**** BEGIN LOGGING AT Tue Jul 31 02:59:56 2007 Jul 31 03:01:19 How can I use the oe built to nfs boot my target Jul 31 03:04:37 That will probably vary depending on what you are building with OE, but you need to make your packaging build a rootfs in a form that you can extract to an NFS server, and adjust your kernel params to build a kernel that can boot via NFS. Jul 31 03:07:11 mwester, All I want to know is where does it install rootfs Jul 31 03:07:29 rest all I am able to do as I boot other rfs's too Jul 31 03:09:41 Depends on what you're building with OE -- each distro can create its own images, such as a jffs2 flash image or similar. They will end up in the tmp/deploy directory usually. Jul 31 04:12:51 03mickeyl * r856 10bitbake/lib/bb/fetch/__init__.py: fetch/__init__.py: fix error in error reporting statement ;) Jul 31 05:11:08 03mickeyl 07org.oe.dev * r2330acf3... 10/ (6 files in 4 dirs): move libmnetutil and minisip to nonworking. neither snapshots nor svn is fetchable Jul 31 05:11:13 03mickeyl 07org.oe.dev * rb0ac6c31... 10/ (18 files in 9 dirs): minisip and related libraries: fix SRC_URI and bring svn versions back from nonworking Jul 31 05:11:22 03mickeyl 07org.oe.dev * r05893dd0... 10/ (1 packages/minisip/minisip-video_svn.bb): minisip-video: fix SRC_URI Jul 31 05:11:32 03mickeyl 07org.oe.dev * r9c3d959d... 10/ (3 files in 3 dirs): move simple-firewall to nonworking (unfetchable and no sign of alternate download location) Jul 31 05:11:37 03mickeyl 07org.oe.dev * rfbe92b16... 10/ (3 files in 3 dirs): move fis to nonworking (unfetchable) Jul 31 05:11:42 03mickeyl 07org.oe.dev * r2f8414c3... 10/ (4 files in 3 dirs): simple-firewall -> nylon/ Jul 31 05:11:50 03mickeyl 07org.oe.dev * r5aa52a23... 10/ (4 files in 3 dirs): fis -> redboot-utils Jul 31 05:35:15 03mickeyl * r857 10bitbake/lib/bb/fetch/svn.py: fetch/svn.py: use LC_ALL=C LANG=C (LC_ALL= LANG= doesn't work on all distributions) Jul 31 05:36:41 yawn Jul 31 05:36:45 i should really go to bed nw Jul 31 05:36:47 now Jul 31 05:37:31 mickey|zzZZzz : good morning :) Jul 31 05:41:24 hardly Jul 31 05:41:26 couldn't sleep Jul 31 05:41:40 ended up doing OE work Jul 31 05:41:41 heh Jul 31 05:42:58 mickey|zzZZzz: you found an alternate location for fis? Jul 31 05:43:13 www.angstrom-distribution.org/unstable/sources/fis.c Jul 31 05:43:58 no i didn't. guess we should drag it into the repo then Jul 31 05:44:02 * mickey|zzZZzz does it Jul 31 05:44:14 is it just one file? Jul 31 05:45:16 no, that won't work Jul 31 05:45:24 it's svn://svn.chezphil.org/ that doesn't work Jul 31 05:45:30 and fis is using libpbe from it Jul 31 05:47:57 rwhitby: btw., is it me or has the amount of nslu2 commits drastically decreased over the last months? Jul 31 05:48:40 mickey|zzZZzz: it has drastically decreased. Jul 31 05:48:54 I've been working on openwrt support for nslu2 :-) Jul 31 05:49:09 oh. why's that? Jul 31 05:49:33 well, actually openwrt for nas100d, fsg3 and dsmg600 Jul 31 05:49:49 Unslung is pretty much at the end of where it can go, reasonably, so there's been few commits for that... Jul 31 05:49:57 we're looking for a good replacement for ucslugc, and openwrt seems to be it Jul 31 05:50:08 (i.e. lots of packages in internal flash) Jul 31 05:50:14 i see. so the uClibC support in OE is not up to par? Jul 31 05:50:22 can we do something to improve that? Jul 31 05:51:04 the thing is that there are too few people in OE focusing on uclibc, and none of the nslu2 people know enough about it to change that, so we went to openwrt, cause they specialise in it. Jul 31 05:51:23 And SlugOS is moving towards just being a build of Angstrom. Jul 31 05:51:34 So that's why there are very few nslu2-specific checkins. Jul 31 05:51:55 (oh, we've been concentrating on kernel stuff too, which is in our svn repo) Jul 31 05:52:09 and openmoko keeps me busy too Jul 31 05:52:15 nod i was just curious as to whether we had something done to offend you guys Jul 31 05:52:28 i was away for a while and I'm trying to catch up Jul 31 05:52:44 oh, thanks for asking. I appreciate that. No, nothing more than koen's normal customer service attitude ;-) Jul 31 05:52:49 ;) Jul 31 05:52:55 (but we're well used to that by now) Jul 31 05:53:19 I was sad to see Rolf leave. That's a big loss. Jul 31 05:54:12 mickey|zzZZzz: did you see we added SD card support to QEMU today, and a mokomakefile target 'qemu-copy-package-%' to copy a newly built ipkg to the virtual SD card ready for ipkg install in the emulator? Jul 31 05:54:30 i agree. but not all hope is lost. i kind of not accept that there is no way back. we need to improve our social skills. period. Jul 31 05:54:35 "... SD card support to how MokoMakefile runs qemu" I should say. Jul 31 05:54:42 rwhitby: oh no, i didn't see that. cool! Jul 31 05:54:46 really cool Jul 31 05:54:51 * mickey|zzZZzz has to give that a try Jul 31 05:54:55 cesarb did the hard work. Jul 31 05:54:59 I just merged it :-) Jul 31 05:55:04 :) Jul 31 05:55:20 hey I tried bitbake world and it is missing some deps Jul 31 05:55:38 like Jul 31 05:55:40 preferred version 1.0.3 of xkbcomp not available (for item xkbcomp) Jul 31 05:56:04 some 9 packages report this Jul 31 05:57:25 i'm afraid world is not buildable most of the times. it would take over a week and several hundred gigs of harddrive space to build world anyways. and you would run into dependency problems like packages needing different versions of libraries which then conflict in staging directory. you better go for individual packages, task-packages, image packages etc. Jul 31 05:57:35 anyways, trying again to sleep now Jul 31 05:57:36 g'night Jul 31 05:57:59 mickey|zzZZzz, gute nacht I will Jul 31 06:01:12 mickey|zzZZzz : good night Jul 31 06:41:12 morning Jul 31 06:53:16 on fedora 7 I am getting this error when running bitbake Jul 31 06:53:27 ERROR: Importing sqlite3 and pysqlite2 failed, please install one of them. A 'python-pysqlite2' like package is likely to be what you need. Jul 31 06:54:04 I have tried to install all pysqlite related packages but of no use Jul 31 06:54:08 any help? Jul 31 07:01:30 too recent change for me to help Jul 31 07:21:12 03koen 07org.oe.dev * r39032111... 10/ (1 conf/distro/include/sane-srcdates.inc): disapproval of revision '595f85835a210f4f204d60ffa60017c0d342736b' Jul 31 07:27:17 khem: You work near my companies headquarters (Nellingen) :) Jul 31 07:28:01 khem: I had the same problem, then I realised debian was installing python2.5 modules without actually installing the python2.5 binary... Jul 31 07:28:35 khem: you could aslo use the 1.8.6 tarball release of bitbake Jul 31 07:32:43 koen, my problem is that when I use native python all is ok Jul 31 07:34:08 koen, but I added openembedded/tmp/staging/i686-linux/bin Jul 31 07:34:33 to my path and then it started picking up the newly built python which caused this issue Jul 31 07:35:06 koen, the above I have to do to workaround bug 2554 Jul 31 07:35:29 khem: do not add staging to PATH Jul 31 07:36:06 hrw, but then what is solution for bug 2554 Jul 31 07:36:24 !oebug 2554 Jul 31 07:36:26 * * Bug 2554, Status: NEW, Created: 2007-06-23 11:51 Jul 31 07:36:27 * * utx(AT)penguin.cz: broken cross build environment / fix ncurses build Jul 31 07:36:28 * * http://bugs.openembedded.org/show_bug.cgi?id=2554 Jul 31 07:36:56 one day I will have to do some test builds on fedora box Jul 31 07:48:31 hrw, for now what is other sol to bug 2554 Jul 31 07:48:49 morning Jul 31 07:51:17 khem: no idea now Jul 31 07:54:54 03koen 07org.oe.dev * r24aef90d... 10/ (3 files in 3 dirs): openmoko: move NOW srcdates to its own include file and only include it in the openmoko distro Jul 31 07:56:21 * khem goes to bed Jul 31 08:12:16 morning