**** BEGIN LOGGING AT Mon Dec 03 02:59:56 2007 Dec 03 03:03:46 Who is 'pfalcon' in IRC? pslash is now failing to build. Dec 03 03:03:56 psplash even Dec 03 03:04:58 ;) it's been "BrokenEmbedded" and "BrokenMoko" this weekend, for sure. Dec 03 03:05:21 * mwester got angstrom to build, but he had to restart it twice after fixing problems. Dec 03 03:05:29 On a positive note, I got Unslung to build. Dec 03 03:10:00 mwester: really? are the changes needed checked in? Dec 03 03:11:14 No, no OE changes (yet). Still trying to compare the old vs new to see what has to be altered. Dec 03 03:12:09 But at least I got something to build by going backwards on both OE and the native compiler. Dec 03 03:12:34 s/compiler/toolchain/ Dec 03 03:13:29 It isn't gcc 4.1.2 that's the problem, i think. Dec 03 03:13:50 I have it building with gcc (GCC) 4.1.2 20070626 (Red Hat 4.1.2-13) Dec 03 03:14:49 It won't build on an updated system with gcc (GCC) 4.1.2 20070925 (Red Hat 4.1.2-27) Dec 03 03:15:08 What is the normal procedure for using Code Sourcery's toolchain for doing builds? Dec 03 03:15:28 Use gcc_csl? Dec 03 04:54:04 * * OE Bug 3417 has been created by autobuild(AT)openembedded.org Dec 03 04:54:06 * * dates-0.4.4-r2-do_compile Dec 03 04:54:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3417 Dec 03 06:54:53 03jeremy_laine 07org.oe.dev * r8100a72f... 10/ (1 conf/machine/mpc8313e-rdb.conf): mpc8313e-rdb: add JFFS2 arguments for rootfs on NAND flash Dec 03 07:33:55 Who is printing this from bitbake? Dec 03 07:33:55 DEBUG: calculating shlib requirements for tinylogin Dec 03 07:33:58 NOTE: Couldn't find shared library provider for libcrypt.so.1 Dec 03 07:33:59 NOTE: Couldn't find shared library provider for libc.so.6 Dec 03 07:34:57 I'm assuming it is from the package making stage of tinylogin Dec 03 07:35:04 cdm: a bitbake class Dec 03 07:35:15 cdm: grep YOURSTRING classes/* Dec 03 07:36:07 urmelZAUS: Good morning. Where should I place the lighttpd modules (in the image .bb file) you suggested on Friday. Dec 03 07:36:47 i'm so confused as to why it can't figure this out. Dec 03 07:39:24 pete: I have added in build/conf/local.conf DISTRO_EXTRA_RDEPENDS += "lighttpd lighttpd-module-indexfile lighttpd-module-dirlisting lighttpd-module-cgi lighttpd-module-staticfile" Dec 03 07:39:51 peta: Then you have to " bitbake -c rebuild task-base" Dec 03 07:42:59 urmelZAUS: Do you think placing them into RDEPENDS of the image .bb would work? Dec 03 07:45:31 so I have a libc.so.6 in a package. Dec 03 07:45:38 I have on in the staging area Dec 03 07:45:51 but the package phase cannot find it, and so it doesn't end up in the rootfs Dec 03 07:49:57 grumble Dec 03 08:24:41 morning Dec 03 08:46:12 03jeremy_laine 07org.oe.dev * r4a23c420... 10/ (1 packages/devmem2/devmem2.bb): devmem2: add package description Dec 03 09:28:35 bonjour Dec 03 09:28:56 yo Genesis Dec 03 09:29:08 hello XorA Dec 03 09:34:38 petu: I don't know. The philosophy behind them all is not not as clear as it should be (to me) Dec 03 09:48:52 morning all Dec 03 09:48:56 rwhitby: pong Dec 03 09:50:29 RP: can you think of any reason why git-native would break? Dec 03 09:50:46 I can't think of how your changes would do it, but the timing is suspicious Dec 03 09:50:47 rwhitby: I upgraded it recently :/ Dec 03 09:50:59 rwhitby: What was the failure? Dec 03 09:51:20 hang on, I'll rerun it Dec 03 09:53:04 (of course it will probably work now ...) Dec 03 09:53:24 rwhitby: That is the natural order of things :) Dec 03 09:54:10 It broke about 10 hours ago, and I've done lots since then :-) Dec 03 09:55:02 hmm - nope, that directory is the psplash failure. Dec 03 10:10:09 hmm.... Dec 03 10:10:22 where do I create a login for the wiki? Dec 03 10:10:52 