts: 1. fetching nothing 2. rest Sep 27 10:21:12 XorA: don't you know that 95% of all typos in OE were pushed by me? Sep 27 10:23:48 koen: heh heh Sep 27 10:24:22 djeez Sep 27 10:24:38 I'm getting so pissed off at people that say "OMG BUGZ!!!! YOU SUCKZORS" Sep 27 10:24:44 03rwhitby 07org.oe.dev * r39125d69... 10/ (3 files in 2 dirs): ftpd-topfield: Update to latest version. Sep 27 10:25:02 but the same people don;t wan't to say what to bugs are nor report them in bugzilla Sep 27 10:26:42 RP: how hard would it be to teach OE/bitbake 'bitbake nano:poodle nano:x86 iptables:wrt54g' ? Sep 27 10:27:51 koen: whining is always the easy choice Sep 27 10:30:20 fsck.. rm_work strikes again Sep 27 10:30:44 morning all Sep 27 10:30:52 hey Dirk Sep 27 10:30:58 hi Liam Sep 27 10:31:12 NOTE: package linux-omap1-2.6.x+git: completed Sep 27 10:31:13 uf Sep 27 10:31:35 Crofton: where are you.... Sep 27 10:31:37 hi Dirk Sep 27 10:31:42 hey Marcin Sep 27 10:31:50 hrw|work: maybe crofton is sleeping? Sep 27 10:31:52 hey Dirk Sep 27 10:32:28 hi Koen Sep 27 10:32:31 koen: I hope that he sleep good Sep 27 10:33:01 oz-unstable for omap5912osk progressing Sep 27 10:33:14 3h for getting kernel built is a bit too much Sep 27 10:58:29 koen: Its funny you should ask that. I've been considering quilt;mach=native ... Sep 27 11:15:16 he was sleeeping Sep 27 11:17:11 hrw|work, the problem with the git kernel is the underlying source always changes :) Sep 27 11:17:39 I need to check that the kernel config is OK, since recently some dsp gw changes went in Sep 27 11:22:56 morning Crofton Sep 27 11:23:31 Crofton: can you tag 2.6.18 kernel in it? Sep 27 11:24:37 at some point, I want to add a bb file that builds from a tag Sep 27 11:24:45 to prevent this kind of stuff Sep 27 11:24:49 yep Sep 27 11:24:55 I forget why I didn't do this Sep 27 11:24:57 :) Sep 27 11:25:03 Crofton: I also had problems with deploying Sep 27 11:25:14 I suspect I see changes I want in the OSK soon flowing into git Sep 27 11:25:25 what kind of problem? Sep 27 11:25:33 u-boot is built against a tag Sep 27 11:25:43 btw, I am glad for help :) Sep 27 11:26:00 there seems to be increasing interest in the OSK Sep 27 11:26:56 Crofton: vmlinux was not in arch/arm/boot/compressed/ Sep 27 11:27:38 duh Sep 27 11:27:43 vmlinux is uncompressed Sep 27 11:27:52 vmlinuz would be :) Sep 27 11:32:10 Crofton: did you looked at MACHINE_FEATURES stuff? Sep 27 11:32:39 Crofton: MACHINE_FEATURES = "kernel26 pcmcia usbhost" would be ok for omap5912osk? Sep 27 11:33:22 not yet Sep 27 11:33:55 if you understand them, please update Sep 27 11:34:09 I can't work seriously on it until next week Sep 27 11:34:32 I still have a proposal and a paper to finsh Sep 27 11:35:05 Crofton: ok Sep 27 11:35:08 hrw|work, the deploy has worked recently Sep 27 11:35:15 maybe the kernel build failed Sep 27 11:35:27 Crofton: I will rebuild to check Sep 27 11:35:31 hi Mike Sep 27 11:35:37 thanks Sep 27 11:35:54 I will keep my data updated Sep 27 11:36:05 does busybox still eat your machine when it builds? Sep 27 11:36:14 no Sep 27 11:36:26 I want to do a rebuild, but the machine is also our svn server Sep 27 11:36:37 and it gets really slow when bb rebuilds .... Sep 27 11:37:34 I build now on my home machine (remotely) Sep 27 11:37:55 I need to install on my new home machine when I have time Sep 27 11:47:00 | install: cannot stat `arch/arm/boot/vmlinux': No such file or directory Sep 27 11:47:00 NOTE: Task failed: /home/hrw/devel/build/dev/tmp/work/omap5912osk-linux/linux-omap1-2.6.x+git-r0/temp/log.do_install.14922 Sep 27 11:47:12 bother Sep 27 11:50:38 morning Sep 27 11:51:09 hmmm Sep 27 11:51:16 I have vmlinux in my boot Sep 27 11:51:45 14:09 hrw@home:dev$ find tmp/work/omap5912osk-linux/linux-omap1-2.6.x+git-r0/git/ -name vmlinux Sep 27 11:51:48 tmp/work/omap5912osk-linux/linux-omap1-2.6.x+git-r0/git/vmlinux Sep 27 11:52:12 have you looked at OSK newbies guide? Sep 27 11:53:50 not yet Sep 27 11:54:06 the path is from that, verbatim ... Sep 27 11:54:23 http://www.capgo.com/Resources/SoftwareDev/osk-newbie-guide.pdf one? Sep 27 11:54:44 pretty sure Sep 27 11:54:51 first hit on google is the pdf Sep 27 11:55:12 I am goign to clean and rebuild kernel here Sep 27 11:56:20 ok Sep 27 12:08:45 hrw, mine fails to compile Sep 27 12:09:03 Crofton: lack of compilefix.patch? Sep 27 12:09:03 need to look at that the .config Sep 27 12:09:15 probably Sep 27 12:09:16 Crofton: enable CONFIG_OMAP_DSP Sep 27 12:09:23 are you on the linux-omap list? Sep 27 12:09:26 yeah Sep 27 12:09:26 Crofton: or use my hack Sep 27 12:09:27 no Sep 27 12:09:38 It needs CONFIG_OMAP_DSP Sep 27 12:09:48 we want the DSP :) Sep 27 12:09:56 03koen 07org.oe.dev * r357d97e8... 10/ (1 conf/machine/guinness.conf): guinness: add MACHINE_FEATURES Sep 27 12:10:13 I can do this later today Sep 27 12:10:17 when I get in to school Sep 27 12:10:35 Crofton: do you remember url to linux-omap ML? Sep 27 12:10:39 they split up the DSP gw some between omap1 and 2 stuff very recently Sep 27 12:10:55 Crofton: http://linux.omap.com/mailman/listinfo/linux-omap-open-source one? Sep 27 12:11:02 http://linux.omap.com/mailman/listinfo/linux-omap-open-source Sep 27 12:11:05 yeah Sep 27 12:11:17 It is a bug that it does not compiel withthe gw disabled Sep 27 12:11:24 and should be submitted Sep 27 12:14:36 oooh Sep 27 12:14:41 dspfs looks nice Sep 27 12:14:54 fir the bridge? Sep 27 12:15:27 Crofton: which distro you use for omap work? Sep 27 12:15:35 openomap Sep 27 12:15:45 but it may make sense t use a more standard one Sep 27 12:15:54 Crofton: you maintain it? Sep 27 12:15:59 sort off Sep 27 12:16:08 ok Sep 27 12:16:12 I use it to