**** BEGIN LOGGING AT Sun Nov 19 02:59:56 2006 Nov 19 03:00:56 03lenehan 07org.oe.dev * r22460f91... 10/ (3 files in 3 dirs): Nov 19 03:00:56 cherokee 0.5.5: Minor fix to configure.in. Should be using "no" and not Nov 19 03:00:56 "false" when checking autoconf variables. This prevented you marking Nov 19 03:00:56 sendfile as broken via the site file. Nov 19 04:02:03 a+ Nov 19 05:01:11 I'm trying to build OE but I don't know what I should set my MACHINE variable to... I'm building it for a HP iPaq rx4540 Nov 19 05:01:54 I've looked in the org.openembedded.dev/conf/machine/ but I can't see anything that directly matches... is their a second best to set it to? Nov 19 06:54:13 zezom, have you happend to come across some doccumentation that someone has gotten a working build for the rx series ipaqs? Nov 19 07:35:28 emte, not yet, here was my first poin of checking Nov 19 08:08:47 morning all Nov 19 08:57:21 I've been doing some reading and I still don't know which MACHINE to set in my config file... I have a HP iPaq rx4540 and I've seen support for the hxxxx series but not the rx's... what is the best course for me to take? Nov 19 09:06:15 zezom: find one with a similar CPU and use that Nov 19 09:12:01 koen, thanks Nov 19 09:14:53 if I make my own config file based on another one will it get wiped in future updates? Nov 19 09:15:15 and if it works where can I submit it? Nov 19 09:15:43 zezom: Name it differently and it shouldn't be wiped out by mtn Nov 19 09:16:23 NAbyss, cool thanks Nov 19 09:17:50 zezom: If you're using ext3, and you're ultraparanoid about it being wiped out, you can use chattr to make it immutable, but I think you'd be safe regardless Nov 19 09:18:21 zezom, Nabyss: mtn never wipes out local changes Nov 19 09:18:42 no it's all good, if I was that paranoid I would just back it up some where else :) Nov 19 09:18:42 the worst case scenario is that it will pop up a 3way diff tool to merge your changes Nov 19 09:19:39 koen|away: Yeah, thought as much Nov 19 09:20:02 it isn't like cvs or svn that will mangle your files with 'inline difs' Nov 19 09:20:47 ~lart rebooterola phones Nov 19 09:20:48 * ibot throws rebooterola phones's poor little doggy off a cliff Nov 19 09:20:51 koen|away: Ah, bloody good Nov 19 09:24:05 "but Qtopia desktop doesn't connect under GPE" Nov 19 09:24:09 * koen sighs Nov 19 09:33:18 hi guys Nov 19 09:33:29 hey Spyro Nov 19 09:33:57 * koen sees the trainstation Nov 19 09:33:58 later all Nov 19 09:35:39 koen I ported hh.org to 2.6.19 here :) Nov 19 09:35:48 Im running 2.6.19 on my e740 now :) Nov 19 09:36:21 git is a fantastic tool Nov 19 09:36:48 (is there anyhting it cant do? it already made me my tea and brought me the paper and my slippers... Nov 19 09:36:49 ) Nov 19 09:45:39 I know this is perhaps off topic, but, has anyone here experience with interfacing linux to STN color LCDs? I have a 12bit (4096) color STN hooked up to an s3c2410 processor that I just cannot seem to get the colors "right". Nov 19 09:46:24 cyan shows as yellow, blue as red, but green is ok. Nov 19 09:47:39 I know it has to do with the palette, but I cannot find in the frame buffer code where the palette colors are assigned / built. Nov 19 09:54:53 you sure its the palette and not your register bitmap? Nov 19 09:54:59 15 or 16 bit colour ? Nov 19 09:55:02 12bit Nov 19 09:55:18 sure you got the regs as 444 ? Nov 19 09:55:44 I got the 565 on e740 as 555 once and it looked 'interesting'... Nov 19 09:56:02 according to the samsung datasheet, the 12bit color mode of the lcd controller does not use the hardware palette. I can see where linux is placing the wrong bit values into the framebuffer memory. Nov 19 09:56:47 so it displays 12 bit from 16 bit data ? Nov 19 09:57:59 the LCD controller is in 12bit mode, the display is a 12bit (RRRRGGGGBBBB) model. Nov 19 09:58:40 12bits of framebuffer (.75 wordlen) defines a bit (pixel) Nov 19 09:59:08 12bits of framebuffer (1.5 bytes) defines a bit (pixel) Nov 19 09:59:10 heh Nov 19 10:00:08 I keep looking via google for any docs that would give me some understanding of palette theory, but, not much out there it seems. Nov 19 10:01:55 AFAIK, the palette is a "best match" to the desired color. For example, if you have a reddish-yellowish-green in a 24bit shade, some algorithm would reduce that 24bit shade into the "best match" of the current 4096 palette? Nov 19 10:02:52 some type of dithering process of the 8x8x8 color down to a 4x4x4 or 5x6x5 or 5x5x5 palette reference Nov 19 10:03:57 but, it looks like the palette is not "unlimited" in size, there is only 256 entries x 16bits/per of choices? Nov 19 10:04:21 freaky Nov 19 10:05:04 I really am out my area with this one, palette color / operation is something I've never concerned myself with before. hheh Nov 19 10:05:20 :) Nov 19 10:05:38 I don't think the customer would want an LCD that gives red instead of blue Nov 19 10:05:38 cant you just set linux up to use 444 modes and drive it direct ? Nov 19 10:06:42 that's what I'm trying to figure out, I think you are correct as I am trying to do the same thing. the framebuffer (s3c2410fb.c) is definately building a 5x5x5 or 5x6x5 palette. Nov 19 10:07:38 there is no code to build a 4x4x4 palette. AND I don't want to run this display in 256 color if I can avoid it! Nov 19 10:08:21 I guess I'll just hack the setcol() (palette builder) function and reduce some 5x5x5 code down to 4x4x4 and see what happens, lol Nov 19 10:08:45 Spyro: just when you think you've been everywhere inside the kernel source... Nov 19 10:42:43 03rpurdie * r695 10bitbake/lib/bb/parse/parse_py/BBHandler.py: Nov 19 10:42:43 BBHandle.py: Instead of iterating through data.keys(d) calling getVarFlag each Nov 19 10:42:43 time, memorise the list of handlers in __BBHANDLERS and tasks in __BBTASKS. Nov 19 10:42:43 Against 600 .bb files in poky, this removes 2 million function calls in total of Nov 19 10:42:43 which 700,000 were to getVarTask resulting in a 5-10% speedup Nov 19 10:46:14 aaargh, my eyes are bleeding Nov 19 10:46:21 * koen should stop staring at m4 code Nov 19 10:55:47 good morning Nov 19 11:02:23 hi, all! Nov 19 11:02:56 Is it possible to build tree with dbus 1.0 or 0.94? Nov 19 11:04:34 sure, I just built against dbus 1.0.1 Nov 19 11:06:10 so gnome-vfs-dbus was ok? Nov 19 11:06:25 I'm happy to hear Nov 19 11:07:54 Is there some bitbake target to just fetch sources? I'm on a dial-up line now... Nov 19 11:08:40 bitbake -c fetchall Nov 19 11:14:41 03koen 07org.oe.dev * r82bc227e... 10/ (1 conf/distro/angstrom-2007.1.conf): angstrom 2007.1: prefer dbus 1.0.1 Nov 19 11:21:47 03koen 07org.oe.dev * r0af62f4f... 10/ (3 files in 3 dirs): dbuys 1.0.1: ensure /var/lib/dbus exists before writing to it Nov 19 11:26:43 good morning Nov 19 11:26:48 T0mW: heh Nov 19 11:27:44 Spyro: it has to do with the pseudo_palette Nov 19 11:30:56 hey likewise Nov 19 11:35:12 koen: hey Nov 19 11:35:17 Hello everybody :) Nov 19 11:35:32 hey gerwinin Nov 19 11:35:54 hey gerwinin Nov 19 11:36:00 Pff catched a cold on friday I think Nov 19 11:39:02 morning Nov 19 11:39:49 hey nicolasfr Nov 19 11:40:49 Hi koen. I am out of luck... I just set up a new build env on Edgy, and now everything is broken (error while building gcc-cross-4.1.1-r7). Nov 19 11:41:24 "checking for getpagesize... configure: error: Link tests are not allowed after GCC_NO_EXECUTABLES." Nov 19 11:55:07 morning Nov 19 11:55:42 morning hrw Nov 19 11:56:57 Shoragan: ping Nov 19 12:02:06 * gerwinin puts on some music Nov 19 12:04:47 hey gremlin Nov 19 12:10:19 morning guys Nov 19 12:10:24 hi Gerrath Nov 19 12:10:34 ops .. hi gerwinin Nov 19 12:21:52 What's the proper place for version-specific patches? Sometimes I see "packages/packagedir/version/*.patch", sometimes "packages/packagedir/package-version/*.patch" Nov 19 12:27:52 03hrw 07org.oe.dev * r52565181... 