**** BEGIN LOGGING AT Mon Jun 30 02:59:56 2008 Jun 30 08:24:32 morning Jun 30 08:27:22 hey hrw Jun 30 09:32:22 hi florian Jun 30 09:40:46 good morning Jun 30 09:48:14 03koen 07org.oe.dev * r3a5a0193... 10/ (1 packages/vlc/vlc-davinci_0.8.6h.bb): vlc: add davinci port of vlc, it currently depends on NDA bits, but the goal is to build is against the 'public' codec-engine in the future Jun 30 10:04:15 03koen 07org.oe.dev * rc6470e7c... 10/ (1 packages/vlc/vlc-davinci_0.8.6h.bb): vlc-davinci: enable dbus en up SRCREV Jun 30 10:07:17 hm, perl-native-5.8.8 package failes when bitbaking oe on a SuSE 11.0 system Jun 30 10:07:28 it works on a SuSE 10.3 system Jun 30 10:07:53 it failes because of line 124 in ODBM_File.xs Jun 30 10:08:22 the call to dbmclose() needs a parameter Jun 30 10:09:05 however this is define in line 35, without a paramter Jun 30 10:09:41 if I replace dbmclose() with xdbmclose() in the file, it compiles Jun 30 10:10:04 don't know if this is a good idea Jun 30 10:10:22 but It should be fixed somehow. Jun 30 10:10:54 as bugs concerning perl-native in the bug tracker a half a year old Jun 30 10:11:07 is it worth to file a bug report :) Jun 30 10:11:11 ? Jun 30 10:27:13 sure Jun 30 11:09:43 hello all Jun 30 11:30:06 * * OE Bug 4396 has been created by oliver.eichler(AT)gmx.de Jun 30 11:30:08 * * perl-native_5.8.8 fails to compile Jun 30 11:30:10 * * http://bugs.openembedded.net/show_bug.cgi?id=4396 Jun 30 11:54:36 03koen 07org.oe.dev * r9a1bde94... 10/ (3 files in 3 dirs): linux-neuros: add binutils fix to get uImage that isn't 2GB Jun 30 11:54:36 03koen 07org.oe.dev * r9a1bde94... 10/ (3 files in 3 dirs): linux-neuros: add binutils fix to get uImage that isn't 2GB Jun 30 11:57:21 03koen 07org.oe.dev * r9a1bde94... 10/ (3 files in 3 dirs): linux-neuros: add binutils fix to get uImage that isn't 2GB Jun 30 11:57:21 03koen 07org.oe.dev * r9a1bde94... 10/ (3 files in 3 dirs): linux-neuros: add binutils fix to get uImage that isn't 2GB Jun 30 11:57:23 03koen 07org.oe.dev * r9a1bde94... 10/ (3 files in 3 dirs): linux-neuros: add binutils fix to get uImage that isn't 2GB Jun 30 11:57:26 03koen 07org.oe.dev * r9a1bde94... 10/ (3 files in 3 dirs): linux-neuros: add binutils fix to get uImage that isn't 2GB Jun 30 11:57:29 03koen 07org.oe.dev * r9a1bde94... 10/ (3 files in 3 dirs): linux-neuros: add binutils fix to get uImage that isn't 2GB Jun 30 11:57:32 03koen 07org.oe.dev * r9a1bde94... 10/ (3 files in 3 dirs): linux-neuros: add binutils fix to get uImage that isn't 2GB Jun 30 11:57:35 03koen 07org.oe.dev * r9a1bde94... 10/ (3 files in 3 dirs): linux-neuros: add binutils fix to get uImage that isn't 2GB Jun 30 11:57:38 03koen 07org.oe.dev * r9a1bde94... 10/ (3 files in 3 dirs): linux-neuros: add binutils fix to get uImage that isn't 2GB Jun 30 11:57:41 03koen 07org.oe.dev * r9a1bde94... 10/ (3 files in 3 dirs): linux-neuros: add binutils fix to get uImage that isn't 2GB Jun 30 11:57:52 03koen 07org.oe.dev * r9a1bde94... 10/ (3 files in 3 dirs): linux-neuros: add binutils fix to get uImage that isn't 2GB Jun 30 11:57:55 03koen 07org.oe.dev * r9a1bde94... 10/ (3 files in 3 dirs): linux-neuros: add binutils fix to get uImage that isn't 2GB Jun 30 11:59:24 03koen 07org.oe.dev * r9a1bde94... 10/ (3 files in 3 dirs): linux-neuros: add binutils fix to get uImage that isn't 2GB Jun 30 12:08:38 The md5 file for the latest mtn-database seems to be currupt. Jun 30 12:08:38 mlcy@xev:~/oe/stuff$ cat OE-this-is-for-mtn-0.37.mtn.bz2.md5 Jun 30 12:08:38 03f9c29cf3134251c7f6a1fa01970e11 OE-this-is-for-mtn-0.35.mtn.bz2 Jun 30 12:09:37 0.35 vs. 0.37 Jun 30 12:09:41 Yes. Jun 30 12:10:02 and if you are into security use mtn db check and other means of mtn to verify the integrity of the db Jun 30 14:31:19 lrg, broonie: Who of you implemented mainstone-wm97xx.c wm97xx_acc_pen_up() ? Jun 30 14:36:55 florian: lrg but I last looked at the code. Jun 30 14:37:12 * lrg hides Jun 30 14:37:29 :-) Jun 30 14:38:32 I'm guessing you have some sort of problem with it but... Jun 30 14:39:17 BTW, spyro was asking after the WM9705 ASoC stuff - looks like you and eh might be duplicating your efforts there. Jun 30 14:40:18 broonie, lrg: Line 119 - this looks wrong since 'count' is signed. Jun 30 14:40:41 broonie: ah, that's good to know indeed. Jun 30 14:40:49 ibot: seen spyro? Jun 30 14:40:55 spyro was last seen on IRC in channel #handhelds, 13h 51m 51s ago, saying: 'gnight!'