set tool chan version mainly Sep 27 12:16:37 by having a custom distro, it minimizes external change Sep 27 12:16:53 sure Sep 27 12:16:57 but on the other hand, I am not opposed to using something like angstrom, if it works ok Sep 27 12:18:13 Crofton: you use nfs, pcmcia, usbhost on devboard? Sep 27 12:19:26 yes Sep 27 12:19:37 pcmcia for microdrive Sep 27 12:19:41 usb for USRP Sep 27 12:19:46 k Sep 27 12:19:54 root over nfs is very handy Sep 27 12:35:44 opie-mediaplayer1: "cannot find -lqtaux2" - does anyone know what causes that offhand? Sep 27 12:36:20 RP: what unit is ramdisk_size in? Sep 27 12:37:00 XorA: no idea... Sep 27 12:37:10 XorA: on the kernel commandline? Sep 27 12:37:32 RP: yeah, just if you knew off top of head Sep 27 12:37:57 XorA: If I had to guess I'd say it would depend on the postfix M for MB... Sep 27 12:39:03 hrw|work, Sep 27 12:39:13 Am goin to start my project in OMAP ..."Porting linux on OMAP" Sep 27 12:39:14 plz any one help me how to start with....what are the things required... Sep 27 12:39:14 like bootloaders,toolchain and etc.... Sep 27 12:39:27 from l-o-o-s list Sep 27 12:40:03 l-o-o-s-er list? Sep 27 12:40:08 (too easy ;)) Sep 27 12:40:13 heh Sep 27 12:40:18 too long a name .... Sep 27 12:40:53 all the cool people are on the davinci list anyway :) Sep 27 12:43:01 monotone has been updated to 0.30, which is using a new database and workspace format ( http://www.venge.net/monotone/UPGRADE ). Are there plans to migrate to 0.30? Sep 27 12:43:22 netsync is compatible Sep 27 12:43:37 you can upgrade, even if the main servers do not Sep 27 12:43:59 Crofton, nice, thanks for the info :) Sep 27 12:44:07 I asked yesterday Sep 27 12:44:14 I haven'tdone the upgrade though Sep 27 12:45:29 i'm testing it right now and it seems to work (i haven't changed the db-forma locally though) Sep 27 12:45:37 main pain with 0.30 upgrade is roster regeneration (not even migration!) - took >3h for me ;-) Sep 27 12:46:38 well, i think i should stick with 0.29 db and use mtn-0.30 (as this combo seems to work) Sep 27 12:46:46 koen: pointed out there are snapshots of 0.30 databases on oe.rg Sep 27 12:46:50 oe.org Sep 27 12:47:15 a day too late for me unfortuneately, I think he owes me a NBP :-) Sep 27 12:47:24 EvilDevil, you need to migrate your db file. just to it at night ;-) Sep 27 12:47:49 the filenames in http://www.openembedded.org/snapshots/ are fairly self explanatory Sep 27 12:48:47 hi Sep 27 12:49:19 i have a prb compiling a simple nano ... anyone can help me ? i have followed step by step the wiki Sep 27 12:49:25 koen, can i use the snapshot for nslu2-linux.mtn too? Sep 27 12:50:16 EvilDevil: it contains org.oe.dev, not the bitbake mirrors Sep 27 12:55:48 re Sep 27 12:55:51 Crofton: ? Sep 27 12:59:40 RP: MACHINE_FEATURES = "keyboard" Sep 27 13:01:43 hrw|work: and 'touchscreen' as well? Sep 27 13:02:03 kergoth, thanks for the help with glibc las night.. you were correct, somehow the uclibc got build over the libc. I rebuilt it and everything works now. I'm not sure it happened as I know I never manualy built it.. Sep 27 13:02:11 MACHINE_FEATURES = "puntable" Sep 27 13:02:58 koen: good point Sep 27 13:05:55 devboards not always has a keyboard so keymaps+console-tools are not necessary Sep 27 13:07:13 man, 'cross' compiling for i686 takes a while Sep 27 13:10:34 hrw|work: I guess it we should indicate such features... Sep 27 13:11:00 I bumped a thread about task-base on the mailinglist today Sep 27 13:28:08 how do you think - ipkg shoudl be in task-base or just added into rootfs? Sep 27 13:28:30 if rootfs is built with ipkg packages then add ipkg or not (in other words) Sep 27 13:31:10 03koen 07org.oe.dev * r5a911227... 10/ (1 classes/multimachine.bbclass): multimachine: remove spurious hyphen in staging-kernel-dir Sep 27 13:31:26 pushing some task-base changes... Sep 27 13:31:52 03hrw 07org.oe.dev * r14f3e63a... 10/ (8 files in 3 dirs): zaurus machines: renamed zaurus-clamshell-2.6.conf into zaurus-2.6.conf as it is used in all 2.6 powered models Sep 27 13:31:58 03hrw 07org.oe.dev * r25bb53fb... 10/ (1 conf/machine/include/zaurus-2.6.conf): zaurus-2.6.conf: added keyboard, touchscreen into features, added tosa info Sep 27 13:35:51 any volunteers for documenting task-base in the user-manual? Sep 27 13:37:25 koen: before or after oedem? Sep 27 13:37:48 right now would be a good start :) Sep 27 13:38:09 03hrw 07org.oe.dev * reef15408... 10/ (1 packages/tasks/task-base.bb): task-base: added support for keyboard feature Sep 27 13:38:19 Over 4k packages, btw Sep 27 13:38:22 i've a problem compiling gnash with libz Sep 27 13:38:24 http://dark-code.bulix.org/vyu0ie-18454 Sep 27 13:38:42 blaster8: I had >4k weeks ago. Sep 27 13:38:55 hmm Sep 27 13:39:00 if someone is interessting by gnash in OE , i need a bit help Sep 27 13:39:04 ok, I'm pulling org.openembedded.dev only Sep 27 13:39:08 * mickeyl crosses fingers while resizing the main partition via partition magic... Sep 27 13:39:11 03hrw 07org.oe.dev * rf61787de... 10/ (1 packages/tasks/task-base.bb): task-base: bumped PR Sep 27 13:41:24 I will bump PR only once per push - more changesets will follow Sep 27 13:41:35 Genesis: http://rafb.net/paste/results/RS8mtY62.html Sep 27 13:42:15 i made a bb file me too Sep 27 13:42:36 that bb file is a few weeks old Sep 27 13:42:45 gnash insists on compiling opengl support Sep 27 13:42:49 i'ven't it in my OE Tree Sep 27 13:43:04 koen, : you dont need it if you use SDL , isn't it ? Sep 27 13:43:19 i show my bb Sep 27 13:43:23 right, you don't need it, but there's no way to turn it off Sep 27 13:43:48 03koen 07org.oe.dev * rff8f16a8... 