**** BEGIN LOGGING AT Mon Oct 08 02:59:57 2007 Oct 08 05:54:04 good morning Oct 08 05:54:15 good afternoon Oct 08 05:54:36 nah, its morning :-) Oct 08 05:54:41 too early Oct 08 07:31:56 good morning all Oct 08 07:32:06 gm Mr Idiot Oct 08 07:32:13 have you turned yourself in yet? Oct 08 07:32:49 yes Oct 08 07:32:56 "not a case for the german police" Oct 08 07:33:06 Crofton: http://dominion.kabel.utwente.nl/koen/cms/oedem-day-2 Oct 08 07:33:13 heh Oct 08 07:33:39 that is much better then my photo Oct 08 07:34:53 oh, so you're all at the office yet. I just barely got out of bed Oct 08 07:34:59 well Oct 08 07:35:03 well, I'm at the office :) Oct 08 07:35:06 I have wireless in my hotel Oct 08 07:35:13 I bet holger is still snoring Oct 08 07:35:45 koen: ah, thanks for the reminder. I'll bring my "UFL to piece of wire" adaptor Oct 08 07:36:04 * * OE Bug 3133 has been RESOLVED (FIXED) by Oct 08 07:36:06 * * mono-1.2.5.1-r3-do_compile Oct 08 07:36:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3133 Oct 08 07:36:41 good morning everyone Oct 08 07:37:58 hi koen Oct 08 07:38:44 hey djay-il Oct 08 07:39:13 I've been away for a long while Oct 08 07:39:21 need to get back Oct 08 07:39:59 hm..does anyone happen to know of a way to use apt-get to install into a chroot environment? Oct 08 07:40:28 ok, I'm leaving now, will need about an hour to join you Oct 08 07:43:20 gilligan_: like debootstrap? Oct 08 07:47:18 * koen eats his mandelschnitt Oct 08 07:47:52 koen: problem is I have to work with a suse based system.. there is apt-rpm, but debootstrap for suse.. hmmm... ;] Oct 08 07:49:26 koen: wish I could do this on debian.. but well, it's some sort of legacy-support which I have to do at work Oct 08 07:54:52 does anyone uses Tasks from pimlico? Oct 08 08:09:59 morning Oct 08 08:26:25 koen, let me in Oct 08 08:27:40 hmm, http://bugs.openembedded.org/show_bug.cgi?id=3089 Oct 08 08:28:32 that bug also exist for me, h2200 and looks like that it was caused by (removed?) patch, 4.2.1-r7 was built fine Oct 08 08:33:40 Darth_Wader: please give a reasonable amount of information, I am not a mind reader Oct 08 08:33:56 in progress Oct 08 08:35:56 error about foprtran not compiling? Oct 08 08:36:21 http://pastebin.ca/729744 Oct 08 08:36:27 nope Oct 08 08:36:40 fortran error gone with upgrade to opensuse 10.3... Oct 08 08:38:25 hm, looks like i`ve misread gcc version in #3089... but anyway Oct 08 08:41:17 thats a new never seen before error Oct 08 08:43:04 * * OE Bug 3089 has been RESOLVED (FIXED) by Oct 08 08:43:06 * * gcc-cross-4.1.2-r9-do_compile Oct 08 08:43:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3089 Oct 08 08:54:42 XorA: http://dominion.kabel.utwente.nl/koen/duvel.mpg Oct 08 09:10:41 mickey|oedem07: salut! Oct 08 09:11:14 heya Oct 08 09:11:37 shall we wait for some more guys until we start with the "official" program? or do we already have a plan? Oct 08 09:12:51 mickey|oedem07: We should perhaps at least discuss the agenda. I think there is a draft on the wiki? Oct 08 09:14:11 http://www.openembedded.org/wiki/OEDEM2007 Oct 08 09:14:14 right. i have a spiced version of that Oct 08 09:15:39 zecke pasted "hack to make RP scream" at http://paste.lisp.org/display/48821 Oct 08 09:21:34 hi mickey :) Oct 08 