**** BEGIN LOGGING AT Wed Aug 27 02:59:56 2008 Aug 27 07:16:56 morning Aug 27 07:32:02 bonjour Aug 27 07:40:16 morning Aug 27 08:24:16 hi ant Aug 27 08:27:23 morning Aug 27 08:28:47 * ant_work is trying to open an .xls written in chinese...no luck with openoffice...what's wrong ? Aug 27 08:31:08 dont know Aug 27 08:31:12 which version? Aug 27 08:31:14 of xls Aug 27 08:36:19 hi all, how do I get osb-jscore to build from dev? Aug 27 08:36:32 (uclibc build) Aug 27 08:37:00 _or_ is it possible to build full libc for h2200? Aug 27 08:37:42 mzb_d800 hm I think there speaks nothing against a full glibc build Aug 27 08:37:49 just tweak your local.conf Aug 27 08:37:56 what distro you are using? Aug 27 08:38:27 angstrom-2008.1 Aug 27 08:38:39 okay Aug 27 08:38:53 than just remove uclibc line Aug 27 08:38:57 or comment out Aug 27 08:39:01 I ended up changing to TARGET_OS = "linux-uclibc" because it didn't want to build Aug 27 08:39:18 can't recall the problem (so long ago!;) Aug 27 08:39:24 no you shoudnt set TARGET_OS your self Aug 27 08:39:52 ok .. I'll rebuild Aug 27 08:40:52 I also ended up setting: ANGSTROM_MODE="uclibc" Aug 27 08:40:57 I'll remove that too Aug 27 08:41:12 yes Aug 27 08:41:31 and than have 5 hour freetime Aug 27 08:41:33 *g* Aug 27 08:42:03 dm, I remember what it's like :) ... also not in a hurry Aug 27 08:42:28 certainly a lot faster with my new server ... X2 3800+ on raid10-f2 Aug 27 08:42:41 so interesting to watch remotely with gkrellm ;) Aug 27 08:42:48 *g* Aug 27 08:43:30 sort of thing I can set off and come back to when the mood strikes me ;) Aug 27 08:53:07 moin diego Aug 27 08:53:13 minipanda: ping Aug 27 08:54:25 ant_work: oops, are you ping a wrong guy? Aug 27 08:54:41 I hope not... Aug 27 08:54:59 could you tell me which language it is? http://jckspaqj.aqsiq.gov.cn/zwgk/ Aug 27 08:56:00 woglinde: this is the error that encouraged me to try uclibc: http://fastpaste.net/144/ Aug 27 08:57:28 mzb hm Aug 27 08:57:38 let me take a look into the source Aug 27 08:58:00 minipanda: I mean, chinese-trad, chinese-simp... babelfish seems having hard time with that page... Aug 27 08:58:11 ref: http://handhelds.org/hypermail/htc-pxa/current/0181.html Aug 27 08:58:52 ant_work: ah, it's chinese-simp Aug 27 09:00:05 current local.conf: http://fastpaste.net/145/# Aug 27 09:00:10 ah, thx, seems better now Aug 27 09:02:50 hm this is from ansgtroem.conf and looks wrong to me KERNEL_INITRAMFS_DIR = "${TMPDIR}/deploy/uclibc/images/${MACHINE}/" Aug 27 09:03:21 I'll further add, just to complicate things;), that the build is being performed on an amd64 Xen domU running Debian Etch Aug 27 09:03:27 ah Aug 27 09:03:31 hmm Aug 27 09:03:49 could you please do bitbake -i Aug 27 09:04:11 done (doing;) Aug 27 09:04:22 what would you like? Aug 27 09:04:32 mom Aug 27 09:04:36 have to search it Aug 27 09:04:59 getvar ANGSTROM_MODE Aug 27 09:05:15 BB>> getvar ANGSTROM_MODE Aug 27 09:05:16 glibc Aug 27 09:05:35 okay Aug 27 09:05:39 BB>> getvar KERNEL_INITRAMFS_DIR Aug 27 09:05:39 /usr/src/oe/h2200/build/tmp/deploy/uclibc/images/h2200/ Aug 27 09:06:06 ant_work: ah, BTW: what kind of info you're looking for? it's mostly not aboue OE :) Aug 27 09:06:12 yes Aug 27 09:06:19 :-) Aug 27 09:06:21 mzb hm seems the kernel needs a initramfs Aug 27 09:06:31 lol Aug 27 09:06:45 minipanda: I'm trying to download the list of the 'authorized importers' Aug 27 09:06:47 yay! ... made that suggestion years ago for the h2200 ... but that's another story ;) Aug 27 09:06:58 I found some xls but these don't open here.... Aug 27 09:07:05 so the question is how you get it Aug 27 09:07:17 problem is ... why doesn't it work