I can only find the "request new password" thingy, which requires a login to already exist Dec 03 10:10:55 you doint need one Dec 03 10:11:01 ok. Dec 03 10:11:10 so there is no way to create one? Dec 03 10:11:23 RP: nope, can't recreate it now - sorry for the false alarm. Dec 03 10:11:34 there probably is hidden deep away, but you shouldnt need one to edit the wiki Dec 03 10:12:06 Ok, but then I am obviously not smart enough to use this Wiki :-( Dec 03 10:13:02 How do I add a new page? In all other wiki's I've been using, you just add a WikiWord link to the new page, click it, and voila: you are given the option to create this new page. Dec 03 10:13:05 but not here.... Dec 03 10:17:28 jeremy_laine: good morning Dec 03 10:17:30 hello all Dec 03 10:17:38 XorA, something has changed on the wiki .... Dec 03 10:17:53 I have been trying to start a FOSDEM page and can't figure out how Dec 03 10:18:17 Crofton|home: ah well, I barely use it, kick the sys admins :-) Dec 03 10:18:27 XorA, ever watch "The Aazing Race"? Dec 03 10:18:47 Amazing that is Dec 03 10:19:04 Crofton|home: the thing where they had to get eggs across the room? Dec 03 10:19:12 rwhitby: ok, np Dec 03 10:20:13 They have a couple of goths in teh current go round Dec 03 10:20:29 although they drees in pink and black Dec 03 10:22:03 Crofton|home: not seen it Dec 03 10:22:15 * XorA wonders why Crofton|home is talking goth to him :-) Dec 03 10:23:09 likewise: hello! Dec 03 10:23:24 who is admin for the wiki? Dec 03 10:23:32 likewise: I have a fixed version of the LEDs driver for MPC8313E-RDB Dec 03 10:23:39 A wiki where you cannot create new pages are really a sad story.... Dec 03 10:24:47 likewise: there were a couple of problems: 1/ the offsets for the R/Y/G leds were wrong 2/ from the docs it looks as though we can only write to the memory address, so we need to keep a static variable with the current led state Dec 03 10:25:19 likewise: one thing I'm wondering about though is : should we switch all the leds off at init? Dec 03 10:25:43 esben: you have to be registered to create pages but not have to edit existing Dec 03 10:25:55 esben: give me name of new page and I will create one Dec 03 10:26:31 BuildOnCygwin Dec 03 10:26:38 thanks :-) Dec 03 10:27:22 yo likewise Dec 03 10:27:23 hrw: isn't it possible to get an account for the wiki? Dec 03 10:27:34 hi Genesis Dec 03 10:27:57 hey mickeyl , seems there is some news about fosdem oe booth ;) Dec 03 10:27:58 hrw, I can't edit the events page .... Dec 03 10:28:30 jeremy_laine: thanks for the fixes. I don't think the LEDs have any meaning so far, but I'ld keep them enabled. Dec 03 10:28:32 fosem also posted list of main track speakers Dec 03 10:28:55 jeremy_laine: I am building a kernel and rootfs now, have to look for some 12 power brick. Dec 03 10:29:03 s/12/12 V/ Dec 03 10:29:11 * hrw is not familiar with drupal stuff Dec 03 10:29:19 likewise: ah, ehm can you do an mtn pull / mtn update? Dec 03 10:29:44 likewise: I have just committed the LEDs fix, but didn't bump the linux-2.6.23 revision Dec 03 10:30:19 jeremy_laine: no problem Dec 03 10:30:46 likewise: if you're interested I can give you my u-boot env for rootfs on NAND Dec 03 10:31:19 jeremy_laine: always interested :-) Dec 03 10:31:38 has someone ever used a v4l2 device on ARM ? Dec 03 10:33:48 likewise: http://pastebin.com/m48cd376d Dec 03 10:33:50 03jeremy_laine 07org.oe.dev * r9e08f0a0... 