10/ (5 files in 2 dirs): h2200, h5xxx, ipaq-pxa270, netbook-pro: add MACHINE_FEATURES Sep 27 13:44:13 http://rafb.net/paste/results/e8QEJD67.html Sep 27 13:44:23 --enable-gui=sdl Sep 27 13:44:40 did you tried with that ? Sep 27 13:44:59 no, I think the release didn't have that Sep 27 13:45:23 ho Sep 27 13:45:26 try Sep 27 13:45:33 koen: oh, good work Sep 27 13:45:38 if you've some time of course :) Sep 27 13:46:38 i think my actual pb is a OE pb dealing with libz Sep 27 13:46:45 more than a gnash one Sep 27 13:47:40 ( my bb file is for cvs since gnash doesnt release often ) Sep 27 13:48:18 03koen 07org.oe.dev * r2061d23a... 10/ (1 conf/machine/a780.conf): a780: add MACHINE_FEATURES, the 'keyboard' one might be a bit misplaced Sep 27 13:49:25 koen: a780 use keymaps package? Sep 27 13:49:45 it could do that for the keypad Sep 27 13:49:59 RP: something for you Sep 27 13:50:53 03hrw 07org.oe.dev * r1595838e... 10/ (1 packages/tasks/task-base.bb): task-base: portmap is needed for NFS not for core system Sep 27 13:50:58 03hrw 07org.oe.dev * r28236de9... 10/ (1 packages/tasks/task-base.bb): task-base: hostap-cs only if DISTRO has wifi in features Sep 27 13:51:04 03hrw 07org.oe.dev * rab63b160... 10/ (1 packages/tasks/task-base.bb): task-base: if there are usbhost and bluetooth supported then add module for USB BT Sep 27 13:51:09 03hrw 07org.oe.dev * r0320e8a3... 10/ (1 packages/tasks/task-base.bb): task-base: if there are usbhost and irda supported then add module for USB Irda Sep 27 13:51:15 03hrw 07org.oe.dev * r330197c2... 10/ (1 packages/tasks/task-base.bb): task-base: do not add USB BT/Irda into usbhost list - they are part of bt/irda already Sep 27 13:51:20 03hrw 07org.oe.dev * r1e14a41a... 10/ (1 packages/tasks/task-base.bb): task-base: added portmap into task-distro-nfs, added cifs into smbfs Sep 27 13:51:33 koen, : sdl backend from cvs should work Sep 27 13:52:04 03hrw 07org.oe.dev * rd5610a07... 10/ (1 packages/tasks/task-base.bb): task-base: bump PR Sep 27 13:57:23 hrw|work: you forgot one :) Sep 27 13:57:31 03koen 07org.oe.dev * r4b32a45f... 10/ (1 packages/tasks/task-base.bb): task-base: include tslib for 'touchscreen' Sep 27 14:02:48 koen: tslib-conf OK. but not tslib Sep 27 14:03:08 because tslib-conf pull in tslib? Sep 27 14:03:31 koen: if machine has touchscreen then it need config for tslib. it does not need tslib because it will get by dependencies of libs/apps which use tslib Sep 27 14:04:02 opie-image will get tslib because qte use tslib. gpe-image will get tslib because xtscal use it etc.. Sep 27 14:05:50 tslib should probably RECOMMEND tslib-config Sep 27 14:06:00 hrw|work: Looks good at a quick glance :) Sep 27 14:06:24 RP: it should Sep 27 14:07:13 RP: I will left it for you then ;D Sep 27 14:08:33 hrw|work: tslib RDEPENDS on tslib-conf so we're safe Sep 27 14:08:35 about tslib, i've a little bug with it , when i lauch xtscal first , calibration failed, if i lauch after lauching my WM ( matchbox ) , that work fine Sep 27 14:08:47 if someone have an idea :) Sep 27 14:09:02 koen: can you revert last changeset then? :D Sep 27 14:09:34 hrw|work: what is you don't build gpe or opie, but want to have tslib from some reason? Sep 27 14:09:53 RP: why are fuser and setserial in 'core'? Sep 27 14:09:55 koen: Its a lib - it would get picked up by shlibs for anything that needs it Sep 27 14:10:04 koen: then I will install something for usage of TS Sep 27 14:10:13 koen: for example tslib-tests Sep 27 14:10:43 koen: I'm not sure, they were just there traditionally and I didn't want to break compatibility at the time Sep 27 14:11:28 koen: setserial can move to task-serial Sep 27 14:11:40 and busybox has fuser iirc Sep 27 14:11:43 koen: but task-pcmcia, task-usbhost will get them Sep 27 14:11:49 koen: >1.01 only Sep 27 14:25:23 is current cvs broken???? Sep 27 14:26:27 i always got : Sep 27 14:26:27 ERROR: Nothing provides dependency autoconf-native Sep 27 14:26:27 ERROR: dependency autoconf-native (for nano) not satisfied Sep 27 14:26:27 NOTE: no buildable providers for nano Sep 27 14:29:47 gumz, you got different bitbake paths wrong. Read GettingStarted wiki. Sep 27 14:30:08 (or at least my telepathy engines suggests that's the cause) Sep 27 14:31:21 ok i re-check Sep 27 14:32:20 is there anything special I have to do to build OE using nptl as the thread libs? Sep 27 14:32:23 cvs? Sep 27 14:32:59 I've been using the generic distro and it does not seem to have built the libs with nptl. Sep 27 14:33:06 Gerrath: GLIBC_ADDONS = "nptl" Sep 27 14:33:31 koen, in local.conf? Sep 27 14:33:45 or in glibc.bb Sep 27 14:33:53 03hrw 07org.oe.dev * r56ae59ea... 10/ (1 packages/tasks/task-base.bb): disapproval of revision '4b32a45f45bd0328a8607e52c07b92d6efd62e6a' Sep 27 14:34:19 anywhere Sep 27 14:36:33 koen, ok I guess it is already included in the glibc_2.4, I'm using glibc_2.3.90 for some reason.. I'll just rebuild with 2.4 and be up to date :-) Sep 27 14:37:45 gumz: cvs? we don't use cvs Sep 27 14:38:21 ok, BTW, always same prb Sep 27 14:40:52 hi Tom Sep 27 14:42:53 03hrw 07org.oe.dev * r71292004... 