with $ bitbake gpe-image ? Aug 27 09:07:55 actually ... historically the h2200 never used an initramfs because LAB wasn't capable Aug 27 09:07:59 (iirc) Aug 27 09:08:19 hm Aug 27 09:08:20 okay Aug 27 09:08:32 I would suggest as quick solution Aug 27 09:08:36 ant_work: sounds interesting, aqsiq.gov.cn just quite close to my current office Aug 27 09:08:41 edit Aug 27 09:08:45 linux-handhelds-2.6.inc Aug 27 09:08:59 and comment out the initramfs stuff Aug 27 09:09:18 minipanda: I can't believe it! Aug 27 09:09:30 linux-handhelds-2.6.inc is in packages/linux/ Aug 27 09:10:05 ant_work: well, i just live in Beijing, the whole city just hold Olympic game Aug 27 09:10:13 yeah, I've got it ... which section? Aug 27 09:10:47 the IF statement starting with: if [ -n "${KERNEL_INITRAMFS_PATH}" -a "${ANGSTROM_MODE}" = "glibc" ]; then Aug 27 09:10:48 ?? Aug 27 09:11:31 ant_work: and could you get my message in query window? i think it's off OE topic here, then we can talk privately Aug 27 09:11:42 mzb yes Aug 27 09:11:45 and in deploy_append? Aug 27 09:11:49 if [ -n "${KERNEL_INITRAMFS_PATH}" -a "${ANGSTROM_MODE}" == "glibc" ]; then Aug 27 09:12:46 yes comment this too Aug 27 09:13:29 ok, thx ... we'll see how it goes Aug 27 09:15:00 hmm ... zipped through that ... we'll see how the rest goes Aug 27 09:15:27 I'll have to see if I can remember how to do|create LAB Aug 27 09:21:44 either way, I assume that means dev is broken for h2200? Aug 27 09:24:16 mzb hm seems so Aug 27 09:24:29 because I dont have h2200 I cant test something Aug 27 09:27:20 yep, makes sense Aug 27 09:27:52 h2200 is weird because of LAB, I guess Aug 27 09:28:06 hm what is LAB? Aug 27 09:28:13 Linux As Bootloader Aug 27 09:28:16 doenst it work with haret? Aug 27 09:28:20 ah hm Aug 27 09:28:33 iirc, Joshua Wise wrote it ... really nifty Aug 27 09:29:12 not interested in haret ... decided I finally want to do useful things with my original h2200 ... testing is over ;) Aug 27 09:29:23 time to flash it's worries away! Aug 27 09:30:22 I did have a 2nd one ... headless ... but it committed suicide by jumping out of it's cradle (battery exploded after a year of the hh being used as an AP) Aug 27 09:30:59 ^^ that one was flashed, and running a custom kernel with modular pcmcia+ide_cs (to run Sandisk Connect+) Aug 27 09:31:14 h2200#1 is still a virgin ;) Aug 27 09:31:43 battery exlpode Aug 27 09:31:44 nice Aug 27 09:33:03 well ... "expanded violently" probably more correct ;) Aug 27 09:33:17 haven't checked it since then ... might even still work Aug 27 09:36:31 prob. broke the clip on the battery compartment, but I might have another one floating around Aug 27 09:37:20 a headless ipaq is probably no longer as useful as it was then Aug 27 09:58:57 hi florian Aug 27 10:00:31 lunch time Aug 27 10:03:06 good morning Aug 27 10:16:21 florian: good morning Aug 27 10:20:58 hi Aug 27 10:21:31 having trouble while buildint gpsdrive on dev branch, with dependency libwww-perl Aug 27 10:21:35 make: *** No rule to make target `/usr/lib/perl/5.8/Config.pm', needed by `Makefile'. Stop. Aug 27 10:21:41 do_compile failer :s Aug 27 10:21:44 failed Aug 27 10:22:12 boris_Omega: don't worry...it will fail later on mysql ! Aug 27 10:22:33 (buildall) Aug 27 10:23:21 http://bugs.openembedded.net/show_bug.cgi?id=4432 Aug 27 10:28:36 xD Aug 27 10:29:52 i was thinking that mysql had a good issu ;.. Aug 27 10:30:14 thank you ant_work.so i cannot hope to have gpsdrive on angstrom ? Aug 27 10:32:39 you have hopes ! Aug 27 10:33:08 the problem is the broken perl status...I thought XorA did already some fix Aug 27 10:33:26 the big bug is http://bugs.openembedded.net/show_bug.cgi?id=3403 