**** BEGIN LOGGING AT Wed Jun 15 23:59:57 2005 Jun 16 00:23:09 morning Jun 16 01:08:03 hmm Jun 16 01:08:17 any maemo rom images yet? ;) Jun 16 01:08:38 ljp: keen to try out the competition? :) Jun 16 01:10:02 of course Jun 16 01:10:12 if you can even call it competetion :) Jun 16 01:10:38 heh Jun 16 01:11:42 hi Jun 16 01:12:04 ljp: koen has maemo-image.bb Jun 16 01:12:11 hi hrw|work Jun 16 01:12:35 hi Dirk Jun 16 01:13:20 morning all Jun 16 01:15:43 hi ade Jun 16 01:17:13 hey hrw|work, Jun 16 01:19:09 hi hrw|work, ade|desk, do13 Jun 16 01:20:47 hrw|work: i looked at the wrt54 stuff last night, as far as i can tell there are patches to make uclibc has locale support, but the uclibc used has limited support, but uclibc svn should be better.. am still looking at it though Jun 16 01:21:56 hi bluelightning Jun 16 01:24:27 ade|desk: ok Jun 16 01:25:16 morning all Jun 16 01:25:24 morning Jun 16 01:25:24 hi RP Jun 16 01:25:32 hi BigAl Jun 16 01:25:42 do13: for reference suspend/resume appears to be broken on 2.6.12-rc6-mm1 Jun 16 01:26:46 RP: it looks like every -rc break ARM Jun 16 01:27:02 hrw|work: That about sums it up Jun 16 01:27:23 hrw|work: There were some arm power management changes in -rc6 so they're prime suspect Jun 16 01:27:24 morning RP Jun 16 01:30:00 http://www.iral.com/~albertr/linux/psion5mx/blue/ - how to hack usb BT dongle into uart BT Jun 16 01:30:51 RP: I found the same yesterday. But I have a known resume problem related with the charging stuff. So it's hard to find this. Jun 16 01:33:12 do13: At some point I'll try and track it down as it would be nice to have it working for 2.6.12. I suspect that is due for imminent release though Jun 16 01:34:54 RP: Any ideas? Jun 16 01:37:01 hrw|work: Nice. I thought this is only possible via SPI interface. Maybe I can repogramm my module to support BCSP instead of H4 mode. Jun 16 01:39:01 do13: glad to find it usefull Jun 16 01:39:21 hmm.. I have to configure BT network here at work Jun 16 01:39:27 at home I finally have wifi working Jun 16 01:43:45 do13: I suspect the patch to mach-pxa/pm.c that went into -rc6 Jun 16 02:03:16 hi all Jun 16 02:04:34 hi pb Jun 16 02:26:31 /dev/hda5 18G 18G 16K 100% /home/hrw/zaurus/oe/build Jun 16 02:26:35 nice.. Jun 16 02:28:27 good morning all Jun 16 02:29:15 hi koen Jun 16 02:29:24 hey hrw|work Jun 16 02:31:09 hi koen Jun 16 02:31:57 hey do13 Jun 16 02:32:11 hi guys Jun 16 02:32:13 latest project: figure out how to add bt to a sparc :) Jun 16 02:34:25 should be pretty easy to use a usb or rs232 dongle on a sparc, I imagine. Jun 16 02:35:46 rs232 would be the easiest way Jun 16 02:35:58 (no usb) Jun 16 02:36:13 do you have pci? if so, it'd be pretty easy to add a usb card. Jun 16 02:36:29 if you're on an sbus machine, rs232 is probably your only option. Jun 16 02:39:00 only sbus Jun 16 02:39:34 moin Jun 16 02:39:47 I'd like to have it internal, so I'm trying to locate the rs232 pins Jun 16 02:39:50 hey CoreDump|home Jun 16 02:41:33 w00t tomorrow arrives my new akita :) Jun 16 02:42:03 CoreDump|home: great Jun 16 02:44:29 Hi. I am still trying to build ppp, but now I get this: http://24.110.54.195/errors/ppp-bitbake.err what does any of it have to do with ppp? I see ipaq-sleeve in there and orinoco modules? Jun 16 02:44:59 did it just fail because of the patches? Jun 16 02:45:26 the patch failures are on ipkg-link though Jun 16 02:45:58 stealth: creal cache Jun 16 02:46:04 clear cache I mean Jun 16 02:46:11 how? Jun 16 02:46:34 rm /stuff/build/tmp/cache/* ? Jun 16 02:47:25 so people, sup? Jun 16 02:48:28 fighting with oe Jun 16 02:48:29 stealth: rm tmp/cache -rf - yes Jun 16 02:49:29 wow, argument list too long Jun 16 02:49:44 stealth: use "rm tmp/cache -rf" Jun 16 02:49:49 ill just do the whole dir and mkdir it Jun 16 02:49:49 yea Jun 16 02:50:03 dont need to mkdir it - bitbake will create it Jun 16 02:50:51 ohyea Jun 16 02:50:55 coo Jun 16 02:51:06 thanks.. lets try this again.. Jun 16 03:05:05 pb_: compiling abiword and gnumeric breaks with the gcc_csl, but not the 'regular' gcc, can I lower the default_pref for csl? Jun 16 03:07:34 depends what the breakage is Jun 16 03:10:03 symbol clashes Jun 16 03:10:19 http://pastebin.ca/13266 Jun 16 03:11:09 that looks more like a bug in abiword. I guess you'll have to debug it a bit to find out what's going on. Jun 16 03:31:24 03koen 07 * r1.3541.2.1 10openembedded/ (3 files in 2 dirs): Jun 16 03:31:24 preferred-gpe-versions.inc: use other xextensions Jun 16 03:31:24 xserver-kdrive_cvs.bb, xserver-kdrive_20050207.bb: change some default_prefs Jun 16 03:31:24 Net result: 'old' kdrive, but tslib works again. Jun 16 03:47:20 it would seem this is the right place to ask my questions, forgive me if it's not... i want to make an embedded linux system, with just a browser and mail reader, to fit into 64 mb of ram and go as fast as possible, where do i start? manuals, pointers? Jun 16 03:48:07 Noid: X11 based? Jun 16 03:48:39 yes, it's for end users Jun 16 03:49:01 x86? Jun 16 03:49:19 i was thinking of something in the lines of xfce+firefox+putamailreaderhere+gaim Jun 16 03:49:21 yes Jun 16 03:50:05 we have that soft in OE now - sylpheed as mailer Jun 16 03:50:23 it's for a celeron 300 with 64 mb of ram, some models have 128 mb of ram and/or better processor Jun 16 03:50:29 really? Jun 16 03:50:35 good (: Jun 16 03:51:05 set DISTRO="generic", MACHINE="x86" and run "bitbake firefox sylpheed gaim" Jun 16 03:51:19 I dont remember how xfce packages are named Jun 16 03:51:37 ok - "bitbake xfce-image" :) Jun 16 03:51:39 and instead of xfce what would be a good alternative? Jun 16 03:52:10 iirc we have xfce and matchbox - no other WM Jun 16 03:52:17 but I dont build X11 so.. Jun 16 03:52:20 hmmm you have everything already under control hehehe Jun 16 03:52:58 many thanks, gonna delve into your wiki and i'll come back to pester you if i don't understand something, thanks (: Jun 16 03:53:43 OE is easy to use when you want just build.. some more work when you want to add something Jun 16 03:54:58 i think what i mentioned is pretty much it Jun 16 03:55:17 sure Jun 16 03:55:35 we try to sell this ancient computers so we must optimize them hehe Jun 16 03:55:45 ;) Jun 16 03:56:03 I was going to buy pII/300 for server/router but got wrt54gs Jun 16 03:56:26 CIA-9: kdepimpi upgraded to 2.16 Jun 16 03:57:11 hrw|work: try using uclibc svn for it Jun 16 03:57:22 hardware solutions are much better, when they are readily configurable Jun 16 03:57:58 ade|desk: ok Jun 16 03:58:47 ade|desk: 20050613 be enough? Jun 16 04:21:05 U uClibc Jun 