**** BEGIN LOGGING AT Tue Sep 29 02:59:58 2009 Sep 29 03:14:52 hey khem, i have committed everything with semi-descriptive messages Sep 29 03:20:09 there are like 50+ commits :( Sep 29 03:20:13 which is good or bad Sep 29 03:20:20 depending on how good your patch editor is Sep 29 06:57:09 good morning Sep 29 06:57:19 oe is really interesting!!! Sep 29 07:01:31 recalcati: of course :D Sep 29 07:28:08 jo Sep 29 07:34:44 thebohemian: I'm convincing my bosses to use it Sep 29 07:35:07 recalcati: what are you using atm? Sep 29 07:35:52 hi robert Sep 29 07:36:51 angstrom, akita, to have gcc 4.2.2 and get some fix for my original pxa270 compilation, first of all iwmmxt fir mplayer Sep 29 07:37:03 gcc 4.2.4 Sep 29 07:37:08 morning Sep 29 07:37:14 hi hrw Sep 29 07:37:14 hrw: gm Sep 29 07:37:24 thebohemian: Jalimo is now fully merged into .dev? Sep 29 07:37:32 03Robert Schuster  07org.openembedded.dev * rd017723ad8 10openembedded.git/recipes/jamvm/ (files/jamvm-1.5.3-jni_h-noinst.patch jamvm-native_1.5.3.bb): jamvm-native 1.5.3: Prevent evil jni.h from being installed. Sep 29 07:37:46 hrw: not yet. but most of it yes Sep 29 07:37:54 thebohemian: cool Sep 29 07:37:55 hrw: you need the branch right? Sep 29 07:38:11 hrw: what revision is of interest to you? (the branch to use it with poky) Sep 29 07:38:12 thebohemian: any plans to merge it to stable/2009? Sep 29 07:40:41 thebohemian: we use r700 Sep 29 07:41:10 hrw: no plans atm. I have some stuff on the agenda (qtjambi etc) and want to get everything into OE .dev first Sep 29 07:41:46 hrw: there are still some issues with java-native (sometimes cacao breaks, sometimes its jamvm) Sep 29 07:42:00 k Sep 29 07:42:10 hrw: I am currently sorting this out with those project's developers Sep 29 07:42:38 thebohemian: for next buglinux we use HEAD of meta-jalimo now Sep 29 07:42:49 woglinde: and you want to send me your icedtea6-native patches :) Sep 29 07:43:09 hm Sep 29 07:43:13 it doesnt work Sep 29 07:43:36 woglinde: yeah, this is where I will continue Sep 29 07:43:55 hrw: svn head + OE .dev? Sep 29 07:44:10 hrw: I mean Jalimo SVN head + OE .dev? Sep 29 07:44:16 thebohemian: OE stable/2009 + jalimo svn head Sep 29 07:44:24 I have icedtea6-native_1.5.bb Sep 29 07:44:58 woglinde: please sent it to me by mail (plus the patches) Sep 29 07:45:15 jupp Sep 29 07:45:18 will do now Sep 29 07:45:49 hrw: this will mean trouble since many recipes aren't in OE stable Sep 29 07:45:59 and the Jalimo SVN head either Sep 29 07:46:57 and .dev is fast moving target Sep 29 07:48:27 robert done Sep 29 07:56:21 thebohemian: so going to .dev will be suggestion? Sep 29 07:57:18 hrw: well, you could also merge the recipes you need from .dev to .stable Sep 29 07:58:39 thebohemian: I think that first I will create repo at buglabs svn which will contain them and then test and request merges Sep 29 08:03:57 hrw: good approach. Sep 29 08:04:53 hrw: btw. I merged the recipes in the order they depend on each other. so you should not have troubles pulling things in one by one Sep 29 08:05:28 ok Sep 29 08:05:39 Hi there :) Here is a challenge for ya - can anybody compile madwifi-ng?? Sep 29 08:05:44 robert what is not merged yet? Sep 29 08:05:57 woglinde: jna and icedtea Sep 29 08:06:00 thebohemian: other way would be creating own stable/2009 fork but it is ugly Sep 29 08:06:02 hasse the atheros driver from kernel are not enough? Sep 29 08:06:21 woglinde: ath5k does not provide AP mode AFAIK Sep 29 08:06:42 (at least you need a very recent kernel to have it) Sep 29 08:06:49 thebohemian: it does AP with 2.6.31 Sep 29 08:07:08 hrw: which is quite recent ;) Sep 29 08:07:10 hm I should update wrap to 2.6.31 Sep 29 08:07:35 woglinde: I'm kind of a novice - what's atheros? Sep 29 08:07:35 I have 2.6.31-rc on alix and ath5k AP worked somehow Sep 29 08:07:44 Hasse: what is madwifi-ng? Sep 29 08:07:57 hrw: "ath5k has enabled AP mode" ... I see (from http://linuxwireless.org/) Sep 29 08:08:18 hrw: don't know, but I need it to build x11-image and also console-image Sep 29 08:08:49 hasse which machine? Sep 29 08:09:04 * XorA hopes ath5k has finally reached some semblance of working these days then Sep 29 08:09:05 woglinde: i686-generic Sep 29 08:09:19 xora it works Sep 29 08:10:03 woglinde: wow! Sep 29 08:10:23 xora when did you test last time? Sep 29 08:10:29 ubuntu stupidly chose ath5k for 9.04 Sep 29 08:10:39 youd be better off with damp string Sep 29 08:10:55 hrw did you test the iw command with 2.6.31? Sep 29 08:10:57 but ubuntu seem to be the kings of stupid decisions Sep 29 08:11:04 hm Sep 29 08:11:14 I have no problems with ubuntu Sep 29 08:11:58 woglinde: check the bug trakcer :-D thousands would disagree, I had to install madwifi for a friend as ath5k just wasnt wifi capable Sep 29 08:12:29 woglinde: but its probably just certain chipsets its busted on Sep 29 08:12:37 jepp Sep 29 08:13:07 but ubuntu 9.04 is just fail all round for lack of testing Sep 29 08:13:32 woglinde & hrw: This is the error I get: http://pastebin.ca/1583959 Sep 29 08:15:39 argggssss Sep 29 08:15:54 hm Sep 29 08:16:04 generic hardcodes vesa and kdrive Sep 29 08:16:19 03Petri Lehtinen  07org.openembedded.dev * r7e769c740b 10openembedded.git/ (conf/checksums.ini recipes/python/python-flup_1.0.1.bb): Sep 29 08:16:19 python-flup: Bump to version 1.0.2 Sep 29 08:16:19 Signed-off-by: Petri Lehtinen Sep 29 08:16:31 03Graham Gower  07org.openembedded.dev * rb7fcc8134d 10openembedded.git/recipes/gnome/libsoup-2.4_2.27.91.bb: (log message trimmed) Sep 29 08:16:31 libsoup: add gnome-keyring dependency. Sep 29 08:16:31 libsoup's configure.in added a gnome-keyring check since the last version. Sep 29 08:16:31 Failure looks like this: Sep 29 08:16:31 | checking whether to build libsoup-gnome... yes Sep 29 08:16:32 | checking for GNOME_KEYRING... no Sep 29 08:16:34 | configure: error: Could not find gnome-keyring devel files. Sep 29 08:16:38 03Eric BĂ©nard  07org.openembedded.dev * rdfcc18bf49 10openembedded.git/recipes/tslib/ (tslib/tslib-pluginsld.patch tslib_1.0.bb): (log message trimmed) Sep 29 08:16:41 tslib_1.0: Apply patch from r56 to link plugins to libts Sep 29 