**** BEGIN LOGGING AT Thu Aug 17 02:59:56 2006 Aug 17 04:10:32 03lenehan 07org.oe.dev * r762d15ce... 10/ (1 site/i386-linux site/i486-linux site/i586-linux): Aug 17 04:10:32 site/i{3,4,5}-linux: Add entries for ipsec-tools so it will build. Aug 17 04:10:32 i686-linux was already ok, but i{3,4,5}86-linux were missing the appropriate Aug 17 04:10:32 entries. Note that ipsec-tools does not currently build with gcc 4.x. Aug 17 07:34:58 morning Aug 17 07:37:43 yo Aug 17 07:39:41 poushag: look not in tags but in branches (for bitbake) Aug 17 07:43:48 morning Aug 17 07:43:58 mickey|thesis: x86_64 Im your man :-) Aug 17 07:44:15 morning hrw|work XorA Aug 17 07:47:52 Morning Aug 17 07:48:21 morning sirfred Aug 17 07:48:55 XorA: Hello. I had some problems yesterday trying to build mplayer. Aug 17 07:50:16 XorA: Strange, in the do_install stage it said: Aug 17 07:50:16 strip: Unable to recognise the format of the input file `/home/mteira/oeroot/build/tmp/work/armv5te-linux/mplayer-1.0pre8-r2/image/usr/bin/mplayer' Aug 17 07:50:27 XorA: As if it tried to use the native strip to strip the binary? Aug 17 07:54:25 sirfred: wierd, maybe try adding install to DEPENDS line Aug 17 07:54:47 sirfred: its possibly one of the apps that needs our hecked install Aug 17 07:54:51 install-native I mean Aug 17 07:56:20 03lenehan 07org.oe.dev * rea4038ae... 10/ (7 files in 2 dirs): Aug 17 07:56:20 site/*: Updates for slrn. Added required entries for slrn to Aug 17 07:56:21 i{3,4,5,6}-linux and sh4-linux-uclibc. Remove some unused entries from Aug 17 07:56:21 sh4-linux. Aug 17 07:59:48 morning all Aug 17 07:59:52 hey RP Aug 17 08:00:00 morning all Aug 17 08:00:06 hey ade|desk Aug 17 08:01:03 hi RP Aug 17 08:04:22 * hrw|work want do13 Aug 17 08:05:06 RP: Morning Aug 17 08:05:13 hi RP Aug 17 08:18:28 03hrw 07org.oe.oz354x * r59b12544... 10/ (1 packages/meta/gpe-image.bb): gpe-image: resolve gpe-task-connectivity problem - close #1293 Aug 17 08:24:38 hey Aug 17 08:24:50 morning all Aug 17 08:25:50 hi dirk Aug 17 08:26:04 hey Marcin Aug 17 08:26:20 do13_: can you look at my recent changesets in .oz354x? Aug 17 08:26:51 hrw|work: which one? Aug 17 08:27:35 do13_: 1. disabled usbctl in /etc/network/interfaces 2. set ohci-hcd to be (un)loaded on suspend/resume Aug 17 08:27:47 do13_: I also want to make ohci-hcd autoloaded Aug 17 08:27:58 I'm just trying to set up ipkg on my target system so that I can upgrade packages easily from my own feed urls... Aug 17 08:28:16 but when running ipkg update, I get the following error: wget: unrecognized option `--passive-ftp' Aug 17 08:28:38 it's trying to use busyboxes wget, which seems to be too basic for ipkg Aug 17 08:28:46 anyone know a way round this? Aug 17 08:28:51 ipkg install wget Aug 17 08:29:16 OK, so I need to override busybox's wget with the real one Aug 17 08:29:26 tkp: yeah, f thats the problem Aug 17 08:29:46 I heard rumours busybox 1.2.1 wget works, but havent verified Aug 17 08:30:03 oh ok. I'm on 1.2.0 atm Aug 17 08:30:58 do13_: what do you think about it? Aug 17 08:31:29 hi woglinde Aug 17 08:31:32 hi hrw Aug 17 08:32:03 what's the url of the repo? I thought it was available for browsing at http://www.openembedded.org/repo/ but it seems I'm getting 'Access forbidden!' Aug 17 08:32:28 http://www.openembedded.org/filebrowser is another way Aug 17 08:32:44 hrw|work: thanks Aug 17 08:33:37 tkp: there is no diff between 1.2.1 and 1.2.0 of busybox for wget Aug 17 08:34:45 ade|desk: ah, ok. well then wget does not work (or, it works, but still doesn't have enough options for use with ipkg) Aug 17 08:35:23 which options are missing ? Aug 17 08:35:30 --passive-ftp Aug 17 08:36:41 the command it tried to issue was `wget --passive-ftp -q -P http://www......` Aug 17 08:46:21 hi Aug 17 08:46:48 hi cobelius Aug 17 08:47:03 if a software requires nptl and I'm using glibc 2.3.2 how can I configure that the glibc is compiled with nptl support? Aug 17 08:47:10 hi woglinde Aug 17 08:58:39 hrw|work: sorry, I was called away Aug 17 08:58:56 hrw|work: sounds fine Aug 17 08:59:14 cobelius I think there is no ntpl patch for this old glibc use a newer one Aug 17 09:11:33 are there any existing helpers/classes/examples that show how you could automate the maintanance of an ipkg feeed uri? Aug 17 09:12:00 Basically, I want a dev feed, which is basicaly a live index of my tmp/deploy/ipk directory Aug 17 09:14:30 tkp: I just rsync my tmp/deploy/ipk to my webserver Aug 17 09:14:49 XorA: yeah, that's kinda what I was thinking too Aug 17 09:15:12 The problem is that I have to use this version because other software that should run on the system is compiled for this version. The Debian libc have this support. I tried to avtivate it with GLIBC_ADDONS = "nptl" and GLIBC_EXTRA_OECONF = "--with-tls" Aug 17 09:15:23 but I got errors when I try to compile it Aug 17 09:15:28 thought there may be a bb classs which would update an index after every build Aug 17 09:15:45 cobelius than include the debian-patches Aug 17 09:16:50 I've been away for a couple of weeks, and would like to re-pull and update my checkout Aug 17 09:17:08 However, my checkout contains quite a few local modificatios Aug 17 09:17:29 What's the deal with updating a checkout that contains local modifications? Aug 17 09:17:55 I usually use SVN for my version control Aug 17 09:18:25 tkp read the montone phrasebook Aug 17 09:18:31 on the oe website Aug 17 09:18:47 montone will warn you if it can not merge something Aug 17 09:19:04 tkp: there s, I cant remeber which one, I just call ipkg-make-index directly Aug 17 09:21:06 tkp: bitbake package-index Aug 17 09:21:28 hrw|work: cool, thanks Aug 17 09:31:56 so... just to confirm... I need to mtn pull & mtn merge rather than mtn update? Aug 17 09:32:18 oh no... forget that Aug 17 09:32:24 read the wrong bit Aug 17 09:35:10 morning all Aug 17 09:36:22 hi Liam Aug 17 09:37:59 RP: is the poodle hp resume related to it being interptreted as a keypress ? Aug 17 09:38:38 brb Aug 17 09:38:44 hi lrg Aug 17 09:38:45 hi rp Aug 17 09:39:10 lrg: I wasn't aware of a poodle hp resume issue. That sounds possible though Aug 17 09:40:49 http://emertxe.com/training_modules.html someone could give me his idea about this ? Aug 17 09:42:57 hey Liam Aug 17 09:43:14 foo Aug 17 09:43:47 hi kergoth Aug 17 09:44:03 hey Aug 17 09:46:08 RP: it's a new bug [1325]. I just noticed it about 5 mins ago Aug 17 09:46:11 poodle make noise when touching touchscreen Aug 17 09:46:48 hrw|work: sounds like more board noise :-/ Aug 17 09:47:03 hrw|work: are you playing audio when this happens ? Aug 17 09:47:14 lrg: no Aug 17 09:47:20 hrw|work: Is the mic on? Aug 17 09:47:31 lrg: I think I heard this effect