16 04:21:05 Checked out revision 10543. Jun 16 04:21:06 this may sound weird, but reading the wiki it seems i need the bitbake tool but i don't know where to find it Jun 16 04:21:30 read GettingStarted? Jun 16 04:21:32 ~bitbake Jun 16 04:21:34 from memory, bitbake is a simple tool for the execution of tasks. It is derived from Portage, which is the package management system used by the Gentoo Linux distribution. It is most commonly used to build packages, and is used as the basis of the OpenEmbedded project. See http://developer.berlios.de/projects/bitbake/ and http://openembedded.org/. Jun 16 04:22:03 oh, a bot, nice Jun 16 04:22:32 yeah ok, i found the link to the manual, but i saw no link to the tool itself and i didn't link it in my mind Jun 16 04:22:35 ok Jun 16 04:24:45 fsck.. uclibc defconfig is too old Jun 16 04:25:28 -rw-r--r-- 1 hrw hrw 1887694848 Jun 16 13:24 log.do_configure.18568 Jun 16 04:30:05 1.8G... Jun 16 04:34:40 ade|desk: built uclibc 20060613 - now bulding gcc-cross Jun 16 04:42:15 ARGHHHHHHHHHHHHHH.............. Jun 16 04:42:24 gcc-cross 3.4.4 failed again ;( Jun 16 04:42:52 I'm near to give up Jun 16 04:43:52 documentation.conf:linux and linux-uclibc. When using linux-uclibc you might want to set USE_NLS_gcc-cross to \"no\"." Jun 16 04:43:56 hmm.. Jun 16 04:44:30 uclibc isn't big on i18n Jun 16 04:44:42 http://www.vanille.de/temp/unfall.jpg Jun 16 04:44:43 hi mickeyl Jun 16 04:44:45 good morning mickeyl Jun 16 04:44:49 morning folks Jun 16 04:45:06 13:44 hrw@home:wrt54$ rm -rf tmp/ Jun 16 04:45:06 13:44 hrw@home:wrt54$ cp -a archive/ tmp Jun 16 04:45:10 another attempt Jun 16 04:45:20 mickeyl: I hope that isn't your car Jun 16 04:45:48 my mothers car after yesterday's accident. she's alright, just broke 2 rips. her friend has broken 12 ribs and the lounge was touched :/ Jun 16 04:46:01 ouch Jun 16 04:46:13 ouch Jun 16 04:46:38 auch Jun 16 04:47:03 yah. thank god it could've been worse. it was a bus from the fire brigade Jun 16 04:47:31 with a smaller car they would probably been dead Jun 16 04:50:45 is http://www.vanille.de/personality/captainslog.spy supposed to list a lot of pythoncode? Jun 16 04:55:37 mickeyl: when you will move to wordpress? :) Jun 16 04:58:36 koen: had to resetup my webserver yesterday. It's broken :/ Jun 16 04:58:45 mickeyl: ouch :( glad your mother is ok after that... Jun 16 04:58:53 bluelightning: thanks Jun 16 04:59:06 was that an E class? Jun 16 04:59:22 mickeyl: I did the same yesterday, apache2 broke everything Jun 16 04:59:36 bluelightning: yeah, '92 200E Jun 16 05:00:15 mickeyl: Ouch! :-( Hope they both GWS! Jun 16 05:01:09 thanks RP Jun 16 05:05:49 "I've built the kernel from openembedded.org and it doesn't work. Can you help" Jun 16 05:05:52 * RP sighs Jun 16 05:06:58 heh Jun 16 05:07:16 You don't want to know how much mails like that I get Jun 16 05:07:38 it doesn't work? fire it, of course Jun 16 05:08:01 "I cannot get collie flashed to 3.5.3 - can you help me? I can pay for this." - user from .au Jun 16 05:08:32 lunch time Jun 16 05:08:33 people usually stop reading docs if they can find and email address Jun 16 05:08:37 NOTE: [14:03] package gcc-cross-3.4.4: completed Jun 16 05:08:43 YES!!!!!!!!!! wrt54 build Jun 16 05:09:15 I vote to set USE_NLS="no" in gcc when TARGET_OS ="linux-uclibc" Jun 16 05:09:36 isn't that set already? Jun 16 05:09:54 seems like it is not Jun 16 05:10:04 I found that by mistake Jun 16 05:10:32 did "grep uclibc opeembedded/conf/* -r" and it was in documentation.conf Jun 16 05:10:49 so finally I will be able to build stuff for my router Jun 16 05:10:56 btw - wake-on-lan roxx ; Jun 16 05:13:25 you mean r0x? Jun 16 05:13:40 if so, you must be using gentoo Jun 16 05:13:41 * koen hides Jun 16 05:19:33 mickeyl: Ohh. Gute Besserung für beide Jun 16 05:19:40 danke dirk Jun 16 05:40:13 heh.. now I know what for one of my network cards has DB9 connector... Jun 16 05:40:28 aui? Jun 16 05:41:02 STP from tokenring networks Jun 16 05:42:23 probably 4mbit Jun 16 05:42:32 most 16/4 cards use utp Jun 16 05:43:18 they say that on STP it can do 100Mbps Jun 16 05:43:29 ~lart br1 for lzma Jun 16 05:43:29 * ibot whacks br1 upside the head for lzma Jun 16 05:43:42 I have some hstr cards here Jun 16 05:43:59 but those are pretty rare Jun 16 05:45:34 I only have them because 3com was giving them away and they included RC cars Jun 16 05:47:11 you are hardware collector? Jun 16 05:50:20 no, I like RC cars :) Jun 16 05:50:37 I do collect Unix machines Jun 16 05:51:08 http://linuxdevices.com/news/NS4736786928.html Jun 16 05:52:23 I dont have place Jun 16 05:53:24 nice Jun 16 05:53:27 after I moved I didn't have place too Jun 16 05:53:38 but I threw some stuff away -> more space Jun 16 05:54:12 sooner or later I will have to buy laptop to have 2 computers at home (2 normal ones) Jun 16 05:57:24 hey mickeyl Jun 16 05:57:37 your and the opie people are in germany right? Jun 16 05:58:23 someone can try to build lzma-native? Jun 16 05:59:42 NOTE: Applying patch 'lzma-406-zlib-stream.patch'ERROR: function do_patchcmd failed Jun 16 05:59:49 ok Jun 16 06:00:03 ~br1 Jun 16 06:02:29 found his email ;) will bug Jun 16 06:05:56 Assignee: Bruno Randolf did not match anything ;( Jun 16 06:06:27 argh.. it cannot even CC :( Jun 16 06:08:04 hrw|work: http://thread.gmane.org/gmane.comp.handhelds.openembedded/4702 Jun 16 06:08:42 yes.. I remember that Jun 16 06:09:53 forwared it to him Jun 16 06:49:35 RP: do you know if kdrive supports xvmc? Jun 16 06:50:20 koen: I'm afraid I dont Jun 16 06:50:50 * koen uses the source to find out Jun 16 06:51:23 RTFS Jun 16 06:51:56 it does have xv Jun 16 06:52:11 so if you figure out overlays on the w100 Jun 16 06:52:42 koen: It should be possible to use the mplayer code to do that Jun 16 06:53:08 hrw|work: well, i have just downloaded the oo sources... this is looking good so far Jun 16 06:54:36 Noid: ok Jun 16 06:55:36 if i say DISTRO="debian" will it create a debian like system? Jun 16 06:56:04 no Jun 16 06:56:17 see openembedded/conf/distro for valid options Jun 16 06:56:33 okis Jun 16 07:03:47 hmmm is it hard to make new packages? Jun 16 07:05:05 not really Jun 16 07:05:49 see http://article.gmane.org/gmane.comp.handhelds.openembedded/4416/match=license Jun 16 07:06:33 i assume every directory in openembedded/packages is one package, right? Jun 16 07:07:15 usually Jun 16 07:07:28 but look at gnome/ and maemo/ Jun 16 07:07:39 or gcc/ and glibc/ Jun 16 07:08:20 is it safe to ignore the "ERROR: kaffe.inc:76: unparsed line: 'pkg_postinst_${PN}...." errors? Jun 16 