08:16:43 The patch sent to the list was integrated into tslib Sep 29 08:16:45 in november 2008: Sep 29 08:16:47 http://svn.berlios.de/viewcvs/tslib/trunk/tslib/plugins/Makefile.am?r1=51&r2=56 Sep 29 08:16:49 Please find attached a patch to add this patch to the tslib's recipe. Sep 29 08:16:53 http://patchwork.openembedded.org/patch/1078/ Sep 29 08:16:55 03Petri Lehtinen  07org.openembedded.dev * r97d95eda70 10openembedded.git/recipes/python/python-django_1.0.2.bb: Sep 29 08:16:58 python-django: Update to work with Python 2.6 Sep 29 08:17:00 Signed-off-by: Petri Lehtinen Sep 29 08:17:09 he zecke Sep 29 08:20:28 woglinde: hey! Sep 29 08:20:42 hasse are you using ansgtroem as distro? Sep 29 08:20:53 woglinde: yeah :) Sep 29 08:21:03 woglinde: 2008.1 Sep 29 08:21:29 okay Sep 29 08:21:35 do you need wifi? Sep 29 08:22:25 hi florian Sep 29 08:22:57 woglinde: not really at the moment, no Sep 29 08:23:09 woglinde: (but would be nice in the future though...) Sep 29 08:23:35 good morning Sep 29 08:23:41 woglinde: can you tell me what is / could be wrong? Sep 29 08:23:54 hasse I could tell you a quick hack Sep 29 08:24:02 woglinde: sure! Sep 29 08:24:04 for building not the wifi stuff Sep 29 08:24:11 woglinde: oki Sep 29 08:24:27 woglinde: I do need PCI ... Sep 29 08:24:30 edit conf/distro/include/angstrom.inc and remove the wifi Sep 29 08:25:13 woglinde: ... a ha... that sounds clever! :) Sep 29 08:27:34 its only hack Sep 29 08:27:45 we should find out why madwifi dont builds Sep 29 08:28:53 woglinde: can't you get it to build as well - or haven't you tried yet? Sep 29 08:28:55 ~curse lua recipe for bugerring tmp/ Sep 29 08:28:56 May you be reincarnated as a Windows XP administrator, lua recipe for bugerring tmp/ ! Sep 29 08:31:18 03Koen Kooi  07org.openembedded.dev * r2d880e1dfe 10openembedded.git/recipes/gnome-mplayer/ (gnome-mplayer.inc gnome-mplayer_0.9.8.bb): gnome-mplayer: add 0.9.8 Sep 29 08:31:20 03Koen Kooi  07org.openembedded.dev * re273b9781e 10openembedded.git/ (3 files in 2 dirs): x264: update to a more recent snapshot Sep 29 08:43:05 03Graham Gower  07org.openembedded.dev * ra29424d444 10openembedded.git/recipes/t1lib/t1lib_5.0.2.bb: (log message trimmed) Sep 29 08:43:05 Fix t1lib dependency. Sep 29 08:43:05 Failed with: Sep 29 08:43:05 | mipsel-angstrom-linux-libtool: compile: mipsel-angstrom-linux-gcc -c Sep 29 08:43:05 -isystem/home/grg/oe/tmp/staging/mipsel-angstrom-linux/usr/include Sep 29 08:43:07 -isystem/home/grg/oe/tmp/staging/mipsel-angstrom-linux/usr/include Sep 29 08:43:11 -fexpensive-optimizations -frename-registers -fomit-frame-pointer -O2 -ggdb3 Sep 29 08:43:15 03Graham Gower  07org.openembedded.dev * r9a081b9c5a 10openembedded.git/recipes/lesstif/lesstif_0.95.0.bb: (log message trimmed) Sep 29 08:43:18 Fix lesstif compilation. Sep 29 08:43:20 Failed like this: Sep 29 08:43:22 | /bin/sh ../mipsel-angstrom-linux-libtool --tag=CC --mode=link Sep 29 08:43:24 mipsel-angstrom-linux-gcc Sep 29 08:43:26 -isystem/home/grg/oe/tmp/staging/mipsel-angstrom-linux/usr/include Sep 29 08:43:28 -fexpensive-optimizations -frename-registers -fomit-frame-pointer -O2 -ggdb3 Sep 29 08:52:59 Anyone: How do I "clean" a distro without cleaning everything? Sep 29 08:53:28 bitbake kim-and-aggy Sep 29 08:54:21 XorA: was that for me?? Sep 29 08:54:57 Hasse: heh, you need to watch british tv to get it Sep 29 08:56:01 XorA: Oh, sadly I don't - or can't... Sep 29 08:56:28 Is it possible to clean a distro? Sep 29 08:58:09 Hasse: that show you dont want to Sep 29 09:01:54 XorA: Oh, now I get it :D .. so funny :P Sep 29 09:02:00 Hasse: you can remove tmp/staging tmp/stamps tmp/work if your distro uses packaged staging Sep 29 09:02:16 Where does oe's initramfs get its network configuration from? Sep 29 09:03:15 Hasse: I guess you used google then :-D Sep 29 09:03:15 XorA: .. but there isn't a command for that? Sep 29 09:03:21 Hasse: no Sep 29 09:03:23 XorA: ;) Sep 29 09:03:27 Hasse: apart from rm Sep 29 09:03:44 XorA: apart from rm Sep 29 09:04:30 hello all, would this be a proper place to ask questions about the zipit z2 or is there a specialized channel here for it? Sep 29 09:05:04 XorA: I might as well clean everythin' then Sep 29 09:05:49 (reads topic) whoops (would like to know where to be pointed to) Sep 29 09:08:21 03Robert Schuster  07org.openembedded.dev * rd295fc9fef 10openembedded.git/recipes/jamvm/ (jamvm-initial_1.4.5.bb jamvm-initial_1.5.0.bb): jamvm-initial 1.x.y: Use 512MB for Xmx by default, install libjvm in private directory. Sep 29 09:08:22 03Robert Schuster  07org.openembedded.dev * r2676a3c919 10openembedded.git/recipes/classpath/ (3 files): classpath-native: Use --enable-tool-wrapper (increased PRs in respective BB files). Sep 29 09:11:08 Hasse: packaged staging means rebuild is lightning quick Sep 29 09:11:22 Hasse: I measured 7 minutes for openmoko instead of 4 hours Sep 29 09:12:03 Cprossu: not sure what a zippit is, but Ive heard people here and on #beagle talk about it Sep 29 09:12:36 XorA they started a own repos of oe Sep 29 09:12:40 or overlay Sep 29 09:12:42 cannt remeber Sep 29 09:13:27 these days if its not a pandora or free I tend to ignore it :-) Sep 29 09:13:31 XorA: it's basically a little pda thing based on a 300mhz xscale 270 with a color.. the kicker is it's cheap.. I picked one up for $26 today Sep 29 09:13:37 *color lcd Sep 29 09:13:53 Cprossu: qvga screen? Sep 29 09:13:55 it has the same wireless g card as in the iphone Sep 29 09:13:58 hrw: yes Sep 29 09:14:07 I would like to store the network configuration on a separate partition and make initramfs load it from there. Sep 29 09:14:17 Cprossu: oh its that beast Sep 29 09:14:23 Cprossu: I tend to ignore toys with hehe Sep 29 09:14:31 this is indeed a toy Sep 29 09:15:13 but for the price, and the idea of having a little cheap rugged that I could care less about than even my netbook was an amusing thought Sep 29 09:15:20 03Koen Kooi  07org.openembedded.dev * r39a8f9ef39 10openembedded.git/conf/distro/include/angstrom-2008-preferred-versions.inc: angstrom: switch to a more recent gtkmm Sep 29 09:15:31 03Koen Kooi  07org.openembedded.dev * r418dde34e9 