**** BEGIN LOGGING AT Fri Sep 23 02:59:56 2005 Sep 23 03:00:59 Bernardo: i know Sep 23 03:01:01 :) Sep 23 03:01:40 Maybe this: open("/usr/lib/python2.4/site-packages/psyco/__future__module.so", O_RDONLY|O_LARGEFILE) = -1 ENOENT (No such file or directory) Sep 23 03:01:40 open("/usr/lib/python2.4/site-packages/psyco/__future__.py", O_RDONLY|O_LARGEFILE) = -1 ENOENT (No such file or directory) Sep 23 03:01:40 open("/usr/lib/python2.4/site-packages/psyco/__future__.pyc", O_RDONLY|O_LARGEFILE) = -1 ENOENT (No such file or directory) Sep 23 03:01:41 ? Sep 23 03:02:19 Bilbolodz: paste it in pastebin and let me take a look Sep 23 03:03:46 Its too big to paste, but it's here: http://pc-bilbo.man.lodz.pl/temp/out Sep 23 03:04:15 Bilbolodz: psyco is python JIT - useful but bitbake work without it too Sep 23 03:05:30 Bilbolodz: are you use you changed BBPATH like I asked? Sep 23 03:06:15 hrw: Yep: [bilbo@pc-bilbo OE]$ set |grep BB Sep 23 03:06:15 BBPATH=/home/bilbo/OE/:/home/bilbo/OE/org.openembedded.dev/ Sep 23 03:06:47 Bilbolodz: BBPATH=/home/bilbo/OE/build/:/home/bilbo/OE/org.openembedded.dev/ Sep 23 03:06:56 Bilbolodz: or move build/conf to OE/conf Sep 23 03:07:43 hhehe Sep 23 03:07:50 hrw:work| YOU ARE BIG! It's working! Sep 23 03:07:56 Bilbolodz: are you sure your local.conf Sep 23 03:08:10 is in the build/conf? Sep 23 03:08:14 Bilbolodz: I'm big - 190cm Sep 23 03:08:21 lol Sep 23 03:08:47 Bilbolodz: stat64("conf/local.conf", 0xbfd51688) = -1 ENOENT (No such file or directory) Sep 23 03:09:06 It was may mistake, sorry nad thanks for any help Sep 23 03:09:20 "/build/" is "the key" Sep 23 03:09:22 Bilbolodz: np Sep 23 03:11:07 In local.conf as distro should I put "opensimpad" or "opensimpad-0.9.0"? Sep 23 03:11:26 numbered Sep 23 03:13:25 hrw|work: My local.conf is: http://pastebin.ca/23679 Please review it Sep 23 03:14:01 Bilbolodz: # NOTE: Do NOT use $HOME in your paths, BitBake does NOT expand ~ for you. If you Sep 23 03:14:06 Bilbolodz: DL_DIR = "${HOME}/sources" Sep 23 03:14:56 Bilbolodz: you can also set TMPDIR to /home/bilbo/OE/build/tmp/ Sep 23 03:15:37 Bilbolodz: parallel_make is good thing if your machine is not pII/400 Sep 23 03:16:01 Bilbolodz: looks ok when you change DL_DIR :) Sep 23 03:17:00 fixed Sep 23 03:30:49 where is libintl.h in uclibc build - someone know? Sep 23 03:33:59 nowhere after you've rebuilt uclibc Sep 23 03:34:03 It comes from gettext Sep 23 03:35:10 ok Sep 23 03:47:14 someone built console-tools with uclibc? Sep 23 03:48:24 03rpurdie 07org.oe.dev * r8df2259d... 10/packages/linux/ (5 files): Sep 23 03:48:24 linux-oz-2.6: Add a fix for spitz resuming the backlight whilst performing Sep 23 03:48:24 charging operations and suspended. Upgrade development series to 2.6.14-rc2-mm1. Sep 23 03:48:24 Testing of -mm kernel welcome - untested as yet but can't be more broken than it Sep 23 03:48:24 was. Sep 23 03:49:06 thought RP was in a tent? Sep 23 03:50:24 its raining so I'm putting off leaving for a bit :) Sep 23 03:50:41 RP: :-D you need waterproof Z Sep 23 03:51:09 XorA: Its not the only waterproof thing I need :-/ Sep 23 03:52:25 long ethernet cable too? :) Sep 23 03:53:26 Yes. Chances are they won't have power either though.... Sep 23 03:53:32 any idea why the apps that i built from OE segmentation fault? Sep 23 03:53:53 RP: time to get cpufreq in to give max battery hours :-D Sep 23 03:53:54 anyway to trace it in OZ in my Z? Sep 23 03:54:20 zuery: strace? Sep 23 03:54:25 gdb? Sep 23 03:55:36 i tried to build dosbox 0.63.. it successfully built.. but when i run it... it segmentation fault.. Sep 23 03:56:48 ok.. i'll try to work on it... Sep 23 04:14:02 is there a way to make the OE built the image and all the modules in one file... Sep 23 04:14:53 bootstrap-image separate all the modules to individual ipk... Sep 23 04:15:12 zuery: create own image which will contain all kernel modules Sep 23 04:15:22 ipkg install kernel-modules Sep 23 04:15:25 and it is not bootstrap-image which separate modules Sep 23 04:15:32 that's a meta-pacakge containing all the modules Sep 23 04:16:05 ok... got it.. Sep 23 04:16:13 go back first.. Sep 23 04:16:20 talk to u all later... Sep 23 04:16:29 thanx .. Sep 23 04:16:39 any doc about how to create a custom image ? Sep 23 04:17:02 alan|home: cd packages/meta/; less *-image* Sep 23 04:17:07 whole doc Sep 23 04:17:12 ok. ^^ Sep 23 04:21:17 looks like c7x0 image builds fine with uclibc except console-tools Sep 23 04:23:42 hrw|work: any idea on size decrease? Sep 23 04:25:50 XorA: not tested yet Sep 23 04:34:06 Whow, I've sucessfully built nano package! Sep 23 04:40:15 Bilbolodz: you use simpad for own pleasure or at work? Sep 23 04:47:50 mainly pleasure Sep 23 04:48:17 but now I want to make "small portable workstation" Sep 23 04:48:34 and use it to work on trips Sep 23 04:51:21 Bilbolodz: tried to add BT+SD hack? Sep 23 04:51:26 Now I'm building gpe-image (just for try) and next step will be to "compose" own gpe, based image with packages for workstation (without any PDA/PIM specific applications). Sep 23 04:52:25 not yet, at first: I don't have SD slot in Simpad. Simpad has PCMCIA and I've PCMCIA BT card working under linux on my Laptop Sep 23 04:53:05 Bilbolodz: http://externe.net/zaurus/simpad-bluetooth/ Sep 23 04:53:29 Bilbolodz: BT+SD hack homepage Sep 23 04:54:47 Interesting Sep 23 04:55:13 But this BT chip is VERY HARD to buy in Poland Sep 23 04:55:18 sad guyhlem wastes so much Sep 23 04:55:29 Bilbolodz: drive to Berlin, and buy it there ;) Sep 23 04:55:51 What's price? Sep 23 04:55:53 and drive through Poznan so we can meet ;) Sep 23 04:57:36 http://www.beecon.de/projekte/SIMpadBT/index_de.html Sep 23 04:59:06 actually I don't really need this: I've GC75 (PCMCIA GSM/GPRS card) Sep 23 04:59:50 But adding SD/MMC without taking PCMCIA is good idea Sep 23 04:59:56 Bilbolodz: but sd hack would be usefull - how much can be put in 32M of flash.. Sep 23 05:01:23 exactly Sep 23 05:01:35 Is this hack included in OE? Sep 23 05:01:53 Bilbolodz: no Sep 23 05:07:20 hi zecke_ Sep 23 05:07:47 weekend is soo close ;) Sep 23 05:11:09 OK SD/MMC will be feature, Now I've to know how to build custom image. Studuing meta/*-image told something but how to EXCLUDE some programs from "parent image"? Sep 23 05:16:59 Bilbolodz: create own image with only needed pacakges Sep 23 05:19:05 Bilbolodz: http://home.hrw.one.pl/oe/hrw-opie-image.bb is my last image (but it lack dropbear) Sep 23 05:19:39 Bilbolodz: cp gpe-image.bb bilbo-gpe-images.bb and start removing stuff Sep 23 05:19:55 then do bitbake bilbo-gpe-images Sep 23 05:20:25 I ran into a problem when I tried to bitbake task-bootstrap Sep 23 05:20:41 | checking for arm-linux-gcc... ccache arm-linux-gcc -march=armv5te -mtune=arm926ej-s Sep 23 05:20:42 | checking for suffix of object files... configure: error: cannot compute suffix of object files: cannot compile Sep 23 05:20:42 | See `config.log' for more details. Sep 23 05:20:42 | FATAL: oe_runconf failed Sep 23 05:20:42 NOTE: Task failed: /opt/oe/build/tmp/work/glibc-2.3.5+cvs20050627-r0/temp/log.do_configure.13785 Sep 23 05:20:42 NOTE: package glibc-2.3.5+cvs20050627-r0: task do_configure: failed Sep 23 05:20:44 ERROR: TaskFailed event exception, aborting Sep 23 05:20:46 NOTE: package glibc-2.3.5+cvs20050627: failed Sep 23 05:20:48 ERROR: Build of task-bootstrap failed Sep 23 05:21:26 I checked on google, and another person had a similar problem because it seems he didn't specify target_arch, however mine is specified (and you can see it's using the arm-linux- gcc binaries) Sep 23 05:22:02 OK e.x. how to remove PIM stuff (contacts, task etc)? Sep 23 05:22:51 I'm also not sure where config.log is... I don't see it Sep 23 05:23:42 Bilbolodz: line 39 of meta/gpe-image.bb Sep 23 05:24:47 invalid Python installation: unable to open /usr/lib/python2.4/config/Makefile (No such file or directory) Sep 23 05:24:49 Bilbolodz: that specifies what ends up in the image, you can see the contents of the task in meta/meta-gpe* Sep 23 05:24:49 Whop: problem building gpe-image for Simpad http://pastebin.ca/23689 Sep 23 05:24:52 anyone seen that? Sep 23 05:25:24 Bilbolodz: which version of gcc? Sep 23 05:25:35 Koen: OK thanks, Sep 23 05:26:03 [bilbo@pc-bilbo OE]$ gcc -v Sep 23 05:26:05 Using built-in specs. Sep 23 05:26:05 Target: i386-redhat-linux Sep 23 05:26:05 Configured with: ../configure --prefix=/usr --mandir=/usr/share/man --infodir=/usr/share/info --enable-shared --enable-threads=posix --enable-checking=release --with-system-zlib --enable-__cxa_atexit --disable-libunwind-exceptions --enable-libgcj-multifile --enable-languages=c,c++,objc,java,f95,ada --enable-java-awt=gtk --with-java-home=/usr/lib/jvm/java-1.4.2-gcj-1.4.2.0/jre --host=i386-redhat-linux Sep 23 05:26:05 Thread model: posix Sep 23 05:26:07 gcc version 4.0.1 20050727 (Red Hat 4.0.1-5) Sep 23 05:27:16 Bilbolodz: kernel cvs HEAD is broken atm Sep 23 05:28:40 pH5: so??? Tell me more please? Treat me as "green" please. Sep 23 05:28:42 :-) Sep 23 05:29:48 Bilbolodz: cross compiler please Sep 23 05:30:03 ls Sep 23 05:30:04 ah well ;) Sep 23 05:30:14 Hi everyone. "internal compiler error: Segmentation fault" Does this sound like bad memory/temperature problems? Sep 23 05:30:30 It can be Sep 23 05:30:51 Bilbolodz: if you want something that works 90%, go to bugs.openembedded.org and get you the 2.4kernel patches for simpad Sep 23 05:31:45 Bilbolodz: if you want a 2.6 kernel, you'll have to checkout with a cvs tag Sep 23 05:32:34 Bilbolodz: look at difference between handhelds-pxa-2.6_cvs.bb and handhelds-pxa-2.6_2.6.12-hh3.bb Sep 23 05:32:48 Bilbolodz: and do the same to handhelds-sa-2.6_cvs.bb Sep 23 05:33:19 Probabbly I need 2.4 kernel (I think there is no working 2.6 kernel on simpad) Sep 23 05:33:38 How to switch to 2.4? Sep 23 05:38:06 Bilbolodz: I don't know. Sep 23 05:38:33 How about this: Sep 23 05:38:35 # Uncomment this to select a particular kernel version if supported by Sep 23 05:38:35 # your MACHINE setting. Currently only supported on Zaurus Clamshells. Sep 23 05:38:35 # KERNEL_VERSION = "2.6" Sep 23 05:41:08 note to self, add perl to build machine , doh! Sep 23 05:42:00 Bilbolodz: where should we take the human resources from? (see the bug report) Sep 23 05:43:36 I can help but please tell me how ;-) Sep 23 05:44:11 Bilbolodz: applying the patch from the bug tracker takes one minute Sep 23 05:44:24 Bilbolodz: maintaining and testing will take up my time Sep 23 05:44:26 OK but where is this patch Sep 23 05:44:34 Bilbolodz: bugs.openembedded.org Sep 23 05:45:23 Bilbolodz: search for 'simpad' Sep 23 05:48:53 OK patch applied Sep 23 05:56:54 zecke: something is working, suppose that it will work, how I can do something for community and upload this patch to OE serwer? Sep 23 05:58:43 Bilbolodz: the issue is not in uploading it once Sep 23 05:58:55 Bilbolodz: I could do that within two minutes once I'm home Sep 23 05:59:17 Bilbolodz: but I will have to maintain and support it. Sep 23 05:59:26 Buu: I've got another error: http://pastebin.ca/23690 Sep 23 05:59:46 Bilbolodz: I've very limited time and would prefer to work on new stuff (finish the 2.6 port) Sep 23 06:00:12 Bilbolodz: well ;) Sep 23 06:00:29 zecke: OK, that's fine and this is great news Sep 23 06:00:30 Bilbolodz: go to openembedded/packages/linux find the 2.4 kernel file for SIMpad Sep 23 06:00:45 Bilbolodz: and there is a line with "3.3.3" change it to "3.3.4" Sep 23 06:03:12 file: opensimpad_2.4.27-vrs1-pxa1-jpm1.bb line: KERNEL_CCSUFFIX = "-3.3.3" Sep 23 06:03:13 ? Sep 23 06:04:21 yes Sep 23 06:04:56 zecke: Done, working...... ;-) Sep 23 06:05:30 I mean I just started compiling we will see the results Sep 23 06:07:53 Bilbolodz: gcc 3.3.3 can not be compiled with gcc 4.0.2 Sep 23 06:11:29 someone built console-tools with uclibc? Sep 23 06:12:09 hrw|work: am just trying Sep 23 06:12:38 zecke: suppose that it works I have another question: How to include into image only one (or maybe two) locale? e.x en-en, and pl-pl? Sep 23 06:12:49 hrw|work: at gettext-native, so maybe some time :( Sep 23 06:13:04 Bilbolodz: IMAGE_LINGUAS can be set IIRC Sep 23 06:14:19 opie-image.bb:export IMAGE_LINGUAS = "en,pl"? Sep 23 06:14:21 e.x. Sep 23 06:14:22 ? Sep 23 06:14:30 ade|desk: ok - maybe you will find what is going on there Sep 23 06:15:16 Bilbolodz: "en pl" rather Sep 23 06:15:24 hrw|work: at this rate perhaps tomorrow Sep 23 06:15:43 ade|desk: ok - I wont test it sooner probably Sep 23 06:16:22 hrw|work: work/university work take priority today Sep 23 06:17:31 ade|desk: sure Sep 23 06:19:24 * ade|desk grumbles something about last ever assignment this year not going to plan Sep 23 06:55:39 Is it me, or has the OE list really quieted down lately? Sep 23 06:58:21 psst Sep 23 07:00:18 we are about to get bug free Sep 23 07:13:44 morning Sep 23 07:22:57 NOTE: package meta-opie-1.0-r28: task do_build: completed Sep 23 07:23:05 for uclibc oz 3.5.4 Sep 23 07:23:11 quite nice Sep 23 07:24:38 some of libxine patches should be really pushed upstream... Sep 23 07:26:43 koen in train? Sep 23 07:31:40 wrobbie: yes Sep 23 07:31:49 * france is away: Away Sep 23 07:36:27 hrw|work: why is uclibc useful ? Sep 23 07:36:43 alan|home: it's much smaller as glibc Sep 23 07:36:51 about 1 - 2 MB Sep 23 07:36:58 depending on configuration Sep 23 07:38:17 zecke: did you get OE working on osx? Sep 23 07:38:50 koen|train: I've fought with KisMac (stupid guys forgot to put a complete xcode project file into the project...) Sep 23 07:38:58 koen|train: and I've struggled with Xephyr Sep 23 07:39:21 * koen|train does monotone pull Sep 23 07:39:26 and crosstool gave me a good laugh Sep 23 07:39:47 crosstool is like buildroot: very simple Sep 23 07:40:04 koen|train: yeah but too hard to fix ;) Sep 23 07:40:23 * zecke goes home Sep 23 07:40:25 cya later Sep 23 07:40:28 cu zecke Sep 23 07:40:30 and it can't order pizza like OE can ;) Sep 23 07:40:40 * zecke prepares for playing Mario 3 with Al Sep 23 07:40:40 cu zecke Sep 23 07:40:57 the good ole NES Sep 23 07:41:02 * pH5 searches for the pizza ordering feature in his checkout Sep 23 07:41:20 koen|train: there is a bitbake dominoes-pizza recipe? Sep 23 07:41:39 XorA: I think that's in kergoth's branch Sep 23 07:41:48 koen|train: wicked :-) Sep 23 07:43:00 and glibc is big as... ? Sep 23 07:43:19 an elephant's backside Sep 23 07:45:15 uclibc is ~1/3 of glibc (but lack iconv which is another 1/3) Sep 23 08:05:57 arf... with opie-image, i have an error : "No rule to make target « -lpcre », needed for « konqueror ». Stop." Sep 23 08:06:53 i think i need some help... Sep 23 08:06:59 alan|home: install libpcre-dev on host Sep 23 08:07:27 hrw|work: dam, that was sooooo easy... pfffff... Sep 23 08:08:21 alan|home: thats a bug in konq/e build Sep 23 08:08:52 thanks a lot hrw|work Sep 23 08:09:33 np Sep 23 08:14:23 hi everybody... Sep 23 08:16:44 i'm still getting too many open file error when compiling freetype-native.. Sep 23 08:17:38 anyone kind enough to tar bzip2 a copy of its freetype-native... in the work directory? Sep 23 08:18:07 heh.. I just shutdown build machine.. Sep 23 08:21:57 I wonder how a uclibc/thumb gpe/opie image would perform Sep 23 08:22:30 a gpe image will likely not work for the most part Sep 23 08:22:43 unless you fix gettext support Sep 23 08:22:54 it probably would fail to compile Sep 23 08:23:26 reenoo|OL: large parts of gpe have USE_NLS ifdefs Sep 23 08:24:08 nice Sep 23 08:24:18 gtk+ etc. however doesn't Sep 23 08:24:33 koen|train: glib-2.0 use it in nice way... is USE_NLS=no then skippackage Sep 23 08:24:48 jinidog was able to build his uclibc/gentoo image with a working libgpewidget Sep 23 08:24:58 hey hey Sep 23 08:25:21 koen|train: well. apps that used gettext would segfault Sep 23 08:25:40 ah, just like regular gentoo Sep 23 08:25:47 it segfaults *really fast* Sep 23 08:25:51 *cough* Sep 23 08:25:58 you need to have gettext provide the NLS bits normally provided by glibc Sep 23 08:26:14 it's perfectly possible but I'm not sure it works right now Sep 23 08:27:27 I see Sep 23 08:27:27 bbl Sep 23 08:40:24 cu all - have a nice weekend Sep 23 09:06:03 Are there any tricks to maintaining a separate OE branch? mt propagate is giving me lots of merge errors for files I never modified. Sep 23 09:16:11 Is the PXA255 conventionally little or bigendian? Was the SA1110 little-endian? I thought the l (L) on the end of the arm4l meant little endian? Sep 23 09:22:00 lardman|afk: little Sep 23 09:22:23 lardman|afk: like the SA1100 as well Sep 23 09:22:32 mickeyl: Wierd, in both R and Octave it's identified as big endian Sep 23 09:22:42 well, you can choose on arm Sep 23 09:22:57 on a per program basis? Sep 23 09:23:04 no Sep 23 09:23:12 i think the kernel handles that Sep 23 09:23:30 Very strange, I'll have to look into this Sep 23 09:24:01 I would think it even stranger, if userspace programs could change endianness at will Sep 23 09:24:10 Quite Sep 23 09:24:35 But Octave works, and the endianness tests say it's bigendian, so something strange is going on Sep 23 09:25:03 arm has some byte alignment issues that I've seen. Sep 23 09:25:09 (actually that's not quite true, the tests don't work completely, hence my question) Sep 23 09:25:13 code to detect endianness sometimes seems to fail. Sep 23 09:25:47 code that assumes that unaligned memory access works is broken. Sep 23 09:26:09 *nod* Sep 23 09:26:51 kergoth: any idea if this was related to the (very old) problem of python not working correctly on arm? Sep 23 09:27:53 what does unaligned memory access entail for fp handling then? Could I not read an unaligned section of a 64bit number, but instead would have to read the whole byte & do some OR'ing? Sep 23 09:28:04 seem to remember either debian (or python) developer actually coming out and saying it was broken by design, and they didn't expect to have any luck making it run in the future (circa 1998 iirc) Sep 23 09:28:18 hi mickeyl Sep 23 09:28:21 hi kergoth Sep 23 09:28:24 hey Sep 23 09:28:32 lardman|work: read the alignment page on www.arm.linux.org.uk Sep 23 09:28:42 kergoth: Thanks Sep 23 09:29:08 I didn't really pay much attention, other than about a year later, the problem seems to have resolved :/ Sep 23 09:32:08 USB question: on h2200, if USB isn't plugged in when the kernel boots, it won't work later when it is plugged in. If I change "dev->vbus = is_usb_connected();" to "dev->vbus = 1;" it works whether or not it was plugged in at boot time. Has anyone else seen this problem? Comments, suggestions? Sep 23 09:41:52 who has btscanner ipk for h3600/h3900? Sep 23 10:01:32 Are there name-mangling issues between major versions of GCC (i.e. GCC 3.3.0 and 3.4.4)? Sep 23 10:02:58 that didn't quite come out right (major/minor), but the versions are correct Sep 23 10:21:44 hi there :-) Sep 23 10:23:34 hey benoit Sep 23 10:24:12 i'm trying to use oe to build something for my brand new PDA Sep 23 10:24:19 however, i need to build monotone Sep 23 10:24:24 which need libboost ... Sep 23 10:24:37 and that seems a bit difficult Sep 23 10:24:50 i'm recompiling monotone with a newly compiled libboost Sep 23 10:25:01 (i'm on debian btw, where monotone seems outdated) Sep 23 10:25:13 sarge? Sep 23 10:25:24 http://venge.net/monotone/downloads/monotone_0.22-0sarge1_i386.deb Sep 23 10:25:33 or sid? Sep 23 10:25:33 http://venge.net/monotone/downloads/monotone_0.22-0.1_i386.deb Sep 23 10:25:45 those should be statically linked versions Sep 23 10:26:01 unstable :-) Sep 23 10:26:05 ah. interesting. Sep 23 10:35:15 night all Sep 23 10:49:35 hummm ..... libboost recompiled. Sep 23 10:49:44 monotone is doing testing ... and it seems to work this time :-) Sep 23 10:52:14 Hi all Sep 23 10:53:37 Quick question! I have two hosts, where I want to use the same monotone repository on. Is it possible to avoid the initial pull process by copying e.g. oe.db and org.openembedded.dev dir ? Sep 23 10:54:50 uv1: yes Sep 23 10:55:02 pH5: Is that all ? Sep 23 10:55:18 uv1: yup, everything you need is in the .db Sep 23 10:55:40 pH5: Thanks, will try that tomorrow morning ;-) Sep 23 10:55:47 bye Sep 23 10:55:50 uv1: np Sep 23 10:56:22 * CosmicPenguin shakes his head Sep 23 11:25:57 mhm, who has mono ipk? Sep 23 12:20:02 * benoit has succeeded in installing monotone 0.22 (after recompiling libboost) Sep 23 12:24:15 hi Sep 23 13:13:54 is that normal if monotone is taking a lot of CPU ? Sep 23 13:14:46 yep Sep 23 13:14:52 mt seems to be a bit of a dog.....but it works OK Sep 23 13:15:01 kergoth: got a sec? Sep 23 13:15:17 it download 31M and say : "verifying new revisions".... Sep 23 13:16:14 benoit: yep, thatll takea bit Sep 23 13:16:15 MSpin: yep Sep 23 13:17:33 is there a preferred way to maintain a private branch with mt? I did one, but everytime I propagate from org.oe.dev to mine I get tons of files needing merging...even though I didn't touch 'em Sep 23 13:20:08 heh, no idea actually. i havent messed with oe much lately, partly due to mt Sep 23 13:22:00 damn Sep 23 13:22:02 kergoth: oz is looking nice :) Sep 23 13:22:42 I agree! I finally got kismet_qt ported over..I should submit the package Sep 23 13:23:01 please do! Sep 23 13:24:16 raduga: i bet. the 2.6 stuff is especially cool Sep 23 13:25:46 kergoth: i'm afraid to run 2.6 on a pxa270 so far... Sep 23 13:25:55 reports are, keyboard still not supported? Sep 23 13:26:00 is that accurate? Sep 23 13:31:44 mhm, who have got mono ipk? Sep 23 13:36:59 mono install error: :( Sep 23 13:37:01 http://rafb.net/paste/results/fPccTm59.html Sep 23 17:01:33 wheres our man JustinP ? Sep 23 17:01:48 i am wondering why he is using oesources for e17cvs stufff ... Sep 23 17:16:20 i've been wondering Sep 23 17:16:43 is there a reason we cant vacuum the working dir after building a package? **** BEGIN LOGGING AT Sat Sep 24 00:19:53 2005 Sep 24 00:20:15 for what>? Sep 24 00:20:36 emte: before build opie-image Sep 24 00:20:40 task-bootstrap is a base for most images Sep 24 00:20:51 opie-image will call it Sep 24 00:21:00 and build what it needs Sep 24 00:21:01 that's what I through Sep 24 00:21:28 but it fail and saying can't find task-bootstrap ... however I can bitbake task-bootstrap Sep 24 00:21:38 emte: any idea? :) Sep 24 00:22:25 no, it seems liek something has changed Sep 24 00:22:33 you sure your in the right directory? Sep 24 00:22:51 I think so :o Sep 24 00:23:05 ... Sep 24 00:23:09 * AvengerMoJo confused Sep 24 00:24:20 so am i Sep 24 00:24:32 i need sleep Sep 24 00:24:33 