**** BEGIN LOGGING AT Wed Oct 21 02:59:59 2009 Oct 21 05:33:16 good morning. I'm trying to build a custom image for the beagleboard. Starting with console-base-image.bb, I want to add the packages I need. I tried adding mplayer and omapfbplay to IMAGE_INSTALL, but both fail with libavcodec errors when run. Oct 21 05:49:30 03Klaus Kurzmann  07shr/import * rfcefc06877 10openembedded.git/recipes/shr/e-wm-theme-illume-gry_git.bb: Oct 21 05:49:30 e-wm-theme-illume-gry_git.bb: new recipe Oct 21 05:49:30 Signed-off-by: Klaus Kurzmann Oct 21 05:49:40 03Klaus Kurzmann  07shr/import * r60e292753f 10openembedded.git/recipes/shr/elementary-theme-gry_git.bb: Oct 21 05:49:40 elementary-theme-gry_git.bb: RRECOMMEND e-wm-theme-gry Oct 21 05:49:40 Signed-off-by: Klaus Kurzmann Oct 21 06:27:59 tasslehoff: and the question is, is that the way I should add packages? :) Oct 21 06:32:52 tasslehoff: It is the correct way, according to http://bec-systems.com/site/177/the-correct-way-to-add-packages-to-an-openembedded-image Oct 21 06:34:48 tsjsieb: ah. thanks. then the maybe the problem is in dev. I'm trying the same build from stable now to check if that works out better. Oct 21 07:37:25 seems it's an ffmpeg bug Oct 21 07:45:25 what decides which version of ffmpeg is built? I can't find any references to it in sane-srcdates.inc or sane-srcrevs.inc Oct 21 07:47:49 ffmpeg.bb Oct 21 07:49:04 XorA|gone: hmm. could it be more obvious? :) Oct 21 07:49:26 is the files I mentioned only used to lock code that oe checks out from version control= Oct 21 07:49:29 ? Oct 21 07:50:12 that was original intent, never really worked out Oct 21 07:51:01 XorA|gone: hmm. so that's not the mechanism to use? Oct 21 07:52:20 XorA|gone: I can't find any ffmpeg.bb. I have for instance ffmpeg_git.bb and ffmpeg_0.5.bb. Oct 21 07:54:44 ffmpeg_git these days I think Oct 21 07:55:54 XorA|gone: ok, but where is that decided? if another recipe depends on ffmpeg, where does it say which version will be built? Oct 21 07:56:42 good morning Oct 21 07:57:52 morning Oct 21 08:04:52 its a balance of defaul_pref and pref_version Oct 21 08:07:18 03Klaus Kurzmann  07shr/import * ree8fcddfa7 10openembedded.git/conf/distro/include/shr-autorev-unstable.inc: Oct 21 08:07:18 shr-autorev-unstable.inc: set e-wm-theme-illume-gry Oct 21 08:07:18 and elementary-theme-neo to AUTOREV Oct 21 08:07:18 Signed-off-by: Klaus Kurzmann Oct 21 08:07:25 03Klaus Kurzmann  07shr/import * r2c31f5e921 10openembedded.git/recipes/shr/e-wm-theme-illume-gry_git.bb: Oct 21 08:07:25 e-wm-theme-illume-gry_git: fix wrongly spelled RRECOMMENDS Oct 21 08:07:25 Signed-off-by: Klaus Kurzmann Oct 21 08:09:58 XorA|gone: ok. I found angstrom-2008-preferred-versions.inc. where is the default prefs set? Oct 21 08:10:18 in the .bb Oct 21 08:11:54 otavio_: pong Oct 21 08:11:57 morning Oct 21 08:25:57 XorA|gone: which .bb? Oct 21 08:26:20 03Marcin Juszkiewicz  07org.openembedded.dev * r8967b7d7c1 10openembedded.git/recipes/linux/linux-bug_2.6.27.2.bb: Oct 21 08:26:20 linux-bug: added extra staging, bumped SRCREV Oct 21 08:26:20 Signed-off-by: Marcin Juszkiewicz Oct 21 08:33:28 XorA|gone: it searches ffmpeg*.bb and looks for one with DEFAULT_PREFERENCE="1" ? Oct 21 08:35:04 highest number wins Oct 21 08:36:54 XorA|gone: another piece of the puzzle :) Oct 21 08:38:33 So versions are decided from sane-srcrevs, sane-srcdates, preferred_versions and default_preference?? Oct 21 08:45:54 * tasslehoff guesses: if a preferred version is set, use that on, if not, use the one with highest default preference. if the winner is versioncontrolled, the sanesrc-files are checked Oct 21 08:50:25 which package will provide me the a2ps tool ? Oct 21 08:51:20 pb_: which holiday inn is the nearest to the venue? Oct 21 08:51:26 good morning, btw. Oct 21 08:55:26 mickey|office: the "express" one. Oct 21 08:55:29 (good morning) Oct 21 08:57:03 I think the link on the wiki page goes to the right place. it should say it is in "coldhams park" or "norman way". Oct 21 08:57:27 if it talks about clifton road or cattle market or impington, you are looking at the wrong one Oct 21 09:01:56 hi all Oct 21 09:02:58 is there any small device that can run linux which have 1ethernet port, usb port, whole system is on external memory card (SD, etc..), has more ram than WRT devices and is cheap? Oct 21 09:03:37 AlHafoudh: how cheap is cheap? Oct 21 09:04:18 hrw: $50 ? :-/ Oct 21 09:04:45 good morning Oct 21 09:05:44 AlHafoudh: There should be some cheap routers that meet these requirements. Or the FriendlyARM mini2440 Oct 21 09:05:50 AlHafoudh: no idea. asus wl500g could be intereseting to check Oct 21 09:05:56 pb_: oki, thanks. Oct 21 09:06:34 hi gremlin[it] did you read my email? Oct 21 09:07:05 pb_: hmm. d'oh. lufthansa doesn't seem to serve Stansted. seem to support LCY only Oct 21 09:08:00 florian: Mini2440 ARM9 Board looks very good, but it is not in case, is it? Oct 21 09:09:34 AlHafoudh: you're askig for impossible stuff Oct 21 09:09:39 mckoan: i dont think so :) Oct 21 09:09:56 mckoan, yes ... but i know yestarday probably i'm stucked at a cardiac pacing congress on the same fri/sat of OEDEM ... Oct 21 09:09:58 AlHafoudh: good, let us know once you find it ;-) Oct 21 09:10:31 ~curse cardiac pacing congress Oct 21 09:10:35 May the fleas of a thousand camels infest your most sensitive regions, cardiac pacing congress ! Oct 21 09:10:44 AlHafoudh: so you need to grab some 4th hand used stuff Oct 21 09:11:29 AlHafoudh: I have a couple of NSLU2 here :-D Oct 21 09:12:07 AlHafoudh: yes... andahammer.com might even have cases for these Oct 21 09:12:59 mickey|office: doh. I'm pretty sure germanwings fly FRA-STN if that is any good. Oct 21 09:13:30 I would prefer Lufthansa too :) Oct 21 09:13:46 oh, hm, no, germanwings seem to have discontinued FRA service Oct 21 09:13:51 mckoan, yes... damn ... Oct 21 09:14:10 yeah Oct 21 09:14:14 they serve from cologne only Oct 21 09:14:17 which kind of sucks from FRA Oct 21 09:14:31 yeah Oct 21 09:14:44 for me cologne is ok Oct 21 09:15:01 i can get to LCY or Heathrow Oct 21 09:15:13 but that's about 2 more hours by train right? Oct 21 09:15:50 a bit less than that, but definitely further away than STN Oct 21 09:15:55 mickey|office: I checked how long the whole trip would take by train... from here its was > 10 hours :) Oct 21 09:16:09 florian: wah Oct 21 09:16:42 are you all staying at hostel? Oct 21 09:16:46 STN to cambridge is about 25-30 minutes on the train, it's a direct link. LHR to cambridge is somewhere between 1.5 and 2.5 hours depending on how much you want to spend and how lucky you are with the connections. Oct 21 09:17:03 LCY about the same as LHR, not much to choose between them Oct 21 09:17:19 ok, thanks. will continue to look around then Oct 21 09:17:33 LTN is closer geographically though transport links not that great Oct 21 09:17:46 why don't you have an airport? :) Oct 21 09:17:56 a public one, i mean Oct 21 09:18:04 i know there's a private one Oct 21 09:18:09 heh. we do, and in fact our office is right next to it. but I think the last scheduled airline ceased service last year. Oct 21 09:18:16 You men you don't have a jet? Oct 21 09:18:17 if only you had your own plane, you would be all set Oct 21 09:18:19 mean even Oct 21 09:18:23 hehe Oct 21 09:18:24 mickey|office: so all you need is a plane ;) Oct 21 09:18:28 hehe Oct 21 09:18:29 Goodness, they'll let anyone into OE these days. Oct 21 09:18:33 or a helicopter would do at a pinch, even Oct 21 09:18:45 good point Oct 21 09:18:51 where's airforce one when you need it... Oct 21 09:18:55 * florian adds this to his personal wishlist Oct 21 09:19:06 I think that we just need OE Jet Oct 21 09:21:18 We need rich sponsors first I guess Oct 21 09:21:55 hmm Oct 21 09:22:01 * mickey|office sighs and books FRA->LCY Oct 21 09:22:16 hrw: florian: how about both? ;-) Oct 21 09:22:17 going for a nice train ride Oct 21 09:22:20 yeah, that looks like as good an option as any Oct 21 09:22:42 ryanair does not make sense to me as I'm living 10 minutes from FRA Oct 21 09:22:45 it looks like you either have to ride the ICE to Stuttgart or Cologne, or ride the train when you get to the UK Oct 21 09:22:49 yep Oct 21 09:23:00 :-) Oct 21 09:23:04 yeah, or go to Hahn, but that sounds like an awful option for every reason Oct 21 09:23:19 *nod* Oct 21 09:23:47 might just as well lay a trail of broken glass to our office and then crawl here on hands and knees Oct 21 09:23:53 hahaha Oct 21 09:23:55 pb_: I guess the train to Stuttgart would be easier... Hahn is.... well... at the countryside ;) Oct 21 09:23:59 ya, it's about that bad Oct 21 09:24:15 at least i get some bonus miles that way... Oct 21 09:34:03 is this http://bec-systems.com/site/456/capture-oe-source-changes the recommended way to do oe source changes?? Isn't there a workflow that could let me to "edit -> compile -> test -> edit -> compile -> test" again and again and then finally make a patch when I'm