10/ (2 files in 2 dirs): linux-2.6.23: fix LEDs driver for mpc8313e-rdb Dec 03 10:34:30 likewise: I use a TFTP server where all the files are in an "mpc8313e-rdb" subdirectory Dec 03 10:35:01 for now I'm still using the u-boot provided by FSL on mpc8313e-rdb Dec 03 10:35:32 the device tree, kernel and rootfs are all generated by OE though Dec 03 10:39:22 Good morning Vietnaaaaaam! Dec 03 10:40:45 Any committers around? Dec 03 10:41:05 Nah, they've all been committed ;) Dec 03 10:42:10 hi Dec 03 10:44:57 hi Dmitry Dec 03 10:45:28 hi hrw Dec 03 10:47:17 I've updated the tc6393xb patches to make them more suitable for kernel submission. Do you want to review them? Dec 03 10:50:37 lumag: send them to lakml etc - there you will find better review then mine Dec 03 10:51:06 hrw: I've sent them to Ian and Dirk. And hope to submit them to lakml for review on wednesday or thursday. Dec 03 10:59:31 jeremy_laine : hi ! Dec 03 10:59:41 steliosk: hi! Dec 03 10:59:50 jeremy_laine : Did you have any time to check the mac address issue ? Dec 03 11:03:26 Can anyone please review bugs #3188 and #3343 ? Dec 03 11:03:48 We would love to get rid of our "bugfix" packages on top of OE ;) Dec 03 11:12:16 ~oebug 3188 Dec 03 11:13:32 Hi, i want to compile angstrom distribution with uclibc. I have added TARGET_OS="linux-uclibc" at local.conf and DISTRO="angstrom2007.1" but Dec 03 11:13:48 it doesn't work Dec 03 11:13:50 !oebug 3188 Dec 03 11:13:52 * * Bug 3188, Status: NEW, Created: 2007-10-21 07:07 Dec 03 11:13:53 * * antti.andreimann(AT)mail.ee: \[PATCH\] fix libx11-native package Dec 03 11:13:54 * * http://bugs.openembedded.org/show_bug.cgi?id=3188 Dec 03 11:13:56 ah Dec 03 11:14:02 could somebody help me?? Dec 03 11:14:12 !oebug 3343 Dec 03 11:14:14 * * Bug 3343, Status: NEW, Created: 2007-11-20 15:27 Dec 03 11:14:15 * * antti.andreimann(AT)mail.ee: \[PATCH\] fix libungif download location Dec 03 11:14:16 * * http://bugs.openembedded.org/show_bug.cgi?id=3343 Dec 03 11:14:38 movilinex, start by pastebin'ing the error Dec 03 11:15:52 :Crofton the problem is that it continues compiling with glibc Dec 03 11:16:10 hmm Dec 03 11:17:11 try ANGSTROM_MODE = "uclibc" Dec 03 11:17:21 thank you Dec 03 11:17:33 I do not do uclibc builds, but I see that in the gumstix stuff I ahhve .... Dec 03 11:17:34 i will try it Dec 03 11:17:38 gl Dec 03 11:19:56 * Crofton|home wonders why he is bitbaking libungif and not looking at defence slides .... Dec 03 11:21:00 hmm one of the mirrors had the source :) Dec 03 11:37:05 * * OE Bug 3419 has been created by koen(AT)openembedded.org Dec 03 11:37:07 * * qmake2-native is broken due to target dependance Dec 03 11:37:09 * * http://bugs.openembedded.org/show_bug.cgi?id=3419 Dec 03 11:49:11 The libx11-native recipes have a line "PE = "1"" Dec 03 11:49:18 shouldn't that be PR = 1? Dec 03 11:49:18 and? Dec 03 11:49:21 no Dec 03 11:49:31 PE = Package Epoch Dec 03 11:49:32 What is PE? Dec 03 11:49:34 Crofton|home: AFAIK NO. Dec 03 11:49:38 PE is Package Epoch Dec 03 11:49:41 ah Dec 03 11:49:42 thanks Dec 03 11:50:01 it's because old versions were like libx11-X11R7.0-1.0.1 Dec 03 11:50:18 without PE this is newer than libx11-1.0.1 Dec 03 11:50:35 ok, I have never stumbled across it befiore Dec 03 11:50:59 if the 1.0.3 builds for me I will go ahead and push Dec 03 11:51:25 great Dec 03 11:54:12 org.openembedded.dev appears to have acquired multiple heads. Dec 03 11:54:26 broonie: cd into it, run mtn merge Dec 03 11:54:33 broonie: w8 a moment (10 minutes) then? Dec 03 11:54:43 broonie: or do as hrw says Dec 03 11:55:06 Last time I checked the web site said to mention it here. Dec 03 12:00:39 ARGHHHH Dec 03 12:01:00 when there will be a day when OE will get WORKING mtn server.. Dec 03 12:01:00 ? Dec 03 12:01:05 mtn: error: I/O failure while talking to peer monotone.openembedded.org, disconnecting Dec 03 12:01:14 yeah, I just had that on e.... Dec 03 12:05:44 hrw: just use monotone.nslu2-linux.org - it's always up :-) Dec 03 12:06:17 rwhitby: two questions: does it still automerge? and how