10/ (8 files in 5 dirs): openembedded-essential: added gawk Nov 19 12:29:12 hrw: How do you package up openembedded-essential for a host distro (such as Debian)? Nov 19 12:29:37 likewise: using debian tools Nov 19 12:29:38 hrw: Because I am interested in having a Dapper and Edgy openembedded-essential as well. Nov 19 12:30:05 Add this line to APT sources: Nov 19 12:30:05 deb http://www.openembedded.org/dl/ packages/ Nov 19 12:30:33 mm I am getting this error : ld: cannot find -lqte-mt-mt-mt Nov 19 12:30:48 mtmtmt? Nov 19 12:31:11 Yeah I think this should be ld -lqte -mt Nov 19 12:31:19 -lqte-mt rather Nov 19 12:31:34 hrw where can I correct this ? Nov 19 12:31:55 looks like something adds -mt, and that something gets called several times Nov 19 12:32:11 gerwinin: check qte* recipes Nov 19 12:33:41 likewise: I wrote about it on ML Nov 19 12:33:56 likewise: "Re: [oe] Debian and RPM packages for easier start with OpenEmbedded" Data: Nov 19 12:34:00 30/10/2006 10:43 Nov 19 12:34:37 hrw: thanks, cathcing up Nov 19 12:34:44 hrw I get it while compiling libsdl-qpe Nov 19 12:35:05 gerwinin: I would like to kill sdl-qpe at all Nov 19 12:35:10 hrw: I am still missing a handful of packages (now that I install oe-essential) :-) Nov 19 12:35:24 likewise: show me output Nov 19 12:35:50 hrw: no, I mean, I still didn't have all dependencies installed that OE (might) need. Nov 19 12:36:14 likewise: apt-get install openembedded-essential should fetch all of them Nov 19 12:36:40 hrw: yup, it still brings up packages I forgot during manual installation of the packages. Nov 19 12:49:24 hrw: me too Nov 19 12:52:12 fsck. I installed bitbake in system.. Nov 19 12:55:53 removed. Nov 19 13:07:44 each time when I visit OE homepage I have a feeling that they need redesinging Nov 19 13:14:24 hrw I added this bug to bugzilla Nov 19 13:19:39 I am having troubles with a fresh OE install on edgy. I am targetting x86 platform and gcc-cross-4.1.1 does not build... It breaks on zlib in "multilib subdirs 64". Any idea? This worked fine before I upgrade to the latest OE revision Nov 19 13:21:08 it breaks just after this notice: NOTE: SITE files /opt/OE/org.openembedded.dev/site/endian-little /opt/OE/org.openembedded.dev/site/ix86-common Nov 19 13:21:40 probably some vars are missing Nov 19 13:21:41 Nicolas did you check i86-common ? Nov 19 13:22:09 Nicolas what is your host ? Nov 19 13:22:29 Linux desktop 2.6.17-10-generic #2 SMP Fri Oct 13 18:45:35 UTC 2006 i686 GNU/Linux Nov 19 13:23:09 I checked ix86-common but can't see anything wrong... Nov 19 13:23:30 Target platform is i486 Nov 19 13:24:08 Nicolas I had a similar problem Nov 19 13:24:26 Nicolas did you check bugs.openembedded.org ? Nov 19 13:25:01 gerwinin: I did, but could not find a related bug... Nov 19 13:25:33 Nicolas can you pastbin the error ? Nov 19 13:26:04 http://pastebin.ca/252411 Nov 19 13:27:04 gcc-cross-intial did build fine Nov 19 13:27:56 Nicolas I am not building for angstrom but it is saying i686-angstrom Nov 19 13:28:23 And I think you should have a look in ix86 common Nov 19 13:30:50 I am looking at ix86-common but I can't see what to do with it :( Nov 19 13:32:02 Nicolas what are you exactly making ? Nov 19 13:33:32 I am just trying to setup the OE environement for a x86 target (Tatung webpad - distro Angstrom). It did work fine until today (I have upgraded my distro to Edgy and to the latest OE...) Nov 19 13:34:36 I should always remember that "if it works, don't break (upgrade) it" Nov 19 13:34:47 Nicolas I must say that I am building for x86 as well , let me give you my local.conf Nov 19 13:34:59 But I am having a different target Nov 19 13:35:18 do you use the latest OE revision? Nov 19 13:35:23 yes Nov 19 13:35:54 could you try to "mv build/tmp build/tmp.bak && bitbake nano"? Nov 19 13:36:05 to see if gcc-cross builds? Nov 19 13:38:04 gcc-cross builds Nov 19 13:38:50 damn... Nov 19 13:39:32 I will pastbin my local.conf Nov 19 13:39:58 Nicolas I am having some other problem I want to fix Nov 19 13:40:12 which problems? Nov 19 13:40:23 Yeah it is with libsdl-qpe Nov 19 13:44:26 So far I am building matchbox-x11 only... Nov 19 13:45:52 Nicolas I will propose to make a seperate brand for x86 Nov 19 13:46:16 brand = branch Nov 19 13:48:03 why do you think it is needed? Nov 19 14:13:18 Spyro: got it! palette is okay now Nov 19 14:16:06 T0mW: congrats Nov 19 14:16:27 koen: 8 lines of code in only 7 hours Nov 19 14:16:27 "Running configure in multilib subdir 64". I don't have a 64bit CPU :/ Nov 19 14:19:39 ~kergoth Nov 19 14:19:44 hmm... kergoth is Chris Larson, an open source developer currently working for Openedhand Ltd (o-hand.com), remotely, from