**** BEGIN LOGGING AT Thu Sep 14 02:59:56 2006 Sep 14 03:00:27 psokolovsky, OK. the building continues. :) Sep 14 03:01:42 leoncamel, maybe you need to subscribe to openembedded-commits@lists.openembedded.org is you want to get commit mails (traffic is noticeable though) Sep 14 03:02:36 psokolovsky, hmm, OK. I am reading you patch from the website. :) Sep 14 03:03:42 03pfalcon 07org.oe.dev * re0dd8288... 10/ (1 packages/libpng/libpng_1.2.12.bb): Sep 14 03:03:42 libpng 1.2.12: Bump PR so seconf fix to #1326 was propogated. Sep 14 03:03:42 * Fixes paths in libpng-config. Sep 14 03:05:12 psokolovsky, I have a question about binclasses. Sep 14 03:05:23 psokolovsky, get_binconfig_mangle returns a string for "sed" using ? Sep 14 03:05:44 leoncamel, yep, it seems ;-) Sep 14 03:06:19 psokolovsky, hmm, how about your modification on libpng package ? Sep 14 03:06:32 psokolovsky, the lastest commit of yours. Sep 14 03:06:42 psokolovsky, what's that ? Sep 14 03:07:02 leoncamel, that's required for all other people who have broken version of libpng to upgrade automatically Sep 14 03:13:06 psokolovsky, here is another problem . http://rafb.net/paste/results/qqiSSi97.html Sep 14 03:13:31 psokolovsky, I builded opie-image failed. Sep 14 03:13:40 psokolovsky, could you mind have a look at it ? Sep 14 03:16:56 leoncamel, some dependency problem Sep 14 03:17:29 psokolovsky, hmm, maybe the opie-task-* stuff. any clue ? Sep 14 03:20:32 leoncamel, please report this as bug, and as workaround, try to build missing packages manually Sep 14 03:21:02 leoncamel, I really don't know what that can be, so far I haven't built opie-image yet (only gpe) Sep 14 03:22:46 http://rafb.net/paste/results/XehLqF46.html Sep 14 03:22:57 nothing builded . Sep 14 03:23:13 psokolovsky, and do you rebuild gpe-image recently ? Sep 14 03:23:24 psokolovsky, it is also failed building gpe-image. Sep 14 03:23:40 psokolovsky, pls have a look at this :http://rafb.net/paste/results/XtdCfL66.html Sep 14 03:27:09 leoncamel, what target are you building for? I don't have that lib in workdir Sep 14 03:28:11 psokolovsky, you mean the "machine" option ? Sep 14 03:28:24 leoncamel, yes, "distro" too Sep 14 03:29:15 the distro is based on generic-unstable.conf, and add my favoriate option. Sep 14 03:29:59 leoncamel, so, you should what you select, why you need it, and why it breaks ;-) Sep 14 03:30:09 angstrom doesn't build that lib Sep 14 03:30:30 psokolovsky, and the machine is based on openzaurus. Sep 14 03:30:45 should know I meant Sep 14 03:31:30 the likely cause is that you nead a proper distro or define stuff to build more precisely yourself Sep 14 03:31:33 psokolovsky, hmm, yes. Sep 14 03:32:25 psokolovsky, I will check it. Sep 14 03:33:29 psokolovsky, but the pacakge liblbxutil-X11R7.1-1.0.1. is either buggy ? right ? Sep 14 03:34:16 psokolovsky, and it should reported to bugzilla ? right ? Sep 14 03:35:00 leoncamel, yes, that would be nice Sep 14 03:35:35 ok, time to get some sleep Sep 14 03:35:58 Have a nice day/morning, people! Sep 14 03:37:27 psokolovsky, :) have a nice dream .. ;) Sep 14 05:14:06 * v8jlene is back Sep 14 05:19:15 03lenehan 07org.oe.dev * rc6382b7d... 10/ (1 packages/clamav/clamav_0.88.4.bb): Sep 14 05:19:15 clamav 0.84: Inherit binconfig so that the clamav-config script is staged, Sep 14 05:19:15 so that building applications that depend on libclamav works. Also update Sep 14 05:19:15 some descriptions. Sep 14 07:13:51 03pH5 07org.oe.dev * r75b12926... 10/ (1 packages/avahi/avahi_0.6.14.bb): Sep 14 07:13:51 avahi: add 0.6.14 Sep 14 07:13:51 - add avahi-autoipd Sep 14 07:14:01 03pH5 07org.oe.dev * r83d3a002... 10/ (14 files in 8 dirs): udev: remove needless machine specifications Sep 14 07:25:36 03clarson 07org.oe.dev * r0cb00e95... 10/ (1 classes/package.bbclass): Sep 14 07:25:36 package.bbclass: Sep 14 07:25:36 * Avoid premature use of the PKG_* variables. We don't need to make use of Sep 14 07:25:36 the debian.bbclass (or otherwise) renamed package names until the very end Sep 14 07:25:36 of the packaging. This was necessary in order to enhance my Sep 14 07:25:37 depchain/correspondantdeps stuff, and doesn't seem to harm anything. Sep 14 07:39:13 kergoth: you are landing depchain today? Sep 14 07:39:27 good morning all Sep 14 07:39:45 yup, finally got it only adding rrecommends on -dev and -dbg packages that actually exist Sep 14 07:39:52 morning Sep 14 07:46:01 * bigd0g restarts X Sep 14 07:46:59 morning Sep 14 07:47:10 hey Sep 14 07:50:16 hey XorA Sep 14 07:50:22 http://avahi.org/wiki/AboutAvahi#Distributions <- mentions OE Sep 14 07:56:34 Anyone have a clue how Firmware Loading works in OE? Sep 14 07:56:56 koen: nice Sep 14 07:57:23 blaster8: afaik via udev if you have a recent 2.6 kernel Sep 14 07:57:30 indeed Sep 14 07:57:40 and how udev works for networking devices :) Sep 14 07:57:42 morning all Sep 14 07:58:00 anyway, I'm digging around for any documentation I can find Sep 14 07:58:08 hey RP Sep 14 08:06:07 morning folks Sep 14 08:06:17 morning ade|desk Sep 14 08:06:20 http://ibot.rikers.org/stats/oe.html.gz sad that i dont even make the top 25 active talkers in here anymore :P Sep 14 08:07:23 hmm. 5th :) Sep 14 08:08:01 heh Sep 14 08:08:10 the "most used words" is funny Sep 14 08:08:25 kergoth: ah the trick is to be the top lurker Sep 14 08:08:50 RP brings happiness to the world. 26.8% lines contained smiling faces. :) Sep 14 08:08:51 hehe Sep 14 08:09:54 It seems that lrg's shift-key is hanging: 5.5% of the time he/she wrote UPPERCASE. Sep 14 08:09:54 For example, like this: RP: :) Sep 14 08:09:57 hahah Sep 14 08:12:53 are those times UTC ? Sep 14 08:18:44 kergoth: just accdently dump bitbake source in the channel, that should make your stats interesting :-) Sep 14 08:44:06 kergoth: Your change to package.bbclass has broke everything here. See http://pastebin.ca/169974 for an example. Sep 14 08:47:21 v8jlene: oops, sorry about that, a delta got in that i didnt catch. checking in the fix Sep 14 08:49:02 v8jlene: pushed, try that Sep 14 08:49:26 (too many trees) Sep 14 08:49:29 03clarson 07org.oe.dev * r48ef0434... 