Aug 27 10:34:36 IIRC some perl modules have been fixed... Aug 27 10:35:58 XorA: can you add some comment ^^? Aug 27 10:37:14 re Aug 27 10:37:21 boris_OmegA: the fix for liburi-perl: http://gitweb.openembedded.net/?p=org.openembedded.dev.git&a=search&h=HEAD&st=commit&s=liburi-perl Aug 27 10:37:29 wb hrw Aug 27 10:39:18 boris_OmegA: ah, it was ordon...here the fix for libwww-perl: http://gitweb.openembedded.net/?p=org.openembedded.dev.git&a=search&h=HEAD&st=commit&s=libwww-perl Aug 27 10:39:33 s/ordon/Gordon/ Aug 27 10:40:36 ant_work: I checked in a generic fix for all perl modules into CPAN.bbclass Aug 27 10:40:51 ah, ok, thx Aug 27 10:40:53 * ant_work thinks a search in gitweb is always worthwhile Aug 27 10:41:49 thx ant_work !! Aug 27 10:42:18 but the commits are 20 days old...didn't you pull lately ? Aug 27 10:42:28 ;-) Aug 27 10:52:48 doesn't work ... Aug 27 10:52:54 but i will try to fix it later Aug 27 10:53:09 ++ good apetite ;) Aug 27 10:56:32 re Aug 27 11:31:09 he zecke Aug 27 11:52:38 woglinde: I was comparing the instructions for setting PATH and BBPATH given Aug 27 11:52:40 here http://wiki.openembedded.net/index.php/Getting_Started Aug 27 11:52:47 and here http://www.angstrom-distribution.org/building-angstrom Aug 27 11:53:17 I have some conceirn about BBPATH=/OE/:... Aug 27 11:55:49 hm Aug 27 11:55:55 never looked at it Aug 27 11:56:02 better discuss this with laibsch Aug 27 11:56:12 something is getting wrong in Angstrom, and I fear the problem could be common Aug 27 11:56:16 http://lists.linuxtogo.org/pipermail/angstrom-distro-users/2008-August/002101.html Aug 27 12:00:06 ant this is only an example how to setup local.conf Aug 27 12:00:27 yes, but moving things out of /build seems not trivial... Aug 27 12:01:18 (QA issues - workdir in path) Aug 27 12:01:27 hi ant Aug 27 12:01:28 hi woglinde Aug 27 12:01:30 hi zecke Aug 27 12:01:41 re pb Aug 27 12:01:47 hello pb___ Aug 27 12:02:05 hm, seems I have a bit too much tail today Aug 27 12:02:22 that's better Aug 27 12:03:24 fish tail Aug 27 12:03:26 *g* Aug 27 12:04:48 mm, I guess that would be more like "pb-<" Aug 27 12:04:57 I prefer my snail tail Aug 27 12:45:55 http://www.cl.cam.ac.uk/~sps32/mcu_lock.html , very interessting Aug 27 12:58:50 ant_work: I've email you cmdline 'searcher' :) Aug 27 12:59:39 Jay7: great, thx Aug 27 13:00:37 it's not a real parser, it just search cmdline for tags specified Aug 27 13:01:00 I'm thinking that is enough for our needs Aug 27 13:07:22 Jay7: we asked for a little gun..you gave us a bazooka ! Nice code... Aug 27 13:08:23 ant_work: It's just a Kalashnikov ;) Aug 27 13:08:40 bazooka is real big parser :) Aug 27 13:08:53 ok, I'll test it later Aug 27 13:09:17 I have to send you the u-boot framework diffs, in case smthg has changed Aug 27 13:09:35 ok Aug 27 13:09:54 btw, is zaurus-2.6.inc updated in mtn? :) Aug 27 13:10:01 not yet ... Aug 27 13:10:20 I'd propose a rfc for all th neded patches ('framework') Aug 27 13:10:34 all machines.conf Aug 27 13:10:35 good! :) Aug 27 13:11:05 about kernel size..did you edit defconfig? Aug 27 13:11:25 or just gzipped the kernel ? Aug 27 13:11:33 ant_work: no, I'm not test it yet Aug 27 13:11:37 ah, ok Aug 27 13:11:58 seems I needed some manual :) Aug 27 13:11:58 remember to tell about FBANGLE Aug 27 13:12:18 well' it's tricky if PR is not bumped... Aug 27 13:12:40 1) clean kexecboot initramfs-kexecboot and linux-kexecboot Aug 27 13:12:52 2) rebuild linux-kexecboot Aug 27 13:12:57 should suffice Aug 27 13:14:17 you get a big initramfs-kernel (1,4mb) in /deploy Aug 27 13:15:52 I want emergency image too :) Aug 27 13:16:17 the first