emte: thanks anyway Sep 24 00:24:38 good night :) Sep 24 00:24:54 trying to learn how to program sockets ... Sep 24 00:25:32 I see Sep 24 00:25:53 :) Sep 24 00:53:27 error: unknown field `attach' specified in initializer Sep 24 00:53:31 for hostap Sep 24 01:45:30 good morning all Sep 24 02:11:41 http://ewi546.ewi.utwente.nl/OE/ will from now on contain daily diffs too Sep 24 02:38:23 hi koen|away Sep 24 02:38:47 hostap get problem with the new 2.6 pcmcia interface ..... I'm trying to write a patch here Sep 24 02:41:59 Didn't RP already solve that? Sep 24 02:42:58 packages/hostap/hostap-modules-0.4.4/kernel_updates.patch Sep 24 02:43:00 ? Sep 24 02:49:39 oh Sep 24 02:49:45 I didn't see that patch Sep 24 02:49:49 when did it happen? Sep 24 02:49:51 :P Sep 24 02:50:26 * AvengerMoJo spend 2 hours on something that someone already did :( Sep 24 02:50:33 hhe Sep 24 02:50:34 hehe Sep 24 04:39:56 good morning Sep 24 04:40:30 [09:17] but it fail and saying can't find task-bootstrap ... however I can bitbake task-bootstrap Sep 24 04:40:38 update your metadata, i fixed that some days ago Sep 24 04:40:43 you should update more often :D Sep 24 04:41:12 or at least mirror the ciabot to see what kind of updates appear Sep 24 04:41:21 s/mirror/read/ Sep 24 04:44:42 ~seen pigi Sep 24 04:44:48 pigi was last seen on IRC in channel #oe, 2d 15h 51m 27s ago, saying: 'bbl'. Sep 24 04:52:14 good morning pb_ Sep 24 04:52:54 pb_: can you explain to me what's the deal with the thumb functionality on recent arm variants? Sep 24 04:52:56 good morning mickeyl Sep 24 04:53:23 can you be any more specific? I'm not sure which particular deal you are thinking of. Sep 24 04:54:07 I read in an arm architecture overview paper that one of the great additions in armv4 was the thumb instruction set Sep 24 04:54:17 i wonder what excactly is improved by that? Sep 24 04:55:14 i'm not an arm expert, but i think the size of instruction is reduced (from 32->16?) right? Sep 24 04:55:33 hmm so it's just about more compact code? Sep 24 04:55:42 pretty much, yeah Sep 24 04:55:48 in ARM, all instructions are 32 bits in size. Sep 24 04:55:58 in Thumb, many instructions are reduced to 16 bits. Sep 24 04:56:52 i see. so it's up to the toolchain to support that. I wonder which gcc version introduced that support Sep 24 04:57:18 either 2.8 or 2.95, I think Sep 24 04:57:57 I'm pretty sure it wasn't in 2.7. Sep 24 05:17:01 mickeyl: you have the user-mode-linux 5ki11z, right? Sep 24 05:17:59 it's been a while, but i used to work with it, yeah Sep 24 05:18:58 if I have a disk image of a regular Fedora installation, and I want to boot it inside a uml instance, do you know what I would need to change? Sep 24 05:21:49 you can either convert it to an uml cow image Sep 24 05:21:57 or try to boot it directly Sep 24 05:22:24 okay. what's a cow image? Sep 24 05:22:41 copy on write, that's the one of the native uml file systems Sep 24 05:22:52 uml comes with userland utilities to setup a cow image Sep 24 05:23:16 ah right, cute Sep 24 05:23:26 okay, I'll have a go at that. Sep 24 05:23:46 one of our servers is running some old fedora at the moment. I want to convert it to debian, but I need to keep the old os install around for a while. Sep 24 05:24:12 hail zecke Sep 24 05:24:13 yeah, uml is pretty nice for these things Sep 24 05:24:39 hey Sep 24 05:24:52 recent versions should also be able to boot from ext2 loop images and parts of the hostfs, but i never used this functionality. COW is the oldest and most stable uml fs. Sep 24 05:24:55 hi zecke Sep 24 05:24:59 * mickeyl shopping Sep 24 05:25:15 mickey|city: cya later Sep 24 05:25:22 WEEK END yeah, yipee Sep 24 05:26:07 mickey|city: I'd rather have a CHECK ?= "${@bb.fatal('Remove this line if you really, *really* want to build a inversioned distro')}" in the unversioned distro.conf Sep 24 05:27:02 and CHECK = "OK" in the versioned distro.conf Sep 24 07:42:44 Hi all - anyone here ? Sep 24 07:42:53 nope Sep 24 07:44:27 Hi koen Sep 24 07:45:12 I had a problem few hours ago, to get OE fetch data behind a http/ftp proxy (in my office). Is there a way arround that ? Sep 24 07:54:39 mhm when i run `bitbake minimo` i have error (404) of minimo's sources Sep 24 07:54:58 univac: well then minimo has errors? Sep 24 07:55:24 it can;t download minimo's sources (404) Sep 24 07:56:25 uv1, I'd like to know too Sep 24 07:56:33 uv1, what I did was... Sep 24 07:57:03 uv1 used transconnect to ssh through the http proxy to an outside machine Sep 24 07:57:23 used danteserver on the outside machine to have a socks4 proxy Sep 24 07:57:47 then used proxychains on the development machine to use the socks4 proxy on my outside machine through an ssh tunnel Sep 24 07:57:55 pretty nasty hack Sep 24 07:58:54 NOTE: Task failed: Fetch failed: http://ftp.info-zip.org/pub/infozip/src/zip231.tar.gz Sep 24 08:00:48 uv1, I think (not sure) that oe respects my http_proxy environment variable. unfortunately, it didn't use it for ftp Sep 24 08:01:44 you could also try having http_proxy not set and using proxychains to connect out of your http proxy Sep 24 08:01:54 mhm who has minimo ipk? Sep 24 08:01:57 newest Sep 24 08:04:48 univac: that would be a question for the channel of your distro Sep 24 08:05:00 minimo is in cvs, so a 404 is very strange Sep 24 08:05:37 MP0: Thanks, I will try the proxy settings on monday morning. Any hints for cvs and svn ? How did you get arround the ftp problem ? Any short description ? Sep 24 08:05:52 MP0: I could provide the info on www.uv-ac.de/openembedded then Sep 24 08:07:48 My personal solution was Sep 24 08:07:53 OE uses wget, so if you can wget to work, OE will work too Sep 24 08:07:58 (for http/ftp) Sep 24 08:08:50 uv1, like koen says, you should try setting the http_proxy first... see if your corporate proxy will let ftp through Sep 24 08:08:59 uv1: did you post to the oe ml? Sep 24 08:09:08 export http_proxy=http://someproxy:someport/ Sep 24 08:09:29 (make sure you add the trailing / as sometimes wget doesn't accept it otherwise) Sep 24 08:10:35 as for svn... my solution is to have a proxy on the outside Sep 24 08:10:55 I used transconnect to get an ssh tunnel to the machine on the outside Sep 24 08:11:22 lardman: Yes. Sep 24 08:11:26 it's rather slow, but other than changing the structure of IT, what can I do. Sep 24 08:12:41 MP0: Any hint, how to setup your box that for this special case ? Sep 24 08:13:07 uv1, either set up your proxy to work on port 443 Sep 24 08:13:32 and use proxytunnel to go through both your corporate proxy and your outside proxy Sep 24 08:13:43 or Sep 24 08:13:55 (what I did, since it was already semi-set up) Sep 24 08:14:13 I had my outside machine running ssh on port 443 (always handy to have) Sep 24 08:14:28 ie, I put "port 443" in /etc/sshd/blahblahblah Sep 24 08:14:35 whatever the config file is called Sep 24 08:14:40 and restarted sshd Sep 24 08:14:53 since sshd is running on port 443, I can create a tunnel Sep 24 08:15:22 now, ssh has a neat feature Sep 24 08:15:26 the -D option Sep 24 08:15:42 it acts like a socks4/socks5 server Sep 24 08:16:05 so Sep 24 08:16:29 using transconnect or proxychains Sep 24 08:16:46 MP0: Thanks, will try that as well. Sep 24 08:16:53 proxychains ssh -D 8080 MP0@someserver Sep 24 08:17:11 the only problem I have with that approach is Sep 24 08:17:17 (ie, using ssh) Sep 24 08:17:41 is I need to use a seperate program to get the ssh connection through the corporate proxy Sep 24 08:18:04 and one to get oe to use my ssh proxy Sep 24 08:18:36 that's why I use both proxychains and transconnect... one of them is configured to get ssh out of the corprorate proxy, the other is configured to make programs use my ssh proxy Sep 24 08:19:33 if you have your outside proxy server proxying on port 443 already (instead of ssh on 443), then you can use proxychains alone without another program. Sep 24 08:20:04 cuz proxychains, as its name implies, allows you to chain proxies... get through the corprorate proxy, and get through your outside proxy Sep 24 08:20:25 MP0: May I ask you to send me an email with that info ? Sep 24 08:20:45 what's your email address? Sep 24 08:21:07 MP0: w1@schulte-ac.de, Thanks again. Sep 24 08:23:25 http://paul.bluereboot.net/proxy.log Sep 24 08:24:02 (you can email that url to yourself if you want) Sep 24 08:28:04 MP0: ;-) Sep 24 08:52:13 I have a problem to get a x86 OE build running. The build process comlains glibc cannot be build. Heres the error message. Sep 24 08:52:14 /data/scdev2/build-x86/tmp/work/glibc-2.3.5+cvs20050627-r0/build-i386-linux/linuxthreads/libpthread_pic.a(pthread.os): In function `pthread_initialize': Sep 24 08:52:14 pthread.c:(.text+0x1eb): undefined reference to `_res' Sep 24 08:52:14 /data/scdev2/build-x86/tmp/work/glibc-2.3.5+cvs20050627-r0/build-i386-linux/linuxthreads/libpthread_pic.a(pthread.os): In function `__pthread_reset_main_thread': Sep 24 08:52:15 pthread.c:(.text+0x91b): undefined reference to `_errno' Sep 24 08:52:17 pthread.c:(.text+0x933): undefined reference to `_h_errno' Sep 24 08:52:19 pthread.c:(.text+0x939): undefined reference to `_res' Sep 24 08:52:20 Any ideas ? Sep 24 08:53:27 http://handhelds.org/hypermail/oe/24/2490.html Sep 24 08:55:01 http://216.239.59.104/search?q=cache:YnUuq6v6s3wJ:ibot.rikers.org/%2523gpe/20050809.html.gz+oe+pthread.c:(.text%2B0x91b):+undefined+reference+to+%60_errno%27&hl=en&client=safari Sep 24 08:56:26 to quote "I don't think linuxthreads is supported on x86 anymore. Try with NTPL. have a look at epia.conf; I think it has the bits" (pb_) Sep 24 09:00:06 MP0: Will try, thanks again. Sep 24 09:05:39 morning Sep 24 09:08:32 back here. Sep 24 09:11:31 * zecke considers buying a Triton Starter Kit Sep 24 09:13:14 How much does one of those cost? Sep 24 09:13:50 pb_: They have no prices, I send an inquiry Sep 24 09:14:07 pb_: the cheapest DIMM Module (PXA255) costs 88 Euro Sep 24 09:21:15 zecke: I found some on a UK site, but they seem quite expensive Sep 24 09:21:16 http://shop.directinsight.co.uk/catalog/index.php?cPath=67 Sep 24 09:22:25 pb_: yes quite expensive, maybe I can get some discount Sep 24 09:22:25 * zecke considers exploting his contacts... Sep 24 09:22:35 hrw|gone: ping Sep 24 09:23:06 good plan Sep 24 09:23:34 pb_: actually I want something where I can hook a disk on and actually start applications (I am not too keen on qemu-arm) Sep 24 09:23:46 pb_: contacts are alwasys good Sep 24 09:24:04 right Sep 24 09:24:18 I could use a fast arm machine with a local disk, too. Sep 24 09:24:38 at the moment the best one I have is the old netwinder, which is only 275MHz. Sep 24 09:28:05 hi guys Sep 24 09:28:28 cheers Bernardo Sep 24 09:28:49 hey mickeyl Sep 24 09:28:51 pb_: yeah, I could deserve some SHARK firmware Sep 24 09:28:59 pb_: so I could actually boot something on it... Sep 24 09:29:03 finally got my desktop machine up Sep 24 09:29:28 and monotone just finished pulling Sep 24 09:32:11 zecke: heh. can't you download the shark bits from the internet? Sep 24 09:33:05 Bernardo: as we're approaching OZ 3.5.4, i'd be grateful if you could bring it in shape for the C1000 Sep 24 09:33:16 pb_: sadly not (not the firmware) Sep 24 09:34:20 ah, that is unfortunate Sep 24 09:34:44 you could port the ohh bootldr to the shark, I guess. Sep 24 09:34:47 :-} Sep 24 09:35:18 pb_: yeah, I got bored after self testing all four sharks (with Open Firmware) Sep 24 09:36:37 mickeyl: do we set a preferred gcc provider for OZ 3.5.4? Sep 24 09:36:43 yep Sep 24 09:36:53 iirc we have set one already in the conf Sep 24 09:36:57 i vote for 3.4.4 Sep 24 09:37:04 or is it 3.3.4? Sep 24 09:37:08 3.4.4 Sep 24 09:37:10 anyway, the latest < 4.0 Sep 24 09:37:30 hi zecke mickeyl Sep 24 09:37:51 yo chouimat Sep 24 09:38:08 mickeyl: my move to Mac OS X takes longer than I had expected Sep 24 09:38:13 mickeyl: the linker drives me nuts... Sep 24 09:38:16 heh Sep 24 09:38:22 does Opie run yet? Sep 24 09:38:51 mickeyl: I'm currently fighting with getting Qt3Mac' QVFB running Sep 24 09:39:01 mickeyl: well eilers and harlekin already did most of the work Sep 24 09:39:16 mickeyl: but libtool makes sense on OS X Sep 24 09:39:49 mickeyl: will try, yes Sep 24 09:39:53 mickeyl: binutils/gcc utilize SO_NEEDED to find the needed libs Sep 24 09:40:02 mickeyl: OS X simply errors... Sep 24 09:40:34 CoreDump does a excelent job anyway, much better than what I did Sep 24 09:40:57 right now I have to see why calling bitbake gives me a python error Sep 24 09:41:19 Bernardo: switched from python 2.3 to 2.4? Sep 24 09:42:05 zecke: I have 2.4.1 here Sep 24 09:43:34 OT:2.4.2 is around the corner... Sep 24 09:44:08 mickeyl: any idea why ./setup bdist_rpm fails? (sehr ungenau...) Sep 24 09:44:32 ./setup.py ? Sep 24 09:44:43 what's the traceback? Sep 24 09:45:06 mickeyl: it wants config/Makefile (somewhere in the python directory) Sep 24 09:45:13 bitbake meta-opie Sep 24 09:45:13 Traceback (most recent call last): Sep 24 09:45:13 File "/mnt/other/oe/bitbake/bin/bitbake", line 27, in ? Sep 24 09:45:13 from bb import utils, data, parse, debug, event, fatal Sep 24 09:45:13 File "/mnt/other/oe/bitbake/lib/bb/data.py", line 137 Sep 24 09:45:14 <<<<<<< .mine Sep 24 09:45:16 hmm yes Sep 24 09:45:16 ^ Sep 24 09:45:18 SyntaxError: invalid syntax Sep 24 09:45:23 Bernardo: hehe merge error? Sep 24 09:45:36 you need to have python-dev installed for distutils to work Sep 24 09:46:05 I have my dau on my lap now, bbl Sep 24 09:46:27 have py from source, full headers and libs Sep 24 09:46:34 mickeyl: I thought I had them... Sep 24 09:46:55 mickeyl: but dpkg -S did not reveal any package (I could have switched to Mandriva though) Sep 24 09:46:59 * Bernardo is one fing typing Sep 24 09:48:25 zecke: will remove bitbake dir and get it again Sep 24 09:49:54 Bernardo: why not do svn revert? Sep 24 09:51:16 zecke: because it didn't change anything... :) Sorry by the terseness above, my daughter wanted to check what I was doing Sep 24 09:51:47 * zecke won his game cube Sep 24 09:52:15 Bernardo: well svn status says what? or did you delete and started over again Sep 24 09:53:34 zecke: I was just impacient now, and moved the dir and pulled it again, and now it works... Sep 24 09:54:09 Bernardo: it is obvious it works Sep 24 09:55:07 something got corrupted... but now meta-opie is starting to build Sep 24 09:55:21 will see later if the fact I am building on amd64 will break anything Sep 24 09:58:35 Bernardo: corrupted == merge conflicts? Sep 24 10:00:25 zecke: no idea, probably I still had some CoW patches applied to the bitbake dir Sep 24 10:02:36 now it failed building gnu-config-native, seems it can't find quilt - time to check again if I have the base stuff... Sep 24 10:03:29 I'm getting disconnected now (need to get monitor mode for his airport working) Sep 24 10:50:43 zecke: it seems I have a too recent version of autoconf. I get "/usr/bin/autoreconf: unrecognized option `--exclude=libtoolize'" building libtool-native Sep 24 10:51:07 Bernardo: what fails to build? Sep 24 10:51:18 Bernardo: autoconf-native should get build Sep 24 10:52:40 patcher-native builds, gnu-config-native, and then libtool-native fails Sep 24 10:52:48 brb - got to help with the dinner Sep 24 11:01:31 The last SDLmain bits got kicked out Sep 24 11:01:45 I'm waiting for my toolchain to be compiled Sep 24 11:01:55 rock Sep 24 11:16:21 hrw|gone: ping Sep 24 11:17:04 zecke: autocof-native had built, I just rebuilt it, will test Sep 24 11:40:53 kergoth: ping Sep 24 11:45:32 hi Sep 24 11:45:46 hi hrw Sep 24 11:47:17 libxine patches really need cleanup Sep 24 11:47:32 some looks like being ready to post upstream Sep 24 11:47:33 hrw: I wanted to ask that Sep 24 11:47:44 NOTE: package libxine-1.1.0-r0: task do_compile: started Sep 24 11:48:05 zecke: my 1st build of libxine 1.1.0 progressing Sep 24 11:48:34 ok. first bug appeared Sep 24 11:49:25 libxine-libavcodec.patch need some serious update Sep 24 11:49:43 zecke: bitabake -c clean autoconf-native, then bitbake autoconf-native, and now meta-opie is building... No idea what went wrong first time around Sep 24 11:49:45 or synchro with ffmpeg Sep 24 11:53:05 ~lart monotone update Sep 24 11:53:06 * ibot slams monotone update against a large cement Tux Sep 24 11:53:12 monotone: misuse: 227 missing files Sep 24 11:53:35 looks like I would need clean checkout.. Sep 24 11:55:21 yep.. libxine-libavcodec.patch should be replaced with current ffmpeg Sep 24 12:00:09 AFK.. Sep 24 12:03:53 hi Bernardo ! you there? Sep 24 12:05:09 hi benoit Sep 24 12:05:27 yep, starting to build a image for my c1000 after a few months Sep 24 12:05:43 Bernardo: i just started using oe, i just build 1058 packages doing bitbake nano :-) Sep 24 12:05:56 Bernardo: c1000 is your zaurus? Sep 24 12:06:00 yes Sep 24 12:06:30 benoit: I've just started now, the last I looked I was now building gcc-cross Sep 24 12:06:58 Bernardo: great. we are at the same level then :-) Sep 24 12:07:48 just found freetype failed to build because it wounldn't find libtool... Sep 24 12:07:54 seems the amd64 Sep 24 12:08:06 has some pecularites still... Sep 24 12:08:37 i'm building on a plain x86 for arm. Sep 24 12:09:52 I finally got my 64 to boot in 64-bits native, using that as my build machine Sep 24 12:10:12 got to go, brb, my daughter (19 months) wants to dance... Sep 24 12:44:53 :o Sep 24 12:44:54 hi Sep 24 12:50:44 Hi all, Sep 24 12:55:09 how is going Bilbolodz|home Sep 24 12:56:34 a pull form scratch: Sep 24 12:56:36 real 473m43.999s Sep 24 12:56:36 user 462m25.557s Sep 24 12:56:36 sys 7m44.445s Sep 24 12:57:30 AvengerMojoFine, thanks, I've successfully build gpe image, but something is wrong: touch screen is not working Sep 24 12:57:54 (i'm building image for Simpad) Sep 24 12:58:40 .....koen are you very patient Sep 24 12:58:55 Bilbolodz|home: hum .... did you try calibrate? Sep 24 12:59:18 AvengerMoJo: that's on a remote server Sep 24 12:59:24 I cant calibrate screen It don't react on stylus! Sep 24 13:00:12 (on "first" run calibrate screen is lunched automatically) Sep 24 13:00:38 I see cross but when I'm touching it nothing happend Sep 24 13:01:00 koen I see Sep 24 13:01:19 Bilbolodz|home: I don't get Simpad ... but what kind of kernel you are using? Sep 24 13:01:30 2.4 Sep 24 13:01:39 hum .... that should work I assume Sep 24 13:01:47 ask around and see Sep 24 13:01:54 :) Sep 24 13:01:56 good luck Sep 24 13:02:13 Bernardo|away: did you apply all patches? Sep 24 13:04:29 strange thing that Devernay images (http://devernay.free.fr/hacks/opensimpad/) are working! I asked him what magic trick is he doing but I've got no answer yet. Sep 24 13:04:46 I see Sep 24 13:04:57 .... Sep 24 13:04:58 Bilbolodz|home: did you apply all patches? Sep 24 13:05:05 zecke: ewi is up again Sep 24 13:05:06 Bilbolodz|home: applying all patches could be the magic Sep 24 13:05:12 koen: thanks for your botmail Sep 24 13:05:22 koen: does ewi run java as well? Sep 24 13:05:24 I applied 2 patches I found for Simpad Sep 24 13:05:35 zecke: sablevm is installed Sep 24 13:05:50 Bilbolodz|home: what does your /ets/ts.conf say? Sep 24 13:05:51 zecke: and my job was to get cvm working in OE Sep 24 13:07:37 zecke: simpad login: root Sep 24 13:07:37 root@simpad:~# more /etc/ts.conf Sep 24 13:07:37 module_raw input Sep 24 13:07:37 module pthres pmin=1 Sep 24 13:07:37 module variance delta=30 Sep 24 13:07:37 module dejitter delta=100 Sep 24 13:07:39 module linear Sep 24 13:08:40 Bilbolodz|home: well then apply all patches ;) Sep 24 13:09:11 what should I look? tc.conf? Sep 24 13:09:15 ts.conf? Sep 24 13:09:36 tslib? Sep 24 13:10:09 Bilbolodz|home: how many patches did you apply? Sep 24 13:11:34 zecke: 2 patches 302 (simpad: allow 2.4 as well as 2.6 kernel) and 305 (patch: tslib simpad - support 2.4 and 2.6 kernels) Sep 24 13:12:06 Bilbolodz|home: and the patch got a line with +module_raw ucb1x00 Sep 24 13:12:17 Bilbolodz|home: but your ts.conf says module_raw input Sep 24 13:12:53 zecke: pah, ucb1x00 is so last year Sep 24 13:12:53 Bilbolodz|home: so something is not right(tm) Sep 24 13:13:21 I bet the patch didn't bump PR Sep 24 13:13:24 pb_: don't tell me, I don't understand why people waste time with the 2.4er linux kernel... Sep 24 13:13:39 koen: well I do not intend to apply the patch anyway Sep 24 13:13:43 OK I will checkt it. Sep 24 13:14:01 zecke: yeah, those crazy retrocomputing guys Sep 24 13:17:17 going back to the tv set Sep 24 13:17:51 zecke: I've checked sources of 3 files changed in this patch all file looks good Sep 24 13:17:57 zecke: did you crack the key yet? Sep 24 13:18:21 I "touched" this files and trying again Sep 24 13:18:47 koen: hehe, no I did not succeed in cracking my own key Sep 24 13:19:01 koen: and the viha modules seem to not work with my Airport2 card Sep 24 13:20:30 Bilbolodz|home: bitbake -b .../tslib_cvs.bb -cclean Sep 24 13:20:32 Hello. I have a question. How can i make OE to use -msoftfloat (the gcc software emulation) and not FPA ( looks like it is used by default) ? Sep 24 13:20:34 Bilbolodz|home: bitbake -b .../tslib_cvs.bb Sep 24 13:20:48 krokas__: TARGET_FPU = "soft" Sep 24 13:21:14 k. let me try... Sep 24 13:24:25 zecke: moment I'm building whole image Sep 24 13:28:47 pb_: are the eabi patches for glibc already writen and waiting for a blessing, or are we waiting for csl/nico to write them? Sep 24 13:38:16 Bilbolodz|home: no need to create a whole image Sep 24 13:38:26 Bilbolodz|home: simply upload and install the new tslib package Sep 24 13:47:35 is that possible to shorten the startup time? Sep 24 13:49:00 which starup time? Sep 24 13:49:23 wow beat spain, yeah... actually spain and the referees... Sep 24 13:52:59 :) zecke the system startup Sep 24 13:53:41 zecke: I'm wondering if we can start qpe after the kernel is started before all the modules and system checking Sep 24 13:54:33 AvengerMoJo: well do not use sysvinit Sep 24 14:04:10 hi reenoo Sep 24 14:04:28 anyone familiar with monotone's hooks? Sep 24 14:05:26 hey pb_ Sep 24 14:13:15 http://pastebin.ca/23803 Sep 24 14:13:23 re Sep 24 14:13:46 grrr, now I've problem with bootloader: Scanning flash JFFS2 = Offset looping infinitely, any idea how to fix it? Sep 24 14:13:47 that's what I get when bitbake meta-opie gets to freetype Sep 24 14:14:38 zecke: what do you mean? .... not using it? Sep 24 14:16:05 Bernardo: libtool on amd64? maybe its broken? Sep 24 14:16:20 hrw on amd64, yes Sep 24 14:16:44 but the native stuff should all work, after all, it's all built from source