done? Oct 21 09:34:48 Im trying to modify httpd.conf (apache2) in my overlay but when I build and the modified file is not included in my image. Isn't the apache2 recipe using that file? Oct 21 09:35:02 pb_: is it advisable to book the train ride upfront or can i do it after arriving in london? Oct 21 09:43:51 mickey|office: you can do it on the spot. if you book in advance then you might get a very slightly cheaper fare, but personally I wouldn't bother. Oct 21 09:44:19 pb_: ok, thanks. Oct 21 09:44:43 It's also very painful to book .uk train travel outside .uk (or was last time I looked, anyway). Oct 21 09:45:14 from LCY, half the journey is on the subway anyway and you can't book that in advance. it's only the train from kings cross to cambridge that would be pre-bookable, and those are so frequent that you are probably better off just buying your ticket on the spot. Oct 21 09:45:46 last time when I had to take a train in UK it was the same price one-month before and on spot Oct 21 09:45:54 if you pre-book you generally have to fix the exact time when you want to travel, which means that you run the risk of either hanging around for ages at the station (if you arrive earlier than expected) or missing the train altogether and having to buy a new ticket. Oct 21 09:48:45 ok, makes sense Oct 21 09:54:16 k, hotel and flight done Oct 21 09:54:23 very good Oct 21 10:21:45 is there a way to prevent the checksum checking on do_fetch ? Oct 21 10:24:38 I don't think there is a way to prevent it altogether, other than hacking base.bbclass. Oct 21 10:25:03 There are variables you can set to make it less fatal for packages to not have a checksums.ini entry at all, but if the entry is there then I think it is always checked. Oct 21 10:25:04 thx pb_ Oct 21 10:25:24 see OE_ALLOW_INSECURE_DOWNLOADS and OE_STRICT_CHECKSUMS. I am not quite sure what the difference is. Oct 21 10:25:34 thx nice Oct 21 10:25:41 IIRC one only moanes, the other bails out Oct 21 10:30:45 ah yeah, that could be Oct 21 10:30:53 ~hail git send-email Oct 21 10:30:56 * ibot bows down to git send-email and chants, "I'M NOT WORTHY!!" Oct 21 10:31:08 * pb_ goes to light a fire Oct 21 10:31:13 suddenly the weather has gone all winter-like Oct 21 10:33:08 yeah Oct 21 10:33:10 * mickey|office freezing Oct 21 10:33:14 03Klaus Kurzmann  07org.openembedded.dev * rcd473395eb 10openembedded.git/recipes/libgee/libgee-native_git.bb: Oct 21 10:33:14 libgee-native: new recipe Oct 21 10:33:14 Signed-off-by: Klaus Kurzmann Oct 21 10:33:15 03Michael 'Mickey' Lauer  07org.openembedded.dev * r7dd5a642d0 10openembedded.git/conf/distro/include/fso-autorev.inc: fso-autorev.inc: add libgee-native Oct 21 10:33:16 03Klaus Kurzmann  07org.openembedded.dev * rbfcebb7e0c 10openembedded.git/recipes/freesmartphone/vala-dbus-binding-tool-native_git.bb: Oct 21 10:33:19 vala-dbus-binding-tool-native: depend on libgee-native Oct 21 10:33:21 Signed-off-by: Klaus Kurzmann Oct 21 10:33:25 03Michael 'Mickey' Lauer  07org.openembedded.dev * rc6dd121742 10openembedded.git/recipes/vala/vala-native_0.7.6-fso2.bb: vala: delete obsolete version Oct 21 10:33:36 pb_: Get a fire with remote control :) Oct 21 10:33:45 yeah, I heard from my friend in frankfurt that you had -3C there last week. Oct 21 10:34:02 brrrr Oct 21 10:34:11 it isn't quite that cold here at the moment but it is dark and wet. Oct 21 10:34:21 we had +3°C yesterday ~10:00 and that was enough Oct 21 10:34:27 florian: heh, that would be a nice idea. Oct 21 10:34:35 I guess I would need a small robot to bring the logs from outside Oct 21 10:35:18 I always wanted to build a small robot to bring coffee to the office... Oct 21 10:35:27 ah.. coffee Oct 21 10:35:37 my router has double coffee Oct 21 10:35:41 in mac address Oct 21 10:36:04 hrw: You ar using a Mr. Coffee as a router? ;) Oct 21 10:36:19 florian: no, C0:FF:EE:C0:FF:EE as mac address Oct 21 10:36:27 heh Oct 21 10:36:33 03Frans Meulenbroeks  07org.openembedded.dev * ra35ce39b0d 10openembedded.git/ (2 files in 2 dirs): Oct 21 10:36:33 kernel: changed defconfig so it has RTC as module; bumped MACHINE_KERNEL_PR Oct 21 10:36:33 still needs testing of things like sgx, dsplink, cmem, dmai, sdma and friends before this can be made Oct 21 10:36:33 the main/default version Oct 21 10:36:34 03Frans Meulenbroeks  07org.openembedded.dev * r74094fba9f 10openembedded.git/: Merge branch 'org.openembedded.dev' of git.openembedded.org:openembedded into org.openembedded.dev Oct 21 10:37:58 hehe Oct 21 10:38:41 the machine I'm getting coffee from, does run Angstrom :) Oct 21 10:39:22 with fresh beans Oct 21 10:39:28 tsjsieb: sounds pretty cool... what kind of machine is this?! Oct 21 10:39:29 Does anyone know where apache2 gets it's httpd.conf from as is doesnt seem to be the one from apache/files/httpd.conf Oct 21 10:40:16 well, it's 'n office coffee machine, containing embedded pc module, but it's still in development Oct 21 10:40:37 jovox: apache and apache2 are quite separate, it's unlikely that the latter would be taking files from the former Oct 21 10:40:53 from a quick look at the apache2.bb file it looks like httpd.conf is probably coming from the upstream tarball Oct 21 10:41:06 tsjsieb: Send a link once it is public :) Oct 21 10:41:26 pb_ ok, what's your suggestion if I want to change it? Oct 21 10:42:00 I don't think there would be any harm in shipping it as a loose file from recipes/apache2/files like apache1 does. Oct 21 10:42:59 pb_ and change the apache2.bb to use that file instead.. oki.. I'll try that Oct 21 10:43:25 right Oct 21 10:43:42 tsjsieb: as soon as we can buy one put me down for it ;-) Oct 21 11:00:24 03Michael 'Mickey' Lauer  07org.openembedded.dev * r29507043d5 10openembedded.git/ (3 files in 2 dirs): libgee: use 'require' rather than duplicating the recipe; also add a sane srcrev into sane-srcrevs.inc Oct 21 11:01:05 hi, i needed php and noticed it was not on the angstrom feed, so tried to compile it myself and I bumped into something odd: Oct 21 11:01:28 the problem is that I get error: 'struct utsname' has no member named 'domainname' Oct 21 11:02:07 this decl is surrounded by #ifdef __USE_GNU in sys/utsname.h, otherwise it is __domainname Oct 21 11:02:24 the odd thing is that if I add -D__USE_GNU to the compile command, I still get the error Oct 21 11:02:42 adding -Ddomainname=__domainname results in a working system Oct 21 11:02:55 pb_: btw., if we don't get any more candidates for the TSC, we might as well elect all of 'em. Didn't we say 4-6 as optimal number? Oct 21 11:03:05 Why is __USE_GNU on the cmd line not honoured, someone explicitly switching it off? Oct 21 11:03:13 and how is this best resolved ? Oct 21 11:03:24 eFfeM: you should define _GNU_SOURCE, not __USE_GNU Oct 21 11:03:39 mickey|office: yah, that might make sense Oct 21 11:06:51 pb_ thanks, that indeed worked Oct 21 11:08:29 testing rebuilding the full recipe will commit if it works Oct 21 11:14:22 hi mickey|office Oct 21 11:14:23 err Oct 21 11:14:27 hi mithro Oct 21 11:14:28 heh Oct 21 11:14:31 ~lart tab completion Oct 21 11:14:31 * ibot slaps tab completion upside the head with a wet fish Oct 21 11:14:38 and schizophrenie Oct 21 11:14:49 mickey|office: hey Oct 21 11:14:52 long time no see Oct 21 11:15:08 indeed. what are you up to these days? Oct 21 11:16:29 I work at the big G Oct 21 11:16:40 uh oh! Oct 21 11:16:50 hmm... should i say "congrats"... Oct 21 11:16:54 well, i guess yeah Oct 21 11:16:57 so, congrats Oct 21 11:16:58 been a year and a half now Oct 21 11:17:07 which department, if i may Oct 21 11:17:08 ? Oct 21 11:17:41 group called Partner Solutions Organisation - we work mainly with the big partners Oct 21 11:18:38 my work consists of part sys-admin, part software development Oct 21 11:19:04 I use my 20% time for working with local FOSS groups like SLUG and Open Australia Oct 21 11:19:18 still leading my game project Thousand Parsec too Oct 21 11:19:31 sounds good. so you're relatively busy and have a stable jog Oct 21 11:19:33 err Oct 21 11:19:34 job, that is Oct 21 11:20:04 way too busy :P Oct 21 11:20:18 how goes OE and yourself? Oct 21 11:20:34 I see you finally moved to git? Oct 21 11:22:23 yes, monotone kind of collapsed Oct 21 11:22:53 we're pretty happy w/ git Oct 21 11:22:58 people branch and merge! Oct 21 11:23:03 yay Oct 21 11:23:05 (which is quite an achievement) Oct 21 11:23:09 we moved to git a long time ago too :P Oct 21 11:23:23 (Thousand Parsec that is) Oct 21 11:23:33 at work we use perforce - which is cringe worthy Oct 21 11:24:04 where you working these days? Oct 21 11:24:25 Still freelancer. After Openmoko pulling the plug on FSO, I had to do more iphone work, but looking for more FOSS jobs again. Oct 21 11:25:02 do you do much OE work? Oct 21 11:26:29 testing patchstatus of 12 kernels for many devices takes time ;( Oct 21 11:26:40 depends. the devices i care about run OE userlands, so I have to regularly update my recipes in OE Oct 21 11:27:03 and