faintly on corgi when we were doing some testing last week Aug 17 09:47:37 RP: nothing changed in alsamixer Aug 17 09:47:56 hrw|work: poodle is still not complete. I have afew issues to fix Aug 17 09:48:13 RP: I flashed 3542rc2 (unofficial) opie and touched screen few times Aug 17 09:48:21 RP: first touch is ok, second make noise Aug 17 09:48:33 hrw|work: touchscreen clicks weren't turned on? Aug 17 09:48:58 RP: rather no - will check evening Aug 17 09:49:28 hrw|work: There every other time would be the speaker enable/disable bug which is known about but not debugged Aug 17 09:49:51 ok Aug 17 09:50:15 anyway I feel that pxa250 is fscking slow Aug 17 09:51:12 so, whhen doing a mtn update, it's telling me "mtn: misuse: path _MTN/tmp/1 already exists" Aug 17 09:51:52 I tried moving the entire _MTN directory out of the way, and then it complains about a million and one missing files Aug 17 09:52:07 i have seen this before... Aug 17 09:52:19 ~lart monotone Aug 17 09:52:20 * ibot decapitates monotone conan the destroyer style Aug 17 09:53:00 ~lart monotone? Aug 17 09:53:00 * ibot --purges monotone? Aug 17 09:54:22 anyone know what I'm supposed to do about this? Aug 17 09:54:41 I never have senn this Aug 17 09:54:51 rm _MTN/tmp -r Aug 17 09:54:53 maybee I doing something wrong Aug 17 09:55:21 hrw|work: but If I remove it, it thhinks all the files are missing and prompts me to revert them Aug 17 09:55:35 which I don't want to do as some of them have local modifications Aug 17 09:56:02 ah... _MTN/tmp... Aug 17 09:57:40 grrr "mtn: misuse: path 'packages/busybox/busybox-1.2.0/mnci' already exists, cannot create"... :( Aug 17 09:57:52 * tkp hits monotone with a large stick Aug 17 10:00:39 when I got this (few times) I usually made new checkout Aug 17 10:01:13 damn it. I'm going in circles here... I remove _MTN/tmp and run mtn update, then it complains that a file exists and can't be created, so I remove the file. Then it complains about loads of missing files, so I mtn revert tthem, then mtn update, and then back to "mtn: misuse: path _MTN/tmp/1 already exists"!! Aug 17 10:01:35 so if I make a new checkout, how am I ment to get all my mods in there? Aug 17 10:01:41 diff/patch I guess? Aug 17 10:02:50 woglinde : Hi !. Did the build with debug and no strip and...asterisk does not seg faults.... Aug 17 10:03:53 ifaistos lol Aug 17 10:03:59 so strip strip something out Aug 17 10:04:31 tkp make a new pull and checkout and convert your changes over Aug 17 10:04:31 woglinde : bute where ? gcc ? uclibc ? asterisk ? Aug 17 10:04:48 ifaistos If I would know I would say it to you Aug 17 10:04:57 woglinde: hmm, but that means remembering every single file I have changed! Aug 17 10:05:19 tkp why? diff ist your friend and always bee Aug 17 10:06:13 woglinde: right, but diff is going to show up all changes in oe over the last 3 weeks! (not just my local changes) Aug 17 10:06:33 tkp hm right Aug 17 10:14:11 re katossi Aug 17 10:25:51 http://www.nat.org/apology-999x720.jpg Aug 17 10:25:57 haha, I need those as well Aug 17 10:27:35 sweet Aug 17 10:28:10 time to compile inkscape Aug 17 10:30:39 ;)) Aug 17 10:30:47 Humm, I have in my local.conf: Aug 17 10:30:48 CVS_TARBALL_STASH = "http://ewi546.ewi.utwente.nl/mirror/www.openzaurus.org/official/unstable/3.5.4/sources" Aug 17 10:30:51 * kergoth wonders which machine he put his mtn key on Aug 17 10:31:04 someone use(d) SonyEriccson D750i cellphone? Aug 17 10:31:08 kergoth: that is the wrong question Aug 17 10:31:12 But, bitbake keeps trying into : Aug 17 10:31:12 NOTE: fetch http://www.oesources.org/source/current/glib-2.12.1.tar.bz2 Aug 17 10:31:18 And fails Aug 17 10:31:21 kergoth: your question must be: Did I place it on an Andrew FilSystem Aug 17 10:31:29 What I'm doing wrong? Aug 17 10:31:33 sirfred: check does it is not overwritten somewhere Aug 17 10:31:53 hrw|work: In some bb file? Aug 17 10:32:00 sirfred: oesources.org is a general mirror as well Aug 17 10:32:05 sirfred: see base.bbclass at the bottom Aug 17 10:32:54 It has these two lines: Aug 17 10:32:55 ftp://.*/.*/ http://www.oesources.org/source/current/ Aug 17 10:32:55 http://.*/.*/ http://www.oesources.org/source/current/ Aug 17 10:33:08 this means as last resort, try oesources.org Aug 17 10:33:13 sirfred: which distro you use? Aug 17 10:33:22 hrw|work: openzaurus-unstable Aug 17 10:35:49 Where is that variable CVS_TARBALL_STASH supposed to be used? Aug 17 10:35:59 sirfred: oz-unstable already has ewi and oesources Aug 17 10:36:38 hrw|work: Where do you see that? Aug 17 10:37:17 conf/distro/openzaurus-unstable.conf ? Aug 17 10:37:19 sirfred: conf/distro/ Aug 17 10:37:22 yes Aug 17 10:38:53 I'm going to comment that line in my local.conf, and try again. Aug 17 10:42:07 hi renoo Aug 17 10:42:19 hey woglinde Aug 17 10:42:25 morning all Aug 17 10:43:08 03florian 07org.oe.oz354x * r71a9ca9a... 10/ (1 packages/fbreader/fbreader_0.7.3d.bb): fbreader: Depends on bzip2 Aug 17 10:46:04 reenoo: hi Aug 17 10:46:10 reenoo: Familiar 0.8.4 today? Aug 17 10:48:18 03clarson 07org.oe.dev * rf876e12a... 10/ (1 packages/tslib/tslib_1.0.bb): Aug 17 10:48:18 tslib: Aug 17 10:48:18 * Add a bitbake recipe for tslib 1.0, which is newly tagged on the recently Aug 17 10:48:18 populated tslib subversion repository at berlios. Aug 17 10:49:08 (yes, it took long enough) Aug 17 10:49:27 kergoth: HEY!!!! Aug 17 10:49:46 kergoth: now it is time to report the emberassing bug to force you to release a 1.0.1 Aug 17 10:49:49 used tailor to pull the history over from the cvs :) Aug 17 10:49:53 hehe, naturally Aug 17 10:50:19 kergoth: why did you decide against cvs2svn? Aug 17 10:50:20 We already did that for bitbake 1.6 (sorry zecke) :-/ Aug 17 10:50:35 kergoth: This is really good to see though :) Aug 17 10:50:39 RP: I'm used to that Aug 17 10:50:44 zecke: no particular reason, just had already played with tailor in the past Aug 17 10:51:13 hrw|work: doesn't seem likely right now. seems like someone is having fun with his botnets again Aug 17 10:51:17 kergoth: wow. Aug 17 10:51:19 reenoo: ;( Aug 17 10:51:20 kergoth: I didn't like tailor too much but I think one can not fail converting from a flat format to a flat format Aug 17 10:51:50 well, the mechanism for determining which cvs deltas belong to a given "revision" is important Aug 17 10:52:14 kergoth: tailor is trying that? Aug 17 10:52:58 it looks at the changelog of a given file delta and uses that + a given timeframe between the checkin of the file deltas to match them up from a single cvs commit Aug 17 10:53:01 afaik Aug 17 10:53:10 i havent been in that code in a while, but back when i looked at it, thats how it worked Aug 17 10:53:45 * kergoth gets