10openembedded.git/recipes/opencv/ (opencv-samples_svn.bb opencv_svn.bb): Sep 29 09:15:31 opencv: update to latest svn, tweak buildsystem a bit Sep 29 09:15:31 * the autofoo buildsystem has problems with exporting symbols Sep 29 09:15:31 * the cmake buildsystem introduces bad RPATHs into the python extensions Sep 29 09:15:37 Cprossu: I counted ~20 linux devices in this room last time Sep 29 09:16:09 don't make me count my routers and other random stuff that runs on linux too heh Sep 29 09:16:43 Cprossu: I have just one router Sep 29 09:17:10 I have several here.. mostly for when I need them on a customer site.. Sep 29 09:17:42 I have a full blown pc with an outdated version of smoothwall at the front of my network though Sep 29 09:18:32 can't beat it though, uptime is right this moment 260 days, 8 hours, 42 minutes Sep 29 09:25:57 well I'll give it a go anyway.. basically the wifi firmware is proprietary for this device and thus gives a headache... the bootloader made for it extracts the firmware but I don't know where to put it... all the instructions say is "do this" -> http://sourceforge.net/apps/mediawiki/openzipit/index.php?title=Getting_Started_with_Open_Embedded_and_the_Z2#WIFI_firmware <- which is where I am stuck Sep 29 09:26:10 XorA: My server is still processing the rm command - guess it's gonna take a bit longer to build than you've measured :) Sep 29 09:27:17 Hasse: an old trick, mv tmp tmp.old ; rm -rf tmp.old & Sep 29 09:27:34 then tmp gets removed in space IO in the background Sep 29 09:32:35 morning cbrake Sep 29 09:33:04 XorA: I know, but it's still removing in the background ;) Sep 29 09:33:17 XorA: Thx anyways :) Sep 29 09:36:16 hi Liam Sep 29 10:02:59 hey Marcin Sep 29 10:27:47 hi thesing Sep 29 10:32:56 03TK, Pratheesh Gangadhar  07org.openembedded.dev * r53e7eb0ef3 10openembedded.git/recipes/gstreamer/ (2 files in 2 dirs): Sep 29 10:32:56 gst-plugin-gles: add git version Sep 29 10:32:56 * the upstream git repo is corrupted, so we employ some tricks to build it Sep 29 10:32:56 * and it has an RPATH problem... Sep 29 10:41:28 hi all Sep 29 10:41:39 II'm moving from my old distro to angstrom. By now, due to time to market pb, I'm using old distro with angstrom gcc 4.2.4 . But I have little merge problems. Sep 29 10:42:22 It is possible to create a packet, for examples for modutils, that contain all its dependencies ? Sep 29 10:47:29 recalcati: you can do it outside OE easy enough Sep 29 10:47:35 recalcati: assuming using ipks Sep 29 10:48:09 ar x name.ipk Sep 29 10:48:17 tar xzf data.tar.gz Sep 29 10:48:34 recalcati: I thought you wanted to go the other way? Sep 29 10:48:58 use opkg-cl in offline root mode, opkg-cl -o offline install modutils Sep 29 10:49:02 I need to use my old distro by now Sep 29 10:49:05 offline will then contain all its dependencies Sep 29 10:49:21 can then be tarred back up Sep 29 10:49:42 opkg-cl: command not found Sep 29 10:49:52 where is it? please Sep 29 10:50:43 recalcati: tmp/staging/HOST_ARCH/usr/bin probably Sep 29 10:51:11 ok, after lunch I'll try, THX Sep 29 10:52:25 whats the beef with tzcode? why does it download 2 versions Sep 29 10:52:33 2009m is no linger available is why I ask Sep 29 10:53:49 oh I see Sep 29 10:55:00 hi lrg Sep 29 10:55:06 hi florian Sep 29 10:55:10 * pb__ back in uk now Sep 29 10:58:22 pb__: how was the flight with young baby? Sep 29 10:59:06 ant_work: the flight was fine, she spent about half the time asleep and the rest of the time making friends with the cabin crew. Sep 29 10:59:33 very good Sep 29 10:59:44 on the first leg she seemed to find the air pressure changes a bit upsetting but she was ok coming home. Sep 29 11:02:13 cbrake, u there ? Sep 29 11:02:23 rob_w, bit early for him .... Sep 29 11:02:39 ah well. too bad for him :-) Sep 29 11:02:46 whats his timezone ? Sep 29 11:02:53 US CDT Sep 29 11:02:57 it is 0600 there Sep 29 11:03:31 yup .. i just read a nice article about embedded printing from him . Sep 29 11:03:46 brining hplip into oe .. but i dont see a recipe yet :-( Sep 29 11:06:22 rob_w: hello Sep 29 11:06:33 there he goes ^^ lol Sep 29 11:06:46 :-) Sep 29 11:06:50 still awake or a early bird ? Sep 29 11:07:30 early -- I'm usually up, but don't always start "work" until a bit later Sep 29 11:08:09 rob_w: got your email -- let me dig out the recipe ... Sep 29 11:08:13 i enjoyed your blog article about embedded printing ( leaving ps away and doing it via cairo into hplip) i would love get more beside that article Sep 29 11:08:20 yep that was me :-) Sep 29 11:08:44 i just checked the site author and wow , i remebered that irc nick name :-) Sep 29 11:09:10 and is there also something shareable in that application you mentioned ? ;-) Sep 29 11:11:07 hey pb__ Sep 29 11:11:24 rob_w: let me check in the recipes if they are clean enough, and I'd be glad to give you a high level overview of how the app works. Sep 29 11:11:44 wow cool Sep 29 11:12:06 03Koen Kooi  07org.openembedded.dev * r331fb40cd2 10openembedded.git/ (conf/checksums.ini recipes/vlc/vlc_1.0.2.bb): vlc: update to 1.0.2 Sep 29 11:15:59 rob_w: I'll check in hplib once I determine it builds -- perhaps you could update to the latest version :-) Sep 29 11:16:19 yep i can look for that then Sep 29 11:24:02 rob_w: we used the IJS client code: http://www.linuxprinting.org/ijs/ Sep 29 11:24:35 rob_w: its in OE, but we ended up just compiling the client modules directly into our app. Sep 29 11:25:05 ok i see Sep 29 11:26:13 (guess most of that is in the article) Sep 29 11:28:34 so then ... how is it i actually render some cairo stuff and pump it into that ijs which then puts it into hplip `? Sep 29 11:34:28 03Cliff Brake  07org.openembedded.dev * re6e71d56f1 10openembedded.git/ (3 files in 3 dirs): hplip: add recipe Sep 29 11:35:15 rob_w: there seems to be a configure issues where cups needs to be built before hplip, but I don't think there are any cups dependencies (I could be wrong) Sep 29 11:35:37 rob_w: anyway, we might want to add to DEPENDS, or fix the hplip configure Sep 29 11:39:39 is there a way to pull only that checkin or do i need to git pull ( i dont like to do that on my