07:08:24 hmmm i see Jun 16 07:08:46 I think you need the svn version of bitbake Jun 16 07:09:08 that fixes a lot of bugs Jun 16 07:09:23 it gives me a few of those but it is going on Jun 16 07:10:22 RP: I just realize that we probably can't use the w100 accell Jun 16 07:10:34 RP: since kdrive keeps breaking :( Jun 16 07:14:02 anyone mind if I cut and hack the link above into the bb style guide? Jun 16 07:14:46 Crofton: please do Jun 16 07:14:53 koen: Something isn't running right on my gpe-image I've just baked :-(. Not sure if its kdrive or someting else... Jun 16 07:15:06 It will look bad, until I/others get around to formatting :) Jun 16 07:15:19 damn, Why are there so few women in this field Jun 16 07:15:28 Crofton: you can use http://ewi546.ewi.utwente.nl/tmp/d/keys.html if you get lost Jun 16 07:15:38 Crofton: err - because we're a bunch of ugly dorks? Jun 16 07:15:41 Crofton: better a few than none Jun 16 07:16:06 why are we computer people classified automatically as ugly? Jun 16 07:16:34 because it's true? Jun 16 07:17:16 I have gf Jun 16 07:17:23 erhm... i'm not mr perfect, but i think i'm fair looking d-: Jun 16 07:20:43 who will push 0x100 revision of bitbake? Jun 16 07:21:08 my money is on mickeyl Jun 16 07:21:38 we already have bug nr 100 in bugzilla, btw Jun 16 07:23:04 heh Jun 16 07:23:20 since mickeyl promised to fix the PN_PV bug Jun 16 07:23:30 did I? Jun 16 07:23:37 Thats what I heard Jun 16 07:23:38 :) Jun 16 07:23:44 hah traitor Jun 16 07:24:24 self.__keys = {} Jun 16 07:24:27 it's a nasty one, i wonder when we broke that Jun 16 07:24:37 it create array in python? how to sort it? Jun 16 07:24:48 {} is not an array, it's a dictionary (hashmap) Jun 16 07:24:57 ah.. Jun 16 07:24:58 [] is a list Jun 16 07:25:07 aka 1-dimensional array Jun 16 07:25:12 OK some more vars in the bb wiki page Jun 16 07:25:12 mickeyl: is there a way to sort alphabetically that dict? Jun 16 07:25:44 hrw|work: the keys in a dictionary have no order by design Jun 16 07:25:48 you can just use Jun 16 07:25:59 for i in keys.keys().sort(): Jun 16 07:26:05 print keys[i] Jun 16 07:26:07 ok Jun 16 07:26:20 03koen 07 * r1.3549.1.1 10openembedded/packages/xserver/xserver-kdrive_20050207.bb: xserver-kdrive_20050207.bb: depend on xextensions-1.0.1 Jun 16 07:29:30 koen, at some point we should link the vars to a detailed description like the page you refered me to Jun 16 07:29:32 python is not for me language Jun 16 07:29:33 woot. mod_python is heaps faster than fcgi Jun 16 07:30:06 Crofton: indeed, but that url is just the output from bitdoc Jun 16 07:30:28 bitdoc? Jun 16 07:30:30 quite ugly output Jun 16 07:30:34 heh Jun 16 07:30:57 bitdoc -o /path/to/output Jun 16 07:30:57 true, but it is a good idea for the ultimate format Jun 16 07:31:19 having it as a devhelp book would be nice Jun 16 07:31:55 must stop procrastinating real work Jun 16 07:34:06 why python doc is not made like php ones.. easier to find etc Jun 16 07:38:58 hi! hrw here? Jun 16 07:41:11 hi br1 Jun 16 07:41:36 hi! strange, the lzma-native patches work for me... Jun 16 07:42:22 lucky one.. Jun 16 07:43:06 br1: mobilemesh cannot be fetched Jun 16 07:43:12 svn: PROPFIND of '/svn/application/mobilemesh': 405 Method Not Allowed (http://meshcube.org) Jun 16 07:43:43 oh, shit. yeah. i'll have to change the SRC_URI Jun 16 07:43:57 br1: I will push some changes to distro/wrt54oe.conf in next week Jun 16 07:44:00 hi pgas Jun 16 07:44:02 ~lart webdav Jun 16 07:44:02 hi Jun 16 07:44:02 * ibot cuts off webdav's head with a halberd that could have been a little bit sharper Jun 16 07:44:16 funny, i changed to the svn version of bitbake and i get worse errors Jun 16 07:44:33 pgas: do I remember corretly - you are pdaX'rom' user? Jun 16 07:44:39 ehe: yes Jun 16 07:44:44 Noid: rm tmp/cache -rf Jun 16 07:44:52 hrw: i use wrt54 with distro = nylon Jun 16 07:45:13 br1: ah... Jun 16 07:45:23 shorewall need update too Jun 16 07:45:47 hrw|work: first thing i did, after getting the first errors Jun 16 07:46:01 br1: can you recommend easy textmode firewall manager? Jun 16 07:46:14 vi Jun 16 07:46:16 Noid: current OE, current bitbake? strange Jun 16 07:46:22 koen: h.. Jun 16 07:46:25 hrw: though I am thinking to flash oz/gpe 3.53 these days and I also have an asus wl-hdd that is waiting for me to play with Jun 16 07:46:46 hrw|work: i just did a svn to get bitbake, and oo Jun 16 07:46:54 oe i mean Jun 16 07:46:57 pgas: c7x0 machine? Jun 16 07:47:04 hrw|work: yes 860 Jun 16 07:47:22 hrw: hmm... "mc" maybe, but i dont use it Jun 16 07:47:25 pgas: better use post 3.5.3 images so you will get 2.6 kernel Jun 16 07:48:09 3.53 doesn't have 2.6 kernel? I've seen an update on the arm kernel page is this what you mean? Jun 16 07:48:45 pgas: OZ 3.5.3 with gpe does not have 2.6 on c7x0 because some things was broken in gpe then Jun 16 07:49:34 hrw: i currently work mostly with the snapshot of oe that i released as nylon stable, but i'll look at these things soon Jun 16 07:49:59 hrw|work: it works fine on my corgi Jun 16 07:50:15 koen: post 3.5.3? yes Jun 16 07:50:23 koen: It was broken when 3.5.3 shipped Jun 16 07:50:44 ehe, do I need to rebuild the whole thing or I just need to use the kernel found here http://www.rpsys.net/openzaurus/packages/ with the 3.53? Jun 16 07:51:16 pgas: It was the xserver that was broken Jun 16 07:52:59 br1: ok - I have some progress too Jun 16 07:56:39 hrw: for the wrt i currently build nylon-image-mini, i don't know if i't necessary to make a distro for it Jun 16 07:57:05 br1: ok - I will try with nylon Jun 16 07:57:18 hrw|work: it was a path problem, it seems Jun 16 07:57:31 br1: maybe we should just remove wrt54oe.conf from conf/distro or add there info "use nylon" Jun 16 07:57:36 so what do you advise me to do? download and install the latest release and use ipkg upgrade? or do you have bootleg images somewhere? (I'm not ready to go through oe ...yet) Jun 16 07:57:51 hrw|work: see boxer.conf Jun 16 07:57:58 just make it print out and fail Jun 16 07:58:11 hrw|work: the openwrt guys may be willing to switch to oe. might want to pull one of them over here and show them what we've got. Jun 16 07:58:22 yeah, i think so. not every piecoe of hardware needs it's own distro :) Jun 16 07:58:30 kergoth: thats also good idea Jun 16 07:58:31 less is more Jun 16 07:58:37 heh Jun 16 07:58:44 good morning mickeyl Jun 16 07:58:48 ehm Jun 16 07:58:52 good morning kergoth Jun 16 07:59:00 ~mornings Jun 16 07:59:02 Mornings MUST be destroyed! (see also http://www.destroymornings.com/) Jun 16 07:59:06 ~botsnack Jun 16 