AZ. Core developer on OpenZaurus, OpenEmbedded/BitBake, and TSlib. Often the resident BOFH, and often an ass. Newbies and the clueless beware. Nov 19 14:20:26 T0mW: what was it ? Nov 19 14:20:42 ~hrw Nov 19 14:20:43 hmm... hrw is Marcin Juszkiewicz - OpenEmbedded/OpenZaurus/OPIE developer, OpenZaurus 3.5.4.x release maintainer - contact via homepage: http://www.hrw.one.pl/ but also Human Rights Watch organization (both not related) Nov 19 14:21:57 Spyro: on the s3c2410, you have to set fb_info->fix.visual = FB_VISUAL_DIRECTCOLOR (does not use a hardware palette), then set the fb_info->psuedo_palette 16 color values to the appropritate 4:4:4 RGB values Nov 19 14:22:23 cool Nov 19 14:23:19 found out why SD wont work on e750 and e800 Nov 19 14:23:21 Hrw: seems I cannot get over this bug in libsdl-qpe Nov 19 14:23:24 the chips not being reset right. Nov 19 14:23:50 Spyro: instead of setting the palette entries as RRRRGGGGBBBB you have to set them as BBBBGGGGRRRR! Nov 19 14:24:41 RP: alive? Nov 19 14:25:07 I'm digging in bitbake source to find list of copyrights Nov 19 14:25:43 Could I use libsdl_ttff instead of libsdl-qpe ? Nov 19 14:28:59 hi Nov 19 14:29:06 hey Genesis Nov 19 14:29:41 yop mister kooi Nov 19 14:31:00 Genesis: hi Nov 19 14:32:51 hey genesis Nov 19 14:33:08 hey :) Nov 19 14:33:19 there is a lot of people from nl here :) Nov 19 14:35:16 Genesis yes :) Nov 19 14:35:27 And a lot from Germany Nov 19 14:35:31 RP: you here ? Nov 19 14:35:41 Genesis where you from ? Nov 19 14:35:45 france Nov 19 14:36:11 Where in France ? Nov 19 14:36:16 normandy Nov 19 14:36:29 I had holliday there once :) Nov 19 14:36:32 At the coast Nov 19 14:36:33 RP: I could use a word with him - seems hes done a lot of work on the tc6393xb stuff but I want to know where his most reecent stuff is and why my name vanished from the code... Nov 19 14:36:45 gerwinin : i'm in Caen Nov 19 14:37:00 Koen : Can you help me with some issue with libsdl-qpe ? Nov 19 14:37:27 gerwinin: yes, switch to a sane libsdl, like libsdl-x11 Nov 19 14:38:36 Koen okay I will Nov 19 14:39:02 03koen 07org.oe.dev * r861ef939... 10/ (1 packages/tinymail/tinymail_svn.bb): tinymail: update iconv stuff to latest svn Nov 19 14:39:48 Koen: I will get the server for oe next week :) Nov 19 14:40:30 gerwinin: a mirror, or build server? Nov 19 14:46:16 Likewise: a mirror Nov 19 14:46:37 gerwinin: source mirror? Nov 19 14:46:56 koen: what's the traffic from the OE server now? Nov 19 14:47:05 Koen : if I define preferred provider in my local.conf it sticks with libsdl-qpe Nov 19 14:47:15 koen: not *now* now, but the last few months? Nov 19 14:47:17 hrw: source miror Nov 19 14:47:30 likewise: you'd have to ask ggilbert_ Nov 19 14:47:35 koen: k Nov 19 14:48:15 likewise: *.oe.org point to various boxes over the world Nov 19 14:48:25 gerwinin: good. so we will probably have 2 source mirrors soon Nov 19 14:48:49 source mirror is a synced mtn + source package achives? Nov 19 14:49:06 hrw: didn't someone offer to write an RFC about uploading the source when adding a recipe? Nov 19 14:49:47 koen: let me guess - that one Polish dev? Nov 19 14:49:57 on oedem? Nov 19 14:50:24 hrw: I remember only that is got discussed at OEDEM, not who Nov 19 14:50:53 koen: first we have to have source mirrors with upload functionality. Nov 19 14:51:07 koen: that one which I work on will does not allow uploading rather Nov 19 14:51:37 03koen 07org.oe.dev * ra5476e98... 10/ (1 packages/gnome/libbacon_cvs.bb): libbacon: add floating cvs version Nov 19 14:51:49 hrw: I would suggest specifying a generic upload method first (FTP, SFTP, WEBDAV, rsync, etc) Nov 19 14:52:03 hrw: so that all (future) mirrors have the same interfaces Nov 19 14:53:32 likewise: rsync sounds ok Nov 19 14:54:44 Shoragan: bitbake 1.6.2-3.0 built - I will send you updated Debian diff later today Nov 19 14:57:38 03koen 07org.oe.dev * ra55caa03... 