dev tree right now ) Sep 29 11:40:44 ah never mind Sep 29 11:40:46 rob_w: git fetch Sep 29 11:40:54 rob_w: git cherry-pick Sep 29 11:41:15 rob_w: in cairo, create a surface and draw on it. Then get the raw buffer from cairo and convert to the buffer format used for printing (I think they are different) Sep 29 11:42:11 rob_w: as far as starting the IJS driver, etc, check out the following source: http://svn.ghostscript.com/viewvc/trunk/gs/src/gdevijs.c?revision=5891&view=markup&sortby=author&pathrev=5891 Sep 29 11:42:30 rob_w: this should describe how to use the ijs client to communicate with the ijs driver, which is a separate process Sep 29 11:43:32 wow thanks alot Sep 29 11:44:26 rob_w: starts with ijs_invoke_server(), then a bunch of ijs_client_set_param, then ijs_client_send_data_wait(), etc ... Sep 29 11:44:31 so then my other guys just need to figure out the actual cairo drawing Sep 29 11:45:46 rob_w: yes, you just a raster image to print -- cairo is an easy way to render text and other 2d objects Sep 29 11:47:11 so i will have a line like : db data -> render via cairo -> recieve the buffer -> load that into ijs via ijs client -> will talk to hplip -> talking cups -> talking to printer or so ;-) Sep 29 11:47:19 rob_w: the other thing that is a bit tricky is you have to start the hp backend: /usr/lib/cups/backend/hp Sep 29 11:47:37 why is that tricky ? Sep 29 11:47:41 rob_w: yet another process Sep 29 11:47:47 right Sep 29 11:49:14 rob_w: there is some magic with pipes, etc. There should be code in cups or hplip you can reference. Sep 29 11:51:55 hmm Sep 29 11:51:58 kk Sep 29 11:53:25 i might come back with some more stupid questions but thanks sofar .. Sep 29 11:57:05 03Thomas Zimmermann  07shr/import * rf5183b615e 10openembedded.git/recipes/ (efl1/etk_svn.bb python/python-etk_svn.bb): Sep 29 11:57:05 etk, python-etk: etk and python etk bindings moved to trunk/OLD Sep 29 11:57:05 Signed-off-by: Klaus Kurzmann Sep 29 12:15:54 XorA: opkg-cl needs anyway a rootfs ? now I have not one. Sep 29 12:17:42 mornig all Sep 29 12:18:50 recalcati: obviously not Sep 29 12:19:09 recalcati: OE uses opkg-cl to make rootfs if using ipk, so it doesnt require one at all Sep 29 12:19:42 mmm opkg-cl -offline mydirempty install modutils Sep 29 12:19:50 Could not obtain administrative lock Sep 29 12:20:30 XorA: hey Sep 29 12:20:43 look at rootfs_ipk, there are a couple of fiddles needed, but it does work, narcisuus uses it to build images that way as well Sep 29 12:21:12 XorA: can you check linux-kexecboot on tosa from current oe.dev? Sep 29 12:21:18 I try, thx Sep 29 12:21:45 Jay7: not for a while Sep 29 12:21:52 well, np Sep 29 12:22:29 * XorA is basically not in his house this week at all Sep 29 12:22:51 XorA: btw, I've relative in Spain now, so I can grab your GTA :) Sep 29 12:22:53 ah.. well Sep 29 12:23:39 Jay7: giving it away at local lug on thurs Sep 29 12:23:51 even better :) Sep 29 12:23:59 its new student time Sep 29 12:24:03 let them suffer :-) Sep 29 12:24:04 no gta - no problems :) Sep 29 12:27:22 ;) Sep 29 12:34:29 I need to make changes in some files like /etc/fstab, /etc/networking/interfaces etc.. Ie, I wanna do this before bitbaking the image. How can this be accomplished? Sep 29 12:36:41 jovox: override them in base-files Sep 29 12:37:07 ok, just by adding my own? Sep 29 12:37:17 yes Sep 29 12:38:50 should I consider base-files/angstrom being the root of my fs? Sep 29 12:53:32 or is it only the fstab that I can override in base-files? Sep 29 12:55:04 hi Sep 29 12:55:56 Im trying to start with OE, someone can sacrify few minutes to explain me some things ? Sep 29 12:57:18 a moze po polsku ? :P Sep 29 12:57:42 voyo: just ask Sep 29 12:58:58 ok. I'v build "something" already. kernel + some filesystem images (jffs2 and/or ext2). Sep 29 12:59:02 voyo: if you want in Polish then /query me and wait 15 minutes for answers Sep 29 12:59:45 I have my own kernel (not only config, but lot of patches), which I need to integrate Sep 29 13:00:26 and some my own applications. I looked into receipes folders. but this seems soooo messy for me :-/ Sep 29 13:01:20 yes.. there is quite a lot of stuff in there :-) Sep 29 13:02:41 voyo: start with a machine definition for your target device. the easiest is to copy an existing one from a similar device and adapt it to your needs. Sep 29 13:06:10 yeah, but I dont understand everything from "recipe" files, to use it as example. ie. - when I build bootstrap-image or helloworld-image , I'v found files helloworld-image.bb and bootstrap-image.bb (I assume its starting point) Sep 29 13:07:07 both are very simple. there is line " inherit image" , which probably point to something else - I dont know where :) Sep 29 13:07:38 voyo: 1. do not try to understand all in one run Sep 29 13:07:58 voyo: 2. start from machine config for your device Sep 29 13:08:03 voyo: the "classes" directory - but that's something you should not start with... usually you don't need to touch these Sep 29 13:08:07 voyo: 2.1 share machine info with us Sep 29 13:08:39 voyo: 3. create set of patches for kernel of your device and incorporate them into recipes/linux/linux_KERNELVERSION.bb recipe Sep 29 13:09:45 its arm from conexant, but for now I'll use qemuarm for it. Sep 29 13:10:02 which arm? url maybe? Sep 29 13:16:01 conexant CX946XX , its new chip. Im not sure documentation is open already Sep 29 13:16:34 I dont have the device yet, so I playing with qemu-arm Sep 29 13:17:32 I edit the files in base-files but it didnt affect my image at all.. Anything else I need todo? Sep 29 13:19:49 so I think I have to do most of things by myself, I need "general" info , where to start with it :( Sep 29 13:21:52 voyo: so you want to use qemuarm with other kernel? Sep 29 13:22:05 for now - yes Sep 29 13:22:26 I see in qemuarm.conf (machine definition) is PREFERRED_PROVIDER_virtual/kernel = "linux-rp" Sep 29 13:22:36 this is kernel for this machine, right ? Sep 29 13:22:55 voyo: yes, it is Sep 29 13:23:13 voyo: with some work you can get something more recent booted Sep 29 13:24:54 there is few directories in recipes/linux Sep 29 13:24:56 ./linux-rp-2.6.23: Sep 29 13:24:56 ./linux-rp-2.6.24: Sep 29 13:24:56 ./linux-rp-2.6.25+2.6.26-rc4: Sep 29 13:24:56 ./linux-rp-2.6.26: Sep 29 13:25:07 which one is used for that ? Sep 29 13:25:10 voyo: don't list them all!!! Sep 29 13:25:13 pls Sep 29 13:25:28 not all. few only :P Sep 29 13:26:08 these contain the version-specific patches Sep 29 13:26:37 03Florian Boor  07org.openembedded.dev * r0834a71b91 10openembedded.git/recipes/linux/ (3 files in 2 dirs): linux-mini2440: Update to 2.6.31 and follow some more standards. Sep 29 13:31:36 morning Sep 29 13:31:59 so, assuming this line from qemuarm.conf, where is kernel Im using now ? Sep 29 13:32:54 you use linux-rp 2.6.26 iirc Sep 29 13:33:32 actually it must be time for linux-rp to die, I think pretty much all Z is upstream now Sep 29 13:33:50 XorA: that needs someone with hw to make it Sep 29 13:34:03 hrw: Im working on a collie recipe Sep 29 13:34:09 hrw: and I will on tosa as well Sep 29 13:34:14 hrw: just not this week Sep 29 13:35:02 hrw: clamshell should follow soon Sep 29 13:35:11 linux-rp.inc -> linux.inc Sep 29 13:37:25 hrw: so using linux-rp.inc file as example, I need to create similar file for myself ? Sep 29 13:38:07 voyo: no, if you want other kernel you should use linux_VERSION.bb ones Sep 29 13:43:47 PREFERRED_PROVIDER_virtual/kernel = "linux-mykernel" , then create recipes/linux/linux-mykernel.bb , using some other .bb as example. Sep 29 13:46:34 voyo: should I repeat myself for 3rd time but this time in Polish? Sep 29 13:47:09 sure , if this can help me ;) Sep 29 13:53:45 Hello, can someone please tell me how to satisfy dependencies if I have all of the needed *.ipks copied to my file system....but when installing them one-by-one is a bit of a pain as the package manager wants me to install the dependent *.ipks first. I know I have all of the needed ipks in my sdcard. Sep 29 13:55:12 Is there a way to modify /etc/opkg.conf to say something like "all the dependent *.ipks are in the sdcard.....please look there" Sep 29 13:55:35 re Sep 29 13:55:36 nhg: "opkg install this that those etc" Sep 29 13:56:40 hrw: what if I just want to do "opkg install this" where I want it to pickup "that and those" automatically from the sdcard. Sep 29 13:57:06 if they are dependednt of course Sep 29 13:58:39 I have 50 *.ipks I have to install...so not very practical. Sep 29 13:59:10 nhg: file:// is a valid url for opkg Sep 29 13:59:20 nhg: just need to generate a Packages file Sep 29 13:59:25 ipkg-make0index Sep 29 14:01:52 XorA: You mean I need to put them on some webserver....like the angstrom repository? or I can still choose to have them all in my sdcard? Sep 29 14:03:57 nhg: still have them on SD Sep 29 14:04:18 src sd-feed file:///media/card/feed/ Sep 29 14:04:31 unless that broke in the last year Sep 29 14:05:43 03Graeme Gregory  07org.openembedded.dev * r279d3f6e11 10openembedded.git/ (4 files in 3 dirs): Sep 29 14:05:43 tzcode-native_2009k.bb : 2009m is no longer available Sep 29 14:05:43 tzdata_2009n.bb : 2009m is no longer available so make 2009n Sep 29 14:05:52 03Graeme Gregory  07org.openembedded.dev * r20b087cc63 10openembedded.git/: Merge branch 'org.openembedded.dev' of git+ssh://git@git.openembedded.org/openembedded into org.openembedded.dev Sep 29 14:07:15 I see also some definition for kernel in minimal.conf (for distro), should I touch it , assuming I want to use my own kernel? (KERNEL = "kernel26" , MACHINE_KERNEL_VERSION = "2.6") Sep 29 14:08:12 all this is soo messy. too many places to look.. my eyes are bleeding :P Sep 29 14:12:37 ignore those vars Sep 29 14:12:57 03Marcin Juszkiewicz  07org.openembedded.dev * rab7a2fd1fb 10openembedded.git/recipes/netbase/ (netbase/bug/interfaces netbase_4.21.bb): netbase: on BUG eth0 is always built-in Ethernet not wifi Sep 29 14:15:02 hi kergoth Sep 29 14:15:09 hey Sep 29 14:35:00 anything changed in using collection in OE for bitbake 1.81.12 (using BB_ENV_EXTRAWHITE). bitbake doesnt seem to recognize files from oe collection in my custom one (require console-image.bb not found from a recipe in another collection) Sep 29 14:36:21 pb__: would you pls. bless this (committed patch)? http://tinyurl.com/ycl3kx7 Sep 29 14:37:13 zoid__: that's never worked the way you seem to think it does. 'require' and 'include' go based on BBPATH, along with the location of the recipe doing the including. the OE repo doesn't have 'console-image.bb' in the root of its repository, so you'd want 'require recipes/images/console-image.bb' to pick it up from the oe collection, most likely Sep 29 14:42:00 kergoth: ok , thx, thought they would search in BBFILES. used gumstix verdex boards before (with their gumstix repository), and it seemed to work their (just looked at the configs, and did find nothing special in BBPATH). it is working now with the full relative path though. Sep 29 14:42:20 BBFILES is how bitbake finds recipes, not how recipes include other files Sep 29 14:42:22 ant_work: that looks a bit weird. I'll have to see it in context to understand it properly. Sep 29 14:43:05 kergoth: thx, got it now ;) and it works perfect! Sep 29 14:43:10 great Sep 29 14:47:05 pb__: configure.ac http://pastebin.com/m6e0dea20 Sep 29 14:49:09 pb__: uname does emit the full string e.g. armv5tel Sep 29 14:50:13 *utsname Sep 29 14:53:43 03Florian Boor  07org.openembedded.dev * r93c311d422 10openembedded.git/recipes/linux/linux-mini2440_2.6.31+git.bb: linux-mini2440: Get rid of floating revision. Sep 29 14:54:40 florian, thanks :) Sep 29 14:54:51 Crofton|work: yw Sep 29 14:55:18 ant_work: I looked at the rest of the source and I guess that patch is okay Sep 29 14:55:27 pb__: the idea was using lif defined (__arm__) Sep 29 14:55:35 it still does seem weird but then the kexec sources are weird. :-} Sep 29 14:55:44 imagine for me...;) Sep 29 14:56:16 it's a funny recipe Sep 29 14:56:39 it is a bit suboptimal that all the other arches will have to do this strcmp() as well, which will never match, but I guess kexec is not exactly performance critical at that level. Sep 29 14:57:05 yea, that code should not slip in Sep 29 14:57:58 but I think sh would need it too Sep 29 14:58:55 possibly multiple sh4|sh4a|sh3|sh won't match utsname.