07:59:07 kergoth: aw, gee Jun 16 07:59:50 kergoth: that would be good, yes. but i'd rather wait until we have a state that's compareable to openwrt Jun 16 07:59:59 * kergoth nods Jun 16 08:00:09 atm no modules, no networking, no shh, no login ;( Jun 16 08:00:10 CIA-9: moved mtd snapshot to 20050615 Jun 16 08:00:37 but that's just a matter of fixing the init scripts Jun 16 08:01:27 another thing is: it would be nice to be able to use sstrip to make even smaller binaries, but that breaks the automatic dependencis for shlibs Jun 16 08:01:45 br1: the question is.. sysvinit (like OZ etc) or busybox init (like openwrt)? Jun 16 08:01:51 call sstrip after shlibs? Jun 16 08:02:07 koen: yeah would be good Jun 16 08:02:15 * mickeyl votes for sysvinit just for the sake of flexibility Jun 16 08:02:32 sysvinit is a lot cleaner, even if you dont use the multiple runlevels Jun 16 08:02:32 * br1 thinks sysvinit is oversized Jun 16 08:02:33 heh Jun 16 08:02:37 yep, it is Jun 16 08:02:51 one thing is important - 4M flash on wrt54g Jun 16 08:03:00 * CosmicPenguin wants to see somebody do something important with init-ng Jun 16 08:03:08 * mickeyl agrees with CosmicPenguin Jun 16 08:03:19 wrt54gs has 8M - ~6M free under openwrt Jun 16 08:03:25 if you want to fit things on 2MB you dont want to waste any byte Jun 16 08:03:42 I need to get a wrt54gs type thingy Jun 16 08:03:59 <[g2]> what is the status of wrt54g / openwrt in OE ? Jun 16 08:04:02 jffs2 garbage collector still needs 0.5M free for itself to behave? Jun 16 08:04:05 I need to convince myself it's worth E75 Jun 16 08:04:24 [g2]: you can build pacakges but not replace firmware yet iirc Jun 16 08:04:33 hrw|work: the wrt would be a really good candidate for unionroot/unionfs. Jun 16 08:04:34 koen: or get it donated like I got Jun 16 08:04:51 sd card hack! Jun 16 08:04:56 (jffs2) looks like it. it does not even want do delete anything atm Jun 16 08:04:59 <[g2]> hrw|work, cool, who's driving the work ? Jun 16 08:05:05 kergoth: you have any idea where things can go wrong when upon 'bitbake nano-1.2.1' 1.3.5 gets built? Jun 16 08:05:33 i have no idea where to start debugging that Jun 16 08:05:53 mickeyl: that isnt handled specially, thats just a provider. bitbake.conf has ${PN}-${PV} in PROVIDES by default. nano 1.3.5 should not have that in provides. Jun 16 08:05:55 [g2]: br1 currently Jun 16 08:05:57 g2: i did a couple of things recently Jun 16 08:06:00 so thats the direction i'd look in Jun 16 08:06:06 kergoth: ah right, thanks for the pointer Jun 16 08:06:09 np Jun 16 08:06:18 [g2]: I'm starting - flashed openwrt two days ago Jun 16 08:07:31 br1: one thing which we can do in other way then openwrt - set root pass in image and do not have telnetd at all Jun 16 08:08:33 anyway - I have to go home (its too hot here) - be back in ~20-30 minutes Jun 16 08:08:38 cu soon Jun 16 08:08:49 * koen farts on telnet Jun 16 08:08:50 <[g2]> neat... Jun 16 08:09:09 hrw: agree Jun 16 08:09:39 <[g2]> it was about 1.5 years ago when I worked on openwrt Jun 16 08:10:02 sweet, tinc is in OE Jun 16 08:10:06 <[g2]> nearly 0 for the last 10+ months, since nslu2-linux got rolling Jun 16 08:10:21 koen: sure ;) Jun 16 08:11:03 <[g2]> hrw|gone, br1 I think they've been playing with CPE replacement Jun 16 08:11:35 g2: would be nice, since CPE is quite big Jun 16 08:11:58 but i have not seen anything in this direction in current openwrt.experimental Jun 16 08:12:04 cpe? Jun 16 08:12:24 <[g2]> beewoolie's APEX bootloader for the ARM can now load the kernel from jffs2, that's what I'm starting to use with Openslug Jun 16 08:12:32 koen: the boot loader of wrt's Jun 16 08:12:40 aha Jun 16 08:12:42 <[g2]> cfe Jun 16 08:12:44 ~curse bootloaders Jun 16 08:12:46 May you be reincarnated as a Windows XP administrator, bootloaders ! Jun 16 08:12:52 <[g2]> common firmware environment ? Jun 16 08:13:32 <[g2]> it's been awhile and I don't miss the broadcom or closed arch much Jun 16 08:13:41 <[g2]> s/much/at all/ :) Jun 16 08:14:27 ah yeah, it's CFE Jun 16 08:23:26 re Jun 16 08:27:00 anyone familiar with the opie-image target here? Or is this a better question for #opie? Jun 16 08:27:11 Crofton: ask Jun 16 08:27:34 I need to build a very simple opie image Jun 16 08:28:02 I have some code written here that works on opie, and I need to work on getting everything in the OSK Jun 16 08:28:15 Crofton: you can create own image if needed Jun 16 08:28:19 of course I could just put most of it on the microdrive Jun 16 08:28:34 yeah, what I am wondering is what bits I should keep Jun 16 08:28:51 Crofton: atleast task-opie-base Jun 16 08:29:11 Crofton: or just install minimal and then ipkg install task-opie-base Jun 16 08:30:56 so, once i have built everything in the build directory, i just make an image out of that? Jun 16 08:31:30 i can't seem to find much more info besides bitbaking packages... i'm new regarding embedded stuff Jun 16 08:31:55 Noid: look into packages/meta/*-image.bb Jun 16 08:32:00 hrw: have been afk for the last ~5 hours .. any news ? Jun 16 08:32:37 hrw: ahhh i see Jun 16 08:32:41 ade|desk: USE_NLS_gcc-cross = "no" and it works Jun 16 08:34:18 hrw: well, i'm building xfce-image, that should bring along the base system then? i just bitbake the rest of the packages and it should be ready to go? Jun 16 08:35:11 Noid: bitbake xfce-image Jun 16 08:35:49 hrw: yes, i'm halfway through that, i then do bitbake firefox sylpheed gaim to install those as well, no? Jun 16 08:36:06 no, that will just build them. that doesnt put them in an image. Jun 16 08:37:38 hmmmm is there a tutorial about the concepts of all this? i'm a bit lost Jun 16 08:39:30 bitbaking a package builds it and emits an ipk or ipks of it. the -image .bb files take a given set of packages, builds them, then installs those ipks into an image and creates an image file of the format you request (jffs2, tar, whatever). to add more packages to _that_ image, there are generally variables you can manipulate in your local.conf to do so Jun 16 08:39:30 Noid: OE build stuff and package it. you can create rootfs (image) with some packages and then install any extra Jun 16 08:42:46 hmmm so i can bitbake xfce-image, create a rootfs and then add firefox and friends to it? Jun 16 08:44:15 bitbaking xfce-image does create a rootfs and an image of that rootfs. Jun 16 08:44:30 ah nice Jun 16 08:44:47 add firefox to DEPENDS and RDEPENDS in local.conf Jun 16 08:45:08 Crofton: no Jun 16 08:45:22 woops Jun 16 08:45:33 Crofton: BOOSTRAP_EXTRA_(R)DEPENDS Jun 16 08:45:34 BOOTSTRAP_EXTRA_DEPENDS Jun 16 08:45:36 