recently we did more work on the e.V. as well Oct 21 11:27:07 (finally...) Oct 21 11:27:31 you might have seen that i dropped out of OE for a while Oct 21 11:27:54 e.V? Oct 21 11:28:04 I havn't been watching OE at all for years how Oct 21 11:28:11 registered association to support further development of OE Oct 21 11:28:13 mickey|office: by the way, my company is interested in finding someone to develop an iphone app for us. if you would like to do that, or know someone else who might, maybe you could send me an email. Oct 21 11:28:20 hrw: hello Oct 21 11:28:33 hrw: can you send me your slides? Oct 21 11:28:34 otavio_: you pinged me during my night Oct 21 11:28:48 pb_: oh cool, will send you something soon. thanks Oct 21 11:28:48 otavio_: http://marcin.juszkiewicz.com.pl/download/presentations/ Oct 21 11:28:58 hrw: cool; thx Oct 21 11:29:22 mickey|office: great, thanks. by the way, did you respond to that guy yesterday? Oct 21 11:29:29 I cringe everytime the android guys try and solve an issue that the oe solved years ago Oct 21 11:29:30 mithro: the good thing is OE is still around and there are hardly any alternatives. Even the big names are starting to use OE now Oct 21 11:29:59 pb_: yeah Oct 21 11:30:03 excellent, thanks Oct 21 11:30:21 big names? Oct 21 11:30:22 mickey|office: true, it's not a bad time to be using and building on OE. Oct 21 11:30:32 mithro: Montavista Oct 21 11:30:37 bbiab, gotta debug our broken vacuum cleaner now. Oct 21 11:30:40 mithro: Palm, Montavista Oct 21 11:30:53 mithro: TI Oct 21 11:30:58 and looks like Garmin uses Poky for last devices Oct 21 11:31:10 mithro: Atmel Oct 21 11:31:50 cool Oct 21 11:33:06 hrw: really, ok, that's nice. Oct 21 11:33:52 whats Nokia using these days? Oct 21 11:34:01 mithro: same as before - scratchbox Oct 21 11:47:38 doh, inadequate tools Oct 21 11:47:53 ~lart dyson for using torx screws in deep recesses Oct 21 11:47:53 * ibot beats dyson severely about the head and shoulders with a rubber chicken for using torx screws in deep recesses Oct 21 11:58:08 how can I check *why* a certain package is being built? bluez4 for example Oct 21 11:58:33 I tried reading depends.dot, but I don't see anything depending on bluez4. Oct 21 11:59:01 it won't appear on my image either, but it's still getting built Oct 21 12:05:52 oh, it's libpcap requiring bluez-libs. Oct 21 12:06:47 it's just that depends.dot doesn't explicitly state the complete dependency chain from libpcap -> blues-lib -> bluez4. Oct 21 12:10:08 someone should figure out how to build Androd user space with OE :) Oct 21 12:19:16 Is "hotplug" in MACHINE_FEATURES enough to make my device laod drivers automatically ? Oct 21 12:20:48 Crofton|work: 0x1abs did Oct 21 12:23:37 Android plus bitbake - that'll be quick. Oct 21 12:25:06 uf.. sorted most of kernel patching problems Oct 21 12:25:09 right, better port bitbake to java Oct 21 12:25:48 .net Oct 21 12:26:54 03Marcin Juszkiewicz  07org.openembedded.dev * r6bbdd3395a 10openembedded.git/conf/checksums.ini: checksums.ini: added some kernel patches Oct 21 12:31:13 oh god cups and friends is quite a nightmare to setup Oct 21 12:32:34 Am I right in saying for the most part these days RPROVIDES and PROVIDES are a bad idea? Oct 21 12:38:01 I don't think they are any more of a bad idea now than they ever have been in the past. Both of them are things that you probably want to use sparingly but they do have their uses. Oct 21 12:39:34 Well in the context of a recipie that provides what it says in the filename ;-). Just noticed an odd thing. If I have a recipe that RPROVIDES and PROVIDES the same as it's filename the resulting package does not get shipped in the image but it seems removing the RPROVIDES and PROVIDES fixes things (I notice a lot of old OE recipies seem to RPROVIDES and PROVIDES the same as the filename). Oct 21 12:40:15 why is cups still calling gs even it is already a *.ps file .. Oct 21 12:40:29 rob_w: er, gs is used to translate .ps to something else Oct 21 12:40:38 not to create .ps files Oct 21 12:40:57 right ok makes sense Oct 21 12:41:21 DJWillis: that does sound odd. Oct 21 12:41:25 as i am using a *.ps file. so gs is still needed ? yet into another format Oct 21 12:41:45 setting "RPROVIDES" (literally) is almost always a bad idea. generally, you only want to set RPROVIDES_${PN} or something similar Oct 21 12:41:57 rob_w: that depends on what languages your printer understands natively Oct 21 12:42:09 if you have postscript input, and your printer can interpret postscript for itself, you don't need gs. Oct 21 12:42:29 if your printer doesn't have a postscript interpreter onboard, you do need gs in order to print postscript input Oct 21 12:42:38 yep, so thats why ive setup up hplip Oct 21 12:42:46 maybe I missed something, today I get this new error ERROR: Unable to parse conf/bitbake.conf (Could not include required file conf/distro/include/glibc-${TOOLCHAIN_TYPE}.inc) Oct 21 12:42:54 pb_: yep, taken me a few days to track down, just checking more (takes a while to build up a suitable image) but it does seem to do that on my env. anyway. Infact I think it is just RPROVIDES that may be the odd thing. Not simple PROVIDES Oct 21 12:43:29 small "mobile" printers from hp dont have "that luxury" of knowing ps Oct 21 12:43:29 right, so you do need gs in that case Oct 21 12:43:42 pb_: you have confirmed my thoughts that RPROVIDES (without a PN) is a bad idea anyway. Oct 21 12:43:58 pb_, i am also doubting the need for gs or not yet know where cups decides to use it Oct 21 12:44:19 or iam hoping to at least drop this step somehow Oct 21 12:45:06 rob_w: I don't quite understand what you are saying. Oct 21 12:45:33 if you want to print postscript input (i.e. you have a .ps file) and your printer does not understand postscript for itself, you need to use gs. Oct 21 12:45:37 there is no way to avoid that. Oct 21 12:46:05 conversely, if you don't want to use gs, you need to either get a printer that can interpret ps for itself, or arrange for your input file to arrive in a pre-rasterised format. Oct 21 12:48:08 most linux programs expect to generate postscript output when printing, so the latter course might be a bit of a struggle, but if your target application set is small then it is probably a viable route. Oct 21 12:48:37 yeah , so far i got this in the way your saying that, as far as i see the cups "war machine" at one point uses gs and gives a with a paramter a optional rasterizer , which in my case is hplip Oct 21 12:50:03 more or less, yeah Oct 21 12:50:19 gs -q -dBATCH -dPARANOIDSAFER -dQUIET -dNOPAUSE -sDEVICE=ijs Oct 21 12:50:20 -sIjsServer=hpijs is what is called actually inside the foomatic-rip-hp perl script Oct 21 12:50:23 the ijs output plugin isn't really a rasteriser, though: it just takes raster images from the gs core and translates them into ijs format Oct 21 12:50:51 hpijs alone is not capable of interpreting postscript, it is just an output driver to be called by ghostscript Oct 21 12:51:14 yep Oct 21 12:52:16 so i see cups pulling all those helper (where this foomatic-rip-hp script errror) already Oct 21 12:52:56 actually this calles gs and gs dies a early, cruel "fatal error" at interp.c Oct 21 12:53:19 that is rather sad Oct 21 12:53:24 that might be the call to hpijs Oct 21 13:03:30 Crofton|work: thx Oct 21 13:03:35 do i need to make gs aware of future hpjs usages at compile time ? Oct 21 13:04:27 yeah i do and i see ijs as available device Oct 21 13:05:10 03Marco Cavallini  07org.openembedded.dev * r9273927fff 10openembedded.git/conf/distro/kaeilos.conf: Oct 21 13:05:10 conf/distro/kaeilos.conf added new distro options Oct 21 13:05:10 * added TOOLCHAIN_TYPE management Oct 21 13:05:10 * switched to glibc 2.9 Oct 21 13:05:10 * selected xserver-xorg as default, since kdrive has been EOL'ed Oct 21 13:05:11 * bumped PR Oct 21 13:10:55 03Marcin Juszkiewicz  07stable/2009 * r6222a39133 10openembedded.git/recipes/linux/linux-bug_2.6.27.2.bb: Oct 21 13:10:55 linux-bug: added extra staging, bumped SRCREV Oct 21 13:10:55 Signed-off-by: Marcin Juszkiewicz Oct 21 13:10:55 Acked-by: Philip Balister Oct 21 13:11:58 03Martin Jansa  07org.openembedded.dev * r9566f5fa67 10openembedded.git/conf/checksums.ini: Oct 21 13:11:58 Missing few xorg checksums.ini Oct 21 13:11:58 Signed-off-by: Marcin Juszkiewicz Oct 21 13:15:30 Regarding local file. netbase fx. has file://interfaces in its SRC_URI. Apparently it is possible to have that file change depending of the platform. fx recipes/netbase/netbase/qemux86. I can't get it to work with my local recipes. Are there any special tricks I should do? Oct 21 13:16:44 ./obj/echogs -e .dev -w- -l-obj ./obj/pdffonts -ps gs_mex_e gs_mro_e gs_pdf_e gs_wan_e Oct 21 13:16:44 | ./obj/echogs: ./obj/echogs: cannot execute binary file that is from gs-8.64.bb do_compile Oct 21 13:17:17 because gs people told me to use this version rather then 8.16 (but that one compiles) Oct 21 13:18:59 that must mean that