all Jul 31 08:13:30 RP : Hi Jul 31 08:13:31 any1 here who can help me a bit with timezone configuration ? Jul 31 08:13:55 putumutukas : go on Jul 31 08:14:19 i have angstrom distro in my small omap5912 Jul 31 08:14:30 installed openntpd and it works great Jul 31 08:14:41 but now i want to configure it to use timezone Jul 31 08:14:46 03rpurdie 07bitbake-1.8 * r858 10/lib/bb/fetch/ (__init__.py svn.py): Add fixes from trunk to 1.8 branch Jul 31 08:14:53 installed timezones-Europe, linked correct city Jul 31 08:16:06 but now i don't know how to tell that from now on let's use localtime instead of UTC Jul 31 08:16:38 good morning Jul 31 08:16:46 morgning :) Jul 31 08:16:48 timezone and ntp are two different things Jul 31 08:16:53 yes Jul 31 08:17:00 timezone is an OS thing Jul 31 08:17:04 yes Jul 31 08:17:28 and now i want my OS to display my local time instead of universal Jul 31 08:18:03 change the TZ enviroment variable to the timezone you want the time to be displayed Jul 31 08:19:22 in os variables ? Jul 31 08:20:09 TZ is a linux enviroment variable that hold the timezone Jul 31 08:21:20 Hi, what is the bitbake command to load the latest Handheld kernel ? Jul 31 08:21:41 ariadni:~# date -u; TZ=UTC+2 date Jul 31 08:21:41 Tue Jul 31 08:21:27 UTC 2007 Jul 31 08:21:41 Tue Jul 31 06:21:27 UTC 2007 Jul 31 08:21:52 hey florian Jul 31 08:23:34 koen: Do the SRCREV ideas look ok to you? Jul 31 08:23:39 hmm Jul 31 08:24:00 RP: yes, my only comment was about the patches mindate and maxdate Jul 31 08:24:05 but can i also use the localtime (due the daylight changes) Jul 31 08:24:08 the rest looks ok Jul 31 08:24:20 koen: Yes, a maxrev minrev should be straightforward to implement Jul 31 08:25:01 putumutukas : TZ can be set to any of the timezones Jul 31 08:26:12 ariadni:~# date -u; TZ=Europe/Athens date Jul 31 08:26:12 Tue Jul 31 08:25:57 UTC 2007 Jul 31 08:26:12 Tue Jul 31 11:25:57 EEST 2007 Jul 31 08:26:23 :D Jul 31 08:26:31 that's the one i was looking for Jul 31 08:26:34 thank you :) Jul 31 08:27:35 putumutukas : You can also set the TZ in the initscripts so you get the same timezone when you reboot Jul 31 08:28:01 03koen 07org.oe.dev * r755c1ef8... 10/ (3 files in 3 dirs): openmoko-rssreader: rename to openmoko-feedreader2 Jul 31 08:28:02 which script should i edit ? /etc/bashrc ? Jul 31 08:31:04 putumutukas : etc/profile Jul 31 08:31:08 found it already in /etc/profile Jul 31 08:31:27 thank you very much steliosk Jul 31 08:33:19 putumutukas : just on e last note (it took me several days to figure it out) the GMT timezone are "wrong" i.e they are the reverse of what you expect Jul 31 08:34:16 i.e gmt+2 will display the time difference from that point and not the other way around Jul 31 08:35:06 for example AThens is gmt+3 but to get the correct time you have to set it to gmt-3 Jul 31 08:35:20 i discovered it already that gmt+2 is utc-2 :D Jul 31 08:36:29 * steliosk thinks that its nice to have so many standards to choose from Jul 31 08:36:58 ...as they guarantee that you will make a mistake Jul 31 08:37:27 especially funny is that i'm from estonia, and few years ago our government thought that it would be great no to use summer/winter time Jul 31 08:37:53 then after few years they overthought and now we use summertime Jul 31 08:38:30 but in many applications where timezones are involved estonia is without summer/wintertime :D Jul 31 08:38:57 so i live in estonia but in most of the cases i have to use finland as timezone Jul 31 08:39:09 although estonia is available to choose Jul 31 08:42:38 but thanks again and bye folks ! Jul 31 08:46:12 gah Jul 31 08:46:18 * koen runs into another ipkg bug Jul 31 08:47:18 03koen 07org.oe.dev * rf563ab05... 10/ (1 packages/openmoko2/openmoko-icon-theme-standard_svn.bb): openmoko-icon-theme: provide upgrade path Jul 31 08:47:51 koen: oh dear... we need to rename it anyway ;) Jul 31 08:54:41 koen: OE has two openmoko-icon-theme-standard recipes now ;( Jul 31 08:54:51 and both are same rather Jul 31 08:54:53 03koen 07org.oe.dev * rf0acec47... 10/ (1 packages/openmoko2/openmoko-icon-theme-standard_svn.bb): disapproval of revision 'f563ab05310af7faf00f83e7bebcae99a5ccb4d8' Jul 31 08:54:57 hrw: I know, complain to mickey|zzZZzz Jul 31 08:54:59 03koen 07org.oe.dev * r3ec7cab8... 10/ (1 packages/openmoko2/openmoko-theme-standard2_svn.bb): openmoko-theme-standard: fix typo in PV Jul 31 08:55:08 koen: and he will complain to me Jul 31 08:55:28 as openmoko2 recipes in OE are based on OH internal ones written by me Jul 31 08:55:34 the whole openmoko2 thing in OE feels like it was never ever tested Jul 31 08:55:58 I sent all of them to mickeyl and icon-theme was one of them as we needed that Jul 31 08:56:03 like SRC_URIs not making sense till mickeyl moves them on the svn server a while later Jul 31 08:56:12 koen: in OE was never - I used them with Poky Jul 31 08:56:24 ARGH.. mickeyl and svn move ARGHHH Jul 31 08:56:48 it would be nice if people actually tried to build recipes before adding them Jul 31 08:58:04 * koen is thinking of adding an angstrom-openmoko-image Jul 31 09:00:10 03daniel 07org.oe.dev * r946f6774... 10/ (1 packages/tasks/task-openmoko.bb): task-openmoko.bb: Remove openmoko-panel-mainmenu as it depends on libmokoui(1) Jul 31 09:00:48 it would also be nice if the openmoko people would make up their minds about parallel installability and keeping 2007 in OE or not Jul 31 09:01:23 I try to fix things to the bst of my knowledge, but it seems things change by the minute Jul 31 09:04:53 koen: Well afaik the goal last night was to have some kind of image that shows the new apps in angstrom. Jul 31 09:06:09 image? I have a movie ;) Jul 31 09:07:45 koen: Heh. That's because we're still painting the picture :-) Jul 31 09:08:31 alphaone: could you hint at using task-base-extended instead of MACHINE_TASK_PROVIDER? Jul 31 09:08:48 alphaone: that will give bluetooth commandline utils and other stuff Jul 31 09:08:49 Hi, i m encountering problem building the x11 angstrom image. The problem seems to be related to gtkcombobox, known issue ? Jul 31 09:11:08 it is related to the package gtk2.10.14 Jul 31 09:11:13 koen: I'll ask mickey|zzZZzz about it. Jul 31 09:11:40 that would stop me from creating an angstrom-openmoko-image.bb Jul 31 09:14:03 koen: I'll tell him. Jul 31 09:25:15 koen: ignore 2007 - use 2007.2 Jul 31 09:25:51 koen: neo1973 has bt feature Jul 31 09:26:26 hrw: ah, right, I was confused again Jul 31 09:26:31 * koen hasn't had coffee Jul 31 09:33:04 trying here : Jul 31 09:33:07 I cannot build angstrom x11 image, it related to the package gtk+-2.10.14 Jul 31 09:33:08 Someone encountered this problem Jul 31 09:33:08 ? Jul 31 09:33:29 hello Jul 31 09:33:41 I can't see any reference on the bugtracker Jul 31 09:35:24 PsyKro: without logs my glassball does not works Jul 31 09:35:56 koen: basically 2007.2 OM apps are the one to use. I do not know is there someone who want to work on 2007 ones Jul 31 09:36:32 Oh it is not the newest glassball model, then. Ok should i report the issue on the bugtracker ? Jul 31 09:36:52 PsyKro: pastebin log and then we will try Jul 31 09:37:08 | gtkcombobox.c: In function 'gtk_combo_box_size_request': Jul 31 09:37:08 | gtkcombobox.c:1859: warning: unused variable 'font_desc' Jul 31 09:37:08 | gtkcombobox.c:1858: warning: unused variable 'metrics' Jul 31 09:37:08 | gtkcombobox.c:1857: warning: unused variable 'context' Jul 31 09:37:08 | gtkcombobox.c:1855: warning: unused variable 'arrow_size' Jul 31 09:37:08 | gtkcombobox.c:1854: warning: unused variable 'font_size' Jul 31 09:37:10 | gtkcombobox.c: In function 'gtk_combo_box_size_allocate': Jul 31 09:37:12 | gtkcombobox.c:1962: error: 'arrow_size' undeclared (first