09:21:37 http://www.openembedded.org/oedem2007.minutes Oct 08 09:26:40 hi all Oct 08 09:27:07 hey, you just said your nick was florian ;) Oct 08 09:27:26 ;) Oct 08 09:28:35 ;D Oct 08 09:29:49 hey Liam Oct 08 09:29:53 hi lrg Oct 08 09:30:22 ~fish lrg Oct 08 09:30:23 * ibot slaps lrg around with a large trout Oct 08 09:30:40 hey koen, XorA, florian Oct 08 09:34:05 shhh, mickeyl is talking ... Oct 08 09:38:34 * RP screams Oct 08 09:38:42 * XorA is defeaned Oct 08 09:38:43 DEPENDS = "libnfnetlink libtool-cross (>=1.5.24)" Oct 08 09:38:50 does this work ? Oct 08 09:39:01 no Oct 08 09:39:08 * koen stops being elitist Oct 08 09:39:10 Crofton: It should be accepted by the parser but doesn't actually do anything Oct 08 09:39:14 (yet) Oct 08 09:39:14 ok Oct 08 09:39:18 03hrw 07org.oe.dev * r1d14c551... 10/ (1 conf/machine/cm-x270.conf): cm-x270: added iwmmxt into archs and features Oct 08 09:39:19 do we plan for = to work? Oct 08 09:39:23 03xora 07org.oe.dev * r55671170... 10/ (1 packages/openmoko2/libmokojournal2_svn.bb): openmoko2/libmokojournal2_svn.bb : gtk-doc patch Oct 08 09:39:26 libtool-cross-1.5.24 works but is a gross hack Oct 08 09:39:28 03koen 07org.oe.dev * r9cb28749... 10/ (3 files in 3 dirs): linux 2.6.21: replace gumstix 18bpp patch with pxa-overlay patch from RP Oct 08 09:39:50 libtool is a gross hack Oct 08 09:39:57 and it does not work for efika :( Oct 08 09:40:17 but I need it for libnetfilter_queue Oct 08 09:40:58 Crofton: I mean DEPENDS = libtool-cross-1.5.24 Oct 08 09:41:07 Crofton: I don't see why we couldn't have = work Oct 08 09:42:05 mtn: updating along branch '' Oct 08 09:42:06 mtn: misuse: cannot determine branch for update Oct 08 09:42:13 how would I go about fixing that? Oct 08 09:42:19 mtn merge Oct 08 09:43:25 hi vivijim Oct 08 09:43:48 florian: gutten morgen Oct 08 09:43:51 http://www.openembedded.org/bonsai/view/scm/6/ Oct 08 09:44:57 http://www.openembedded.org/bonsai/view/scm/6/ Oct 08 09:56:08 RP maybe <= should work :) Oct 08 09:56:26 the specific case is recent gnuplot picks up c++ and older does not Oct 08 09:56:55 so problem solution was to use older version to keep stdc++ lib off the image Oct 08 09:57:20 03koen 07org.oe.dev * r766b6ff7... 10/ (1 packages/dropbear/dropbear.inc): dropbear: familiar: DIE DIE DIE Oct 08 09:58:38 lol Oct 08 10:00:46 /wi 42 Oct 08 10:00:56 err Oct 08 10:03:17 koen: "distro_debug or familiar" thing? Oct 08 10:04:06 hrw: yes Oct 08 10:19:47 And link to http://xkcd.com/303/ in the policy :-) Oct 08 10:20:37 someone has to finish the minutes for the discussion as you were not able to pause them while I left the room ;) Oct 08 10:20:57 alphaone: and http://xkcd.com/169/ Oct 08 10:21:05 alphaone: hehe, I was thinking of the same yesterday Oct 08 10:22:02 zecke|OEDEM2007: just insert Oct 08 10:22:15 XorA: I don't know the syntax for drupal Oct 08 10:24:52 when you use OE for more than one arch, do you build everything within the same build directory, or do you use one dir per arch instead ? Oct 08 10:25:45 nud: the same dir Oct 08 10:25:58 multimachine.bbclass takes cares of most details Oct 08 10:26:19 and if you use angstrom as distro it will take care of even more details :) Oct 08 10:26:25 and what about local.conf ? Oct 08 10:27:02 just