it's trying to run a target binary on the build host Oct 21 13:19:12 you probably need to patch the makefiles to build echogs with $(BUILD_CC) Oct 21 13:19:51 sgh: what exactly is not working for you? Oct 21 13:21:08 there aren't any special tricks, though you might want to look at the definition of ${FILESPATH} in bitbake.conf if you haven't already. Oct 21 13:22:12 I'm trying to understand how the bb files works. I took a look in the simple initrdscripts recepies. all they do is to install some files from the "files" directory. In the beginning of the bb files the SRC_URI poits out the file to be installed. What I don't understand is that the URI doesnt contain the "files" directory. It looks like this: SRC_URI="file://85-blockboot.sh". How does it know that the file is in the "files" directory? The reason I'm asking is be Oct 21 13:22:43 jovox: see what I just said to sgh Oct 21 13:23:29 and, also, the fine manual: http://bitbake.berlios.de/manual/ch03s02.html Oct 21 13:27:49 Ok, but I don't see either FILESPATH or FILESDIR being set in the recipe, but that might be set 'globally'? Oct 21 13:28:03 there aren't any special tricks, though you might want to look at the definition of ${FILESPATH} in bitbake.conf if you haven't already. Oct 21 13:29:11 Ok, thanks Oct 21 13:34:14 pb_: i placed the file "interface" in my local recipe-dir undet netbase/netbase/mpc5121ads it still fetches the default interface. Copying the mpc5121ads-dir to the main recipe-dir under openembedded makes netbase grab the interface-file in the mpc5121ads-dir... Oct 21 13:35:15 by "local recipes" you mean you are using a collection? Oct 21 13:36:52 03Koen Kooi  07org.openembedded.dev * rd57e14b9e2 10openembedded.git/recipes/u-boot/ (14 files in 6 dirs): u-boot: fix fw_env logic, update beagle and touchbook Oct 21 13:36:53 03Koen Kooi  07org.openembedded.dev * r926ae17102 10openembedded.git/recipes/linux/ (6 files in 2 dirs): linux-omap 2.6.29: update touchbook patches Oct 21 13:36:54 03Koen Kooi  07org.openembedded.dev * rd81d0ec74e 10openembedded.git/recipes/pam/libpam_1.0.2.bb: libpam: add meta package for all plugins Oct 21 13:36:56 03Koen Kooi  07org.openembedded.dev * rd2fcef1e0b 10openembedded.git/ (conf/checksums.ini recipes/xdg-utils/xdg-utils_1.0.2.bb): xdg-utils: add 1.0.2 Oct 21 13:36:59 03Koen Kooi  07org.openembedded.dev * r85c20c24bc 10openembedded.git/recipes/ffmpeg/ (ffmpeg_svn.bb omapfbplay_git.bb): Oct 21 13:37:02 ffmpeg: bump SRCREV Oct 21 13:37:04 omapfbplay: bump PR Oct 21 13:37:06 03Koen Kooi  07org.openembedded.dev * r59f089e8ec 10openembedded.git/ (3 files in 3 dirs): mplayer: bump SRCREV, remove pld patch that went upstream Oct 21 13:37:09 03Koen Kooi  07org.openembedded.dev * re51c625a06 10openembedded.git/recipes/totem/totem-pl-parser_2.28.1.bb: totem-pl-parser: update to 2.28.1 Oct 21 13:37:12 03Enric Balletbo i Serra  07org.openembedded.dev * ra52956b5aa 10openembedded.git/ (3 files in 3 dirs): Oct 21 13:37:15 linux-igep2: Add 2.6.28.10 for IGEP v2 machine Oct 21 13:37:22 mickey|office: welcome back Oct 21 13:37:23 Signed-off-by: Enric Balletbo i Serra Oct 21 13:37:25 Signed-off-by: Koen Kooi Oct 21 13:37:27 03Enric Balletbo i Serra  07org.openembedded.dev * r54eab44990 10openembedded.git/conf/machine/igep0020.conf: Oct 21 13:37:27 thanks Oct 21 13:37:30 igep0020: Add IGEP v2 support to OE Oct 21 13:37:32 Signed-off-by: Enric Balletbo i Serra Oct 21 13:37:34 Signed-off-by: Koen Kooi Oct 21 13:37:36 03Koen Kooi  07org.openembedded.dev * r7097dc9b69 10openembedded.git/ (conf/checksums.ini recipes/libgmime/gmime_2.4.10.bb): gmime: add 2.4.10 Oct 21 13:37:39 03Koen Kooi  07org.openembedded.dev * rdfba140ac1 10openembedded.git/conf/machine/igep0020.conf: igep0020: use omap3.inc Oct 21 13:37:42 03Koen Kooi  07org.openembedded.dev * r9fb7e7b2d2 10openembedded.git/recipes/gnome/gnome-python-extras_2.25.3.bb: gnome-python-extras: fix up patch to python stuff Oct 21 13:37:51 03Koen Kooi  07org.openembedded.dev * ra4f21f295b 10openembedded.git/contrib/angstrom/sort.sh: angstrom feed sorter: add igep2 support Oct 21 13:37:54 03Koen Kooi  07org.openembedded.dev * r91f7f50d7e 10openembedded.git/ (conf/checksums.ini recipes/gnome/libgdata_0.4.0.bb): libgdata: add 0.4.0 Oct 21 13:37:57 03Koen Kooi  07org.openembedded.dev * r4bdc304d25 10openembedded.git/recipes/totem/ (16 files in 4 dirs): totem: clean up totem dir Oct 21 13:38:02 03Koen Kooi  07org.openembedded.dev * r57e9e4b5b2 10openembedded.git/recipes/gnome/gnome-screensaver_2.28.0.bb: gnome-screensaver: fix packaging Oct 21 13:38:05 03Koen Kooi  07org.openembedded.dev * r88f07770b5 10openembedded.git/recipes/totem/totem_2.28.1.bb: totem: update to 2.28.1 Oct 21 13:43:03 pb_: yes Oct 21 13:44:33 ah, I see. I'm not quite sure how collections and FILESPATH interact nowadays; you used to have to modify FILESPATH but that might not be the case anymore. Oct 21 13:44:41 kergoth is the expert on collections, you might ask him. Oct 21 13:50:48 morning Oct 21 13:53:46 pb_: thanks you. Oct 21 14:08:52 I'm building glibc_2.9.bb but it claims FATAL: kernel too old http://tinderbox.openembedded.net/packages/296471/ which kernel is it referring to? Oct 21 14:09:21 mckoan: most likely kernel headers Oct 21 14:09:26 linux-libc-headers or such Oct 21 14:10:05 I think that error message with "FATAL" in means that your qemu-advertised kernel is less than glibc was expecting. Oct 21 14:10:07 zecke: in OE build tree? or in HOST machine? Oct 21 14:10:13 it isn't usually quite so shrill about header problems Oct 21 14:10:30 thanks, I'm wrong. didn't see it fails in do_package Oct 21 14:11:11 pb_: what do you mean with "qemu-advertised kernel"? Oct 21 14:11:25 the kernel version that qemu reports to the apps that run under it Oct 21 14:13:08 pb_: and do you know how to solve it? is possible to inform qemu to use a difefrent one? Oct 21 14:13:48 yes, it is. see the mailing list archives, there was a long-ish discussion about this a few months ago. Oct 21 14:14:18 pb_: keyword is qemu or glibc ? Oct 21 14:15:31 03Michael 'Mickey' Lauer  07org.openembedded.dev * r7100155e21 10openembedded.git/recipes/freesmartphone/ (4 files): fso: add some new dependencies Oct 21 14:15:32 03Michael 'Mickey' Lauer  07org.openembedded.dev * r62f3874f4a 10openembedded.git/recipes/pulseaudio/libcanberra_0.17.bb: libcanberra: depends on libtool Oct 21 14:15:33 03Michael 'Mickey' Lauer  07org.openembedded.dev * receec52d44 10openembedded.git/recipes/tasks/task-fso2-compliance.bb: task-fso2-compliance: add libfsoresource and fsogsmd Oct 21 14:15:34 03Michael 'Mickey' Lauer  07org.openembedded.dev * re8576cce84 10openembedded.git/conf/distro/include/ (fso-autorev.inc sane-srcrevs-fso.inc): Oct 21 14:15:35 sane-srcrevs-fso.inc: update to current cornucopia SRCREV Oct 21 14:15:37 fso-autorev.inc: add libfsoresource Oct 21 14:15:48 mckoan: see http://projects.linuxtogo.org/pipermail/openembedded-devel/2009-May/010999.html and the ensuing thread Oct 21 14:16:37 I'm not sure there was ever a completely satisfactory resolution to that but it should at least shed some light on what is going on. Oct 21 14:19:22 pb_: looks like it could be "[oe] [RFH] Stable qemu-native BROKEN" Oct 21 14:19:33 ah, could be that as well Oct 21 14:19:46 I tend to ignore any mails which mention the stale branch. Oct 21 14:23:35 :-) Oct 21 14:29:01 pb_, thatnks for your support Oct 21 14:30:11 heh Oct 21 14:30:52 not that I have anything against it, I just don't personally have any interest in using it. Oct 21 14:31:03 pb_, at this point, when I think about the TSC, I can't see voting for people who are not supportive of tha goals of the stable branch Oct 21 14:31:15 03Michael 'Mickey' Lauer  07org.openembedded.dev * r9043ab7d88 10openembedded.git/ (3 files in 2 dirs): serial-utils: fetch updated versions from FSO git and bump sane-srcrevs Oct 21 14:31:43 it is one thing to be disinterested, another to seem hostile Oct 21 14:44:47 good morning Oct 21 14:45:27 morning Oct 21 14:47:04 gm Oct 21 14:47:05 I'm starting to work on DM365 EVM. I'd like to avoid !!! MV . The question is: video codecs and general kernel support are at a good point? Oct 21 14:47:13 Crofton: gm Oct 21 14:47:48 recalcati_, there is ongoing work in that area, but i do not know the exact status Oct 21 14:48:43 hi kergoth Oct 21 14:52:20 morning kergoth Oct 21 14:54:34 03Frans Meulenbroeks  07org.openembedded.dev * rd9cb7ebc9d 10openembedded.git/recipes/php/ (php-5.2.6/string.c.patch php.inc php_5.2.6.bb): php: fixed build errors Oct 21 14:57:23 pb_: about qemu error is ti solved or still open? Looks like it has been solved only in stable branch Oct 21 15:00:41 that'd be surprising. as of our policies stuff can only go into stable once it's in OE Oct 21 15:00:43 err Oct 21 15:00:44 .dev Oct 21 15:02:09 hmm, that is in the policy, but over time it becomes less workable Oct 21 15:02:11 * kergoth thinks we should think about renaming the branches to git conventions