caffeine Aug 17 10:54:40 hi kergoth Aug 17 10:54:58 why tslib 1.0 has SRCDATE inside set to 'now'? Aug 17 10:55:06 hrw|work: do we have free flash in the oz gpe images? Aug 17 10:55:41 hrw|work: because its still svn, and afaict from the fetcher, an srcdate set to anything else (i.e. by a distro .conf) will result in a -r on the svn commandline, which we dont want, we want the tip of the tag Aug 17 10:55:53 so i forced the matter Aug 17 10:56:13 if you've got a better idea, by all means implement it :) Aug 17 10:56:20 when I try to compile glibc 2.3.2 with nplt support i get the follwing error message: http://pastebin.de/11048 Aug 17 11:00:27 kergoth: That sounds like a bug in the fetcher Aug 17 11:01:01 kergoth: If a tag is specificed, it should drop the srcdate Aug 17 11:01:39 its svn, there is no "tag". its just a path. the artifical module= thing annoys me Aug 17 11:01:45 imho the svn fetcher should be scrapped and redone Aug 17 11:01:53 if nobody else gets to it, i will Aug 17 11:02:42 cobleius urgs old gcc old glibc Aug 17 11:02:45 kergoth: Having it use a cached checkout directory like cvs would also be good Aug 17 11:03:10 hm Aug 17 11:03:15 why he leaves Aug 17 11:04:10 florian_kc: I already cut them a lot for 3.5.4.2-rc Aug 17 11:05:08 hrw|work: ah, what did you change? Aug 17 11:05:47 hrw|work: in the devices with enough free flash it might be good to replace gpe-plucker by fbreader Aug 17 11:08:30 lrg: do you happen to know what the differences are between the WM8721 and the WM8725? Aug 17 11:08:36 hmm my powersuply must be broken... Aug 17 11:09:14 zecke: suck Aug 17 11:09:48 pb_: I finally have a Karo board (www.karo-electronics.de) and searched for a power supply the last hour... Aug 17 11:09:54 florian_kc: 1M will be enough? Aug 17 11:09:58 pb_: and now something does not work... Aug 17 11:10:12 zecke: oh dear Aug 17 11:10:27 pb_: this design doesn't even have a LED Aug 17 11:10:30 hrw|work: yes, i expect about 650k overhead Aug 17 11:10:48 ah, and we need to fix gpe-screenshot... Aug 17 11:17:17 pb_ : hi ! Did the build with debug + no strip and guess what... asterisk does not seg faults... Aug 17 11:20:05 Ifaistos: was that a debug and no strip from scratch or just for the asterisk bb ? Aug 17 11:20:28 Ifaistos: ah, that is strange Aug 17 11:20:30 ade|desk : from scratch Aug 17 11:21:27 Ifaistos: can you try it without debug and with stripping but then inhibit stripping for asterisk only ? Aug 17 11:22:03 speaking of debug... just before I went on holiday, this new .debug package thing was introduces to contain unstripped versions of packages (or so I understood it)... But there was a little problem which caused .debug packages to be installed onto the target system regardless of weather or not they were asked for... Aug 17 11:22:09 did that get resolved? Aug 17 11:24:52 ade|desk : will try it and post the result. The problem though seem to come from a call to glob() i don't *think* is an asterisk problem Aug 17 11:29:32 Ifaistos: what cpu ? Aug 17 11:30:04 ade|desk : host is an amd64 target is i386 Aug 17 11:30:52 ah ok, was remembering there was some uclibc stuff about glob() not that long ago but was for mips Aug 17 11:30:53 ade|desk : the same config works without a problem on ppc Aug 17 11:34:38 My problem persists. I've tried with an interactive bitbake: Aug 17 11:34:54 BB>> getvar CVS_TARBALL_STASH Aug 17 11:34:54 http://ewi546.ewi.utwente.nl/mirror/www.openzaurus.org/official/unstable/3.5.4/sources/ http://www.oesources.org/source/current/ Aug 17 11:35:04 BB>> build opie-image Aug 17 11:35:12 ... Aug 17 11:35:21 NOTE: fetch http://www.oesources.org/source/current/glib-2.12.1.tar.bz2 Aug 17 11:35:29 But it doesn't try with ewi Aug 17 11:36:00 Have to leave now. Aug 17 11:36:28 sirfred: bitbake 1.4.3+ needed Aug 17 11:38:45 On another mater a number of site files are missing especially for uclibc targets. anyone has a quick 'howto' ? Aug 17 11:55:47 03florian 07org.oe.oz354x * rbb69d0b3... 10/ (3 files in 3 dirs): gpe-conf: install suid to be able actually to configure things. Aug 17 11:56:29 florian_kc: pulling, will add to upgrades Aug 17 11:57:13 hrw|work: ah yes, that is a quite useful one :-) Aug 17 11:58:09 the original idea was to let the distribution choose (e.g. familiar folks were not amused havine suid binaries at all) but for gpe-conf its essential. Aug 17 11:58:12 florian_kc: I would add do_install_append() for that suid thing ;) Aug 17 11:58:41 hrw|work: that would be another solution Aug 17 12:11:41 florian_kc: not for stuff that introduces security holes you could drive a truck through Aug 17 12:13:18 do13_: ping Aug 17 12:15:50 hrw|work: pong Aug 17 12:30:14 do13_: what do you think about switching tosa to 2.6 in .dev? Aug 17 12:30:26 hrw do it Aug 17 12:30:33 thats what dev are for Aug 17 12:30:37 * XorA likes the ROAD S101 phone Aug 17 12:30:42 hrw|work: ack Aug 17 12:30:45 woglinde: Dirk is tosa maintainer Aug 17 12:31:11 hrw oh okay Aug 17 12:31:13 sorry Aug 17 12:31:29 xora yes zecke has its finger in it Aug 17 12:31:30 XorA: it is awesome (the phone) Aug 17 12:31:37 woglinde: s/has/had/ Aug 17 12:31:42 his Aug 17 12:31:45 dammit Aug 17 12:31:49 when will I learn Aug 17 12:31:57 zecke: I saw it on The Register Aug 17 12:32:11 XorA: what I can say, regardles of the Photoshop+Outlook pictures, this hardware is real and working Aug 17 12:32:53 zecke: looks very qte in those photos Aug 17 12:32:59 zecke: is the road phone available yet? Aug 17 12:33:01 +Konquerer Aug 17 12:33:20 XorA: ah. on www.pro-linux.de they still show the photoshop+outlook version Aug 17 12:33:31 do13_: I haven't been in their offices since fosdem Aug 17 12:33:31 zecke: http://www.reghardware.co.uk/2006/08/17/road_handypc_targets_nokia/ Aug 17 12:33:39 do13_: but they did a press release Aug 17 12:34:05 oh they have some new icons Aug 17 12:34:11 and yes it is looking like the real thing Aug 17 12:36:15 zecke: the display on the picture says: 30. Sept 2005 Aug 17 12:37:32 they should get better fonts Aug 17 12:37:43 I have just seen the icons Aug 17 12:37:48 I think I like this style Aug 17 12:38:05 it is a bit 80's but still it looks nice Aug 17 12:38:17 damn my phoneapplet still shows that bad antenna sign... Aug 17 12:38:46 do13_: on site shots have few days only - show mails from 2006.08.09 Aug 17 12:38:52 * XorA will buy that phone if he can afford it Aug 17 12:39:06 http://www.road-gmbh.de/en/produkte/pro_s101_11.html Aug 17 12:39:42 but in one place there is pxa264 in other its pxa263 Aug 17 12:39:45 fine, also the calendar shows 29. Jul 2006 Aug 17 12:41:40 xora it will 1000 euro Aug 17 12:41:53 woglinde: where did they print the price? Aug 