Sep 24 14:17:08 AvengerMoJo: use a better suited init system Sep 24 14:17:16 heh Sep 24 14:17:26 hrw|tv: all fresh and clean, reinstalled last weekend Sep 24 14:17:31 If I had a nickle for every time a gentoo person said that Sep 24 14:17:41 koen: not gentoo Sep 24 14:17:43 "but it's built from source!" Sep 24 14:17:47 Bernardo: I know Sep 24 14:18:10 Bernardo: sourcemage, right? Sep 24 14:18:19 zecke: what will be my choose? Sep 24 14:18:23 choice Sep 24 14:18:28 and I've just checked the base system (ldd check, md5 check, symlinks, the works) Sep 24 14:18:35 koen: yes Sep 24 14:19:40 AvengerMoJo: use another init system (probably even inittab based) Sep 24 14:19:57 ok ... Sep 24 14:20:08 zecke: I bet we could use bitbake as initsystem Sep 24 14:20:09 zecke: so I just basically skip all the init Sep 24 14:20:20 koen what do you mean Sep 24 14:20:22 zecke: let's patent that, sell it to CELF and get filthy rich Sep 24 14:20:51 * zecke deserves food Sep 24 14:21:12 * koen scp food zecke:stomach Sep 24 14:21:24 AvengerMoJo: use init=qpe :) Sep 24 14:22:18 hrw|tv: .... Sep 24 14:22:41 hrw|tv: I will wonder ... how dead my system will be .. if I do that Sep 24 14:22:55 :) Sep 24 14:23:20 it will be *really* fast Sep 24 14:23:36 beacuse init is much faster when it accesses the framebuffer directly Sep 24 14:23:39 *cough* Sep 24 14:23:43 but not working ;) Sep 24 14:24:43 koen: well DirectFB is fast ;) Sep 24 14:25:22 zecke: speaking of direcfb ..is there any port for qt on directfb Sep 24 14:25:23 * koen mv OE DirectOE Sep 24 14:25:37 AvengerMoJo: no recent one AFAIK Sep 24 14:25:42 zecke: I kind of want to make standalone application Sep 24 14:26:06 AvengerMoJo: then QtE is a good candidate Sep 24 14:26:14 qte Sep 24 14:26:20 AvengerMoJo: with it any application can be the GUI server Sep 24 14:26:34 AvengerMoJo: and it starts without showing that black/white X image Sep 24 14:26:36 meaning no qpe ... just install the qte library Sep 24 14:26:51 AvengerMoJo: just install qte Sep 24 14:27:04 AvengerMoJo: maybe you want to link to libopie2 as well but that is optional Sep 24 14:27:12 zecke: any doc on how to do that Sep 24 14:27:30 AvengerMoJo: doc.trolltech.com Sep 24 14:27:33 zecke: I just want to make a stand alone application that start right after the kernel started Sep 24 14:27:35 zecke: you can start X without that b/w screen Sep 24 14:27:48 AvengerMoJo: create your app, start it with -qws Sep 24 14:28:03 AvengerMoJo: or use QApplication app(arcg,argv, GUIServer) Sep 24 14:28:06 zecke: oh ... Sep 24 14:28:16 koen: well but you need a client to connect to it Sep 24 14:28:16 zecke: that will work already? Sep 24 14:28:21 AvengerMoJo: yes Sep 24 14:28:32 zecke: thanks let me try Sep 24 14:28:35 AvengerMoJo: and using qvfb you can test and develop your app on your desktop system Sep 24 14:28:40 hmm NOTE: package libxine-1.1.0-r0: task do_compile: completed Sep 24 14:28:52 zecke: did siemens release that dvb-h yet? Sep 24 14:28:54 but idct not optimized on arm Sep 24 14:29:05 koen: hehe, that was a project ;) Sep 24 14:29:11 koen: eventually we will see a product Sep 24 14:29:19 zecke: man ..... Sep 24 14:29:29 zecke: you save me a lot of time Sep 24 14:29:31 :( Sep 24 14:29:34 AvengerMoJo: too much input at once? Sep 24 14:29:34 * AvengerMoJo feeling stupid Sep 24 14:29:45 AvengerMoJo: no reason to do that Sep 24 14:30:01 zecke: thanks ... I can start my application without start the qpe Sep 24 14:30:11 then this pretty much solve all my problem Sep 24 14:30:25 I can start my application in inittab ... Sep 24 14:30:32 or rc whatever Sep 24 14:30:40 make it look much faster Sep 24 14:30:40 AvengerMoJo: http://doc.trolltech.com/2.3/qvfb-qws.html Sep 24 14:30:42 AvengerMoJo: right Sep 24 14:30:57 zecke: because I already make a faker Sep 24 14:31:05 AvengerMoJo: http://doc.trolltech.com/2.3/running-qws.html Sep 24 14:31:20 zecke: thanks alot Sep 24 14:31:27 zecke: I've decided I want a porsche, so we have to hurry with getting filthy rich Sep 24 14:31:40 koen: get a VW, it is owned by Porsche Sep 24 14:31:46 koen hehe Sep 24 14:32:00 koen: one year and I hope to be done with uni :} Sep 24 14:32:01 zecke: a VW touareg would be fine too Sep 24 14:32:10 damn Sep 24 14:32:15 2.5 years over here Sep 24 14:32:16 koen: I've a beatle, I would exchange it with a better one Sep 24 14:32:29 (cabriolet, red, us/california edition) Sep 24 14:32:30 * koen doesn't have a car Sep 24 14:32:43 koen: I would not call that a car ;) Sep 24 14:32:49 VW suxx Sep 24 14:33:00 hrw|tv: their homepages or also the cars? Sep 24 14:33:30 their homepage, Polish distributor, their materials for web Sep 24 14:33:45 also german VW webmasters - those ones suxx more Sep 24 14:33:59 NOTE: package libxine-1.1.0-r0: task do_build: completed Sep 24 14:34:17 zecke: which libxine-plugin-vo does opie need? Sep 24 14:35:40 hrw|tv: none Sep 24 14:35:48 hrw|tv: op2 implements a custom vo plugin Sep 24 14:35:51 looks like my libxine 1.1.0 builds atleast Sep 24 14:36:03 zecke: for the doc .. I can even use virtual display ... Sep 24 14:36:30 gprs sucks Sep 24 14:36:51 zecke: can I do standalone app on display 1 and then when the system started .. qpe started... using display 2 ... and i just add some icon for changing the display? Sep 24 14:36:52 what is dbos? Sep 24 14:37:26 AvengerMoJo: right Sep 24 14:37:35 koen: dbos, dunno? you mean dbs? Sep 24 14:37:36 or better, what is bluez-utils-nodbos? Sep 24 14:37:44 hehe Sep 24 14:38:11 R DJ strikes back Sep 24 14:38:19 lol Sep 24 14:38:22 jbowler: hey Sep 24 14:38:29 hi zecke Sep 24 14:38:36 jbowler: can I reject a commit/netsync_receive with a lua hook? Sep 24 14:38:49 ibot: kill monotone update Sep 24 14:38:51 * ibot shoots a hyper-charged electron gun at monotone update Sep 24 14:39:18 jbowler: a fresh monotone pull org.openembedded.dev is 52MB Sep 24 14:39:18 monotone: misuse: 206 missing files Sep 24 14:39:26 zecke: I believe so, there are certainly hooks for each new revision Sep 24 14:39:45 it only took 8 hours to do a complete checkout Sep 24 14:40:04 koen: 8h on 2xOpteron? Sep 24 14:40:06 koen: that is the problem... Sep 24 14:40:18 hrw|tv: yes Sep 24 14:40:29 hrw|tv: but it's a single thread Sep 24 14:40:41 koen: my solution is to have a separate slave database which is automatically pushed just org.openembedded.{dev,dreambox} then make the snapshot from that. Sep 24 14:41:02 but a 1.8 ghz opteron beats a 3.6 ghz xeon with monotone Sep 24 14:41:13 monotone: I HATE you! Sep 24 14:41:16 jbowler: I'm working on that too Sep 24 14:41:28 argh... after last pull/merge my working copy is fucked Sep 24 14:42:13 jbowler: http://ewi546.ewi.utwente.nl/OE/ will contain daily tarballs, diffs between daily tarballs and a daily updated db-snapshot Sep 24 14:42:27 the first two are a cronjob now Sep 24 14:42:41 I'm working on the db-snapshot Sep 24 14:43:53 koen: I'm starting to wonder if, maybe, we have some big binary checked in to something somewhere (maybe something from org.nslu2-linux.*?) Sep 24 14:44:39 170MByte == 3*52MByte, that seems wrong unless there is some big thing accidentally in there. Sep 24 14:45:02 it's because monotone doesn't have empty merges Sep 24 14:45:32 all the cset that are part of a merge are duplicated by the merge Sep 24 14:46:05 add some extra branches in the mix Sep 24 14:46:13 and a lot of propagates Sep 24 14:46:20 shake it well Sep 24 14:46:25 170 MB Sep 24 14:47:56 what is the firewire storage driver? Sep 24 14:48:09 the snapshot is the original db Sep 24 14:48:26 so it contains all branches (dev, dreambox, nslu2, zecke, koen, etc) Sep 24 14:49:03 I guess that makes sense, though my head starts to spin thinking about it... Sep 24 14:51:11 not having empty merges annoys me a lot Sep 24 14:54:13 cya tomorrow Sep 24 14:55:10 'night all Sep 24 14:57:20 'night guys Sep 24 14:58:17 heh.. I just noticed that some patches from 1.0 I forgot Sep 24 15:04:12 hrw|tv: cya tomorrow Sep 24 15:04:41 cu zecke Sep 24 15:06:39 my sbp2 is not responding ... Sep 24 15:13:40 CIA-4: ping Sep 24 15:30:46 can I update the hdimage1.tgz and make it only install like /lib/modules/* ? Sep 24 15:33:09 03koen 07org.oe.dev * rc9078f4b... 10/packages/gpe-today/ (3 files in 2 dirs): packages/gpe-today/: add (disabled) cvs version of gpe-today Sep 24 15:33:13 03jbowler 07org.oe.dev * r328c68f4... 10/packages/perl/perl-5.8.7/config.sh-armeb-linux: perl: add libgcc.a to libperl.so on armeb to support thumb Sep 24 15:33:17 03jbowler 07org.oe.dev * r5a20b56f... 10/packages/sane-backends/sane-backends_1.0.15.bb: Sep 24 15:33:18 sane-backends: add gphoto2 to DEPENDS Sep 24 15:33:18 Without this if gphoto2 doesn't get built the sane-backends build fails Sep 24 15:33:18 because, curiously, the libtool hack in the .bb file is *not* required. Sep 24 15:33:29 03jbowler 07org.oe.dev * rff5c6f58... 10/packages/linux/ (4 files in 2 dirs): nslu2-kernel: initial .bb files for 2.6.13.2 and 2.6.14-rc2 Sep 24 15:33:35 03jbowler 07org.oe.dev * r9ef230b9... 10/packages/meta/ (ucslugc-native.bb ucslugc-packages.bb): ucslugc-packages: add streamripper, sane-backends and perl back in Sep 24 15:33:49 03freyther 07org.oe.dev * r04efeed3... 10/packages/linux/ (2 files): Sep 24 15:33:51 OpenSIMpad: Sep 24 15:33:53 -Switch to gcc 3.3.4 as this can be compiled with uclibc and gcc 4.0 Sep 24 15:33:55 03freyther 07org.oe.dev * r12532c1a... 10/packages/libsdl/libsdl-qpe-1.2.7/pygame.patch: Sep 24 15:33:57 libsdl-qpe: Sep 24 15:33:59 -Patch -lSDLmain away. We do not require it anymore Sep 24 15:34:01 03hrw 07org.oe.dev * rc335b4e3... 10/packages/libxine/ (12 files in 2 dirs): (log message trimmed) Sep 24 15:34:03 added libxine 1.1.0 - someone more familiar with assembly should check it Sep 24 15:34:05 DEFAULT_PREFERENCE = "-1" as it is not tested on device Sep 24 15:34:07 patches status: Sep 24 15:34:09 - cpu-1.0.patch renamed to cpuid.patch Sep 24 15:34:11 - configure-1.0.patch renamed to configure.patch Sep 24 15:34:13 - fix-syntax.patch dropped as fix-syntax-xine-vorbis-decoder.patch do the same Sep 24 15:50:28 mount: Mounting usbdevfs on /proc/bus/usb failed: No such device Sep 24 15:50:30 .... Sep 24 15:50:31 ? Sep 24 15:51:15 OE kernel? Sep 24 15:51:41 hrw|tv: right Sep 24 15:51:46 spitz? Sep 24 15:52:13 right Sep 24 15:52:23 hrw|tv: is the usbcore the only thing I need? Sep 24 15:52:33 AvengerMoJo: no Sep 24 15:52:59 usb-ohci-pxa27x is needed - dont know how module is named Sep 24 15:53:15 hrw|tv: right .. but in 2.6 I didn't see that build :( Sep 24 15:53:46 you use 2.6 now? Sep 24 15:54:24 right Sep 24 15:54:40 usb-ohci then Sep 24 15:55:00 I see the ohci-pxa27x.c file .. but I don't see the usb-ohci driver Sep 24 15:55:36 does it build seperate? Sep 24 15:55:56 man.. ever tried to read 'make menuconfig' help texts? Sep 24 15:56:46 CONFIG_USB_OHCI_HCD=m Sep 24 15:56:51 this is in the .config file Sep 24 15:56:56 but my image doesn't get it :( Sep 24 15:57:22 hrw|tv: is there a way to compile the kernel not using bitbake? Sep 24 15:57:45 sure Sep 24 15:58:32 hrw|tv: :) Sep 24 15:58:40 but I think that you SHOULD look into tmp/deploy/ipk/ Sep 24 15:58:53 and browse in kernel-modules Sep 24 15:59:05 hrw|tv: right Sep 24 15:59:12 and then you will find one thing: Sep 24 15:59:13 hrw|tv: because I see ./drivers/usb/host/ohci-hcd.ko Sep 24 15:59:22 but it is not in the root image? Sep 24 15:59:30 2.6 kernel on Sl-Cxxx is EXPERIMENTAL Sep 24 16:00:03 it is not release quality so it does not have to be good configured Sep 24 16:00:08 hrw|tv: I see ..... so how can I add the othe build install for all those kernel-modules Sep 24 16:00:23 they are build and packaged Sep 24 16:00:32 you just lack them in machine config Sep 24 16:00:57 look which modules are they and add them into conf/machine/zaurus-clamshell-2.6.conf probably Sep 24 16:01:15 and try to learn how OE works Sep 24 16:01:37 and forgive that I'm not polite now - I have to go sleep Sep 24 16:02:19 good night guys ! Sep 24 16:02:28 hrw|tv: no :) thanks alot Sep 24 16:06:48 night Sep 25 03:05:52 moin Sep 25 03:14:50 hail zecke Sep 25 03:41:33 Morning all! Sep 25 03:41:40 Anyone here ? Sep 25 04:15:08 hi Sep 25 04:28:54 do13[offline]: ping Sep 25 04:36:07 do13[offline]: any idea why when applying tosa-machine-base-r9.patch Sep 25 04:36:20 do13[offline]: the last 6 lines do not get applied? Sep 25 04:51:02 ibot: test Sep 25 04:51:05 Test Passed! Sep 25 05:00:52 <_alwin_> hi Sep 25 05:47:52 morning Sep 25 05:49:06 morning mickeyl Sep 25 06:05:58 03mickeyl 07org.oe.dev * ra4aff455... 10/packages/tcltk/ (files/disable-xim.patch tcl_8.4.11.bb tk_8.4.11.bb): add tcl (tool command language) and the tk (toolkit) extension Sep 25 06:06:03 03mickeyl 07org.oe.dev * rd9e0576b... 10/packages/ (3 files in 2 dirs): add python-pyflakes, a source code checker similar to pychecker Sep 25 06:06:08 03mickeyl 07org.oe.dev * r3eb5b8ba... 10/packages/ (3 files in 2 dirs): Sep 25 06:06:08 add python-pyvisa, a package that enables you to control all kinds of measurement equipment through various buses (GPIB, Sep 25 06:06:08 RS232, USB) with Python programs. Sep 25 06:06:15 03mickeyl 07org.oe.dev * r09997038... 10/packages/tcltk/ (tcl_8.4.11.bb tk_8.4.11.bb): tcl/tk: install header files into $STAGING_DIR Sep 25 06:06:21 03jbowler 07org.oe.dev * r6c47e659... 10/packages/linux/ (45 files in 5 dirs): nslu2-kernel, openslug-kernel: 2.6.14-rc2 (ixp425-eth will not build with this) Sep 25 06:06:40 03jbowler 07org.oe.dev * r6934e0a4... 10/packages/ixp425-eth/ (files/2.6.14.patch ixp425-eth_1.1.bb): ixp425_eth: commit fixes for 2.6.14 kernels Sep 25 06:06:43 03dyoung 07org.oe.dev * r67751448... 10/ (2 files in 2 dirs): OpenSlug Packages: Add support for the Happauge PVRUSB2 to openslug builds. Sep 25 06:31:25 03mickeyl 07org.oe.dev * r60a9adaf... 10/packages/python/ (3 files in 2 dirs): Sep 25 06:31:25 python: add patch to fix Tcl/Tk detection. python-core now correctly emits Sep 25 06:31:25 python-tkinter including lib-dynload/_tkinter.so Sep 25 08:25:54 mickeyl: ping Sep 25 08:31:07 pöng Sep 25 08:32:29 I was just going to clear up the endian issues I was asking you about (lest you think I am mad) Sep 25 08:33:57 Although we run in little-endian mode, the FPA soft-float implementation always stores its numbers in big-endian format - this is why R requires ac_cv_bigendian=true as a configure argument, and why I've been pulling my hair out trying to work out what's wrong in Octave Sep 25 08:34:52 interesting Sep 25 08:34:56 Though I've a feeling that the endianness is not split between writing to arrays/disk & handling the fp numbers, so both will probably be broken Sep 25 08:35:20 It says this in include/bits/endian.h Sep 25 08:35:21 03mickeyl 07org.oe.dev * r9227cfee... 10/packages/tcltk/ (4 files in 2 dirs): tcl/tk: add SONAME patches, ship .so in $PN, not $PN-dev Sep 25 08:36:01 i guess we should try to ask the R folks about that Sep 25 08:36:36 I'll get onto it eventually, I've got to get it packaging first (after I finish my PhD too) Sep 25 08:49:52 heh, you should just configure openzaurus to use a sensible byteorder. Sep 25 08:50:58 pb_: The only way to do that would be to switch to the vfp floating point system whose byte order matches the processor's Sep 25 08:51:17 indeed, and that is precisely what you should do Sep 25 08:51:21 Or run the PXA in bigendian mode I suppose Sep 25 08:52:12 I had been meaning to ask why we use FPA - the pdaXrom lot use VFP. Anyone know why (other than FPA is the default for soft-float)? Sep 25 08:55:38 Hmm, where are the toolchain experts? I don't even know the difference. If there is a benefit in changing endianness or floating point processing style, i'm all ears. Sep 25 08:56:05 I think that VFP is faster, and can perform array operations, but my understanding is pretty foggy too Sep 25 08:56:26 the array op thing must be on actual hardware though Sep 25 08:57:21 The easy answer is that there is never any good reason to use FPA in a soft-float system, except for backwards compatibility with legacy code. Sep 25 08:58:03 I just read something which says that vfp won't work on arm4, though if there's no physical fp unit it should still be possible to emulate the vfp handling in code Sep 25 08:58:55 hi :) Sep 25 08:59:53 pb_: I assume that vfp is the actual instruction set used by arm5 physical fp units? and fpa is the equiv for arm4 units? Sep 25 09:03:13 kind of, yeah. It isn't actually anything to do with armv4 or armv5; the floating point instruction set is orthogonal to the processor architecture. Sep 25 09:03:49 there seem to be a lot of different options - presumably because you can combine lots of different fp units? (http://www.redhat.com/docs/manuals/enterprise/RHEL-4-Manual/gnu-assembler/arm-dependent.html) Sep 25 09:04:06 To be exact: softfpa, fpe, fpe2, fpe3, fpa, fpa10, fpa11, arm7500fe, softvfp, softvfp+vfp, vfp, vfp10, vfp10-r0, vfp9, vfpxd, arm1020t, arm1020e, arm1136jf-s and maverick Sep 25 09:04:09 It's true that FPA is an old, mostly obsolete standard, and VFP is the one that is included in modern systems. Sep 25 09:04:51 03ph5 07org.oe.dev * r46f618fc... 10/packages/avahi/ (avahi-0.5/alignment-fix.patch avahi_0.5.bb): avahi: memory alignment fix Sep 25 09:04:55 03ph5 07org.oe.dev * r0dd3f47a... 10/packages/libnss-mdns/ (files/alignment-fix.patch libnss-mdns_0.6.bb): libnss-mdns: memory alignment fix Sep 25 09:07:15 How does one go about voting to change the fp format in OpenZaurus then? Sep 25 09:08:08 I think one types "mickeyl: please change the fp format to " Sep 25 09:08:18 I like your thinking Sep 25 09:08:40 mickeyl: please change the fp format to softvfp, please Sep 25 09:08:44 ;) Sep 25 09:19:41 uhm Sep 25 09:19:45 * mickeyl clueless Sep 25 09:20:00 mickeyl: join the club! Sep 25 09:21:45 even if I would know where to tweak that, I wouldn't dare to do it. a) I think this involves tweaking toolchain packages doesn't it? b) we're near 3.5.4 Sep 25 09:22:36 hi koen Sep 25 09:22:50 I think the tweak may be to remove the 'soft' in the conf file (reading gcc-3.4.0-arm-softfloat.patch) Sep 25 09:23:19 though I agree that this may be a bit of a big change just before release just to satisfy me Sep 25 09:23:28 hmm interesting. my understanding was, that reenables hard float Sep 25 09:23:35 which is heaps slower than soft float Sep 25 09:23:40 at least i know that. Sep 25 09:23:44 heh Sep 25 09:23:56 From a first read, it looks like softvfp should be the default with the arm toolchain Sep 25 09:25:49 if that's true, then familiar has been using soft-float for the last three releases.... which i have a hard time to believe :) Sep 25 09:26:13 Yes, it would appear that passing -soft-float enables softfpa, -hard-float enables a real fp unit and nothing enables softvfp Sep 25 09:26:32 hi ! Sep 25 09:27:03 mickeyl: It'd dependent on the version of the gcc-3.4.0-arm-softfloat patch though Sep 25 09:27:10 s/It'd/It's Sep 25 09:27:11 is vfp the thing an arm9 core has? Sep 25 09:27:48 koen|train: The docs I can see are none to clear about any of it Sep 25 09:28:09 s/to/too - I should stop typing I think Sep 25 09:28:17 doesn't eabi solve all this? Sep 25 09:28:33 I don;t evn know what it is.... off to google Sep 25 09:28:53 should sd cards work with a 2.6.14-rc1-r2 kernel ? Sep 25 09:29:12 alan|home: Should do Sep 25 09:29:17 then i did something wrong with my config... Sep 25 09:29:20 thanks. Sep 25 09:29:34 alan|home: Try ejecting it, then putting it back in & wait for a few seconds Sep 25 09:30:24 lets just postpone this discussion until we a) have released OZ 3.5.4 and Fam 0.8.3 and b) we are a little more educated about the options. Sep 25 09:30:31 lardman: the probleme is not in accessing the files in the sd card, it's when i try suspending... Sep 25 09:31:04 mickeyl: I'm going to set bitbake baking after removing the soft option and I'll see what I get Sep 25 09:31:16 alan|home: What happens? Sep 25 09:31:22 lardman: k Sep 25 09:32:36 lardman: it just won't work.... hey... it does now !!! Sep 25 09:33:22 what the hell did happen ? i didn't do anything... and it wouldn't work yesterday even after 3 reboots... Sep 25 09:33:43 ley lines Sep 25 09:33:51 that, or evil gnomes Sep 25 09:33:52 :D Sep 25 09:34:20 Good evening! Sep 25 09:34:32 koen|train: : ^^ Sep 25 09:35:29 mickeyl: Do I still require the TARGET_FPU var? What happens if it hasn't been found, will there be an error in the parsing, or will is just be assumed to be empty? Sep 25 09:35:50 empty Sep 25 09:36:27 cool Sep 25 09:36:29 mickeyl: did you get my note on the distro.confs and using bberror? Sep 25 09:36:54 I have problems to build gcc-4.0.0 (x86). MACHINE=x86, PREFERRED_VERSION_gcc="4.0.0".bitbake builds gcc-3.4.4-initial-cross ! Any help appreciated. Sep 25 09:36:59 koen|away: i read it. i will implement something like that shortly Sep 25 09:37:25 i won't introduce another variable though. i will just check DISTRO_VERSION Sep 25 09:38:11 I'm reluctant to add any more variables. After our next set of releases I'm going to be on a eliminate and unify variables and methods crusade Sep 25 09:38:55 mickeyl: I'd like to have an easy way to build unversioned distros without touching $distro-$version.conf Sep 25 09:39:06 to check the latest and greatest in OE Sep 25 09:39:09 yeah Sep 25 09:39:31 which was how it was intended Sep 25 09:39:32 how about the easy way is a) removing the check from the unversioned file or b) setting DISTRO_VERSION to a dummy variable. Sep 25 09:39:50 i just don't like that CHECK thing Sep 25 09:40:06 b) is cleaner, but harder to implement I guess Sep 25 09:40:51 that was just an example Sep 25 09:41:10 something like DUMMY := "${bb.data.getVar("DISTRO_VERSION",d,1) or bb.fatal('Remove this line or set DISTRO_VERSION if you really want to build an unversioned distro')}" Sep 25 09:42:08 mickeyl: when do we branch off? After RP has beaten the c7x0 kernel into shape? Sep 25 09:42:44 koen: let's do it on wednesday Sep 25 09:42:56 mickeyl: I also think we need to come up with guidelines on where to keep patches Sep 25 09:43:06 sorry, dinner time _now_ Sep 25 09:43:07 :) Sep 25 09:43:14 keeping patches on random webservers has broken to many times IMO Sep 25 09:43:29 ~bon appetit Sep 25 09:43:31 somebody said bon appetit was smacznego. Guten Appetit. Eet Smakelijk. God Appetitt. Buon Appetito Sep 25 09:51:08 hy guys Sep 25 09:52:11 benoit: you're here? Sep 25 09:52:41 Bernardo: yep! Sep 25 09:53:06 benoit: your oe machine, is it running sourcemage? Sep 25 09:53:25 nope. Sep 25 09:53:38 it's running debian/testing Sep 25 09:53:45 but i'm running out of space ... Sep 25 09:53:55 i'm trying to use evms to make a large partition at the moment Sep 25 09:54:29 ok... I am having problems building on my amd64 with smgl, pcre tries to link with the system's libgcc Sep 25 09:54:39 * lardman needs a computer which can rm -rf tmp faster! Sep 25 09:54:41 I'm wondering if it is smgl related Sep 25 09:54:47 http://pastebin.ca/23865 Sep 25 09:55:01 lardman: that goes fast on a sata disk... Sep 25 09:55:25 too bad it also fails fast Sep 25 09:55:32 Bernardo: Ah, that's mean buying a computer though rather than 'inheriting one' :) Sep 25 09:56:00 yay! finished Sep 25 09:56:13 Bernardo: cannot access your url :-( Sep 25 09:57:05 Right, I'm setting a bitbake gpe-image running, hopefully it'll build with softvfp, we'll see tomorrow (if I'm lucky) when it's finished building Sep 25 09:57:10 bye all Sep 25 09:57:15 lardman: bye Sep 25 09:57:34 benoit: I was there just now Sep 25 09:59:31 what I am seing here is that when building pcre, -L/usr/lib is passed as a argument Sep 25 10:00:55 and then cross-libtool tries to link with /usr/lib/libgcc_s.so and fails Sep 25 10:02:25 I'll try the usual - dispel pcre, bitbakte pcre, and then cast pcre after bitbake finishes... usually works Sep 25 10:03:32 "/usr/lib/libgcc_s.so: file not recognized: File format not recognized/usr/lib/libgcc_s.so: file not recognized: File format not recognized" Sep 25 10:03:42 Bernardo: that is your problem i think. ok you found it :-) Sep 25 10:07:55 benoit: didn't work... and I do have a tmp/cross/arm-linux/lib/libgcc_s.so Sep 25 10:08:21 what is this file ? do a file on it to know. is it 32 or 64 bits? Sep 25 10:08:25 jbs Sep 25 10:08:29 are you running pure 64 bits or multilib ? Sep 25 10:08:38 opps, wrong window Sep 25 10:09:13 I'm running pure 64, and this file should be arm (the one in tmp/cross), the one in /usr/lib should be x86_64 Sep 25 10:09:40 and what are they really? Sep 25 10:10:36 exactly that... well, libgcc_s.so.1 is the file, libgcc_s.so is a link Sep 25 10:10:42 but they are in the correct places Sep 25 10:11:32 the strange thing is that -L/usr/lib is passed to libtool, instead of tmp/cross/amr-linux/lib Sep 25 10:12:55 well, I'll recheck later...have to go Sep 25 10:24:59 Anyone here, who had success building a gcc 4.0.0 for x86 machine ? Sep 25 10:25:09 Anyone here, who had success building a gcc 4.0.0 for x86 machine ? Sep 25 10:25:18 oops ;-) Sep 25 10:25:38 re Sep 25 10:36:36 Ping ? Sep 25 10:38:15 Last question. Will BUILD_ARCH != TARGET_ARCH build a gcc-cross ? Sep 25 10:40:51 uv1: i think so ... except i don't know what is BUILD_ARCH and TARGET_ARCH exactly Sep 25 10:41:24 03mickeyl 07org.oe.dev * rd36b86d3... 