10/ (1 packages/tasks/task-base.bb): task-base: added modules needed for WEP/WPA support into wifi section Sep 27 14:56:46 I have ajusted the configurations of several OE packages to meet my specific needs, what is the best way to keep track of these changes and also apply them to a fresh OE download? Should I generate a monotone patch or just use normal diff or is there a better way? Sep 27 14:57:23 Gerrath: generate patches and submit them to us Sep 27 14:57:28 'mtn diff . > ../foo.diff ; patch < ../foo.diff' ? Sep 27 14:57:29 then pester us to get them included Sep 27 14:57:44 koen: quilt will be better Sep 27 14:58:55 hrw|work: good morning Sep 27 14:58:58 hrw|work, no I don't think these patches should be submitted as they pertain to things like specific PHP mods to be compiled and not to be compiled. They are very "individual requirement" dependent. Sep 27 14:59:53 hrw|work, although there is one (which is a real bug) I submitted as a bug report and I'm not sure what is the best way to fix it. I would love some input: http://bugs.openembedded.org/show_bug.cgi?id=1418 Sep 27 15:00:36 koen, thanks. Sep 27 15:00:36 Gerrath: if you want to keep track of local changes commit them to your mtn db Sep 27 15:01:04 Gerrath: I suggest using your own branch and propagatig from org.openembedded.X to your branch when you pull Sep 27 15:01:06 JustinP, ok, thats what I thought, I just wanted to make sure it was the best method. Sep 27 15:01:09 Gerrath: no idea Sep 27 15:04:48 koen: you mentioned in the wiki edit that the task-bootstrap was soon to be changed in favor of task-base? IOW, that the strategy of the task-.bb vs. -image.bb is about to go an rework? Sep 27 15:05:16 how is the rootfs generated in OE? Is there a skeleton directory somewere that is copied into tmp/rootfs during build and then the apps are installed into it? Sep 27 15:05:32 Gerrath: no, and yes Sep 27 15:05:39 Gerrath: atually, the entire rootfs is created by packges Sep 27 15:05:54 s/packges/packages/ Sep 27 15:06:08 CosmicPenguin, ok, that makes sense.. Sep 27 15:06:13 T0mW: machine config need to have MACHINE_TASK_PROVIDER = "task-base" and then you build any image Sep 27 15:06:32 Gerrath: the strategy is that the task-bootstrap.bb for example will collect info on what ipk's should be installed, then the image-bootsrap.bb actually installs them into the rootfs image tree. Sep 27 15:06:38 T0mW: task-base will get built instead of task-bootstrap. your machine and distro has to be ready for it Sep 27 15:07:05 T0mW: machine need to have MACHINE_FEATURES listed and distro has DISTRO_FEATURES Sep 27 15:07:13 * CosmicPenguin boggles Sep 27 15:07:16 T0mW!!! Sep 27 15:07:36 You are the last guy I expected to find in here Sep 27 15:07:36 hrw|work: what I'm seeing is that, for example, task-bootstrap.bb is too specific. for example, someone put hotplug, portmap and dropbear into that task. I don't want those packages! Sep 27 15:07:39 well - second to last Sep 27 15:07:39 T0mW: when I will find time then I will add it into user-manual Sep 27 15:07:42 but still, thats pretty good Sep 27 15:07:55 I use tmpfs for my var directory so there is minimum flash writing.. I wanted to blitz all files/directories in the var directory so there would be no extra space wasted on my final rootfs.jffs2 image and was trying to figure out the cleanest way to do that.. Sep 27 15:07:58 ~lart Jordon for making assumptions Sep 27 15:07:58 * ibot moos at Jordon for making assumptions Sep 27 15:08:36 CosmicPenguin: yeah, you take what works, then hassle people to make it better. heheh Sep 27 15:08:38 T0mW: portmap is only installed if nfs is in features, dropbear and hotplug are by default now Sep 27 15:08:53 hrw|work: right, they shouldn't be Sep 27 15:09:08 I guess I should be able to blitz it in my main meta file. Sep 27 15:09:16 T0mW: then propose better solution? Sep 27 15:09:28 hrw|work: that leaves the distributor no choice, the dropbear + hotplug should be part of the distro. Sep 27 15:09:56 hrw|work: IMO, the task-builder should be more generalized. Sep 27 15:10:19 T0mW: we can/should probably move hotplug into pcmcia/usbhost tasks Sep 27 15:10:43 hrw|work: I've had to create my own task-bootzipit.bb and bootzipit-image.bb to get around some default assumptions as to what packages would be in an OE based distro. Sep 27 15:10:44 T0mW: or add newer feature: hotplug Sep 27 15:11:52 hrw|work: the zipit is a closed-box, IOW, there is no usb, etc., that devices would plug into it. So hotplug is just a time waster for the initial boot. Sep 27 15:12:03 T0mW: task-bootstrap is hopelessly bloated Sep 27 15:12:10 koen: agreed Sep 27 15:12:13 T0mW: that's why OE is moving to task-base Sep 27 15:12:45 T0mW: 2.6 kernel? Sep 27 15:13:05 T0mW: you do not use udev at all? Sep 27 15:13:13 koen: yeah, we go from specific --> generalized --> specific again when traversing from machine + distro + local confs to the tasks/images. odd Sep 27 15:13:28 hrw|work: I have to use a 2.4 kernel in this box. Sep 27 15:13:36 ahok Sep 27 15:13:44 hrw|work: I do use devfs though Sep 27 15:13:58 T0mW: then task-base should not give you any hotplug Sep 27 15:14:20 hrw|work: the limitation on the kernel version is a binary only wifi module tied to 2.4.21 Sep 27 15:15:21 hrw|work: I'm almost finished with this phase of the project, just have some fbcon funnies to clear up, then I'll want to commit what I have into OE. Sep 27 15:17:09 ok Sep 27 15:17:19 koen: It almost seems that the task-*'s themselves should be done away with or merged into the *-image's. There seems to be a peer-to-peer relationship that exists there that shouldn't? Sep 27 15:17:29 I'm thinking about adding 'modules' into features Sep 27 15:17:50 so machines without kernel modules will not get module-init-tools/modutils at all Sep 27 15:17:55 T0mW: actually, no Sep 27 15:18:13 by using tasks I can have a user install extra stuff when he updates Sep 27 15:18:51 T0mW: MACHINE_FEATURES = "kernel26 alsa pcmcia keyboard touchscreen irda usbhost usbgadget" + DISTRO_FEATURES = "" gives very small rootfs Sep 27 15:19:56 cu guys Sep 27 15:20:05 hrw|work, have you updated to defcong for the osmap? Sep 27 15:20:16 Crofton: not touched it Sep 27 15:20:19 ok Sep 27 15:20:23 I'll look now then