17 12:42:05 zecke hehe Aug 17 12:42:29 reenoo: but is security of suid utils on a pda all that important ? Aug 17 12:42:40 zecke at the first annouces as I remeber correctly Aug 17 12:43:24 so two Qtopia phones announced for september Aug 17 12:43:48 if it is still true the ROAD will be reflashable as well Aug 17 12:44:30 hm where manucfatred the green phone Aug 17 12:44:35 I bet trolltech not Aug 17 12:45:06 woglinde: What I assume (guess) is that they licensed the design+softwarestack from one of their chinese customers Aug 17 12:52:31 Arrr Aug 17 12:55:55 there is no way to override c++ header search path ones compiler has been hardcoded with --with-gxx-include-dir= ? Aug 17 12:56:30 ie similar to what gcc -isystem does for normal c Aug 17 12:57:35 nchip: right :( Aug 17 12:57:41 nchip: -isystem does not do the trick Aug 17 12:57:57 nchip: you can add the CXX stuff manually but then you need to play a lot of silly games Aug 17 12:58:04 nchip: not even -B is helping here Aug 17 12:58:18 nchip: you could take a look at cross-dev and their 'fix'embedded' paths Aug 17 13:02:19 however, if the (cross)toolchain has been compiled with -sysroot , it seems to work Aug 17 13:05:37 bitbake 1.6.1 (r570) has one NASTY bug in bbshell Aug 17 13:05:47 NOTE: package binutils-cross-2.17-r0: task do_compile: failed Aug 17 13:05:47 ERROR: TaskFailed event exception, aborting Aug 17 13:05:47 NOTE: package binutils-cross-2.17-r0: task do_clean: started Aug 17 13:06:00 why it do do_clean() after error? Aug 17 13:06:10 hrw|work: cause you have tinderclient? Aug 17 13:06:23 zecke: I have tinderclient Aug 17 13:06:37 zecke: but that is not an answer for me Aug 17 13:06:39 it is an 'autoevict' feature for builds Aug 17 13:06:45 it has an... Aug 17 13:06:57 when a Package fails, it removes the package Aug 17 13:06:57 zecke: so I should first do build and then enable tinderclient? Aug 17 13:07:02 to recompile it on the next run Aug 17 13:07:21 now if I have tinderclient then I cannot check what is wrong Aug 17 13:07:34 in other way then by disabling tinderclient Aug 17 13:07:41 hrw|work: Yes/No. You can check the log Aug 17 13:07:54 hrw|work: but this 'feature' is for auto builders Aug 17 13:08:09 hrw|work: where you want to rerun fetch, configure, patch without bumping the PR of the package Aug 17 13:08:13 e.g. as one dependency got fixed Aug 17 13:08:24 ok Aug 17 13:08:27 hrw|work: we could add another TINDER_ option to disable auto evicting Aug 17 13:08:29 03hrw 07org.oe.dev * r1a4a3b2b... 10/ (1 conf/machine/tosa.conf): tosa: linux 2.6 by default Aug 17 13:08:48 zecke: TINDER_AUTOBUILD = 1/0 Aug 17 13:09:08 e.g. Aug 17 13:09:17 okay Aug 17 13:09:28 I'm hacking tinderclient.bbclass ATM anyway Aug 17 13:09:35 (working around -k on autobuilds) Aug 17 13:11:33 nchip: how does sysroot change anything? Aug 17 13:11:42 NOTE: preferred version 2.4 of glibc not available Aug 17 13:12:03 this message is funny (after binutils-cross) - it will build glibc 2.4 anyway Aug 17 13:13:35 zecke: can you add some kind of 'bitbake tell-tinderclient-that-I-really-stopped-build'? Aug 17 13:14:46 hrw|work: I don't know how :( Aug 17 13:15:08 ;( Aug 17 13:15:19 I will have to think about it Aug 17 13:17:33 ERROR: Nothing provides dependency tell-tinderclient-that-I-really-stopped-build Aug 17 13:17:37 it worked Aug 17 13:17:51 http://ewi546.ewi.utwente.nl/tinderbox/showlog.pl?machine_id=93&logfile=20060817152906.log Aug 17 13:18:42 it worked as you started a new build Aug 17 13:18:44 and it failed Aug 17 13:19:04 bitbake quilt-native would work as well Aug 17 13:19:18 ok Aug 17 13:19:26 just wanted to stop it Aug 17 13:24:37 hmm now gettext is running amock Aug 17 13:33:13 hrw|work: TINDER_AUTOBUILD is on by default Aug 17 13:33:59 anyone seen this before: /lib/libc.so.6: undefined reference to `_dl_out_of_memory@GLIBC_PRIVATE' ? Aug 17 13:34:10 I'm getting it when trying to build mplayer Aug 17 13:34:49 it used to build ok before I updated my checkout and rebuilt everything from scratch Aug 17 13:34:55 zecke: ok Aug 17 13:35:46 tkp: do you use a different version of binutils now? Aug 17 13:36:07 RP: we need an option for BitBake to write a PREFERRED_PROVIDER,_VERSION file Aug 17 13:36:27 zeckie: no (2.15.94.0.1) Aug 17 13:36:45 tkp: also using /lib/libc.so.6 sounds wrong Aug 17 13:37:07 indeed it does Aug 17 13:45:36 zecke, doesnt bitbake -s >> file work for you? Aug 17 13:45:55 bitbake -s | less works Aug 17 13:46:24 hms Aug 17 13:46:32 should read everything Aug 17 13:48:29 hrw|work: i built an "extended" gpe image for h3900 - with gpe-mini-browser, gaim, gpe-file-manager and fbreader Aug 17 13:48:35 22mb Aug 17 13:49:31 florian_kc: it has 32M rootfs? Aug 17 13:50:04 xora i can confirm that busybox 1.2.1 wget works with passive ftp Aug 17 13:50:20 hrw|work: yes Aug 17 13:50:37 florian_kc: poodle, tosa, c7x0 has 20-23M rootfs Aug 17 13:50:53 i also tried 1.2.0 and had the same wget problem Aug 17 13:50:56 hrw|work: how is the flash partitioned? Aug 17 13:51:21 that tkp was reporting Aug 17 13:51:28 florian_kc: 64M: 7M (rescue), 1.2M (kernel), 20-23M (rootfs), 32M (home) Aug 17 13:52:06 poushag: 1.2.1 does work? cool Aug 17 13:52:20 hrw|work: oh, what is "rescue" for? Aug 17 13:52:20 tkp i have a feed of my oe built ipks - are you still having problems setting it up? Aug 17 13:52:31 florian_kc: flashing for example Aug 17 13:52:42 florian_kc: it has 2.4 kernel and minimal rootfs Aug 17 13:52:59 poushag: I got distracted updating my checkout and am now having some problems with other packages Aug 17 13:53:16 I'll get on to that again shortly (hopefully!) Aug 17 13:53:29 hrw|work: i see... you can't have the system for flashing on an external card, right? Aug 17 13:54:05 florian_kc: pxa-zauruses use rescue system to run flash process Aug 17 13:54:34 hrw|work: the h3900 is flashed from the bootloader so there is no need to waste expensive flash for this. Aug 17 13:55:56 florian_kc: if we will switch to u-boot then we can reorganize flash Aug 17 13:56:17 hm we should port uboot to simpad to Aug 17 13:56:51 hrw|work: sounds good, does it support suspend/resume now? Aug 17 13:57:23 florian_kc: dont know but rather yes - sashz (pdaX developer) ported u-boot to c7x0 and cxx00 zauruses Aug 17 13:57:40 woglinde: what is uboot? Aug 17 13:58:03 woglinde: a new bootloader? Aug 17 13:58:03 mr_nice_slacker: bootloader Aug 17 13:58:31 hrw|work: ah, that guy th hh.org guys had to kick from the SoC team... Aug 17 13:58:43 florian_kc: same Aug 17 14:00:07 hrw|work: well, we'd need to check... if it works that would be great but in fact i could imagine that they maybe could live without this Aug 17 14:00:16 mr_nice