10/ (1 classes/package.bbclass): package.bbclass: yank an unintentional change. Sep 14 08:49:30 kergoth: no problems. I'll will applogise to the cpan class though... which isn't broken afterall ;) Sep 14 08:49:43 hehe Sep 14 08:51:08 ~irc Sep 14 08:51:13 methinks irc is Internet Relay Chat or a massive time sink Sep 14 08:51:17 ~irclog Sep 14 08:51:19 [irclog] http://www.handhelds.org:8081/irclogs Sep 14 08:51:41 kergoth: All working again now. Thanks. Sep 14 08:51:43 ~ibot Sep 14 08:51:45 it has been said that ibot is a blootbot written in perl run by TimRiker on his server. logs on http://ibot.rikers.org// , ibot, jbot, apt are all the same process. It uses sqlite, but mysql or other SQL storage is also supported. Sep 14 08:53:18 the oe packaging stuff really has a lot of cruft from old ways of doing things Sep 14 08:53:24 needs a cleanup Sep 14 08:59:36 if only &*$#&#@ uclibc would build, I'd be a lot closer to building a nice installer for Ã…ngström Sep 14 09:00:18 good morning Sep 14 09:17:44 kergoth: Its not the only thing :-/ Sep 14 09:26:30 v8jlene: hullo, seems someone else meet gettext-native breakage yesterday. :-) Sep 14 09:33:26 koen: i think that your uclibc problem is a toolchain error not the library itself Sep 14 09:34:13 the toolchain works fine when building glibc Sep 14 09:34:20 minipanda: Yeah, seem to be something gentoo have done with their java setup. Sep 14 09:34:30 the daily compile logs for uclibc svn show it is working for all arch but bfin and hppa Sep 14 09:34:35 ade|desk: and it has virtually all patches from buildroot/uclibc applied Sep 14 09:34:45 ade|desk: 'arm' is a bit vague Sep 14 09:35:07 ade|desk: OABI, EABI, vfp, fpa, softfpa, softvfp, iwmxt? Sep 14 09:35:12 true Sep 14 09:35:18 and their permutations Sep 14 09:35:33 i doubt EABI is used in the compile Sep 14 09:36:05 vfp,fpa,soft etc .. should be little difference Sep 14 09:37:23 maybe, but the buildlog doesn't mention it Sep 14 09:37:32 koen what arch are you trying for and i will take a look tomorrow Sep 14 09:37:34 so for arm the buildbot is pretty useless Sep 14 09:38:26 armv5te-linux-uclibcgnueabi Sep 14 09:38:46 DISTRO=angstrom-2007.1 MACHINE=h2200 ANGSTROM_MODE=uclibc Sep 14 09:38:51 koen in the same way i{3,4,5,6}86 is only reported as i386 in the logs Sep 14 09:40:07 * v8jlene is away Sep 14 09:40:09 with http://ewi546.ewi.utwente.nl/tmp/uclibc.diff Sep 14 09:40:10 ok so eabi arm little endian with thumb as well ... don't ask too much of your toolchain then ;) Sep 14 09:40:50 KERNEL_SOURCE="/data/build/koen/OE/build/tmp/angstrom/cross/arm-angstrom-linux-uclibcgnueabi" looks weird to me Sep 14 09:40:56 koen: USE_BX might help Sep 14 09:41:15 but the whole uclibc/ dir in OE seems to be on crack for the uninitiated Sep 14 09:41:57 i totally agree Sep 14 09:42:52 does uclibc need 'kernel sources' or 'linux-libc-headers' ? Sep 14 09:44:29 good point, pre 2.6.15 uclibc should use the linux-libc-headers, as on 2.6.18 its better to use the kernel source to make the headers first Sep 14 09:46:25 hmmm Sep 14 09:46:33 that would make uclibc depend on the kernel Sep 14 09:56:52 03MartinDietze 07org.oe.dev * r619a276e... 10/ (3 files in 2 dirs): xorg-lib: added nylon-specific solutions to bugs #1411, #1412 (can well be made common by maintainers) Sep 14 09:57:01 03MartinDietze 07org.oe.dev * rb4bf51e9... 10/ (5 files in 4 dirs): kismet: added kismet configuration for mtx-1, mtx-2 for version 2005-01-R1 Sep 14 09:57:03 03MartinDietze 07org.oe.dev * r1482c6f1... 10/ (1 classes/nylon-helpers.bbclass conf/distro/nylon.conf): nylon: added nylon-helpers.bbclass (helper functions for working on the svn trunk), updated preferred providers in nylon.conf. Sep 14 09:57:09 03MartinDietze 07org.oe.dev * r1549b228... 10/ (7 files in 4 dirs): busybox: added nylon setup for versions 1.2.0 and 1.2.1 Sep 14 09:57:20 03MartinDietze 07org.oe.dev * r14f21ea0... 10/ (1 packages/ppp-dsl/ppp-dsl_0.1-monolithic.bb): ppp-dsl: added ppp-dsl version for monolithic kernels. Sep 14 09:57:29 03MartinDietze 07org.oe.dev * ra9647afc... 10/ (4 files in 3 dirs): mtn: added configuration for cvs20041113 version of mtd-utils which works for nylon/mtx-[12]. Sep 14 09:57:39 03MartinDietze 07org.oe.dev * r178ad077... 10/ (1 packages/meta/nylon-feed.bb packages/meta/nylon-feed.inc): Sep 14 09:57:39 nylon-feed: changed some defaults, added packages Sep 14 09:57:39 * use linux-hotplug instead of hotplug-ng Sep 14 09:57:39 * use ntp instead of ntp-bin Sep 14 09:57:41 * add python Sep 14 09:57:43 * add boost Sep 14 10:01:01 oh dear Sep 14 10:01:28 the nylon people are starting to use unnecessary _nylon overrides again Sep 14 10:11:13 RP: could you have a look at http://ewi546.ewi.utwente.nl/tmp/multi-machine-weirdness.diff ? Sep 14 10:11:41 ade|desk: I pointed uclibc to the kernel sources Sep 14 10:11:48 ade|desk: lets see how far it gets Sep 14 10:12:00 s/kernel sources/staged kernel headers/ Sep 14 10:12:29 w00t Sep 14 10:12:35 Packaged contents of uclibc into /data/build/koen/OE/build/tmp/angstrom/deploy/ipk/uclibc_0.9.28+svn20060914-r2_armv5te.ipk Sep 14 10:14:51 koen: quick build an angstrom image and see if it dies :-) Sep 14 10:18:10 morning all Sep 14 10:18:19 hey do13_ Sep 14 10:18:33 hi koen Sep 14 10:28:19 ~lart stupid windows for making a different USB port == different device Sep 14 10:28:20 * ibot executes killall -HUP stupid windows for making a different USB port == different device Sep 14 10:31:39 koen: That is a bit of a problem although understandable. The kernel is machine specific and if uclibc has machine specific depends, it should arguably also be machine specific. Sep 14 10:31:56 koen: Really, we need a set of clean headers for uclibc, or a dummy kernel package Sep 14 10:32:43 koen: or, kernel staging shoud not be machine specific Sep 14 10:33:35 uclibc is built within minutes, so making it machine specific wouldn't be a big problem Sep 14 10:33:53 koen: It shouldn't be machine specific though :) Sep 14 10:34:54 I'd vote for fixing up our linux-libc-package, but that is probably a wasted effort Sep 14 10:35:39 koen: Better off making a headers package from a 2.6.18-rc kernel source Sep 14 10:36:11 I saw a 'make headers' command in some git branch earlier Sep 14 10:36:51 koen: It should be in 2.6.18-rc Sep 14 10:37:27 hey mickeyl Sep 14 10:37:49 hi guys Sep 14 10:37:49 morning mickeyl Sep 14 10:38:16 RP: are you into eds? i.e. do you know about the status of the 200601 variant? Sep 14 10:38:27 (which is the one in OE) Sep 14 10:38:43 mickeyl: I know people who are :) Sep 14 10:39:04 mickeyl: ross did the dbus port of eds and is working on getting that merged with upstream Sep 14 10:39:27 (ross being @o-hand.com) Sep 14 10:39:27 RP: lets say I have the opportunity to recommend basing a PIM application suite for a mobile phone. Those guys are already using Gtk+ 2.6. Do you think eds 200601xx is stable enough for a product? Sep 14 10:39:38 morning Sep 14 10:39:50 mickeyl: Nokia use it in the 770 for the addressbook Sep 14 10:40:05 which I