change MACHINE Oct 08 10:28:47 mmh ok Oct 08 10:29:46 is it ok to have more than one distro at the same time ? Oct 08 10:30:01 not in one builddir Oct 08 10:32:13 XorA: mamona's debootstrap script: http://dev.openbossa.org/svn/mamona/repository/trunk/debootstrap/mamona Oct 08 10:33:49 vivijil: sweet Oct 08 10:36:41 can we get koen C lessons so he has to code these ideas Oct 08 10:36:41 mickeyl: http://www.packagekit.org/wiki/index.php/Main_Page Oct 08 10:37:06 zecke|OEDEM2007: Please note that url in the minutes ;-) Oct 08 10:37:07 fork ipkg to opkg Oct 08 10:37:36 upkg Oct 08 10:37:41 u = ugly Oct 08 10:37:56 * florian offers some free C lessons Oct 08 10:38:41 We could remove all this text file parsing and use sqlite then too :) Oct 08 10:38:44 RP: multimedia? Oct 08 10:44:24 mickeyl: OE & community minutes from last year: http://rafb.net/p/QJX0gb97.html Oct 08 10:55:48 RP: linux-libc-headers doesnt work for multimachine (x86/x86_64 meaning) I think simple fix is to copy include/asm* not just the include/asm include/asm-generic that we do now, can you think of downfalls? Oct 08 10:57:00 03stefan 07org.oe.dev * rac8a5e02... 10/ (1 packages/hal/ohm_git.bb): Oct 08 10:57:00 ohm_git.bb: Touch gtk-doc.make and disable xauth to let it build again. Oct 08 10:57:00 (--without-xauth should be fine on single user systems) Oct 08 10:57:04 03jluebbe 07org.oe.dev * rbed6bec9... 10/ (3 files in 3 dirs): linux-openmoko-devel: change CONFIG_LEDS_S3C24XX from module to compiled-in. Oct 08 11:08:57 03jluebbe 07org.oe.dev * r1deac361... 10/ (1 packages/linux/linux-openmoko-devel/squashfs-2.6.23.patch): linux-openmoko-devel: add missing squashfs port to 2.6.23 Oct 08 11:09:50 more mamona stuff: http://labs.vivi.eng.br/blog/ Oct 08 11:12:38 what does this mean? Oct 08 11:12:39 mtn: multiple update candidates: Oct 08 11:12:39 mtn: 1deac361ca5b23f9321f50e6bb4143327fefbdd7 jluebbe@openembedded.org 2007-10-08T11:02:38 Oct 08 11:12:43 mtn: ac8a5e02cb4b87019cb729953ff361387cf30945 stefan@openembedded.org 2007-10-08T10:20:31 Oct 08 11:12:51 which one should I choose? Oct 08 11:13:01 ynezz: any of the too Oct 08 11:13:09 ok Oct 08 11:13:14 ynezz: or wait some time until it is merged Oct 08 11:13:39 ah, now I understand Oct 08 11:14:19 just mtn merge Oct 08 11:20:20 hi Oct 08 11:20:53 hi woglinde Oct 08 11:22:41 he florian Oct 08 11:22:50 not time for oedem this year? Oct 08 11:23:03 woglinde: he is at oedem Oct 08 11:23:08 ah Oct 08 11:23:26 I can only attend tomorrow Oct 08 11:23:35 *sigh* Oct 08 11:23:57 happens Oct 08 11:23:59 hi hrw btw. Oct 08 11:24:18 * florian is happy that he finally made it Oct 08 11:24:35 hehe is mickeyl arrived? Oct 08 11:24:44 mickeyl, is here Oct 08 11:24:46 http://people.openmoko.org/henryk/mamona.pdf Oct 08 11:24:58 he crofton Oct 08 11:25:08 hi Oct 08 11:25:13 You don't have permission to access /henryk/mamona.pdf on this server. Oct 08 11:25:17 again Oct 08 11:25:37 You don't have permission to access /henryk/mamona.pdf on this server Oct 08 11:25:53 hm Oct 08 11:25:59 woglinde: yes you do. talk sense into your cache :) Oct 08 11:26:15 I have no cache Oct 08 11:26:25 butnow it works Oct 08 11:27:35 hm I hope tha I get the toolchain for qtopia 4.3 ready than I can show some jave apps on the