10/conf/distro/ (5 files): distro configurations: revert unversioned distro configurations back to standard names and add a warning if they're chosen Sep 25 10:42:39 mickey|dinner: you just broke familiar Sep 25 10:42:57 in how far? Sep 25 10:43:06 ah right Sep 25 10:43:10 don't tell m Sep 25 10:43:10 e Sep 25 10:43:11 familiar-0.8.3.conf: include.... Sep 25 10:43:13 sure Sep 25 10:43:17 * mickeyl on it Sep 25 10:43:25 and probably oz too ;) Sep 25 10:43:29 heh right Sep 25 10:43:30 :D Sep 25 10:44:16 but thanks for looking into the problem and coming up with a solution that doesn't sacrifice functionality Sep 25 10:44:43 even if justinp claims to know the intended behaviour Sep 25 10:44:48 hehe Sep 25 10:46:35 03mickeyl 07org.oe.dev * r01ec2955... 10/conf/distro/ (familiar-0.8.3.conf openzaurus-3.5.4.conf): distro configurations: catch up versioned distro configurations with last change Sep 25 10:47:33 benoit: It seems. Until I have choosen TARGET_ARCH == BUILD_ARCH it tries to build gcc-4.0.0 instead of gcc-3.4.4-cross. I have to find out, why that happens, because I need a gcc-4.0.0-586. Sep 25 10:47:54 Any idea, why gcc-4.0.0 is configure to build in fortran ? Sep 25 10:48:13 because people were complaining we didn't enable it in previous versions Sep 25 10:49:43 mickeyl: So what. Just removed tmp and rebuild gcc. Can you tell me, how to configure for 586 gcc on 686 machine ? Sep 25 10:50:26 sorry, i have no idea about toolchains Sep 25 10:50:40 gb2, kergoth, and pb_ are the toolchain experts iirc Sep 25 10:51:04 mickeyl: Thanks, If I have no success, I will write a mail to the list. Sep 25 10:51:25 yeah Sep 25 10:53:02 psst, mickeyl Sep 25 10:53:14 I spot a small error Sep 25 10:54:22 where? Sep 25 10:54:54 watch the next commit Sep 25 10:55:50 I bet you will feel pretty stupid ;) Sep 25 10:56:24 03koen 07org.oe.dev * rec33cbe7... 10/conf/distro/ (familiar-0.8.3.conf openzaurus-3.5.4.conf): conf/distro: make oz and fam include distro.conf *after* setting DISTRO_VERSION Sep 25 10:56:40 *cough* Sep 25 10:56:44 * mickeyl whistles Sep 25 10:57:19 i guess i better break something else ;) Sep 25 10:57:55 mickeyl: did you see the problems I'm having with pcre? Sep 25 10:58:49 yeah. looks like we should patch the autofoo macros in pcre a bit Sep 25 10:58:58 or Makefile.in for that matter Sep 25 10:59:16 what bot is CIA-4? Sep 25 10:59:30 ~ciabot Sep 25 10:59:42 ~ciabot Sep 25 10:59:43 http://cia.navi.cx/ Sep 25 10:59:45 ah Sep 25 10:59:46 mickeyl: I'll try "hacking it by hand" then Sep 25 10:59:51 I thougt it is a command Sep 25 10:59:54 just to see if I can get past it Sep 25 11:00:02 Bernardo: yep Sep 25 11:00:46 03koen 07org.oe.dev * r5a76331c... 10/conf/distro/opensimpad-0.9.0.conf: conf/distro/opensimpad-0.9.0.conf: include distro.conf after setting DISTRO_VERSION Sep 25 11:02:05 mickeyl: what kernel should I build for the akita? 2.4 or 2.6? Sep 25 11:02:31 Bernardo: 2.4 I'm afraid Sep 25 11:02:54 we may ship 2.6 as an extra for the CXK, but since there are features missing, we can't use it as default yet Sep 25 11:03:18 ok Sep 25 11:09:25 BINDIR = /usr/bin Sep 25 11:09:25 LIBDIR = /usr/lib Sep 25 11:09:25 INCDIR = /usr/include Sep 25 11:09:25 MANDIR = /usr/share/man Sep 25 11:09:40 this seems to be the problem in Makefile... Sep 25 11:13:22 that's ok so far. the problem is that some Makefiles do CFLAGS+=-I$(INCDIR) Sep 25 11:13:31 which is wrong and needs to be patched Sep 25 11:15:06 have a look at a couple of patches in packages/efl Sep 25 11:16:19 ok, thanks Sep 25 11:33:51 mickeyl: ping Sep 25 11:38:34 pöng Sep 25 11:39:52 i love you Sep 25 11:40:07 hmm i forgt the first part Sep 25 11:40:11 oh well Sep 25 11:42:00 mickeyl: bitbake bootstrap-image with a tosa kernel fails Sep 25 11:42:13 mickeyl: it wants to install orinoco and hostap modules Sep 25 11:42:17 mickeyl: but these are not built Sep 25 11:43:24 sounds like someone fiddled with that Sep 25 11:43:28 it worked the last time I touchedit Sep 25 11:43:45 who did the last change? Sep 25 11:44:02 and is this 2.6 or 2.4? Sep 25 11:45:01 from the looks of it, tosa-2.6.conf seems a bit brief Sep 25 11:45:50 yeah orinoco-modules are missing in zaurus-clamshell-2.6.conf Sep 25 11:45:56 or rather, orinoco Sep 25 11:47:09 hey people Sep 25 11:47:22 anyone here going to Linux Kongress? it appears I will be attending it Sep 25 11:48:35 mithro, you will be happy to know ... not me Sep 25 11:48:53 mickeyl: 2.6 ;) Sep 25 11:49:18 but I still have no cross compiler on my mac... Sep 25 11:51:20 anyone here used getaddrinfo() yet? Sep 25 12:34:39 mickey|bbl: do we need to restart monotone? Sep 25 12:35:15 hi Sep 25 12:36:54 hi :) Sep 25 12:41:22 i got a problem compiling libmpeg2-0.4.0b from org.oe.dev pkgs with "undef ref to 'main'" ( http://pastebin.com/373972 ) is this supposed to be a bug in the bb? Sep 25 12:51:03 <_chronic> is monotone ill? Sep 25 12:51:06 <_chronic> monotone: 31.0M | 1.0M | 9450 | 2445 | 2367 Sep 25 12:51:06 <_chronic> monotone: read from fd 5 (peer monotone.vanille.de) failed, disconnecting Sep 25 12:55:01 zecke: how can I build the zaurus kernel not using bitbake? but my hand? Sep 25 12:59:00 :o Sep 25 12:59:33 AvengerMoJo: that saw him off Sep 25 12:59:42 lol Sep 25 12:59:44 :P: Sep 25 13:00:10 AvengerMoJo: what was your question? Sep 25 13:37:47 zecke: how can I build the zaurus kernel not using bitbake? but my hand? Sep 25 13:40:41 in 2.6 is the name of usbdevfs change to usbfs? Sep 25 13:41:32 Yes. Sep 25 13:42:07 Noodles: so I mount usbfs insteand of usbdevfs right Sep 25 13:42:50 Yes. Sep 25 13:43:08 I thought recent 2.4s allowed both in preparation for the move to usbfs, bicbw. Sep 25 13:43:20 ok :) Sep 25 13:43:37 I think hotplug need to update that Sep 25 13:48:55 now my error using 2.6 for my libusb become ioctl(15, USBDEVFS_IOCTL, 0xbedcb32c) = -1 ENOTTY (Inappropriate ioctl for device) Sep 25 13:49:08 instead of Function not find Sep 25 13:49:39 AvengerMoJo: well the kernel is always right ;) Sep 25 13:54:02 zecke: hehe Sep 25 13:54:32 AvengerMoJo: what effect do you expect? Sep 25 13:54:40 zecke: I think it is not the host problem ... devio report this .. and I the connection info under proc is stay the same in my laptop Sep 25 13:54:57 zecke: you should come to linux kongress :P Sep 25 13:55:10 zecke: well basically libusb try to talk to the device I'm connecting to Sep 25 13:55:19 mithro: you are coming? Sep 25 13:55:22 zecke: but some how this ioctl won't go Sep 25 13:57:35 open("/proc/bus/usb/001/003", O_RDWR) = 3 Sep 25 13:57:35 ioctl(3, USBDEVFS_IOCTL, 0xbeb2a758) = -1 ENOTTY (Inappropriate ioctl for device) Sep 25 13:57:40 this is the strace result Sep 25 13:59:42 AvengerMoJo: is that usbproc (or procusb) or plain procfs? Sep 25 14:00:02 zecke: yes Sep 25 14:00:17 zecke: got my plane and ticket booked Sep 25 14:00:26 hey mickey|bbl, you should come too! :P Sep 25 14:00:37 they have information on better flash file systems :P Sep 25 14:02:01 mithro: is dw*mw*Ă coming? Sep 25 14:02:21 zecke: who is that? Sep 25 14:02:55 author of jffs2 Sep 25 14:03:34 hrm dunno Sep 25 14:03:36 Logfs - finally a scalable flash file system by Jörn Engel Sep 25 14:04:16 logfs is a jffs3 split of Sep 25 14:04:16 * mithro is also intrested in the cluster stuff Sep 25 14:05:06 you should come too koen :P Sep 25 14:05:28 VFS based Union Mounts for Linux by Jan Blunck Sep 25 14:06:17 that could be intresting too Sep 25 14:07:16 ..... Sep 25 14:07:27 mithro: hehe, what about sponsoring? Sep 25 14:07:38 zecke: it is so crazy .... I look at my strace in the working version ... Sep 25 14:08:27 zecke: it is exactly the same but ... after the ENOTTY ... it will read the usb bus again in the working one... the zaurus just stop Sep 25 14:08:54 AvengerMoJo: time to consult the source? Sep 25 14:09:34 zecke: i couldn't afford to sponsor anyone Sep 25 14:10:06 mithro: how does that come? Sep 25 14:10:25 zecke: right I looking at it right now Sep 25 14:11:07 not working ---> ioctl(3, USBDEVFS_IOCTL, 0xbeb2a758) = 4 Sep 25 14:11:11 zecke: i could not parse your last sentance Sep 25 14:11:13 working ---> ioctl(3, USBDEVFS_IOCTL, 0xbfc82ff0) = 3 Sep 25 14:11:28 anyone know the = 4 and = 3 ...what is that ? Sep 25 14:11:39 I don't really what the number mean Sep 25 14:11:51 when is it -> http://www.linux-kongress.org/2005/ October 11-14, 2005 at University of Hamburg, Germany Sep 25 14:12:02 AvengerMoJo: it is a FileDescriptor Sep 25 14:12:08 AvengerMoJo: look for 'open' it returns the number Sep 25 14:12:13 zecke: ok thanks Sep 25 14:12:22 AvengerMoJo: 0,1,2 should be the default (stdin,stdout,stderr) Sep 25 14:12:29 zecke: I think he means the return value Sep 25 14:12:34 oh Sep 25 14:12:39 pb_: you are right Sep 25 14:12:44 hey pb_ Sep 25 14:12:47 I'm look at the = 4 and = 3 Sep 25 14:12:49 hi mithro Sep 25 14:12:52 not the 3 in the ioctl Sep 25 14:12:58 what you up to? Sep 25 14:13:20 would you be intrested in coming to/speaking at LCA Embedded Miniconf 2006? Sep 25 14:13:54 nothing much at the moment, just lounging around. Sep 25 14:14:16 yeah, possibly, but I'd have to think about how much it would cost. Sep 25 14:14:22 LCA is in New Zealand next year, right? Sep 25 14:14:48 pb_: yeah Sep 25 14:15:05 LCABAINZ Sep 25 14:15:46 i most probably won't be at the LCA, however I have been roped into organising the miniconf again Sep 25 14:16:23 mithro: how is your swedish nowadays? enough to go shopping at IKEA? Sep 25 14:16:37 zecke: everyone at IKEA speaks english :P Sep 25 14:16:48 cheater Sep 25 14:17:15 is there any chance anyone know where is this up_read function in devio.c come from? Sep 25 14:17:16 i failed Beginner Swedish, so i'm doing it again Sep 25 14:18:06 oh ... Sep 25 14:18:13 pb_: however it's in January instead - which means everything is in a rush Sep 25 14:18:26 mm, expedia says it would cost me 3500 GBP. Sep 25 14:18:32 that does seem rather expensive Sep 25 14:19:07 pb_: that seems a bit expensive - i know that you can get from sydney to Dunedin for like $400 AUS Sep 25 14:19:17 pb_: qantas has a special fare for that period Sep 25 14:19:27 it does require a stop in australia Sep 25 14:19:27 mithro: We've been here before :) Sep 25 14:19:28 03koen 07org.oe.dev * r5762bb54... 10/packages/linux/linux-h6300-omap1-2.6.12-rc5/defconfig: packages/linux/linux-h6300-omap1-2.6.12-rc5/defconfig: apply patch from #343 Sep 25 14:19:36 some people got really really cheap sydney to dunedin for like $200 AUS :P Sep 25 14:19:37 evening all btw Sep 25 14:19:43 hey RP Sep 25 14:19:44 RP: hey Sep 25 14:19:44 koen: ah, I'll have a look Sep 25 14:19:50 come to the Linux Kongress! :P Sep 25 14:19:54 hey RP Sep 25 14:20:03 RP: shoudln't you be camping right now? Sep 25 14:20:11 I don't remember how much it cost to fly to Sydney last time, but I don't think it was that much. Sep 25 14:20:19 koen: No, I've done enough of that... Sep 25 14:20:28 Ice on tents insn't funny :-/ Sep 25 14:20:35 pb_: i seem to rember it been cheaper then I thought it should have been Sep 25 14:20:37 RP: wb Sep 25 14:20:40 anyway got to run Sep 25 14:20:52 you should all come to Linux Kongress and LCA 2006! :P Sep 25 14:21:00 koen: You mentioned me beating the c7x0 kernel into shape? Have there been problems? Sep 25 14:21:01 oh Sep 25 14:21:10 zecke: I think I find the problem Sep 25 14:21:12 RP: not that I know off Sep 25 14:21:22 koen: Ah, that's ok then :) Sep 25 14:21:24 zecke: in my working one is using uhic instead of ohic Sep 25 14:21:28 RP: I just wanted to know an ETA for the branch Sep 25 14:21:28 RP: surely ice on tents goes with the territory in the frozen north Sep 25 14:21:56 pb_: I was in the South (Yorkshire) ;-) Sep 25 14:22:04 pb_: http://www.qantas.com.au/regions/dyn/nl/specials/ActionFares?lk=prnl Sep 25 14:22:12 pb_: I hope that applies for the UK too Sep 25 14:22:19 koen: I'd consider the kernel to be fairly well beaten... Sep 25 14:22:30 RP: :) Sep 25 14:22:35 RP: heh Sep 25 14:23:39 RP: do you know what's the different between the ohic and uhic? Sep 25 14:24:53 AvengerMoJo: You mean OHCI and UHCI? They're two different USB host interface standards Sep 25 14:25:02 I mean OHCI ... :) and UHCI right :) Sep 25 14:25:28 RP: does libusb only able to use UHCI? but not OHCI? Sep 25 14:25:43 AvengerMoJo: I have no idea Sep 25 14:25:47 koen: hm, qantas is offering me 2800 GBP to fly to Sydney in January, or 5400 GBP (!) to fly to Brisbane. Sep 25 14:25:57 I'd expect both to present the same sysfs interface Sep 25 14:26:06 libusb talks to /proc/bus/usb I thought, so should work with both. Sep 25 14:26:10 pb_: ouch Sep 25 14:26:14 RP: thanks Sep 25 14:26:24 Noodles: that's what I think ... Sep 25 14:26:33 Noodles: and the same procfs interface ;-) Sep 25 14:26:37 Noodles: but somehow it is not working the same way it should be Sep 25 14:26:56 oh well Sep 25 14:26:58 heh Sep 25 14:27:20 pb_: You should see the mess when you try and add a flight from Newcastle in ;-) Sep 25 14:27:31 Anyhow, I'm just passing the computer - back tomorrow Sep 25 14:28:32 RP: anyone else using libusb on zaurus before? Sep 25 14:28:37 ops Sep 25 14:28:48 sorry RP hehe didn't mean to put your name there hehe Sep 25 14:47:51 03jbowler 07org.oe.dev * r8980f66e... 10/packages/ixp425-eth/files/2.6.14.patch: ixp425-eth: extra patch for 2.6.14+ kernels with netfilter Sep 25 14:48:39 it is so funny... lsusb work just fine ... Sep 25 14:51:15 03jbowler 07org.oe.dev * r95f5e803... 10/packages/linux/ (6 files in 2 dirs): nslu2-kernel: working 2.6.14-rc2 kernel Sep 25 14:57:58 what is usb.ids? Sep 25 14:58:27 it maps numeric USB IDs to text strings Sep 25 14:59:29 I see ... Sep 25 14:59:51 pb_: do you think that will affect the result of libusb not using the device? Sep 25 15:00:01 no Sep 25 15:01:23 pb_: from what I seen ... it seem my application is not consider the /proc/bus/usb/xxx as the device ... wonder why ... Sep 25 15:02:06 AvengerMoJo: what does mount say? Sep 25 15:03:53 zecke: everything is working now ... usbfs and the usbutils return correct info Sep 25 15:04:09 zecke: except my application ... and ptpcam and libgphoto Sep 25 15:04:20 wondering ... what is missing Sep 25 15:09:02 * AvengerMoJo is confused ... Sep 25 15:09:35 well good luck and good night Sep 25 15:09:42 zecke: thanks :) Sep 25 15:09:46 good night Sep 25 15:39:14 ? I'm running into problems with bitbake opie-image; konqueror won't build Sep 25 15:42:15 here's my log.do_compile file: http://pastebin.ca/23897 Sep 25 15:44:05 libgphoto pack .a file install of .so ... Sep 25 15:49:00 however ... /home/oe/OpenEmbedded/build/oetmp/staging/arm-linux/shlibs/libgphoto2-camlibs.list this should .so file instead... Sep 25 16:13:26 .... Sep 25 16:13:30 it is working now :( Sep 25 16:13:59 I think it is because my library was not compile the way bitbake is compiling before Sep 25 17:16:57 03jbowler 07org.oe.dev * rb412bd6f... 10/packages/openslug-init/ (openslug-init-0.10/conffiles openslug-init_0.10.bb): Sep 25 17:16:57 openslug-init: make /etc/fstab a 'preserve' file Sep 25 17:16:57 This is so that a reflash does not update the file - if it is necessary Sep 25 17:16:57 to patch the file in the future then it will be necessary to write a Sep 25 17:16:57 post-install script for whatever package requires the patch. Sep 25 17:18:44 03jbowler 07org.oe.dev * r642f125f... 10/packages/linux/ (3 files in 2 dirs): nslu2-kernel_2.6.14-rc2: remove nslu2-part.c from the kernel Sep 25 17:23:30 03jbowler 07org.oe.dev * r7f6b7120... 10/conf/distro/ (openslug-packages.conf ucslugc-packages.conf): Sep 25 17:23:30 openslug,ucslugc: python now requires tcl and tk Sep 25 17:23:30 Added packages to the .conf file Sep 25 17:37:28 03jbowler 07org.oe.dev * r607764b0... 10/conf/distro/openslug-packages.conf: openslug-packages.conf: add x11, required by tk Sep 25 17:38:29 huh? x11? Sep 25 17:38:41 jbowler-away: ping Sep 25 17:40:19 NAiL: pong Sep 25 17:40:52 tk is a graphics toolkit, it depends on x11 Sep 25 17:41:31 yeah Sep 25 17:41:42 isn't it possible to avoid building the tk bindings? Sep 25 17:42:06 I feel building x11 just for tk support is a tad overkill ;) Sep 25 17:42:23 Zap mickeyl's DEPENDS addition in python.bb so that on nslu2 it doesn't include either tcl or tk Sep 25 17:43:04 Blargh... it's still broken - now I'm getting PREFERRED_PROVIDER errors. This is dump, I'm going to kill the change on nslu2 Sep 25 18:57:03 03jbowler 07org.oe.dev * r5873b927... 