use in this function) Jul 31 09:37:13 kurwa! Jul 31 09:37:14 | gtkcombobox.c:1962: error: (Each undeclared identifier is reported only once Jul 31 09:37:16 ~pastebin Jul 31 09:37:16 | gtkcombobox.c:1962: error: for each function it appears in.) Jul 31 09:37:17 rumour has it, pastebin is a place to paste your stuff without flooding the channel - try http://pastebin.ca, or http://channels.debian.net/paste, or http://rafb.net/paste/, or http://pastebin.com is usually painfully too slow and unresponsive to use, use one of the other pastebin sites, or dpaste.com is a very nice pastebin as well Jul 31 09:37:18 | gtkcombobox.c:1965: error: 'font_desc' undeclared (first use in this function) Jul 31 09:37:22 | gtkcombobox.c:1966: error: 'context' undeclared (first use in this function) Jul 31 09:37:24 | gtkcombobox.c:1967: error: 'metrics' undeclared (first use in this function) Jul 31 09:37:25 PsyKro: fsck you (in other way) Jul 31 09:37:26 | gtkcombobox.c:1969: error: 'font_size' undeclared (first use in this function) Jul 31 09:38:13 excuse me i m not english.. what is the meaning of pastebin then ? Jul 31 09:38:22 31 11:41 < ibot> rumour has it, pastebin is a place to paste your stuff without flooding the channel - try Jul 31 09:38:56 Ok... Jul 31 09:39:04 sorry ... Jul 31 09:40:01 and paste whole log into pastebin not only end of it Jul 31 09:43:51 hrw, ok sorry for flood, here is the link: http://pastebin.com/d1a5eb67 Jul 31 09:45:32 hmm.. no idea sorry Jul 31 09:46:00 I am updating Oe now and will build it Jul 31 09:46:35 I built it from scratch last night Jul 31 09:46:41 * koen checks for local diff Jul 31 09:47:00 ok Jul 31 09:47:34 Had no problem last week, i managed to build it last week Jul 31 09:48:48 morning Jul 31 09:49:07 is there an irc channel where about oe/dreambox ? Jul 31 09:49:16 or did i found i already :) Jul 31 09:50:13 Aim2: no, it's not here. I'm not sure wheree the channel is though (enigma build). I'm interested as well, I ordered a dreambox 600 pvr last week :-) Jul 31 09:50:33 likewise: dvb-c, dvb-t, dvb-s ? Jul 31 09:51:02 well im more intrested in developingon oe/dreambox Jul 31 09:51:18 but sadly im try to make a build, without any suc6 Jul 31 09:51:23 trying* Jul 31 09:51:37 dvb-t Jul 31 09:51:47 * koen spots a dutchism :) Jul 31 09:51:49 but the 600 has exchangeable tuners -t -c or -s Jul 31 09:51:51 koen: check :) Jul 31 09:52:06 likewise: that's why I asked :) Jul 31 09:52:35 koen: yup, this model and newer have exchangeable tuners, and they even the same :-) Jul 31 09:52:46 i.e. one can bring this tuner to the newer model Jul 31 09:53:11 269 EUR for a 600pvr (w/o disk) w/ DVB-S tuner. EUR 50 extra for a DVB-T module. Jul 31 09:53:31 koen: any other questions? :-) Jul 31 10:07:23 03pH5 07org.oe.dev * r171b7760... 10/ (7 files in 3 dirs): Jul 31 10:07:23 libgsmd: update gsmd patches Jul 31 10:07:23 - alive-start-if-interpreter-ready.patch and Jul 31 10:07:23 vendor-qc-v0.patch went upstream Jul 31 10:07:23 - add mlbuf-static-bss.patch and Jul 31 10:07:25 Andrzej Zaborowski's sms-hacks.patch and libgsmd-tool-fix.patch Jul 31 10:57:42 steliosk, ping Jul 31 11:00:28 how cain i move the qemu version of angstronto from 0.8 to 0.9.0+cvs ? Jul 31 11:00:32 s/cain/can Jul 31 11:00:39 re Jul 31 11:01:57 NOTE: package gtk+-2.10.14-r1: task do_compile: completed Jul 31 11:03:20 ok i got... Jul 31 11:13:52 03koen 07org.oe.dev * r711eee81... 10/ (1 packages/gnash packages/gnash/gnash_0.8.0.bb): gnash: add 0.8.0 Jul 31 11:25:39 Hi folks, need some quick help. In my tmp/work directory, I have arm5te... and i686... In the bb directory for my package (lua), I see lua-5.0.2 and lua-native-5.0.2. What does the native designate? Jul 31 11:27:25 native runs in your oe-builder machine Jul 31 11:28:25 psycoman: That's what I thought. However, when I do a bitbake lua, it builds in the i686 directory first. Jul 31 11:28:52 (looks like it builds both) Jul 31 11:30:21 yeap Jul 31 11:34:07 koen: you make my eyes bleed Jul 31 11:34:28 XorA: need some bandages? Jul 31 11:34:52 koen: yeah, after that gnash .bb Jul 31 11:35:02 drmikecrowe: lua probably depends on lua-native Jul 31 11:35:33 XorA: I did say it had broken autotools :( Jul 31 11:35:47 RP: Then if i'm upgrading from 5.0.2 to 5.1.2 I need to do both? Jul 31 11:35:54 * XorA guesses he never heard of pkg-config Jul 31 11:35:56 drmikecrowe: probably Jul 31 11:36:07 RP: Thanks Jul 31 11:36:25 * RP is surprised nobody's mentioned the SRCREV "errrors" from bitbake yet Jul 31 11:36:26 koen: no wonder he was struggling to get a .bb working Jul 31 11:36:44 ERROR: SRCREV was used yet no valid SCM was found in SRC_URI Jul 31 11:36:45 NOTE: : while evaluating: Jul 31 11:36:47 ${@bb.fetch.get_srcrev(d)} Jul 31 11:36:56 koen: yes, that one :) Jul 31 11:36:57 OMG BITBAKE ERROR!!!!11!!one Jul 31 11:37:02 there you go :) Jul 31 11:37:19 its OMFG surely koen Jul 31 11:37:33 ~omfg Jul 31 11:37:34 methinks omfg is "oh my fluffy gerbil" Jul 31 11:37:42 RP: I had to report it but decided that if you left them then it is ok Jul 31 11:37:58 The fact that happens proves other bits of the bitbake core are doing really inefficient things :-( Jul 31 11:38:21 They're totally harmless but show there is a problem deep in the core :( Jul 31 11:38:24 useful errors are a good thing Jul 31 11:39:05 RP: did you see my messages to the oe-dev mailinglist yet about some worse problems? Jul 31 11:39:08 Folks, what's a diff command I can use to generate a patch for a Makefile? searching for "diff" on oe.org finds every page Jul 31 11:39:24 RP: isnt this the time the hero volunteers to put on the asbestos suit and delve into the core Jul 31 11:39:53 * XorA suddenly gets Meat Beat Manifesto in the brain Jul 31 11:40:18 XorA: I know what the fix is, I just haven't dared rewrite the emit_vars function yet Jul 31 11:41:09 koen: hmm Jul 31 11:41:35 koen: I guess this is where we need cache policy Jul 31 11:42:30 I'm going to write a script that parses bitbake -s and writes out a list of SRCREV_pn-=bar for the angstrom release Jul 31 11:42:43 to should solve most of the evil firewall builds Jul 31 11:43:10 s/to/that/ Jul 31 11:46:37 koen: I've just replied as such to the list. A script like that could be useful in general Jul 31 11:46:58 koen: As for the other issues you mention, thanks for that test, its interesting. I need to look into it ;-) Jul 31 11:47:00 03daniel 07org.oe.dev * raa90cc7b... 10/ (3 files in 3 dirs): openmoko-session2: Change mb session file so today and dialer are started Jul 31 12:03:51 Folks, is there a howto somewhere on patching? I need to patch the Makefile. I created a patch similar to how the prior version did it. The prior version patched Makefile against Makefile~make. I assumed bitbake would create Makefile~make, but apparently not. Jul 31 12:04:22 drmikecrowe: look in patches/ in the unpacked source dir Jul 31 12:04:31 drmikecrowe: and learn to use quilt Jul 31 12:05:40 XorA: When it downloads the master package, there isn't a patches/ directory. Jul 31 12:05:50 (will do on the quilt) Jul 31 12:05:59 drmikecrowe: I mean in the previouys version Jul 31 12:06:23 drmikecrowe: but the patch you want will be beside the .bb in files/ or name-version/ or name/ directories Jul 31 12:07:11 XorA: I put the make.patch in the files/ directory of the recipe. Jul 31 12:07:41 drmikecrowe: and add it to SRC_URI file://make.patch;patch=1 Jul 31 12:09:23 hi Jul 31 12:09:33 XorA: Right, did that to. Made it look just like the prior version. Jul 31 12:09:44 XorA: doh, here's the error: missing header for unified diff at line 3 of patch Jul 31 12:09:47 Let me fix that Jul 31 12:09:56 drmikecrowe: sounds like you writing your own howto there :-) Jul 