greenphone Oct 08 11:29:46 woglinde: bring the GP, I need to review some apps :} Oct 08 11:30:04 http://en.wikipedia.org/wiki/Athens_Wireless_Metropolitan_Network Oct 08 11:30:24 zecke I have flashed the 4.3 preview Oct 08 11:33:30 http://www.flickr.com/photos/koenkooi/1514546310/ Oct 08 11:34:51 Thanks for http://www.openembedded.org/oedem2007.minutes, it's nice to see what's ongoing even if I was not able to attend. Oct 08 11:36:38 noglitch: I'm looking at the wiki now Oct 08 11:42:45 ~lart glibc Oct 08 11:42:45 * ibot whacks glibc with the cluebat Oct 08 11:46:14 koen: yes... Oct 08 11:49:29 koen: BTW, I compile a snapshot (x11 image) last week and it behaves good on 9263 Oct 08 11:49:40 noglitch: great! Oct 08 11:52:42 noglitch: when I get back home I'll update uboot and install an image to flash to do some more testing Oct 08 11:53:45 koen: You framed me! Oct 08 11:54:57 koen: ok, do not hesitate to tell me what is wrong or can be improved Oct 08 11:56:24 koen: I will also follow closely what you are doing in OEDEM : it interests me a lot ! Oct 08 11:57:05 I just mentioned the at91.com wiki :) Oct 08 11:57:17 koen : | ../sysdeps/powerpc/powerpc32/fpu/s_isnan.c:1: warning: -msoft-float and -mlong-double-128 not supported Oct 08 11:57:18 | In file included from ../sysdeps/powerpc/powerpc32/fpu/s_isnan.c:1: Oct 08 11:57:18 | ../sysdeps/powerpc/fpu/s_isnan.c:59: warning: conflicting types for built-in function 'isnanf' Oct 08 11:57:18 | ../sysdeps/powerpc/fpu/s_isnan.c: In function '__isnan': Oct 08 11:57:18 | ../sysdeps/powerpc/fpu/s_isnan.c:45: error: impossible constraint in 'asm' Oct 08 11:57:18 | ../sysdeps/powerpc/fpu/s_isnan.c:48: error: impossible constraint in 'asm' Oct 08 11:57:30 steliosk|OEDEM: yes, that one Oct 08 11:57:32 check the first line Oct 08 11:57:43 heh Oct 08 11:58:36 damn fpu stuff Oct 08 12:00:56 koen: when it will be published, it will appear at linux4sam.org (no more this preview location...) Oct 08 12:02:36 koen: (I must go to a meeting : good afternoon) Oct 08 12:02:46 ok Oct 08 12:02:49 good luck :) Oct 08 12:24:09 YEAH g++ -emit-llvm -S -o foo.o /home/zecke/foo.cxx :) Oct 08 12:24:31 zecke|OEDEM2007: yay! Oct 08 12:26:02 koen: I'm a bit mean and copy three glibc files into the build dir... Oct 08 12:35:04 * * OE Bug 3109 has been RESOLVED (FIXED) by philippedeswert(AT)scarlet.be Oct 08 12:35:06 * * bitbake gpe-image fails on gpe-mini-browser Oct 08 12:35:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3109 Oct 08 12:35:48 Philippe: great! Oct 08 12:38:30 florian: I just haven't seen the mtn commit passing by... Oct 08 12:39:27 at least I am slightly active again :-) Oct 08 12:40:07 gah, gcc-cross doesnt know to look in lib64 for libgcc Oct 08 12:40:41 03philippe 07org.oe.dev * rdea5e351... 