yep Jun 16 08:45:42 I just checked mine Jun 16 08:45:58 * Crofton still has to look to be sure about things Jun 16 08:47:30 hrw: which gcc/binutils version have you been using for the wrt? Jun 16 08:47:42 To control what actually makes it to the image in opie-image.bb, I need to edit the IPKG_INSTALL line? Jun 16 08:48:25 hmmm i have no such var in local.conf... the most similar thing i've found so far is XFCE_DEPENDS in xfce-image.bb Jun 16 08:48:37 You may need to add it Jun 16 08:48:53 look at bootstrap-image.bb Jun 16 08:49:01 task-bootstrap.bb Jun 16 08:49:32 koen: any idea where to add sstrip? Jun 16 08:49:38 That is where I started from Jun 16 08:49:55 br1: I built gcc 3.4.4 and binutils 2.15.94.0.1 Jun 16 08:50:08 br1: but didnt checked anything on wrt yet Jun 16 08:50:45 cool! i got gcc 3.3.4 and binutils 2.14.90.0.7 Jun 16 08:51:01 gcc 3.4.4 supposedly build smaller binaries Jun 16 08:51:27 br1: I just said OE to build gcc-cross in version which I know from arm Jun 16 08:51:37 iirc gcc-cross 4.0.0 failed Jun 16 08:51:53 and have to check gcc-cross with java build Jun 16 08:52:54 any ideas how to avoid depmod? it's too big Jun 16 08:53:17 i think i've found it in machine/x86.conf file... Jun 16 08:53:36 br1: nope - we had that problem with Zaurus year ago Jun 16 08:53:55 should i add the packages to both BOOSTRAP_EXTRA_ variables? Jun 16 08:54:09 Noid: DEPENDS are for build, RDEPENDS are for image Jun 16 08:54:19 so both it is Jun 16 08:54:25 hrw: it's not very likely that modules change on a device like the wrt Jun 16 08:54:43 br1: but I can install module later Jun 16 08:55:04 true... Jun 16 08:55:36 br1: perhaps we can force remove depmod from rootfs after creation and if someone will try to install kernel module then depmd will be fetched also Jun 16 08:56:18 hrw: we do't have to install it in the first place. but then /ectc/init.d/modutils.sh fails Jun 16 08:56:42 br1: if you have depmod in the kernel module ipk deps, then its in the deps of the kernel module packages that got put into the rootfs at image creation time as well. Jun 16 08:56:51 br1: which is why you'd have to force its removal, which is what hrw just said Jun 16 08:57:09 hrw: great :) Jun 16 08:57:13 kergoth: ah right Jun 16 08:57:26 I think it'd be reasonable to demote that dependency to a RECOMMENDS. Jun 16 08:57:31 or RRECOMMENDS, even Jun 16 08:57:34 yeah Jun 16 08:57:43 after all, the modules will still work without depmod, you just can't use modprobe to load them. Jun 16 08:58:38 pb_: or user can download/install depmod into /tmp/ and run it by hand Jun 16 08:59:50 brb.. Jun 16 09:00:07 we'd need another /etc/init.d/modutils.sh then Jun 16 09:01:19 we can make it Jun 16 09:02:31 we could even make the current one able to hande situations where depmod and modprobe is not there Jun 16 09:02:47 modprobe is in busybox Jun 16 09:03:15 i know but without modules.dep it does not work Jun 16 09:03:23 ofcourse Jun 16 09:06:13 I don't believe you have pulled the cross-build depmod patches for big endian back into oe-devel (they're in nslu2-linux). Not sure about that - (module-init-tools/files/depmod-byteswap.patch) Jun 16 09:06:30 Without that you can't run depmod on the compile host for a cross-byte-sex target. Jun 16 09:06:53 Without that the system won't boot unless /etc/init.d/modutils.sh is able to run depmod - because modprobe is required for boot. Jun 16 09:20:03 ok, so you agree if i make a modutils.sh which also works with insmod instead of depmod + modprobe Jun 16 09:20:09 ? Jun 16 09:20:23 basically yes Jun 16 09:20:39 I can even call you wrt54 maintainer Jun 16 09:20:58 and has everybody here registered with bugzilla yet? Jun 16 09:21:20 i don't want to be officiall wrt maintainer Jun 16 09:21:39 i can do the work but i cant do it officially Jun 16 09:22:08 ah.. ok Jun 16 09:22:13 wrt is competition for meshcube Jun 16 09:22:47 ah.. forgot that you work there Jun 16 09:23:35 koen: yes, i'm registered with bugzilla Jun 16 09:23:57 br1: cool Jun 16 09:24:07 it's not that bad, but at least i cant do it with the 4g email address Jun 16 09:24:41 hrw: I got a mail from Xora "I'm on honeymoon, but you can add me to bugzilla if there are bugs" Jun 16 09:24:58 koen: ;) Jun 16 09:25:15 my response: "You're on honeymoon, stop thinking about bugs" Jun 16 09:25:24 exactly Jun 16 09:25:50 * koen throws some rice in Xora's direction Jun 16 09:25:57 * koen throws some rice in Florian's direction Jun 16 09:26:03 any gcc hackers here? I want to disable Java for mipsel Jun 16 09:26:04 lot's of people getting married Jun 16 09:26:20 its "lot's" or "lots" in english? Jun 16 09:26:28 don't ask me :) Jun 16 09:26:35 lots Jun 16 09:26:37 those apostophes are killing me Jun 16 09:26:47 +r somewhere Jun 16 09:27:03 ~spell apostrophes Jun 16 09:27:05 'apostrophes' may be spelled correctly Jun 16 09:27:07 apostrophes are used for contractions and possessives Jun 16 09:27:30 and the think nobody (almost) understands is that contractions have a higher priority Jun 16 09:27:35 jacques: one sentence and two unknown words - nice:) Jun 16 09:27:58 hence, it's is a contraction for "it is" and not the posessive Jun 16 09:28:14 possessive Jun 16 09:29:13 ppl also frequently screw up their they're and there as well as your and you're Jun 16 09:29:31 br1: you built gcc-cross with java support or without? Jun 16 09:30:12 hrw: umm, dunno, without i guess Jun 16 09:30:18 anyway to answer the original question, AFAIAA "lot's" is not a word Jun 16 09:30:21 who needs java ;) Jun 16 09:30:30 or fortran Jun 16 09:30:31 br1: I do Jun 16 09:30:36 but only for coffee :) Jun 16 09:30:36 or objc Jun 16 09:30:45 or c++ Jun 16 09:30:45 or ada Jun 16 09:30:46 br1: I want to disable it for mipsel (it does not build with uclibc anyway) Jun 16 09:31:25 hrw: i didn't do anything to disable it, but it's disabled for 3.3.4 Jun 16 09:31:33 * jacques goes back to reading about wine. Jun 16 09:32:01 br1: atleast was Jun 16 09:32:25 anyone know if the kernel on simpad has USB net support ? Jun 16 09:32:49 2.4 or 2.6? Jun 16 09:32:52 2.4 or 2.6? Jun 16 09:32:57 heh Jun 16 09:33:13 Spyro: 2.6 does not have usb gadget for strongarm Jun 16 09:33:20 under 2.4 I dont know Jun 16 09:38:29 hrw: what i had to do is USE_NLS_gcc-cross = "no" to build with uclibc Jun 16 09:38:39 br1: me too Jun 16 09:39:33 couldn't that be set automatically when TARGET_OS==linux-uclibc? Jun 16 09:40:29 rather could Jun 16 09:43:00 br1: kismet fails: | checking host system type... Invalid configuration `mipsel-linux-uclibc': system `uclibc' not