at some point in the future Oct 21 15:02:27 since dev diverges Oct 21 15:03:33 so are you all able to build for arm at the moment? Oct 21 15:03:59 using glibc_2.9 I mean Oct 21 15:04:22 kergoth, what conventions might those be? Oct 21 15:04:47 "org.openembedded.dev" is awfully monotoneish. we call that "master" :P Oct 21 15:05:10 heh Oct 21 15:05:18 kergoth, agreed Oct 21 15:05:54 and it is easier to check out master Oct 21 15:06:42 less typing is a good thing. i've found git's shell completion to be rather slow, since it often doesnt know what you're trying to complete Oct 21 15:06:43 heh Oct 21 15:06:46 bother meta-toolchina fails for me Oct 21 15:07:33 Crofton|work: where? binutils? Oct 21 15:07:43 urg Oct 21 15:07:49 let me find a tinderbox link Oct 21 15:09:19 http://tinderbox.openembedded.net/builds/35077/ Oct 21 15:09:21 grrr Oct 21 15:10:00 mckoan: stable/2009 angstrom uses 2.6.1 glibc Oct 21 15:10:20 mckoan: and it works properly with stable/2009 qemu-native for armv6 target Oct 21 15:11:12 hrw: and what about .dev ? Oct 21 15:11:50 mckoan: I do armv5te, armv4t, armv6, armv7a with dev angstrom glibc 2.9 and qemu-native still works Oct 21 15:12:35 Crofton|work: aeh? anyway, time to sleep Oct 21 15:13:56 I have a fix in bintuls-cross to put .debug into the right -dbg package Oct 21 15:13:57 bye Oct 21 15:14:34 hrw: this issue sounds too difficult to solve in a reasonable amount of time, looks like would be better I revert to PREFERRED_VERSION_glibc ?= "2.6.1" although I wouldn't like to withdraw Oct 21 15:15:46 mckoan: its probably because you blindly copied angstrom which has some kludgws to fix Oct 21 15:15:54 have to go now Oct 21 15:17:53 XorA|gone: yes may well be true, although I couldn't figure out which kludgws to fix Oct 21 15:18:44 * mckoan would like to use Angstrom if it wasn't so fast forwarding, I need something more stable Oct 21 15:19:29 stable = not changing every minute Oct 21 15:19:38 :-D Oct 21 15:21:38 03Koen Kooi  07org.openembedded.dev * rf48b7c01ac 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.org:openembedded into org.openembedded.dev Oct 21 15:21:38 03Koen Kooi  07org.openembedded.dev * r42a4fc9e98 10openembedded.git/recipes/tasks/task-perl-module-all.bb: task-perl-module-all: switch it to RRECOMMENDS Oct 21 15:21:39 03Koen Kooi  07org.openembedded.dev * r48239648e9 10openembedded.git/recipes/ti/ (ti-dmai/touchbook-support.patch ti-dmai_svn.bb): ti-dmai: add igepv2 support Oct 21 15:21:40 03Koen Kooi  07org.openembedded.dev * r12603e25de 10openembedded.git/recipes/totem/totem-2.28.1/gst-detect.diff: totem: add missing patch Oct 21 15:22:22 heh, i think zecke has a fix for my problem, but the tinderbox messages are not clear Oct 21 15:22:34 not critical, hopefully he pushes soon Oct 21 15:56:56 mckoan: welcome to the club, I know a lot of people suffereing from these fast changes Oct 21 16:08:28 florian: I thought that stable was created to avoid this pain, but I was wrong Oct 21 16:10:26 florian: I'd like to discuss this topic at OEDEM Oct 21 16:10:54 yes, I think that would be a good agenda item for OEDEM Oct 21 16:11:12 pb_: great Oct 21 16:11:18 mckoan: yes right... we need to find out why doesn't work like we handle it currently Oct 21 16:11:31 mcgeagh: yes that's a good idea Oct 21 16:11:46 s/mcgeagh/mckoan Oct 21 16:11:54 florian: :-D Oct 21 16:12:01 mckoan: please go ahead and add it to the agenda on the wiki page Oct 21 16:12:46 pb_: at what time? looks like all is already scheduled Oct 21 16:13:13 there goes my idea of adding a branch between stable and devel Oct 21 16:13:21 perhaps i get more support for it this time Oct 21 16:13:22 moreover consider that most of us will leave after lunch time Oct 21 16:13:23 :) Oct 21 16:13:34 leave after lunch time on Sunday Oct 21 16:13:36 mckoan: just add it somewhere, we will reshape the agenda Oct 21 16:13:40 (on the first day= Oct 21 16:13:44 mickeyl: I proposed that as well - something like Debian has with testing Oct 21 16:13:52 florian: *nod* Oct 21 16:14:23 another thing is about Pre-meeting beer session at the Red Lion, Histon Oct 21 16:14:42 too fat from city centre for who (like me) is on foot Oct 21 16:15:58 mckoan: ah, hm. let me check the agenda again. Oct 21 16:16:13 apropos the beer session, there is a direct (albeit not especially speedy) bus if you don't want to walk Oct 21 16:17:09 mckoan: 14:00-18:00 on Sunday is still free. Put it in there for now; like mickey says, we can move things around later. Oct 21 16:19:33 pb_: my flight is leaving at 1:45 on Sunday Oct 21 16:19:43 pb_: my flight is leaving at 18:45 on Sunday Oct 21 16:20:01 pb_: so I won't be there on Sunday afternoon Oct 21 16:20:21 some topics can perhaps be shrinked Oct 21 16:20:33 I added something in http://wiki.openembedded.net/index.php/Oedem/2009 Oct 21 16:20:33 and we can also vote for an order on site Oct 21 16:20:39 * mckoan must go now Oct 21 16:20:47 have a nice rest of the day Oct 21 16:24:29 mickeyl: yeah, I just gave all of them an hour in the first instance, but I imagine they will be differing lengths in reality Oct 21 16:24:49 I suspect we will have to make a lot of the detail of the agenda up as we go along, but it is nice to have a list of items if not a firm schedule. Oct 21 16:26:02 time for another session with the vacuum cleaner now that I have a bigger screwdriver. Oct 21 16:26:04 * pb_ -> Oct 21 16:30:09 hehe Oct 21 16:30:11 good speed Oct 21 16:42:11 pb_: This Oedem/2009, could that be interesting for or beginners as well? I happen to be in London that week Oct 21 16:52:23 doesnt looks like it when I read the agenda :) Oct 21 16:57:53 jovox: you are welcome to come along but I suspect you would find most of the agenda quite boring. Oct 21 17:02:04 is there a way to stop sane-srcrevs from overriding SRCREV = "${AUTOREV}" in a recipe? Oct 21 17:04:14 I don't think there's any way to stop it, as such, but you can override it back again. Oct 21 17:04:25 SRCREV_pn-${PN} = "${AUTOREV}" Oct 21 17:05:02 ok. that's what I was considering but it felt like there might be a better way Oct 21 17:06:57 pb_, we can move the infrastructure session to Sunday afternoon Oct 21 17:07:36 we need to get the autorevs out ASAP ..... Oct 21 17:08:13 what's the alternative if you want to a recipe to build SVN head though? Oct 21 17:08:37 well, AUTOREC is ok in your local data Oct 21 17:08:43 but is bad in the public repo Oct 21 17:09:13 ah right. yes agreed. i thought you were implying that the feature would be dropped from bitbake Oct 21 17:09:14 what happens is the repo AUTOREV points at goes out, and everyones builds fail Oct 21 17:09:17 no Oct 21 17:09:29 apparently some got pushed though Oct 21 17:09:44 also, sometimes poeple do builds when they are not connected to the net .. Oct 21 17:10:51 it's very useful for local builds because of the way it automatically rebuilds dependencies if someone makes a commit Oct 21 17:15:22 drat, just got my car back from being repaired and now I have a puncture. Oct 21 17:15:25 how tiresome Oct 21 17:17:07 hi all Oct 21 17:21:16 03Koen Kooi  07org.openembedded.dev * r66a55e59bd 10openembedded.git/recipes/gnome/gnome-bluetooth_git.bb: gnome-bluetooth git: bump SRCREV Oct 21 17:21:26 03Koen Kooi  07org.openembedded.dev * r240441d73b 10openembedded.git/recipes/gnome/gedit_2.28.0.bb: gedit: fix packaging so that the .pc ends up in -dev and we don't get the depchain code adding deps to gtksourceview-dev Oct 21 17:21:27 03Koen Kooi  07org.openembedded.dev * reaafa4bd4a 10openembedded.git/ (6 files in 3 dirs): gtksourceview: update to latest release, fix packaging Oct 21 17:21:57 Hi. I'd like to compile Scheme code using Chicken (http://www.call-with-current-continuation.org) on OE. Chicken compiles Scheme to C, then uses a C compiler to compile C code to native code. In the end, I'd like to have recipes which are Scheme code to be compiled to native code to the target platform. What's the best way to do that? I suppose I'd need to have Chicken to be installed on the staging area. Oct 21 17:23:00 mike_cw, I've done that, but it means keeping locol diffs Oct 21 17:32:15 mario-goulart, i think you need a ckicken-native recipe and a chicken bbclass Oct 21 17:33:11 playya: is there any similar compiler I can use as kind of guide? Oct 21 17:33:38 maybe vala Oct 21 17:34:01 vala compiles genie/vala files to c code Oct 21 17:34:08 but it uses autotools Oct 21 17:35:02 Hmmm. Interesting. Thanks a lot for the tip. I'm gonna take a look at vala. Oct 21 17:41:44 right, infrastructure session moved to sunday morning, stable branch now set for sunday 11am Oct 21 17:41:58 er, infrastructure moved to sunday afternoon, not morning Oct 21 17:42:35 mickeyl: what time are you planning to arrive in cambridge? Oct 21 17:43:52 pb_: i'll arrive at LCY @ 16:25 on friday. Depending on connections, I'd hope to be in the hotel around 19:00 then Oct 21 17:44:03 righto Oct 21 17:44:19 I'll leave on monday @ 15:55 from