often it sync? Dec 03 12:06:38 two answers: no, every 10 minutes. Dec 03 12:07:02 ok Dec 03 12:07:36 hrw, I think it was a transient issue Dec 03 12:08:00 hrw: actually, every 5 minutes. Dec 03 12:08:25 all minutes ending in 4 and 9. Dec 03 12:08:54 (unless a sync is already running) Dec 03 12:08:56 * Crofton|home wonders if all this syncing is driving it wacky ... Dec 03 12:09:31 Crofton|home: it's been like this for about a year, so I doubt it. Dec 03 12:13:31 mtn: 3 heads on branch 'org.openembedded.dev' Dec 03 12:13:34 argh Dec 03 12:14:16 my fault, I did "mtn pull; mtn merge; mtn push" and the pull failed Dec 03 12:14:19 sorry Dec 03 12:20:35 uf.. pushed Dec 03 12:27:44 can ipkg be issued as a remote command? "ssh root@10.0.0.1 'ipkg install foo.ipk'" ? Dec 03 12:30:53 I don't see why not Dec 03 12:31:54 CoreDump, i get "Installing foo (1.0-r0) to root...", some five times, and it exits with "Segmentation fault" (trying with OpenMoko on Neo1973) Dec 03 12:32:16 CoreDump, but, if I install it on the target manually (ipkg install foo.ipkg), it installs fine! Dec 03 12:32:25 lol, nice bug Dec 03 12:32:49 CoreDump, remote installation is not complete (the executable gets installed in /usr/bin); I am able to run the executable Dec 03 12:33:15 CoreDump, because installation is not completed, or some database is not updated, i am unable to uninstall the package remotely or on the target Dec 03 12:34:10 CoreDump, "ipkg list" works fine though; only with install this issue Dec 03 12:34:21 ipkg is crap but we still use it Dec 03 12:35:04 * * OE Bug 3343 has been RESOLVED (FIXED) by philip(AT)balister.org Dec 03 12:35:06 * * fix libungif download location Dec 03 12:35:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3343 Dec 03 13:05:55 03hrw 07org.oe.dev * raf4d0cf9... 10/ (1 packages/perl/libxml-parser-perl_2.34.bb): libxml-parser-perl: add expat dependency (from Poky) Dec 03 13:06:03 03antti.andreimann 07org.oe.dev * r97d05555... 10/ (1 packages/libungif/libungif_4.1.3.bb): libungif : Update source URL per bug #3343. Dec 03 13:06:15 03antti.andreimann 07org.oe.dev * r4b48cc33... 10/ (1 packages/xorg-lib/libx11-native_1.0.3.bb): libx11-native : Add version 1.0.3. Dec 03 13:06:27 03xora 07org.oe.dev * r68aa05a2... 10/ (6 files in 3 dirs): Dec 03 13:06:27 xserver-kdrive_1.3.0.0.bb : more glamo patches, have purposely not bumped Dec 03 13:06:27 PR as this isnt a released device yet. Dec 03 13:06:38 03koen 07org.oe.dev * r3e1b85fb... 10/ (1 packages/gtk-webcore/midori_0.0.12.bb): midori: add 0.12 Dec 03 13:06:46 03hrw 07org.oe.dev * r3c3da8a3... 10/ (1 packages/tasks/task-native-sdk.bb): task-native-sdk: really drop libgcc-s-dev (part of gcc now) Dec 03 13:06:56 03hrw 07org.oe.dev * r21f54c80... 10/ (1 packages/tasks/task-base.bb): Dec 03 13:06:56 task-base: build only parts required by machine/distro combo used in build (like in Poky) Dec 03 13:06:56 Idea come from Poky - no need to build pcmcia manager if machine do not have Dec 03 13:06:56 pcmcia/cf slot. As result time of build is shorter. Dec 03 13:06:56 BT/WiFi stuff needs cleaning - now it is built anyway, should be built only if Dec 03 13:06:58 machine/distro supports it or if distro use task-base-extended then also pci/usbhost Dec 03 13:07:00 needs to be checked. Dec 03 13:07:06 03xora 07org.oe.dev * r7d943741... 