10/ (4 files in 2 dirs): disapproval of revision '7b5746140aa364c513fc814efde39c90df31234e' Oct 08 12:40:46 03philippe 07org.oe.dev * rdd45113e... 10/ (7 files in 2 dirs): Oct 08 12:40:46 gtk-webcore : update default preferences Oct 08 12:40:46 * default preference update so that the correct version gets build for gpe-image Oct 08 12:40:46 * fixes bug #3109 Oct 08 12:48:10 I made few changed to unified updater.sh Oct 08 12:48:31 a bit cleaned and now it also should not flash files bigger then mtd partition Oct 08 13:05:02 hrw: building cm-x270 and testing now ... Oct 08 13:14:28 Hi! I am running Angstrom kernel 2.6.22 on an Akita. I can switch gpio 93 on and off. My Problem is, when I suspend, the gpio ALWAYS assumes high state. I need it low. Any idea? Oct 08 13:38:32 cbrake: hi! Oct 08 13:39:24 djay-il: hello Oct 08 13:42:14 how are you? Oct 08 13:42:40 djay-il: doing well -- pretty much the same routine -- and you? Oct 08 13:46:29 cbrake: I'm ok. getting married in 4 days Oct 08 13:46:43 djay-il hehe gratz Oct 08 13:47:04 ni march I will be married for two years Oct 08 13:47:17 but we cant mess with crofton Oct 08 13:47:19 *g* Oct 08 13:47:29 djay-il: wow, best wishes :-) Oct 08 13:47:53 * cbrake is happily married for 7 years :-) Oct 08 13:49:58 woglinde: thanks Oct 08 13:50:06 cbrake: thanks Oct 08 13:50:51 cbrake ;) Oct 08 13:51:10 the puzzeld 7 year Oct 08 13:51:14 hehe watch put Oct 08 13:51:18 aeh out Oct 08 13:52:23 woglinde: :-) Oct 08 13:56:25 cbrake: hi cliff Oct 08 13:56:36 hrw: hello Oct 08 13:59:19 * hrw -> afk now Oct 08 14:02:14 Crofton: any more progress with the wifi? Oct 08 14:02:46 the blue light blinks Oct 08 14:03:03 we are not having any luck raising an access point Oct 08 14:03:14 I think I need to try at home with the open access point Oct 08 14:03:55 I've only tried with an open AP, and with that it has never given me any problems Oct 08 14:04:48 at least you're getting somewhere with it - there are lots of folks having issues with it on the list Oct 08 14:04:58 crofton I can bring a ap tomorrow Oct 08 14:05:02 an Oct 08 14:05:08 if you want Oct 08 14:06:41 We have ideas, but not too much time atm Oct 08 14:08:21 scruggs: are you using wifistix or the new combo board? Oct 08 14:09:14 sakoman: wifistix-cf Oct 08 14:09:46 scruggs: straight wifistix w/ connex for me Oct 08 14:10:09 seems to be working very reliably for me now w/ wpa-supplicant Oct 08 14:10:53 mine is working really well too - haven't tried with supplicant yet, but its coming Oct 08 14:11:29 we suspect our board can't transmit Oct 08 14:11:43 ��p000�9-0~~~9999999999999999999999999999999999999999999999 Oct 08 14:12:01 scruggs: if size matters make sure to steal the updated wpa-supplicant from my collection. it uses a newer rev that has a tiny tls library built in Oct 08 14:12:26 sakoman: http://www.flickr.com/photos/koenkooi/1513324894/ Oct 08 14:12:30 will save you several hundred K over external libgnutls Oct 08 14:12:59 koen: there was a post to the list today with a patch for a verdex not transmitting - was a DMA issue, but I think it was the smc wired ethernet, not wifi Oct 08 14:13:19 sakoman: right now if you have a mod I use it :) Oct 08 14:13:23 koen: I saw that! I have this insane urge to try to hook up a paper clip antenna Oct 08 14:13:29 :-) Oct 08 14:13:40 I working postcards atm .... Oct 08 14:13:45 sakoman: where is this updated wpa supplicant Oct 08 14:15:25 likewise: http://www.sakoman.net/oe/org.sakoman.dev/packages/wpa-supplicant/ Oct 08 14:17:23 likewise: it is just a version bump + revised defconfig to enable buil-in tls Oct 08 14:17:34 sakoman: 0.6cvs? Oct 08 14:18:49 hrw: not sure what you are asking! Oct 08 14:19:39 wpa-supplicant Oct 08 14:19:56 o.. 0.5.8 