31 12:10:15 LOL -- not what I started. Jul 31 12:11:15 hmm.. building openmoko-devel-image for fic-gta01/angstrom builds mesa?? strange Jul 31 12:11:24 hrw why? Jul 31 12:11:43 why builds or why strange? Jul 31 12:11:46 hrw bitbake -g Jul 31 12:11:58 woglinde: I prefer to not break that build Jul 31 12:12:19 hrw the question is if we should update to mesa 7.0 Jul 31 12:14:57 Crofton : hi ! Jul 31 12:15:11 kdrive depends on sdl which depends on mesa Jul 31 12:15:18 ok Jul 31 12:15:22 thx koen Jul 31 12:15:36 and sdl dlopens() the mesa stuff irrc Jul 31 12:15:42 and only Xsdl depends on sdl Jul 31 12:15:48 steliosk, got the email, I'll grab the papers in a bit Jul 31 12:15:52 so you won't notice anythign at runtime Jul 31 12:16:05 just need to update hal quirks to see if this laptop will suspend/resume Jul 31 12:16:54 Crofton : Great ! Got the first (ieee) paper in pdf, any chance to get the vhdl code mentioned ? or should i contact the authors ? Jul 31 12:17:07 I suspect you will need to contact the authors Jul 31 12:17:12 unless they give an url Jul 31 12:17:22 no url Jul 31 12:17:35 google the authors :) Jul 31 12:17:55 * steliosk wonders on how many secret services databases he will be added by emailing people in Iran Jul 31 12:18:06 as a lapsed IEEE member, I must insist you join the IEEE :) Jul 31 12:18:11 steliosk: just dont go on holiday :-) Jul 31 12:18:17 crofton *g* Jul 31 12:18:18 * Crofton closes his eyes Jul 31 12:18:36 Crofton : I was but stopped Jul 31 12:18:48 heh Jul 31 12:18:58 IEEE is too snooty for me Jul 31 12:19:05 Crofton : I was also an acn member (sigraph etc etc) Jul 31 12:19:08 they have "standards" Jul 31 12:19:13 gets expensive Jul 31 12:19:39 itu? Jul 31 12:20:05 yeah i "invested" the money in other "projects". marriage,children etc :) Jul 31 12:22:23 * psycoman wonder if can make a iwmmxt imagee :) Jul 31 12:23:32 is possible make a iwmmxt image with easy steps ? /hmm Jul 31 12:23:40 yes Jul 31 12:24:07 woglinde: can you tell me ? Jul 31 12:24:46 bbl Jul 31 12:25:55 org.openembedded.dev/conf/machine/include/tune-iwmmxt.conf Jul 31 12:26:43 ah.. include/tune-* reminds me one thing Jul 31 12:26:54 hrw which one? Jul 31 12:27:01 RP: bitbake -s | grep svnr | awk '{print "SRCREV_pn-" $1 "=" $2}' | sed -e s:\+:\=:g -e s:svnr::g| awk -F "=" '{print $1 "=\"" $3}' | awk -F "-r" '{print $1"\""}' Jul 31 12:27:24 koen: neat :) Jul 31 12:27:33 a sample: http://rafb.net/p/xmeptz73.html Jul 31 12:28:43 woglinde: i put it as required in spitz machine to build every thing optimized but is break locate glib things Jul 31 12:29:58 woglinde: PACKAGE_ARCHS Jul 31 12:30:24 psycoman: I think you have to add PACKAGE_ARCH_glibc="armv5te" somewhere or disable qemu binary locale generation. qemu doesn't support iwmmxt. Jul 31 12:30:27 now they are set in machine configs and I think about moving them into tune-* files Jul 31 12:31:03 hrw: please don't Jul 31 12:31:23 hrw: that would mean e.g. spitz owners can't install iwmmxt files Jul 31 12:31:23 * woglinde will going now scrub the kitchen Jul 31 12:31:27 I will not without rfc etc Jul 31 12:31:42 since iwmmxt isn't in PACKAGE_ARCH when building Jul 31 12:31:54 pH5: well qemu accept iwmmxt now in cvs :) but i dont know good knowleged in bitbake system to make a 0.9.0cvs version Jul 31 12:32:19 koen: you can always do require ...\nPACKAGE_ARCHS += "iwmmxt" Jul 31 12:32:46 that doesn't help for all the builds out there Jul 31 12:33:26 PACKAGE_EXTRA_ARCHS should list everything a machine supports besides 'all', '$MACHINE' and '$TARGET_ARCH', not a subset Jul 31 12:33:32 pH5: Get a better version of qemu then ;-) Jul 31 12:33:46 it would be better to fix the machine config files than to add it to the tune files Jul 31 12:34:53 RP: got it, I'm not up to date :) but I don't intend to build complete iwmmxt images again any time soon ;) Jul 31 12:34:59 PACKAGE_EXTRA_ARCHS should list all the archs a system can run, I'm certain of that... Jul 31 12:35:52 tune-iwmmxt would have PACKAGE_EXTRA_ARCHS = "all noarch any arm armv4t armv5te iwmmxt" Jul 31 12:36:01 tune-xscale would have PACKAGE_EXTRA_ARCHS = "all noarch any arm armv4t armv5te" Jul 31 12:36:01 etc Jul 31 12:36:10 that won't work Jul 31 12:36:21 No, we can't do that Jul 31 12:36:31 ok - I never tested that so it was just a thought Jul 31 12:36:58 that would mean a c700 can support iwmmxt by doing 'require tune-iwmmxt' Jul 31 12:37:02 which isn't true Jul 31 12:37:17 let's only have one place where we set PACKAGE_EXTRA_ARCH Jul 31 12:37:21 S Jul 31 12:37:23 whats any for? Jul 31 12:37:25 now it can support iwmmxt by adding 'iwmmxt' in PACKAGE_EXTRA_ARCHS Jul 31 12:37:34 woglinde: shell scripts Jul 31 12:37:40 rp ah okay thanks Jul 31 12:37:48 woglinde: leftovers - same as noarch Jul 31 12:37:57 'any' and 'all' and 'noarch' are basically the same Jul 31 12:43:57 so add tune-iwmmxt to spitz.conf, is not a good way to try build a iwmmxt angstrom image right ? Jul 31 12:45:13 psycoman: It should work if you have a recent qemu Jul 31 12:46:57 RP: i know but i dont understand some bitbake basics, and right now i dotn have skill to change it to 0.9.0 Jul 31 12:47:03 will there be a release of 1.8 for the svn-fetcher fix soon? Jul 31 12:47:56 psycoman: Pinch the files from poky ;-) Jul 31 12:48:06 and now i get: :'module' object has no attribute 'get_srcrev' while evaluating:${@bb.fetch.get_srcrev(d)} Jul 31 12:48:54 psycoman: Just ignore that for now Jul 31 12:49:35 but it stops the build Jul 31 12:49:40 ctrl-c ? Jul 31 12:49:58 It shouldn't stop the build :/ Jul 31 12:50:27 it stoped, olny continue if i hit ctrl-c Jul 31 12:51:27 RP: poky have a mtn branch ? Jul 31 12:51:45 psycoman: no, just svn Jul 31 12:53:48 RP: just a queistion about BB, i many ..b files in qemu, how bb can select 0.8.2+cvs and not 0.9 ? Jul 31 12:54:27 psycoman: PREFERRED_VERSION_qemu = "0.9" Jul 31 12:55:24 in agnstrom.conf not set a version, and it always get 0.8, the question is who it select 0.8 and not 0.9 Jul 31 12:56:08 maybee 0.9 has PREFERRED_bla = "-1" Jul 31 12:56:56 woglinde: PREFERRED_VERSION trumps DEFAULT_PREFERENCE Jul 31 12:57:25 well, sorry my bad english, the question could be, how bitbake choose a version if no version preferred is seted Jul 31 12:57:26 aehm PREFERENCE I meant sorry Jul 31 12:57:28 set'd Jul 31 12:58:36 go bb only 100 packages to parse Jul 31 12:58:45 psycoman: It goes by the PREFERENCE values Jul 31 12:59:34 hmmm Jul 31 13:00:32 R:P where it is se`d ? Jul 31 13:00:52 s/R:P/RP Jul 31 13:03:15 psycoman: PREFERENCE defaults to "1" iirc Jul 31 13:09:58 03rpurdie * r859 10/ (4 files in 4 dirs): data.emit_var() - only call getVar if we need the variable Jul 31 13:11:04 RP: why not move que emu to the qemu*.bb files of poky ? Jul 31 13:11:33 psycoman: Mainly lack of time to deal with any bugs that might cause Jul 31 13:11:52 right. Jul 31 13:12:15 i newbie question, how can i help oe/distros ? Jul 31 13:13:08 look at bugzilla and provide patches Jul 31 13:14:26 i make some patch, im wating the commit now. Jul 31 13:14:53 is dificult see bugtrack many bugs have benn fixed but still open and pending commit Jul 31 13:17:14 pyscoman hm but you cant close the bugs in the bugtracker? Jul 31 13:18:34 how can i check if they are commited ? Jul 31 13:20:03 ah spotted gettext-bug Jul 31 13:20:07 ah spotted gettext-native Jul 31 13:20:12 needs zlib-native Jul 31 13:21:06 hms Jul 31 13:21:16 has to do with my reported bug Jul 31 13:21:23 !oebug 2679 Jul 31 13:21:25 * * Bug 2679, Status: NEW, Created: 2007-07-24 04:54 Jul 31 13:21:26 * * heinold(AT)inf.fu-berlin.de: zlib.h from debian sneaks in when compiling libxml2-native Jul 31 13:21:27 * * http://bugs.openembedded.org/show_bug.cgi?id=2679 