10/ (1 packages/contacts/contacts_0.2.bb): contacts: update to 0.2 Nov 19 14:57:49 03koen 07org.oe.dev * r0b204efb... 10/ (3 files in 2 dirs): tinymail: add missing files Nov 19 14:57:58 gerwinin: how many ppl at your work, working with OE? Nov 19 15:00:56 Likewise: At this moment I am the only one Nov 19 15:02:58 gerwinin: same here, although I will probably need to teach a few people how to work with it for all our embedded dev. soon Nov 19 15:05:02 Likewise: Where are you working ? Nov 19 15:15:29 hrw: there's a bug with the mindate/maxdate stuff: http://www.angstrom-distribution.org/repo/?action=details&pnm=tinymail Nov 19 15:15:46 will see Nov 19 15:15:50 Well, I've finally got my gpe-image compiled. Only two issues: I had to downgrade glib-dbus to 0.71 (because it seems that 0.72 needs a different dbus version) and also an strange problem in the arm-linux staging directory, where libesd.so.0.2.28 was a link to itself. Nov 19 15:17:05 koen: what kind o bug is there? Nov 19 15:17:32 the Source: lists patch(es) that weren't used Nov 19 15:17:46 koen: it list SRC_URI entries Nov 19 15:17:56 koen: not 'I used those files to built' Nov 19 15:18:36 I made manpage for bitbake trunk Nov 19 15:19:10 So, finally, good bye opie, hello gpe Nov 19 15:20:08 sirfred, :) Nov 19 15:22:32 good sirfred :) Nov 19 15:23:00 * CoreDump|home installs a clean image on Poodle Nov 19 15:23:44 It lives Nov 19 15:24:43 * Genesis try to get a wrt54GS V3 ( 32Mo Ram ) Nov 19 15:25:02 Genesis: try wrtsl54gs Nov 19 15:25:19 i can't ship them in europe Nov 19 15:25:29 sure they're better , with usb2 ... Nov 19 15:25:47 perharps i can paid from US webshop Nov 19 15:25:56 i look Nov 19 15:30:50 morning Nov 19 15:30:54 yo chouimat Nov 19 15:31:10 hrw : only one antenna on the SL , i prefer a GS :) Nov 19 15:31:41 Genesis: the SL is better and faster than then GS :) Nov 19 15:32:13 it's for long distance , 2 antennas is better Nov 19 15:33:42 connect better antenna instead of normal one? Nov 19 15:33:43 Genesis: I hope for they actually plugged the second one ... I remember my wrt54g 1.1 2 antennas but holy one was actually plugged inside :) Nov 19 15:50:43 03hrw 07org.oe.dev * r610d8df9... 10/ (7 files in 3 dirs): openembedded-essential: bumped to 1.2, now it is more lintian happy Nov 19 15:51:19 Hum, something strange happened, I'm getting a lot of alignment traps for my programs in the kernel logs. What could be the cause of this? Nov 19 15:51:44 Things like: Alignment trap: rxvt (2398) PC=0x....... INtrs=0x..... Address=0x..... FSR 0x013 Nov 19 15:54:24 CIA-4: kick Nov 19 15:54:24 03hrw * r696 10bitbake/doc/bitbake.1: added bitbake manpage based on --help output Nov 19 15:54:46 hi greentux Nov 19 15:54:53 hoi hrw Nov 19 15:59:54 I need a break.. Nov 19 16:00:00 time for it crowd Nov 19 16:00:06 I think I'm going to try with a 2.6.17 Nov 19 16:18:48 Koen: for yout list: gpe-image and mythtv image compile now Nov 19 16:38:14 I'm also getting those Alignment traps kernel errors with 2.6.17 Nov 19 16:39:23 http://netwinder.osuosl.org/users/b/brianbr/public_html/alignment.html Nov 19 16:44:10 koen: we need to find someone who will redesign OE website Nov 19 16:44:40 I tried to read gettingstarted but those headers make my brain escaped Nov 19 16:45:03 koen: So, this error I'm seeing in the kernel log is something that just happens under X ? Nov 19 16:45:20 it happens in badly written apps Nov 19 16:45:47 For what I'm reading here, it's just a problem of having odd size structures Nov 19 16:46:52 I never had seen this in the old days, when I run opie ;-) Nov 19 16:47:09 It seems it's happening in all the gpe-* programs Nov 19 16:47:15 hrw: fwiw, I want to remove gettingstarted from the wiki and put it in the usermanual Nov 19 16:47:25 sirfred: like I said, badly written apps ;) Nov 19 16:47:41 or a badly written lib (glib) Nov 19 16:47:49 or libdbus Nov 19 16:47:52 koen: The fact is that I'm feeling the whole system slow, could this be the cause? Nov 19 16:48:07 It's dumping a really lot of logs into the kernel msg queue Nov 19 16:50:29 koen: I'll try to talk with coworkers about new design Nov 19 16:54:26 sirfred: one of the possible slowdowns could be a dbus timeout problem Nov 19 16:55:02 sirfred: or you used a buggy gcc that generates tons of alignenment errors Nov 19 17:19:26 hrw, Spyro: here now Nov 19 17:20:06 Spyro: I haven't had anything to do with the tc6393 code - you need to talk to the Tosa developers (Dirk - do13) Nov 19 17:20:13 03pfalcon 07org.oe.dev * r4531506c... 10/ (1 packages/linux/linux-handhelds-2.6/h4000/defconfig): linux-handhelds-2.6 cvs: Update h4000 defconfig for 2.6.18. Nov 19 17:20:24 03pfalcon 07org.oe.dev * rf19342b2... 