recognized Jun 16 09:43:26 03mickeyl * r256 10bitbake/lib/bb/shell.py: Jun 16 09:43:27 - add 'getvar' to print out a variable from the outer environment Jun 16 09:43:27 - 'print' now serves a different purpose, it lists either all bbfiles or all providers Jun 16 09:43:27 - add '|' to allow the output of commands be piped into grep, sort, and friends: Hint try: Jun 16 09:43:27 parse; print files | grep opie Jun 16 09:43:27 - internal: add debugOut() Jun 16 09:43:36 i have not compiled kismet yet with uclibc Jun 16 09:44:03 wooohoooo! Jun 16 09:44:03 and finally mickeyl sent 0x100 rev of bitbake Jun 16 09:44:09 tadaa :)) Jun 16 09:44:14 I won the bet! Jun 16 09:44:44 mickeyl: can I "parse;build `print files|grep opie`"? Jun 16 09:45:01 hrw: ah no, it's not that sophisticated Jun 16 09:45:05 a good idea though Jun 16 09:45:08 * mickeyl adds to IDEAS Jun 16 09:45:22 mickeyl: or just "build *opie*" Jun 16 09:45:28 *nod* Jun 16 09:45:31 * mickeyl adds to IDEAS as well Jun 16 09:45:44 * koen adds a wildcard to IDEAS Jun 16 09:45:59 or Jun 16 09:46:18 build a([b-d,g-u])*-1.[3,5,7] Jun 16 09:46:43 urghs Jun 16 09:46:44 bummer Jun 16 09:46:50 TKOM people arriving Jun 16 09:46:54 my neighbours phone still non worky Jun 16 09:46:59 i may be disconnected every minute now Jun 16 09:47:14 have fun Jun 16 10:09:27 hi [cc]smart Jun 16 10:15:29 mickeyl: do "parse;" in bitbake shell Jun 16 10:16:05 hrw: what do you think? one big jffs2 partition for the wrt or "openwrt style"? Jun 16 10:16:16 br1: one big imho Jun 16 10:16:30 thinks so to ;) Jun 16 10:17:10 "big", well... ;) Jun 16 10:17:16 ;)) Jun 16 10:17:41 I have to reflash my wrt for one jffs2 Jun 16 10:19:01 * chouimat|ibook waves Jun 16 10:21:12 * chouimat|ibook is away: shower Jun 16 10:22:16 just commited a new modutils.sh Jun 16 10:23:00 which can also work without depmod and modprobe Jun 16 10:23:50 cool Jun 16 10:33:46 * chouimat|ibook is back. Jun 16 10:52:33 ok, so we definitly have a problem here: Jun 16 10:52:58 DEBUG: providers for nano-1.2.1 are: ['nano'] Jun 16 10:53:00 that's simply wrong Jun 16 10:53:25 somewhere in the first 50 lines of buildProvider() things go wrong Jun 16 10:54:11 ~seen proti Jun 16 10:54:13 ~seen zecke Jun 16 10:54:13 proti <~seb@calypso.frankengul.org> was last seen on IRC in channel #oe, 10d 9h 4m 5s ago, saying: 'To me, something will be lost in the move.'. Jun 16 10:54:14 zecke <~ich@83-169-170-199-dynip.superkabel.de> was last seen on IRC in channel #handhelds.org, 8d 18h 36m 42s ago, saying: 'good nite'. Jun 16 10:55:09 time to update to 20050615 opie Jun 16 10:56:11 and to fix bbimage again ;( Jun 16 10:56:33 hrw: is there new feeds? Jun 16 10:56:55 gints|wrk: my own one Jun 16 10:57:59 hmm. no - bbimage was ok Jun 16 11:19:02 treke|home: ping Jun 16 11:19:14 treke|home: you should set the default notification email to bitbake-dev@lists.berlios.de Jun 16 11:19:19 instead of greg@treken.net Jun 16 11:19:24 treke.net even Jun 16 11:19:37 (for bugs filed against bitbake) Jun 16 11:24:42 mickeyl: readline support stopped working for me in bbshell sometime. any ideas about that? Jun 16 11:25:13 yeah Jun 16 11:25:20 that's been fixed since a week or so Jun 16 11:25:46 keturn: add "svn up" in enviroment-setup script :) Jun 16 11:26:33 should have tried that first. Thanks. Jun 16 11:27:20 I usually do: cd ~/zaurus/oe;. setup.sh build/0602 - this will update bitbake to newest, bk pull, bk -q co -r, set PATH to get 2.95.3 for zauruskernels, add bitbake/bin/ to PATH, set LC_ALL=C so everyone understand my error messages and say "all done" Jun 16 11:28:08 does it go "beep" to? Jun 16 11:28:21 koen|tv: nice idea.. Jun 16 11:28:21 I have to be careful about the "bk pull" because it sometimes takes a half day of work to integrate the changes Jun 16 11:28:44 keturn: I usually dont keep changes long.. Jun 16 11:29:02 (like when it started giving me new versions of compilers to use that don't work here) Jun 16 11:29:04 koen|tv: it say: echo "Altered environment for BB Zaurus Development ARM" Jun 16 11:29:42 keturn: ever heard about PREFERRED-VERSION_packagename = "the version which I like to use"? Jun 16 11:30:10 I have to go to shop - forget about bread... Jun 16 11:30:11 hrw: yes, and I have that set for gcc now. but I didn't used it. Jun 16 11:30:28 s/used it/used to/ Jun 16 11:30:35 happens Jun 16 11:30:59 until today I didnt have to use USE_NLS flag Jun 16 11:31:37 mickeyl: can you add some magic to gcc3 to set USE_NLS=no when TARGET_OS = linux-uclibc? that python magic scares me Jun 16 11:32:24 just USE_NLS? Jun 16 11:32:27 no export? Jun 16 11:32:58 which gcc3 should I tweak? Jun 16 11:33:17 can someone add it to documentation.conf too? Jun 16 11:37:36 hmm, readline in 'python' expert mode of bbshell is still gone Jun 16 11:39:48 yeah readline in python mode is gone. was it ever there ? Jun 16 11:40:18 not sure for certain Jun 16 11:40:23 i have an idea Jun 16 11:40:26 * mickeyl checks Jun 16 11:42:42 Ciao all Jun 16 11:50:15 03mickeyl * r257 10bitbake/lib/bb/shell.py: Jun 16 11:50:16 - add 'showdata' command Jun 16 11:50:16 - fix readline in python expert mode Jun 16 11:50:16 - add some more IDEAS Jun 16 11:51:43 woot Jun 16 11:51:55 thanks for spotting Jun 16 11:53:16 ibot: botmail to zecke: please look into bb bug #103 Jun 16 11:53:22 ibot: botmail to proti: please look into bb bug #103 Jun 16 11:53:36 * mickeyl sighs and watches TV Jun 16 11:56:27 mickey|tv: in conf/local.conf I have USE_NLS_gcc-cross="no" - without it gcc-cross does not build with linux-uclibc Jun 16 11:56:54 hrw|afk: no problem using a base_conditional. i'll take care about it Jun 16 11:58:00 mickey|tv: thx Jun 16 12:05:33 mickey|tv: read your bitbake mail Jun 16 12:05:41 hi zecke Jun 16 12:05:43 koen|tv: bastard ;) Jun 16 12:05:51 koen|tv: I will get a Nokia 770 as well Jun 16 12:05:53 * zecke uncans his first Gtk+ app Jun 16 12:06:02 ! Jun 16 12:06:06 zecke: cheers Jun 16 12:06:09 go aliens! Jun 16 12:08:12 oops. broke bbshell by doing reload(bb.shell) in expert mode. I have to actually restart it. Jun 16 12:09:35 03pb 07 * r1.3536.1.21 10openembedded/packages/lirc/lirc-modules_0.7.1.bb: use $D, not ${D} Jun 16 12:12:50 evening Jun 16 12:13:14 hi reenoo_ Jun 16 12:13:43 03pb 07 * r1.3542 10openembedded/packages/xserver/xserver-xorg_6.8.99.10.bb: round up escaped Xprint bits Jun 16 12:13:51 heh Jun 16 12:14:08 someone has been watching too much westerns Jun 16 12:14:38 given the list returned by cooker.status.providers[foo], how do I find which of those is preferred? Jun 16 12:15:31 hey pb_ Jun 16 12:17:04 mickey|tv: do not blackmail me ;) Jun 16 12:20:43 koen, Should I send the bb files to oe ML ? Jun 16 12:22:55 attach them to bugs.openembedded.org bugzilla Jun 16 12:23:00 I will Jun 16 12:23:02 that makes sure they wont get lost Jun 16 12:25:47 I want to disable Java for mipsel - any objections? Jun 16 12:26:15 hrw: why? Jun 16 12:26:41 reenoo_: does not build - atleast for mipsel-linux-uclibc Jun 16 12:30:26 * zecke needs to start with the assignment Jun 16 12:31:21 ~hail drw Jun 16 12:31:22 * ibot bows down to drw and chants, "I'M NOT WORTHY!!" Jun 16 12:31:57 wellenreiter with OResource does not require magnify glass to press buttons Jun 16 12:32:07 fluff Jun 16 12:32:21 hrw: hmm. odd. Jun 16 12:32:57 uclinux: ultra castrated linux Jun 16 12:33:12 that was the first time I heard about it Jun 16 12:33:37 reenoo_: but true :( Jun 16 12:35:56 hrw: well. odd as in I've been told gcj worked on/for mipsel Jun 16 12:38:01 reenoo_: http://bugs.openembedded.org/show_bug.cgi?id=68 Jun 16 12:39:49 oh, that's the argz thing? Jun 16 12:39:58 someone just needs to debug why configure is making a mistake. Jun 16 12:40:07 I don't think that should be an intractable bug. Jun 16 12:40:15 ok Jun 16 12:40:48 you might try gcc 4.0, too, unless there is a compelling reason to stick with 3.3.4. Jun 16 12:40:50 er, 3.4.4 Jun 16 12:40:54 koen, are the login for bugs.openembedded.org the same that I use on bugs.hh.org or I should create a new account ? Jun 16 12:41:07 Pigi: you'll need to create a new account. they are completely separate. Jun 16 12:41:24 ok thx Jun 16 12:42:01 pb_: in monday I can try Jun 16 12:43:07 will we go back to using rcs manually soon? Jun 16 12:43:35 zecke: no, we'll switch to bk Jun 16 12:43:45 koen: Noooooooo! Jun 16 12:44:07 zecke: but the license requires you to never eat dutch food again Jun 16 12:44:15 and speak french all the time Jun 16 12:44:30 zecke: is mq200 accelerated ? Jun 16 12:44:46 Spyro: in contrast to a SA1100fb? Jun 16 12:44:52 zecke: yeah Jun 16 12:44:55 Spyro: no it is not, but we could Jun 16 12:45:01 zecke: we have docs ? Jun 16 12:45:18 Spyro: one can find docs, and the Lineo Jp driver have hooks for hw Jun 16 12:45:28 ~google mq200 pdf Jun 16 12:45:31 zecke: cool. I'll have to investigate Jun 16 12:45:40 zecke: is mq200 a pure GPU or a SoC ? Jun 16 12:45:53 yeah, a kaa driver for mq200 would be a nice project for someone. Jun 16 12:45:54 Spyro: pure GPU Jun 16 12:46:10 zecke: I'll get on that once I start hacking X Jun 16 12:46:12 Spyro: but you can drive two displays Jun 16 12:46:16 Im doing w100 first, hopefully Jun 16 12:46:17 it'd be interesting to see if it can share any acceleration code with the mq1100, though I suspect probably not. Jun 16 12:46:32 mq11xx accell!! Jun 16 12:46:35 pb_: maybe the fifo locking logic could be shared Jun 16 12:46:48 if I ever get that 2200 board and LCD I will have a crack at mq11xx accel Jun 16 12:47:16 zecke: so how do you 'debrick' a simpad? whats the hardware organisation? Jun 16 12:47:31 Spyro: Best practice is not to brick it Jun 16 12:47:34 zecke: from what I read (I have a t-sinus pad) the thing uses 16 bit flash ? Jun 16 12:47:57 Spyro: the flash has the same width Jun 16 12:48:02 zecke: I want to work on bootloader / 2.6 on simpad Jun 16 12:48:04 Spyro: one has two and one has one bank Jun 16 12:48:17 zecke: is it 16 bit or 32 ? Jun 16 12:48:25 Spyro: 16bit width IIRC Jun 16 12:48:36 ok so XIP is pretty worthless on it then... Jun 16 12:48:37 Spyro: talk to Chicken if the hh.org jtag tools are sufficent Jun 16 12:48:50 ~seen Chicken Jun 16 12:48:51 zecke: heh Jun 16 12:48:51 chicken is currently on #opie (56m 33s) #handhelds.org (56m 33s) Jun 16 12:49:08 well, not that mq1100 acceleration is really all that exciting. Jun 16 12:49:37 zecke: ok, so assuming I had a totally blank simpad (corrupt flash or something), what bootloader would I need to flash (opensimpads site is rather vague) Jun 16 12:49:40 mq11x8 supposedly can to xv Jun 16 12:49:42 mickey|tv: pyflakes says: in pasteBin and pasteLog functions there is reference to undefined symbol "response" Jun 16 12:49:56 how can I add the new bb files in bugzilla if I don't have an "attach" ( or similar ) button ? Jun 16 12:49:56 someone know does mickey|tv tested his pcmciaapplet on c7x0? Jun 16 12:50:09 Spyro: bootldr only Jun 16 12:50:12 Spyro: using jtag Jun 16 12:50:15 Pigi: after you submit the bug, go back and click "attach a file to this bug" Jun 16 12:50:30 it's a bit broken that you can't attach a file during the initial submission, but there you are. Jun 16 12:50:33 eheh I suspected, but I wouldn't seems dumb :) Jun 16 12:50:34 Pigi: "Create a New Attachment (proposed patch, testcase, etc.)" link exist Jun 16 12:50:41 Spyro: we're completely Free of any proprietary firmware Jun 16 12:51:02 Spyro: but why not finish lab for SA? Jun 16 12:51:37 zecke: Id rather work on the 'main' kernel personally - I want to *use* my simpad (unlike the toshibas, which I want to make do cool things) Jun 16 12:52:04 zecke: does current bootldr source compile for simpad ? or is it modded ? Jun 16 12:52:07 Spyro: ok not too much is missing in the 2.6 kernel Jun 16 12:52:20 Spyro: it is modded and contains propietary mq200 code Jun 16 12:52:24 zecke: I'll make a page on my site about all this once I get my head round it... Jun 16 12:52:43 zecke: do we have an open source setup sequence for the mq200 yet ? Jun 16 12:52:45 Spyro: that can be removed, get back to me tomorrow and I send you a diff without the mq200 stuff Jun 16 12:53:03 Spyro: yupp, I found a Lineo Jp. mq200 driver Jun 16 12:53:05 can it use the display with non-proprietary mq200 code ? Jun 16 12:53:14 Spyro: it should Jun 16 12:53:16 or is that a TODO ? Jun 16 12:53:19 cool Jun 16 12:53:41 Spyro: it survies the resume, so it can completely initialize itself Jun 16 12:53:56 so then, basically, what works under 2.6 - afaict we have keypad, LCD, PCMCIA, smartcard, UDC, serial, audio Jun 16 12:54:11 Spyro: sorry I need to do my math assignment... Jun 16 12:54:18 zecke: I understand Jun 16 12:54:29 zecke: I look forward to your patch Jun 16 12:54:33 PCMCIA, Serial are handled by generic SA1100 code Jun 16 12:54:42 cool Jun 16 12:54:43 IrDa is not tested but generic SA1100... Jun 16 12:54:57 UDC in hh.org seems to work but should be converted to the gadget stuff Jun 16 12:55:01 when you send the patch, could you include a defconfig please? Jun 16 12:55:07 nfsroot did not work out of the box Jun 16 12:55:20 bye Jun 16 12:55:25 cu monday Jun 16 12:55:30 