Oct 08 14:20:30 0.6.0 has been released Oct 08 14:20:39 yes, it's just a quick hack to the existing recipe to try out the new built-in tls lib Oct 08 14:21:23 koen: that was fast! last I looked it was 0.5.9 and I saw some complaints about bugs on that rev Oct 08 14:21:34 hence my use of 0.5.8 Oct 08 14:21:34 koen: as development not stable Oct 08 14:22:27 koen: yes, what hrw says! I'm having plenty of issues with dev versions :-) Oct 08 14:23:22 koen: I keep rolling back till I find one that works. sort of like how you feel about udev 0.92 :-) Oct 08 14:24:14 Hmmm Oct 08 14:24:50 imlib doesn't distribute imlib.m4 in staging Oct 08 14:25:06 Even though the file can be found in the imlib-dev packages. Oct 08 14:25:38 sakoman: so give information how goes testing of wpa-supplicant with internal TLS implementation Oct 08 14:25:49 This makes an old version of gqview fail, which isn't bad, but I was wondering if this would be as serious as to demand a bug report Oct 08 14:26:38 koen : the gcc 4.1.1 patches for softfloat fail to apply as is on 4.1.2. Need to look what actually fails Oct 08 14:27:11 morning Oct 08 14:28:46 hrw: haven't done exhaustive testing, but have used it to connect with Apple Airport and D-Link access points Oct 08 14:28:53 steliosk|OEDEM: do you know how to make gcc use /lib /usr/lib on 64bit architectures? Oct 08 14:29:28 sakoman: wep? wpa1? wpa2? psk? tkip? Oct 08 14:30:13 hrw: will do more testing after I get the rest of my sw moved over to oe. right now just tried open and wep. Oct 08 14:30:32 ok Oct 08 14:30:47 cbrake: what are your plans about 2.6.23 for cm-x270? Oct 08 14:31:06 hrw: so basically just confirmed that things don't blow up, not that the builtin tls works :-) Oct 08 14:32:15 hrw: I'm already using 2.6.23-rc on projects, but was waiting for 23 to be released before pushing Oct 08 14:32:54 03koen 07org.oe.dev * r3ce2953f... 10/ (1 packages/linux/linux.inc): linux.inc: add RPSRC Oct 08 14:37:06 cbrake: in OH we considering adding 2.6.23-rc for now (until 2.6.23 release) Oct 08 14:42:57 stop being elitist :-D Oct 08 14:44:04 koen: stop being elitist Oct 08 14:44:19 ahhaha Oct 08 14:44:39 hi guy Oct 08 14:44:41 s Oct 08 14:47:42 what is the mplayer command I need to run Oct 08 14:48:53 hrw: that should be fine. The patches apply fairly cleanly to the latest kernels. Oct 08 14:50:14 hrw: 2.6.23-rc has high res timer support in mainline for PXA, so that is nice Oct 08 14:50:47 hrw: I had to apply the rt patch as well to get the system to perform well enough for me, but it seems to working Oct 08 14:50:56 cbrake: base needed two small changes and it should builds Oct 08 14:52:27 Crofton: mplayer -vo pxa -ac none Oct 08 14:52:36 of -vo fbdev or -vo fbdev2 Oct 08 14:52:46 -nosound is better then -ac none iirc Oct 08 14:53:50 I'm not an mplayer expert :) Oct 08 14:55:52 * XorA gets confused by .dpath files Oct 08 14:55:55 .dpatch Oct 08 14:56:14 debian patches? Oct 08 14:56:17 yeah Oct 08 14:56:35 they can be easily canged to normal patches iirc Oct 08 14:56:49 yeah, but this one creates a script I dont understand Oct 08 14:56:52 with dpatch-edit? Oct 08 14:58:09 ah actually I see what it does Oct 08 15:00:30 roh: oe_runmake Oct 08 15:00:38 how many times can you write on