version (ramdisk) is done Aug 27 13:16:39 but I'm now finishing the initramfs version Aug 27 13:16:51 it is 2.6mb IIRC Aug 27 13:17:08 with u-boot it's not a problem :) Aug 27 13:17:30 does emergency start from kexecboot? Aug 27 13:17:35 well, the emergency image from pdaxrom can now be used...the mtdparts are respected :-) Aug 27 13:17:50 (you can use it in the meanwhile) Aug 27 13:18:20 just edited bootargs for that image in u-boot Aug 27 13:18:32 trivial change Aug 27 13:19:01 you should clean and rebuild u-boot-git Aug 27 13:19:35 in order to have last version Aug 27 13:20:07 and remember to edit ROOT_FLASH_SIZE in akita.conf as you need Aug 27 13:20:29 otherwise it defaults to standard partitioning Aug 27 13:20:43 do you making split emergency initramfs or you extend kexecboot's initramfs with some tools? Aug 27 13:21:10 here is too many 'remembers' now :) Aug 27 13:23:22 !oebug 3053 Aug 27 13:23:23 * * Bug 3053, Status: NEW, Created: 2007-09-22 11:03 Aug 27 13:23:23 * * cesarb(AT)cesarb.net: ruby-1.8.5 fails to build with dash as /bin/sh Aug 27 13:23:24 * * http://bugs.openembedded.net/show_bug.cgi?id=3053 Aug 27 13:24:43 what's the better status to say it's oki with the patch , it just needs to be commited ? Aug 27 13:24:57 keyword patch i think Aug 27 13:44:00 Jay7: eh, atm I'm splitting in two Aug 27 13:49:22 ant_work: I think we can make 'fork' of kexecboot initramfs with busybox and other things.. Aug 27 13:49:33 specially for use with u-boot Aug 27 13:49:40 Jay7: I started with initramfs-minimal Aug 27 14:17:05 * * OE Bug 4525 has been created by mad_soft(AT)inbox.ru Aug 27 14:17:07 * * New package - glista-0.2 (simple to-do manager) Aug 27 14:17:10 * * http://bugs.openembedded.net/show_bug.cgi?id=4525 Aug 27 14:51:41 lrg: is there any mechanism in asoc for DAC drivers to indicate that they need the I2S bus clocks to be present at all times? Aug 27 14:52:01 pb_: yeas, there is a gated flag Aug 27 14:52:24 ah great, thanks Aug 27 14:52:57 pb_: sorry, wrong answer Aug 27 14:53:03 oh, not so great Aug 27 14:53:05 no thanks :-} Aug 27 14:53:41 pb_: I assume you mean you need clocks when there is no playback Aug 27 14:53:49 right Aug 27 14:54:00 we have some dacs from one of your competitors which fall into disarray if the clocks are ever stopped Aug 27 14:54:24 pb_: my ex companies competitors ;) Aug 27 14:54:30 I'm frrelance now Aug 27 14:54:52 pb_: your mach driver will need to force them on at probe() Aug 27 14:55:31 oh right, when did you leave wolfson? Aug 27 14:55:39 right, yeah, that's what I feared Aug 27 14:55:40 oh well Aug 27 14:56:01 the alternative would be that if the asoc core can tell me when the clocks are going to be stopped, I could put the dac into reset and then re-initialise it again when playback recommences Aug 27 15:05:45 pb_: I take it that this is the LRCLK and BCLK and they're being stopped by the platform driver? Aug 27 15:06:18 ASoC doesn't explicitly stop the clocks itself. Aug 27 15:07:51 pb_: I left 2 weeks ago - got a good lump sum - now looking for interesting Linux work. Aug 27 15:07:54 broonie: right Aug 27 15:08:11 Hello ... when is the GIT move planned? I have a long queue of patches to send Aug 27 15:08:24 also potentially the MCLK, though in practice most platforms seem to leave that one running anyway Aug 27 15:08:33 pb_: enable in your mach driver probe() Aug 27 15:09:33 lrg: right, thanks Aug 27 15:10:05 Your machine driver may need some persuasion to let you do this. Aug 27 15:10:36 s/machine/platform/ sorry Aug 27 15:12:26 yeah, I suspected as much Aug 27 15:12:47 never mind, I will try to make one of my colleagues fix this instead