10/ (3 files in 3 dirs): Dec 03 13:07:06 xserver-kdrive-common : Xglamo now works well enough to be default for Dec 03 13:07:08 GTA02. Thanks koen for most of the legwork on this. Dec 03 13:33:23 r21f54c80... should be good for many people Dec 03 13:34:04 * * OE Bug 3421 has been created by razvanm(AT)musaloiu.com Dec 03 13:34:06 * * Oops on boot Dec 03 13:34:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3421 Dec 03 13:45:05 * * OE Bug 3188 has been RESOLVED (FIXED) by philip(AT)balister.org Dec 03 13:45:07 * * fix libx11-native package Dec 03 13:45:09 * * http://bugs.openembedded.org/show_bug.cgi?id=3188 Dec 03 13:46:28 ~curse psokolovsky BADLY Dec 03 13:46:29 May the fleas of a thousand camels infest your most sensitive regions, psokolovsky BADLY ! Dec 03 13:47:22 XorA: it looks like openmoko patch for psplash do not apply anymore Dec 03 13:47:40 hrw: I know trying to get round to that Dec 03 13:48:41 good morning all Dec 03 13:49:34 arse, the _angstrom patch and the _openmoko patch collide, and of course both apply for openmoko Dec 03 13:50:04 hi koen Dec 03 13:50:10 hi koen Dec 03 13:50:44 XorA: can someone in openmoko also fix desktop files for openmoko apps? insane informs what is wrong Dec 03 13:51:12 hrw: openmoko apps are written by your company, cant you just ask the guys direct? Dec 03 13:51:27 XorA: openmoko-terminal too? Dec 03 13:51:42 ~blame mickeyl then :-) Dec 03 13:51:43 * ibot blames mickeyl then :-) (and Canada) for all the evil in the world Dec 03 13:52:19 * XorA decides psplash is junk after just 5 seconds looking at source Dec 03 13:53:18 XorA: tell your boss to not use it Dec 03 13:53:19 ;) Dec 03 13:53:57 hrw: heh heh Dec 03 13:55:59 cbrake: koen just mentioned you are using openocd with a pxa270 cpu... I'm just trying to make this combination work. Dec 03 13:56:42 ~blame koen Dec 03 13:56:43 * ibot blames koen (and Canada) for all the evil in the world Dec 03 14:00:56 koen: Played with packaged staging yet? Dec 03 14:01:17 RP: not yet, too busy with uni work for large projects like that :( Dec 03 14:01:46 koen: In theory its no longer a large project, should near enough work ;-) Dec 03 14:01:49 RP, koen: what's the final determination on DISTRO_PACKAGE_MANAGER? Dec 03 14:02:22 rwhitby: This is to tell the different between ipkg and dpkg+apt Dec 03 14:02:24 ? Dec 03 14:03:14 rwhitby: ok, if it also touches conf/distro/includes/angstrom-package-*.inc to keep the current behaviour Dec 03 14:04:22 RP: the idea was for the distro to define that variable, so that an image can choose to include the distro's preferred package manager without having to specifically name which one. Dec 03 14:05:38 rwhitby: My concen is that regardless of what the distro's choice there are currently only two scenarios that work - image built by rootfs_ipk with ipkg in rootfs or image built by rootfs_deb with dpkg+apt in the rootfs Dec 03 14:05:45 The latter doesn't even work properly :( Dec 03 14:06:11 koen: isn't current behaviour that ipkg and ipkg-collateral get included in angstrom rootfs regardless of those files? Dec 03 14:06:31 rwhitby: ipkg has been included in the deb images due to the postinsts Dec 03 14:06:39 rwhitby: yes, but only because .deb needs ipkg as well Dec 03 14:06:57 rwhitby: I have a .bb called run_postinsts.bb in poky which used awk to do the same thing, removing the need for ipkg Dec 03 14:07:13 koen: so where does dpkg get included in the image? Dec 03 14:07:55 afaik using rootfs_deb.bbclass, which needs to get patched to leave to the respective image Dec 03 14:08:23 Hmm - so package_deb.bbclass adds dpkg to DISTRO_EXTRA_RDEPENDS, but package_ipk.bbclass does *not* add ipkg. That seems a bit lop-sided? Dec 03 14:08:39 rwhitby: It is added in poky Dec 03 14:09:07 hrw: http://bugs.openembedded.org/show_bug.cgi?id=3422 Dec 03 14:09:08 * RP suspects a merge oversight :/ Dec 03 14:09:54 the classes shouldn't add it, or provide a clean way to remove it from things like minimal-image Dec 03 14:10:35 Agreed. I guess rootfs_*.bbclass should be the things setting PACKAGE_MANAGER Dec 03 14:10:35 in any case, DISTRO_EXTRA_RDEPENDS is not the right