a NAND flash drive ? Oct 08 15:01:14 nud: 3 Oct 08 15:01:29 Once. Oct 08 15:01:36 (until you erase it, of course) Oct 08 15:01:50 hrw: don't be so elitist Oct 08 15:01:54 I meant an order of magnitude Oct 08 15:01:56 koen: ;D Oct 08 15:02:04 to fix a fscked Makefile org.openembedded.dev/packages/vnc/tightvnc_1.3dev7.bb Oct 08 15:02:06 nud: it depends on chip etc Oct 08 15:02:08 nud: if you're asking about lifetime, it depends on the flash itself, but usually 1000 - 10,000 times Oct 08 15:02:32 nud: youu can kill nand chip during one day Oct 08 15:02:36 heck, developping on a flash drive is bad for it, then Oct 08 15:03:04 I mean, writing images again an again on a device Oct 08 15:03:16 nud: but it's complicated because some manufacturers will guarantee part of the flash for 10x longer than the rest -- so you place your critical data there, and manage bad flash blocks in the rest. Oct 08 15:03:42 nud: there's a difference between flash drive and chip Oct 08 15:03:52 mickeyl: could we run through http://dominion.kabel.utwente.nl/koen/talks/ELC2007.pdf on the beamer Oct 08 15:04:11 the typical drive already has wear-leveling built in, so it lasts many many times longer than the chip without wear leveling. Oct 08 15:04:28 I'm interested in, for instance, the flash in an embedded device like a N770, FIC or, in my case, cm-x270 Oct 08 15:05:31 Those should be bare chips, the load-levelling needs to be done by the software (jffs2 for example). Oct 08 15:05:39 nud: 10k writes. but how are you going to check? Oct 08 15:06:17 nud: mounting using "-o noatime" is very useful if you mount r-w. Oct 08 15:06:51 sakoman: thanks for the wpa .bb. Will probably test it on powerpc w/ madwifi soon Oct 08 15:07:10 I was just wondering, because currently I'm erasing and writing new images on my board something like 10 times a day Oct 08 15:07:45 * mwester has always used the rule of thumb of one flash per day on average for the non-jffs2 parts of the flash, so 1,000 is about 3 years of constant reflashing. Oct 08 15:08:50 as koen doesnt code, he can write instead :-D Oct 08 15:08:59 pfffff Oct 08 15:09:14 shit this thing sound so much fun Oct 08 15:09:27 so is it better to ipkg install or to update the whole image for the device lifetime ? Oct 08 15:09:37 or it doesn't matter ? Oct 08 15:10:21 mickeyl: Shoragan: http://www.openembedded.org/screencasts/org.openembedded.htm Oct 08 15:10:28 nud: ipkg install will only install a small amount of data and use jffs2 for wear leveling Oct 08 15:11:29 ok so ipkg is better for the common case Oct 08 15:11:29 nud: but in practice, you probably won't ever be able to wear out flash if you write the entire image 10 times a day Oct 08 15:11:57 nud: 100,000 writes spec, that is 20+ years Oct 08 15:11:57 wear out ? Oct 08 15:12:04 ah ok Oct 08 15:12:12 can somebody at oedem tell zecke to submit his foss.in paper soon? ;) Oct 08 15:12:13 so "don't care" Oct 08 15:12:16 MLC is sadly 10.000 writes :( Oct 08 15:12:51 do you rebuild regularly a fresh new OE build or do you stay with the same build for months ? Oct 08 15:13:11 nud: misc - I often use one build for months Oct 08 15:13:21 nud: both, depends Oct 08 15:13:51 LaF0rge: #openmoko - zecke|OEDEM2007 answered Oct 08 15:14:00 ok Oct 