of me :-} Aug 27 15:13:25 lrg: well, congratulations on your new non-job :-) Aug 27 15:14:56 otavio: not sure... maybe don't wait for it. mtn works too... welll sort of ;) Aug 27 15:15:10 florian: mm, "works" Aug 27 15:15:40 It does an awful lot of "work" according to my CPU usage figures :) Aug 27 15:15:50 heh Aug 27 15:15:54 florian: hehe ... I don't intend to learn how to use mtn and git is working well to keep it up to date. If someone wants to push them for me, no problem ;-) Aug 27 15:16:22 otavio: file them in bugzilla and someone will ign^H^H^Hdeal with them appropriately. Aug 27 15:25:36 bbl Aug 27 15:26:09 hehe Aug 27 15:26:33 hi florian Aug 27 15:27:55 fyi florian the register has a piece on bug that mentions OE: http://www.theregister.co.uk/2008/08/26/bugs_labs_second_release/ Aug 27 15:30:25 kgilmer_: hey great! Aug 27 15:31:42 kgilmer_: heh... I have to convince some people to use OE instead of LTIB tomorrow :) Aug 27 15:31:56 what is LTIB? Aug 27 15:32:00 kgilmer_, thanks! Aug 27 15:32:16 Crofton: http://savannah.nongnu.org/projects/ltib Aug 27 15:33:20 the latest news being Aug 18, 2006 should be a sign Aug 27 15:33:54 "200 userspace packages selectable" Aug 27 15:33:59 LTIB is another approach Aug 27 15:34:02 using spoofing Aug 27 15:34:13 it's from freescale Aug 27 15:34:15 These guys say somethign about OE used Scratchbox in their presentation :-/ Aug 27 15:34:54 OE can build binaries that run inside scratchbox :D Aug 27 15:35:13 via its LDFLAGS hack ... Aug 27 15:36:07 I mean LD_RUN_PATH hack Aug 27 15:38:59 hiya Aug 27 15:39:07 who manages task-base.bb ? Aug 27 15:40:17 I am trying to move to a 2.6.26 kernel, and unfortunately task-base-wifi causes madwifi-ng-modules to be built and fail Aug 27 15:40:44 what machine? Aug 27 15:40:53 mpc8313e-rdb Aug 27 15:41:08 (the .bb for linux-2.6.26 is local until I actually test it) Aug 27 15:41:43 don't select wifi in machine features? Aug 27 15:42:34 well the machine feature that triggers madwifi is "pci" Aug 27 15:42:41 jeremy_laine: use newer snapshot of madwifi? Aug 27 15:43:03 hrw|gone: does it event make any sense since ath5k seems to be moving along nicely? Aug 27 15:43:06 urg Aug 27 15:44:21 jeremy_laine: no idea? Aug 27 15:53:47 mickeyl: good morning Aug 27 15:53:52 morning pb_ Aug 27 15:54:27 join #angstrom Aug 27 15:54:32 re Aug 27 15:54:34 hi mickeyl Aug 27 15:54:41 ~curse rpm Aug 27 15:54:42 May the fleas of a thousand camels infest your most sensitive regions, rpm ! Aug 27 15:54:46 heya hrw Aug 27 15:54:48 how's life? Aug 27 15:55:12 mickeyl: *lot* of things happening Aug 27 15:55:43 rkirti|afk: thanks for the suggestion Aug 27 15:55:53 hrw: good things? bad things? :) Aug 27 15:55:53 re Aug 27 15:55:59 woglinde: wb Aug 27 15:55:59 hrw: i think i know 'bout some... Aug 27 15:56:20 he mickeyl Aug 27 15:56:30 hm seems linuxthreads is working in uclibc Aug 27 15:56:53 heh, linuxthreads is so last century Aug 27 15:57:12 pb khem will not fix nptl for oabi Aug 27 15:57:15 so for simpad Aug 27 15:57:21 I have to stay with it Aug 27 15:57:30 actual uclibc uses linuxthreads.old Aug 27 15:57:30 hehe Aug 27 15:58:01 pb_ : I am so sorry for the typo *sheepish look* Aug 27 15:58:13 woglinde: yah, I was only joking. linuxthreads is probably the right answer for oabi systems. Aug 27 15:58:56 mickeyl: both Aug 27 15:59:00 hrw: A lot mor offsprings? ;) Aug 27 15:59:26 florian: no ;) Aug 27 15:59:28 woglinde: there are very few applications that actually depend on the extra correctness that you get with nptl, and I suspect nptl will be significantly slower on most arm platforms before armv7. Aug 27 15:59:48 change of