place to add it too, cause an image might want package support but not the rest of DISTRO_EXTRA_RDEPENDS. Dec 03 14:10:46 RP: you'll like this one: http://dominion.thruhere.net/koen/cms/fun-with-vendor-patches-part-2 Dec 03 14:11:00 however if you switch rootfs class the package managers must not be hardcoded in Dec 03 14:11:14 rwhitby: and DISTRO_EXTRA_RDEPENDS is only in task-base, right? Dec 03 14:11:27 OK, I'm going to back out of my patch, and step away, cause rootfs changes are out of my league :-) Dec 03 14:11:40 rootfs_*.bbclass changes that is. Dec 03 14:11:50 koen: yes, only in task-base. Dec 03 14:12:28 koen: I'll just bury my head in the sand on that ;-) Dec 03 14:12:28 koen: task-base looks ok here Dec 03 14:12:58 rwhitby: Its a complex change and both OE and Poky don't have it right Dec 03 14:13:56 * rwhitby reverts his local tree Dec 03 14:14:13 hrw: it started doing that since r16a637ca82811bb2723a8c4a408e5af7734ca326 Dec 03 14:14:26 hrw: did you overlook task.bbclass? Dec 03 14:14:45 argh Dec 03 14:15:14 thats at least two "arrghs" from hrw today Dec 03 14:15:17 rwhitby: Thinking about this, its actually relatively simple. Have rootfs_deb set PACKAGE_MANAGER_INSTALL ?= "dpkg apt", rootfs_ipk set PACKAGE_MANAGER_INSTALL ?= "ipkg etc." and then change the images that need a package manager to IMAGE_INSTALL =+ "${PACKAGE_MANAGER_INSTALL}" Dec 03 14:15:43 rwhitby: and merge run-postinsts from poky and add that to the debian version too Dec 03 14:15:57 simple for you perhaps :-) Dec 03 14:16:38 If we add IMAGE_FEATURES support, the addition of PACKAGE_MANAGER_INSTALL becomes a conditional of the image features containing "package-manager" Dec 03 14:18:19 I like the sound of image_features. Dec 03 14:18:39 rwhitby: Did you have a look at poky-image.bbclass? Dec 03 14:18:49 nope, I'm off to bed now. Dec 03 14:19:08 rwhitby: Probably a good move ;-) Dec 03 14:19:27 I could merge some of this stuff from poky if people wanted... Dec 03 14:21:26 It is a fairly funamental change though which makes me nervous... Dec 03 14:24:53 have to go, bye! Dec 03 14:27:15 koen: check now and close if it does not happen Dec 03 14:27:51 * koen pulls Dec 03 14:47:46 re Dec 03 14:48:02 XorA: fwiw, we'll replace psplash soon :) Dec 03 14:48:21 mickeyl: good, hopefully for something with a config file :-) Dec 03 14:48:25 you bet Dec 03 14:48:45 mickeyl: it currently doesnt compile because angstrom config collides with openmoko config Dec 03 14:49:05 ah right, someone else also found it a bit too ohand-centric :) Dec 03 14:49:21 Anybody seen this? - http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=025621f9a7c0efe1139d43d246136a0f3e8ea675 Dec 03 14:53:07 "gcc-cross.do_package" -> "external-toolchain.do_package" ?? Dec 03 14:54:26 slapin: hehe Dec 03 14:56:23 psplash really was just a quick hack and needs making more generic if others are going to use it... Dec 03 14:57:43 how different is psplash to splashy ? Dec 03 14:57:53 /leave Dec 03 14:58:57 HopsNBarley: hi Dec 03 15:00:20 03pfalcon 07org.oe.dev * r5a239919... 10/ (1 conf/distro/angstrom-2007.1.conf): angstrom-2007.1.conf: Upgrade linux-handhelds-2.6 to 2.6.21-hh19. Dec 03 15:00:41 where does external-toolchain come from? grep -rn -e .*external-toolchain.