10/packages/python/python_2.4.1.bb: python: remove the tcl and tk depends on MACHINE=nslu2 Sep 25 21:30:12 MOrning Sep 25 21:35:59 hi Sep 25 21:46:29 AvengerMoJo: Hi, can you help me regarding gcc-4.0.0 for x86 ? Sep 25 21:49:15 AvengerMoJo: Sorry no more time ... Have to go to work now ;-(, Bye Sep 25 23:19:51 hi all! Sep 26 00:12:57 03jbowler 07org.oe.dev * r42b3af1b... 10/packages/upslug/ (upslug2-native_5.bb upslug2_5.bb): upslug2: fancy new version with no support for LKG MAC ids (no other change) Sep 26 01:01:56 morning Sep 26 01:11:06 morning Sep 26 01:12:11 morning all Sep 26 01:23:06 morning folks Sep 26 02:02:31 morning Sep 26 02:23:26 RP: morning, how was the trip? Sep 26 02:30:01 XorA: ok, thanks. Ice on the outside of the tent one morning :) Sep 26 02:31:01 RP: wow, hasnt been that cold in this part of Scotland yet Sep 26 02:31:20 we had frost saturday night Sep 26 02:31:35 (cambridgeshire) Sep 26 02:32:05 This was Friday night/Saturday morning... Sep 26 02:32:31 come to Edinburgh, the tropical city :-D Sep 26 02:34:05 Also came across a bike accident and had two of our group collide (at slow speed) :-/ Sep 26 02:38:49 RP: pedal or motor? Sep 26 02:39:00 XorA: Motor Sep 26 02:39:03 RP: eeek Sep 26 02:39:09 hi all, Sep 26 02:41:01 hi Bilbolodz Sep 26 02:42:38 I have a problem with bootloader: Scanning all of flash for your JFFS2 convenience and offset looping...... Any idea how to deal with this? Sep 26 02:50:25 ask bootloader people Sep 26 02:56:56 How is "bootloader people" and where I can find it? Sep 26 02:57:49 simpad use hh.org bootloader? Sep 26 03:05:55 http://lists.gnu.org/archive/html/monotone-devel/2005-09/msg00297.html Sep 26 03:07:57 nice Sep 26 03:08:10 I hope that gtkmm does not need gnome Sep 26 03:09:40 hello Sep 26 03:09:43 it's called gtkmm Sep 26 03:09:51 not gnomemm Sep 26 03:11:03 I noticed that Sep 26 03:11:21 hrw: yes hh.org bootloader Sep 26 03:11:58 Bilbolodz: so maybe ask on #handhelds.org or look in hh.org wiki? Sep 26 03:20:32 djeez Sep 26 03:20:43 how freaking hard is it to get OE isn't a distro Sep 26 03:21:29 does it show I read the oesf forums again? Sep 26 03:21:58 and remind me why we were OK with ZUG using our name? Sep 26 03:23:58 koen: I dont remember does anyone asked us when creating oesf Sep 26 03:24:14 me neither Sep 26 03:24:47 http://www.oesf.org/forums/index.php?showtopic=15143&view=findpost&p=97018 Sep 26 03:24:53 that is what pissed me off Sep 26 03:25:50 argh.. Sep 26 03:29:49 that thread http://www.oesf.org/forums/index.php?showtopic=15167 looks interesting to me - some strange soft (x11) replacements for stuff from sharprom Sep 26 03:30:13 yeah Sep 26 03:30:34 but they haven't done any real searching Sep 26 03:30:54 they haven't found the patched balsa or our patched sylpheed Sep 26 03:33:37 its sharprom users... Sep 26 03:43:49 sudo apt-get remove --purge opera Sep 26 03:57:52 Here's the output of gcc when compiled with TARGET_FPU= "": http://pastebin.ca/23942 Sep 26 03:58:46 It looks like it is probably still using softfpa :( - any other suggestions? (I think I need to re-read the soft-float patch to see how to enable softvfp) Sep 26 04:00:11 does that mean that gcc default to softfpa when nothing is specified, or it will use softfpa when you pass -msoftfloat Sep 26 04:00:14 ? Sep 26 04:18:26 koen: not sure, Sep 26 04:23:13 What does the following do?: %{msoft-float:-mfpu=softfpa} %{!msoft-float:-mfpu=softvfp} Sep 26 04:24:05 I thought it meant that if msoft-float were specified then -mfpu=softfpa would be used and if there's no mention of msoft-float then -mfpu=softvfp will be used? Sep 26 04:26:30 hrw: there is a dead silence on #handheld.org Sep 26 04:27:39 Bilbolodz: happens Sep 26 04:48:56 morning Sep 26 04:48:59 hey Sep 26 04:49:41 hey reenoo_ & zecke Sep 26 04:50:02 hi zecke Sep 26 04:50:04 yo koen Sep 26 04:50:32 Actually I may have been wrong about my new toolchain: %{msoft-float:-mfpu=softfpa} - it might work if I don;t specify -msoft-float... Sep 26 04:51:10 Anyone know how I can tell if a binary has been compiled with soft-float? Sep 26 04:51:14 hi zecke, reenoo Sep 26 04:51:39 man KARO is expensive... Sep 26 04:52:00 lardman: not offhand, objdump is a good candidate Sep 26 04:52:22 hey hrw|work Sep 26 04:52:50 zecke: Thanks, I'll try that Sep 26 04:54:08 lardman: I have no clue, but I would look at the div symbol Sep 26 04:54:45 someone tested libxine 1.1.0 Sep 26 04:54:47 ? Sep 26 04:55:21 not yet (OE on OS X is not yet working) Sep 26 04:58:49 Hmm, objdump gives the following: private flags = 2: [APCS-32] [FPA float format] [has entry point], not what I wanted to see Sep 26 05:00:28 and a grep shows no signs of it having been compiled with -soft-float which would be the convenient explanation Sep 26 05:04:34 Oh, I wonder if the objdump line above actually means I've compiled for hard-float... Sep 26 05:05:44 yeah it does, drat Sep 26 05:05:57 * lardman goes for lunch to think about patches... Sep 26 05:18:56 hi offroadgeek Sep 26 05:31:15 so you all coming then? Sep 26 05:31:29 where? Sep 26 05:33:19 LCA or Kongress? Sep 26 05:34:07 sure Sep 26 05:39:39 both Sep 26 05:48:32 Kongress would be cool, LCA would be cooler Sep 26 05:52:02 Hi Sep 26 05:52:10 do13: hey Sep 26 05:52:31 do13: any idea why the last 6 lines of the tosa-base-machine patch are not getting applied? Sep 26 05:52:39 hi zecke_ Sep 26 05:53:30 zecke_: tosa-machine-base-r9.patch is old. which kernel did you try to compile? Sep 26 05:54:44 do13: stock linux-openzaurus from OE Sep 26 05:54:53 zecke_: Try linux-openzaurus-2.6.14-rc1 Sep 26 05:55:06 do13: I did Sep 26 05:55:19 do13: it tried to applied version r7 of the patch (and succeeded) Sep 26 05:55:29 do13: but the last four lines of tosa.c were missing Sep 26 06:00:27 zecke_: Just a moment: I will try a newer one. Sep 26 06:02:12 do13: and the missing lines were in the patch, but they did not get applied Sep 26 06:02:19 I have never seen such thing before Sep 26 06:10:09 zecke_: I put my working version online. please download Sep 26 06:15:20 do13: I will do it tonight (once I'm home) Sep 26 06:17:52 zecke_: To build an opie-image you need this patch http://www.do13.de/oz/temp/defconfig-tosa.patch Sep 26 06:23:46 do13: I'm going to push that tonight as well Sep 26 06:25:47 OT: wow I can do up to four OE builds in parallel Sep 26 06:33:37 zecke_: on osx? Sep 26 06:33:52 koen: on my P4 here at ROAD ;) Sep 26 06:49:31 koen: I will sanitize Disconnects MacOS X patch Sep 26 06:53:59 morning Sep 26 06:54:08 morning Sep 26 07:00:15 hi guys Sep 26 07:24:53 zecke_, do13: That kind of patch error occurs when the number of lines listed as being in a patch is incorrect Sep 26 07:27:55 ~users Sep 26 07:27:57 so this is what we get for sharing our unpaid volunteer work with you? complaints and accuses? well done, this clearly supports our motivation to continue working on open source projects. Sep 26 07:28:06 vanille.de is up again Sep 26 07:28:45 still no word from CELF Sep 26 07:29:22 03koen 07org.oe.dev * r8456894b... 10/packages/gpe-mini-browser/gpe-mini-browser_cvs.bb: packages/gpe-mini-browser/gpe-mini-browser_cvs.bb: recommend some pixbuf loaders Sep 26 07:29:37 libxine do_configure() part takes time... Sep 26 07:30:03 * koen starts on doing his taxes Sep 26 07:31:47 I do taxes once per year Sep 26 07:34:16 me too, but healthcare and rentsupport are getting reformed Sep 26 07:37:49 we choose(d) new parlament yesterday - who knows what will change... Sep 26 07:38:00 healthcare for sure - it always change Sep 26 07:39:10 hrw|work: I saw the right wing won? Sep 26 07:39:16 yep Sep 26 07:39:23 you'll probably lose healthcare Sep 26 07:39:32 here they trashed it when they were in power Sep 26 07:39:49 Bernardo: here they cannot - too much pressure from people Sep 26 07:40:09 but public healthcare is ;( Sep 26 07:40:30 well, here they "outsourced" hospital management Sep 26 07:41:07 reduced sick leave (50% salary for short periods, going to a max 70 for 1 year or more), etc. Sep 26 07:50:47 03mickeyl 07org.oe.dev * rf081336d... 10/packages/resistorui/resistorui_0.9.bb: resistorui: fix DESCRIPTION and drop MAINTAINERship Sep 26 07:52:04 mickey|bbl: ping Sep 26 07:52:16 pöng Sep 26 07:52:48 mickeyl: tried libxine 1.1.0? Sep 26 07:53:06 hrw: i'm building it atm, Sep 26 07:53:22 ok Sep 26 07:53:27 hey mickeyl Sep 26 07:53:30 hi koen Sep 26 07:53:31 hi all Sep 26 07:53:53 'online taxes' Sep 26 07:54:04 they actually mean 'download windows 9x binary' Sep 26 08:00:01 mickeyl: didn't have the time to "fix" pcre Sep 26 08:03:52 mickeyl: if you build libxine 1.0 before then you need to remove shlibs file - libxine contra libxine1 names Sep 26 08:09:55 koen: what kind of response from CELF are you waiting for? Sep 26 08:10:30 "Lovely, koen here is the money to make you flithy rich" Sep 26 08:10:45 'We are going to hire someone to implement feature XYZ from your list" Sep 26 08:10:52 zecke_: that would be ok too :) Sep 26 08:11:07 hrw|work: oops. right. Sep 26 08:11:08 I told you when the message first came out, don't hold your breath Sep 26 08:11:36 CosmicPenguin: well, we don't hold it Sep 26 08:11:48 ~lart berlios Sep 26 08:11:49 * ibot installs a bad bootloader on berlios and turns berlios into a brick Sep 26 08:13:20 zecke_: if I post a patch against bitbake 1.3.2, is that new enough for your to apply to the trunk? Sep 26 08:13:37 CosmicPenguin: likely Sep 26 08:14:18 I can't get svn to work right now Sep 26 08:14:24 that goes along with everything else Sep 26 08:14:31 what a crappy day already Sep 26 08:14:50 hmm.. mplayer-atty -vo w100 looks like broken Sep 26 08:16:29 CosmicPenguin: what did you patch? Sep 26 08:16:45 CosmicPenguin: if I don't apply it within days, just beat me to Sep 26 08:18:25 zecke_: I figured out my SVN Sep 26 08:18:33 I'm going to post it anway, because it breaks core functionality Sep 26 08:18:50 CosmicPenguin: whell why do you do it, if it breaks things? Sep 26 08:18:59 well even... Sep 26 08:19:21 break is a strong word Sep 26 08:19:35 it adds http proxy support for CVS, and adds a perforce fetcher Sep 26 08:19:47 so it doesn't break anything, but its a major change to a sensitive area... :) Sep 26 08:19:56 and my python is only slightly better then my german Sep 26 08:21:30 * CosmicPenguin starts German lessons on Wednesday - yay! Sep 26 08:23:12 Viel Erfolg Sep 26 08:23:44 zecke_: Any news on getting bitbake to export the dependency information? Sep 26 08:24:16 one more unfinished thing... Sep 26 08:24:22 I have not hacked on it the last week Sep 26 08:30:03 03koen 07org.oe.dev * r7edab77a... 10/packages/gpe-themes/gpe-theme-industrial_0.1.bb: packages/gpe-themes/gpe-theme-industrial_0.1.bb: add tweaked industrial theme to OE Sep 26 08:33:57 I'm fast coming to the conclusion that I don;'t know how to enable vfp Sep 26 08:38:43 ~seen justinp Sep 26 08:38:51 justinp is currently on #openzaurus (21h 8m 43s) #oe (21h 8m 43s) Sep 26 08:40:06 i see there is a e-image meta file. Does it mean i could have an openzaurus running enlightenment just by running "bitbake e-image" ?? Sep 26 08:41:30 yes, I think so, but it's rather large - so best bake e-image-core iirc Sep 26 08:47:18 Anyone know what the state of open source clash is? is there an open source player? Sep 26 08:47:22 err Sep 26 08:47:24 Flash* Sep 26 08:48:00 mickeyl: we still lack poodle developer? Sep 26 08:48:16 hrw|work: yes Sep 26 08:49:01 lardman|work: mmm... i guess that means the answer is the same for xfce-image... i wonder what it will look like... let's try that ! (^_^) Sep 26 08:49:04 I could dig for right-arrow fix for poodle (someone wrote it) but it wont be tested Sep 26 08:49:14 the xfce-image needs a lot of love Sep 26 08:49:24 love ? Sep 26 08:49:24 as do the xfce* packages Sep 26 08:49:26 care Sep 26 08:49:27 tender Sep 26 08:49:29 work Sep 26 08:49:39 hehe Sep 26 08:50:07 what about the e one ? Sep 26 08:50:17 that's ready for consumption Sep 26 08:50:24 anyone ever tested it here ? Sep 26 08:50:51 yea Sep 26 08:50:58 oh dear, I forgot to re-enable TARGET_FPU, that could be why it doesn't work - damn! Sep 26 08:51:15 lardman|work: heh Sep 26 08:51:56 mickeyl: http://www.oesf.org/forums/index.php?act=Attach&type=post&id=894 is a patch Sep 26 08:52:01 the more i try to understand how bitbake works, the more i discover how powerfull this tool is... and sometimes it's scarry. ^^" Sep 26 08:52:01 brb Sep 26 08:52:08 mickeyl: I've got my fingers crossed that it will work, but looking here (where the softfloat patch came from): http://www.andric.com/cross/patches/, there are lots of vfp specific ones, so it might be more complex I fear Sep 26 08:52:37 hrw|work: looks ok so far Sep 26 08:53:32 lardman|work: yeah, it's a bit unfortunate that our toolchain experts are on hiatus :D Sep 26 08:54:11 mickeyl: Who are they? Sep 26 08:54:35 gb2, kergoth, pb_ Sep 26 08:54:42 ok, cool Sep 26 08:55:03 jbowler looks like he's an expert as well, we may pester him about these things when we have a chance Sep 26 09:01:50 mickeyl: I don't suppose you know where/whether gcc-3.4.0-arm-softfloat.patch is even applied? I just grep'ed in packages/gcc and it doesn't appear (so my changes aren't going to do much!) Sep 26 09:02:38 doesn't look like it gets applied Sep 26 09:02:39 looks like nowhere Sep 26 09:02:49 at least not to 3.4.x Sep 26 09:02:59 odd, I wonder how we get any softfloat in that case Sep 26 09:02:59 the only occurences are 3.3.3 and 3.3.4 Sep 26 09:03:10 perhaps it hit mainline? Sep 26 09:03:51 NOTE: package openzaurus-pxa-2.4.18-rmk7-pxa3-embedix-r43: task do_patch: completed Sep 26 09:03:53 uf. Sep 26 09:04:00 nice Sep 26 09:05:32 http://www.dilbert.com/comics/dilbert/archive/images/dilbert2005092116285.jpg Sep 26 09:06:45 hehe Sep 26 09:07:05 looking at sharp kernel compile log message is.... Sep 26 09:07:41 ../usbd-arch.h:198: warning: #warning SETTING DEFAULTS FOR PXA Sep 26 09:07:42 ../usbd-arch.h:387: warning: #warning CONFIGURING FOR POODLE Sep 26 09:07:53 or my favorite: /home/hrw/devel/build/0915/tmp/work/poodle-linux/openzaurus-pxa-2.4.18-rmk7-pxa3-embedix-r43/linux/include/asm/proc/cache.h:118: warning: passing arg 1 of `cpu_pid_va_to_mva' makes integer from pointer without a cast Sep 26 09:07:58 ;) Sep 26 09:08:10 anyway it builds Sep 26 09:08:18 NOTE: package openzaurus-pxa-2.4.18-rmk7-pxa3-embedix-r43: task do_build: completed Sep 26 09:08:25 tomorrow I'll push it as r44 Sep 26 09:08:26 cu Sep 26 09:08:32 time to go home Sep 26 09:08:39 cu Sep 26 09:08:58 bye Sep 26 09:09:05 mickeyl: is the cx000 kernel still built with 2.95.3, or can it be built with a later toolchain (3.x for example)? Sep 26 09:09:08 bye hrw|gone Sep 26 09:10:23 2.4.20-crapix Sep 26 09:10:24 ups Sep 26 09:10:32 2.4.20-embedix can't be built with 3.x Sep 26 09:10:33 heheh Sep 26 09:10:52 looking at gcc-3.4.4 before it's patched, it appears that softfpa is the default, and that patch hasn't been applied directly Sep 26 09:13:24 any idea what toolchain tetsu uses, as he enables the wmmx (or whatever the name is) extensions? Sep 26 09:14:27 no idea. but 2.4.20-dontworkix is history soon anyway thanks to Richard... Sep 26 09:14:47 random patch which looks like it enables vfp for 3.4.4: http://gentoo.kems.net/gentoo-portage/sys-devel/gcc/files/3.4.4/gcc-3.4.4-softfloat.patch Sep 26 09:15:08 lardman|work: might be worth a try for us Sep 26 09:15:19 mickeyl: I do hope so... Sep 26 09:20:04 mickeyl: just added it to the .bb, fingers crossed, etc. Sep 26 09:24:53 mickeyl: Hello, I am having trouble to get pygtk working, could you please give me a hint which .bb to bake ? pygtk2_2.4.1.bb, pygtk-0.6.12.bb, pygtk_2.2.0.bb, pygtk2_2.6.2.bb ? getting qt to work is a lot smoother ;-) cheers Sep 26 09:27:57 i'd try the latest and greatest Sep 26 09:28:06 2.6.2 Sep 26 09:28:13 _if_ you want to use gtk2 Sep 26 09:28:16 otherwiese 0.6.12 Sep 26 09:29:24 I'm off, c u all tomorrow Sep 26 09:29:30 mickeyl: thank you, the pygtk ipk which is on you server seems to be broken Sep 26 09:31:21 in how far...? Sep 26 09:32:30 mickeyl: when I try to install it I get the error: ' ipkg: invalid magic ' Sep 26 09:32:38 hmm Sep 26 09:33:00 which file is this exactly? Sep 26 09:33:22 mickeyl: http://www.vanille.de/projects/packageinfo.spy/python-pygtk_2.2.0-r3_arm.ipk Sep 26 09:33:33 uhm Sep 26 09:33:40 this is not a link to an ipk Sep 26 09:33:46 this is a link to an information page Sep 26 09:34:03 http://vanille.de/feed/python/modern/python-pygtk2_2.6.2-r2_arm.ipk Sep 26 09:34:05 would be a valid link Sep 26 09:34:56 hoops --- shame on me, I just saw the ending ipk and thought it would be a valid ipackage Sep 26 09:35:02 heh, nm Sep 26 09:35:44 but I just tried to bake python-pygtk2_2.6.2.bb and I fails with the ERROR: function do_configure failed Sep 26 09:36:00 I will update my sources from the server and try again Sep 26 09:36:05 yeah Sep 26 09:36:16 if it doesn't build, add an entry to the bt Sep 26 09:36:35 will do - thanks for your help Sep 26 09:40:30 np Sep 26 09:49:50 mickeyl: did you need me? Sep 26 09:49:57 mickeyl: been very busy lately... Sep 26 09:51:47 JustinP: hi. no hurry, just wanted to check if you have any plans to upgrade the E stuff yet. Sep 26 09:54:24 03koen 07org.oe.dev * r85d9528f... 10/packages/gpe-today/gpe-today_cvs.bb: packages/gpe-today/gpe-today_cvs.bb: fix PV Sep 26 09:59:05 03koen 07org.oe.dev * r0a9631b6... 10/packages/gpe-today/gpe-today_0.10.bb: packages/gpe-today: add 0.10 Sep 26 09:59:09 03koen 07org.oe.dev * r58382d22... 10/conf/distro/preferred-gpe-versions-2.7.inc: conf/distro/preferred-gpe-versions-2.7.inc: prefer gpe-today 0.10 Sep 26 10:01:00 mickeyl: upgrade to? Sep 26 10:01:14 mickeyl: e doesn't work for me right now Sep 26 10:01:24 mickeyl: and updatign everything to CVS caused other problems Sep 26 10:03:40 mickeyl: I see that there may be some new "releases" of things...I'll try those... Sep 26 10:04:09 yeah, these "releases" are what I meant Sep 26 10:04:14 i added a bug entry to remind you Sep 26 10:04:18 (or us) Sep 26 10:04:49 i'll test a build of E asap Sep 26 10:05:50 mickeyl: did you manage to get the spinning crosshairs working? Sep 26 10:06:20 yes, but they didn't look good, then i trashed them Sep 26 10:06:54 I see Sep 26 10:07:11 mickeyl: thanks, I'd appreciate it. I have had to switch back to GPE....I decided to switch to an e-image, which had worked fine previously just fine but which decided to break when I tried it... Sep 26 10:07:52 koen: calibration is done. next thing to do is to make it look for all possible boot devices and then hand control back over to a part of altboot. Sep 26 10:10:15 oh yeah, and gui wise i need to implement a nice selection widget Sep 26 10:10:27 i want something more fancy than a listbox :) Sep 26 10:10:51 ok, i'm gone for a beer Sep 26 10:10:52 bbl Sep 26 10:10:53 mickeyl: how big would a statically linked bootmanager be? Sep 26 10:11:06 you mean against _all_ libs? Sep 26 10:11:29 only E at the start Sep 26 10:11:36 maybe uclibc(++) in the end Sep 26 10:11:53 to make it more robust Sep 26 10:12:23 all runtime dependencies of angstrom-bootmanager are atm. less than 3MB. that includes everything EFL-fb, freetype, libpng, libjpeg Sep 26 10:12:26 03mickeyl 07org.oe.dev * r05dd4ce9... 10/packages/tcltk/tk_8.4.11.bb: tk: DEPEND on libx11, not x11 Sep 26 10:12:40 bbl Sep 26 10:40:00 Good evening! Sep 26 10:40:19 Is anyone here for toolchain problems ? Sep 26 10:40:22 are people unhappy with berlios? is the svn server down too much? Sep 26 10:40:43 it is Sep 26 10:40:58 being down often isn't the real problem Sep 26 10:41:03 whats the issue then? Sep 26 10:41:05 * CosmicPenguin just needs to be smarter then the fucking software Sep 26 10:41:19 being down for prolonged periods of time (a week) is the problem Sep 26 10:41:24 gotcha Sep 26 10:41:42 uv1: just ask Sep 26 10:41:59 well i just set up a project on http://opensvn.csie.org/, which seems like a nice, simple svn service provider only. no SF-like crap, just svn. andt hey provide 'trac' for the projects Sep 26 10:42:12 so i'm gonna commit the bitbake repo history there and see how it behaves Sep 26 10:42:25 kergoth: if you can not resist ;) Sep 26 10:42:37 Have problems to build x86 gcc-4.0.0. Package build gives up due to unmatched gmp fortran depends. Sep 26 10:42:52 trac is pretty sweet Sep 26 10:43:18 indeed Sep 26 10:43:19 uv1: well it is not smart to use gcc4.0.0 anyway Sep 26 10:43:26 uv1: fwiw, gcc-4.0.0 is broken anyway Sep 26 10:43:34 Also I get a gcc-cross, when BUILD_ARCH != TARGET_ARCH (e.g. 586 != 686). Is that normal ? Sep 26 10:43:55 So which gcc is recommended ? Sep 26 10:44:41 zecke: Why not use ? Problems with code ? Sep 26 10:45:18 uv1 is broken Sep 26 10:45:30 uv1: gcc4.0.0 is broken Sep 26 10:45:54 zecke: You mean the bb recipe or gcc itself ? Sep 26 10:46:02 gcc itself Sep 26 10:46:08 it's being said that 4.02 is pretty sweet Sep 26 10:46:23 uv1: use 3.4.4 unless you want to fix both gcc and all the other source code which will need to be fixed Sep 26 10:46:39 JustinP: ;-) Sep 26 10:46:56 Thanks for info. Will go on with gcc-3.4.4 Sep 26 10:47:28 hi Sep 26 10:48:49 how to start the tell the standalone application ( not with opie ) to start as landscape or portrait Sep 26 10:49:00 how to start the standalone application ( not with opie ) to start as landscape or portrait Sep 26 10:49:27 AvengerMoJo: look into the Qt documentation Sep 26 10:49:42 zecke thanks Sep 26 10:49:43 AvengerMoJo: QWS_DISPLAY=Transformed:RotXXX Sep 26 10:49:57 AvengerMoJo: Rot0, Rot90, Rot180, Rot270 Sep 26 10:50:16 oh thanks Sep 26 10:54:47 .... Sep 26 10:55:14 zecke not working :) ... I'm starting my application like this ... application -qws -display :3 Sep 26 10:56:00 AvengerMoJo, zecke: FYI, this is neither #opie nor #qpe or whatever Sep 26 10:56:14 reenoo_: thanks Sep 26 10:58:02 reenoo_: relax Sep 26 10:58:04 pb_: hey Sep 26 10:58:13 pb_: karo wants 1600 Euro for a starter kit... Sep 26 10:58:25 pb_: is that a usual price? Sep 26 10:59:32 hail zecke Sep 26 10:59:47 that probably is about normal for a starter kit, yeah Sep 26 10:59:51 those things tend to be expensive Sep 26 11:00:32 <_alwin_> hi Sep 26 11:01:40 pb_: where could I else put a SODIMM module in? Sep 26 11:02:06 zecke: I guess you could construct your own baseboard with a DIMM socket Sep 26 11:02:30 * zecke tries to talk his Uni into buying one... but lacks a reason Sep 26 11:02:38 starter kit? Sep 26 11:02:58 mithro: I want some current ARM board where I can attach some sort of disk to (CF, IDE) Sep 26 11:03:28 zecke: http://www.giantshoulderinc.com/hardware.html ? Sep 26 11:03:41 (that's armeb) Sep 26 11:04:23 koen: price? Sep 26 11:04:35 $239 ish Sep 26 11:04:41 ask [g2] :) Sep 26 11:05:30 zecke: ahh Sep 26 11:07:18 <[g2]> koen currently $249 with board, ps and case Sep 26 11:07:42 <[g2]> however, I'm thinking about going 533Mhz which would be a $20-$25 bump Sep 26 11:08:06 <[g2]> the protos running @ 533 just seem so much zippier Sep 26 11:08:44 <[g2]> zecke I've booted to the unofficial armbe debian port on the board with the 2.4 kernel Sep 26 11:08:45 * XorrA|gone hunts for a spare $300 Sep 26 11:08:57 <[g2]> straight to the microdrive and CF Sep 26 11:09:34 128MB would be another $30? Sep 26 11:09:49 <[g2]> XorrA|gone it's got 2 ethers, USB 2.0 and POE lots more Sep 26 11:10:09 <[g2]> koen I'm checking the pricing on that too Sep 26 11:10:19 [g2]: yes, I looked, thats a sweet bit of kit, would be nice toy to play with Sep 26 11:10:36 [g2]: maybe even replace one of my PC's with Sep 26 11:10:37 <[g2]> it's OH some much fun to play with Sep 26 11:11:17 <[g2]> it's such a sane environment, serial, bootloader with network support, JTAG there if you need it Sep 26 11:11:38 <[g2]> I've just about got all the patches for 2.6.11.12 ready to push upstream Sep 26 11:12:29 <[g2]> the two big item are the CF/IDE driver in 2.6 which is installed by being debugged right now (literally) Sep 26 11:12:50 <[g2]> and the USB 2.0 chips gets interrupts by is not fully happy yet Sep 26 11:12:57 <[g2]> s/by/but/ Sep 26 11:13:25 <[g2]> The POE is quite fun Sep 26 11:13:38 whats a miniPCI? Sep 26 11:13:53 <[g2]> it's a small form factor PCI connector Sep 26 11:14:01 mainly used in laptops Sep 26 11:14:06 <[g2]> and embedded Sep 26 11:14:10 [g2]: ah, so cant use normal pci cards Sep 26 11:14:35 <[g2]> XorrA|gone, the box is only 6.3x4.2x1.5 inches Sep 26 11:14:48 <[g2]> the board is just 4x6 Sep 26 11:15:11 CosmicPenguin: ping Sep 26 11:15:18 [g2]: had visions of strapping a Matrox g450 on top and having a small tv streamer Sep 26 11:15:19 CosmicPenguin: there is no need to use os.environ Sep 26 11:15:25 CosmicPenguin: use bb.data.getVar Sep 26 11:15:29 <[g2]> I would have liked to have a PCI edge connector on the end like the soekris, but it wasn't in the cards Sep 26 11:15:50 CosmicPenguin: also + with string operations is slow Sep 26 11:16:04 CosmicPenguin: use" %s %s" % (px0, px1) Sep 26 11:17:01 <[g2]> XorrA|gone, did you see http://www.yuan.com.tw/en/products/vdo_mpc622.html in the links section ? Sep 26 11:17:21 <[g2]> XorrA|gone, what's wrong with an external USB 2.0 driver ? Sep 26 11:17:33 <[g2]> drive Sep 26 11:18:57 [g2]: got one which does DvB? Sep 26 11:19:22 mithro: DVB and USB2? Sep 26 11:19:34 no MiniPCI Sep 26 11:19:35 <[g2]> the real benefits to this box is 1) non-x86, 2) low-power ~3-4W typical, 3) fully open Sep 26 11:20:04 [g2]: got an url again, and order form? Sep 26 11:20:08 any moving parts? Sep 26 11:20:56 [g2]: that card is tv-in only, I wanted ok quality tv-out, hence the G450 Sep 26 11:20:58 <[g2]> zecke I'm getting ready for the first production build after verifying the USB 2.0 Sep 26 