Sep 27 15:20:27 Crofton: I do not have experience with board at all Sep 27 15:20:28 cu Sep 27 15:20:31 ye Sep 27 15:20:32 bye Sep 27 15:22:30 koen: hmmm, for a beginner like me, the double-ended confusion of creating / customizing a machine.conf, distro.conf and local.conf, is compounded by complexities at the other end by choosing / creating a task-.bb and -image.bb. Shouldn't there be a 'task-clad_in_stone.bb' and 'clad_in_stone-image.bb' ? Sep 27 15:23:47 koen: then those far-end task/image could be affected by choices in machine/distro/local.conf Sep 27 15:24:16 I see no reason why limiting the flexibility would be good Sep 27 15:25:10 koen: yeah, true. I thought of that after I typed it. Maybe move those out of the package path into the conf dirs so that it is more obvious what they are there to do. Sep 27 15:25:32 I guess if I want to change my rootfs some I should be able to add do_rootfs_prepend() to my main metafile and put in all my changes to that function? Sep 27 15:25:41 koen: just thinking out loud here. I like what you guys have put together, don't misunderstand. Sep 27 15:25:54 OE is already 'too easy' in a way that people just stop using common sense trying to build something Sep 27 15:25:58 or I guess it would really be do_rootfs_append Sep 27 15:27:37 and the current heavy users of OE have been using it for some years, so it's hard to find a 'fresh look' at OE Sep 27 15:27:46 * koen -> food Sep 27 15:29:10 koen: one other thing that I found to be a near show stopper for my using OE in the beginning. That of the local.conf files. For me, it was very confusing as to what combination of machine + distro confs to choose that "worked" together. Something representative that I could build and then customize to meet my needs. Those local.confs seem to be only in the hands of the developers of the particular machines. Sep 27 15:29:57 Sep 27 15:32:28 machine and distro should be orthogonal Sep 27 15:32:38 that's the idea of OE Sep 27 15:32:42 nod Sep 27 15:33:32 koen|away: but, for example, there is a machine called h3600 but no distro called h3600 Sep 27 15:34:42 that is the point Sep 27 15:34:56 I should use amgstrom for the omap5912osk machine Sep 27 15:35:12 but instead I use openomap so I can fiddle with it Sep 27 15:35:25 but I intend to move to a "standard" distro one day Sep 27 15:36:02 now if I were to get an omap machine, I would not intuitively know that angstrom would be the distro to use on it. Sep 27 15:36:55 distro and machine are orthogonal Sep 27 15:37:05 I should be able to any distro on any machine Sep 27 15:37:14 true Sep 27 15:37:17 distro is tool version policy Sep 27 15:37:27 and available packages Sep 27 15:37:36 of course it isn't that simple :) Sep 27 15:38:06 I was attempting to build for h1940 but seemed to never get a build to finish out except for bootstrap-image Sep 27 15:39:52 the Zipit was a test case to learn OE by doing something useful. My real project is an S3C2410 based board similar to the h1940 handheld. I would like to use gpe on it, but that won't fully build. Sep 27 15:41:43 i should do a build of the gpe dev stuff and check what goes wrong Sep 27 15:41:52 what doesn't build (not that I know anything about gpe ...) Sep 27 15:42:07 can you use angstrom on the machine? Sep 27 15:42:40 T0mW: booted my samsung board today :-) Sep 27 15:42:46 Crofton: heh, I don't recall that was about two weeks ago that I was messing with that. The last two weeks are a blur. heh Sep 27 15:42:56 XorA: ah, cool! Sep 27 15:43:14 XorA: nice to log into it, isn't it? Sep 27 15:43:30 T0mW: well I have troubles with rootfs, but that'll be fixed tomorrow Sep 27 15:43:50 florian_kc: that'd help. gpe is in really bogus state in .dev :/ Sep 27 15:44:45 alot of OE work consists of patiently waiting for the perosn who can solve your problem to have time to solve it .... Sep 27 15:44:59 The OE gods help those who help themselves Sep 27 15:45:02 and being able to explain very clearly what the problem is ... Sep 27 15:45:12 CosmicPenguin, part II of my answer Sep 27 15:45:13 I've got one more issue to resolve with the zipit, then I can put it to bed for a while. There is some framebuffer funnies I have to track down. Something is not properly wrapping lines when you cursor back over the line you typed. Sep 27 15:45:26 CosmicPenguin: heh Sep 27 15:45:39 hrw did give me the specific action I needed to do to make the omap1 git kernel build again, I hope Sep 27 15:46:20 CosmicPenguin: "patches gratefully accepted" Sep 27 15:49:33 re Sep 27 15:54:14 hrw, compiling with new defconfig now Sep 27 15:54:20 ok Sep 27 15:54:57 hmmm no vmlinux ... Sep 27 15:55:31 hmmm Sep 27 15:55:37 it moved to ./git now Sep 27 15:55:57 yes Sep 27 15:56:06 wb koen Sep 27 15:56:38 also getting some kind of error in modules install ... Sep 27 15:57:40 mickeyl: now that we've finally are moving away from obsolete stuff (dbus 0.2x) gpe is in a pretty good state Sep 27 15:57:51 mickeyl: DISTRO=angstrom-2007.1 does wonders :) Sep 27 15:58:09 Crofton: you NEED to set KERNEL_RELEASE = "2.6.18-omap1" to get correct modules dependencies Sep 27 15:58:25 in kernell bb file? Sep 27 15:58:31 yes Sep 27 15:58:36 ok Sep 27 15:58:48 I am going to mod the file to take vmlinux from ./git Sep 27 15:58:56 and add KERNEL_RELEASE Sep 27 15:59:07 Crofton: ok. if you can also use tag for git... Sep 27 15:59:11 hrw, thanks for the help btw Sep 27 15:59:26 when there is a tag that is good :) Sep 27 15:59:41 I'll look once we get these issues sorted out Sep 27 15:59:44 Crofton: no problem - I will do some work on omap stuff Sep 27 16:00:37 it is handy having someone who knows the OE ins and outs