10/ (4 files in 3 dirs): linux-handhelds-2.6 cvs: Drop patches merged upstream. Nov 19 17:30:55 re Nov 19 17:31:33 RP: I was digging through all bitbake files to gather list of authors/copyrighters Nov 19 17:43:20 hrw: ah, ok Nov 19 17:43:57 RP: I cant find a moment when subversion fetcher was added Nov 19 17:44:15 will see in pre-bitbake times Nov 19 17:45:05 found. Nov 19 17:45:11 http://oe-devel.bkbits.net:8080/oe.deprecated/user=hrw/cset%401.780?nav=!-|index.html|stats|!+|index.html|ChangeSet Nov 19 17:45:27 bkbits.net heh Nov 19 17:45:33 now that's old Nov 19 17:46:26 27 months ago Nov 19 17:46:49 I was wondering about adding itself into authors file in bitbake Nov 19 17:49:23 to left any sign of myself there ;D Nov 19 17:49:41 but now time is for food making Nov 19 17:53:27 hrw: Feel free to add a mention on the header of the file :) Nov 19 18:09:52 RP: which o-hand dude is the best one to ask about eds-dbus building troubles in OE? (bug #1603) Nov 19 18:12:17 koen: ross Nov 19 18:12:45 thanks Nov 19 18:13:53 koen: I suspect upgrading dbus in poky would help with this which is on oh's todo list at some point... Nov 19 18:14:08 :) Nov 19 18:14:29 It annoys me that on one machine dbus works, and on the other it doesn't Nov 19 18:16:18 03hrw * r697 10bitbake/lib/bb/fetch/svn.py: Nov 19 18:16:18 svn fetcher: added my (C) as this is code which I wrote in 2004 (pre-bitbake times) Nov 19 18:16:18 http://oe-devel.bkbits.net:8080/oe.deprecated/cset%401.780 Nov 19 18:16:59 03hrw * r698 10bitbake/AUTHORS: AUTHORS: added myself Nov 19 18:16:59 oemake Nov 19 18:17:07 oemake2 Nov 19 18:17:09 koen: oemake2 Nov 19 18:17:10 bitbake Nov 19 18:17:14 oemake2 was better Nov 19 18:17:32 koen: oemake -> oemake2 -> oemake -> bitbake Nov 19 18:18:01 cool Nov 19 18:28:55 2cu Nov 19 18:39:49 03koen 07org.oe.dev * rcc4e8cb9... 10/ (1 packages/dbus/dbus-native_1.0.1.bb): dbus-native: update to latest stable (1.0.1) Nov 19 18:41:27 tinymail is pretty fast on the hx Nov 19 18:52:25 RP: Ross will look at it next week Nov 19 19:43:04 koen: cool :) Nov 19 19:55:09 RP: I also added contacts 0.2 + libbacon to OE today Nov 19 19:55:31 koen: That was quick. It was only released last week if I remember rightly :) Nov 19 19:55:39 today :) Nov 19 19:55:47 or rather, I read Ross' blog today Nov 19 19:56:28 Friday :) Nov 19 19:56:37 It still refuses to import vcards, so I guess I'll add some bugs to o-hand bugzilla next week Nov 19 19:56:43 I doubt its make poky yet... Nov 19 19:56:57 I checked, it isn't in poky yet Nov 19 19:57:27 We have some new people working on contacts and now that release it out the way, I expect a few improvements Nov 19 19:57:48 yeah, I also read Thomos' blog today :) Nov 19 19:58:08 :) Nov 19 19:58:22 I should really read them, then I might have more of an idea what was going on... Nov 19 19:58:37 Hopefully I will be syncing poky <-> OE this week Nov 19 19:58:42 http://planet.closedfist.co.uk/ Nov 19 19:59:00 mallum swears that planet has nothing to do with o-hand ;) Nov 19 19:59:24 It isn't mallum's domain ;-) Nov 19 20:03:15 If all goes well people can use o-hand PIM and MokoPIM on their toys without worrying about moving their data Nov 19 20:03:33 That would be nice :) Nov 19 20:04:57 * koen spots some nokia dudes hanging out in #openmoko Nov 19 20:05:12 that should increase OE's visibility a bit in .fi Nov 19 20:06:18 koen: I have a habit of running Poky images on my 770 ;-) Nov 19 20:06:45 I'm still waiting for a de-hildonified maemo-mapper to do that Nov 19 20:07:32 I should have another go at that... Nov 19 20:08:16 I started one a while back... Nov 19 20:08:16 it's irritating that someone already got it working, but refused to release the patch Nov 19 20:08:16 Try asking again Nov 19 20:08:16 I forgot the nick...... Nov 19 20:12:01 We should try and move packaged staging forward again soon... Nov 19 20:14:12 tomorrow i'll to package mozplugger Nov 