. Jun 30 14:41:03 florian: Yes, that's probably wrong. Jun 30 14:41:37 broonie: I guess that reads a little bit more than necessary to empty the fifo :) Jun 30 14:41:38 Given that count will always be 16 on the way in... Jun 30 14:42:11 I've no PXA25x hardware to test on, though, so I'd rather not touch it myself. Jun 30 14:42:23 I'm building images for multiple machines under one OE tree - for the most part it looks like the TMPDIR has subdirs for arch/machine where needed - is this always the case? I notice rootfs does not have an arch/machine subdir but then I think its rebuilt every time an image is built correct? Jun 30 14:43:37 broonie: I based my platform driver on this one and I am currently trying to fix the quite odd behaviour. Currently xkbd is unusable with it... Jun 30 14:43:49 tharvey: yes Jun 30 14:44:42 florian, yes to both? I don't have to worry about changing TMPDIR to include MACHINE right? Jun 30 14:45:06 florian: At a guess it should just be a for loop. Jun 30 14:45:15 tharvey: right, I do the same here Jun 30 14:45:32 cool... thanks Jun 30 14:45:55 broonie: indeed... changed this here already. Jun 30 14:47:46 broonie: I'm testing it here right now (with a pxa255 of course). It does not seem to behave that different... so the source of the issues I see are differnt. Jun 30 14:47:59 What issues are you seeing? Jun 30 14:50:12 broonie: In X it looks like this: I seem to get some wrong coordinates from time to time. There are two different effects: Jun 30 14:51:32 First is that my pointer ends up in the correct position, but before this I seem to get an event for the location where the previous tap occured. Jun 30 14:52:37 Second is that some times I get an event that is far off from the location I touched. Jun 30 14:52:39 That should be stale coordinates in the FIFO - fixing acc_pen_up() ought to help that. Jun 30 14:53:27 broonie: That's what I thought and why I looked at this function :) Jun 30 14:53:28 Also try fiddling with the fuzz and sensitivity settings and seeing if that helps. Jun 30 14:54:25 Try using polling mode as well to narrow it down to being acceleration-releated. Jun 30 14:55:20 That's polling mode, using the interrupt it was even worse. Jun 30 14:55:44 OK, in that case the various module options are a good place to start looking. Jun 30 14:55:45 broonie: look back in git history, I was once sent a patch for similar effects, I had to revert it as it totally broke the mainstone at wolfson Jun 30 14:57:09 broonie: One thing I can confirm is that the old (non-soc) driver worked perfect. Jun 30 14:57:43 florian: Eh? You mean if you don't use ASoC then the touch driver is fine? Jun 30 14:58:01 florian: The commit XorA is talking about is 933fab0f105e3acfb97d2538109220aafec15bd3 Jun 30 14:58:07 broonie: yes, I just tried Jun 30 14:59:59 Then I'd suggest looking at the differences between the two pxa2xx-ac97 drivers. Jun 30 15:00:42 They may not have exactly the same workarounds for the issues with the PXA AC97 controllers. Jun 30 15:01:44 XorA: That patch looks well dodgy, it may return samples from different readings... Jun 30 15:02:30 broonie: the guy who wrote it swore it fixed problems like florian sees Jun 30 15:02:37 broonie: but it fscked up ts for me Jun 30 15:02:47 broonie: but maybe it shows where there is an issue Jun 30 15:02:56 broonie: I committed to git by mistake Jun 30 15:04:01 Yeah. It's trying to work around dodgy readings (and doing some... interesting duplicate suppression too). Jun 30 15:06:01 florian: In any case, if this is a difference between the ASoC and non-ASoC AC97 drivers then I'd be looking there to see what's different about the hardware access. Jun 