[machine] Sep 29 14:59:16 no idea about sh kernels Sep 29 14:59:58 eh..the match should be on 2 chars, even Sep 29 15:01:41 pb__: but if the first test fails the second (inline) is not done, isnt gcc so optimized? Sep 29 15:02:22 03Marcin Juszkiewicz  07org.openembedded.dev * r4ffd506b52 10openembedded.git/conf/checksums.ini: checksums.ini: some new entries Sep 29 15:04:06 ant_work: yes, it will skip the second half of the && if the first half fails Sep 29 15:04:32 ok, at least this :) Sep 29 15:04:53 thx Sep 29 15:06:53 I've been trying to find some docs about overlay in oe but it seems that the docs are out of date. I don't see anything about it in the oe docs either. Sep 29 15:35:29 hi, wondering if someone could answer a quick question? Sep 29 15:35:57 that rather depends on the question Sep 29 15:37:03 it's innocent, I have been trying to build the qt embedded sdk as highlighted on the OE wiki but the recipe is not avaialable it would seem, can anyone shine some light on this? Sep 29 15:40:31 I see the recipe in .dev Sep 29 15:40:47 http://cgit.openembedded.org/cgit.cgi/openembedded/tree/recipes/meta/meta-toolchain-qte.bb Sep 29 15:42:02 but not in stable/2009 Sep 29 15:43:00 should this of been in my cloned tree? Sep 29 15:50:54 Hello all, Quick question about creating a recipe for a later version of jamvm. I need jamvm_1.5.3 but the latest in openembedded repo is 1.5.2 the bb file for 152 mainly just uses jamvm.inc. I tired to keep the same strategy and added a jamvm_1.5.3.bb but on trying to build it I get told nothing provides jamvm_1.5.3.bb Can anybody point me in the right direction. I've read a few things out there in the Internet but can't see to find anything usefull to this p Sep 29 15:50:54 articular case. there was me thinking it would be easy :-( Sep 29 15:54:39 Arigead: you're using bitbake wrong. it isn't 'bitbake jamvm_1.5.3.bb'. it's bitbake jamvm, or bitbake jamvm-1.5.3. read the bitbake manual, or --help Sep 29 15:59:27 kergoth: I've tried bitbake jamvm but having previously built an earlier version of jamvm bitbake jamvm thinks it's got nothing to do. I'd hoped it would pick up the later version as the jamvm dir is full of recipes. I've missed something important obviously Sep 29 15:59:55 if jamvm_1.5.3.bb is in the correct place, it will be built when you bitbake jamvm. Sep 29 16:00:04 bitbake finds recipes using BBFILES. look at what yours is set to Sep 29 16:04:26 kergoth: thanks that's what I thought should happen. I'm building for a gumstix system. I've added to the directory already being processed so BBFILES is correct. Sep 29 16:11:24 re Sep 29 16:14:54 hi wogline Sep 29 16:15:02 woglinde, even Sep 29 16:17:14 03Klaus Kurzmann  07shr/import * rfc36a64b7a 10openembedded.git/recipes/freesmartphone/misc-vapi_git.bb: Sep 29 16:17:14 misc-vapi: do not replace DEPENDS set by cornucopia.inc Sep 29 16:17:14 Signed-off-by: Klaus Kurzmann Sep 29 16:17:48 he pb Sep 29 17:41:05 03Khem Raj  07org.openembedded.dev * r8c662717cc 10openembedded.git/recipes/eglibc/ (eglibc-package.bbclass eglibc-stage.inc eglibc.inc): (log message trimmed) Sep 29 17:41:05 eglibc: Dont package timezone data Sep 29 17:41:05 * Depend on USE_LDCONFIG to package ldconfig. Sep 29 17:41:05 * Delete usr/share/zoneinfo from install tree. Sep 29 17:41:06 * Delete extra binaries installed by getconf. Sep 29 17:41:10 POSIX_V6_ILP32_OFF32 POSIX_V7_ILP32_OFF32 XBS5_ILP32_OFF32 Sep 29 17:41:12 POSIX_V6_ILP32_OFFBIG POSIX_V7_ILP32_OFFBIG XBS5_ILP32_OFFBIG Sep 29 18:02:24 re Sep 29 18:02:44 wb Sep 29 18:27:47 florian: wb Sep 29 18:59:44 anyone feel like glancing at this bbclass which implements more granular debug file packaging? Sep 29 19:00:58 sure, bring it on. sounds like a good thing. Sep 29 19:01:50 http://kergoth.pastey.net/125876 Sep 29 19:04:22 if using python 2.4, that needs another commit that adds the any() method to base.bbclass, since it was introduced in 2.5 Sep 29 19:04:24 heh Sep 29 19:04:31 I wonder if you could ditch PACKAGE_DBG_DIRS, just consider all dirs with binaries in them. Sep 29 19:04:55 or, actually, looking at what it's doing with that var, just consider all binaries in the first place :-} Sep 29 19:05:43 it just needs to determine what .debug files exist which are associated with the binaries in a given subpackage. going by actual binaries rather than a path list would definitely be an improvement, didn't feel like checking to see how package.bbclass gets the binaries Sep 29 19:06:20 * kergoth thinks __find/__package_get_files should probably be put in package.bbclass, also Sep 29 19:06:30 right, that would make sense Sep 29 19:06:40 * kergoth <3 generators Sep 29 19:07:24 if not os.path.islink(file) and not os.path.isdir(file) and isexec(file) Sep 29 19:07:33 oh, i guess i can remove that any() check entirely. adding .debug paths for non-binaries won't actually hurt anything Sep 29 19:07:35 seems to be the condition that package.bbclass uses for "is this a binary" Sep 29 19:07:44 don't need to limit it to binaries, since the code that splits to debug already does that Sep 29 19:08:10 ah right, yes Sep 29 19:11:05 http://kergoth.pastey.net/125878 .. there, thats removed. otherwise seems sane? will go ahead and mail to the list if so. at least its opt-in, since its a new class Sep 29 19:15:09 * kergoth looks over other crap in his backlog Sep 29 19:15:23 yah, looks reasonable to me Sep 29 19:15:58 okay, thanks Sep 29 19:16:06 might as well go ahead and send it to the list. I think mailman is operating at peak efficiency nowadays. Sep 29 19:19:07 kergoth: does it works ok with recipes which prepends to PACKAGES? Sep 29 19:20:03 kergoth: or with wireless-tools which defines own -dbg subpackages Sep 29 19:20:25 don't see any reason why not. will verify though. it only sets the subpackage vars if they're unset, so the recipe can override Sep 29 19:23:06 just in case I asked Sep 29 19:23:17 have a nice day/night Sep 29 19:23:19 * kergoth nods, appreciate it Sep 29 19:23:42 wireless-tools recipe can drop those dbg subs anyway Sep 29 19:23:46 * hrw -> sleep