must add is a piece of crap Sep 14 10:40:12 RP: *nod* Sep 14 10:40:19 it crashes after importing a vcard Sep 14 10:40:31 gconf-dbus is worrying me a bit Sep 14 10:40:33 it doesn Sep 14 10:40:46 't seem to be accepted in mainstream Sep 14 10:41:07 mickeyl: It should end up in mainstream eventually Sep 14 10:41:49 morning Sep 14 10:42:40 reenoo, morning Sep 14 10:42:46 morning reenoo Sep 14 10:43:40 mickeyl: I just asked ross and dbus-glib will be merged for 2.18 Sep 14 10:43:50 koen: The vcard issue isn't eds's fault ;-) Sep 14 10:44:03 hi leoncamel Sep 14 10:44:06 hey RP Sep 14 10:44:12 RP: could be, but the 770 is a bad example to use Sep 14 10:44:14 RP: 2.18 of what? Sep 14 10:44:24 ah, gnome Sep 14 10:44:28 yes Sep 14 10:44:33 excellent, thanks Sep 14 10:44:57 i have some more questions to ask you, but i'm afraid i need to do that via email... *cough* Sep 14 10:45:16 mickeyl: ross also mentioned the addressbook is well tested and stable, the calander a bit less though dates does work with it :) Sep 14 10:45:58 RP: i see. i've noticed TRUNK needs gtk+ 2.10 now, so we're kind of stuck with the january version Sep 14 10:45:58 If the calendar did have any issues, I suspect there'd interest in fixing them but dates should have caught any major issues Sep 14 10:46:00 mickeyl: and keep in mind that eds(-dbus) will give you syncing with opensync for free Sep 14 10:46:03 hola :) Sep 14 10:46:17 koen: *strike* eds was a very good recommendation. i like it Sep 14 10:46:28 mickeyl: That is due to the syncing/merging process with mainline eds - you can't win really :-/ Sep 14 10:46:45 koen: so we just need to come up with a solution for *cough* outlook *cough* Sep 14 10:47:02 RP: of course. you know though, when you have to make a product, you need to fixate versions Sep 14 10:47:04 didn't ximian/novell open-source the connector? Sep 14 10:48:18 koen yup put it depend on specific version of outlook web interface Sep 14 10:48:23 s/put/but Sep 14 10:49:18 ade|desk: could you have a look at http://www.openembedded.org/~koen/OE/uclibc-svn.diff ? Sep 14 10:49:34 RP: on a slightly related note... do you know how the performance situation is w/ gtk+ 2.10. is that sorted out yet? Sep 14 10:49:48 mickeyl: That's a very related note :) Sep 14 10:49:56 righto, it is Sep 14 10:50:08 mickeyl: There has been dicussion but not a lot of action. We can't get the cario people interested Sep 14 10:50:15 bummer Sep 14 10:50:43 jorn@o-hand did spend some time on it and there are ideas of how to solve the problem floating around :-/ Sep 14 10:51:14 o-hand do want to see that addressed Sep 14 10:53:45 mickeyl: It might be worth you talking to some people at o-hand, if you send me an email, perhaps I can introduce you to some people Sep 14 10:54:32 I do most things except gtk/apps here which means I'm probably not the best person to talk to about this :) Sep 14 10:55:13 Morning all Sep 14 10:55:13 RP: that would be really helpful. we might have a chance to cooperate business wise as well here :) Sep 14 10:55:27 RP: I'll prepare this mail today. thanks for the offer Sep 14 10:55:46 mickeyl: Indeed :). ok, thanks Sep 14 10:56:53 Warning: unable to open an initial console. Sep 14 10:57:09 minipanda: no /dev/console in the rootfs? Sep 14 10:57:10 but seems /dev/console was there ... Sep 14 10:57:48 *sigh* Sep 14 10:57:50 NOTE: preferred version 2.4 of glibc not available (for item virtual/arm-angstrom-linux-uclibcgnueabi-libc-for-gcc) Sep 14 10:58:11 heh Sep 14 10:58:12 that again Sep 14 10:58:13 koen: oh, no /dev/console in latest build Sep 14 10:58:15 s/again/still/ Sep 14 10:58:28 mickeyl: this has a different twist Sep 14 10:58:47 koen: so what shall i do with generic-unstable? i mean howto fix it Sep 14 10:59:11 mickeyl: PREFERRED_PROVIDER_virtual/arm-angstrom-linux-uclibcgnueabi-libc-for-gcc = "uclibc" Sep 14 11:01:40 koen, mickeyl: I just made the message a little more helpful to show why bitbake gets confused. 50% of the time it picks the wrong one which leads to weird breakage :-/ Sep 14 11:01:59 very good Sep 14 11:02:20 RP: in this case the uclibc check in glibc needed to get fixed Sep 14 11:02:52 koen: well, is that lack of whole /dev is caused by udev package? Sep 14 11:03:24 minipanda: it should be created by base-files or makedevs or something Sep 14 11:04:02 03koen 07org.oe.dev * rf7550674... 10/ (1 packages/glibc/glibc-package.bbclass): glibc: adjust check for uclibc Sep 14 11:06:24 later all Sep 14 11:06:35 03koen 07org.oe.dev * r647f04c8... 10/ (3 files in 3 dirs): angstrom: some more uclibc stuff Sep 14 11:06:57 hmm Sep 14 11:06:59 undefined symbol: XmbResetIC Sep 14 11:13:46 | Cannot load module /buildx86/tmp/work/i686-linux/gtk+-2.6.10-r2/gtk+-2.6.10/modules/input/im-xim.la: /buildx86/tmp/work/i686-linux/gtk+-2.6.10-r2/gtk+-2.6.10/modules/input/.libs/im-xim.so: undefined symbol: XmbResetIC Sep 14 11:13:46 | /buildx86/tmp/work/i686-linux/gtk+-2.6.10-r2/gtk+-2.6.10/modules/input/im-xim.la does not export GTK+ IM module API: /buildx86/tmp/work/i686-linux/gtk+-2.6.10-r2/gtk+-2.6.10/modules/input/.libs/im-xim.so: undefined symbol: XmbResetIC Sep 14 11:13:49 ever seen that? Sep 14 11:14:22 * RP hasn't Sep 14 11:14:36 some missing x depends? Sep 14 11:14:51 interestingly the corresponding image built fine for arm Sep 14 11:14:59 i'm building for MACHINE=native atm. Sep 14 11:15:08 probably some header / library confusion Sep 14 11:15:32 guess i should remove X completely from the build machine Sep 14 11:16:22 aah Sep 14 11:16:26 X_HAVE_UTF8_STRING again Sep 14 11:28:08 * mickeyl goes for full-libx11 for now Sep 14 11:28:20 what's the size saving anyway nowadays for diet-x11? Sep 14 11:39:04 koen|away: makedev or makedevs? Sep 14 11:40:40 crap, in $OESTUFF/build/tmp/rootfs/ have /dev, but in bootstrap-image.tar.gz, no! Sep 14 11:41:00 hi rp, hi hrw Sep 14 12:20:04 03rwhitby 07org.oe.dev * r006e08a6... 