Jul 31 13:21:53 seems gettext-native took all the time the hostsystem zlib.h Jul 31 13:22:15 not the oe one Jul 31 13:24:55 03pH5 07org.oe.dev * r15d74dae... 10/ (5 files in 3 dirs): Jul 31 13:24:55 libgsmd: update gsmd patches again Jul 31 13:24:55 - talloc mlbuf in struct gsmd instead of having a static global variable Jul 31 13:24:55 - adapt sms-hacks.patch Jul 31 13:25:46 talloc? Jul 31 13:27:09 woglinde: gsmd has a copy&paste version of the samba trivial allocation library built-in. Jul 31 13:27:21 ah Jul 31 13:27:29 I remember darkly now Jul 31 13:37:29 RP: could you try changing the PV of libgsmd to use SRCREV and then doing a bitbake -b ? Jul 31 13:55:20 args Jul 31 13:55:38 file-native made the error not gettext Jul 31 13:56:07 koen: Why do I suspect I don't want to know about this :} Jul 31 13:56:24 I suspect the ;proto=http breaks it Jul 31 13:56:40 steliosk, got IEEE paper Jul 31 13:57:06 eye tripple e Jul 31 13:58:12 ~curse broken svn certs Jul 31 13:58:13 May you be reincarnated as a Windows XP administrator, broken svn certs ! Jul 31 14:02:13 koen: svn info doesn't seem to work with http urls :/ Jul 31 14:02:25 cvs, but better! Jul 31 14:03:39 Mental :-( Jul 31 14:04:33 koen: ah, it depends on your version of svn. It works on one of my boxes and not on the other Jul 31 14:13:20 1.4.4 from debian has working svn info for http Jul 31 14:13:58 svn, version 1.4.2 (r22196) hasn't Jul 31 14:15:00 1.4.3 works too Jul 31 14:15:24 * koen installs 1.4.4 Jul 31 14:16:00 Time for svn-native? :) Jul 31 14:16:13 hmmm Jul 31 14:16:19 still raise MalformedUrl(url) with 1.4.4 Jul 31 14:18:48 (debian) Jul 31 14:19:20 woglinde: ping Jul 31 14:19:35 hi rschuster Jul 31 14:19:59 woglinde: hi. can you send me the patch that adds --with-target-classpath-classes to cacao? Jul 31 14:20:14 rschuster yes I will send you now all patches Jul 31 14:20:37 is a zip file okay? Jul 31 14:20:41 woglinde: yep Jul 31 14:21:03 woglinde: have you asked twisti for inclusion of the target-classpath thingß Jul 31 14:21:10 s/ß/? Jul 31 14:21:16 no Jul 31 14:21:21 not yet Jul 31 14:30:54 hi there Jul 31 14:32:00 woglinde: I have that weird problem that classpath-tool (=> tools.zip) is on my rootfs but classpath-common (glibj.zip) is not. any idea what I could have made wrong? Jul 31 14:33:43 pH5: I lart you for 31 15:29 < CIA-3> libgsmd: update gsmd patches again Jul 31 14:33:54 how can I track the dependencies of a particular package? eg. I've build the bootstrap-image for angstrom distribution but don't won't a kernel with wifi but the image still depends on madwifi-ng (what naturally fails if kernel doesn't support it) Jul 31 14:33:58 pH5: Patch mlbuf-in-gsmd-struct.patch does not apply (enforce with -f) Jul 31 14:35:03 rschuster hm did you use the oe thinks? Jul 31 14:35:07 zarath: bitbake -g can sometimes hint about the dependency tree Jul 31 14:35:27 besides classpath-minimal the other stuff is broken Jul 31 14:35:37 I will fix it the next time Jul 31 14:36:59 ~change 32 gbp to eur Jul 31 14:37:05 i got some error trying that Jul 31 14:37:28 hm Jul 31 14:37:29 *g* Jul 31 14:37:29 ah, ok, but still no clue why there is still a dependency... Jul 31 14:38:56 woglinde: I am using the classpath-minimal (0.95) you gave me but changed it a bit (e.g. classpath-tools need not only the tools.zip but the tool binaries itself, I suppose ) Jul 31 14:39:34 rschuster hm I decision was that classpath-minimal is only for compiling Jul 31 14:39:37 my Jul 31 14:39:47 no packages Jul 31 14:40:11 and as I said the other classpath stuff is broken Jul 31 14:40:30 classpath for example cant be compiled since 0.93 with jikes Jul 31 14:41:11 pH5: please do TEST build before commiting Jul 31 14:41:21 he zecke Jul 31 14:41:55 woglinde: I thought this was the case for classpath-minimal-native (which is ok) Jul 31 14:42:15 RP: is bitbake -c rebuild bootstrap-image sufficient to rebuild the dependency tree? Jul 31 14:42:36 rschuster no I did for both if class-path is really needed on the device it should get the full version Jul 31 14:42:37 zarath: It depends which bit we're talking about Jul 31 14:42:58 zarath: If you altered something the changed one of the packages, you need to rebuild that package Jul 31 14:43:15 rschuster and the cacao_bb I send you will not go in, it needs fixes too and was only a beginning Jul 31 14:44:44 koen: http://rafb.net/p/EX660e67.html Jul 31 14:44:57 woglinde: yep. but I think I found the issue. PACKAGES was listing ${PBN}-examples but that one is not defined Jul 31 14:45:40 woglinde: I can live with the fact that cacao cannot go in until they fix all their crosscompiling issues. there is still jamvm Jul 31 14:46:07 woglinde: classpath-minimal has its purpose. Eg on routers you probably want no gtk peers and such Jul 31 14:46:25 RP: both work for me Jul 31 14:46:36 Crofton : nice. one down :) Jul 31 14:46:59 koen: both work now for me Jul 31 14:47:08 * RP wonders what on earth this machine is playing at Jul 31 14:47:10 yeah Jul 31 14:47:39 RP: I now get http://ewi546.ewi.utwente.nl/tmp/gsm-svn-error.txt Jul 31 14:47:55 RP: could it be the openmoko proxy acting up? Jul 31 14:48:24 koen: Its entirely possible Jul 31 14:49:24 steliosk, I am beginning to suspect the ""ACM" link referers to a print only journal Jul 31 14:49:33 koen: I'll check a few more things... Jul 31 14:49:39 maybe not Jul 31 14:49:42 morning Jul 31 14:50:26 steliosk, got it Jul 31 14:50:58 with the new openmoko theme: http://www.youtube.com/watch?v=sOnvyXowa1U Jul 31 14:51:10 Crofton|laptop : cool ! Thanks for your rime Jul 31 14:51:19 s/rime/time Jul 31 14:51:52 http://www.springerlink.com/content/q41385g6696n/?p=d3572a9e5775495cb9ffb430ec3d877f&pi=13 Jul 31 14:52:21 koen: your neo is on 100dpi? Jul 31 14:52:31 no, 285 Jul 31 14:52:38 I changed that this morning Jul 31 14:52:52 rschuster dont know why I should run a java programm onto routers Jul 31 14:53:00 ~curse ph5 badly Jul 31 14:53:01 May you be reincarnated as a Windows XP administrator, ph5 badly ! Jul 31 14:53:09 koen: change font size to 4 instead of 5 Jul 31 14:53:59 * koen wants to change the dialer screen fontsize as well Jul 31 14:56:01 where i set PACKAGE_ARCH_glibc="armv5te" for force glibc to not use default iwmmxt in mu build ? Jul 31 14:56:24 Crofton|laptop : the springer link takes me to a page where the paper costs 32$.... Jul 31 14:56:28 setting PACKAGE_ARCH won't help, since it will still use the iwmmxt cflags Jul 31 14:56:50 heh Jul 31 14:56:54 free for me :) Jul 31 14:57:01 so, bye bye iwmmxt build until qemu 0.9.1 is out =( Jul 31 14:57:02 must be working from IP's Jul 31 14:57:09 FSCKING SHIT Jul 31 14:57:10 I need to book mark that page Jul 31 14:58:48 woglinde: pircbot! Jul 31 14:59:07 woglinde: man-di did that on the linksys routers :) Jul 31 14:59:29 rschuster hms Jul 31 15:00:03 okay I will think over how to get classpath in proper shape Jul 31 15:01:31 woglinde: yeah but if it makes to much trouble splitting classpath into pieces then let it out and someone can always add it when the need arises Jul 31 15:02:06 koen: I can fix that problem but we then hit recursion overload. To work out SRCREV we parse SRC_URI. SRC_URI has file:// paths in so FILESDIR and FILESPATH are evaluated. FILESDIR and FILESPATH can contain PV. Loop :/ Jul 31 15:02:24 hmm Jul 31 15:02:46 RP: this one is going to be fun as well: http://www.openembedded.org/repo/org.openembedded.dev/packages/webkit/webkit_svn.bb Jul 31 15:03:24 koen: That will be a good test for SRCREV_FORMAT :) Jul 31 15:05:04 it will probably have one SRCREV as it is one tree Jul 31 15:05:04 I'm going to have to split the urldata init process into two stages :-( Jul 31 15:05:27 hrw: It