* * does not give me anything... Dec 03 15:01:06 yawn. hi likewise Dec 03 15:02:52 ade|desk: Totally different iirc what splashy is Dec 03 15:03:50 HopsNBarley: and good morning :-) Dec 03 15:13:28 Where should the PREFERRED_PROVIDER of (for example) virtual/${TARGET_PREFIX}binutils, be set? Dec 03 15:14:23 in your distro.conf Dec 03 15:17:16 does someone know a tool/way to generate and write partition table into a image file without being root/sudoers/... ? Dec 03 15:20:20 Genesis: look at bootimg.bbclass Dec 03 15:22:21 i've a look ; wrt-image.bbclass also is interessant Dec 03 15:29:38 Good morning! Dec 03 15:29:44 yo DukeOfURL Dec 03 15:30:03 Can someone help me with the openmoko build process? Dec 03 15:30:17 "*** No rule to make target 'arm-elf'" Dec 03 15:30:50 It has built /usr/src/armtoolchain, which is way outside the $OM directory. Dec 03 15:31:05 (owned by root) Dec 03 15:31:13 morning Dec 03 15:33:12 hmm, methinks some config is missing for meta-toolchain, it tries to be used by default Dec 03 15:33:31 eternal-toolchain I mean Dec 03 15:43:48 XorA: Perhaps set its DEFAULT_PREFERENCE -1 or something but this shouldn't happen :/ Dec 03 15:45:10 RP: that makes it happier Dec 03 15:45:24 RP: spoke too soon Dec 03 15:46:26 ive delete recipe for now, dont have time to debug it Dec 03 15:46:42 ok :( Dec 03 15:49:44 RP: how about raising skippackage if a var isn;t set for external-toolchain? Dec 03 15:52:09 Its an option although in theoy we shouldn't need to do that... Dec 03 15:55:22 It used to compile. The only difference is that I have some arm-elf stuff in another directory to prototype an arm7 board. Dec 03 15:55:22 XorA: since no one responded negatively to my version bump request, can you bump it? Dec 03 15:56:04 Crofton|home: RMK just confirmed that the mkimage stuff in the kernel is broken and people should build their own uImage :) Dec 03 15:56:15 heh Dec 03 15:56:19 I was wondering .... Dec 03 15:56:46 I am trying to remember how things tie in with u-boot versus kernel decompressor? Dec 03 15:57:09 do you have a link to the email? Dec 03 15:58:07 http://article.gmane.org/gmane.linux.ports.arm.kernel/36556 Dec 03 15:58:15 RMK: "I'm regretting ever merging the mkimage crap now." Dec 03 15:58:38 Crofton|work_: use the kernel decompressor, that can use MMU and caches Dec 03 15:58:45 and has optimized assembly Dec 03 16:01:09 ok so feed zImage to mkimage in bbfile Dec 03 16:01:17 03koen 07org.oe.dev * r0573649a... 10/ (30 files in 2 dirs): gcc: replace non-functional block o'python with working overrides Dec 03 16:01:21 * Crofton|work_ forgets how this all works regularly Dec 03 16:01:32 RP: there we go :) Dec 03 16:01:32 sadly, my mind is not what it used to be Dec 03 16:02:06 koen: great :) Dec 03 16:02:53 I want to check something with you guys. It looks like package.bbclass relies on the lib.ver/.list files that oe_libinstall populates to find dependancies. Is that right? Dec 03 16:11:53 re Dec 03 16:12:18 if some tool (native tool) is required to build stuff, how can I specify it in the .bb file? DEPENDS would mean that my ipk will depend on it, right? Dec 03 16:12:33 is there something like "build depends"? Dec 03 16:13:02 Jin^eLD: depends is build depends Dec 03 16:13:16 Jin^eLD: rdepends is runtime depends Dec 03 16:13:36 oh ok.. so the ipk will not depend on everything in DEPENDS? Dec 03 16:13:39 thanks Dec 03 16:13:39 koen: welcome in a world of people which do not know python Dec 03 16:13:41 Jin^eLD: So for a native tool, DEPENDS is the right option. Dec 03 16:14:00 roger that, thanks guys Dec 03 16:14:03 hi likewise, btw :) Dec 03 16:14:07 Jin^eLD: hi Dec 03 16:14:22 Jin^eLD: if the file inside the ipk are linked against foo, then foo is auto added to rdepends Dec 03 16:15:04 hrw: I have it when the lines of python outnumber the lines needed for overrides Dec 03 16:15:05 Yes, and you have to specify DEPENDS = "foo" mostly because its header files (-dev) package is needed during the build. Dec 03 16:15:34 ok.. Dec 03 16:17:13 ~lart PyQt Dec 03 16:17:14 * ibot forces PyQt to use Outlook Express Dec 03 16:17:17 ~lart PyQt again Dec 03 16:17:18 * ibot plops PyQt again into a giant vat of herring Dec 03 16:18:43 