11:21:13 [g2]: But Im dreaming as the Mrs never allow expendeture :-) Sep 26 11:21:18 [g2]: ah, you are building them? Sep 26 11:21:41 <[g2]> I've contracted an OEM that builds boards like this Sep 26 11:22:04 pb_: thanks for the fast answer Sep 26 11:22:19 pb_: I had assumed it would resolve all SO_NEEDED (on OS X) Sep 26 11:22:41 pb_: but it does not, or our Opie build system is broken (but that was done by kergoth) Sep 26 11:23:21 ah, that is unfortunate Sep 26 11:23:33 what was the question? Sep 26 11:24:54 whether OS X handles DT_NEEDED at link time in the same way that ELF systems do Sep 26 11:25:29 though, thinking about it some more, I don't think it really matters in this context. Sep 26 11:25:41 ahh Sep 26 11:26:01 if the Opie Games are making direct calls into the opie libraries, they should be linking with -lopie-opie themselves, not relying on SDL to pull it in (either directly or via sdl-config) Sep 26 11:26:09 * kergoth nods Sep 26 11:28:27 pb_: it is more a metter of /opt/Qtopia/apps/Games/foogame.desktop Sep 26 11:28:48 zecke: I don't understand Sep 26 11:29:02 * [g2] hugs kergoth Sep 26 11:29:04 can you explain the problem in small words, as if speaking to someone very stupid? Sep 26 11:29:14 heh, hey [g2] Sep 26 11:29:23 <[g2]> hey hey kergoth ! Sep 26 11:29:28 how ya been? Sep 26 11:29:35 <[g2]> awesome thx Sep 26 11:29:56 <[g2]> I've got proto boards and I'm just about to turn on the production build Sep 26 11:30:02 <[g2]> 200 units Sep 26 11:30:22 nice Sep 26 11:30:39 <[g2]> then I'll see whether this boat float or sinks Sep 26 11:30:54 hehe Sep 26 11:31:05 <[g2]> if it sinks there will be a fire sale in about 6-9 months :) Sep 26 11:31:16 * kergoth is having a fair bit of fun learning about networking and advanced routing in linux Sep 26 11:31:24 * kergoth hugs netfilter Sep 26 11:31:30 pb_: well, currently most of the SDL games, install resources useful to Opie Sep 26 11:31:40 <[g2]> nod Sep 26 11:31:41 pb_: including an icon for our launcher Sep 26 11:31:46 right Sep 26 11:31:58 pb_: what I do not want to see is somebody going over the games and simply remove these resources Sep 26 11:32:05 oh, right, I see Sep 26 11:32:08 yes, I agree with that. Sep 26 11:32:47 there's no reason to remove them, anyway. contrary to popular belief, GPE does not actually explode if it detects files in /opt/QtPalmtop: it will just quietly ignore them and hope that they go away on their own. Sep 26 11:33:04 wow GPE is wonderful Sep 26 11:33:06 hehe Sep 26 11:33:13 yeah, it is a paragon of tolerance Sep 26 11:33:21 zecke: he he Sep 26 11:34:01 koen: did you look at the Qtopia start script? Sep 26 11:34:15 koen: they do rm -rf /etc/rc*.d/S*X* on start up Sep 26 11:34:33 cya later Sep 26 11:34:41 <[g2]> bye Sep 26 11:34:51 zecke: I see Sep 26 11:38:17 * pb_ stabs samba Sep 26 11:58:15 so, does anybody understand how smb browsing is meant to work? Sep 26 11:58:34 pb_: badly Sep 26 11:58:53 * koen was struggling with that this afternoon Sep 26 11:58:57 samba won Sep 26 11:59:54 hehe samba can not be as evil as BSD sed (which just beat me) Sep 26 12:00:37 heh Sep 26 12:00:39 probably not Sep 26 12:00:44 * pb_ hands zecke a copy of gnu sed Sep 26 12:01:50 pb_: hehe, fink was faster Sep 26 12:05:16 man, i like these bluetooth headphones Sep 26 12:05:33 the cell phone being able to ring through the mp3s i'm listening to with my iriver is just sweet Sep 26 12:05:46 kergoth: what dongle did you get? Sep 26 12:05:58 I got a V3 over the weekend, so I'm in the market Sep 26 12:06:20 CosmicPenguin: got my two comments? Sep 26 12:06:36 hi do we get bootsplash for 2.6 kernel? Sep 26 12:06:42 CosmicPenguin: V3? Sep 26 12:06:49 moto razr? Sep 26 12:06:54 zecke: thanks Sep 26 12:07:06 zecke: I thought we used bb.data.getVar pretty much everywhere - where did I miss? Sep 26 12:07:22 CosmicPenguin: i picked up http://www.expansys-usa.com/product.asp?code=114076 for listening to mp3s and the cell interrupting. dont have a seperate bt headset deal for just using the phone yet Sep 26 12:07:26 CosmicPenguin: os.environ for HTTP_PROXY Sep 26 12:07:34 ahhh.... Sep 26 12:07:39 sure Sep 26 12:07:47 CosmicPenguin: but bitbake copies all environment variables into its scope anyway Sep 26 12:08:06 CosmicPenguin: this would even allow you to use seperate proxies for seperate files Sep 26 12:08:24 * CosmicPenguin makes the changes Sep 26 12:10:39 ah, now samba works Sep 26 12:10:47 seems that nmbd just needed to chill out for a while Sep 26 12:11:03 can I patch the 2.6 kernel with the normal bootsplash or I need to use something else? Sep 26 12:17:53 kergoth: wow, those are abit spensive Sep 26 12:18:17 CosmicPenguin: do you like the v3? Sep 26 12:18:29 koen: yeah, so far Sep 26 12:18:43 I don't like that I'm going to need to use Windoze to load mp3s and stuff, but.... Sep 26 12:18:58 i was thinking about a v3, v3x or nokia n90 Sep 26 12:19:08 I need to learn more about how the whole bluetooth/usb/phone thing works Sep 26 12:19:13 (after winning a huge amount of money offcourse) Sep 26 12:19:15 I've pretty much put my head in the ground so far Sep 26 12:23:09 koen: heh, n90 is that -really- expensive one, isnt it? Sep 26 12:23:14 with lots of camera features Sep 26 12:23:32 * kergoth is really happy with his 6260 so far Sep 26 12:23:56 what about the song walkman phone Sep 26 12:24:26 sony? i dont like sony. they engage in way too much proprietary bs. Sep 26 12:24:29 heh Sep 26 12:24:39 kergoth: 600 euros when I last checked it Sep 26 12:24:45 heh, ouch Sep 26 12:24:48 which is waaaaaaaaaaaaaay too expensive Sep 26 12:24:50 :) Sep 26 12:25:07 even if I had the money I wouldn't buy a E600 phone Sep 26 12:25:31 i want to try some symbian apps, see if i can use it for pda functionality effectively Sep 26 12:25:37 I just get a Motorola A780 Sep 26 12:26:29 as long as it can a) make calls and b) do EDGE over bt I'm happy Sep 26 12:26:56 somehow b) is the rare part Sep 26 12:27:07 heh Sep 26 12:27:10 :P Sep 26 12:27:33 koen: the sony walkman phone itune on it Sep 26 12:27:37 all i wanted was: flip phone, standard OS (i.e. not the motorola proprietary crap), bluetooth, mmc/rsmmc, GSM, unlocked Sep 26 12:27:47 finding a phone that met those requirements was a pain Sep 26 12:27:55 kergoth: my motorola is actually linux Sep 26 12:27:57 AvengerMoJo: I have an ipod for that :) Sep 26 12:28:03 cool Sep 26 12:28:17 AvengerMoJo: any idea what processor is in that? Sep 26 12:28:23 xscale Sep 26 12:28:30 ah Sep 26 12:28:34 hum :) Sep 26 12:28:42 they have a hacked opie port Sep 26 12:28:48 but I just get it ... didn't get a time to hack it yet Sep 26 12:28:59 but that was made using some 'discovered' header files Sep 26 12:29:15 god why am i so tired Sep 26 12:29:19 caffeine isnt working Sep 26 12:29:23 grr Sep 26 12:29:29 dehydrated? Sep 26 12:29:38 hmm, could be Sep 26 12:29:54 * AvengerMoJo is still wondering what is the bootsplash-native do and bootsplash working or not? Sep 26 12:30:31 AvengerMoJo: you aren't making a whole lot of sense there. Sep 26 12:30:50 pb_: well there is bootsplash bb in the database Sep 26 12:31:31 pb_: but I don't know what they do ... one is bootsplash-native .. and the bootsplash didn't build ... I won't I can patch that and does it work with 2.6 kernel Sep 26 12:32:51 also ... I think it required freetype .. I don't know I need to build a native freetype for it or not Sep 26 12:33:16 when I build bootsplash it didn't seem to build freetype automagically Sep 26 12:34:10 It looks like it does to me Sep 26 12:34:35 .... Sep 26 12:36:14 oh... may be because I already get freetype build .... Sep 26 12:36:23 * AvengerMoJo kick himself ... Sep 26 12:36:52 so the bootsplash package suppose to work with 2.6? Sep 26 12:40:15 sure Sep 26 12:41:19 CosmicPenguin: but I think my ... get some problem compiling Sep 26 12:41:45 CosmicPenguin: the make file missing including the -I*/include/freetype2 Sep 26 12:42:21 Ok, well there may be some problem building the bootsplash code, but that has nothing to with kernel 2.6 Sep 26 12:42:52 CosmicPenguin: ok :) how can I add -I in the oerunmake ? Sep 26 12:43:03 just add it add the end? Sep 26 12:45:43 Are you saying that the CFLAGS arn't being passed in? Sep 26 12:46:48 CosmicPenguin: missing the -I*/include/freetype Sep 26 12:46:50 CosmicPenguin: missing the -I*/include/freetype2 Sep 26 12:46:56 I mean Sep 26 12:49:43 Well, then you'll need to add that to the CFLAGs Sep 26 12:50:21 CosmicPenguin: I can just put that in the bb file right Sep 26 12:50:25 in the do_compile Sep 26 12:51:08 no - you don't want to override the cflags Sep 26 12:51:11 you want to add to the cflags Sep 26 12:51:41 recall at this point that the CFLAGS variable is just that - a variable Sep 26 12:54:20 ok Sep 26 12:54:45 also fbmngplay.h:17:20: libmng.h: No such file or directory Sep 26 12:55:02 ... wonder my database get problem or just the package not being updated? Sep 26 12:55:37 No - bootsplash is definately broken in that regard Sep 26 12:55:44 it should be marked BROKEN="1" Sep 26 12:56:14 .... ok ... so ... even I get it build I should not able to get it working ? Sep 26 12:56:32 CosmicPenguin: mv bootsplash/ non-working/bootsplash/ Sep 26 12:56:49 that too Sep 26 12:57:04 AvengerMoJo: you'll need to add libmng to the DEPENDS and hope for the best Sep 26 12:57:09 koen: the bootsplash package just the utility right? Sep 26 12:57:26 koen: I still need the kernel patch after that? Sep 26 12:57:33 CosmicPenguin: will do Sep 26 12:57:42 When I did bootsplash, I only used the -native, I didn't care about any of the userland animnations and bs like that Sep 26 12:58:28 CosmicPenguin: what does the -native do? Sep 26 13:00:06 AvengerMoJo: no idea Sep 26 13:00:55 koen: hehe :) I build libmng .. but it didn't put the header and the library in the staging ... Sep 26 13:07:57 koen: do you want me to write a patch for the package? Sep 26 13:08:22 koen: I finish the build but I'm not sure it is working in the system yet Sep 26 13:10:03 AvengerMoJo: sure, go ahead, you can attach the result to bugzilla Sep 26 13:10:37 koen: on Sep 26 13:10:38 ok Sep 26 13:11:58 03koen 07org.oe.dev * r3aa50a83... 10/conf/machine/ (h3600.conf h3900.conf ipaq-pxa270.conf): conf/machine/: add ROOT_FLASH_SIZE for h3600 (16), h3900 (32) and ipaq-pxa270 (80) Sep 26 13:12:43 sorry .... where is the bugzilla hostname Sep 26 13:12:56 url I mean Sep 26 13:12:57 AvengerMoJo: see topic :) Sep 26 13:13:03 ops Sep 26 13:13:04 thanks Sep 26 13:13:30 * AvengerMoJo is getting tired Sep 26 13:16:39 koen: try svn co http://opensvn.csie.org/bitbake/trunk/bitbake Sep 26 13:17:44 kergoth: seems to be working Sep 26 13:17:59 cool. did an svnadmin dump, then load, so all the history is over there Sep 26 13:18:05 Checked out revision 313. Sep 26 13:18:06 powerbook-wlan:~/Projects/OpenEmbedded koen$ Sep 26 13:18:12 even on OSX :) Sep 26 13:18:15 so if we want to ditch berlios, we can Sep 26 13:18:39 can we use svk between berlios and csie? Sep 26 13:21:25 koen: to keep them in sync? probably, havent tried Sep 26 13:22:51 * kergoth tries Sep 26 13:23:00 AvengerMoJo: -native just builds the bootsplash utilities needed to create the initrd Sep 26 13:24:55 CosmicPenguin: thanks Sep 26 13:25:05 CosmicPenguin: so I need to rebuild the initrd? Sep 26 13:25:15 meaning the pivitboot image? Sep 26 13:25:41 I think the utility I built work Sep 26 13:25:58 but I think I will missing the kernel patch hehe :P Sep 26 13:27:30 Yes, you do need the kernel patch, but thats none of bootsplash's business Sep 26 13:27:42 hehe :) Sep 26 13:27:58 koen: okay, yeah, looks like i can Sep 26 13:28:20 CosmicPenguin: wonder the offical patch work for spirtz? Sep 26 13:28:37 what framebuffer does spirtz Sep 26 13:28:37 don't know, don't care - I don't have any of those devices Sep 26 13:28:51 CosmicPenguin: I see :) thanks anyway Sep 26 13:29:48 kergoth: nice Sep 26 13:30:12 i could set up a cron job to maintain one of them as a backup to the other, if we want Sep 26 13:30:17 or just say screw it and abandon berlios Sep 26 13:30:48 judging from the past we'll do teh former and a bit later the latter Sep 26 13:30:51 the* Sep 26 13:36:53 http://www.penny-arcade.com/view.php?date=2005-09-26&res=h Sep 26 13:39:23 skymall? Sep 26 13:40:05 odd little catalog on some airplanes that sells things from tons of other companies. didnt think about it, but maybe us only Sep 26 13:40:12 yeah Sep 26 13:40:23 I had some fun browsing through it Sep 26 13:40:33 i love http://www.penny-arcade.com/view.php?date=2005-07-29&res=l Sep 26 13:40:56 I had some real fun in the real life sharper image stores Sep 26 13:41:04 the amount of useless gizmos is amazing Sep 26 13:41:08 hehe. their stuff is so overpriced Sep 26 13:41:11 yeah Sep 26 13:41:36 'air purifier' 'ionizing air purifier' 'ionizing air purifier with hummer h2 look' Sep 26 13:41:42 lol Sep 26 13:41:56 some people have just too much money Sep 26 13:42:12 since someone must be buying that crap or they wouldnt still be in business Sep 26 13:42:17 indeed Sep 26 13:42:37 the discovery channel stores was much better :) Sep 26 13:42:59 * koen looks at the scale model of the OCC comanche bike Sep 26 13:43:01 yeah, but the stuff in the discovery channel stores are of poorer quality Sep 26 13:43:44 the amount of hummer shaped gadgets is stunning Sep 26 13:44:23 as is the use of cast-iron in said gadgets Sep 26 13:44:49 * koen wonders what the TSA would think of that stuff Sep 26 13:45:03 a large hunk of metal + batteries inside a suitcase Sep 26 13:45:30 RP: are you here? Sep 26 13:48:10 anyone knnow a good place to get custom wired db9 cables? Sep 26 13:48:34 for quantities <5: my desk ;) Sep 26 13:49:07 heh, this is for q > 25 Sep 26 13:49:23 need a db9 3 pin serial cable (no control lines) Sep 26 13:49:37 can't find anyone who sells stuff other than full up 9pin straight through Sep 26 13:49:45 or full custom Sep 26 13:59:03 WTH? Sep 26 13:59:07 no nethack in OE? Sep 26 13:59:32 hehe Sep 26 13:59:40 andersee complains about that once in a while Sep 26 14:00:29 heh Sep 26 14:00:35 you scared him away Sep 26 14:01:06 heh Sep 26 14:01:43 gb2: can we talk you into adding gcc 4.02 when it's released? Sep 26 14:01:48 heh, not even opie-nethack? Sep 26 14:02:03 koen@dominion:/data/build/oe/monotone/org.openembedded.dev/packages$ ls | grep hack Sep 26 14:02:03 matchbox-panel-hacks/ Sep 26 14:02:06 that's all Sep 26 14:02:42 crumbs Sep 26 14:03:03 Hi everyone, I'm building epia machine with uclibc based distro. It tries to build gcc 4 and fails, but I actually need gcc 3.x - how to tell bitbake to build cross 3.x ? Sep 26 14:03:21 lazy_marmot: check conf/distro/familiar-0.8.3.conf Sep 26 14:03:38 Thanks! Sep 26 14:50:53 I thought that gcc 3.4.4 was the default set for DISTRO = "openzaurus-3.5.4"; has this changed, or is my build just going mad (trying and failing to build gcc 4.0.0)? Sep 26 14:52:27 going mad would be my bet Sep 26 14:52:39 yeah, much like me I reckon ;) Sep 26 14:53:42 The other odd thing it did was to build gcc-cross-initial 3.x.x which wasn't the preferred 3.4.4, complaining that it wasn't there (but it was) Sep 26 14:55:42 As a matter of interest, is the gcc-cross-initial there for bootstrapping? It builds libc and the binutils? Is libc then rebuilt using the actual gcc-cross? Sep 26 14:58:39 lardman: yes Sep 26 14:59:01 mickeyl: so initial can be a different version, & differently patched....? Sep 26 14:59:25 can be, yes Sep 26 14:59:27 binutils is built with the native toolchain, not the cross-initial version. cross-initial is used to get enough of a toolchain to get the c library built, since that's needed for the next pass of gcc Sep 26 14:59:39 ah, okay Sep 26 15:00:27 hi mickeyl Sep 26 15:00:42 evening pb_ Sep 26 15:00:49 wow, lots of toolchain gurus all at once! ;) Sep 26 15:01:01 lardman: pester them while they're here Sep 26 15:01:03 heh Sep 26 15:01:45 * kergoth runs away Sep 26 15:01:54 * mickeyl holds him back Sep 26 15:01:55 toolchain gurus: I'd like to get the cross toolchain built using softvfp rather than softfpa for the soft-float implementation; any bright ideas....? Sep 26 15:02:12 heh, not offhand. i dont think our soft float patch makes that easy Sep 26 15:02:33 I'm not sure gcc is even patched with that anymore... not in 3.4.4 Sep 26 15:04:01 what's necessary to make the eabi thing happen? vfp is the default there iirc Sep 26 15:04:25 mickeyl: patches to glibc Sep 26 15:04:42 any ETA ? Sep 26 15:04:49 or is it just about applying them in OE? Sep 26 15:04:50 and to gcc too I'd imagine as that's still softfpa Sep 26 15:05:25 (I've found a patch for gcc 3.4.4 for softvfp, but I do wonder about what else I'd need to patch, etc...) Sep 26 15:05:28 mickeyl: you'd have to ask pb_ or any other arm expert Sep 26 15:05:52 * mickeyl looks to pb_ or any other arm expert Sep 26 15:06:02 * koen reads http://en.wikipedia.org/wiki/ARM_architecture Sep 26 15:10:45 I don't know exactly what the status is. That'd be a question for drow. Sep 26 15:11:07 koen: interesting Sep 26 15:11:10 I haven't been paying too much attention to the EABI since I stopped working at Nexus. Sep 26 15:13:30 vfp vs fpa? Sep 26 15:13:47 shall I...? Sep 26 15:14:20 pb_: vfp is *the* thing to use nowadays? Sep 26 15:14:27 and fpa a relic of the past? Sep 26 15:14:27 indeed Sep 26 15:14:32 both instruction sets (I think) which relate to actual hardware, but which we emulate with our particular soft-float implementation Sep 26 15:15:00 vfp is the thing to go for, and its data has the same endianness as the processor while fpa is always bigendian Sep 26 15:15:25 which is why I noticed it as I was trying to sort out a warning which Octave was throwing up Sep 26 15:16:33 hope that makes some sense and is along the right lines... Sep 26 15:17:25 koen: if I add a new package .... what can I do Sep 26 15:23:27 I've just rm -rf'd my org.openembedded.dev dir, but will try to build one of the earlier version of gcc (to which the softfloat patch is actually applied); am I right in thinking that it will give me softvfp by default and softfpa if I specify -soft-float? Sep 26 15:24:09 You'd probably be better off using gcc 4.0, which has the soft-float bits incorporated already. Sep 26 15:24:34 3.4.4 has them too, but only seems to build softfpa Sep 26 15:24:35 I don't know what the situation is with vfp using the softfloat patch that's in OE. Sep 26 15:24:49 from the patch: %{!mhard-float: %{msoft-float:-mfpu=softfpa} %{!msoft-float:-mfpu=softvfp}} Sep 26 15:25:17 I read that as: no -hard-float & no -soft-float = softvfp? Sep 26 15:26:37 that's correct, but those options don't actually affect the generated code Sep 26 15:27:28 but if I set TARGET_FPU="", won't I get softvfp? Sep 26 15:28:03 dunno. you certainly can't infer that from the bit of specs you pasted. Sep 26 15:28:15 oh Sep 26 15:29:19 worth a try, though Sep 26 15:29:53 I'll have a look at that, but there's a later patch (http://www.andric.com/cross/patches/) which seems to be vfp specific Sep 26 15:30:06 http://www.andric.com/cross/patches/gcc-3.4.3-arm-softvfp.patch.bz2 to be exact Sep 26 15:30:58 it seems a bit unlikely that the default for arm-linux would be softvfp in either 3.4.4 or 4.0, since that would break Familiar, Debian and any other system that expected hard fpa. Sep 26 15:31:04 I take it kernel 2.6 doesn't work for poodle properly yet? Sep 26 15:31:26 Does anything need to be done specifically for it? Hrm Sep 26 15:31:59 df001: correct. see http://www.cs.wisc.edu/~lenz/zaurus/ for what things you could do Sep 26 15:32:01 pb_: in 3.4.4 the default is fpa (not soft) as I tried that Sep 26 15:32:24 df001: a couple of these things may already be done as it seems Sep 26 15:32:39 df001: I suggest building a recent kernel and just having a look at what goes and what not Sep 26 15:32:52 pb_: but 3.4.3 still uses the softfloat patch iirc, so this might (he hopes) be a drop in replacement Sep 26 15:33:37 ah, right Sep 26 15:33:41 yeah, that might work Sep 26 15:34:27 I'll have a look at it tomorrow - I'm just going to rebuild the toolchain now to make sure I get the correct version selected (as it'd gone mad earlier this afternoon) Sep 26 15:34:41 anyway night all, time for bed me thinks Sep 26 15:34:47 g'night lardman Sep 26 15:35:12 Maybe I'll give it a try as soon as OE is set up entirly again. It's probably just a matter of porting over the 2.4.x drivers, isn't it? Sep 26 15:35:34 'night all Sep 26 15:35:38 n8 koen Sep 26 15:35:41 "just", yes. Sep 26 15:35:48 night koen Sep 26 15:36:02 "just" being rewriting a whole lot of stupidity Sep 26 15:36:31 because our goal is not to get something working, but to make it right and submittable to mainline Sep 26 15:36:40 s/not/not just/ Sep 26 15:37:37 Right Sep 26 15:39:34 Wait, stupidity? Sep 26 15:39:58 to call it stupidity would be an understatement Sep 26 15:40:12 2.4.20-holyshitix Sep 26 15:40:31 what will the image directory name called? Sep 26 15:40:37 tmp/deploy/images Sep 26 15:47:40 Does the kernel still need gcc 2.95 to compile? haha Sep 26 15:48:48 2.4.{18|20}-ancientix, yes Sep 26 15:52:29 xserver-kdrive_cvs is failing because randrproto.pc is missing. what's the fix? Sep 26 15:52:57 sounds like adding DEPENDS=xrandr or so Sep 26 15:53:17 ah, I'll try that. thanks mickeyl Sep 26 15:55:33 g'night Sep 26 15:55:40 night mickeyl Sep 26 16:37:38 grrrr Sep 26 16:37:55 mickey|zzZZzz: newest version of evas doesn't link correctly with freetype....argh! Sep 26 16:49:27 sweet - xml parsing is a breeze with python Sep 26 16:58:10 mickey|zzZZzz: ok, I "fixed" it for evas-x11...