LCY Oct 21 17:45:00 okay, very good Oct 21 17:52:15 pb_, cbrake moved the docs from amethyst yesterday Oct 21 17:52:28 we are down to patchwork and the website on amethyst Oct 21 17:52:48 well, dns wise :) Oct 21 17:58:42 righto Oct 21 17:59:17 I've added a bit more travel information to the wiki page as well. Oct 21 18:01:32 pb_, can I collect T shirts for US guys who can't make it Oct 21 18:01:40 khem would like a shirt Oct 21 18:02:26 hi Crofton, wow oe T-shirts...could I buy one at fosdem if I can make it to there? Oct 21 18:02:40 "Crofton went to OEDEM and all I got was this lousy t-shirt" Oct 21 18:02:47 ah ok Oct 21 18:02:59 rofl Oct 21 18:03:02 not sure Oct 21 18:03:06 so no oe t-shirts? Oct 21 18:03:24 ah, I'm not sure exactly what is going to happen with t-shirts. right now, so few people seem to be interested that it is not going to be economic to get them printed. Oct 21 18:03:39 hmm Oct 21 18:04:12 he many would you need to make it economical? Oct 21 18:04:41 probably 15-20, thereabouts. I think at present the total stands at about four. Oct 21 18:05:25 I'd like one Oct 21 18:05:28 XL Oct 21 18:05:40 well, if USsizes map to .uk sizes Oct 21 18:05:58 and khem asked if I could bring one back Oct 21 18:07:13 XL size seems to be about 44" chest in uk speak Oct 21 18:07:16 XXL is 46" Oct 21 18:08:11 hmm Oct 21 18:08:24 I need a tape measure Oct 21 18:08:24 ah, hm, varies a bit with vendor. fruits of the loom is saying XL=44/46, XXL=47/49 Oct 21 18:09:01 and gildan says XL=46/48, XXL=50/52 Oct 21 18:09:07 heh Oct 21 18:10:09 what about using something like cafepress? Oct 21 18:10:12 I'll try and find out some prices and lead times tomorrow. Oct 21 18:10:48 kergoth, we should have florian see if he can get a cafepress account for the ev Oct 21 18:10:54 hi Oct 21 18:11:06 I wonder if cafepress works internationally? Oct 21 18:11:18 * Crofton really needs to work on slides Oct 21 18:11:21 good question, never used them myself, just seen others use them Oct 21 18:11:22 hm what was cafepress? Oct 21 18:11:26 cafepress does have a uk branch but I think their tshirts are all laser printed Oct 21 18:11:28 the t-shirt maker? Oct 21 18:11:31 ah Oct 21 18:11:47 they're also quite expensive for what you get Oct 21 18:12:35 for something like the oe logo, you'd get a better result with screen printing, and probably cheaper too if you have a reasonable quantity to do Oct 21 18:13:38 screen setup is about £20 per colour (so £40 for oe logo) but then the shirts come out cheaper than laser transfer ones. Oct 21 18:14:32 ahh, cool Oct 21 18:16:12 yeah, but it gets us out of shipping Oct 21 18:16:27 and makes it "easy" for people to provide financial support to the ev Oct 21 18:23:52 yeah, cafepress might be a fine idea for general oe eV merchandise Oct 21 18:29:42 by all accounts, the cafepress printed mugs (and other ceramics) seem to be pretty good. I don't think I've ever actually seen one myself but they seem to be fairly well thought of. Oct 21 18:30:10 personally I would avoid their clothing although I guess the convenience does have its attractions. Oct 21 18:33:07 fwiw, my favoured printing place is http://www.4imprint.co.uk/, might be worth a look for your ev bits. Oct 21 18:33:38 you'd still have to do the shipping, but their stuff is (in general) quite a lot cheaper than somewhere like cafepress and in most cases better. Oct 21 18:46:56 03Koen Kooi  07org.openembedded.dev * re8a1700ea4 10openembedded.git/recipes/geoclue/geoclue_git.bb: geoclue: use the correct virtual for gconf Oct 21 18:46:56 03Koen Kooi  07org.openembedded.dev * r5411ee5d6f 10openembedded.git/recipes/maemo4/libgpsbt_0.1.bb: libgpsbt: fix QA errors Oct 21 18:58:56 pb_, maybe we should look at selling shirts at FOSDEM Oct 21 19:00:07 jo florian Oct 21 19:00:15 hm android cellphone from acer Oct 21 19:00:45 yes, that's not a bad idea Oct 21 19:00:47 hm and spring 3.0 is out Oct 21 19:02:12 morning florian Oct 21 19:02:28 re Oct 21 19:02:30 hi rkirti Oct 21 19:02:39 hi rkirti Oct 21 19:02:49 * woglinde is working on chinook compat Oct 21 19:02:51 hello woglinde Oct 21 19:03:17 * rkirti peeks in Oct 21 19:42:21 kergoth: Hi... pb_ told me that you know the handling of collections well. Oct 21 19:43:42 did you have a question about it? Oct 21 19:47:57 kergoth: yes .... it relates to netbase which installs /etc/network/interfaces. I can see that it is possible to install a different file depending of the current platform. It works well as long as I place a dir named in recipes/netbase/netbase/. But what I really want is to place it in my local collection, and that does not work the same way apparentl ;(. Oct 21 19:53:02 sgh: http://article.gmane.org/gmane.comp.handhelds.openembedded/25159 Oct 21 19:53:37 sgh: if you change your FILESPATHBASE that way, then it will pick up file:// files from your collection Oct 21 19:53:46 alternatively, set it based on COLLECTIONS rather than BBPATH Oct 21 19:53:58 oh, it does, nevermind :) Oct 21 19:54:06 just copy/paste that FILESPATHBASE Oct 21 19:56:01 kergoth: will it then pick up files from openembedded if it does not find them in my local overlay ? Oct 21 19:56:33 of course. Oct 21 19:56:51 FILESPATHBASE defaults to the location of the recipe, which is why it looks relative to there by default Oct 21 19:57:32 that version of it just prepends the other paths in your collections to that. it doesn't remove it Oct 21 19:58:05 kergoth: ehm ..... ok,, I will try it. Thanks. Oct 21 19:58:24 just make sure the files are the same location relative to the root of the collection Oct 21 19:58:29 i.e. recipes/foo/ Oct 21 19:58:54 kergoth: it wud be nice if there was a sample example in OE Oct 21 19:59:09 these files could be committed Oct 21 20:01:23 may just add the FILESPATHBASE change to collections.inc, so it works out of the box when you use COLLECTIONS Oct 21 20:01:41 was going to do that for mvl6 out of the box, but ran into an unrelated bitbake bug.. never had a chance to revisit it Oct 21 20:01:48 one of these days Oct 21 20:43:54 kergoth: How about amend-recipes.inc it should also make sense to have it committed in conf/ I guess Oct 21 20:44:22 kergoth: my COLLECTIONS is this : /home/sgh/svn/OpenEmbedded/build-ppc /home/sgh/svn/OpenEmbedded/openembedded Oct 21 20:44:38 enabling it results in FILESPATH being expanded twice during each parse, which means any recipes that haven't yet switched away from base_set_filespath will get warned multiple times Oct 21 20:44:43 if any are left, anyway Oct 21 20:44:49 if not, might as well enable it by default i'd think Oct 21 20:45:00 sgh: okay, and? Oct 21 20:45:28 kergoth: I think there are recipes which still use FILESPATH Oct 21 20:46:05 FILESPATH is used everywhere, its the ones that use base_set_filespath rather than changing FILESPATHPKG that issue the warning Oct 21 20:46:27 03Stanislav Brabec  07org.openembedded.dev * rbd9190c531 10openembedded.git/recipes/ (17 files in 9 dirs): Oct 21 20:46:27 many recipes: Recipes must not depend on gconf-dbus directly, only dbus can be referred: Oct 21 20:46:27 * Only PREFERRED_PROVIDER_gconf in distro definition can refer to gconf-dbus. Oct 21 20:46:27 * Both gconf-dbus and gconf provide than same interface. Oct 21 20:46:27 * New gconf packages already provide defaults settings via dbus. Oct 21 20:46:30 * Increment PR to rebuild against correct package. Oct 21 20:46:35 kergoth: ehm .../.../..buid-ppc contains conf,recipes,tmp and .../../../openembedded is openembedded itself. Is it correct, or should COLLECTIONS be setup differently ? Oct 21 20:47:44 that's fine Oct 21 20:48:55 kergoth: ok ... then I do not understand, because that is the setup that does not work. Oct 21 20:49:09 "does not work" tells me nothing Oct 21 20:49:18 i misplaced my crystal ball, i'm sorry to say Oct 21 20:50:22 kergoth: I know .... details comming up .. :D Oct 21 20:50:25 -rw-r--r-- 1 sgh users 129 21 okt 15:00 build-ppc/recipes/netbase/netbase/mpc5121ads/interfaces Oct 21 20:50:25 -rw-r--r-- 1 sgh users 282 29 sep 08:54 openembedded/recipes/netbase/netbase/qemux86/interfaces Oct 21 20:51:13 MACHINE is mpc5121ads? what's FILESPATHBASE set to, exactly? Oct 21 20:51:18 run bitbake -e | grep \^FILESPATHBASE= Oct 21 20:51:41 FILESPATHBASE="conf" Oct 21 20:52:09 so you didn't do what i told you to do Oct 21 20:52:20 i said to set it the way it's set in that link i gave you Oct 21 20:52:25 and you didn't, so that's why it isn't finding it Oct 21 20:53:51 kergoth: where is the search for the files done ? Oct 21 21:00:50 kergoth: I will have a decent look at it. Thank you. Oct 21 21:01:09 every file:// url is found via FILESPATH Oct 21 21:01:26 FILESPATH is defined in bitbake.conf, like nearly every other important variable Oct 21 21:02:50 RP: kergoth bitbake -c clean hangs if I use bitbake master Oct 21 21:03:02 its after recent commits from RP Oct 21 21:03:12 previously it worked ok Oct 21 21:06:04 i'd suggest doing a bisect to find the exact