19 20:15:47 a good tip i think to package something on OE is to look after the gentoo ebuild Nov 19 20:16:21 hi Nov 19 20:23:44 18.4 seconds for a full rescan of poky using bitbake trunk :) Nov 19 20:24:17 hi Bernardo Nov 19 20:26:55 poky tree is still lighter ? Nov 19 20:27:01 600 recipes , isnt it ? Nov 19 20:27:06 yes Nov 19 20:29:03 RP : lot of improvement in the last trunk ? Nov 19 20:29:14 'ill compare tommorow on OE Nov 19 20:29:45 Genesis: In the most part improvements are shared with OE Nov 19 20:30:07 i was speaking about bitbake Nov 19 20:30:51 Genesis: Yes, lots of improvements (although all are in bitbake trunk too) Nov 19 20:31:08 Genesis: Tehcnically poky is now using an unstable bitbake version :} Nov 19 20:31:31 I'd much rather fix them than deal with issues in bitbake 1.4 though Nov 19 20:31:34 RP : how is the market target for OH ? Nov 19 20:31:51 Genesis: How do you mean? Nov 19 20:31:57 we want to launch a company with a friend Nov 19 20:32:19 Genesis: Doing what? Nov 19 20:32:39 linux integration , software devlopement / custuzation Nov 19 20:32:53 perharps hardware stuff if we find the right person Nov 19 20:33:36 donno , i need to study the market Nov 19 20:33:37 Genesis: good luck :) Nov 19 20:33:50 hey , gives me some advise :) Nov 19 20:34:41 i'd like to target automation process Nov 19 20:35:15 Genesis: Some of it is being in the right place at the right time. You need a good founding project to build a repuation on Nov 19 20:37:05 RP : do you know who work on a UI project for OE ? Nov 19 20:47:48 koen: Thanks for your tips. Nov 19 20:48:25 Do you think these king of warnings could be the symptom of that align problems? Nov 19 20:48:32 dbus-0.92-r1/temp/log.do_compile.745:dbus-marshal-byteswap.c:55: warning: cast i Nov 19 20:48:33 ncreases required alignment of target type Nov 19 20:48:45 sirfred: I've been dogfooding .dev images the past two weeks, so I know what annoyance people might encounter Nov 19 20:49:20 sirfred: dbus is a notorious offender for alignment stuff, but recent versions have gotten better Nov 19 20:49:45 koen: Perhaps my need to downgrade to glib-dbus 0.71 could have something to do Nov 19 20:53:19 koen: About the amount of alignment errors, they should be a lot, as every offending program is filling the whole fifo Nov 19 20:53:39 you can influence the behaviour Nov 19 20:54:19 /etc/init.d/alignment.sh tells how verbose the kernel is Nov 19 20:54:33 Humm, I see. Nov 19 20:54:40 What are the possible values? Nov 19 20:54:49 For /proc/cpu/alignment ? Nov 19 20:55:42 sirfred: see http://www.nslu2-linux.org/wiki/Info/Alignment Nov 19 20:55:50 koen: Thanks Nov 19 20:56:00 the default is '3', you might want '2' Nov 19 20:56:04 which is a workaround Nov 19 20:56:56 Let's see Nov 19 20:57:22 but the amount of alignment traps are a bit worrying Nov 19 20:58:20 koen: How could I know the reason? Nov 19 20:58:30 I recall some problems with binutils and gcc Nov 19 21:11:46 It seems that even no glib or dbus linked programs are also suffering alignment problems, for example: xstroke Nov 19 21:26:09 do13: hello Nov 19 21:27:06 * koen finishes his e-mail spree Nov 19 21:27:17 ~lart e-mail Nov 19 21:27:18 * ibot blasts e-mail to oblivion with a kamehameha wave Nov 19 21:27:31 ~lart snail-mail Nov 19 21:27:32 * ibot drops a baby grand on snail-mail Nov 19 21:27:57 hey really , what's the status for an UI in OE ? Nov 19 21:28:17 Genesis: You mean a UI for bitbake? Nov 19 21:28:26 yeap Nov 19 21:28:28 do13: u there ? Nov 19 21:28:51 Genesis: Being thought about. We have some ncurses code thanks to mickeyl Nov 19 21:29:07 * Spyro prods d013 with some bare mains cable Nov 19 21:29:36 Spyro: At least it wasn't live ;-) Nov 19 21:29:49 who says? Nov 19 21:30:30 RP: Could you please sync pcmanfm in OE with poky (OE version keeps segfaulting...)