08 15:14:55 LaF0rge: oh and the CfP restarted Oct 08 15:16:08 could a crash in ld-linux.so (with nothing else in the stacktrace, as if it crashed while reading the library file) be related to something else than a borked toolchain ? Oct 08 15:24:29 nud: the NAND on the cm-x270 modules I have is SLC (100,000K) erase cycles, but some of the newer modules have flash with 16KB erase blocks -- no idea what those are Oct 08 15:25:01 100.000, not 100.000k :) Oct 08 15:25:06 why did the compulab-pxa270 get renamed cm-x270 ? Oct 08 15:25:42 nud: because this is how board is named Oct 08 15:26:03 nud: compulab-pxa270 does not tell does it is cm-x270 or em-x270 or sbc-x270 Oct 08 15:26:16 ok Oct 08 15:27:11 http://people.openmoko.org/henryk/oedem20071008-1.jpg Oct 08 15:31:54 henryk: can you please remove the environment partition image from your home Oct 08 15:32:28 henryk: it's really not a good idea if people download your environment into their device (which will very likely have bad blocks at different location than your device) Oct 08 15:32:44 LaF0rge: already did Oct 08 15:32:48 henryk: ah, thanks. Oct 08 15:33:12 LaF0rge: sorry for the confusion Oct 08 15:34:01 henryk: no problem. I just wanted to halt it before it spreads further ;) Oct 08 15:35:20 LaF0rge: I was discussing this with roh on the way to lunch. a better fix would be to have a more complete uboot default env, because right now it's a catch-22: when your environment is messed up you need the uboot console to recreate it, but you need a full environment in order to have the menu option to set the console to usb. Oct 08 15:35:27 I'll try to come up with a patch Oct 08 15:37:16 koen: right :-) Oct 08 15:38:22 henryk: yes, we can have usbtty as the default, no problem with that Oct 08 15:38:40 or at least make it usbtty once you press AUX Oct 08 15:39:36 LaF0rge: that would be good. because most device owners will not have a serial console per default Oct 08 16:07:55 okay, I will just do releases ;) Oct 08 16:08:15 don't be so elitist Oct 08 16:09:35 koen: this is why I want more people to use OE Oct 08 16:11:59 they hijacked that point... Oct 08 16:26:23 * koen pulls mallum's beard Oct 08 16:27:29 koen: and it is very pullable atm Oct 08 16:27:42 :D Oct 08 16:28:06 koen: in berlin ? Oct 08 16:28:12 yes Oct 08 16:28:27 mallum: http://people.openmoko.org/henryk/oedem20071008-1.jpg Oct 08 16:29:37 hi mallum Oct 08 16:30:10 koen: that picture lacks stefan ;) Oct 08 16:31:03 koen|away: ah excellent - RP's photo avoiding skills are now even tuned to group photo's Oct 08 16:31:04 hi mallum :} Oct 08 16:31:05 * zecke|OEDEM2007 hides Oct 08 16:31:12 hey zecke|OEDEM2007 Oct 08 16:31:20 zecke|OEDEM2007: are you still coming to UK btw ? Oct 08 16:31:23 hey hrw Oct 08 16:31:57 mallum: I still want to, but I have ni date in mind... I got stuck at my family in the netherland :( Oct 08 16:32:31 zecke|OEDEM2007: okeys, well just let me know. Your very welcome to visit some time Oct 08 16:32:49 WTF architecture is x64_64 Oct 08 16:33:34 mallum: awesome, I love o-hand and I didn't even know you hired pippin Oct 08 16:33:42 XorA: ? Oct 08 16:33:44 * XorA blames koen Oct 08 16:33:54 in insane.bbclass x64_64 Oct 08 16:33:57 * koen is being elitist Oct 08 16:35:57