will take the revisions for each directory Jul 31 15:05:57 Foo None Jul 31 15:05:57 NOTE: Psyco JIT Compiler (http://psyco.sf.net) not available. Install it to increase performance. Jul 31 15:06:00 woglinde: could you have a look at http://www.openembedded.org/monotonephrasebook ? Jul 31 15:06:06 what for is that 'Foo None'? Jul 31 15:06:10 woglinde: I will work on classpath-minimal and classpath today. if you want I can send you my bb files. I hope that would save us some double work. Jul 31 15:06:19 koen yes Jul 31 15:06:20 woglinde: your last 2 commit messages differ from our usual format Jul 31 15:06:47 stefan_schmidt: could you have a look at http://www.openembedded.org/monotonephrasebook, your last 3 commit messages differ from our usual format Jul 31 15:07:26 alphaone: your last 2 commit messages are following our usual format :) Jul 31 15:07:37 koen: ah, that one I was searching for. I take a look before the next commits. Jul 31 15:07:46 hrw: hmm. I don't remember adding anything like that... Jul 31 15:07:54 stefan_schmidt: basically "recipe name: changes" Jul 31 15:08:07 koen: ok, that's easy. :) Jul 31 15:08:10 koen: I hope so. :-) Jul 31 15:08:30 koen: Previous to that they didn't? Jul 31 15:08:52 alphaone: no idea, the changelog generator only picked up two for you :) Jul 31 15:09:05 stefan_schmidt: you can see the difference at http://rafb.net/p/DZZd8I27.html Jul 31 15:09:46 koen: Oh, changelog generator. Nice :-) Jul 31 15:09:47 Crofton|laptop : got the emails Jul 31 15:09:53 koen: yeah, looks nice. Screams for a small wrapper to automate this. Jul 31 15:11:36 03rpurdie * r860 10/ (4 files in 2 dirs): Various fetcher bugfixes Jul 31 15:11:54 alphaone: it's in contrib/mtn2cl Jul 31 15:12:04 we used to have issues with people running crawlers on the IEEE site Jul 31 15:12:22 this really pissed off the library and the IEEE Jul 31 15:17:55 Crofton|laptop : I think i got the 2.6.19 "vanilla" kernel to build the xilinx board Jul 31 15:18:08 * steliosk keeps fingers crossed Jul 31 15:18:29 cool Jul 31 15:18:57 getting the correct EDK/bsp/kernel stuff can be a nightmare :( Jul 31 15:19:02 RP: as you guessed: RuntimeError: maximum recursion depth exceeded, after your last commit Jul 31 15:19:20 koen: Can you tell roughly how many packages cause that? Jul 31 15:19:39 I guess anything with a svn + file url will :-( Jul 31 15:19:45 yeah Jul 31 15:20:29 koen: I need to think about this one a bit ;-) Jul 31 15:21:15 RP: http://lists.linuxtogo.org/pipermail/openembedded-devel/2007-July/002623.html would eliminate a few of those :) Jul 31 15:22:03 koen: Don't you mean add some? :) Jul 31 15:23:24 no, since only their patches are in svn Jul 31 15:23:31 that script puts them in mtn Jul 31 15:23:51 koen: Thereby adding file:// urls, not removing them ;-) Jul 31 15:24:11 replacing svn:// with file:// Jul 31 15:24:21 ah, but the base package is source based, right :) Jul 31 15:24:28 koen, what are the chances of using openmoko to create an encrypted voice phone Jul 31 15:25:00 Crofton|laptop: analog encryption is icky Jul 31 15:25:15 so we only have access to the analog data? Jul 31 15:25:30 http://pastebin.ca/640641 Jul 31 15:25:35 my friends comments Jul 31 15:25:36 afaik it puts 'analog' data into the gsm module Jul 31 15:26:05 Crofton|home: ask XorA or lrg about how the modem is connected to the codec Jul 31 15:26:08 I'd ask XorA Jul 31 15:26:12 of course, then the only people he could talk to would be other paranoid openmoko users :) Jul 31 15:26:24 will doo Jul 31 15:28:32 Crofton|home: analogue lines from codec go to/from modem Jul 31 15:28:41 Crofton|home: the description is in the wiki somewhere Jul 31 15:29:16 You can probably do one way encryption Jul 31 15:29:27 bye Jul 31 15:31:46 when will someone implement lzip based transmission for gsm? Jul 31 15:33:42 hrw|gone: I'm very sorry, it applies cleanly against svn head. It seems harald's latest commits were today, not yesterday. Jul 31 15:34:06 that curse was cruel, though. Jul 31 15:41:58 pH5: curse is standard Jul 31 15:43:57 bye Jul 31 15:44:07 hrw|gone: ok :) bye Jul 31 15:46:16 hm. I tried http://en.pastebin.ca/640664 and get recursion depth exceeded RuntimeErrors while evaluating "${@bb.fetch.get_srcrev(d)}" Jul 31 15:46:39 pH5: that's what RP and I have been taling about :( Jul 31 15:47:10 pH5: Pick an svn package which doesn't have file:// urls Jul 31 15:47:28 [17:02] RP: koen: I can fix that problem but we then hit recursion overload. To work out SRCREV we parse SRC_URI. SRC_URI has file:// paths in so FILESDIR and FILESPATH are evaluated. FILESDIR and FILESPATH can contain PV. Loop Jul 31 15:47:28 [17:02] koen: hmm Jul 31 15:47:38 This is resisting all my attempts to fix it :-( Jul 31 15:48:52 oops Jul 31 15:50:21 mickeyl: good morning! Jul 31 15:50:47 yawn Jul 31 15:51:01 RP: fixed some buglets yesterday nite Jul 31 15:51:04 morning mickeyl Jul 31 15:51:12 RP: i'm using it now. it works cool. Jul 31 15:51:14 mickeyl: I saw, thanks. I applied them to the 1.8 branch Jul 31 15:51:38 mickeyl: We've found a few gremlins since then. Some are fixed, some aren't :/ Jul 31 15:51:55 RP: anything serious? Jul 31 15:52:06 does bitbake have an 'unset' function (e.g. delvar)/ Jul 31 15:52:09 ? Jul 31 15:52:23 afaik not Jul 31 15:52:37 mickeyl: The one which breaks any package with both an svn:// and a file:// in the SRC_URI is Jul 31 15:52:57 koen: There is a delVar function call in bb.data if thats what you mean Jul 31 15:53:17 RP: ah well, this shouldn't be used anyways Jul 31 15:53:20 [just kidding] Jul 31 15:53:46 mickeyl: It causes an "recursion depth exceeded" :} Jul 31 15:53:52 eek, nasty Jul 31 15:54:08 RP: didn't use trunk for a while. i noticed you made it really quiet Jul 31 15:54:14 i kind of miss some NOTEs Jul 31 15:54:34 i.e. the one that gives me the filename of the ipk it packaged Jul 31 15:54:45 mickeyl: Its due to what I did with the logging. We need to tweak it to look more like the old one Jul 31 15:54:52 mickeyl: The logs are all there, just in files Jul 31 15:56:19 It probably generates too many logs files since it outputs one for every anonymous function, most of them empty :/ Jul 31 15:57:17 mickeyl: the 1.8 branch has all these fetcher changes btw, I want this to work with 1.8 Jul 31 15:57:27 RP: can't you just cut all file:// URIs from SRC_URI before parsing it in get_srcrev? Jul 31 15:57:47 RP: nod Jul 31 15:57:51 pH5: Yes and no. It breaks my caching algorithms Jul 31 15:58:40 RP: your patch already had some good effects to OE. found some invalid SVN URIs at once :) Jul 31 15:59:08 my sed script also picked up some :) Jul 31 15:59:14 mickeyl: :) Jul 31 15:59:36 cvs:// recipes that set 0.0+svn and stuff like that Jul 31 15:59:38 e.g. http://svn.opensync.org/plugins/syncml-client/plugin/tags/release-0.22+svn20070731 (invalid URI) Jul 31 15:59:47 mickeyl: btw, what is 'srvn'? Jul 31 16:00:08 a typo. fixed locally here and will end up in the next commit Jul 31 16:00:23 mickeyl: nice :} Jul 31 16:00:46 mickeyl: I think poky might even have that one actually, I saw something like that earlier! Jul 31 16:05:39 mickeyl: FWIW, there was a bug where we were checking a lot more svn heads than we should have been. That is fixed now (but means you won't see quite as manu invalid errors) Jul 31 16:06:44 bye bye guys, cya Jul 31 16:06:51 03woglinde2 07org.oe.dev * r142fee5c... 10/ (1 packages/gettext/gettext-native_0.14.1.bb): Jul 31 16:06:51 -spotted by renaming my host zlib.h to see why bug 2679 happens Jul 31 16:06:51 -I think it was never spotted before, because all developers have Jul 31 16:06:51 a host-zlib.h installed Jul 31 16:06:56 03woglinde2 07org.oe.dev * r55cb6151... 