10/ (1 packages/ixp4xx/ixp4xx-npe_2.3.bb): ixp4xx-npe: Added support for version 2.3 microcode (but mc_grab doesn't extract it properly) Sep 14 12:25:59 hmm, hostap results in a hard lock that not even Magic Sysreq will get out of Sep 14 12:31:18 hi Sep 14 12:31:25 * hrw is in inet cafe Sep 14 12:31:46 its good to have account with 0.00/h rate Sep 14 12:37:49 this really interesting to read http://ktown.kde.org/~seli/memory/ Sep 14 12:40:02 hey. hrw. Sep 14 12:40:22 hrw, do you know why about this ? http://rafb.net/paste/results/qqiSSi97.html Sep 14 12:40:26 hrw, any clue ? Sep 14 12:41:10 leoncamel: bitbake meta-opie Sep 14 12:41:32 leoncamel: looks like BUILD_ALL_DEPS is not working now Sep 14 12:41:37 hi dirk Sep 14 12:42:00 hey Marcin Sep 14 12:43:34 cu Sep 14 12:52:51 RP: in kernel hostap seems to be shafted in linux-openzaurus_2.6.17 Sep 14 12:59:01 hola Sep 14 12:59:04 juas Sep 14 12:59:07 soy kepen? Sep 14 12:59:50 hello Sep 14 12:59:53 hola Sep 14 13:00:00 con quien hablo? Sep 14 13:00:18 puedes decirme quien eres? Sep 14 13:00:48 kepen918: English please Sep 14 13:01:07 oh sorry, i no spike english Sep 14 13:01:25 kepen918: you are from spain? Sep 14 13:03:25 yes Sep 14 13:03:30 in catalonia Sep 14 13:03:41 i from catañonia Sep 14 13:03:48 03rwhitby 07org.oe.dev * ra3a176fe... 10/ (1 packages/images/slugos-image.bb): slugos-image: reverted blaster8's changes, as they break the current builds. Sep 14 13:03:57 help, my english its very bad Sep 14 13:04:37 kepen918: sorry, I can't speak spanish Sep 14 13:05:21 bye bye Sep 14 13:15:04 03rwhitby 07org.oe.dev * rc9d0318a... 10/ (1 packages/ixp4xx/ixp4xx-npe_2.3.bb): ixp4xx-npe: fixed the extraction Sep 14 13:26:58 RP: what do you think about this: http://en.pastebin.ca/170230 (mount.sh whitelist and partitioned blkdevs) Sep 14 13:29:42 03rwhitby 07org.oe.dev * r8156b74b... 10/ (1 packages/meta/slugos-packages.bb): slugos-packages: Demoted bwmon for ucslugc as it fails to build Sep 14 13:29:57 somenody interested in a greenphone to build support in oe? Sep 14 13:39:26 i wouldn't pay for it, but i would be interested Sep 14 13:39:46 dunno whether we it will come with an open source qtopia btw. Sep 14 13:39:50 s/we// Sep 14 13:43:42 greentux: I'd interested in one for a 2.6 kernel port and I'd suspect base OE support would result :) Sep 14 13:43:44 i am a newbie in embedded. I was reading a microcontroller spec and could not understand wht RTT meant ? Sep 14 13:44:28 CAN, Sep 14 13:44:28 USB 2.0, Sep 14 13:44:28 Ethernet 10/100, Sep 14 13:44:28 RTT, Sep 14 13:44:38 oops.. sorry !! Sep 14 13:45:19 pH5: Does that script work with partitions or does a race situation result? Sep 14 13:45:41 greentux: I suspect most OE devs would be interested in a greenphone to make it really 'open' ;) Sep 14 13:46:57 XorA: It works for me? :-/ Sep 14 13:52:10 RP: ooops with hard lock here, and I dont have serial cable Sep 14 13:52:33 http://cbs5.com/topstories/local_story_256204954.html Sep 14 13:55:17 any idea where to find libiberty on a debian system? Sep 14 13:55:39 koen, ok i can organise one or perhaps two Sep 14 13:56:06 sounds good Sep 14 13:56:14 koen but which of the devs can do some useful things? Sep 14 13:56:47 greentux: that depends on what you'd find usefull Sep 14 13:57:00 koen kernel 2.6 is most important for me... Sep 14 13:57:05 greentux: I think almost every OE dev can add device support and patch the appropriate packages Sep 14 13:57:27 for kernel 2.6 I'd recommend contacting a kernel dev like RP Sep 14 13:57:28 * chouimat dreams of a device that will clean his apartment for him ... and cost lest than 100$ Sep 14 13:57:53 chouimat: what about a little vietnamese lady? :P Sep 14 13:58:25 giel: I say less than 100$ ... and this include the maintenance cost :P Sep 14 13:58:43 heh Sep 14 13:59:27 giel: I was joking ... it's just I don't feel like cleaning and do last week dishes today ... Sep 14 13:59:50 chouimat: you might have noticed i wasn't completely serious either Sep 14 13:59:59 I noticed Sep 14 14:00:03 good. Sep 14 14:01:07 mickey|bbl: libiberty.a is in binutils-dev Sep 14 14:01:34 aah Sep 14 14:01:35 thanks pH5 Sep 14 14:01:43 i'm still a debian newby Sep 14 14:01:48 i have no idea how to query the package repository Sep 14 14:01:53 koen ok will organise the devs Sep 14 14:01:58 ~lart having to use debian because of professional reasons Sep 14 14:01:58 * ibot pries having to use debian because of professional reasons's back open with a screwdriver and flashes a new bootldr to having to use debian because of professional reasons Sep 14 14:02:23 mickey|bbl: packages.ubuntu.com is your friend :) Sep 14 14:02:32 since packages.debian.org is down most of the time Sep 14 14:02:33 greentux: a greenphone for some of the Opie guys would maybe bring new life to Qt/Embedded and friends in OE Sep 14 14:02:36 koen: k Sep 14 14:02:43 koen, but nobody knows if the sources you get with the device are useful enough... Sep 14 14:02:56 mickey|bbl: yup Sep 14 14:03:03 mickey|bbl: apt-cache search Sep 14 14:03:08 greentux: I'm 100% sure the sources are either missing or crappy licensed Sep 14 14:03:19 greentux: unless I pay $$$$ Sep 14 14:03:37 XorA: thanks Sep 14 14:03:41 koen ok so i will order one community version and will report here... Sep 14 14:04:14 greentux: sounds great Sep 14 14:15:21 weirdness Sep 14 14:15:43 gcc-cross doesn't compile, and the patch to fix it is already applied Sep 14 14:30:02 koen: through binary only modules or worse? Sep 14 14:30:59 RP: it's mvista, so anything is possible Sep 14 14:31:45 koen: Hmm. mvista should know better. Plenty of people to lean on there I guess though Sep 14 14:32:30 RP: mv employs people who's only job is to backport 2.6 stuff to 2.4 Sep 14 14:32:43 so they can lock people into 2.4-mv forever Sep 14 14:32:52 and at $40k per seat.... Sep 14 14:35:09 koen: I know... Sep 14 14:36:12 It's probably cheaper to buy you a rust-free MG :) Sep 14 14:36:58 koen: Don't believe the photos, its not that bad ;-). Not that I'd object to one mind ;-) Sep 14 14:37:30 For rust free I have my nice shiny Daytona 675 :) Sep 14 14:38:18 koen: mv will use 2.4 because Linus said binary modules on 2.4 were not a GPL volation. He changed his mind for 2.6 Sep 14 14:39:00 Surely if they're illegal for one, they're illegal for the other? Sep 14 14:40:23 not if Linus(TM) says it Sep 14 14:40:27 *cough* Sep 14 14:40:52 koen: What Linus says don't matter. Its what the licence says... Sep 14 14:41:59 RP: that was a long and painful argument in lkml, Linus point was 2.4 interfaces were simple and public and therefore binary modules could claim to not be derivative, he sad that changed in 2.6 Sep 14 14:42:26 RP: but any statements along those lines are going to be used by dodgy companies true or not Sep 14 14:42:38 XorA: I know :-( Sep 14 14:43:32 and us little guys cant afford to sue normally Sep 14 14:43:51 XorA: There are few people who can :-( Sep 14 14:50:14 Hello Sep 14 14:50:21 hi sirfred Sep 14 14:51:04 pH5: hi Sep 14 14:51:24 hey sirfred Sep 14 14:51:47 RP: what kind of race conditions? I'm still fighting with the ipaq to test this (it starts loosing interrupts after the first pcmcia card removal) Sep 14 14:52:59 pH5: When you insert a card, /dev/hda