-fb and -native will probably still fail... Sep 26 17:04:45 damn it...why are there autoconf problems all of a sudden! Sep 26 17:15:48 o_o Sep 26 17:15:50 configure.in:106: error: possibly undefined macro: AC_DEFINE Sep 26 17:18:31 damn.....things...... Sep 26 17:31:12 Hey. I want to build the OZ kernel for a pxa device. bitbake openzaurus-pxa Sep 26 17:31:12 I want it to give me menuconfig so I can enable/disable a few things of my own Sep 26 17:31:16 is there any way to do that? Sep 26 17:32:07 openzaurus-pxa_2.4.18-rmk7-pxa3-embedix20031107.inc maybe I have to edit this file? Sep 26 17:33:08 or i can just overwrite the .config file for poodle..hmm Sep 26 17:43:22 'night all Sep 26 19:11:01 * Geo_KM is away: Away at the moment Sep 26 19:40:13 your up erly raster Sep 26 19:43:31 :o Sep 26 19:44:00 well i finally figured out how to make getaddrinfo() work Sep 26 19:44:16 well the ipv4 part anyway Sep 26 21:21:58 kergoth: you here? Sep 26 21:22:18 Are the OE .15 orinoco modules no good? Sep 26 21:22:39 im getting a bunch of weird BAP offset errors in dmesg Sep 26 21:25:32 someone was mentioning problems in certain setups Sep 27 00:09:48 * Geo_KM is away: Away at the moment Sep 27 00:10:38 Hrm Sep 27 00:10:43 didnt someone add a bitlbee package to OE? Sep 27 01:03:28 morning all Sep 27 01:20:10 good morning all Sep 27 01:20:36 morning koen Sep 27 01:21:08 hey RP Sep 27 01:21:23 morning ! Sep 27 01:21:31 hey alan|home Sep 27 01:21:40 oy koen Sep 27 01:29:39 hey Sep 27 01:29:41 can anyone confirm that there exists no linux drivers for this -> http://www.lifeview.com.tw/html/products/external_tv/q_dvbt.htm Sep 27 01:29:46 i did a quite extensive search but came up with nothing Sep 27 01:30:58 RP: I get some question for build kernel for spritz Sep 27 01:31:41 RP: can you give me some direction so I can try custom the kernel a little Sep 27 01:32:01 RP: I'm trying to patch in the bootsplash for my kernel Sep 27 01:33:06 morning Sep 27 01:33:17 hi hrw|work Sep 27 01:33:56 RP: if you get some info for step configure the kernel please let me know :) Thanks Sep 27 01:34:16 hmm.. I need to clean staging a bit to get libsdl working correctly Sep 27 01:34:27 hrw|work: oh :) Sep 27 01:34:41 hrw|work: you are using libsdl on framebuffer? Sep 27 01:34:48 morning Sep 27 01:34:52 ~lart ADSL Sep 27 01:34:52 * ibot tries to shut ADSL up Sep 27 01:34:53 hi XorA Sep 27 01:35:28 AvengerMoJo: no - over 3wire serial Sep 27 01:35:37 oh Sep 27 01:35:49 hrw|work: interesting Sep 27 01:36:34 hrw|work: wtf? Sep 27 01:38:00 joking :) Sep 27 01:38:21 AvengerMoJo: libsdl-qpe Sep 27 01:38:23 hehe Sep 27 01:38:32 hrw|work: I see Sep 27 01:38:51 mithro: this one should work: http://tvde.terratec.net/modules.php?op=modload&name=News&file=article&sid=238 Sep 27 01:39:02 according to zecke Sep 27 01:39:19 but using libsdl on top of qpe will be better then libsdl on directfb? Sep 27 01:39:36 AvengerMoJo: depends on enviroment Sep 27 01:39:43 * AvengerMoJo never try directfb on zaurus Sep 27 01:40:00 hrw|work: what do you mean .. Sep 27 01:40:14 AvengerMoJo: under opie libsdl-qpe is better, under x11 it will be libsdl-x11 Sep 27 01:40:23 AvengerMoJo: under plain FB I dont know which one Sep 27 01:40:24 I see Sep 27 01:40:56 libsdl-x11 is the normal libsdl Sep 27 01:41:07 and that can paint directly to the framebuffer Sep 27 01:41:40 koen: so libsdl use linuxfb directly Sep 27 01:41:49 it can Sep 27 01:41:55 sdl has multiple backends Sep 27 01:42:08 koen: I try the sdl on directfb on via C3 Sep 27 01:42:21 that should work Sep 27 01:42:21 but the cost is too high Sep 27 01:42:28 slow Sep 27 01:42:44 so I end up using directfb directly Sep 27 01:42:45 | /home/hrw/devel/build/0912/tmp/cross/lib/gcc/arm-linux/3.4.4/../../../../arm-linux/lib/crt1.o(.text+0x30): In function `_start': Sep 27 01:42:47 doesn't X have accelerated drivers for that? Sep 27 01:42:48 | init.c: undefined reference to `main' Sep 27 01:42:50 ;( Sep 27 01:43:03 libsdl-(image,mixer,ttf) give me that after rebuild of libsdl-qpe ;( Sep 27 01:43:04 oh Sep 27 01:43:18 that look like what I have seen before .... Sep 27 01:43:37 ..... forget what is that problem Sep 27 01:43:57 hrw|work: sounds like you still got the -Dmain=SDLMain sdl-config lieing around Sep 27 01:44:48 AvengerMoJo: I'm afraid I don't understand what you're asking about the spitz kernel Sep 27 01:45:00 yep Sep 27 01:45:42 XorA: removed sdl-config.. rebuilding Sep 27 01:46:15 RP: since I'm build the kernel with your bb file ... I'm not sure how do you configure your kernel :) Sep 27 01:46:57 RP: I want to patch the kernel and add more configuration Sep 27 01:47:49 AvengerMoJo: You need to edit the defconfig-spitz file then... Sep 27 01:48:10 oh Sep 27 01:48:13 RP: thanks Sep 27 01:48:20 AvengerMoJo: But be warned - the devices have a limit as to how big the kernel can be Sep 27 01:48:28 RP: right Sep 27 01:48:45 RP: I need some briefing :) if you get time want I need to be care about Sep 27 01:49:36 ops ... what I need to know .. I mean Sep 27 01:50:26 when i patch CPAR050218.patch in openzaurus-pxa27x... i notice that i can't compile it with old embedix kernel anymore... -mcpu=iwmmx not recognize... Sep 27 01:51:02 do i need gcc 3.4.x to compile the kernel? Sep 27 01:51:10 zuery: ask tetsu Sep 27 01:51:27 zuery: in OE noone maintain 2.4.(18|20)-crapix kernels Sep 27 01:52:04 ya.. i know.. but i'm using akita.. so no choice... 'yet'... Sep 27 01:52:09 koen: that is more expensive and a lot bigger Sep 27 01:52:18 plus i don't have a supplier here in Sweden Sep 27 01:53:57 zuery: You do have a choice - 2.6 is mostly working for the cxk models... Sep 27 01:54:15 zuery: http://www.rpsys.net/openzaurus/ Sep 27 01:58:29 zecke_: ping Sep 27 01:58:54 or anyone who understand libsdl-qpe Sep 27 01:59:29 it still has Dmain=SDL_main but not link with libSDLmain Sep 27 02:05:42 morning Sep 27 02:05:53 hi rene Sep 27 02:06:03 hey Marcin Sep 27 02:06:48 hrw|work: Are you in a position to help with this: http://lkml.org/lkml/2005/9/24/23 ? Sep 27 02:07:25 w8 Sep 27 02:08:08 RP: it will take two days but I can Sep 27 02:08:46 I dont have armv4l toolchain now Sep 27 02:09:34 :) Sep 27 02:10:18 hrw|work: My problem is I don't understand how to get bluetooth into the base image with the size restricitions on collie... Sep 27 02:10:34 ops. I have armv4l - oz353 Sep 27 02:10:48 hrw|work: He is actively working on the collie 2.6 kernel so it would be good to help him if we can Sep 27 02:10:52 RP: we can remove some stuff like ipv6, ipsec Sep 27 02:10:56 RP: sure Sep 27 02:12:14 RP: bootstrap isn't 14MB Sep 27 02:12:31 RP: ipkg install kernel-module-bluetooth bluez-utils-nodbus Sep 27 02:13:28 RP: rm /etc/init.d/check-version.sh Sep 27 02:13:52 RP: it seems Pavel has a real problem either reading docs or using google Sep 27 02:14:29 koen: or Ctrl-C instead of rm Sep 27 02:14:46 and without PIM+games the image is a lot smaller too Sep 27 02:14:47 koen: Feel free to reply to the mail and explain that - don't be too nasty as he could be quite useful though ;-) Sep 27 02:15:08 NOTE: package minimal-opie-image-1.0-r1: task do_rootfs: started Sep 27 02:18:05 RP: how big the kernel could be for spirtz? Sep 27 02:18:43 hrw|work: I added ROOTFLASHSIZEs to the ipaqs too, so opie can install more crud on newer models Sep 27 02:18:56 morning all Sep 27 02:19:01 hey lardman Sep 27 02:19:07 h ikoen Sep 27 02:19:10 oops Sep 27 02:19:14 koen: nice Sep 27 02:19:53 koen: you were right about my build, it had gone mad - I removed everything and re-pulled and it built fine last night Sep 27 02:21:26 heh.. looks like I need to build from scratch arm4l Sep 27 02:21:59 morning lardman Sep 27 02:22:07 AvengerMoJo: Read the kernel .bb file Sep 27 02:22:27 Hi RP, nice weekend? Did the rain stop for you? Sep 27 02:23:37 lardman: Not on the way there but it was mostly dry once I got there :) Sep 27 02:31:28 I'm doing a pull from the monotone repository and I get errors like Sep 27 02:31:29 monotone: warning: discarding revision data packet 5bbc6ec4525557db317977a31210230c82e55137 with unmet dependencies Sep 27 02:31:33 is that normal? Sep 27 02:33:39 ~lart queue muzak Sep 27 02:33:40 * ibot stuffs queue muzak into a shiny new tin can and vacuum seals it Sep 27 02:36:47 RP: thanks Sep 27 02:40:11 hi folks Sep 27 02:40:18 hey ade|desk Sep 27 02:47:12 ~lart cron Sep 27 02:47:13 * ibot forces cron to use Outlook Express Sep 27 02:47:52 3 days of changes in OE is a 78k compressed diff Sep 27 02:50:03 where you host those diffs? Sep 27 02:53:38 http://ewi546.ewi.utwente.nl/OE/ Sep 27 02:55:05 http://ewi546.ewi.utwente.nl/OE/OE.db.bz2 is now 29MB Sep 27 02:55:15 and it only contains org.openembedded.dev Sep 27 02:57:22 nice Sep 27 02:57:40 I adjusted the wording in GettingStarted Sep 27 03:00:52 hmm.. I have to archive some mailboxes - mainly ML which I'm not subscribed now Sep 27 03:01:08 doh, my wife just found my secret stash of money i was saving up to buy a c1000. she is planning on getting a set of chairs instead :( Sep 27 03:01:27 ade|desk: ;( Sep 27 03:03:58 03jbowler 07org.oe.dev * rb5619393... 10/conf/distro/openslug.conf: openslug: 2.7 in preparation for a source release Sep 27 03:04:02 03jbowler 07org.oe.dev * r63b3233a... 10/conf/distro/openslug-packages.conf: openslug: updated .conf file correct for 2.7 release Sep 27 03:04:07 03jbowler 07org.oe.dev * r1a5ea175... 10/conf/distro/openslug-packages.conf: openslug-packages.conf: undo accidental removal of irrsi/* Sep 27 03:04:11 * Geo_KM is back Sep 27 03:04:13 03mickeyl 07org.oe.dev * r5d8bca6d... 10/packages/tcltk/tk_8.4.11.bb: disapproval of revision 05dd4ce9d2ee0995688e45e60a21534114fbc45f Sep 27 03:04:17 03jbowler 07org.oe.dev * r8d51140c... 10/packages/freeze/ (files/freeze freeze.bb): freeze: add extra package/version syntax so irssi works Sep 27 03:04:21 03jbowler 07org.oe.dev * r578bbee3... 10/conf/distro/openslug.conf: openslug: revert accidental change to the preferred kernel version Sep 27 03:04:25 03jbowler 07org.oe.dev * r2cc82e33... 10/conf/distro/openslug-packages.conf: openslug-packages: remove tcltk for release Sep 27 03:04:30 03jbowler 07org.oe.dev * rbbe482d8... 10/conf/distro/openslug.conf: openslug: move head to openslug-2.8 Sep 27 03:04:35 03jbowler 07org.oe.dev * r745a0b2f... 10/ (2 files in 2 dirs): Sep 27 03:04:37 ucslugc: updated/corrected packages list Sep 27 03:04:39 This fixes the misleading perl-with-broken-thumb mistake as well as adding Sep 27 03:04:41 those packages recently added to openslug which actually compile - see the Sep 27 03:04:43 broken packages list for other packages from openslug which probably have Sep 27 03:04:45 very simple compile problems. Sep 27 03:04:59 03jbowler 07org.oe.dev * r776f76fe... 10/ (3 files in 3 dirs): ucslugc: move to kernel 2.6.14-rc2 and ucslugc-3 Sep 27 03:05:07 http://monotone.vanille.de/viewmtn/headofbranch.psp?branch=org.openembedded.oz354fam083 Sep 27 03:07:24 koen: will it be lot of pulling to get it? Sep 27 03:07:39 7k certs, but only one rev Sep 27 03:07:46 and about 3MB of communication Sep 27 03:07:49 ok Sep 27 03:09:23 27 12:05 < koen> http://monotone.vanille.de/viewmtn/headofbranch.psp?branch=org.openembedded.oz354fam083 Sep 27 03:09:26 ops Sep 27 03:11:50 pulled Sep 27 03:13:35 mail sent to oe@ Sep 27 03:15:57 so future upgrades to 3.5.4 or 0.8.3 goes into oz354fam083 branch Sep 27 03:16:03 yes Sep 27 03:16:16 not propagate but copy/add Sep 27 03:16:24 yes Sep 27 03:16:28 ok Sep 27 03:16:52 unless monotone supports cherrypicking Sep 27 03:17:01 but last I heard, that doesn't work well Sep 27 03:17:55 I'll talk to mickeyl tonight and after that we'll probably remove 354 and 083 from the .dev branch Sep 27 03:18:46 RP: you can break the kernel now Sep 27 03:18:46 next distro conf changes should allow usage of plain 'openzaurus' or 'familiar' (not numbered) I think Sep 27 03:19:30 I suspect mickeyl and I will go for 360 and 090 Sep 27 03:19:37 but we'll discuss that tonight Sep 27 03:19:59 echo 'BROKEN="1"'>>packages/linux/linux-openzaurus_2.6.14-rc1.bb;mt commit --message "kernel is broken";mt sync Sep 27 03:21:42 vanillle, ewi and dominion are serving the oz354fam083 branch Sep 27 03:26:55 koen: Contry to popular belief, I don't aim to break them :) Sep 27 03:27:03 :) Sep 27 03:27:17 It's just luck ;) ? Sep 27 03:27:22 RP: no? buu Sep 27 03:28:35 RP: pcmcia ids hit mainline? Sep 27 03:29:44 Mother of god. OZ is terrible on poodle Sep 27 03:30:23 I can make it somewhat work by building stuff in OE and installing it overwriting whats broken -_- Sep 27 03:30:40 hrw|work: I haven't pushed our extra ones yet Sep 27 03:31:20 df001: poodle lacks testers - if we know what's wrong, we can fix it... Sep 27 03:31:48 oh, I can test. Sep 27 03:32:26 kernel, IGNORE_USER_SUSPEND disabled, holding right on the dpad, or sometimes letters causes the PDA to suspend, sometimes it won't wake back up Sep 27 03:32:42 IGNORE_USER_SUSPEND enabled, holding cancel shuts it off, but the other buttons wont Sep 27 03:33:21 df001: You need to file bug reports on the problems - make sure to mention which kernel+image versions you're using Sep 27 03:33:26 Allright Sep 27 03:34:06 df001: could you test kernel for me? with right-arrow fix Sep 27 03:34:30 You have it fixed properly? Sep 27 03:34:58 And yeah, I can. is there any way to flash a new kernel without having to flash everything else as well? Sep 27 03:34:58 not me but it should be ok Sep 27 03:35:11 put updater.sh and zImage.bin on card and flash Sep 27 03:35:14 no initrd.bin Sep 27 03:35:14 if I don't put initrd.bin on the SD card as well, the device will no longer work Sep 27 03:35:32 it just sits at the OZ splash screen Sep 27 03:35:41 df001: http://www.hrw.one.pl/tmp/poodle/ have a kernel Sep 27 03:35:57 Allright Sep 27 03:36:00 df001: Sounds like you flashed a kernel that was incompatible with your filesystem Sep 27 03:36:10 RP: not the case. If I flash both, it's fine Sep 27 03:36:21 I never used to have this problem. Is there only one updater.sh? Sep 27 03:37:15 hrw: can I also suggest a kernel be built with cache erratica fix set to N? Any modern poodle has a 255 in it, and having that workaround enabled slows the device down a lot Sep 27 03:37:38 df001: No, you can't suggest that as it would break for anyone with the 250 Sep 27 03:37:58 df001: we would rather give two kernels then disable it Sep 27 03:38:04 yeah Sep 27 03:38:05 Asking for two kernels, one with the fix and one without would be much better Sep 27 03:38:05 thats what I mean Sep 27 03:38:19 like zimage-poodle-255 zimage-poodle-250 Sep 27 03:39:19 df001: prepare and submit recipes for it Sep 27 03:39:39 hrm. Sep 27 03:40:18 haha. I'm really bad with build systems. I hacked the bb file in do_configure_prepend to overwrite the .config file with my own Sep 27 03:40:42 easier would be to replace defconfig-poodle ;) Sep 27 03:41:04 df001: fetch kernel and try - I want to know does that fix is correct Sep 27 03:41:17 http://www.openzaurus.org/official/unstable/3.5.3/sl5600/ will the updater.sh here allow me to flash the kernel alone? Sep 27 03:41:31 i don't wanna brick my PDA again Sep 27 03:41:59 should Sep 27 03:42:15 K. just zImage.bin then, correct? Sep 27 03:42:31 zimage.bin and updater.sh Sep 27 03:43:19 allright downloading Sep 27 03:45:12 flashing Sep 27 03:45:35 Also later today I'll prolly have a bb to submit. bitlbee compiled and worked Sep 27 03:45:53 its an IRC to various chat protocal gateway Sep 27 03:46:14 Add it to the bugtracker as an enhancement Sep 27 03:46:34 k Sep 27 03:46:36 the moment of truth Sep 27 03:46:44 OpenZaurus splash... Sep 27 03:47:04 Good itl work Sep 27 03:47:57 at home I have 2799 recipes, at work 2802 - have to clean a bit in hrw-packages Sep 27 03:48:42 hrw: can I have the bb for this kernel if it works Sep 27 03:49:05 it works, holding down right doesn't cause it to suspend Sep 27 03:49:25 df001: I'll push it Sep 27 03:49:29 hrw|work: nice work :) Sep 27 03:49:31 wait Sep 27 03:49:31 dont Sep 27 03:49:44 huh weird Sep 27 03:49:48 holding down cancel froze it Sep 27 03:49:53 it unfroze shortly after Sep 27 03:50:00 then i did it again and it suspended holding down cancel Sep 27 03:50:08 ..uh.. Sep 27 03:50:18 RP: eccerr0r 'ported' collie fix Sep 27 03:50:39 df001: it has to be like that - patches welcome Sep 27 03:51:19 when I read Sep 27 03:51:29 if I hold down cancel, it freezes, then it's fine Sep 27 03:51:37 click cancel, it'll suspend Sep 27 03:51:39 wake it up Sep 27 03:51:43 then itl suspend again by itself Sep 27 03:51:52 i guess thats better than holding down right and having it suspend Sep 27 03:51:54 Werner Schulte mails I want to mail something about topposting etc.. Sep 27 03:52:21 hrw|work: Turning on ignore_user_suspend, should that make it ignore the cancel button totally? Sep 27 03:52:39 I think it used to Sep 27 03:53:14 df001: dont ask kernel related questions to me - I'm not qualified to answer Sep 27 03:53:34 Ok :p Sep 27 03:53:40 df001: especially when it comes to crapix kernels from sharp Sep 27 03:53:48 ~lart sharp and lineo Sep 27 03:53:48 * ibot throws sharp and lineo's poor little doggy off a cliff Sep 27 03:53:59 I second that. Sep 27 03:54:07 I want to try porting over their 2.4 stuff to 2.6 Sep 27 03:54:13 I see theres a project page on it already Sep 27 03:55:54 hrw: is it pushed? Will I get it if I update OE now? Sep 27 03:56:10 df001: All patches for 2.6 poodle gratefully received Sep 27 03:56:38 :) Cool. Thanks Sep 27 03:57:11 2.6 should work an awful lot better than 2.4. Its know to work well on c7x0 and cxx00 devices... Sep 27 03:57:40 Yeah Sep 27 03:58:33 I saw a chat kergoth and someone else were having about the orinoco modules Sep 27 03:59:09 something like... version .15 spits garbage into dmesg about unexpected packets Sep 27 03:59:10 df001: I have to pull first - it take some time Sep 27 03:59:29 hrw|work: oh Sep 27 04:01:57 * Twiun peeks at SkyEye Sep 27 04:02:02 i dont think im ever going to buy anything sharp again. I've had this thing for over a year, and it's given me nothing but problems Sep 27 04:02:02 interesting] Sep 27 04:03:33 damn. seems that SkyEye is win32... wtf? Sep 27 04:04:26 or is just that the screenshots are only of that port? Sep 27 04:04:28 RP: hmm. how to build rootfs for collie/2.6 when our kernel does not build ;) Sep 27 04:05:09 hrw: is there anything special you did for your kernel? my machine is poodle, distro "openzaurus" Sep 27 04:05:18 hrw|work: copy the c7x0 defconfig over the collie one to make it compile - Pavel will have his own kernel Sep 27 04:05:26 my kernel just causes the PDA to sit on startup Sep 27 04:05:36 unless i flash it with initrd.bin Sep 27 04:05:38 df001: distro = openzaurus-3.5.4 might work better...? Sep 27 04:05:43 Hrm. Sep 27 04:06:01 3.5.3? Sep 27 04:06:10 3.5.4 Sep 27 04:06:28 I'll try it, although I thought thats not out yet Sep 27 04:07:20 nearly, anyway you are using an up to date oe.db I presume so you should go for the most up to date distro Sep 27 04:07:26 RP: will do Sep 27 04:07:49 is oe.db supposed to take eons to update? Sep 27 04:07:58 df001: yes Sep 27 04:08:10 it seems to slow down drastically after like 5-6 megs Sep 27 04:08:21 a watched kettle... Sep 27 04:10:25 Oh wow, I just realized monotone uses a lot of CPU Sep 27 04:15:35 df001: pull Sep 27 04:16:20 03hrw 07org.oe.dev * rebb1f3b5... 10/packages/linux/ (2 files in 2 dirs): Sep 27 04:16:20 openzaurus-pxa kernel: added APM fix for 'suspend on right-arrow' problem Sep 27 04:16:20 patch 'ported' to poodle by 'eccerr0r' from collie fix Sep 27 04:17:21 now I will have to push it into oz354fam083 branch Sep 27 04:18:09 I guess the point behind having that branch would be testing stuff first Sep 27 04:19:50 reenoo_: and df001 tested that kernel 0.5h ago Sep 27 04:20:08 its poodle for which we lack testers ;( Sep 27 04:20:36 damn, this DVB-T reciver is damn nice - but it doesn't have any linux support :/ Sep 27 04:21:21 hrw|work: and mickeyl need to adjust his cia script Sep 27 04:22:31 hrw|work: " my kernel just causes the PDA to sit on startup" <- doesn't sound like a successful test to me Sep 27 04:23:27 reenoo_: ok Sep 27 04:24:06 'sit' is better as 'play dead' ;) Sep 27 04:30:20 NOTE: Unpacking /home/hrw/devel/oe/org.openembedded.dev/packages/linux/openzaurus-pxa-2.4.18-rmk7-pxa3-embedix20031107/defconfig-poodle255 to /home/hrw/devel/build/0701/tmp/work/poodle255-linux/poodle255-kernel-2.4-embedix-2.4.18-rmk7-pxa3-embedix-r44/ Sep 27 04:30:25 nice Sep 27 04:31:14 reenoo_: he was talking about one he'd flashed before the hrw pushed the fix wasn't he? Sep 27 04:32:38 reenoo: _MY_ kernel Sep 27 04:32:41 As in one I compiled Sep 27 04:32:45 his kernel worked fine Sep 27 04:32:51 mithro: The Terratec Cinergy T2 is the best DVB-T receiver ;) Sep 27 04:33:02 Also this monotone stuff isn't working very well for me Sep 27 04:33:05 it's just sitting there Sep 27 04:33:19 df001: It does that, it should wake up eventually Sep 27 04:34:03 df001: good news for you - poodle and poodle255 kernels will be after my next push Sep 27 04:34:14 Awesome. Thanks Sep 27 04:35:48 df001: you have pxa255? Sep 27 04:35:51 Yeah, I do Sep 27 04:36:00 hrw|work: Is it not possible to enable/disable the bug work around using sysfs or similar (what is there in 2.4 - /proc?) so one kernel could be used Sep 27 04:36:22 lardman|work: I dont know and patches welcome Sep 27 04:36:24 haha. Sep 27 04:36:28 That would be nice, hmm Sep 27 04:36:37 hrw|work: ;) Sep 27 04:36:47 2.4 has sysfs? Sep 27 04:37:03 but yes, you could do it using proc Sep 27 04:37:04 Thing is, would that patch be usable on 2.6 as well? Sep 27 04:37:08 lardman|work: /proc/cpuinfo only iirc Sep 27 04:37:22 df001: same location as before - poodle255 kernel Sep 27 04:39:35 hrw: flashing Sep 27 04:41:07 hrw: it's sitting there..i wonder why it does this Sep 27 04:41:15 I think I might have to reflash initrd Sep 27 04:41:37 http://handhelds.org/scap/land.2569.png Sep 27 04:42:09 df001: ;( Sep 27 04:42:17 * koen thinks about doing 'ipkg install monotone' Sep 27 04:42:37 df001: Try reflashing with hrw's kernel first Sep 27 04:42:47 df001: I'd guess there is just something wrong with your kernel Sep 27 04:43:05 RP: i did, no dice, sits there, but it'll prolly work when i reflash it with initrd Sep 27 04:44:45 Ok, it works Sep 27 04:44:48 i had to reflash initrd Sep 27 04:44:56 http://pastebin.ca/24032 Sep 27 04:44:59 I'll try flashing the kernel solo again as soon as it boots up Sep 27 04:47:21 http://www.midwinter.com/%7Elch/programming/m7/ Sep 27 04:47:39 hrw: trying a solow reflash again Sep 27 04:48:54 hrw|work: No good. I have to flash it with initrd.bin or it won't work Sep 27 04:49:26 hrw|work: any ideas why it does that? Sep 27 04:49:39 it seems to be specific to the cache workaround fix Sep 27 04:49:44 maybe modules Sep 27 04:50:06 oh, yeah maybe Sep 27 04:50:55 that was a problem like that with r21 kernel on collie Sep 27 04:51:26 Is there a fix? Or is the fix just to install all the module ipks it builds? Sep 27 04:52:26 Can you throw those on your site? Sep 27 04:52:52 i can try initrd + kernel, then ipkg install those modules, then flash the kernel solo Sep 27 04:53:05 df001: same location - dir 'modules' and you need to add 'poodle255' into 'arch' lines in ipkg.conf Sep 27 04:53:26 