30 15:06:29 broonie: ok, will do that now Jun 30 15:40:49 hello ... folks ... in last days i've been fighting with X. I've added the localization packages, X xmodmap for our keyboard and properly set the system to work in pt-br. However my input doesn't work fine in X. If I try to type éçáó... it shows nothing. (yes, libx11-locale is installed) Jun 30 15:43:47 does someone know what I'm doing wrong? Jun 30 15:43:54 * otavio went to have lunch Jun 30 16:13:37 broonie: The basic drivers do not differ much. I guess its the accelleration code in the device specific driver. Jun 30 16:14:12 That's identical between the two. Jun 30 16:16:10 maybe I missed some important difference between the mainstone and my device Jun 30 16:20:08 Like I say, the module parameters may be worth playing with. I'd get it working well in polling mode first then look at the interrupt stuff separately. Jun 30 16:23:39 Well... I just compared the behaviour in polling and interrupt mode again. The baheviour is different: Polling shows the effects I described and in interrrupt mode it seems to miss some events. Jun 30 16:25:47 Like I say, I'd recommend getting polling mode working first. Jun 30 16:26:12 ok :) Jun 30 16:33:00 morning all Jun 30 16:33:15 thesing: morining :) Jun 30 16:34:48 thesing: I was able to do some testing on the kexec inintramfs this weekend. I got a minimal version of linux-rp to boot :) Jun 30 16:35:54 thesing: I also noticed that the defconfig in linux-rp is set with ?= , therefore it can be overridden in local.conf Jun 30 16:36:48 hvontres|work: so you build a kernel with build-in initramfs that kexecs another kernel from flash? Jun 30 16:37:27 thesing: yup. i tried playing with inintramfs-bootmenu, but it just drops straight to inint :( Jun 30 16:39:58 hvontres|work: the bootmenu stuff has to be rewritten for use with klibc. it uses "expr". I can't add it to klibc-utils becaus klibc doen't support regular expressions. Jun 30 16:40:17 thesing: One thing I noticed is that there is no sizecheck on the kernel+initramfs Jun 30 16:40:26 thesing: ahh... Jun 30 16:40:33 hvontres|work: I think the best thing would be to rewrite the bootmenu stuff in plain C + klibc. Jun 30 16:41:37 thesing: do you know if ncurses cna compile with klibc? I wrote a small frontend for altboot in curese that could be re-used for that Jun 30 16:41:41 hvontres|work: could you open a bug for the missing sizecheck? Jun 30 16:41:55 thesing: sure Jun 30 16:42:31 It looks like we add the task for adding the initramfs after the sizecheck has run Jun 30 16:43:17 hvontres|work: I don't know. But curses has probably to many dependencies. Jun 30 16:43:38 hvontres|work: we build two kernels in one run. And only the one without initramfs is checked atm. Jun 30 16:48:06 bye Jun 30 17:22:21 hvontres|work: did you encounter any other problems with building initramfs kernels? Jun 30 17:23:39 thesing: no.The hardest part was trying to figure out what to put in my local.conf Jun 30 17:25:48 how can I echo something to the terminal from within sh script? - looking for bb.debug equivalent Jun 30 17:57:27 re Jun 30 18:10:13 ls -l Jun 30 18:10:16 ops Jun 30 18:10:39 Can someone help me with input issues? Jun 30 18:11:10 I've been fighting with that for quite a few days and I have no idea what I'm doing wrong. Jun 30 18:15:02 Did someone manage to get input with non-ascii chars working (latin1 or other)? Jun 30 18:17:45 I'm looking for a cheap, existing, minimal board that runs Linux, has a bit of ram and flash, and a way to connect a daughterboard. Needs a port to write the initial image to the flash. That's it. Anyone know anything like that? Jun 30 18:18:20 elver: gumstix? Jun 30 18:19:00 gumstix has too much. more minimal. Jun 30 18:22:12 elver: See olimex cs-e9302 or TS-7200 Jun 30 18:22:24 elver: hmm, more minimal, sorry. Jun 30 18:22:39 elver: no ethernet? Jun 30 18:22:47 elver: ngw100 is too big? Jun 30 18:22:47 Don't need it. Jun 30 18:23:27 How much is that? Jun 30 18:24:03 60$ Jun 30 18:24:10 or something Jun 30 18:24:13 