flat, job etc Aug 27 16:04:36 hm whats sratchbox-compat good for? Aug 27 16:13:03 woglinde: its allows you to build binaries that run native in a scratchbox environment Aug 27 16:13:09 like gcc Aug 27 16:13:24 hm okay Aug 27 16:13:51 thebohemian: which "scratchbox environment"? Aug 27 16:14:47 hrw: you only need it if you deal with sbox. think of it as a sideuse of OE Aug 27 16:16:05 re kgilmer Aug 27 16:16:24 hey woglinde :) Aug 27 16:16:26 thebohemian: there are many different use of sbox Aug 27 16:16:37 hrw: e.g. set distribution to chinook-compat, target to x86 and add scratchbox-compat: voila your x86 binaries will run inside nokia's sbox for maemo Aug 27 16:16:47 ah. that way Aug 27 16:17:12 florian, irt LTIB, i have some things to say about it. overall i think it's a good tool for a specific task: building a linux-base hardware application. ie something static. Aug 27 16:17:33 OE on the other hand is something else Aug 27 16:17:56 so I think it really depends on what you want to get done, etc. Aug 27 16:18:09 LTIB is well supported in the ml fyi Aug 27 16:18:56 who uses it? Aug 27 16:19:11 it ships with freescale imx products Aug 27 16:19:18 we used it up until recently Aug 27 16:19:28 why did you switch? Aug 27 16:19:29 but i don't have a list of people that use it Crofton|work Aug 27 16:19:54 because our product is a general purpose computer, not a static "appliance" or whatever. Aug 27 16:20:00 we needed a real distro Aug 27 16:20:20 there are other reasons, but this was a primary one. Aug 27 16:20:24 ok Aug 27 16:20:46 always good to know what is important to people Aug 27 16:20:48 we also wanted the opportunity for more collaboration, which OE provides. Aug 27 16:20:57 for example Jalimo Aug 27 16:21:32 so you think your sales will increase by using OE, due to the larger dev community? Aug 27 16:22:25 perhaps. but essentially i think it's just a better product with OE. It's more compelling on many levels. Aug 27 16:22:56 being an engineer, i try to avoid sales speculations whenever possible :) Aug 27 16:23:08 heh Aug 27 16:23:33 we all need to eat :) Aug 27 16:24:15 hehe yes Aug 27 16:24:33 absolutely :) I guess to better answer your question, we did not move to OE to sell units to OE devs. We did it in part to give customers more choice in what they run on the device. Aug 27 16:42:13 hackable device with OE support is nice thing Aug 27 16:42:21 indeed Aug 27 16:42:35 more companies welcome Aug 27 16:43:43 So OE gives customers choices :) Aug 27 16:44:09 * Crofton|work thinks we also do pretty good for the static cases Aug 27 16:44:11 if they get OE to use Aug 27 16:44:25 bbl Aug 27 16:44:33 even if the vendor provides feeds, there are loads of choices Aug 27 16:44:36 I do not think that labquest users really use their Poky for adding new stuff Aug 27 16:44:44 hm where is our site which gives an overview who is using OE in product ;) Aug 27 16:45:00 choice means different things to different people :) Aug 27 16:45:54 woglinde: and how they use it.. Aug 27 16:46:04 woglinde: did you saw recent Garmin devices? Aug 27 16:46:12 hrw no Aug 27 16:46:19 they are using OE? Aug 27 16:46:29 poky at least Aug 27 16:46:34 so another player next to tomtom is using linux Aug 27 16:46:35 http://butterfeet.org/?p=57 Aug 27 16:46:54 woglinde: Garmin use Poky + own addons Aug 27 16:47:12 woglinde: at least it looks like it Aug 27 16:47:25 and their developer was quite active on Poky ML Aug 27 16:47:30 heh... but we don't have yet usable world-wide navigation software.. Aug 27 16:47:43 hrw they only have one dev? Aug 27 16:47:56 or only one dev of them is communicate with you Aug 27 16:47:58 