commit Oct 21 21:06:25 yeah I was planning on doing that Oct 21 21:16:14 kergoth: currently COLLECTIONS="/home/sgh/svn/OpenEmbedded/build-ppc /home/sgh/svn/OpenEmbedded/openembedded" - according to your link it should also include "recepies" - is that correct? Oct 21 21:16:23 no. Oct 21 21:16:26 that isn't what i said Oct 21 21:16:43 what i said is to make sure that the paths to the files relative to the collection are the same Oct 21 21:16:56 so if the files are "recipes/foo/bar" in openembedded, then you'd want your files in build-ppc/recipe/foo/bar Oct 21 21:17:08 s/recipe/recipes/ Oct 21 21:19:38 kergoth: sure, I know that. And I actually believe that I have that setup. bitbake knows my local recipes and all, so it seems strange to me. Oct 21 21:19:51 once again, set FILESPATHBASE the way it's set in the link i gave you Oct 21 21:19:53 it's not that hard Oct 21 21:19:57 copy/paste into your local.conf Oct 21 21:32:04 kergoth: hey there. Still battling with packaged-staging...there are a couple of things I don't understand in staging_helper () Oct 21 21:32:28 i'd suggest asking the person that wrote it :) Oct 21 21:34:06 perhaps I'm dumb, but the opkg.conf seems wrong, causing the other issues Oct 21 21:35:04 I'll ask pb__ to put this in the agenda for the 'hacking-break' ;) Oct 21 21:35:17 RP is supposed to be there Oct 21 21:35:27 * kergoth nods, probably a good idea Oct 21 21:38:00 kergoth: the copy/paste sure changed FILEPATHBASE, but still it was not finding my local file. I'll mess with it some. Thanks for your help. Oct 21 21:38:12 check FILESPATH. Oct 21 21:38:28 bitbake -e | grep \^FILESPATH= Oct 21 21:42:37 kergoth: /home/sgh/svn/OpenEmbedded/build-ppc/conf:/home/sgh/svn/OpenEmbedded/openembedded/conf:conf Oct 21 21:42:56 seems ok. At least both are represented there now :) Oct 21 21:43:15 bitbake -e netbase|grep \^FILESPATH Oct 21 21:43:22 to see the ones for that recipes Oct 21 21:43:24 s/s$// Oct 21 21:44:55 kergoth: ah ok... oe-dirs are place first, damn Oct 21 21:45:17 COLLECTIONS is in priority order, highest to lowest Oct 21 21:45:33 kergoth: so my BBFILES should just be reversed.... Oct 21 21:45:40 not sure wha tyou mean Oct 21 21:45:47 the order of files in bbfiles is irrelevent Oct 21 21:45:50 not a factor Oct 21 21:46:28 kergoth: ok, but it is where the paths come from. Oct 21 21:47:12 i wrote the filespath code, and i wrote collections.inc, you don't need to tell me what comes from where, thanks Oct 21 21:47:22 again, the order of files in bbfiles is irrelevent Oct 21 21:47:30 the order of entries in collections and bbpath are what matters Oct 21 21:47:58 kergoth: sorry ... missed a questionsmark there :) Oct 21 21:50:37 bitbake parses everything in bbfiles, but the order there isn't used, it obeys the priorities set with the BBFILE_COLLECTION variables to decide between them Oct 21 21:59:24 kergoth: I'm not sure we are actually talking about the same thing. I might not have described the issue correctly. It is late here - may I write to you at a later time? Oct 21 22:00:33 you asked about collections/overlays, so that's what i told you about. you don't have to use collections.inc, it just automates setting bbfiles and the collection variables so you can copy a recipe from OE into your collection and have it use the correct one. Oct 21 22:01:07 if all you want to mess with is individual files rather than recipes, you don't need to mess with any of that, and can just set FILESPATHBASE based on BBPATH rather than COLLECTIONS, or set just COLLECTIONS and FILESPATHBASE, should be fine Oct 21 22:01:53 that FILESPATHBASE you copied is the same order as the entries in COLLECTIONS, so as long as your collection is first in COLLECTIONS, it'll be before oe in FILESPATH. Oct 21 22:04:27 kergoth: ok. Those vars are just new to me. Thank you for your time. Oct 21 22:05:32 np Oct 21 22:05:48 again, you can read the definitions of the FILESPATH variables, they'll make sense if you know python Oct 21 22:09:42 kergoth: I'm a python pre-newbie :( But I intend to learn it. Oct 21 22:20:58 kergoth: do we need to modify BBPATH with collections Oct 21 22:21:28 the collections only need to be in BBPATH if you have conf/classes there that you need to use Oct 21 22:21:41 recipes just need bbfiles + bbfile_collection vars Oct 21 22:22:05 i see Oct 21 22:22:08 collections.inc will change bbpath if anything is missing and re-execute itself with the new path Oct 21 22:22:12 it's disgusting :D Oct 21 22:22:26 hmm Oct 21 22:22:38 mvl6 uses tarballs for its collections, so it was necessary Oct 21 22:22:40 fun stuff Oct 21 22:24:15 hmm gcc 4.4.2 came out last week Oct 21 22:24:32 meh, is the week over yet? Oct 21 22:25:00 expecting another release? :) Oct 21 22:25:25 http://gcc.gnu.org/bugzilla/buglist.cgi?bug_status=RESOLVED&resolution=FIXED&target_milestone=4.4.2 Oct 21 22:33:20 grg: I plan to move gcc 4.4.1 to 4.4.2 Oct 21 22:33:51 actually I can also take just the diff Oct 21 22:41:50 hmm, I need to get mru to look at 4.4.2 on armv7 Oct 21 22:42:36 khem, i haven't looked at it yet, so you may as well do it Oct 21 22:42:59 I have tested it a bit here Oct 21 22:43:08 there are like 25 bug fixes Oct 21 22:43:34 I was just thinking of backporting the diff to 4.4.1 Oct 21 22:44:24 would that be ok ? I dont think it makes sense to add more clutter Oct 21 22:44:43 or I could rename the existing recipes for 4.4.1 to 4.4.2 Oct 21 22:44:48 I am looking for some fixes of bad codegen for armv7 Oct 21 22:47:55 I dont think any of the fixes are specific to arm let alone armv7 Oct 21 22:48:09 and half of total fixes are prolly for fortran Oct 21 22:48:25 and ada Oct 21 22:57:22 yeah Oct 21 22:57:48 not sure it is worth messing with from my point of view :( Oct 21 22:59:36 well, chances are 4.4.3 will get released in a few months. Oct 21 22:59:54 and you'll want to get those fixes in too Oct 21 23:00:27 renaming makes things clearer for people using the recipe than a backported diff Oct 21 23:00:48 * Crofton grg agreed Oct 21 23:01:08 either way, it means everyone using 4.4.1 has to rebuild their whole tree Oct 21 23:01:26 and its the default in a lot of places now, right? Oct 21 23:01:36 only if the distros bump their preferred versions :) Oct 21 23:01:58 Angstrom is still on 4.3 for most targets Oct 21 23:02:04 oh ok Oct 21 23:02:07 word is codegen in 4.4 is bad for armv7 Oct 21 23:02:23 bad luck eh :( Oct 21 23:02:33 mips codegen is muchly improved Oct 21 23:02:35 I want the NEON intrinsics in 4.4, but not that bad Oct 21 23:02:37 ah Oct 21 23:02:44 good to know Oct 21 23:03:03 I think we can set compiler on a per machine basis Oct 21 23:03:15 but not sure how many mips users are active in Angsrtom Oct 21 23:03:32 count me as one Oct 21 23:04:20 bring it up on the list, most of the active people use arm stuff Oct 21 23:06:08 also, enabling --with-mplt on mips builds is a good thing on gcc 4.4 Oct 21 23:06:44 have you ever looking in the distro and machien dirs to see how all this gets set? Oct 21 23:06:59 i'm just looking now to find out where i put that :) Oct 21 23:07:28 I can help you find it, but not right now :) Oct 21 23:10:10 03Andrea Adami  07org.openembedded.dev * r243600420c 10openembedded.git/recipes/makedevs/ (makedevs-1.0.0/makedevs.c makedevs_1.0.0.bb): Oct 21 23:10:10 makedevs: fix interpretation of device table (count). Oct 21 23:10:10 Signed-off-by: Antonio Ospite Oct 21 23:10:22 03Andrea Adami  07org.openembedded.dev * r2d7dbf672c 10openembedded.git/recipes/kexecboot/device_table-oldmmc.txt: device_table-oldmmc: adjust count fields according to makedevs fixes. Oct 21 23:10:23 03Andrea Adami  07org.openembedded.dev * ra61d083034 10openembedded.git/files/ (6 files): Oct 21 23:10:23 device_table(s): fix the 'count' field in device_tables. Line up to the new Oct 21 23:10:23 'count' field semantics in makedevs, it now means "number of device nodes to Oct 21 23:10:24 create" opposed to the old meaning "number of device nodes to create as if the Oct 21 23:10:26 'start' field was 0". Oct 21 23:10:28 Signed-off-by: Antonio Ospite Oct 21 23:13:19 mickeyl: I tested the patch before committing it Oct 21 23:15:19 ant__: excellent, thanks Oct 21 23:15:55 i meant to do it, but didn't have a chance Oct 21 23:16:36 actually for kexecboot we reScan the devices in the initramfs..so we don't need that file anymore Oct 21 23:16:51 but still I've seen was a 2005 uclibc bug :/ Oct 21 23:17:20 well, http://lists.busybox.net/pipermail/uclibc-cvs/2005-September/019451.html Oct 21 23:18:54 mickeyl: funny lines btw: ' /* FIXME: MKDEV uses illicit insider knowledge of kernel * major/minor representation... */ ' Oct 21 23:21:18 heh Oct 21 23:21:23 seems someone knew that since long Oct 21 23:21:29 but didn't bother to fix Oct 21 23:23:29 03Michael 'Mickey' Lauer  07org.openembedded.dev * r805bb50ef0 10openembedded.git/recipes/freesmartphone/libfso-glib_git.bb: libfso-glib: remove bogus additional prefix in SRC_URI Oct 