please dont use familar when building images try openzaurus Aug 17 14:00:58 woglinde: if familiar is not maintained in .dev then why it is not removed? Aug 17 14:01:21 florian_kc: maybe angstrom could use it - oz will not Aug 17 14:01:43 woglinde: openzaurus? - well ok - i do what ever you say :) Aug 17 14:01:47 hrw but took some poitions from the angstroem file for my local.conf Aug 17 14:01:52 will now head home Aug 17 14:02:03 woglinde: but why don't use familiar? Aug 17 14:02:13 fix up my libpng patches will see if gcc-4.1.1 and glibc2.4 works on simpad Aug 17 14:02:41 hrw|work: ok Aug 17 14:03:43 mr_nice look at hrw Aug 17 14:03:52 its not maintained in .dev Aug 17 14:04:17 woglinde: a ok thats the reason why it breaks that often Aug 17 14:04:50 thx Aug 17 14:05:24 in which tree is it maintained? Aug 17 14:05:32 *sigh* Aug 17 14:05:48 erik i belive is going to bring it up after 0.8.4 ships Aug 17 14:05:49 none Aug 17 14:06:00 the GIT hh.o one Aug 17 14:06:02 becaue familar dont use oe anymore Aug 17 14:06:12 yes it does Aug 17 14:06:18 emte hm Aug 17 14:06:43 poushag: cool Aug 17 14:06:44 they forked bitbake and the tree to avoid issues so they could finish 0.8.4 Aug 17 14:06:59 no Aug 17 14:07:03 no? Aug 17 14:07:05 no Aug 17 14:07:06 hm and now the switching? Aug 17 14:07:18 so whats going on here Aug 17 14:07:20 *g* Aug 17 14:08:57 so whats right now? Aug 17 14:08:57 reenoo is the man with the master plan Aug 17 14:09:43 dunno now, i was only following the hh.o wiki stuff with the gui interface :) Aug 17 14:10:59 h, Aug 17 14:11:01 http://familiar.handhelds.org/build-setup.html Aug 17 14:11:31 yup, that one Aug 17 14:11:44 only that zenity is broken on my debian testing... Aug 17 14:11:50 and I do not have a gui anyway :) Aug 17 14:11:53 zenity is pretty neat when you want to add more stuff Aug 17 14:12:00 zecke, i use it on etch ... Aug 17 14:12:16 emte: it choked here some month ago Aug 17 14:12:30 emte: do you know if the zenity syntax is compatible to dialog/kdialog? Aug 17 14:12:42 hmm... so I removed all my local modifications and reverted to the stock bb file (mplayer) but am still getting | /lib/libc.so.6: undefined reference to `_dl_out_of_memory@GLIBC_PRIVATE' Aug 17 14:12:53 is there some known problems with binutils 2.15.94.0.1 then? Aug 17 14:12:55 zecke, dunno, i try to avoid KDE Aug 17 14:13:18 emte: I just wanted to point out that it is possible to write a dialog GUI and having a compatible syntax... Aug 17 14:13:37 emte: this means you create a variable $DIALOG in your script and it works on GUI and TUI Aug 17 14:13:40 ... Aug 17 14:13:40 wouldnt that just mean using the console ? :P Aug 17 14:13:52 yeah i know what you mean Aug 17 14:14:09 tkp: well. Try to pass the turing test you are exposed to Aug 17 14:14:17 i was only playing with zenity for more verbose kernal stuff Aug 17 14:14:23 tkp: you opened the window and closed it but it is still using /lib/libc.so.6 Aug 17 14:14:35 tkp: and opening and closing the window doesn't change that? Aug 17 14:14:41 so that i could watch the build and errors etc Aug 17 14:14:46 tkp: might be wise to use the source? Aug 17 14:14:54 turing test? Aug 17 14:15:08 tkp: take a look at the command line just above the error Aug 17 14:15:24 tkp: open a gcc manpage and search for linker/library path Aug 17 14:15:36 tkp: and see if there is a 'bogus' entry? Aug 17 14:15:43 like for /lib? Aug 17 14:15:45 zeckie: there is no reference to /lib in there Aug 17 14:15:56 check which application is used to link it Aug 17 14:16:06 although I do ee -L/usr/lib Aug 17 14:16:18 boomer Aug 17 14:16:25 boomer? Aug 17 14:16:27 baby boomer? Aug 17 14:16:29 zecke, i am using zenity 2.14.2-1 Aug 17 14:17:08 emte: my buildbox is gone, so I can not check the version of zenity Aug 17 14:17:10 till later Aug 17 14:17:38 emte: but using zenity was the reason we do not have autobuilds for familiar Aug 17 14:17:49 'We' == I did not setup Aug 17 14:18:36 zecke: well a with toolchain built in a sysroot, it appears c++ respects -isystem Aug 17 14:18:47 the zenity stuff is pretty superficial ... Aug 17 14:19:33 emte: well, if a layer of convience adds that much inconvience I don't setup a build... Aug 17 14:20:05 zecke: stop being silly. you're well capable of writing a local.conf yourself Aug 17 14:20:25 reenoo: hehe Aug 17 14:20:26 maybe i am missing something, but building with or without zenity running makes no difference Aug 17 14:21:05 the zenity stuff was just an attempt to deal with the ton of mail I got from people asking me to help them adjust my local.conf for their needs Aug 17 14:21:15 reenoo: limited time is/was another reason, but a non working zenity pissed me off :) Aug 17 14:21:41 zecke: you'll have to take that up with the debian gnome maintainers I guess Aug 17 14:21:42 anyway now I'm searching for a new server and I will setup builds for familiar once I found a free pizza box in the rack Aug 17 14:22:20 reenoo: do you know if zenity is compatible with the dialog syntax? Aug 17 14:22:31 dunno. unlikely Aug 17 14:22:36 has a gdialog wrapper ... Aug 17 14:25:20 reenoo: Right, IIRC it was once again a scroll-keeper issue and encodings... so totally unrelated to familiar Aug 17 14:26:23 OK, well I have the problem tracked down now Aug 17 14:26:36 tkp: thumb up Aug 17 14:26:56 tkp: is it using a wrong gcc? or is something from /usr/lib pulling libc.so? Aug 17 14:27:00 I hadn't completely removed my customizations, and it's one particular one that I adding incorrected linker locations Aug 17 14:27:05 --enable-x11 Aug 17 14:27:23 without that option all is fine Aug 17 14:27:32 tkp: mplayer(?) must be at fault Aug 17 14:27:37 yep Aug 17 14:27:43 needs patching I guess Aug 17 14:27:46 tkp: consult the config.log to see why it tries to use that Aug 17 14:27:54 tkp: patching is good, sending patches upstream is better Aug 17 14:28:19 well, if I get it fixed, Ill send the patch to the bugtracker Aug 17 14:28:32 of mplayer Aug 17 14:28:42 ah I see... Aug 17 14:28:46 can do Aug 17 14:29:00 maybe they already have a patch for it? Aug 17 14:29:21 presumably it'll be fixed in svn, but thats no good for me Aug 17 14:29:28 tkp: maybe Aug 17 14:29:43 tkp: but you can always 'backport' the fix (if they have one) Aug 17 14:29:52 * tkp skips off to #mplayer Aug 17 14:31:23 morning Aug 17 14:31:30 tkp: I never went as far as fixing x11 compiles as mplayer in OE uses SDL to access either X11 or Qt/e without needed special versions of mplayer Aug 17 14:34:29 XorA: I didn't realize mplayer could access X11 without enabling support for it Aug 17 14:34:41 tkp: SDL is a good library :-) Aug 17 14:34:43 * tkp doesn't know about SDL Aug 17 