woglinde: one to contact with upstream Poky Aug 27 16:48:12 ah okay Aug 27 16:48:29 jay7 on n810 you have navicore Aug 27 16:48:56 uh 2.6.17 kernel Aug 27 16:49:02 adding tomtom support into OE is not hard task I think. But what for use TT device if you cannot run TTN on it? Aug 27 16:49:05 smells like montavista Aug 27 16:49:33 woglinde: nokia used 2.6.16/18/21 for long time Aug 27 16:50:10 yes true Aug 27 16:50:20 I hope garmin fixed all security bugs Aug 27 16:50:21 *g* Aug 27 16:50:25 ;D Aug 27 16:50:43 I have device here with 2.6.16, other with 2.6.18/24... Aug 27 16:50:51 both are custom devices Aug 27 16:51:21 sandmann time Aug 27 16:51:25 $WORK-1 was using 2.6.12 last time I checked. :( Aug 27 16:56:48 fresh ;D Aug 27 16:58:00 I still have 2.4.30 on one device here Aug 27 17:08:11 btw - does someone know board with 2-3 ethernets and 2 minipci slots? (other then pc engines alix boards) Aug 27 17:22:57 hrw, gateworks has some - the avila GW2348 comes to mind but perhaps something in the Cambria family as well Aug 27 17:23:06 hrw, ixp4xx based Aug 27 17:50:56 thx Aug 27 17:51:00 have a nice evening Aug 27 18:02:00 gn Aug 27 18:05:17 anyone else notice the overlap in functionality between tools like openembedded and tools like cruisecontrol? Aug 27 18:11:21 cruise control is some form of automated testing application? Aug 27 18:13:24 hmmm Aug 27 18:39:57 * kgilmer hasn't had much good luck with cruisecontrol Aug 27 18:41:01 kergoth`work: looks like CC is intended as a wrapper around your build system rather than the build system itself... Aug 27 18:41:03 what i mean is, conceptually, they both build a set of projects for integration and release. some build & release tools can handle dependency, just few can scale to what OE does Aug 27 18:41:08 hvontres|work: which is exactly what OE is. Aug 27 18:41:19 just cc's scale is different Aug 27 18:41:33 CC is more about the lifecycle of continuos automation testing Aug 27 18:41:43 internally it has no dependency logic Aug 27 18:41:56 it relies on Ant for dependency resolution Aug 27 18:41:58 i'm talking about enterprise level build & release management tools in general Aug 27 18:42:00 cc is just one Aug 27 18:42:15 gotcha. Aug 27 18:42:22 how does OE do this? Aug 27 18:42:28 this is something I'm struggling with now Aug 27 18:42:39 we have integration/testing/prod branches Aug 27 18:42:45 which can include recipes as well as sources Aug 27 18:42:56 in CC i would create three seperate environments Aug 27 18:43:03 and use branches or tags for checkout Aug 27 18:43:23 I guess the same thing could be achieved by passing in tags or branches into the SRC_URI variable. Aug 27 18:43:30 I'd like to keep it as light as possible. Aug 27 18:43:44 yeah, thats probably the best way to handle it Aug 27 18:44:33 one difference we have is that the recipe for something lives in a different place than the sources and patches Aug 27 18:44:37 this causes extra complexity Aug 27 18:45:32 * kergoth`work would like to see oe metadata in the upstream source tree working better, seperating the metadata & patches from the source makes the life of the application developer more difficult Aug 27 18:46:10 exactly Aug 27 18:46:20 my next task is to dig into buildbot Aug 27 18:46:35 see if it has a happy-path for testing automation workflow Aug 27 18:46:50 that's one of my biggest gripes with oe right now, it neither scales down to the app developer or up to the level of buildbot or similar tools. people end up wedging it into their environments because its scope is relatively narrow Aug 27 18:47:30 hmm... what's an example of something that scales like that? Aug 