Sep 29 19:46:55 re Sep 29 20:43:29 hi bluelightning Sep 29 20:43:46 hi woglinde Sep 29 20:47:31 hi woglinde Sep 29 20:47:48 jo khem Sep 29 20:47:54 Angie is chancellor again ! Sep 29 20:48:38 jo ant Sep 29 20:49:12 hello Sep 29 20:55:30 Hi..I'm not a developer but I'm trying to build get beagleboards to communicate though MPI....and I found out about OE....I just got it and need some help writing a recipe, actually, I wrote one but I get a buildong failure...could someone help me Sep 29 20:55:31 ? Sep 29 20:56:33 please pastebin ther error Sep 29 20:56:37 ~pastebin Sep 29 20:56:38 [~pastebin] A "pastebin" is a web-based service where you can paste anything over 3 lines without flooding the channel. Here are links to a few : http://www.pastebin.com , http://pastebin.ca , http://channels.debian.net/paste , http://paste.lisp.org , http://www.rafb.net/paste Sep 29 20:56:53 ok Sep 29 20:56:57 and of course the recpie too Sep 29 20:58:41 http://pastebin.com/m17240ddd this is the recipe Sep 29 21:00:17 geez....lost the error message I'd saved...gonna come back later after trying to rebuild it Sep 29 21:00:47 thanks for the time Sep 29 21:13:05 crap, I wanted to see the error :) Sep 29 21:14:08 crofton? Sep 29 21:14:12 hi robert Sep 29 21:14:45 I have been emailing him about MPI pacakges Sep 29 21:15:01 he will comeback later Sep 29 21:15:06 do you already have one? Sep 29 21:18:15 no Sep 29 21:18:39 we tried openmpi, but it needs some assmebly done for armv7 atomic ops :) Sep 29 21:18:50 so we are trying another Sep 29 21:18:57 hehe Sep 29 21:19:03 atomic ops isnt that hard Sep 29 21:19:19 yeah Sep 29 21:23:23 crofton you could try the recipe meanwhile Sep 29 21:24:55 hmm, he didn't past the recipe Sep 29 21:25:09 < Saul_> http://pastebin.com/m17240ddd this is the recipe Sep 29 21:25:57 entries for checksum.ini is missing Sep 29 21:26:16 ok Sep 29 21:26:23 we can fix that later :) Sep 29 21:26:35 yes Sep 29 21:30:41 ok building Sep 29 21:31:11 well parsing :) Sep 29 21:41:19 http://tinderbox.openembedded.net/public/logs/task/2990188.txt Sep 29 21:41:24 this ring a bell with nayone/ Sep 29 21:42:30 hm I have to look at the .m4 file Sep 29 21:43:44 yeah Sep 29 21:43:47 type in m4 Sep 29 21:43:55 AC_DEFUN(PAC_LANG_FORTRAN90, Sep 29 21:43:56 [AC_REQUIRE([PAC_PROG_F90]) Sep 29 21:43:56 define([AC_LANG], [FORTRAN90])dnl Sep 29 21:44:07 hmm Sep 29 21:44:24 maybe not .... Sep 29 21:44:55 hm Sep 29 21:45:06 the svn version looks diffrent Sep 29 21:45:35 AC_DEFUN([PAC_PROG_F90],[AC_PROG_F90]) Sep 29 21:45:35 AC_DEFUN([PAC_LANG_FORTRAN90],[AC_LANG_PUSH(Fortran 90)]) Sep 29 21:46:09 hmm Sep 29 21:46:47 latest version is mpich2-1.1.1p1.tar.gz] Sep 29 21:46:53 http://www.mcs.anl.gov/research/projects/mpich2/downloads/index.php?s=downloads Sep 29 21:47:40 hm Sep 29 21:47:41 MPICH2 currently does not support heterogeneous platforms. Users who need to run applications on such platforms are advised to use MPICH-1 instead. Sep 29 21:50:20 as long as you run beagle only you are ok Sep 29 21:50:41 *g* Sep 29 21:51:35 anyone had issues with libnet? Sep 29 21:54:06 http://tinderbox.openembedded.net/public/logs/task/2990332.txt Sep 29 21:55:01 crofton whats autom4te.cache/traces.0t lookslike? Sep 29 21:55:54 only requests in that dir Sep 29 21:56:24 orbarron: what error? Sep 29 21:58:28 XorA|gone, collie already uses the generic linux recipe Sep 29 21:58:47 running a beagleboard-demo-image but getting stuck on libnet_1.1.2.1.bb looks like fetch failed Sep 29 21:59:33 crofton could you try to remove config.guess config.sub install-sh in confdb dir? Sep 29 22:00:02 hm Sep 29 22:00:03 noe Sep 29 22:00:08 that wouldnt work Sep 29 22:00:45 If I have some free time, I'll look tomorrow Sep 29 22:01:04 hm I think thats the problem builtin(include,aclocal_am.m4) Sep 29 22:01:11 in aclocal.m4 Sep 29 22:01:42 XorA|gone, just as a note. I read about removal of linux-rp Sep 29 22:02:29 thesing: I knew that, needs updated to 2.6.31 unless you snuck in and did it Sep 29 22:03:43 not yet. I want to push some collie patches upstream first. Sep 29 22:04:27 thesing: any luck yet with my SD problems? Sep 29 22:04:52 did some cards not work? Sep 29 22:05:01 thesing: that all cards did not work Sep 29 22:05:21 thesing: would get about 5s into boot then just return errors Sep 29 22:05:25 Ah. I remember. Sep 29 22:05:46 thesing: no hassles if you never got to look at it, we can retest with 2.6.31 Sep 29 22:06:43 thesing: it would just be fun to have working collie again :-) Sep 29 22:07:30 the current oe kernel should have everything but sound working. (If I remember correctly) Sep 29 22:07:55 collie .. is it strongarm ? Sep 29 22:08:00 but I did have some problems booting from sd-card. Sep 29 22:08:02 yes. Sep 29 22:08:18 oh you oabi folks :) Sep 29 22:08:56 btw. there is a glibc patch missing from oe for oabi. (some #include ) Sep 29 22:08:57 is it sa1110 Sep 29 22:09:13 thesing: Ive got a fix for that I keep meaning to email for Ack as its core Sep 29 22:09:24 thesing: Ill try and do it tomorrow Sep 29 22:09:27 XorA|gone: post it I will review it Sep 29 22:09:37 khem, I think so. Sep 29 22:09:44 khem: its actually a patch that looks accidently dropped Sep 29 22:09:54 yeah I remember that patch Sep 29 22:09:57 khem: the patch is already in OE, its just not applied Sep 29 22:10:04 right Sep 29 22:10:11 are you using glibc or eglibc Sep 29 22:10:18 glibc Sep 29 22:10:32 hehe Sep 29 22:10:36 I am oabi too Sep 29 22:10:47 oh oabi++ Sep 29 22:10:53 reminds me to try build an image for simpad Sep 29 22:11:09 although when Angstrom moves to gcc 4,4 I think we can drop oabi completely Sep 29 22:11:11 XorA|gone: I think I took care of it for eglibc Sep 29 22:11:23 woglinde, my simpad has flash problems and my jtag cables don't work for some reason.. Sep 29 22:11:29 but glibc I dont give a lot of attention to it anymore Sep 29 22:11:31 XorA|gone, why? Sep 29 22:11:47 thesing luckily I could worked around my flash problems Sep 29 22:11:54 may be the cables are old enough and broken ;) Sep 29 22:11:54 thesing: gcc 4,4 + binutils 2.19.X can do eabi ish on armv4 Sep 29 22:12:08 thesing: binutils automatically drops the bx Sep 29 22:12:12 yes Sep 29 22:12:14 zecke and I removed the problematic section from bootloader ones Sep 29 22:12:37 xora yes Sep 29 22:12:37 probably should shift angstrom to eglibc be default Sep 29 22:12:50 woglinde: what? Sep 29 22:12:57 yes you should :) Sep 29 22:12:58 woglinde, how? Sep 29 22:13:03 binutils gcc-4.4 arm4 eabi Sep 29 22:13:17 tehsing dont ask its long time since then Sep 29 22:13:24 * XorA|gone will have to talk to koen about the next Angstrom red flag day Sep 29 22:14:44 actually its the interworking that was problem Sep 29 22:14:58 XorA|gone, I think we should have a better x11-image for angstom. Not gpe but some newer stuff. Sep 29 22:15:08 so now gcc can generate armv4 compatible interworking code and comply to eabi Sep 29 22:15:17 Maybe steal something from openmoko or so. Sep 29 22:15:47 thesing: resurecting om2007 stack would be good Sep 29 22:15:55 thesing: OH did a good job on that Sep 29 22:16:01 thesing: OM pissed it away Sep 29 22:16:56 I think some environment with an alive upstream would be better. Sep 29 22:19:00 thesing: e17 then Sep 29 22:19:26 is collie fast enough for it? Sep 29 22:20:55 thesing: gotta be faster than gta02 :-) Sep 29 22:23:11 hildon? ;) Sep 29 22:23:56 XorA|gone: this is the patch http://gcc.gnu.org/ml/gcc-patches/2008-09/msg00048.html Sep 29 22:24:04 that enable eabi for armv4 Sep 29 22:24:33 in essence with gcc 4.4.1 we can get rid of oabi Sep 29 22:25:16 khem: that agrees with what I was reading anyway, which is good Sep 29 22:25:46 here is binutils part http://sourceware.org/ml/binutils/2008-02/msg00196.html Sep 29 22:27:29 time for some sleep. Night all Sep 29 22:27:37 gn Sep 29 22:29:08 so everybody who is interested in continued support for armv4 in oe should seriously consider moving to gcc 4.4.1+binutils 2.19+ eabi for their machines Sep 29 22:29:48 I would still think that uclibc is not armv4 eabi ready ( the asm parts) Sep 29 22:30:04 but we can easily fix that Sep 29 22:42:34 khem: hi. I could solve the linker options, but I still don't understand the Makefile :/ CFLAGS and LDFLAGS mixed together.. Sep 29 22:42:44 (klcc) Sep 30 00:21:10 03Andrea Adami  07org.openembedded.dev * r62a3c30d2f 10openembedded.git/recipes/kexec/files/kexec-tools-2-klibc.patch: Sep 30 00:21:10 kexec-tools-klibc-static_2.0.1: set LD options for static. Fix Sep 30 00:21:10 fuzzyness. WIP Sep 30 00:25:25 how does one go about getting a personal git tree? i have about 10 recipes to submit (updates + new). Sep 30 00:31:17 (is it more than just "git clone"?) Sep 30 01:05:33 you can make branches in your clone Sep 30 01:05:42 then extarct the patches and submit them Sep 30 01:07:49 crofton: ok thanks. i thought it was that simple, but the FAQ on OE's site made it sound more complicated. Sep 30 01:08:14 heh Sep 30 01:08:32 do you know a Ko Takimazawa? Sep 30 01:08:53 why do i feel like i'm being setup for a knock-knock joke ... Sep 30 01:09:20 no i do not Sep 30 01:09:23 well, I stalk people habitually Sep 30 01:09:37 excellent hobby, so long as the cops don't interfere Sep 30 01:09:48 and there is a match in linkedin for your name with a friend f mine Sep 30 01:09:59 well the callsign is easy Sep 30 01:10:06 k4ep btw Sep 30 01:10:15 bizarre Sep 30 01:10:16 nice to meet you :-) Sep 30 01:10:29 yeah linkedin has really proved that there are only 3 degrees of separation in the world. Sep 30 01:11:59 how good are your git skillz Sep 30 01:12:27 i'm a novice. i can clone a tree and get upstream updates, but that's it. i primarily work in svn. Sep 30 01:12:35 looking forward to learning :-) Sep 30 01:12:45 ah Sep 30 01:12:50 yeah I came from that Sep 30 01:12:59 I do not like to go back now ... Sep 30 01:13:54 others have expressed the same feeling. i don't quite understand the attraction ... the versioning seems more complex. Sep 30 01:14:06 certainly more flexible Sep 30 01:14:17 well, it is easier to do local work and get it organized before sending back upstream Sep 30 01:14:28 we are starting to use it with gnuradio Sep 30 01:14:50 http://gnuradio.org/trac/wiki/DevelopingWithGit Sep 30 01:15:03 is an example of how you can publish your work Sep 30 01:15:21 nice, thanks! that'll definitely help get the recipes started. Sep 30 01:15:22 in that case their is a central gatekeeper controlling the main repo though Sep 30 01:15:49 we could do the same, but the load on the gatekeepers woudl be huge Sep 30 01:16:04 and we do not have someone who could do that job, without going broke :) Sep 30 01:16:32 haha, yeah the money is critical to maintain Sep 30 01:17:23 i think the socketcan folks are transitioning over to git, too. i've been reading all the pro/con arguments on their mailing list ... sounds like git will win out, mostly because the kernel uses it. Sep 30 01:18:15 the central server model is troublesome Sep 30 01:18:42 I dread having to work on projects in svn now, because I can easily do my work and create patches to test in OE Sep 30 01:20:58 does OE have a policy about using the stable or devel trees for recipe patches? Sep 30 01:22:55 submit against either tree Sep 30 01:23:13 sounds good to me. :-) thanks for the help! Sep 30 01:23:16 basically stable only accepts patches with two other devs agreeing it shoudl go in Sep 30 01:23:33 and we do not let "major disturbances" in Sep 30 01:23:56 dev is much loser Sep 30 01:24:13 of course you can wake up and find some ugly change slipped in ..... Sep 30 01:24:15 looser Sep 30 01:24:22 something like that :) Sep 30 01:24:56 makes sense. there are some older packages that i've created updates for (notable apache, msmtp, inetutils, openldap, and tzcode/data. figured others might be interested. Sep 30 01:24:57 zecke, gm Sep 30 01:25:02 yeah Sep 30 01:25:09 that would be great Sep 30 01:25:39 there are lots of packages, hard for us to keep up with changes Sep 30 01:25:52 gotta run ... my hardware keeps locking up for no reason, gotta go kick it hard. :-) thank you again, crofton Sep 30 01:26:15 gn Sep 30 01:27:36 Crofton: hey, how are you? Sep 30 01:28:31 good **** ENDING LOGGING AT Wed Sep 30 02:59:57 2009