paying attetnion Sep 27 16:02:43 2;) Sep 27 16:05:09 ~change 995 pln to eur Sep 27 16:05:15 995.00 Zloty (PLN) makes 250.613 Euro (EUR) (from http://www.xe.com/) Sep 27 16:05:59 hrw, the last tag is 7 days old and the dspgw updates went in 6 days ago Sep 27 16:06:39 we could try a build based on the 2.6.18 tag Sep 27 16:06:54 or see if the patch is available and a gitless version Sep 27 16:07:02 but I really care about the dspgw Sep 27 16:07:09 Crofton: currently I just need it to build as I do not have such board Sep 27 16:07:21 :) Sep 27 16:07:29 Crofton: you are omap maintainer so it is up to you how to solve it. Sep 27 16:07:52 I hope there is a usable tag made soon :) Sep 27 16:08:16 the other guy I have talked to is also interested in the dsp ge Sep 27 16:08:20 er gw Sep 27 16:08:39 isn't dspfs going to obsolete dspgw? Sep 27 16:08:47 or am I reading the mails wrong? Sep 27 16:08:52 I am not sure Sep 27 16:09:07 if it obsoletes or compliments Sep 27 16:09:34 there are still updates going in for dspgw Sep 27 16:11:26 RP: any plans to improve 'addtask'? Sep 27 16:12:02 hrw: improve it to solve which problem? Sep 27 16:12:03 RP: addtask rm_work right_before do_build Sep 27 16:12:30 now it can be compile/stage/install/rm_work/package/deploy/build Sep 27 16:12:48 or compile/stage/install/package/rm_work/deploy/build Sep 27 16:12:56 hrw: What if two people call right_before do_build ? Sep 27 16:13:38 RP: priorities? task interdependencies? Sep 27 16:14:24 hrw: priorities are a road I'd prefer not to go down... Sep 27 16:14:31 addtask do_deploy after do_package depends on do_package; addtask do_rm_work after do_package before do_build; == package/deploy/rm_work never mind which order they were defined Sep 27 16:15:18 rm_work does not depend on other tasks so it is run as last one Sep 27 16:16:07 addtask do_rm_work after do_package,do_deploy before do_build is what we need? Sep 27 16:16:20 yes Sep 27 16:17:03 hrw: We should be able to support that... Sep 27 16:17:11 now? Sep 27 16:18:03 hrw: Maybe not. I need to look at the code Sep 27 16:18:15 ok Sep 27 16:18:17 thx Sep 27 16:24:37 looks like I will have something to add into oe doc.. Sep 27 16:24:41 Creating a new Distribution Sep 27 16:24:41 This section is a stub, help us by expanding it Sep 27 16:24:50 :) Sep 27 16:25:02 I already have one Sep 27 16:25:07 good evening Sep 27 16:25:12 hey pH5 Sep 27 16:25:13 hi pH5 Sep 27 16:26:29 tetex installing.. Sep 27 16:29:04 thats remind me old times when I was working on my thesis.. Sep 27 16:29:38 does docbook need tetex? Sep 27 16:29:56 docbook-utils need Sep 27 16:33:27 why does do_install look for vmlinux in the wrong place? Sep 27 16:34:15 Crofton: KERNEL_OUTPUT var? Sep 27 16:35:59 KERNEL_OUTPUT = "${KERNEL_IMAGETYPE}" Sep 27 16:36:33 I see KERNEL_OUTPUT = "arch/${ARCH}/boot/compressed/${KERNEL_IMAGETYPE}" Sep 27 16:36:58 I change that Sep 27 16:50:14 how do i compile org.openembedded.oz354x for a collie without the 2.6 kernel? Sep 27 16:51:55 set MACHINE_KERNEL_VERSION = "2.4" and use oz 3.5.4.1 Sep 27 16:53:03 hmm.. I think that we have few things too much in distros... Sep 27 16:53:13 22 # Glibc provides iconv and intl -- adjust, when doing uclibc builds Sep 27 16:53:13 23 PREFERRED_PROVIDER_virtual/libiconv = "glibc" Sep 27 16:53:13 24 PREFERRED_PROVIDER_virtual/libintl = "glibc" Sep 27 16:53:20 can it be set in glibc? Sep 27 16:54:34 hrw, are we sure arch/arm/boot/vmlinux is not hard coded in something? Sep 27 16:54:57 not I Sep 27 16:56:57 KERNEL OUTPUT is set in kernel.bbclass Sep 27 16:57:28 How exactly do override that in the bb file? Sep 27 16:57:32 Crofton: then set it after 'inherit kernel' Sep 27 16:57:38 thanks! Sep 27 17:00:11 np Sep 27 17:01:39 hrw: thanks, that seems to be what is going on :) Sep 27 17:02:39 simcop2387: no problem - I checked that few days ago Sep 27 17:03:05 Crofton: there is no 2.6.18-omap yet Sep 27 17:03:14 ok Sep 27 17:03:18 muru.com? Sep 27 17:03:40 http://www.muru.com/linux/omap/ Sep 27 17:03:42 nope Sep 27 17:10:17 :0 Sep 27 17:10:26 oops. this isn't vi Sep 27 17:15:19 hrw, looks like it ran to completion Sep 27 17:15:37 Crofton: great - push it Sep 27 17:16:46 hey zecke Sep 27 17:16:51 ho Sep 27 17:20:13 hmm.. someone remember why we use module-init-tools instead of busybox? Sep 27 17:20:56 because the busybox ones didn't work with 2.6 Sep 27 17:21:19 ah. Sep 27 17:22:24 03crofton 07org.oe.dev * r629adcaf... 10/ (4 files in 4 dirs): Sep 27 17:22:24 omap5912osk.conf : prepare for task-base stuff Sep 27 17:22:24 defconfig : Update defconfig to follow git and enable dsp gateway Sep 27 17:22:24 linux-omap1_2.6.x+git.bb : Fix problems in build process Sep 27 17:22:48 Crofton: split that kind of stuff in future Sep 27 17:22:58 hmm Sep 27 17:23:02 which stuff Sep 27 17:23:10 machine.conf and kernel.bb Sep 27 17:23:11 the machine file was an accident :) Sep 27 17:23:15 I know Sep 27 17:23:29 I jsut bitches a guy out over a huge diff in our svn :) Sep 27 17:23:36 hrw: ah cool Sep 27 17:23:44 well, not really bitched, but we can all see the problem Sep 27 17:24:03 the machine change is commented out also Sep 27 17:24:15 Crofton: I will merge and look - I have made some changes in machine too Sep 27 17:24:29 ok Sep 27 17:24:51 I really should have reverted my machine file ..... Sep 27 17:25:00 * Crofton kicks Crofton Sep 27 17:26:11 Crofton: then do revert of changes if you need Sep 27 17:27:22 not needed, feel free to update the machine with your worl Sep 27 17:27:24 work Sep 27 17:28:12 hrw: did the commentspam protector on oz.org go bad? Sep 27 17:28:25 koen: ? Sep 27 17:28:37 koen: root@home:/tmp/busybox-1.2.1# ./busybox insmod /lib/modules/2.6.17-beyond3/kernel/lib/zlib_inflate/zlib_inflate.ko Sep 27 17:28:40 works Sep 27 17:28:58 hrw: http://openzaurus.org/wordpress/2006/03/18/openzaurus-354-released/#comment-4703 Sep 27 17:29:00 have to go Sep 27 17:29:05 hrw: that's good news about busybox Sep 27 17:29:11 0.6x didn't work Sep 27 17:29:20 fsck Sep 27 17:29:28 hrw, do you plan to commit osk machine file stuff soon? Sep 27 17:30:47 koen: removed Sep 27 17:31:13 Crofton: now I have to go - will review and push in 2h or tomorrow (20:00 here) Sep 27 17:31:48 ok Sep 27 17:31:59 cu Sep 27 17:32:00 I want to rebuild bootstrap-image soon Sep 27 17:32:02 I will wait Sep 27 17:32:10 * Crofton has planty to do .... Sep 27 17:32:19 Crofton: I also have change to openomap distro for you ;) Sep 27 17:32:50 03kristoffer 07org.oe.dev * rf1eebc81... 