10/ (1 packages/gettext/gettext-native_0.14.1.bb): Jul 31 16:06:56 I am to stupid some times it not comes from gettext, it comes Jul 31 16:06:56 from file-native. Jul 31 16:07:03 03woglinde2 07org.oe.dev * rf8864781... 10/ (1 packages/file/file-native_4.21.bb): file-native/compress.c needs zlib.h so it must depend on zlib-native Jul 31 16:07:07 03koen 07org.oe.dev * r21335221... 10/ (1 contrib/mtn2cl/mtn2cl.sh): contrib/mtn2cl: generate dated changelogs Jul 31 16:07:12 03rpurdie 07org.oe.dev * r8961e177... 10/ (1 classes/base.bbclass conf/bitbake.conf): base.bbclass: Remove warnings for older bitbake users Jul 31 16:08:07 03koen 07org.oe.dev * r56cb2223... 10/ (1 packages/uboot/u-boot-mkimage-gta01-native_oe.bb): u-boot-mkimage-gta01-native_oe.bb: remove conflicting patch Jul 31 16:08:11 woglinde: :) Jul 31 16:08:12 03koen 07org.oe.dev * r449760b0... 10/ (3 files in 2 dirs): edje*: these packages are from cvs, not svn Jul 31 16:08:16 03koen 07org.oe.dev * rfcc8ed3e... 10/ (3 files in 2 dirs): gtkhtml2: fetch from svn Jul 31 16:08:22 03koen 07org.oe.dev * re3756fff... 10/ (1 packages/gaim/pidgin_2.1.0.bb): pidgin: update to 2.1.0 Jul 31 16:08:36 woglinde: read the commit message policy Jul 31 16:08:51 woglinde: if you do stupid things, disapprove yourself, don't create a commit to fix it up Jul 31 16:09:12 zecke: openmoko-feedreader2 doesn't compile because it seems to look for libmokoui in the wrong place Jul 31 16:09:43 koen: hey koen. how are you? Jul 31 16:09:54 sleepy, how are you? Jul 31 16:10:09 * zecke considers doing a rip off of microsoft's slogan "What are you going todo today?" Jul 31 16:10:20 koen: sleepy as well, I'm no longer a homeless jerk, just a jerk Jul 31 16:10:43 zecke: good question? Jul 31 16:10:43 public class jerk implements Homeless ;) Jul 31 16:10:54 s/?// Jul 31 16:13:51 * zecke just compiled Qtopia4.2 for a 640x480 resolution and it is butt ugly Jul 31 16:13:55 how do task-base-wifi and task-base-bt correspond to ADD_WIFI and ADD_BT? In other words, what is task-base-extended for? Jul 31 16:14:00 zecke: :D Jul 31 16:14:09 zecke: aren't they prerendering svg's nowadays? Jul 31 16:14:31 mickeyl: not in the released GPL version or not by default Jul 31 16:14:33 mickeyl: if your devices has pcmcia or usbhost or pci it will include wifi and bt drivers Jul 31 16:14:43 koen: I need to port feedreader2 Jul 31 16:15:00 zecke: yes, please! Jul 31 16:15:19 zecke: somewhere later tonite i can give you an image to get an idea of the new style Jul 31 16:15:34 mickeyl: I tried the one from chris but that doesn't work Jul 31 16:15:49 zecke: odd. it does for me. perhaps you need nand erase manually. Jul 31 16:16:00 dfu-util seems to still have problems with flashing smaller images Jul 31 16:16:07 (when larger ones are installed) Jul 31 16:16:22 that's a long standing bug Jul 31 16:16:25 (see bugzilla) Jul 31 16:19:39 nand erase == factory reset? Jul 31 16:20:01 no. go to uboot menu Jul 31 16:20:10 log in via usb serial emulation Jul 31 16:20:13 (ttyACM0) Jul 31 16:20:16 and then issue Jul 31 16:20:20 nand erase rootfs Jul 31 16:20:26 then, proceed w/ dfu upload Jul 31 16:20:28 zecke: http://people.openmoko.org/alphaone/images/ Jul 31 16:20:52 RP: what task are you using to get all those puzzle games into poky? Jul 31 16:21:09 mickeyl: Its just the puzzles package Jul 31 16:21:18 ah. cool, thanks Jul 31 16:21:44 mickeyl: We don't actually package them separately since they're all quite small Jul 31 16:22:10 mickeyl: There is an OH customised version iain made too (oh-puzzles) Jul 31 16:23:30 i see Jul 31 16:25:14 hmm we should add that Jul 31 16:27:54 with fancy scrolling and puzzles we won't get any work done ;) Jul 31 16:28:14 hehe Jul 31 16:28:41 did anyone try robtaylor's patch for the scrolling? Jul 31 16:28:48 thank god untangle redraws on mouse move events which makes it way unusable with high report rate touchscreens on slow phones. Jul 31 16:29:32 patch? where? Jul 31 16:29:43 pH5: 'untangle'? Jul 31 16:29:53 mickeyl: http://codethink.co.uk/~robtaylor/fingerscroll-more-physics.patch Jul 31 16:29:59 from http://chrislord.net/blog/fancy-scrolling-widgets-pt2.enlighten Jul 31 16:32:06 has anyone tried it on a c7x0 with Ximageon? Jul 31 16:34:21 hmm interesting Jul 31 16:34:48 mickeyl: it's one of the puzzles Jul 31 16:34:49 http://www.chiark.greenend.org.uk/~sgtatham/puzzles/ Jul 31 16:36:01 mmh, is it time to start iwmmxt / dma blitting routines for pxafb? Jul 31 16:36:55 pH5: finishing the sram driver would make a better improvement I think Jul 31 16:55:36 koen: good point. I gave up on the current sram driver temporarily when I noticed that it can't be mmapped easily. Jul 31 16:56:18 I guess I should try to understand the linux vmm before looking at this again. Jul 31 16:59:53 is it valid if a package RDEPENDS on one of its subpackages (eg. foo_0.95.bb: RDEPEND="${PN}-common" ... PACKAGES="${PN} ${PN}-common")? Jul 31 17:00:35 rschuster: as long as you don#t create circular deps?! Jul 31 17:02:59 zecke: sure :) Jul 31 17:21:47 03rpurdie * r861 10/ (2 files in 2 dirs): Work around fetcher recursion problems for now Jul 31 17:21:55 koen: next problem? :) Jul 31 17:25:12 hmm, are the old wiki pages from the start of oe still around, the reasoning behind the original selection of portage as the starting point? Jul 31 17:25:41 * RP doesn't know Jul 31 17:26:02 kergoth: I might have a copy of that at home Jul 31 17:26:33 kergoth: Isn't there an archive of historic websites somewhere? Jul 31 17:26:40 www.archive.org? Jul 31 17:27:26 hmm good idea, will check Jul 31 17:27:50 rp with latest svn bb the first run parsing packages now lasts 5 minute or so Jul 31 17:28:05 woglinde: How long did it used to take? Jul 31 17:28:21 RP: libgsmd fetches now Jul 31 17:28:27 rp 1-2 minutes Jul 31 17:28:35 woglinde: and is that the latest OE.dev? Jul 31 17:28:41 rp yes Jul 31 17:29:02 RP: if it tries to connect to remote sites, I loose since my cablemodem is sucky Jul 31 17:29:09 woglinde: How many SRCREVs have you set? Jul 31 17:29:16 ah Jul 31 17:29:18 hm Jul 31 17:29:40 i was at 85x now 86 Jul 31 17:29:42 x Jul 31 17:29:46 let me try again Jul 31 17:30:05 yeah much better now Jul 31 17:30:26 okay Jul 31 17:30:53 supper now Jul 31 17:30:55 till later Jul 31 17:36:26 koen: :/ Jul 31 17:36:44 koen: If you get a chance, I'd be interested to know if any of the packages outright fail on parsing now... Jul 31 17:37:37 khem: zecke is our external-toolchain master, so if you still have problems with it, ask him :) Jul 31 17:38:05 * khem waves at zecke Jul 31 17:38:44 koen, I tried bitbake task-base Jul 31 17:38:48 koen, Jul 31 17:38:58 it tried to build glibc 2.5 Jul 31 17:39:05 and failed while doing locales Jul 31 17:39:22 * RP back tomorrow now -> afk Jul 31 17:39:41 I have chosen angstrom and I read that you do qemu stuff to generate it Jul 31 17:40:07 yes, since mr drepper decided everyone has at least 128MB of ram Jul 31 17:40:16 localegen eats ram like candy Jul 31 17:40:21 true Jul 31 17:40:29 it didn't use to Jul 31 17:40:34 with eglibc you can generate locales in cross environment Jul 31 17:40:43 nice Jul 31 17:41:20 * koen -> tv now Jul 31 17:41:21 even when I chose to have external toolchain why is it building glibc ? Jul 31 17:42:02 NOTE: generating locale es_NI (UTF-8) Jul 31 17:42:02 NOTE: Task failed: localedef returned an error Jul 31 17:51:58 Crofton : NOTE: package linux-xilinx-ml403-2.6.19: completed Jul 31 18:18:28 hi all Jul 31 18:30:03 re Jul 31 18:35:16 hi mrdata Jul 31 18:35:18 hi florian Jul 31 18:35:24 woglinde: hi Jul 31 18:36:34 woglinde: with redboot, rootfs over nfs is working for simpad now (with ne2k) Jul 31 18:36:43 nice Jul 31 18:38:06 Slims had done a good job Jul 