would appear but /dev/hda[1-9] won't appear until after the partition scanning occurs. Also, udev might delay the device creation/reorder it etc. Sep 14 14:53:16 hi sirfreg Sep 14 14:53:23 hi sirfred even :) Sep 14 14:53:40 RP: Hi. :) Sep 14 14:53:57 RP: Did you have time to ask the framebuffer maintainer about sync ? Sep 14 14:54:29 sirfred: no, I thought you were going to do that... :-/ Sep 14 14:54:54 RP: I missunderstood you, sorry. Sep 14 14:55:10 RP: Do you know him? Sep 14 14:55:48 RP: I understand. In this case the script won't work. The /dev/hda mount would have to sleep a while to wait for partitions to appear. Sep 14 14:55:53 sirfred: kind of. We've exchanged patches :) Sep 14 14:56:08 pH5: Yes, this was my concern... Sep 14 14:56:28 RP: That's a good beginning. :) Sep 14 14:58:03 RP: I suppose you've seen my mails about the SCALE issues with the IMAGEON. I don't know if you're interested anymore into those things. Sep 14 14:58:10 03koen 07org.oe.dev * r3a18c612... 10/ (3 files in 3 dirs): Sep 14 14:58:10 angstrom: prepare for uclibc/NPTL and seperate uclibc and glibc packages Sep 14 14:58:10 * NOTE: DEPLOY_DIR changed, remove all do_packages and do_install stamps Sep 14 14:58:45 sirfred: I've seen them and I am interested, I just have no time :-( Sep 14 14:59:02 * koen looks at the uclibc/EABI/armv5te nano package Sep 14 14:59:05 * koen rejoices Sep 14 14:59:22 koen: but does it run? :) Sep 14 14:59:25 RP: No problem, I just don't want to disturb you with unwanted mails. Sep 14 14:59:28 I have no idea :) Sep 14 14:59:40 sirfred: Please keep my cc'd, I do read them :) Sep 14 14:59:50 RP: of course Sep 14 14:59:56 RP: I decided to idiot-proof the infrastructure first :) Sep 14 15:00:01 sirfred: Do you want me to ask the fbdev maintainer some questions? Sep 14 15:00:11 koen: Always good :) Sep 14 15:00:21 RP: Well, I'm pretty sure that sync is just for syncing hardware/software ops. Sep 14 15:00:31 RP: That's what other drivers do. Sep 14 15:01:01 I think the patch is useful. I know that only applies and accelerates the console. Sep 14 15:01:05 RP: my commit does force the equivalent of a rebuild :( Sep 14 15:01:20 sirfred: I think it also gets called when userspace reads/writes to /dev/fb Sep 14 15:01:25 RP: but it pretty much allows you to mix uclibc and glibc builds :) Sep 14 15:01:42 koen: Mixing the types is cool :) Sep 14 15:01:53 RP: That makes sense, to avoid hardware operations to be out of sync. Sep 14 15:02:16 RP: Writing/reading the framebuffer from userspace can be seen as a non accelerated operation. Sep 14 15:02:26 RP: You must wait for any hw op to finish. Sep 14 15:02:29 multimachine seperates the archs and DEPLOY_DIR = $ANGSTROM_MODE/deploy makes sure images and packages get seperated Sep 14 15:02:30 sirfred: Does your patch account for that? Sep 14 15:03:01 RP: It's the same from the driver perspective. Sep 14 15:03:13 RP: You don't need to hardware sync if your last operation was not hardware. Sep 14 15:03:39 sirfred: ok. Sorry, I just haven't followed the code as well as I'd like :-/ Sep 14 15:03:56 RP: Well, it's a single patch. Sep 14 15:04:06 RP: But I think the console is faster. Sep 14 15:04:20 RP: Have being using it the last week and haven't found any issue. Sep 14 15:04:55 I understand that accelerated console is not a breaking feature, but, I think that if we have it, it should be the best it could. Sep 14 15:05:28 So, accelerated framebuffer or not accelerated at all, but partially accelerated... Sep 14 15:06:03 sirfred: It did look like a good idea to me Sep 14 15:06:26 sirfred: I will test and if it works ok for me too, we can submit it for 2.6.19 Sep 14 15:06:41 RP: That would be very nice. Sep 14 15:06:46 sirfred: I will also try and really speed up the scrolling :) Sep 14 15:06:57 RP: Please keep me informed if you find any problem. Sep 14 15:06:58 does that play well with Xw100? Sep 14 15:07:21 koen: It should, as far as the current w100fb plays well. Sep 14 15:07:29 cool Sep 14 15:07:36 sirfred: I'll let you know when I've tried it one way or another Sep 14 15:07:45 RP: thanks. Sep 14 15:08:49 RP: What's actually making it faster is colour expansion, I think. We're saving a lot of bandwidth. Sep 14 15:09:15 sirfred: I can test any w100 patches f you need extra testng Sep 14 15:09:47 RP: With the current approach, the 1bpp fonts have to be converted to 16bpp before being written to the framebuffer. With this patch we're transfering the 1bpp fonts directly, and colour expansion is being made by the w100 Sep 14 15:10:03 XorA: The patch is available in libw100.sf.net. Feel free to test it. Sep 14 15:10:15 sirfred: That is extremely useful :) Sep 14 15:10:32 RP: Sure Xw100 could also benefit with this. Sep 14 15:10:57 And also alpha blending, that's working fine in the libw100 demos. Sep 14 15:10:59 sirfred: cool Sep 14 15:11:23 XorA: Be patient, because sf is really slow sometimes (most of times) Sep 14 15:11:40 sirfred: Both sound like they could be useful to X although it probably doesn't deal with those bit depths too often :-/ Sep 14 15:11:53 (conversion from 24bpp would be more useful) Sep 14 15:12:19 RP: Conversion is possible, just have to split the 24bpp into a 16bpp non-alpha pixmap and a 4bpp alpha mask. Sep 14 15:12:35 XorA: http://prdownloads.sourceforge.net/libw100/w100_extaccel.patch?download Sep 14 15:13:19 sirfred: would that be more work than just converting it within X in the first place though? Sep 14 15:13:28 XorA: It should apply against the current linux openzaurus 2.6.17 Sep 14 15:13:48 RP: Converting into what? Sep 14 15:14:19 RP: You mean blending in CPU ? Sep 14 15:14:36 RP: Well, excuse me. I was thinking about pixmaps with alpha channel. Sep 14 15:15:07 Not 24bpp non-alpha. The imageon does not support more than 16bpp. Sep 14 15:15:43 sirfred: I shall try and check that out then Sep 14 15:15:51 XorA: Thanks. Sep 14 15:17:09 XorA: I've being playing a little with overlay + scaling in the imageon. Sep 14 15:17:30 XorA: You were interested into those things, if I remember correctly. Sep 14 15:17:46 sirfred: I tried mplayer -vo w100 in Xw100, result is not pretty, but Im sure in time you will have the all singing and dancing solution Sep 14 15:17:59 sirfred: I use my Z for video watchng on occasion Sep 14 15:18:02 sirfred: The next step for the overlay might be an xv extension :) Sep 14 15:18:13 RP: Yes, that would be nice. Sep 14 15:18:21 RP: And XvMC... the promised land. Sep 14 15:18:54 wow Xv / XvMC would make mplayer/xine users happy Sep 14 15:18:55 RP: Unfortunately, it seems there's no information for any ATI chip regarding MC and IDCT. It's a pitty, because the IMAGEON seems to share most register names regarding IDCT Sep 14 15:19:36 sirfred: Its a shame there are no docs... Sep 14 15:19:53 RP: We need to hack ATI servers. :-\ Sep 14 15:20:13 RP: sirfred apparently the DivX people managed to get docs before ATI lost them, but they were under NDA Sep 14 15:20:17 Perhaps it would be less efort than reverse engineer that part. Sep 14 15:20:57 XorA: And do they released any open code, DivX people, I mean? Sep 14 15:21:23 sirfred: from what they said on forums they got it to do the WinCE version of the player, but no source :-( Sep 14 15:21:42 XorA: :-( Sep 14 15:21:58 XorA: I'm going to start playing again with the overlay issues. Sep 14 15:22:03 XorA: What I got by now is: Sep 14 15:22:11 1.