10/ (1 packages/tuxnes packages/tuxnes/tuxnes_0.75.bb): Sep 27 17:32:50 packages/tuxnes : Addition of skeleton tuxnes.bb Sep 27 17:32:50 * Addition of Tuxnes (Nintendo 8bit Emulator) into OE. Sep 27 17:32:50 * Needs fixing to compile. Sep 27 17:52:21 re Sep 27 17:56:46 Anyone else having bad issues with perl builds lately? Sep 27 18:00:35 Kristoffer: check the mailinglist Sep 27 18:05:40 koen, found it. Im experiencing but #1439 Sep 27 18:05:43 bug Sep 27 18:11:01 v8jlene, updated bug 1439 for you. Im having same issues Sep 27 18:41:54 I'm trying to compile generic-unstable distro but it keeps having errors when compiling glibc-intermediate-2.4-r10: Log: http://pastebin.ca/184273 Has anyone compiled it recently? I'm using an Xscale processor (PXA270). Sep 27 18:46:12 Well, who added 4000th .bb? ;-) Sep 27 18:46:14 re .. how i can make a "make clean" like with bitbak ? Sep 27 18:46:28 gumz, -c clean Sep 27 18:46:36 tx Sep 27 18:47:09 "Nothing to do. Use 'bitbake world' to build everything, or run 'bitbake --help'" Sep 27 18:47:35 gumz, bitbake -c clean, of course Sep 27 18:47:52 of course :) Sep 27 18:48:01 Bitbake complains about "no free bitbake after 2005 something". This is with an old oe enviroment I set up somewhere around 2005 for bulding openzaurus apps. Is there a free version today? Sep 27 18:48:19 psokolovsky : work, thanks Sep 27 18:48:41 TuomasT, of course there's Sep 27 18:50:38 TuomasT: not 'bitbake', 'bitkeeper' Sep 27 18:50:55 lol Sep 27 18:52:22 koen|away: just noticed: Sep 27 18:52:23 OE_REVISION = "format_version "1" Sep 27 18:52:23 new_manifest [0000000000000000000000000000000000000004] Sep 27 18:52:23 old_revision [629adcaf78ed27ef81715e0e1b0ba32f4fb9ee20]" Sep 27 18:52:35 OE Build Configuration: Sep 27 18:52:36 BB_VERSION = "1.6.1" Sep 27 18:52:36 OE_REVISION = "format_version "1" Sep 27 18:52:36 new_manifest [0000000000000000000000000000000000000004] Sep 27 18:52:36 old_revision [629adcaf78ed27ef81715e0e1b0ba32f4fb9ee20]" Sep 27 18:52:36 TARGET_ARCH = "arm" Sep 27 18:52:38 TARGET_OS = "linux-gnueabi" Sep 27 18:52:44 yeah Sep 27 18:53:03 OE needs to use mtn automate get_base_revision or something similar Sep 27 18:53:18 I use bbcollections, so I just get Sep 27 18:53:22 TuomasT: see http://www.openembedded.org/wiki/GettingStarted Sep 27 18:55:14 'mtn automate get_base_revision_id' to be exact Sep 27 18:58:08 NOTE: package glibc-2.3.5+cvs20050627-r12: task do_configure: failed : > checking for kernel header at least 2.6.15... too old! Sep 27 18:58:36 i have removed the 2.6.15 kernel header in conf file ... but he always want kernel header 2.6.15 Sep 27 19:01:12 is there some class to modify bb files ? Sep 27 19:02:43 maybe the 'vi' or 'emacs' classes Sep 27 19:07:05 arf Sep 27 19:08:32 i'd like to parse it to make a generator of bb file Sep 27 19:09:49 That sort of precludes the whole "reproducablity" thing Sep 27 19:12:58 first goal will be to add/remove some dependancies of a images bb file Sep 27 19:13:17 in function of what my software need as dependancies Sep 27 19:28:03 does glibc and glibc-intermediate need to be the same version? Sep 27 19:30:26 I just looked in the bb for glibc-intermediate and it requires the glibc of the same version so I guess that answers my question :-) Sep 27 19:37:28 can anyone explain to me the difference between glibc and glibc-intermediate? is the intermediate the one it builds first to be able to build the tools to finally cross-compile the final glibc? Sep 27 19:56:45 almost Sep 27 19:57:12 the intermediate glibc is used to compile the final toolchain Sep 27 19:57:19 including the final glibc Sep 27 19:58:13 blaster8, ok that is what I thought.. do you know of a distro that glibc_2.4 compiles for? Sep 27 20:09:37 Gerrath: angstrom Sep 27 20:15:43 hrw, thanks.. I will give it a try. Sep 27 20:16:05 hrw, generic-unstable does NOT work. Sep 27 20:18:59 Crofton: two changesets for you Sep 27 20:19:36 Gerrath: ;( cant test this Sep 27 20:19:45 03hrw 07org.oe.dev * rd7e328fb... 10/ (1 conf/machine/omap5912osk.conf): omap5912osk.conf: transformed into task-base format (old config commented as part of it will probably need to be added into MACHINE_EXTRA_RRECOMMENDS) Sep 27 20:19:51 03hrw 07org.oe.dev * rb95778d4... 