14:34:56 mplayer -vo sdl under x11 will open an x11 window Aug 17 14:34:58 guess I can try with that then Aug 17 14:34:59 Ifaistos: There's no powerpc-linux-uclibc site file at all, so you'll get a lot of things fail that target... you'll probably need to fix as many as you can and then submit a complete file. Aug 17 14:35:08 I see Aug 17 14:35:19 * tkp wonders if that will work with kaa Aug 17 14:36:59 XorA: mplayer with caca roxx over ssh conection Aug 17 14:37:13 v8jlene: yes i am working on it. from the bugs i posted today i have made patches for all of them, just need to double check them before submission Aug 17 14:37:32 heh... damn sdl fails to build! Aug 17 14:37:43 or rather, libqpe-opie-1.2.2+cvs-20060817-r0 fails to build Aug 17 14:38:15 cannot find -lqte Aug 17 14:38:38 do the nlsu2 distros still merge with OE or do they have their own tree ? Aug 17 14:38:38 Ifaistos: I posted some changes to oe mailing list today to allow merging most of the site files together today so we don't need to repeat all this stuff. If people agree to that it should make things a lot easier. Aug 17 14:38:59 so qte is missing in DEPENDS somewhere then Aug 17 14:39:01 If it ever show up on the mailing list... Aug 17 14:39:33 Ifaistos: merged Aug 17 14:40:16 v8jlene: nice :) . i'll have a look and let you know Aug 17 14:42:00 Ifaistos: I think we merge Aug 17 14:42:55 03freyther 07org.oe.dev * r03841e8e... 10/ (1 classes/tinderclient.bbclass): (log message trimmed) Aug 17 14:42:55 classes/tinderclient.bbclass: Add workaround for bitbake -k and save the status Aug 17 14:42:55 Using the continue option of BitBake can lead to some issues with the Aug 17 14:42:55 Tinderbox. Even if one build failed the color of the result will be green Aug 17 14:42:55 as the BuildCompleted event will be send by BitBake. Aug 17 14:42:55 As a workaround we will store the failure in a tinder-status file in case Aug 17 14:42:59 of failure (e.g. no Provider or PkgFailed) and will reread on the start. Aug 17 14:43:01 03freyther 07org.oe.dev * rbad145ee... 10/ (1 classes/tinderclient.bbclass): Aug 17 14:43:03 classes/tinderclient.bbclass: Fixes. truncate does not work this way... Aug 17 14:43:05 Fixup truncate to forget the the status of previous runs. We will Aug 17 14:43:07 simply write("") to the file and empty it this way. int(f.read()) will Aug 17 14:43:09 fail then Aug 17 14:43:24 03freyther 07org.oe.dev * rcbcfc49a... 10/ (1 classes/tinderclient.bbclass): classes/tinderclient.bbclass: TINDER_AUTOBUILD=0 to disable running do_clean Aug 17 14:44:22 hrw|work: NOTE: package bmpx-0.20.3: completed Aug 17 14:44:28 what's the smallest qte I can use? Aug 17 14:44:38 hrw|work: TINDER_AUTOBUILD is there Aug 17 14:44:46 mickey|thesis: ping? Aug 17 14:49:20 zecke: great Aug 17 14:49:53 zecke: time to pluck it into .oz354x then Aug 17 15:03:31 so, SDL is no good for me Aug 17 15:03:59 kaa (the program I'm trying to link mplayer too) does not support X11 through SDL yet Aug 17 15:04:14 so... what happened to binutils? Aug 17 15:04:40 mplayer used to compile ok with the --enable-x11 option Aug 17 15:14:30 Aug 17 15:28:30 Is there anyway to tell bitbake to use a different local.conf from the file in build/conf/ ? Aug 17 15:29:17 lrg: change BBPATH variable Aug 17 15:32:04 XorA: thanks, I was hpoing for a bitbake switch. I have numerous boards and would like to have been able to do something like bitbake -conf=/path/to/conf Aug 17 15:33:15 lrg: echo 'MACHINE = "spitz"' > conf/auto.conf Aug 17 15:34:04 lrg: Also, the names of the files are at the end of bitbake.conf so you could tweak to include a shell variable Aug 17 15:35:00 RP: ah thanks Aug 17 15:36:59 bah, going to have to find room for both OZ and Angstrom on me poor Z Aug 17 15:38:00 me too, when i get back from my hols i'm going to have to have two roots on diff sd Aug 17 15:38:23 cu Aug 17 15:38:28 Angstrom is currently so ill, I cant tell if I need to fix apps, or fix Angstrom Aug 17 15:39:01 XorA: erm... yes! both Aug 17 15:39:37 once it does work i think debian will be taking stuff from OE rather tahn the other way around for a change ;) Aug 17 15:39:42 XorA: btw, I have Angstrom fixes for Mainstone and i.MX31 Aug 17 15:39:47 cu hrw|gone Aug 17 15:40:22 lrg: currently we are finding stuff which uses shm crashes randomely, Xfbdev dies, loadkeys dies and other stuff segfaults at random Aug 17 15:41:02 * XorA suspects the cse.patch might have side effects for gcc Aug 17 15:41:11 XorA: :-/ Aug 17 15:41:43 I need to free up diskspace so I can do 3 different build trees Aug 17 15:42:00 can we do better things for alsa so that cse patch is not used Aug 17 15:42:42 ade|desk: I need to prove/disprove theory, then we can get back to gcc, it may be fluke that the custage happened around same time that patch was applied Aug 17 15:42:56 s/cust/bust/ Aug 17 15:43:54 crap 800M left in /home again Aug 17 15:44:08 what else changed in the toolchain between breakages ... cse is all i can think of Aug 17 15:44:43 ade|desk: yes, that is all I can think of as well, hence why I would like to build a rootfs without cse, with current repo contents Aug 17 15:45:15 binutils didn't change did it, gcc version (only the patch) glibc is the same Aug 17 15:45:36 anyway I have to catch train home, laters Aug 17 15:46:12 hehe maybe i should give you a login for my employers 4way with a few spare gigs on the san Aug 17 15:46:19 cu XorA|gone Aug 17 15:52:02 its days like these that make you wish you didn't work in a green house, the rain is so loud Aug 17 15:54:11 under opie, konqueror-embedded and xmms-embedded are really old - is there any newer alternative to these or updates for these in oe .dev branch? Aug 17 15:58:21 anyone know a combination of freetype-native/fontconfig-native that builds ok? Aug 17 15:58:52 the default freetype-native version does not build ( freetype-native_2.1.10.bb ) Aug 17 15:59:14 so I got it to build by setting preferered version to 2.2.1 Aug 17 15:59:32 but with that, fontconfig-native fails to build Aug 17 15:59:45 (version 2.3.91.bb - the default) Aug 17 16:00:45 a+ Aug 17 16:08:42 later Aug 17 16:28:16 easy question: on a zaurus 5500, is the usb port a host, device, or capable of being either? Aug 17 16:29:09 and, if it's capable of being a device, do standard linux usb drivers work with it? (I want to use it with a FTDI 245M, for which linux drivers exist) Aug 17 16:37:08 * Ifaistos marvels his new qtek 9000 and wonders where the jtag pins might be Aug 17 16:37:58 this pda needs to be liberated from windows :) Aug 17 16:46:43 mr_nice_slacker: aargh, again no detect Aug 17 16:59:55 x29a: as i sayed try to get your modules working proppary - check the wiki ask parts you don't understand so we can improve the wiki Aug 17 17:05:51 x29a:btw: i am installing my "new fileserver" (p 350 *g) atm ;) - after that i have a to set up my old fileserver (p 700) to a buildserver. than i have enought space (80GB) and resources to build the hole oe and set up that buildserver thing for opensimpad there. Aug 17 17:52:51 . Aug 17 17:55:33 . Aug 17 18:03:06 hi for moment of few Aug 17 18:03:23 KeyserSoze: its usb device only (zaurus 5500 usb port) Aug 17 18:03:32 poushag: no newer versions exist Aug 17 18:06:00 mr_nice_slacker: i assume u use minicom for serial communication? Aug 17 18:06:46 mr_nice_slacker: can u help me set up a serial connection with minicom, so i dont have to bring the board back and forth? Aug 17 18:07:16 x29a soft hard flow of connection speed to 11125000 done Aug 17 18:07:24 thanks hrw Aug 17 18:07:32 and chosse the right device under /dev/ Aug 17 18:08:43 woglinde: do i not have to remove any modemcommands? Aug 17 18:08:45 x29a: yes woglinde is right Aug 17 18:08:51 and which is "the right device"? Aug 17 18:09:38 ttyS0 or ttyS1 Aug 17 18:09:54 if you can read serial1 or serial2 on your host-pc Aug 17 18:09:55 x29a: dmesg | grep tty Aug 17 18:10:55 I just tried both ports and checked which one is correct one Aug 17 18:11:28 x29a: or cat /proc/ioports Aug 17 18:12:14 x29a: to see if the serial ports are detected correctly Aug 17 18:12:58 x29a: could be that your bios is different to it - and if you don't use acpi you should check the bios as well Aug 17 18:16:06 x29a: do you have a at or atx class pc? Aug 17 18:16:16 atx Aug 17 18:17:04 does minicom give feedback of success after initializing modem (M)? Aug 17 18:17:43 mr_nice_slacker: cat /proc/ioports is hardcore a bit - 44 lines Aug 17 18:18:08 x29a: not really Aug 17 18:18:38 hrw: well.. you can use | grep tty ore something Aug 17 18:18:46 mr_nice_slacker: so i start minicom, press strg+a+m and then? Aug 17 18:19:22 x29a: normaly i start it with minicom -s then i set up the line Aug 17 18:19:43 x29a: than i save it as default and exit to minicom Aug 17 18:20:13 x29a: than i get stuff from the simpad Aug 17 18:20:14 did u delete all the modem commands? Aug 17 18:20:30 x29a: no i only set up the serial stuff Aug 17 18:21:47 x29a: here is my setup: A - Serial Device : /dev/ttyS0 B - Lockfile Location : /var/lock C - Callin Program : x D - Callout Program : E - Bps/Par/Bits : 115200 8N1 F - Hardware Flow Control : No G - Software Flow Control : No Aug 17 18:22:04 x29a: that setup is for hh.org bootldr only! Aug 17 18:22:44 x29a: for simpad loader you have to change E to (G: 38400) Aug 17 18:22:56 x29a: and exit to minicom or save it Aug 17 18:23:30 x29a: if you are exiting to minicom it doesn't save it you have to change it every time you use it Aug 17 18:24:18 x29a: i hope i havn't confused you more than i helped Aug 17 18:28:36 anyone knows if the motorola a1200 is already available? Aug 17 18:30:42 katossi no and when only in china Aug 17 18:30:56 katossi and the rokr e2 looks nicer Aug 17 18:31:07 too bad, if I just could get a a780 Aug 17 18:31:25 indeed it looks great, will it be available in europe? Aug 17 18:31:28 mr_nice_slacker: Erase Block Size: 131072 B (128 KiB) Aug 17 18:31:28 Number of Erase Blocks: 128 Aug 17 18:31:50 katoiss go on ebay a780 sold about 200 euros Aug 17 18:31:55 dunno what an erase block is, but can i conclude from that, that one block is 131072B? Aug 17 18:32:04 x29a: yes Aug 17 18:32:13 motorola suggested sommer for the rokr but alway they are late Aug 17 18:32:26 x29a: the image i sent you was 5 blocks large Aug 17 18:33:38 woglinde: no gps on the rokr e2 though Aug 17 18:34:06 hm Aug 17 18:34:19 not everyone needs wifi gprs utms or gps Aug 17 18:34:27 mr_nice_slacker: so i have to erase 128 blocks in order to get it all emtpy Aug 17 18:35:05 x29a: mom - i have to calculate *g Aug 17 18:35:15 i did already Aug 17 18:35:20 and it actually says it Aug 17 18:35:27 number of erase blocks: 128 Aug 17 18:35:33 x29a: then it will fit Aug 17 18:35:42 hmm, i once erased 512 blocks, u think it matters? Aug 17 18:36:01 x29a: i think no ;) Aug 17 18:36:12 im curious where jtag wrote if it only can access the first 16m Aug 17 18:36:28 x29a: we need to hack the jtag code to get a save option Aug 17 18:36:40 save of what? Aug 17 18:36:57 x29a: that we don't write or save mor than 16MB *g Aug 17 18:36:59 wont readmem do the job? Aug 17 18:37:17 oh yeah, or just use($brain); a lil ;) Aug 17 18:37:25 x29a: *g Aug 17 18:37:30 i read too much, i erase to much, seems like a standard issue Aug 17 18:37:35 *u Aug 17 18:37:56 x29a: or we should write in the manual that it is all in byte Aug 17 18:37:58 03mreimer 07org.oe.dev * r78ac43a9... 10/ (4 files in 3 dirs): Aug 17 18:37:58 LAB-kernel: Generate a LAB kernel suitable for flashing using the HTC Aug 17 18:37:58 rescue bootloader. The kernel is installed to tmp/deploy/images and Aug 17 18:37:58 its filename ends with .htc. Aug 17 18:39:57 mr_nice_slacker: rumor has it that some ppl will even read to much if it says _bytes_ Aug 17 18:41:05 x29a: does it says bytes - oh *g Aug 17 18:41:20 mr_nice_slacker: jtag> eraseflash 0 128 should do right? Aug 17 18:41:35 x29a: no Aug 17 18:41:43 why not? Aug 17 18:41:51 x29a: you have to set the second in byte not in blocks Aug 17 18:41:58 no Aug 17 18:42:13 x29a: blocks are alos allowed? Aug 17 18:42:19 only Aug 17 18:42:19 Usage: eraseflash ADDR BLOCKS Aug 17 18:42:34 x29a: a ok the readflash only allowes bytes Aug 17 18:42:43 yeah Aug 17 18:42:46 x29a: than it should do it Aug 17 18:42:51 erasing done Aug 17 18:42:55 x29a: fine Aug 17 18:43:01 just imagine flashing 16M of zeros Aug 17 18:43:11 k, now simpadSL.rom Aug 17 18:43:30 x29a: the bosted one? Aug 17 18:43:57 hmm, didnt find the exact adresses of bost options in simpadSL.rom Aug 17 18:44:07 just edited the menu, should i use that one anyway? Aug 17 18:44:38 x29a: i didn't found the adress in simadSL.rom ,too Aug 17 18:44:43 since simpadSL.rom is the original one and loader_alt is just an update Aug 17 18:44:51 x29a: so we need the code or a disassembler Aug 17 18:44:59 no Aug 17 18:45:06 x29a: but upgrading 2.4 with 2.4?? Aug 17 18:45:22 x29a: or loader_alt 2.5.3 Aug 17 18:45:35 mr_nice_slacker: mission: install simpadSL.rom -> upgrade with loader_alt (which is still 2.4 i think) -> read bootloader via jtag -> edit bits -> reflash Aug 17 18:46:01 i think "loader" is 2.5.3 and "loader_alt" is still 2.4? Aug 17 18:46:22 x29a: no Aug 17 18:46:30 both 2.5.3? Aug 17 18:46:31 x29a: loader_alt is a alternative loader Aug 17 18:46:44 hmm, k Aug 17 18:46:50 x29a: it uses a other memory