31 19:47:02  Jul 31 19:58:36 03florian 07org.oe.dev * r15b74f4a... 10/ (3 files in 2 dirs): starling: Update svn version number and use esd instead of some pulseaudio which does not seem to be available. Jul 31 20:28:34 koen: You already did some work on getting webkit build. Any local changes or is it broken for you too atm? Jul 31 20:31:33 re Jul 31 20:31:52 re hr Jul 31 20:31:53 w Jul 31 20:31:57 hms Jul 31 20:32:07 * woglinde needs something like virtual/qt Jul 31 20:32:07 Hello Jul 31 20:32:17 woglinde: virtual/qt? Jul 31 20:32:20 what for? Jul 31 20:32:22 hi B_Lizzard Jul 31 20:32:30 hrw midpath Jul 31 20:32:35 it can run both Jul 31 20:32:47 qte and qt-x11 Jul 31 20:37:56 fscking svn Jul 31 20:38:59 hrw anymind we should let provide it the qt packages? Jul 31 20:42:04 is there any other app which use this virtual? Jul 31 20:42:28 woglinde: and if midpath can run qt or qte then how it choose it? Jul 31 20:44:17 hrw its only wraps some java calls to qt thinks Jul 31 20:44:51 hrw I dont know no other package doing this Jul 31 20:44:58 but somehow it checks which one is available Jul 31 20:45:13 hm no it only use the fucntion Jul 31 20:45:27 and for now qte and qt-x11 provides the same functions Jul 31 20:45:55 you can set COMPILE flaf -DQWS Jul 31 20:46:01 than it uses qte Jul 31 20:46:05 other wise X11 Jul 31 20:46:13 and it does not link to any of them? Jul 31 20:46:59 for now you have to do this manually Jul 31 20:47:09 midpath has no autotools support yet Jul 31 20:47:18 but I will change this in the future Jul 31 20:47:33 but anyway than you have to provide a virtual/qt too Jul 31 20:47:46 to set up configure correct Jul 31 20:53:39 so does it link to qt or qte during build or not? Jul 31 20:58:27 yes it link Jul 31 20:58:31 sorry Jul 31 20:59:42 so if it link then create midpath-qt.bb midpath-qte.bb so we will be able to build both Jul 31 21:01:12 hm Jul 31 21:04:07 hrw and let the distribution/image choose which to use? Jul 31 21:04:53 yep Jul 31 21:33:56 3A Jul 31 21:33:58 bye Jul 31 21:35:02 hey mickey Jul 31 21:36:21 I have to alter my earlier statement of the but-uglynes of Qtopia. It just has working DPI support. It remains ugly though :) Jul 31 21:36:46 ugly? Jul 31 21:36:58 woglinde: the out of the box GPL version Jul 31 21:37:07 zecke: did you see the 4.3 screenies? Jul 31 21:37:32 much better default theme Jul 31 21:37:37 java is mor ugly Jul 31 21:37:37 ljp: yes, it is progress :) Jul 31 21:37:56 woglinde: java was born to be ugly, Qt(opia) wasn't Jul 31 21:38:01 although qtopia is getting rather bloated Jul 31 21:38:02 * mickeyl chuckles Jul 31 22:00:24 mickeyl, hello Jul 31 22:00:53 mickeyl, what would be a good target to start with ? Jul 31 22:01:16 I want to be able to boot into the system and do some ssh and stuff Jul 31 22:01:26 sounds like a job for angstrom-console-image Jul 31 22:01:31 + DISTRO=angstrom-2007.1 Jul 31 22:02:46 yeah distro is same Jul 31 22:04:08 oh man it fails in glibc Jul 31 22:04:36 I am having glibc external so how can I ask it not to build glibc when using extternal toolchain Jul 31 22:05:13 hmm. you might have luck using ASSUME_PROVIDED = "" Jul 31 22:05:20 but you're on your own setting the paths and stuff then Jul 31 22:05:34 i don't think many people use OE with external toolchains Jul 31 22:05:40 if in doubt, ask on the mailing list Jul 31 22:05:44 I see Jul 31 22:06:01 after all, it's one of OE's assets that it's relatively self-contained Jul 31 22:06:11 and it builds really good toolchains :D Jul 31 22:06:53 hmmm it doesnt do eglibc yet Jul 31 22:07:30 before I figure how I can fiddle with toolchain on OE Jul 31 22:07:42 interesting Jul 31 22:07:46 never heard about that Jul 31 22:07:53 www.eglibc.org Jul 31 22:08:23 its glibc totally but with embedded part taken care of Jul 31 22:08:47 you dont have to use qemu to generate locales Jul 31 22:08:59 it has capability to generate them in cross env Jul 31 22:09:12 wow. sounds pretty good Jul 31 22:09:27 03stefan 07org.oe.dev * r62b5fc0f... 10/ (1 conf/distro/openmoko.conf): openmoko.conf: Use mirrors to balance load and set FEED_URIS Jul 31 22:09:43 and also maintains arches that glibc has vitually kicked out Jul 31 22:09:51 mickeyl, I have ASSUME_PROVIDED += " virtual/libc " Jul 31 22:10:04 certainly thats not honoured Jul 31 22:10:12 you may need to assume provided more Jul 31 22:10:22 take a look at some distro configuration files Jul 31 22:10:35 where will I find them Jul 31 22:10:37 or use bitbake -v to see some actual names of providees Jul 31 22:10:41 OE/conf/distro/* Jul 31 22:14:21 hm I use bitbake -D Jul 31 22:24:22 zecke: webkit: Jul 31 22:24:24 make[1]: c: Command not found Jul 31 22:24:26 ever seen? Jul 31 22:24:54 oehm Jul 31 22:24:56 lol Jul 31 22:25:30 i wonder which variable expansion is breaking there Jul 31 22:25:54 aah Jul 31 22:25:57 darn simple Jul 31 22:26:02 yes? Jul 31 22:26:12 for some reason CXX is not set Jul 31 22:26:30 hm Jul 31 22:26:45 but why is this expanded with c:? Jul 31 22:27:10 good question Jul 31 22:31:54 aha Jul 31 22:32:01 woglinde: it's getting expanded like Jul 31 22:32:07 $(CXX) -c foo Jul 31 22:32:13 hm okay Jul 31 22:32:16 and Make treats -c as a silent c command Jul 31 22:32:30 there's some special treatment for the hyphen iirc Jul 31 22:33:25 re Jul 31 22:34:18 i need a advice about a device, i need a device with bt, wifi, and a good processor/video chip for video play back Jul 31 22:34:23 this device exits ? Jul 31 22:34:29 and usb host Jul 31 22:34:34 hm Jul 31 22:34:44 I dont think it exists Jul 31 22:35:44 another question, where i can find pxafb_overlay ? i have installed angstrom x11 image but i dont have the module Jul 31 22:35:49 and dont find in the repo Jul 31 23:39:24 someone get this issue on spitz, with mplayer ? Jul 31 23:40:08 ioctl(6, FBIOPUT_VSCREENINFO, 0x3e1ed0) = -1 EINVAL (Invalid argument) Jul 31 23:40:23 then mplayer exits Jul 31 23:51:50 03mickeyl 07org.oe.dev * r15cbf81a... 10/ (1 packages/puzzles/oh-puzzles_svn.bb): add oh-puzzles, the Opened Hand Puzzles Collection (taken from Poky) Jul 31 23:52:07 03mickeyl 07org.oe.dev * r33555e3a... 10/ (13 files in 3 dirs): Jul 31 23:52:07 openmoko2: Jul 31 23:52:07 * prepare for SVNREV Jul 31 23:52:07 * add openmoko-calculator2 Jul 31 23:52:16 03mickeyl 07org.oe.dev * r412996f7... 10/ (1 packages/tasks/task-openmoko.bb): task-openmoko: add games, add calculator, remove panel-mainmenu Jul 31 23:52:27 03mickeyl 07org.oe.dev * r202e63b7... 10/ (8 files in 2 dirs): openmoko-panel-plugins: prepare for SVNREV Jul 31 23:52:36 03mickeyl 07org.oe.dev * r1bff33df... 10/ (1 packages/libowl packages/libowl/libowl_svn.bb): add libowl (OpenedHand Widget Library) [taken from poky] Jul 31 23:52:41 03mickeyl 07org.oe.dev * refe503d8... 10/ (1 packages/puzzles/oh-puzzles_svn.bb): oh-puzzles: add libowl to dependencies [should be done in poky as well] Jul 31 23:52:46 03mickeyl 07org.oe.dev * r2ee5368f... 10/ (1 packages/webkit/webkit_svn.bb): webkit: clean up (guys, please don't mix TABs and SPACEs) and make it build Aug 01 00:52:21 NOTE: package module-init-tools-cross-3.2.2: failed Aug 01 00:53:12 If you meant to cross compile, use `--host'. Aug 01 00:55:49 hi all, I'm new to oe, I need to build a software which runs on my pc called usbip for oe, I noticed that oe already has met most of the prerequisites for usbip. But what I don't understand is how to compile usbip without using .bb file on the "virtual oe system" on the host machine. Or, is using a bb file unavoidable? Aug 01 01:29:25 lun, using a bb file is unavoidable Aug 01 01:42:20 Thanks Crofton, I'll try copying from some existing bb file Aug 01 02:44:57 Hi folks **** ENDING LOGGING AT Wed Aug 01 02:59:57 2007