-I was able to overlay from a surface in external memory. Sep 14 15:22:45 Anyway, it seems there's some distortion for YUV surfaces, I don't know why. Sep 14 15:23:03 sirfred: certanly sounds like things are in capable hands with you :-=) Sep 14 15:23:23 2.-Using stretchblt operations seems not to be the way to follow, because the chroma channels are not handled correctly. Sep 14 15:23:56 I tried to stretch separately luma and chroma. We can fake the graphic engine about the luma, saying that it's 8bpp src and dst. Sep 14 15:24:16 But chroma is 4bpp, and I've found no way to say to the graphic engine to stretch a 4bpp plane. Sep 14 15:24:50 Anyway, it's very strange that overlays can not be scaled directly. I want to look at it again. Sep 14 15:25:38 sirfred: who knows what mad ideas ATI had Sep 14 15:25:52 XorA: Well, I was able to scale the overlay directly, but only with an integer factor (by 2, by 3 and by 4), and that's not nice to show a video full screen. Sep 14 15:25:58 XorA: :) Sep 14 15:26:07 sirfred: I would for a year at ST when they were designing gfx chips, and sense didnt seem to feature much Sep 14 15:26:22 XorA: ST? Sep 14 15:26:47 sirfred: ST Microelectronics, they did the Kyro/KyroII and nvidias first chip Sep 14 15:26:56 oh Sep 14 15:27:15 XorA: Well, as you have seen, I've setup a wiki about the imageon (expect ATI not to close it) Sep 14 15:27:29 XorA: There's also a reference library to test functionality, if you're interested. Sep 14 15:27:42 * mrz80 is away: Metaphysically I'm still here; physically, I'm elsewhere Sep 14 15:27:44 sirfred: I did look a few days ago, looked good what was there Sep 14 15:27:49 XorA: I have some demos about overlay, stretching,... Your comments + help would be appreciated. Sep 14 15:27:59 sirfred: are the demos available as well? Sep 14 15:28:09 XorA: Well, there's a lot of information missing, but it's slowly progressing. Sep 14 15:28:17 koen: Yes, they're part of the svn repository. Sep 14 15:28:17 sirfred: I am closing up my job here, so Im actually busier than before I quit :-) Sep 14 15:28:21 koen: In the tests directory. Sep 14 15:28:48 XorA: infect lrg with the 'ooooo! shiny!' virus ;) Sep 14 15:29:17 koen: But please, don't try the 'hostcopyrot' one. :) Sep 14 15:29:27 koen: If you don't want to restart your machine. Sep 14 15:29:50 'rot' for rotated, but it could be for 'rotten' in this case. Sep 14 15:30:03 I'd need to localize my Z first Sep 14 15:30:14 it's ~1.5m from a poweroutlet and within wifi coverage Sep 14 15:30:28 :) Sep 14 15:31:38 and it's running a build with broken alsa so I can't make it beep :) Sep 14 15:34:47 * koen -> food Sep 14 15:34:52 :) Sep 14 15:35:51 NOTE: package xxx-devel-image-1.0-r0: task do_build: completed Sep 14 15:35:55 * mickey|bbl dances a jig Sep 14 15:36:01 that was MACHINE = x86 Sep 14 15:36:05 or rather, i686 Sep 14 15:36:09 no, 'native' even Sep 14 15:36:22 now let's see whether these binaries actually work Sep 14 15:37:24 mickey|bbl: That means that I could create a distro for an x86 machine? Sep 14 15:37:31 yeah! Sep 14 15:37:45 I'm thinking in an old 486 laptop I'm storing somewhere. Sep 14 15:37:47 well, at least you can use OE to make an ad-hoc SDK Sep 14 15:38:19 mickey|bbl: what an accomplishment, making an x86 image Sep 14 15:38:19 heh Sep 14 15:38:20 building a distro for x86 involves a bit more, but OE should do that as well Sep 14 15:38:55 CosmicPenguin: heh, after it took me 8 hours yesterday to build all the stuff manually this comes as quite a nice relief Sep 14 15:39:22 it even didn't work what I built yesterday Sep 14 15:39:27 Wow - what was tripping you up? Sep 14 15:39:36 ran into all kinds of problems w/ local stuff vs. native stuff Sep 14 15:39:50 header files not matching, symbols missing Sep 14 15:39:54 heh Sep 14 15:40:18 i guess some stuff is easier to cross-build than locally Sep 14 15:40:23 s/locally/native/ Sep 14 15:40:46 yeah - I'll agree with that Sep 14 15:40:46 eds-dbus svn killed me :D Sep 14 15:41:00 time to get bus Sep 14 15:41:11 We actually did have problems bulding x86-64 -> x86-64 Sep 14 15:41:48 we really really need to get away with all the -native packages Sep 14 15:41:54 this has to be different Sep 14 15:42:06 * mickey|bbl marks as OEDEM topic Sep 14 15:53:20 traffic is insane out here Sep 14 16:01:36 mickey|bbl: I have ideas about native packages and multimachine... Sep 14 16:01:41 * RP should experiment at some point Sep 14 16:05:09 RP: youll have to invent a pocket universe to give you more time Sep 14 17:27:17 koen: will look at it tomorrow when i get back to the office for real Sep 14 17:31:33 ade|desk: thanks Sep 14 17:32:02 ade|desk: I'll try to figure out how to add proper uclibc.confs for the machines I want to build for Sep 14 17:37:16 hi Sep 14 17:37:21 likewise: hi Sep 14 17:37:36 hey hrw Sep 14 17:37:39 hey likewise Sep 14 17:38:45 hrw: hi there Sep 14 17:38:49 koen: hi there as well Sep 14 17:39:12 likewise: can you remind me your email? Sep 14 17:39:52 I'm too tired now to discuss.. ;( Sep 14 17:40:15 np, hold on Sep 14 17:40:53 hrw: saw your OE mailing list mail as well, there is no hurry. leonw@mailcan.com Sep 14 17:40:56 likewise: the cairo float patch was reverted since it caused troubles on some systems Sep 14 17:41:38 koen: ok, I will test it anyway in the next few weeks. I have also picked up working on a PCI Express card, but my VHDL knowledge is still near-zero. Sep 14 17:42:00 hey mickeyl Sep 14 17:42:08 I want PIM which is able to link between contacts/todos/events/notes/mails... Sep 14 17:42:10 hi Sep 14 17:42:13 hi mickeyl Sep 14 17:42:22 cheers hrw Sep 14 17:42:22 hrw: you mean 'people as first class objects' Sep 14 17:42:36 koen: no. Sep 14 17:43:05 koen: 'finish