10/ (1 conf/distro/openomap.conf): openomap distro: switch to task-base format Sep 27 20:20:02 thanks hrw Sep 27 20:21:15 Crofton: do testing - some modules maybe will need adding Sep 27 20:21:39 Crofton: we also need to fix^Wadd some stuff into task-base probably Sep 27 20:23:09 I will rebuild bootstrap-image and look it over tonight Sep 27 20:23:46 ok Sep 27 20:24:17 does the new mtn use less memory? Sep 27 20:24:40 for commit and log, yes Sep 27 20:24:47 update? Sep 27 20:24:52 yes Sep 27 20:25:00 anything that searches for a 'head' Sep 27 20:25:05 time to upgrade or buy more RAM for this machine Sep 27 20:25:09 I have only 512M here now Sep 27 20:25:11 I am killing it Sep 27 20:25:19 swap? Sep 27 20:25:23 same here, but I run our groups svn server on it Sep 27 20:25:30 second 512M is on desk because I want to test how it will work with one Sep 27 20:25:33 and they bitch when things get slow Sep 27 20:26:01 Crofton: no possibilities to get^Wuse other machine? Sep 27 20:26:01 when I am done with paper, I will set up home machine to do builds Sep 27 20:26:12 only home machine Sep 27 20:26:37 I have been very time limited the past few weeks Sep 27 20:27:32 hrw: will they ship a board to you to test on? Sep 27 20:28:20 hrw, alot of the old modules may b built in with new config Sep 27 20:28:27 I was module happy Sep 27 20:28:59 but it makes no sense unless the thing is remvoeable or rarely used Sep 27 20:29:10 of course I don't have a display on one of my boards Sep 27 20:29:14 and I don't really use it Sep 27 20:29:31 koen: what for? for me it does not make any differ if I have machine on desk or have it remotely Sep 27 20:29:45 ah, right Sep 27 20:29:52 Crofton: 2.6.12-rc5 has more modules Sep 27 20:29:53 forgot about remove access :) Sep 27 20:30:15 remote* Sep 27 20:30:21 the OSK is $295 usd Sep 27 20:30:28 the LCD is like $800 usd Sep 27 20:30:37 ~change 295 usd to pln Sep 27 20:30:42 295.00 United States Dollar (USD) makes 922.255 Zloty (PLN) (from http://www.xe.com/) Sep 27 20:30:46 the imx31ads is $5k Sep 27 20:31:10 955 PLN is sempron64 + mainboard + 2x512 ddr2 Sep 27 20:31:31 much better toy then omap5912osk I think Sep 27 20:32:13 all depends on what you want to do with the ty Sep 27 20:32:57 Crofton: some random builds from time to time... for example omap1510/generic Sep 27 20:33:20 or x86/openzaurus or wrt54/openomap Sep 27 20:33:31 or c7x0/celinux-test Sep 27 20:33:37 or other combo Sep 27 20:34:13 thank god bitbake no longer takes all the RAM in a machine ... Sep 27 20:34:19 or 'lets fetch world and look how much we miss' Sep 27 20:34:34 Crofton: with 4k recipes it would take 2-4G... Sep 27 20:34:38 brb Sep 27 20:36:46 the toolchain don't compile for me ... he want "checking for kernel header at least 2.6.15... too old!" and i see that : NOTE: fetch http://ep09.pld-linux.org/~mmazur/linux-libc-headers/linux-libc-headers-2.6.11.1.tar.bz2 ... what i can do ???? Sep 27 20:37:37 specify an other version for linux-libc-headers Sep 27 20:37:50 where i specify that ? Sep 27 20:40:00 a+ Sep 27 20:50:48 cu Sep 27 20:52:00 :'( Sep 27 21:12:31 bother glibc-intermediate-2.4 build just blew up on me ... Sep 27 21:13:04 hrw, thoughts? Sep 27 21:16:52 Crofton: we are seeing glibc-2.4 failure at wolfson Sep 27 21:17:37 this is the glibc-intermediate, I suppose it ould be related Sep 27 21:18:06 undefined references to __atfct_seterrno type problems? Sep 27 21:18:30 Crofton: sounds familiar, it was lrg's machine that failed. Sep 27 21:19:09 Crofton: is it related to the openat and chownat patches that were added a couple of days ago? Sep 27 21:19:51 not sure Sep 27 21:20:22 I was hoping it would build so I could check what hrw did with the omap stuff :) Sep 27 21:20:54 Crofton: Id try removing those two patches, if that works could you email oe@ Sep 27 21:21:15 grr, I am supposed to working on a paper ... Sep 27 21:23:48 time for me to sleep Sep 27 21:23:52 gn Sep 27 21:24:03 good luck Crofton Sep 27 21:25:49 trying your suggestion now Sep 27 21:35:44 RP: btw, my sugar rootfs is 63M Sep 27 21:37:42 * v8jlene is back Sep 27 21:39:35 Crofton: in unix/sysv/linux/openat.c, comment out the top redundant define block around #define OPENAT - seems to work for me. Sep 27 21:39:46 Don't know enough about OE to suggest a clean way of doing it. Sep 27 21:40:07 Need to do it for both glibc-intermediate and glibc... Sep 27 21:44:52 theuns: clean way would be rediff the patch and submit it to bugtracker Sep 27 21:47:10 removing those two patches solved the problem Sep 27 21:47:18 the patch is in glibc Sep 27 21:47:31 -int... includes the glibc bb file Sep 27 21:47:48 Crofton: sorted, that is good to know Sep 27 21:47:58 Crofton: can fix the build at work tomorrow then Sep 27 21:49:09 hopefully someone sorts the patch out Sep 27 21:49:30 any idea who to bother? Sep 27 21:50:37 Crofton: mickey|bbl submitted I think Sep 27 21:50:50 Crofton: you can always file a bug at bugs.oe.org Sep 27 21:50:51 ok Sep 27 21:50:56 Crofton: Are the problems caused by the fix-fchownat-20060808.patch that mickey submited two days ago? Sep 27 21:51:04 I suppose :) Sep 27 21:51:24 anyway I really shall go sleep now, bye Sep 27 21:55:14 that one or openat-bugzilla Sep 27 21:55:19 I commented both out Sep 27 22:25:14 03blaster8 07org.oe.dev * r90eea54e... 10/ (1 conf/distro/debianslug.conf conf/distro/openslug.conf): Openslug & Debianslug: Add missing sd_mod module Sep 27 22:25:21 03blaster8 07org.oe.dev * r0b0e5efd... 10/ (3 files in 3 dirs): slugos-init: update loadmodules.sh script to autoload sd_mod Sep 27 22:25:26 03blaster8 07org.oe.dev * rdc9bff2d... 10/ (4 files in 3 dirs): Sep 27 22:25:26 slugos-init: first stages of fixing disk boot Sep 27 22:25:26 * Add module loading in disk and ram bootdisks Sep 27 22:25:26 * Should eventually be consolidated - still planning this Sep 27 23:42:12 03lenehan 07org.oe.dev * r47723948... 10/ (5 files in 2 dirs): perl: Add the perl logging modules Log::Dispatch and Log::Log4perl. Sep 27 23:55:45 03lenehan 07org.oe.dev * rd679a6d2... 10/ (3 files in 2 dirs): perl: Add the perl IP address manipulation module Net::IP. Sep 28 00:08:55 * v8jlene is away (cycling) Sep 28 02:44:56 morning ~ **** ENDING LOGGING AT Thu Sep 28 02:59:56 2006