Allright Sep 27 04:56:23 df001: update modules then flash newer kernel Sep 27 04:56:44 df001: just like in http://www.openzaurus.org/wordpress/2005/07/27/howto-get-r21-kernel-on-collie/ Sep 27 04:57:05 hrw: can you tar those modules Sep 27 04:57:19 downloading each individually is a pain..unless wget can somehow magically do it? Sep 27 04:57:37 wget -cr -np http://www.hrw.one.pl/tmp/poodle/modules/ Sep 27 04:58:10 it wants index.html Sep 27 04:58:39 hrw.one.pl/tmp/poodle/modules: No such file or directoryhrw.one.pl/tmp/poodle/modules/index.html: No such file or directory Sep 27 04:58:45 hm Sep 27 04:59:12 a.. you do it on poodle directly with busybox/wget? Sep 27 04:59:15 03hrw 07org.oe.dev * r086de75a... 10/packages/linux/ (5 files in 2 dirs): Sep 27 04:59:15 openzaurus-pxa for poodle: added possibility to build kernel for PXA250 and PXA255 Sep 27 04:59:15 this will need some tweaking in machine config because it add new 'arch' (poodle255) Sep 27 04:59:15 setting for ipkg which is needed only on PXA255 poodles Sep 27 04:59:29 i was doing it on my desktop, but i can do it on my PDA Sep 27 04:59:43 df001: my wget fetch it correctly Sep 27 04:59:50 actually Sep 27 04:59:53 wget 1.10.1 Sep 27 04:59:53 scratch that Sep 27 04:59:57 orinoco is broken :P Sep 27 05:00:03 no net on the pDA Sep 27 05:00:08 until i fix it Sep 27 05:00:23 I'm using 1.9.1 Sep 27 05:02:04 df001: you need orinoco 0.13a? Sep 27 05:02:39 yeah, hold on Sep 27 05:02:41 i'll just fix it Sep 27 05:03:08 whatever ox 3.4.3 ships with is broken, i had to build new modules Sep 27 05:04:36 ill let you know if installing the modules works in a few minutes Sep 27 05:07:25 also, hrw, it has been yelling at me saying it expects a different kernel and that it's going to pause for one minute Sep 27 05:08:30 df001: ignore Sep 27 05:10:40 Kk. Hopefully my wifi card will work now so i can get those ipks Sep 27 05:13:39 no such options -n or -r Sep 27 05:13:43 can you tar them? Sep 27 05:14:59 oki Sep 27 05:15:02 Thanks Sep 27 05:16:04 df001: http://www.hrw.one.pl/tmp/poodle/modules.tar.gz Sep 27 05:16:26 thanks Sep 27 05:17:06 hrw|work: if you want to backport into the stable branch, patch can handle stuff like http://monotone.vanille.de/viewmtn/getdiff.py?id1=ebb1f3b5b2d477b7e6feb7628032eb03920b318d&id2=086de75a7766964d8f97bff13a302c1f7d3c9fde Sep 27 05:18:06 koen: I know Sep 27 05:18:07 so just add poodle255 to ipkg.conf? Sep 27 05:18:50 df001: "arch poodle255 50" Sep 27 05:18:55 kk Sep 27 05:19:11 arch.conf Sep 27 05:20:20 there's unresolved symbols in hci_usb for bluetooth..ill ignore it Sep 27 05:20:22 i dont use that Sep 27 05:20:23 df001:can poodle with 255 kernel use kernel modules built for pxa250? Sep 27 05:20:32 im not sure Sep 27 05:20:40 Actually, i think id sassume yes Sep 27 05:20:43 can you check? Sep 27 05:21:16 Well Sep 27 05:21:28 everything works if i flash initrd and your zImage for 255 Sep 27 05:21:35 so I assume the modules are working, nothing funky in dmesg Sep 27 05:21:54 initrd with normal modules? Sep 27 05:22:00 normal == OZ 3.5.3 Sep 27 05:22:08 Right Sep 27 05:22:20 great Sep 27 05:22:33 Let me see if i can flash the kernel alone now though Sep 27 05:22:36 out-of-kernel ones will require rebuild for r44 kernel probably Sep 27 05:23:17 r44? Sep 27 05:24:08 ok flashed poodle 255 alone Sep 27 05:24:10 lets see if it works now Sep 27 05:24:32 doesn't look like it Sep 27 05:24:49 is there any way to enable like kernel debug output...like kprint or whatever it's called? Sep 27 05:24:52 maybe itl tell us what's wron Sep 27 05:25:21 What a wasted effort =\. It's bricked again. Sep 27 05:26:14 use serial cable Sep 27 05:26:24 I don't have one :\ Sep 27 05:26:35 i have the USB cradle Sep 27 05:27:01 This doesn't make any sense, why would it work if i flash it along with initrd... Sep 27 05:27:08 and not if it's alone Sep 27 05:27:19 with the errata fix enabled its fine either way Sep 27 05:27:43 I know I have the 255 Sep 27 05:27:53 it said so on my box, and cpuinfo I think mentioned it too Sep 27 05:28:22 hrw: is it possible to make an initrd.bin with dd? Sep 27 05:28:35 from device? Sep 27 05:28:45 like dd if=/dev/whatever / is on of=initrd.bin Sep 27 05:28:46 Yeah Sep 27 05:29:00 that way if I goof up my zaurus i have a backup Sep 27 05:29:15 df001: you can try to dd /dev/mtdblockNUMBER - look which one is / - df is your friend Sep 27 05:29:21 yeah Sep 27 05:29:26 you think thatd work? Sep 27 05:30:55 dont know Sep 27 05:31:47 Allright. I'll look into this later. I need some sleep. There's a way to turn on debugging messages, so stuff that goes into dmesg will print to the screen Sep 27 05:31:54 something like printk in /proc Sep 27 05:32:03 If I turn that on in my kernel config maybe itl say why it's sitting there Sep 27 05:35:46 hm Sep 27 05:35:51 its something that goes in CONFIG_CMDLINE Sep 27 05:35:55 printk something Sep 27 05:37:50 oh Sep 27 05:37:50 there Sep 27 05:38:01 console=tty0 instead of serial ttyS0 Sep 27 05:38:03 or whatever Sep 27 05:39:35 I have time I guess, I'll do this now Sep 27 05:46:16 03florian 07org.oe.dev * ra217dce7... 10/packages/gtk+/ (14 files in 2 dirs): gtk-2.6.10: Add GTK+ 2.6.10. Sep 27 05:53:30 console=tty0 no output Sep 27 05:53:33 i wonder what that means Sep 27 05:53:41 maybe ill wait for mickey to come around Sep 27 06:11:48 df001: CONFIG_CMDLINE is overwritten by OR Sep 27 06:11:53 s/OR/OE Sep 27 06:12:08 oH? Sep 27 06:12:25 overwritten by the .config file OE puts in? Or Something else? Sep 27 06:12:47 OE write own CONFIG_CMDLINE Sep 27 06:13:26 :\ Sep 27 06:13:29 Oh. Sep 27 06:13:49 Oh well Sep 27 06:13:52 ill deal with it later Sep 27 06:23:18 It's written in the .bb isn't it? Sep 27 06:52:56 http://handhelds.org/scap/port.19263.png Sep 27 06:53:01 lardman|work, df001: It is written in the .bb Sep 27 06:53:12 err.. wrong channel. sorry Sep 27 06:55:26 reenoo_: nice Sep 27 07:10:52 <[g2]> Ok.. I confess, I've been quite distracted and not doing tons of OE builds Sep 27 07:11:20 <[g2]> Today I kicked one off in Knoppix 4.02 with only adding monotone and it running fine so far Sep 27 07:11:56 <[g2]> did you guys reduce the required sw or did Knoppix 4.02 just include some extra packages ? Sep 27 07:12:02 <[g2]> anyone know ? Sep 27 07:12:09 probably include Sep 27 07:12:32 <[g2]> it was quite cool... Sep 27 07:13:35 <[g2]> s/was/is/ Sep 27 07:27:31 hi guys Sep 27 07:29:53 hi Bernardo Sep 27 07:30:21 hmm.. poodle change broke tosa kernel Sep 27 07:35:06 hey, any German people about who could do me a favor? Sep 27 07:36:18 mithro: looks like noone here alive from Germany currently Sep 27 07:37:39 * mithro ponders why mickey is still asleep :P Sep 27 07:38:55 maybe his PhD bored him so much that he felt asleep... Sep 27 07:39:15 he he Sep 27 07:39:17 oktoberfest is still a few days away, so that isn't the reason... Sep 27 07:39:29 preparation sleeping perhaps Sep 27 07:40:46 * zecke hopefully goes to the Oktoberfest Sep 27 07:40:58 * hrw|work dont Sep 27 07:41:26 I have to collect 150-200 eur and visit Berlin first Sep 27 07:41:36 * lardman|work would like to, but the end of PhD is also nigh Sep 27 07:44:35 hrw|work: ah Berlin is overrated Sep 27 07:45:07 but close to Poznań Sep 27 07:45:14 hrw|work: you should go to hamburg and the Linux Kongress instead! Sep 27 07:46:04 mithro: not with my salary Sep 27 07:49:19 get a better job ;) Sep 27 07:49:46 I'm trying to do that Sep 27 07:51:25 ~lart lineo and sharp for 2.4.18-crapix Sep 27 07:51:25 * ibot farts in lineo and sharp's general direction for 2.4.18-crapix Sep 27 07:51:54 total mess in code Sep 27 07:52:35 I 'love' to edit patches by hand Sep 27 07:53:16 looks like latest patch for oz-pxa 2.4.18 builds for tosa too Sep 27 07:53:23 NOTE: package openzaurus-pxa-2.4.18-rmk7-pxa3-embedix-r44: task do_build: completed Sep 27 07:53:25 yes... Sep 27 07:54:14 * Bernardo still hasn't had time to find why pcre is trying to link /usr/lib/libgcc_s.so... Sep 27 07:59:31 tosa kernel sent for tester Sep 27 08:09:53 hrw|work: I'm building libxine 1.1 now Sep 27 08:10:42 ok Sep 27 08:11:27 03hrw 07org.oe.dev * r89fba6d8... 10/packages/linux/ (2 files in 2 dirs): openzaurus-pxa: update APM patch to apply correct also for TOSA Sep 27 08:15:42 cu Sep 27 08:15:51 bye hrw|gone Sep 27 09:35:35 so I built xserver-kdrive_cvs with the disable-apm and no-serial patches, and it seems to be working well. Any other reasons why we need to stay with 20050207? Sep 27 10:20:00 kergoth`food: I'd like to upload daily tarballs to oe.handhelds.org, is that ok with you? Sep 27 10:37:19 anyone have any idea why this would happen? Sep 27 10:37:21 configure.in:106: error: possibly undefined macro: AC_DEFINE Sep 27 10:37:57 likely an m4 quoting problem in configure.in or acinclude.m4 Sep 27 10:38:06 generally that would indicate insufficient quoting Sep 27 10:49:52 koen: got a vnc client for OS X? Sep 27 10:50:48 zecke_: Chicken of the VNC Sep 27 10:55:13 why does vlc-gpe want to built qte-2.3.10? Sep 27 11:02:43 ah, because virtual/libsdl is set to libsdl-qpe Sep 27 11:02:54 what's the appropriate provider for gpe? Sep 27 11:07:01 looks like libsdl-x11 Sep 27 11:22:09 mreimer: I'd say we can upgrade xserver now the oz+fam releases have branched Sep 27 11:22:29 We can also make the changes to the machine files we talked about Sep 27 11:22:34 ah good Sep 27 11:22:58 I'll check with mickey|zzZZzz first and then do it Sep 27 11:23:08 it == machine files? Sep 27 11:23:17 that'll be good to clean up Sep 27 11:23:52 it == machine files? Sep 27 11:24:16 "then do it" meaning you'll make the changes to the machines files after checking with mickeyl? Sep 27 11:24:19 I mean the changes to sync up how 2.4 or 2.6 kernels are selected and tidy up the Zaurus includes Sep 27 11:24:23 gotcha Sep 27 11:24:48 yes, just check mickeyl is happy with the branch, then start breaking things ;-) Sep 27 11:25:20 that's what branches are for, right? a license to break stuff :-) Sep 27 12:05:49 hi Sep 27 12:06:00 hey hrw Sep 27 12:06:12 RP: did the kernel break yet? Sep 27 12:06:57 koen: No. I'm thinking of making the KERNEL_VERSION changes we mentioned though so it will happen soon enough :) Sep 27 12:07:19 linux26 and linux24, right? Sep 27 12:07:26 yup Sep 27 12:11:41 got a phone Sep 27 12:19:01 ok. now after branch I can look into hrw-packages and do some cleaning there Sep 27 12:31:51 libsdl-qpe/libsdl-x11 suxx Sep 27 12:34:44 it definately does Sep 27 12:35:07 hrw: but it sucks less than before Sep 27 12:35:52 I probably need to start from scratch again to get sdl linked with libsdl-qpe Sep 27 12:41:55 <_chronic> mt --db=/home/adamb/oe/oe.db pull monotone.vanille.de "org.openembedded.{dev,dreambox}" Sep 27 12:41:55 <_chronic> monotone: doing anonymous pull; use -kKEYNAME if you need authentication Sep 27 12:41:55 <_chronic> monotone: connecting to monotone.vanille.de Sep 27 12:41:56 <_chronic> monotone: network error: failed to connect: Connection refused Sep 27 12:42:30 _chronic: then use other mirror Sep 27 12:42:33 _chronic: try one of the backipt servers Sep 27 12:42:43 afk Sep 27 12:42:58 <_chronic> pardon ignorance, are mirrors documented? Sep 27 12:43:07 yes Sep 27 12:43:31 http://oe.handhelds.org/cgi-bin/moin.cgi/GettingStarted Sep 27 12:43:42 <_chronic> oops. ta. Sep 27 12:43:47 chouimat: even on the front page of oe.handhelds.org Sep 27 12:43:58 _chronic: and once on monotone.vanille.de Sep 27 12:44:08 koen: care to restart mickeyls monotone server? Sep 27 12:44:15 zecke_: already did :) Sep 27 12:44:51 <_chronic> (it wasn't doc'ed last time i looked :-) Sep 27 12:45:16 _chronic: hehe, but last time the server was not down?! Sep 27 12:45:40 <_chronic> zecke_: ah heh um no. Sep 27 12:45:57 yipee QtE 2.3.10 linked on OS X Sep 27 12:46:02 but back to my assignment... Sep 27 12:46:14 zecke_: tab complete it Sep 27 12:52:25 re Sep 27 12:53:16 wv Sep 27 12:55:29 if upstream does not release tarballs but keep all in CVS then if PV="0.2.0" then is it ok to set CVSDATE to next day after release? (no tags in cvs) Sep 27 12:56:07 I don't see another way Sep 27 12:56:09 03hrw 07org.oe.dev * r0260fec9... 10/packages/coreutils/ (3 files in 2 dirs): added coreutils 5.3.0 Sep 27 12:56:14 without hosting snapshots Sep 27 12:57:14 ~lart sf.net for their so called services Sep 27 12:57:14 * ibot stamps sf.net on the forehead with the official Troll marker for their so called services Sep 27 12:57:46 cvs.sf.net is .... or ..... but exist Sep 27 13:14:50 bitbake chooses openvpn-2.0rc1 instead of openvpn-2.0... What's the "correct" way to fix this? Sep 27 13:20:14 preferred_version Sep 27 13:20:49 There is a small issue with the 770 images bitbake generates. They don't work on the 770 as /sbin/init is a symbolic link. Sep 27 13:21:18 Does anyone object if I add a function to change this on nokia770 images as a post rootfs command? Sep 27 13:21:26 NAiL: or add DEFAULT_PREFERENCE = "-1" to the 2.0rc1 .bb Sep 27 13:22:11 03hrw 07org.oe.dev * r1e5ed75d... 10/packages/corsair/ (corsair_0.2.0.bb files/corsair-opie.patch): Sep 27 13:22:11 added corsair 0.2.0 - RSS aggregator for Qtopia Sep 27 13:22:11 Maintainer does not wanted to give his name ;( Sep 27 13:23:32 RP: didn't florian have a solution for that? Sep 27 13:23:53 koen|tv: We've just discussed it - his is hack sysvinit Sep 27 13:24:07 not acceptable to OE... Sep 27 13:24:12 ah, ok Sep 27 13:24:31 I recalled something about init and the 770 Sep 27 13:24:56 ok - see you guys Sep 27 13:25:08 hrw: bye Sep 27 13:26:11 * zecke_ wants his N770 Sep 27 13:26:40 zecke_: join the club Sep 27 13:27:02 planet.wait-4-n770.nl? Sep 27 13:27:14 +1 day without my gadget Sep 27 13:50:02 03nail 07org.oe.dev * r6c4836f3... 10/packages/openvpn/openvpn_2.0rc1.bb: openvpn: Set DEFAULT_PREFERENCE to -1 for the rc-version. Sep 27 14:01:57 RP: is your boss reachable? Sep 27 14:08:46 zecke_: he 'available' in jabber Sep 27 14:09:17 insert 'is' in above sentence Sep 27 14:09:44 mallum AT jabber.org? Sep 27 14:10:06 openendhand dot com Sep 27 14:10:43 openedhand* Sep 27 14:11:39 proll Sep 27 14:14:55 any autoconf savvy people around? Sep 27 14:14:57 http://pastebin.ca/24063 Sep 27 14:16:15 wrong version of autoconf? Sep 27 14:16:39 same version I've been using I think... Sep 27 14:16:41 JustinP: looks like normal m4 syntax to me Sep 27 14:16:53 2.59 Sep 27 14:17:29 cya tomorrow Sep 27 14:17:30 shouldn't that be -lts? Sep 27 14:17:50 no idea Sep 27 14:18:01 actually, that should be removed....since we don't want ts support I think... Sep 27 14:18:13 libts-0.0.so.0.1.1 Sep 27 14:18:30 I think that we do want tslib support Sep 27 14:18:41 otherwise mickeyl wouldn't have written it Sep 27 14:20:02 no, we don't need it for ecore-x11 Sep 27 14:20:18 he's using it for ecore-fb for his angstrom boot manager Sep 27 14:20:34 it was working fine without tslib before Sep 27 14:20:53 now the next AC_DEFINE call fails... Sep 27 14:20:56 you didn't mention ecore-x11 Sep 27 14:21:08 sorry, this is what this is for Sep 27 14:21:56 raster wrote the autofoo bits Sep 27 14:22:01 I guess he didn't test them Sep 27 14:22:22 the -ltslib is a strong clue Sep 27 14:22:28 as well as AM_CONDITIONAL and AC_PATH_GENERIC Sep 27 14:23:04 it's alive..... Sep 27 14:23:07 hey mickeyl Sep 27 14:23:10 morning Sep 27 14:23:15 argh Sep 27 14:23:28 this ecore release has a broken configure.in Sep 27 14:23:52 there's no need to disable tslib support in ecore-x11 Sep 27 14:24:06 ::shrug:: ok, I just knew it wasn't needed before Sep 27 14:24:07 since it's only implemented in ecore_fb anyway Sep 27 14:24:24 either way, though, there are still errors... Sep 27 14:24:36 yeah, raster most likely didn't test that Sep 27 14:24:45 first off with AC_DEFINE in the tslib section Sep 27 14:25:03 all calls to AC_DEFINE, AM_CONDITIONAL, and AC_PATH_GENERIC break Sep 27 14:25:42 the comment is also wrong :) Sep 27 14:25:46 AC_DEFINE(HAVE_TSLIB, 1, [Build Ecore_X Module]) Sep 27 14:25:55 that should read: using tslib in ecore_fb Sep 27 14:26:57 ::shrug:: again, that's not the problem Sep 27 14:27:02 sure Sep 27 14:27:05 i can't help you though Sep 27 14:27:15 my autofoo knowledge is 0 Sep 27 14:27:28 and although i like learning new things, i'm reluctant to learn autofoo Sep 27 14:28:21 can't we all use qmake? Sep 27 14:28:22 * mickeyl runs Sep 27 14:28:44 I wish the Linux world used something sane, yes.... Sep 27 14:28:48 mickeyl: SCONS Sep 27 14:28:50 Unix/Linux... Sep 27 14:28:59 mickeyl: it also links against the wrong name Sep 27 14:29:08 heh right Sep 27 14:29:11 -ltslib/-lts Sep 27 14:29:16 ~lart raster Sep 27 14:29:17 * ibot cuts raster into thin stripes Sep 27 14:29:51 damn...lots of changes between 013 and 015.... Sep 27 14:30:44 morning mickeyl Sep 27 14:30:47 hi RP Sep 27 14:31:26 mickeyl: Am I ok to start rearranging the zaurus files to use a KERNEL_VERSION override now we've branched? Sep 27 14:32:57 RP: yep. go ahead Sep 27 14:34:49 mickeyl: I altered your restart script a bit Sep 27 14:35:01 yeah, i've noticed that. Sep 27 14:35:12 koen|tv, mickeyl : I'd guess from mreimer's comments we can also move the xserver CVS date forward a bit as well? Sep 27 14:35:42 in .dev, yes Sep 27 14:35:55 it's actually a default_preference Sep 27 14:36:29 and you'll need to rearrange some xlibs (xextensions iirc) too Sep 27 14:37:47 mickeyl: what to we do with fam-083.conf and oz-354.conf in .dev? Sep 27 14:37:52 Ah, right. I was assuming it was a CVSDATE... Sep 27 14:38:00 * JustinP reverts the configure.in Sep 27 14:39:20 mickeyl: should we remove them? Sep 27 14:39:51 damn it.... Sep 27 14:44:45 03rpurdie 07org.oe.dev * ree063ba1... 10/ (classes/rootfs_ipk.bbclass conf/machine/nokia770.conf): rootfs_ipk: Add a remove_init_link function to the rootfs_ipk oeclass. This is needed for the nokia770 which won't boot with /sbin/init as a symbolic link (and I'm told other machines have this problem). Sep 27 14:46:37 hehe oeclass, old skool ;) Sep 27 14:46:46 hehe Sep 27 14:48:01 kergoth: would you have problems with adding me to proj_oe so I can upload OE tarballs and the snapshot db to oe.hh.org? Sep 27 14:50:34 ~lart autotools Sep 27 14:50:35 * ibot duct-tapes autotools to the floor and drools on him Sep 27 14:50:48 ~lart developers for releasing with broken configure.in Sep 27 14:50:49 * ibot follow's developers with a gauntlet and ... scratch ... HUMILIATION for releasing with broken configure.in Sep 27 14:51:32 mickeyl: have you tried an e-image? Sep 27 14:52:05 mickeyl: it *seems* that something changed and broke the e-image... Sep 27 14:52:23 I thought it was the non-csl compiler possibly but that didn't seem to make a difference Sep 27 14:55:42 ~seen drow Sep 27 14:55:52 koen: i haven't seen 'drow' Sep 27 14:56:15 JustinP: no time yet Sep 27 14:56:30 mickeyl: do you have any knowledgable contacts for EABI? Sep 27 14:56:47 unfortunately not Sep 27 14:57:20 23:36 < koen> mickeyl: what to we do with fam-083.conf and oz-354.conf in .dev? Sep 27 14:57:22 koen: i think we leave the .conf files here Sep 27 14:57:46 or do you think we should remove 'em until we have released Sep 27 14:57:48 ? Sep 27 14:57:56 remove them Sep 27 14:57:57 either way is ok for me Sep 27 14:58:00 k Sep 27 14:58:07 they are unmaintained in .dev Sep 27 14:58:37 shall I remove them? Sep 27 14:59:23 yep Sep 27 14:59:29 .dev is rather useless without them Sep 27 15:00:29 given that stuff blows up randomly without preferred versions Sep 27 15:00:42 s/blows/tends to blow/ Sep 27 15:01:09 *shrug* might as well add the preferred stuff to the unversioned distro Sep 27 15:01:14 OE won't even start to build Sep 27 15:01:27 since most distro.conf have a bbfatal nowadays Sep 27 15:01:56 which is good for the branch but maybe should be removed for the .dev again Sep 27 15:01:58 mickeyl: could you add the branch to the cia script? Sep 27 15:02:04 yeah, will do Sep 27 15:02:23 it will still blow up, so the warning should stay Sep 27 15:02:52 and certain things (preferrec compiler) should be moved to the unversioned conf IMHO Sep 27 15:03:01 JustinP: why? Sep 27 15:03:10 so that it won't blow up Sep 27 15:03:23 if you're removing the versioned ones then we need something in the unversioned to make it work... Sep 27 15:03:27 right...? Sep 27 15:03:29 no Sep 27 15:03:39 03koen 07org.oe.oz354fam083 * r4f334b99... 10/conf/distro/familiar-0.8.3.conf: -----> OpenZaurus and Familiar have branched for release <----- Sep 27 15:03:43 03koen 07org.oe.oz354fam083 * raca49b63... 10/packages/linux/ (2 files in 2 dirs): packages/linux/openzaurus-pxa-2.4.18-rmk7-pxa3-embedix20031107: backport ebb1f3b5b2d477b7e6feb7628032eb03920b318d Sep 27 15:03:46 if you want a working distro, use the branch Sep 27 15:04:07 and people using dev can't build an image.....this is what I'm hearing... Sep 27 15:04:15 or really much of anything Sep 27 15:04:31 koen: I'm not exactly sure I like that concept Sep 27 15:04:44 koen: chances are you'll end up with a fork Sep 27 15:04:51 fork? Sep 27 15:05:02 fork Sep 27 15:05:29 oz354 and fam083 will be effectively unmaintained in .dev for the next few weeks Sep 27 15:05:37 and after that they get removed anyway Sep 27 15:06:24 without proper branching/merging support that's going to be a pain Sep 27 15:06:36 s/support/support in monotone/ Sep 27 15:06:53 branches are working fine for me **** ENDING LOGGING AT Tue Sep 27 15:07:06 2005 **** BEGIN LOGGING AT Fri Sep 30 02:36:11 2005 Sep 30 02:44:48 hrw|work: His git tree is the way he publishes his work Sep 30 02:45:05 hrw|work: We should really try and make some patches from it to add to OE Sep 30 02:45:38 hrw|work: The battery code is horrible... Sep 30 02:46:03 RP: would adding a linux-pawel_git.bb be a solution? Sep 30 02:46:26 since we have a shiney new git fetcher :) Sep 30 02:46:26 koen: If we could add my git patches to bitbake :) Sep 30 02:46:57 I really need to get them rediffed and pester zecke__ :) Sep 30 02:48:18 and we should pester kergoth to get 1.3.x bitbake released Sep 30 02:53:30 nah, better have him enable the release bits on zecke__'s account Sep 30 02:57:34 yep - that would be ideal Sep 30 02:58:09 or both Sep 30 02:58:20 ~lart samba Sep 30 02:58:21 * ibot nukes samba with a single large nuke Sep 30 02:59:15 someone know how to make KDE show names in samba network? now it show IP only.. and they are dynamic so I have to guess does PPOSP is .114 or .107 today.. Sep 30 02:59:29 hrw|work: install and setup winbind **** ENDING LOGGING AT Fri Sep 30 02:59:57 2005