http://fitottawa.com/wordpress/wp-content/artic_troubles.jpg Dec 03 16:18:53 at least we aren't in Canada ... Dec 03 16:19:40 mickeyl: http://bugs.openembedded.org/show_bug.cgi?id=3288 Dec 03 16:19:49 Crofton|work_: I'm there Dec 03 16:20:18 those articulated buses remind me of the bus ride at FOSDEM .... Dec 03 16:20:43 articulated bus, full of nerds, going through the narrow streets Dec 03 16:20:51 the wrong way Dec 03 16:21:22 now we know to take the train .... Dec 03 16:21:29 koen: i'm no longer the resident qmake expert here. it's been too long. zecke needs to fix that Dec 03 16:25:11 Crofton|home: and to not follow koen to the wrong Uni Dec 03 16:26:23 bye Dec 03 16:34:09 florian: yes, I've used openocd with a PXA270 board: http://bec-systems.com/web/content/view/74/9/ Dec 03 16:35:05 florian: it was pretty fussy to get going, and I'm still not sure the openocd support is completely correct as I could not get it to work with a second PXA270 board. Dec 03 16:37:05 cbrake: yeah, I read this today - what I have is quite similar including the openocd configuration. But it looks like the communication with the CPU stops after reset... Dec 03 16:37:46 cbrake: http://pastebin.ca/804456 Dec 03 16:38:47 florian: there is one PXA270 power supply that must be active all through TRST sequence Dec 03 16:39:17 cbrake: Do I have any influence on this? Dec 03 16:39:38 florian: no, it must be set up correctly in HW Dec 03 16:39:48 florian: checking which one ... Dec 03 16:40:30 florian: VCC_BATT. A common mistake is to enable this power using SYS_EN, which goes away when TRST is active Dec 03 16:41:03 florian: the fact that your board is failing very early seems to indicate something basic is wrong -- probably reset or power. Dec 03 16:41:41 I'm having trouble to convert the image to be used with psplash. I don't know why it's getting different collors when displayed on the device Dec 03 16:41:42 cbrake: oh well... that's a regular Mainstone II board Dec 03 16:43:39 cbrake: do you know what packet size the cm-x270 usb is able to handle ? Dec 03 16:43:48 cbrake: hello et al., too ;-) Dec 03 16:44:28 nud: hi, its a regular OHCI USB controller in the PXA270, so its seems it should be able to do fairly standard sized packets Dec 03 16:45:03 Does anyone has any hint about the file format or like to avoid this psplash issue? Dec 03 16:45:13 nud: does your device work on a x86 PC? Dec 03 16:45:50 cbrake: it does Dec 03 16:46:53 nud: hmm, I'd suspect an ARM issue in the driver then. Trying to remember the common pitfalls porting x86 code to ARM ... Dec 03 16:47:34 nud: the PXA270 USB should be pretty robust. If used networking dongles, serial, storage, etc with no problems. Dec 03 16:48:24 cbrake: and what about video devices ? ;-) Dec 03 16:48:42 nud: no :-) Dec 03 16:49:02 in usbvision the current packet size seems to be something like 32 × 640×480×3 Dec 03 16:49:34 nud: but that, is why I suspect an issue in the usbvision driver, and not the PXA270 USB HW, or OHCI USB HC driver Dec 03 16:49:52 hmm gcc-cross-4.1.2 looks busted Dec 03 16:49:56 nud: that is a big packet! Dec 03 16:50:13 cbrake: the pwc driver works, and, according to its author, the only relevant difference is the huge packet size Dec 03 16:52:25 03jeremy_laine 07org.oe.dev * r8e73de6b... 10/ (1 packages/sofia-sip/sofia-sip_1.11.7.bb): sofia-sip: add missing DEPENDS on glib-2.0 and openssl Dec 03 16:53:06 03jeremy_laine 07org.oe.dev * radd871c4... 10/ (1 packages/sofia-sip/sofia-sip_1.12.7.bb): sofia-sip: package version 1.12.7 Dec 03 17:23:51 ~curse koen Dec 03 17:23:52 May the fleas of a thousand camels infest your most sensitive regions, koen ! Dec 03 17:25:05 ouch lets not nice Dec 03 17:25:13 hey Dec 03 17:25:16