27 18:47:31 kergoth`work: heh... that would make for an interesting chicken & egg problem Aug 27 18:49:31 i also think some of this could be solved in tooling but i imagine there is such diversity in what people like to use that it would difficult to build higher level tools that a majority of devs would use. Aug 27 18:50:12 i've noticed one common pattern among OE subprojects and diravatives is having a scripts directory do take care of some common operations. Aug 27 18:50:34 kergoth, have you ever made buildbot dop anything useful? Aug 27 19:06:14 * kgilmer takes that as a 'no' Aug 27 19:15:23 re Aug 27 19:16:43 re Aug 27 19:35:49 ~lart smoking motors Aug 27 19:35:49 * ibot makes a balloon animal out of smoking motors Aug 27 19:47:25 hvontres|work: did you let the magic out of another electric device? Aug 27 19:49:30 * Esben wonders when we will see cross and staging merge into one directory... Aug 27 20:21:22 mwester: yup. cooked a custom three phase motor here at work.... and it was "pilot" error Aug 27 20:24:40 re Aug 27 20:26:39 re ant Aug 27 20:26:58 ciao woglinde Aug 27 22:09:03 * mwester has just completed an audit of the latest SlugOS builds downloads, and discovered that it uses two different versions of the openssl library... sigh. Aug 27 22:10:56 mwester how is this happens? Aug 27 22:11:04 oe dont know about version depens Aug 27 22:11:45 That's what I'm trying to figure out... I just did a "find" to check the last access time of everything in the dl directory after a SlugOS build... Aug 27 22:11:50 Ah. Aug 27 22:12:01 :) one is -native Aug 27 22:12:05 hehe Aug 27 22:12:09 Must be a preferred version somewhere. Aug 27 22:12:50 The real one is for OpenSSH, dunno what's pulling in the -native version. It didn't used to do that. Aug 27 22:28:32 git-native is pulling in the other openssl library. Aug 27 22:28:52 How does everyone else handle that? Still building git, or can we just assume-provided for that one in general? Aug 27 22:29:51 these days I'd say lets ASSUME_PROVIDE it Aug 28 00:34:58 <_hc> I am trying to figure out how to pin a package to a specific version in DEPENDS Aug 28 00:35:05 <_hc> would something like DEPENDS = "vala-native=0.1.5 vte libmokoui2" Aug 28 00:35:08 <_hc> work? Aug 28 00:42:34 _hc: for build-time dependencies, just use "PREFERRED_VERSION_vala-native = 0.1.5" (or whatever the version string is; check in the conf directory for examples of how this is used) Aug 28 00:54:50 <_hc> mwester: thanks, trying it now Aug 28 01:02:08 <_hc> hmm, its building vala-native-0.3.4 again... arg Aug 28 01:10:32 <_hc> is there a way to have both versions of vala-native (0.1.5 and 0.3.4)? Aug 28 01:10:32 No. Aug 28 01:11:00 Use the -g option to bitbake to dump the dependency graphs, it will tell you what is depending upon vala, and will give you an idea what else might want the newer version. Aug 28 01:11:30 Or you can try bitbake with debugging enabled, and check the verbose output -- it might tell you why it chose to override your preferred version. Aug 28 01:14:42 <_hc> I think I forced it by deleteing vala-native_0.3.4.bb... :-/ Aug 28 01:16:06 <_hc> w00t, openmoko-terminal2 built, mission accomplished Aug 28 01:16:10 <_hc> thanks Aug 28 02:50:06 * * OE Bug 4526 has been created by lukas.gorris(AT)gmx.de Aug 28 02:50:08 * * wlags-modules do_compile fails Aug 28 02:50:10 * * http://bugs.openembedded.net/show_bug.cgi?id=4526 Aug 28 02:57:06 * * OE Bug 4158 has been RESOLVED (FIXED) by lukas.gorris(AT)gmx.de Aug 28 02:57:08 * * cs do_fetch fails Aug 28 02:57:10 * * http://bugs.openembedded.net/show_bug.cgi?id=4158 **** ENDING LOGGING AT Thu Aug 28 02:59:56 2008