Nice. Jun 30 18:24:29 likewise_: did you ever made an image that supports (éçá... like) latin1 as input? Jun 30 18:24:42 likewise_: we're in hard time to figure what is wrong Jun 30 18:24:48 otavio: no. but input to what you mean? Jun 30 18:24:56 likewise_: we have libx11-locale installed Jun 30 18:24:58 likewise_: X Jun 30 18:25:04 otavio: I never used X with OE. Jun 30 18:25:14 likewise_: like been able to write on Xterm or whatever Jun 30 18:27:00 likewise_: do you have any hint or knows someone who might help? Jun 30 18:27:24 did you ask on the ml? Jun 30 18:28:37 likewise_: I don't really know how to explain it properly Jun 30 18:28:58 likewise_: I fear people won't figure what we want Jun 30 18:35:18 likewise_: well, I tried to send a mail now. I hope it is clear... Jun 30 18:37:31 03koen 07org.oe.dev * r9a1bde94... 10/ (3 files in 3 dirs): linux-neuros: add binutils fix to get uImage that isn't 2GB Jun 30 18:37:31 03koen 07org.oe.dev * r9a1bde94... 10/ (3 files in 3 dirs): linux-neuros: add binutils fix to get uImage that isn't 2GB Jun 30 18:41:54 03koen 07org.oe.dev * rcce8eab3... 10/ (3 files in 3 dirs): linux-neuros 2.6.15: add patch to build with gcc4, still needs better defconfig and EABI patches Jun 30 18:41:57 03koen 07org.oe.dev * rcce8eab3... 10/ (3 files in 3 dirs): linux-neuros 2.6.15: add patch to build with gcc4, still needs better defconfig and EABI patches Jun 30 18:42:00 03koen 07org.oe.dev * rcce8eab3... 10/ (3 files in 3 dirs): linux-neuros 2.6.15: add patch to build with gcc4, still needs better defconfig and EABI patches Jun 30 18:42:05 03koen 07org.oe.dev * rcce8eab3... 10/ (3 files in 3 dirs): linux-neuros 2.6.15: add patch to build with gcc4, still needs better defconfig and EABI patches Jun 30 18:42:08 03koen 07org.oe.dev * rcce8eab3... 10/ (3 files in 3 dirs): linux-neuros 2.6.15: add patch to build with gcc4, still needs better defconfig and EABI patches Jun 30 18:42:11 03koen 07org.oe.dev * rcce8eab3... 10/ (3 files in 3 dirs): linux-neuros 2.6.15: add patch to build with gcc4, still needs better defconfig and EABI patches Jun 30 18:42:16 03koen 07org.oe.dev * rcce8eab3... 10/ (3 files in 3 dirs): linux-neuros 2.6.15: add patch to build with gcc4, still needs better defconfig and EABI patches Jun 30 18:42:19 03koen 07org.oe.dev * rcce8eab3... 10/ (3 files in 3 dirs): linux-neuros 2.6.15: add patch to build with gcc4, still needs better defconfig and EABI patches Jun 30 18:42:35 yikes Jun 30 18:42:41 cia seems confused Jun 30 18:47:37 Is someone using OE with mips? Jun 30 18:49:51 03koen 07org.oe.dev * r3e5f1873... 10/ (1 conf/distro/include/angstrom-2008-preferred-versions.inc): angstrom 2008: prefer uclibc 0.9.29 Jun 30 18:49:51 03koen 07org.oe.dev * r3e5f1873... 10/ (1 conf/distro/include/angstrom-2008-preferred-versions.inc): angstrom 2008: prefer uclibc 0.9.29 Jun 30 18:54:18 koen says the cia xmlrpc reports "error, not received" on any request and we retry. Jun 30 18:54:25 kergoth: ^ Jun 30 19:17:07 * * OE Bug 4397 has been created by mdarland(AT)pager.net Jun 30 19:17:09 * * syslog-ng needs libnet added to dependency Jun 30 19:17:11 * * http://bugs.openembedded.net/show_bug.cgi?id=4397 Jun 30 19:17:18 * * OE Bug 4398 has been created by likewise(AT)gmx.net Jun 30 19:17:21 * * Add three new patches to gcc-4.2.2 for avr32 Jun 30 19:17:23 * * http://bugs.openembedded.net/show_bug.cgi?id=4398 Jun 30 19:24:06 * * OE Bug 4399 has been created by  Jun 30 19:24:09 * * kernel.bbclass does not run sizecheck on kernel+initramfs Jun 30 19:24:11 * * http://bugs.openembedded.net/show_bug.cgi?id=4399 Jun 30 20:17:29 re Jun 30 21:02:36 anyone around who can check why this page (only one in the entire wiki) give a 403 error ? http://oe.linuxtogo.org/user-manual&dpage=recipes_initscripts Jun 30 21:15:53 anyone building with MACHINE=x86-prescott? I believe the conf file is simply broken in perhaps more ways than one - by defining TARGET_ARCH=i686 and including i686 in PACKAGE_EXTRA_ARCHS rootfs will fail **** ENDING LOGGING AT Tue Jul 01 02:59:56 2008