? Nov 19 21:30:56 nicolasfr: Ah, yes, I think poky's version has some fixes for that kind of thing... Nov 19 21:32:00 How useful does anyone find the DEPENDS and RDEPENDS graphs output of bitbake? Would a combined view of which filename.task depends on which other filename.task be better? Nov 19 21:32:07 RP: that's the kind of fixes I like!! thanks! Nov 19 21:32:55 koen: Was there any reason OE has a dated glib? Nov 19 21:33:12 RP: no idea, iirc I use 2.12 Nov 19 21:34:04 I think we should keep a pre-gslice glibc around and 2 major-versions post-gslice Nov 19 21:34:18 koen: ah, its just glib-native that wasn't updated Nov 19 21:35:23 koen: I know why 2.10 isn't there as you need a patch in poky to make it run on arm... Nov 19 21:36:13 I wish maemo wasn't stuck in 2004 Nov 19 21:36:24 * RP too Nov 19 21:36:58 since lots of gtk/glib/gnome people find the 770 sexy, but are oblivious to the fact that maemo needs outdated stuff to be fast Nov 19 21:37:39 No oblivious, just don't want to do the work in fixing up cario Nov 19 21:37:59 or do the political incorrect thing and removing cairo Nov 19 21:38:06 * koen likes the pangoxft patch Nov 19 21:38:37 koen: I was wondering about adding support for binary packages to OE - pull the deb packages from maemo.org and install into OE. Maybe contain those .bbs in a separate repo/collection and have them autogenerated by a script... Nov 19 21:39:02 koen: Obviously ABI would have to match and we'd have to have some safeguards for that Nov 19 21:39:08 a .dsc parser would do approx the same thing, right? Nov 19 21:39:37 but yeah, packaged-staging would make your idea pretty easy Nov 19 21:39:46 You'd pull in info from the .dsc, the packages themselves and from a suplimentry hints file Nov 19 21:40:42 koen: We could do it now, just with a new class with special a do_unpack for .debs, a null do_configure and null do_compile, special do_stage Nov 19 21:41:10 yeah Nov 19 21:41:44 I want to see zecke's face when he sees that class ;) Nov 19 21:41:56 koen: I'd probably vote for unpacking and repacking with oe's own functions so we have more control over the output... Nov 19 21:42:37 OE's .deb functions are a bit weak Nov 19 21:42:45 you are a bit spoiled by poky :) Nov 19 21:42:47 The advantage is we could then survive a binary only package in maemo. Currently we can't... Nov 19 21:43:13 koen: OE will get deb handling as soon as I'm happy it actually works. At the moment signs are, it doesn't quite Nov 19 21:44:33 if the binary only packages aren't C++ we could even build most stuff using OE Nov 19 21:45:55 koen: That would be the idea Nov 19 21:46:32 if we forget about opera, that only leaves mce and dsp stuff, iirc Nov 19 21:46:42 Any idea of why could be urxvt giving these errors? Nov 19 21:46:47 urxvt: unable to parse color 'rgb:ff/ff/cd', using pink instead. Nov 19 21:47:03 sirfred: XorA and I haven't been able to track down that bug Nov 19 21:47:18 A lot of those, with different RGB combinations, and I eventually get a beautiful pink terminal. Nov 19 21:47:31 koen: OK. Nov 19 21:47:35 sirfred: http://rafb.net/paste/results/xLAedW75.html Nov 19 21:47:48 I've also found that ncurses-terminfo is not installed into gpe-image Nov 19 21:47:52 sirfred: with that .Xdefaults it won't be pink anymore Nov 19 21:47:55 So, at least top doesn't work. Nov 19 21:48:11 koen: thanks Nov 19 21:49:06 koen: Another question. I was not able to instruct bitbake to build xserver-kdrive-w100 Nov 19 21:49:54 koen: bitbake xserver-kdrive-w100 seems not to be enough Nov 19 21:51:52 sirfred: 'bitbake xserver-kdrive' will build all 6 kdrive servers Nov 19 21:52:05 sirfred: 'ipkg install xserver-kdrive-w100' will install it on your device Nov 19 21:52:08 koen: Ouch! Nov 19 21:52:24 sirfred: Its all one build Nov 19 21:52:51 I understand, I thought that only xserver-kdrive-fbdev was built by default, and was trying strange PREFERRED_* spells Nov 19 21:53:05 grr stupid chip :( Nov 19 22:14:50 Good night Nov 19 23:31:31 * Spyro larts the tc6393xb Nov 20 00:32:12 hrw|gone: I see we have a little problem w/ udev.bb in .oz. For some machines (ie: spitz) udev is machine-specific (ie:udev-doc_084-r14_spitz.ipk) and for some it is not (poodle: udev_084-r14_armv5te.ipk). This is a pain to handle when upgrading the distro, so I suggest making udev machine-specific by default. Nov 20 00:32:49 hrw|gone: What's your stand on the matter? Nov 20 02:00:03 has anyone used vpnc on the sl 5500 ? **** ENDING LOGGING AT Mon Nov 20 02:59:59 2006