space Aug 17 18:47:02 in order to get hh.bootldr running? Aug 17 18:47:12 jtag> flashmem 0 /deb_data/JTAG/simpadSL.rom Aug 17 18:47:13 ? Aug 17 18:47:25 x29a: it is to use it to be more confort with linux kernels Aug 17 18:47:41 k Aug 17 18:47:58 x29a: well yes flash it ;) Aug 17 18:48:07 na ill go for simpadSL_bost.rom Aug 17 18:48:11 cant hurt Aug 17 18:48:24 x29a: good thinking *G Aug 17 18:48:36 x29a: bost it! Aug 17 18:48:49 hope so Aug 17 18:49:14 x29a: also hope so Aug 17 18:49:43 x29a: i didn't found time to buy the "steckerleiste" :( Aug 17 18:50:00 dont bother, i really hope i can get it set up Aug 17 18:50:05 x29a: how "steckerleiste" is called in english? Aug 17 18:50:16 plug? Aug 17 18:50:24 pinboard? Aug 17 18:50:28 x29a: no guess Aug 17 18:50:48 http://dict.leo.org/ende?lp=ende&lang=de&searchLoc=0&cmpType=relaxed§Hdr=on&spellToler=on&search=steckerleiste&relink=on Aug 17 18:50:53 edge connector Aug 17 18:51:30 x29a: hehe thx Aug 17 19:02:04 x29a: btw - here is a picute of my new server case :) http://opensimpad.org/index.php/Image:OpenBSD.jpg Aug 17 19:03:04 x29a: i asked my girfriend a lot and then she painted it for me - hehe Aug 17 19:03:47 mr_nice_slacker: dude, thats awesome Aug 17 19:03:56 but we have a shell_exec php error on that site Aug 17 19:04:18 x29a: yes - i know - safemod and mediawiki *arg Aug 17 19:04:48 x29a: with typo3 and safemod i have no problems Aug 17 19:05:22 x29a: and there i use imagemagic as well Aug 17 19:05:58 x29a: if i find time i have to look in the mediawiki code Aug 17 19:36:55 poushag: you can't Aug 17 19:37:03 poushag: at least not from svn Aug 17 19:48:40 03hrw 07org.oe.dev * rf5bdc770... 10/ (1 packages/glib-2.0/glib-2.0_2.12.1.bb): glib-2.0: fix SRC_URI for 2.12.1 version Aug 17 19:49:46 do13: wlan-ng-utils does not build with DISTRO=angstrom-2006.9 Aug 17 19:50:58 hrw|tv: svn version? Aug 17 19:52:48 yes Aug 17 19:53:15 do13: even with only-utils patch removed Aug 17 19:53:27 do13: bad params to linker Aug 17 19:53:41 mr_nice_slacker: started new flash Aug 17 19:53:52 some patches are outdated. bugzilla has an update, but I had no time to look at this Aug 17 19:54:01 re Aug 17 19:54:30 do13: OE bugzilla? Aug 17 19:54:36 yep Aug 17 19:54:53 hrw|tv: a few days ago the promised to publish a new version. Aug 17 19:56:44 but not released Aug 17 19:56:51 hrw|tv: Hmm maybe we should remove the svn bb and add fixed svn snapshots Aug 17 19:57:29 good idea Aug 17 19:57:30 hrw|tv: yeah. I'am waiting for the next release Aug 17 19:59:51 I'm trying to build angstrom images Aug 17 20:00:42 tosa image finished? Aug 17 20:01:48 failed on wlan-ng-utils Aug 17 20:01:55 thx for response hrw and justinp - sounds like opie is practically dead in the water Aug 17 20:02:24 at least as far as oz is concerned Aug 17 20:03:27 ~lart wlan-ng so-called-devs for such broken builddystem Aug 17 20:03:28 * ibot whips out a sword and chops wlan-ng so-called-devs in half for such broken builddystem Aug 17 20:03:35 bitbake -b ~/devel/oe/org.openembedded/packages/linux/linux-openzaurus_2.6.17.bb -ccompile Aug 17 20:03:38 hehe Aug 17 20:03:45 or it will not build Aug 17 20:04:45 I WANT hostap_usb.ko Aug 17 20:05:07 this is the best solution Aug 17 20:05:52 would be Aug 17 20:06:47 but the hostap guys have no interest Aug 17 20:07:59 and we have other things to do Aug 17 20:08:18 NOTE: package wlan-ng-utils-0.2.4+svn20060817-r4: task do_build: completed Aug 17 20:08:51 hi mallum Aug 17 20:09:02 hey florian Aug 17 20:09:18 mr_nice_slacker: unknown flash error, restart Aug 17 20:12:00 do13: modules_svn need drop of pcmcia-driver patch Aug 17 20:19:10 hrw|tv: I suspect these changes are alreay in the svn Aug 17 20:19:37 do13: exactly Aug 17 20:20:53 does anyone know if the kernel headers in staging/arch/include come from the virtual/kernel or from the kernel-headers package ? Aug 17 20:22:00 guess they come from kerne-headers Aug 17 20:23:20 what about opie-mediaplayer? anybody tried that? version1 or ver2? which might work better on collie? Aug 17 20:24:13 poushag: #openzaurus Aug 17 20:26:40 this is still not a distribution support channel Aug 17 20:30:41 woglinde : yes it looks so. they come from the linux-libc-headers package. Aug 17 20:31:07 any idea what variable holds the path to the virtual/kernel headers ? Aug 17 20:31:56 STAGING_KERNEL_DIR iirc Aug 17 20:32:43 do13: I will push .oz354x tosa changes into .dev on days Aug 17 20:33:21 how come the ipkg package does not install a ipkg.conf file (for use with ipkg feeds)? Aug 17 20:33:38 where do other distros tend to do this? Aug 17 20:34:01 hrw|tv: ok. should we think about removing the 2.4 crap from .oz354x? Aug 17 20:34:35 there is a #FIX ME note in the ipkg bb file sugesting that this is yet to have been implemented Aug 17 20:34:41 do13: basically we can do that for tosa and poodle Aug 17 20:35:08 tkp: rootfs_ipk.bbclass do that iirc + ipkg-collateral package Aug 17 20:36:46 ipkg-collateral? I see no ipkg-collateral package Aug 17 20:37:17 * Ifaistos ifaistos is in header hell Aug 17 20:37:32 ah, you mean ipkg-cl? Aug 17 20:37:36 tkp: no Aug 17 20:37:50 tkp: packages/ipkg/ipkg-collateral.bb Aug 17 20:37:55 one of my first packages Aug 17 20:37:57 ahhh Aug 17 20:38:46 cool... thanks Aug 17 20:38:50 ChangeSet@1.1976.1.1 2004-09-02 17:22:22+02:00 hrw@local Aug 17 20:39:33 long time ago in other SCM... Aug 17 20:42:24 ok. time to go sleep Aug 17 20:42:25 cu Aug 17 20:49:31 what is the purpose of FEED_URIS in DISTRO.conf? Aug 17 20:53:24 they will end up in the ipkg configuration files in the images Aug 17 20:53:39 ahh. ok, great Aug 17 20:53:47 I thought it was probably that Aug 18 02:10:01 03lenehan 07org.oe.documentation * r8d82f639... 10/ (11 files in 3 dirs): Aug 18 02:10:01 usermanual: Add details on the image_ipk and rootfs_ipk to the Aug 18 02:10:01 reference section of the document. Add id and xreflabel to all Aug 18 02:10:01 sections in the reference documentation so we can refer to the Aug 18 02:10:01 individual sections (which is used here to refer to the image types). Aug 18 02:27:47 03lenehan 07org.oe.documentation * r30e3810d... 10/ (8 files in 2 dirs): Aug 18 02:27:47 usermanual: Minor grammer fixes. Update the autotools site example to match Aug 18 02:27:47 a recent change to the code. Change the reference to to image_ipkg class in Aug 18 02:27:47 image_types into an actual xref. Aug 18 02:32:54 Question: How do I insert a step at the end of do_populate? Aug 18 02:53:56 do_package_prepend seems not to do anything or get called. Is there a trick? Aug 18 02:59:41 yea! do_stage_append works. Just in case anyone was reading :) **** ENDING LOGGING AT Fri Aug 18 02:59:57 2006