writing reply to *this* mail before 2006.09.14 20:00 CEST" Sep 14 17:43:28 ah, you want a crm system Sep 14 17:43:29 koen: no person info at all in event Sep 14 17:44:33 http://youtube.com/watch?v=64QuJdJmCbA is nice to watch for people that want to design a PIM system Sep 14 17:45:18 koen: it can be called crm, pim, anything but it has to be desktop app which use kde components Sep 14 17:47:59 ~lart flashplayer for !alsa Sep 14 17:47:59 * ibot says "boot to the head" and knocks flashplayer over for !alsa Sep 14 17:50:54 mickeyl: what's the url for the berlin baer? Sep 14 17:51:50 hold on a second Sep 14 17:52:56 http://www.hotel-berliner-baer.de/e/index.htm Sep 14 17:55:42 mickeyl: thanks Sep 14 18:01:25 'evening Sep 14 18:12:29 koen: nice video Sep 14 18:15:14 mickeyl: | /home/hrw/devel/build/oz-dev/tmp/staging/arm-linux/qt2/lib/libqte.so: undefined reference to `vtable for __cxxabiv1::__si_class_type_info' Sep 14 18:15:20 mickeyl: ideas? Sep 14 18:15:35 mickeyl: other then 'rebuild failed package with stdc++'? Sep 14 18:16:36 hrw: edit qmake-base.bbclass Sep 14 18:16:44 s/-fno-rtti/-frtti/ Sep 14 18:16:49 try rebuilding package Sep 14 18:17:24 trying Sep 14 18:18:11 mickeyl: built Sep 14 18:18:26 mickeyl: opie-bluepin, opie-rdesktop fails with it Sep 14 18:18:42 where it = 'undefined reference' Sep 14 18:19:22 http://ewi546.ewi.utwente.nl/tinderbox/showlog.pl?machine_id=115&logfile=20060914195628.log Sep 14 18:19:33 "Hrw-OZ-3542-from-dev" build Sep 14 18:19:35 it shows how simple things can make your like much easier Sep 14 18:19:45 03tmbinc 07org.oe.dreambox * r362d0008... 10/ (3 files in 2 dirs): swig: remove to fix merge conflict Sep 14 18:19:50 03tmbinc 07org.oe.dreambox * r52dfcd4e... 10/ (1 packages/gcc/gcc-4.1.0/zecke-xgcc-cpp.patch): merge: resolve orphaned node conflict Sep 14 18:19:57 03tmbinc 07org.oe.dreambox * r8caf60d4... 10/ (1 classes/package.bbclass): package.bbclass: work around bogus -dbg RDEPENDS on split packages Sep 14 18:20:02 03tmbinc 07org.oe.dreambox * r2f777287... 10/ (6 files in 4 dirs): busybox-1.2.1: port config, keymap endianess patch Sep 14 18:20:16 03tmbinc 07org.oe.dreambox * r9e6b5803... 10/ (1 packages/timezones/timezones-alternative.bb): timezones-alternative: fix do_install Sep 14 18:20:22 03tmbinc 07org.oe.dreambox * r20b54a43... 10/ (1 packages/swig/swig_1.3.29.bb): swig-1.3.29: add disable_exceptions.diff Sep 14 18:20:28 03tmbinc 07org.oe.dreambox * r5f6adad9... 10/ (1 packages/samba/samba.inc packages/samba/samba_3.0.23c.bb): samba: fix smb.conf installation Sep 14 18:20:33 03tmbinc 07org.oe.dreambox * rb83ff04b... 10/ (3 files in 3 dirs): initscripts-opendreambox: fix mount Sep 14 18:20:45 03tmbinc 07org.oe.dreambox * rc04eb040... 10/ (1 packages/gstreamer/gst-plugins.inc): gst-plugins.inc: don't depend on gconf Sep 14 18:20:55 03tmbinc 07org.oe.dreambox * r3c170089... 10/ (1 packages/enigma2/enigma2.bb): enigma2: update to 20060911 Sep 14 18:21:05 03tmbinc 07org.oe.dreambox * rec59c70d... 10/ (1 packages/initscripts/initscripts-1.0/dm7025/bootup): initscripts-opendreambox: fix conditional iframe Sep 14 18:21:17 03tmbinc 07org.oe.dreambox * rc8fda820... 10/ (1 packages/dreambox/dreambox-compat.bb): dreambox-compat: use install -d instead of mkdir Sep 14 18:21:22 hrw: try rebuilding everything with RTTI and see whether opie-rdesktop builds then Sep 14 18:21:39 mickeyl: starting from qte? Sep 14 18:22:02 yeah Sep 14 18:22:11 someone know how much turion64 x2 notebooks are today? :D Sep 14 18:22:18 *g* Sep 14 18:22:29 hi mickeyl Sep 14 18:22:54 could you take a look at something for me. It will only take a minute Sep 14 18:22:59 Ania bought laptop few days ago and I'm starting to feel that this is my last desktop machine Sep 14 18:24:41 we are close to 4K recipes.. Sep 14 18:25:29 and 10k revs in .dev Sep 14 18:26:57 asus a6m with t64x2 1.6GHz is <1000 EUR Sep 14 18:27:53 isn't a turion the AMD celeron? Sep 14 18:28:12 koen yup Sep 14 18:28:42 koen: turion is mobile version of amd64 line Sep 14 18:30:19 remember to put in loads or ram Sep 14 18:30:23 amd celeron is sempron Sep 14 18:30:44 koen: I know - @work have only 768M Sep 14 18:32:40 mickeyl: NOTE: package opie-rdesktop-1.2.2+cvs-20060913-r0: task do_build: completed Sep 14 18:33:14 *nod* Sep 14 18:33:19 do me a favour and compare the size Sep 14 18:33:23 hrw: now mickeyl is fixing native builds, you could install opie ;) Sep 14 18:33:24 rtti vs. non-rtti Sep 14 18:34:21 mickeyl: iirc diet-x11 vs full x11 is ~0.5MB Sep 14 18:34:33 but my memory is a bit hazy Sep 14 18:34:35 shit.. now have to rebuilt no-rtti.. Sep 14 18:35:24 opie-rdesktop is not packaged properly.. Sep 14 18:35:46 rebuilding it first Sep 14 18:36:04 * mrz80 is back (gone 03:08:24) Sep 14 18:36:53 koen: thanks. Sep 14 18:37:04 i have just learned that our device will feature... Sep 14 18:37:10 ... 1G Sep 14 18:37:17 (flash) Sep 14 18:37:26 so I no longer care about diets ;) Sep 14 18:37:33 mickeyl: 256M RAM? Sep 14 18:37:33 GiB or Gib? Sep 14 18:37:55 well, GB Sep 14 18:37:58 hrw: I'm afraid not. Sep 14 18:38:04 let me check memory specs Sep 14 18:38:45 hmm 64MB in the develboard Sep 14 18:38:54 dunno whether this will be increased for the final Sep 14 18:39:16 I don't why I keep buying Asus motherboards. :( Sep 14 18:39:38 * T0mW needs a 12 step program for his Asus habit Sep 14 18:41:43 uf.. mail sent Sep 14 18:43:22 mickeyl: the bär was a bit pricey, so it looks like we'll be staying at the majestic Sep 14 18:43:22 argh Sep 14 18:43:22 the cable modem is being sick again Sep 14 18:44:09 mickeyl: libqpe.so.1.5.0 1093532 1121012, qte looks same size Sep 14 18:44:27 hmm, that's less worse than I've thought Sep 14 18:45:48 time to takeover haerwu.one.pl Sep 14 18:46:49 s/haerwu.one.pl/the world/ :) Sep 14 19:11:00 Doing a fresh build of openslug from the makefile but am getting a "ImportError: cannot import name msg" Is this expected, odd, or me doing something wrong on my end? Sep 14 19:12:32 03pH5 07org.oe.dev * r90bb70bf... 10/ (1 packages/linux/handhelds-pxa-2.6_cvs.bb): handhelds-pxa-2.6_cvs: update HHV Sep 14 19:16:40 sleep well guys Sep 14 19:16:56 hrw: sleep well & thanks Sep 14 19:17:05 np pgfeller Sep 14 19:18:27 joshin: did you updated bitbake ? msg is one bitbake module Sep 14 19:20:05 I thought I did. Am grabbing again. Sep 14 19:20:30 msg looks to be a new module. It might not be part of the openslug master makefile. Not sure yet. Sep 14 19:23:11 joshin: looking at CIA, RP made some change on msg.py. So I guess updating bitbake will solve your issue Sep 14 19:23:44 