22 00:02:55 03Michael 'Mickey' Lauer  07org.openembedded.dev * r7cb1640b06 10openembedded.git/recipes/tasks/task-cli-tools.bb: task-cli-tools: add serial-forward; bring back ltrace Oct 22 00:23:35 grg: Crofton so it seems renaming is what you guys incline too Oct 22 00:23:43 the patch applies cleanly to 4.4.1 too Oct 22 00:24:15 crofton: code gen is better in 4.4 for arm in general but the compiler itself is slow and needs more memory Oct 22 00:24:33 minimal and micro are the distros currently using gcc 4.4.1 as default. Oct 22 00:24:42 if you have a gcc 4.4.1 recipe with the 4.4.2 patch on it, people will probably wont realise that they are using 4.4.2 Oct 22 00:25:49 actually, thats not true, because gcc will install as 4.4.2 wont it... Oct 22 00:26:03 meh Oct 22 00:26:15 i still say move the recipe Oct 22 00:38:46 hi, I need dm9601 cortex, dos anyone have it? Oct 22 00:40:22 none for sale here Oct 22 00:40:39 nops, it is a module Oct 22 00:40:54 I need the arm port Oct 22 00:40:58 dm9601.ko Oct 22 00:44:02 I have a source code I modified to build dm9601.ko (in order to support my usb-ethernet adapter), and it works alright when compiling for x86. however, when I try to cross compile it to cortexA8 code it fails. Could someone please help???? Oct 22 01:02:41 what are folks using for mem leak testing on their embedded devices? Oct 22 01:03:11 i'm almost tempted to do an x86 build of my image just to use valgrind... Oct 22 01:04:32 grg: valgrind is great Oct 22 01:05:18 the on non-x86 options include boehmgc, or porting memprof but I have not used any of them Oct 22 01:06:21 nothing with an OE recipe already? Oct 22 01:09:54 I don't think so Oct 22 01:09:56 I have a source code I modified to build dm9601.ko (in order to support my usb-ethernet adapter), and it works alright when compiling for x86. however, when I try to cross compile it to cortexA8 code it fails. Could someone please help???? Oct 22 01:10:35 lim_, put the logs into a paste bin and someone will look at it Oct 22 01:10:47 ok Oct 22 01:15:47 grg: ok let me move it then Oct 22 01:16:27 grg: mpatrol is cool for memory usage too Oct 22 01:16:43 cheers Oct 22 01:17:10 mtrace(3) looks useful too Oct 22 01:18:03 I have a source code(http://pastebin.ca/1636424 (.c) http://pastebin.ca/1636426(.h) http://pastebin.ca/1636430(makefile)) I modified to build dm9601.ko (in order to support my usb-ethernet adapter), and it works alright when compiling for x86. however, when I try to cross compile it to cortexA8 code it fails(pastebin.ca/1636434 (crash)). Could someone please help???? Oct 22 01:18:20 grg: http://www.hpl.hp.com/personal/Hans_Boehm/gc/leak.html Oct 22 01:18:35 I have a source code(http://pastebin.ca/1636424 (.c) http://pastebin.ca/1636426(.h) http://pastebin.ca/1636430(makefile)) I modified to build dm9601.ko (in order to support my usb-ethernet adapter), and it works alright when compiling for x86. however, when I try to cross compile it to cortexA8 code it fails(http://pastebin.ca/1636434 (crash)). Could someone please help???? Oct 22 01:20:04 lim_: have you ever cross compiled a kernel? Oct 22 01:20:44 zecke_: no Oct 22 01:21:18 zecke_: but I figured that I chould only change the makefile to use the cross compiler Oct 22 01:21:26 *could Oct 22 01:21:33 lim_: maybe google for it. What do you think if something like "i586" comes up as unknown argument Oct 22 01:21:41 yep Oct 22 01:21:58 that happened here Oct 22 01:22:04 and some other flags too Oct 22 01:22:14 lim_: what do you think it could mean? Do you know what i586 is? Oct 22 01:22:22 lim_: you don't need to change the Makefile.. The kernel supports CROSS_COMPILE=arm-your-foo- and ARCH=arm... Oct 22 01:22:38 it is an architecture isn't it? Oct 22 01:23:30 really, I'll try adding these flags to the make file, then Oct 22 01:23:54 lim_: no, treat the Makefile as read-only... Oct 22 01:24:53 hi , is it possible to start OS very fast ? i mean instead of starting each service everytime , i write them on MMC and hibernate OS everytime from MMC on boot ? Oct 22 01:25:30 i dont want hibernate but like that... it will show popup image startup again but boots very fast. Oct 22 01:25:42 read-only? Oct 22 01:25:49 how so? Oct 22 01:26:03 yes read from cache on startup Oct 22 01:27:02 hmm Oct 22 01:27:03 trying Oct 22 01:27:16 i mean... i just copy first boot dump and record it , and everytime os starts , it just resumes first recorded dump Oct 22 01:27:29 lim_: do not change it ARCH=arm CROSS_COMPILE=arm-whatever- make modules (e.g. one more argument to point it to the build tree) Oct 22 01:27:34 no lose time on starting apps... Oct 22 01:27:52 zecke, but I already built a .ko like this Oct 22 01:27:56 JDuke128, that sounds like hibernate to me Oct 22 01:28:07 and it didn't work on angstrom Oct 22 01:28:55 grg , hibernate yes but on OS boots , it shows startup banner again. Oct 22 01:29:19 hibernate but just hibernates first recorded boot not current Oct 22 01:29:50 JDuke128, i've never used hibernate, but I think you could modify the scripts involved to do that Oct 22 01:30:16 it is the closest to what you want, so would be a good thing to look at as a starting point Oct 22 01:30:26 i just wanted to do fastest startup Oct 22 01:30:58 on beagleboard Oct 22 01:31:04 JDuke128, http://lwn.net/Articles/299483/ Oct 22 01:33:02 lim_: You want to cross compile a kernel module so it works on your beagleboard or such? Oct 22 01:33:10 khem, I'll see if I can find reference why we are not using 4.4 for arm Oct 22 01:33:11 yep Oct 22 01:33:30 that's exactly it Oct 22 01:33:39 zecke, I think you have a fix for my meta-toolchain issue, I coudln't fins a good error message Oct 22 01:33:40 lim_: First the kernel buildsystem is quite good in this respect. Never change a makefile, never change CC or such in it Oct 22 01:34:15 zecke_: ok Oct 22 01:34:27 lim_: The build can be controlled with "$ ARCH=XYZ CROSS_COMPILE= make ABC". ARCH and CROSS_COMPILE are environment variables Oct 22 01:34:45 lim_: ARCH can be like "arm", "mips", "x86", "alpha", "sparc"... Oct 22 01:34:54 ERROR: QA Issue with gdb-cross-sdk: non debug package contains .debug directory: gdb-cross-sdk path /work/x86_64-armv7a-sdk-angstrom-linux-gnueabi/gdb-cross-sdk-7.0-r0/install/gdb-cross-sdk/usr/local/angstrom/arm/lib64/.debug/libiberty.a Oct 22 01:34:54 ERROR: QA run found fatal errors. Please consider fixing them. Oct 22 01:35:03 zecke, is that what you fixed/ Oct 22 01:35:15 zecke_: then, where should I put this information if not in the makefile Oct 22 01:35:37 lim_: and CROSS_COMPILE points to the string in front of "gcc" e.g. "arm-unknown-" Oct 22 01:35:39 grr, mayb ethat is not the hard failure Oct 22 01:35:48 lim_: environment variable... Oct 22 01:35:59 Crofton|work: no, I fixed something like this in binutils-cross :) Oct 22 01:36:10 did you push that? Oct 22 01:36:15 zeckeÇ but the line that calls gcc is inside the makefile Oct 22 01:36:55 $CC .... ..... unknown.... Oct 22 01:38:15 lim_, you should remove that CC= line from the dm9601 Makefile Oct 22 01:38:51 zecke_Ç very well '~ll give it a try Oct 22 01:39:07 you probably need to set the KERNELDIR env variable too Oct 22 01:40:08 in fact, you might want a new Makefile altogether Oct 22 01:40:52 lim_, you should read the bits of LDD3 that pertain to building modules. i think its in the 1st or 2nd chapter Oct 22 01:41:58 or.... :) Oct 22 01:42:12 the linux kernel seems to include a dm9000 driver Oct 22 01:42:27 perhaps you should use the in kernel driver instead of the one you have? Oct 22 01:42:55 grg_: that's a thought I'll try too Oct 22 01:42:59 kinda desperate Oct 22 01:43:01 Ç) Oct 22 01:43:03 :) Oct 22 01:43:24 it could even be the same driver.... "Copyright (C) 1997 Sten Wang" Oct 22 01:45:57 grg_ : I'm trying some complicated things...as soon as I try every idea I'll report, thanks guys Oct 22 01:53:15 found a guide http://www.ailis.de/~k/archives/19-ARM-cross-compiling-howto.html Oct 22 01:57:31 seems strange that you would be in the oe channel but not using oe as your cross compilation environment Oct 22 01:58:38 I have oe here Oct 22 01:58:58 but I had way many problems learning how to create bb files Oct 22 01:59:06 too much new things to learn..not so much time Oct 22 01:59:07 rsrs Oct 22 02:00:49 you dont need to create any bb files to build a cross compiler Oct 22 02:01:19 ops, I'm using codesourcery for cross compiling Oct 22 02:02:54 ethe only thing I got from oe was angstro console image Oct 22 02:03:54 you are not looking hard enough Oct 22 02:04:12 the angstrom console image must have been built with something... Oct 22 02:04:23 you probably also got a kernel Oct 22 02:04:27 with modules Oct 22 02:04:33 yes Oct 22 02:04:47 I installed all kernel modules in an image i got from narcissus yesterday Oct 22 02:04:58 9601 not included Oct 22 02:05:21 i'm running a test on bb so i couldn't check wether 9000 is tehre **** ENDING LOGGING AT Thu Oct 22 02:59:56 2009