s/some/some recent/ Sep 14 19:23:56 Yup, thanks! Sep 14 19:29:00 hi Sep 14 19:31:15 * koen mumbles curses at gcc-cross Sep 14 19:34:44 hi, anyone an idea what is wrong if bb tells me: ImportError: cannot import name msg Sep 14 19:35:27 eFfeM: scroll up, updating bitbake worked for joshin Sep 14 19:35:58 cyrilRomain: must see if this is in my log, was offline for 30 mins or so Sep 14 19:35:59 cyrilRomain, It hasn't worked yet. I'm doing a fresh download Sep 14 19:36:11 ah Sep 14 19:36:20 But I expect it will... Sep 14 19:36:52 oops. forgot that I do not have oe logged to disk Sep 14 19:37:06 joshin, what are you attempting? Sep 14 19:37:36 eFfeM: joshin has the same error message Sep 14 19:38:04 I'm trying to figure out why building gcc 4.1.x for openslug dies on my gentoo amd64 box and was trying a fresh install on a P3. Sep 14 19:38:10 just did an mtn update, before that it works Sep 14 19:38:35 eFfeM, what target are you trying to build for? Sep 14 19:38:46 cyrilRomain, I dare you to mtn update. :) Sep 14 19:38:50 joshin there are some issues with amd, don't know the details, but apparently the openslug image is also bigger with amd than when build with intel Sep 14 19:38:57 joshin, openslug Sep 14 19:39:09 Bizarre. AMD vs Intel or 64 vs 32 bit? Sep 14 19:39:38 dunno. I have an D930 cpu (3 Ghz, dual core) Sep 14 19:46:50 Ok, looks like the current version of bitbake grabbed from org.nslu2-linux.bitbake has issues. Sep 14 19:46:54 joshin, thought it is amd64 vs rest of the world (including amd32) not totally sure though Sep 14 19:47:05 hm Sep 14 19:47:45 So cyrilRomain can you risk updating and see if the main repository is fine? If so then the issue is with org.nslu2-linux.bitbake Sep 14 19:50:29 ok, i'm leaving for a while (almost melting from my chair), back later, feel free to leave a msg Sep 14 19:50:43 (might log on from the laptop as well) Sep 14 19:53:23 i would like to compile a local saved kernel tree (mod of handhelds-sa-2.6 kernel ) Sep 14 19:53:26 what's the local.conf syntax to link to "my source tree" instead of the unpacked handhelds-sa-2.6 package in the work dir Sep 14 19:54:33 ok, back from other system Sep 14 19:54:54 * eFfeM is now in the garden with laptop & beer. Much better ..... Sep 14 19:55:12 03eFfeM 07org.oe.dev * ra1d864e3... 10/ (6 files in 3 dirs): asterisk: added 1.2.12.1 on behalf of Corneliu Doban Sep 14 19:55:36 eFfeM: :-) Sep 14 20:00:03 Hi all, - on what do I have to DEPEND if I rely on uinput support? Sep 14 20:01:23 joshin: I'm on my laptop with a fresh OE checkout. but have another issue (I have to configure my local.conf). So I'll tell you asap Sep 14 20:06:13 thanks cyrilRomain! Sep 14 20:06:31 joshin: do you have the error message as soon as you run bitbake ? Sep 14 20:06:37 yup Sep 14 20:07:06 joshin: ah, I have compilation issue on quilt (qemu because I use gcc4) but don't have your bitbake error Sep 14 20:08:17 Thanks. Sep 14 20:08:24 joshin: is org.nslu2-linux.bitbake a bitbake branch (I mean not the bitbake/branches/bitbake-1.6 one) ? Sep 14 20:09:05 I'm not (yet) sure. After I deal with a sick and cranky 2.5 year old and get her to eat her lunch I'll poke at it. Sep 14 20:10:53 pgfeller: on nothing, (R)DEPENDING on kernel-modules is a bug Sep 14 20:11:35 koen|away, does this bitbake thing ring a bell for you ? Sep 14 20:12:34 koen: mhh ... should I check in the preinst script then, if the environment is suited for the package? Sep 14 20:14:13 pgfeller: you can do RRECOMMENDS = "kernel-modules-uinput" Sep 14 20:15:01 koen: thanks .. I'll do so. Sep 14 20:15:36 there's no foolproof way to check for kernel features :( Sep 14 20:16:08 np ... the pgm checks it again anyhow ;-) Sep 14 20:25:01 03rpurdie * r607 10bitbake/lib/bb/runqueue.py: runqueue.py: Convert Deps and RunDeps fields into sets, add debugging code for tracing indriect circular dependencies Sep 14 20:54:20 03clarson 07org.oe.dev * rdb2f843f... 10/ (5 files in 2 dirs): Restructure the subpackage metadata to facilitate use of that metadata by other packages. Sep 14 20:56:16 kergoth: yay! Sep 14 20:58:38 wow - fun Sep 14 21:01:18 CosmicPenguin: http://www.openembedded.org/viewmtn/diff.psp?id1=8156b74b683b22fd001a5c526476023c0a3c9ec8&id2=3a18c612a87aa7ac8ef62ccb9ab6920dbb1cd0de&fname=conf/distro/include/angstrom.inc might interest you Sep 14 21:01:45 CosmicPenguin: if you combine that with multimachine.bbclass you can switch between glibc and uclibc with much pain Sep 14 21:02:08 why would we do something painful? Sep 14 21:03:41 03clarson 07org.oe.dev * rfd4f8197... 10/ (1 classes/package.bbclass conf/bitbake.conf): Sep 14 21:03:41 classes/package.bbclass,conf/bitbake.conf: Sep 14 21:03:41 * Add package "depchains". This facilitates, for example, ensuring that if Sep 14 21:03:41 A depends upon B, then A-dev will RRECOMMENDS B-dev, and the same for the Sep 14 21:03:41 -dbg packages. Sep 14 21:04:05 not as flexible as it could be, but gets the job done Sep 14 21:04:11 koen: but could you do it on a package by package basis? :) Sep 14 21:05:15 CosmicPenguin: not yet Sep 14 21:05:45 CosmicPenguin: I only managed to build nano against uclibc Sep 14 21:16:00 This is where I figure out if I at all manged to make sugar work Sep 14 21:22:23 * koen hands CosmicPenguin some aspartam Sep 14 21:24:18 heh - I suppose I would need python first Sep 14 21:24:25 * CosmicPenguin smacks self upside the head Sep 14 22:01:43 03rpurdie * r608 10bitbake/lib/bb/taskdata.py: taskqueue.py: Add a small cache for task ids for a big speed improvement Sep 14 22:02:08 * v8jlene is back Sep 14 22:53:28 hello all Sep 14 22:55:06 anyone know much about pdaxrom? Sep 14 22:55:20 wrong channel Sep 14 22:56:03 im in the wrong channel? Sep 14 22:56:17 sorry Sep 14 22:56:37 wow - nice Sep 14 22:56:50 hehe Sep 14 22:57:05 watch my *yedimove* Sep 14 22:57:10 he koen Sep 14 23:02:57 hey woglinde Sep 14 23:08:00 * JustinP grins Sep 14 23:10:45 hehe yes but I need to be a regexp monster Sep 14 23:11:13 to fix many stupid regexp in oe Sep 14 23:11:34 * JustinP enjoys the fine taste of a well-made regex Sep 14 23:12:49 heh - I always thought kergoth had some well defined regex-foo Sep 14 23:13:00 He always managed to impress me - I guess I'm simple Sep 14 23:13:03 pb is a regexp master Sep 14 23:26:19 ah nice Sep 14 23:26:26 grouping Sep 14 23:26:30 and backreferenz Sep 15 00:09:41 re Sep 15 00:49:53 03rwhitby 07org.oe.dev * r423d9514... 10/ (6 files in 3 dirs): ixp4xx-npe: Added 2.1 version from blaster8 Sep 15 01:02:52 * v8jlene is away (cycling) **** ENDING LOGGING AT Fri Sep 15 02:59:57 2006