Spyro: As far as the kernel is concerned everything is in the 2.6 hh.org tree Jun 16 12:55:31 cya hrw Jun 16 12:55:35 hrw: cya Jun 16 12:55:59 Spyro: we miss implementation for the 'two' keyboards, SmartCard and Audio Jun 16 12:56:02 zecke in that case, could you send a .config to spyro@f2s.com so that I might play with it ? Jun 16 12:56:22 Spyro: a working one is found in the hh.org tree as well Jun 16 12:56:38 cool. config_sipad or somethign I assume ? Jun 16 12:56:40 simpad Jun 16 12:56:59 Spyro: something like that in arch/arm/*config*/*simpad* Jun 16 12:57:17 zecke: ok, last question for tonight then Jun 16 12:57:25 where do I get the modded bootldr from ? Jun 16 12:57:57 (source would be handy) Jun 16 12:58:15 Spyro : hh.org/~zecke/bootldr_simpad_tmp.tar.bz2 Jun 16 12:58:22 cheers! Jun 16 12:58:24 Spyro: but boot it from ram first Jun 16 12:58:39 Spyro: it was a copy we all worked on... Jun 16 12:58:44 zecke: I will make sure I have JTAG prior to fiddling with it, dont worry Jun 16 13:11:57 so printing DEPENDS gives me a list of things, but it doesn't include e.g. the toolchain. Where is that defined? Jun 16 13:12:08 base.bbclass Jun 16 13:15:58 did anyone fix the kdrive/tslib vt switch crash yet? Jun 16 13:16:16 Im trying a fix now but if anyone already did it Id be happy to hear Jun 16 13:16:29 I don't think anyone fixed it yet. Jun 16 13:16:39 ok. I'll report back in a bit Jun 16 13:16:46 XorA and I talked about it last week, and he was planning to work on a patch after he finished getting married. Jun 16 13:16:54 heh Jun 16 13:17:18 perhaps I can save him the effort :) Jun 16 13:35:35 nigth all Jun 16 13:35:52 'night all Jun 16 14:29:56 03mickeyl * r258 10bitbake/lib/bb/shell.py: - fix pasteLog and pasteBin failure codepaths. Thanks to keturn for spotting. Jun 16 14:34:13 re Jun 16 14:39:20 Good evening Jun 16 14:47:19 Wow - we're going to hit a 100 before long Jun 16 14:47:25 people in the room, that is Jun 16 15:18:21 03pb 07 * r1.3544 10openembedded/packages/mythfront/ (mythfront-config.bb mythfront-config/epia/xorg.conf): update mythfront xorg.conf Jun 16 15:21:39 03pb 07 * r1.3545 10openembedded/packages/dri/drm-kernel_cvs.bb: use KERNEL_CC for drm modules Jun 16 15:25:55 Hi guys... Jun 16 15:26:34 pb_: I can confirm that preventing the double closing of tslib in kdrive cures the suspend / resume issues I have been observing. Jun 16 15:26:56 pb_: I'll try to see why kdrive even bothers to attempt the double closure tomorrow. Jun 16 15:27:28 pb_: sadly, my audio codec goes awol during sleep so now I need to fgure out that mystery before e750 is actually useful :/ Jun 16 15:29:08 righto, very good Jun 16 15:29:28 that matches the reports we've had from other people, at least. Jun 16 15:29:30 pb_: I'll try to merge the wolfsen wm97xx ALSA code tomorrow Jun 16 15:29:53 pb_: with luck, it'll fix my touchscreen bug :-) Jun 16 15:30:11 oh, right, is that not merged already? Jun 16 15:30:24 pb_: I didnt merge it, not aware of anyone else using it... Jun 16 15:30:37 pb_: if it is, I'll be testing instead of merging :) Jun 16 15:30:52 ah, maybe I was thinking of the OSS code. Which is the one that Liam Girdwood maintains? Jun 16 15:31:17 Liam wrote the OSS one and maintains it (a bit) for the clients still needing it Jun 16 15:31:24 I modified it and merged it into CVS Jun 16 15:31:33 Liam since wrote an ALSA one Jun 16 15:31:46 but its newish and not TTBOMK merged Jun 16 15:31:54 right Jun 16 15:32:02 I guess it was the older OSS one that I was thinking of. Jun 16 15:32:20 btw, e750 is literally one audio bug and one backlight driver from being 'ready for end users' Jun 16 15:32:38 I'm fairly sure that suspend/resume used to work OK with the oss driver when I was last using it, though. Jun 16 15:32:49 very good! Jun 16 15:33:01 pb_: Im suspecting the wince loader is nuking the audio codec (poss. a GPIO sleep state issue) Jun 16 15:33:11 ah Jun 16 15:33:53 pb_: theres still much to do - WiFi (needs USB host), nand (worked but is broken atm), PCMCIA (needs to be fixed, poss. a userspace issue) VGA out (not done yet) Jun 16 15:34:06 and audio needs some GPIO magic to turn on amps etc. Jun 16 15:34:34 ah, I see Jun 16 15:34:49 where do you store your rootfs, if neither nand nor pcmcia is working? on sd? Jun 16 15:34:53 but it does work reliably Jun 16 15:34:56 SD Jun 16 15:35:07 tmio_mmc Jun 16 15:35:24 sounds good Jun 16 15:35:42 well, its (looks to be) all downhill from here Jun 16 15:36:17 e740 is in a similar condition now, except that its tmio chip is being scuppered during sleep ( again, probably wince loader GPIO tomfoolery) Jun 16 15:37:01 e400 is a e740 minus PCMCIA and plus NAND, so it should work at about the same time I get e740 working (same SD issue) Jun 16 15:37:27 e330 uses mediaQ for video and is less complete (X doesnt run well at present) Jun 16 15:43:42 Spyro: you're insane ;) Jun 16 15:43:52 zecke|assignment: but in a good way ;-) Jun 16 15:48:19 sure Jun 16 15:48:29 but it is insane nevertheless Jun 16 15:51:26 zecke|assignment: hehe. Jun 16 15:51:30 good night guys :-) Jun 16 15:51:37 cya Jun 16 15:51:42 'night Spyro Jun 16 15:51:55 Anyone know a compiler that works properly for Cxx0? Jun 16 15:52:08 luke-jr_: Intel CC? Jun 16 18:20:29 bleh Jun 16 18:30:59 bleh? Jun 16 19:46:31 meh Jun 16 19:49:11 heya raster Jun 16 19:49:20 how goes it? Jun 16 19:49:33 emte: emmie! Jun 16 19:49:36 ola! Jun 16 19:50:01 yeah end of next week i can put a LOT more time into biohazard Jun 16 19:53:52 how you been? Jun 16 19:55:12 ey, 'd that be raster as-an-example the feller who does imlib2? Jun 16 19:55:51 all of EFL not just imlib2 Jun 16 19:56:04 yeah but imlib2 is preticularly purdy Jun 16 19:56:11 not really Jun 16 19:56:17 well I LIKE IT Jun 16 19:56:19 so there. Jun 16 19:56:26 you using e17? Jun 16 19:56:36 nay Jun 16 19:56:52 try it and then see if your opinion is the same Jun 16 19:57:00 :) Jun 16 19:57:11 my opinion is sweet Jun 16 19:57:17 like big juicy tomato Jun 16 19:57:50 a fruit that thinks it's a vegetable? Jun 16 19:57:54 :P Jun 16 19:58:21 i gave e a look see and it just doesn't compare to WindowMaker + middlewheel scroll for changing desktops + Windows Key bound to mod4 for Z order and manipulation Jun 16 19:58:34 for like, my uses Jun 16 19:59:21 * emte wonders where that came from Jun 16 20:00:07 ... Jun 16 20:00:50 * jnc blinks Jun 16 20:00:58 e17 on an embedded device? Jun 16 20:01:09 whatfor Jun 16 20:01:16 yeah yeah yeah . .. i am working oin it Jun 16 20:01:33 my course load takes up a lot of time Jun 16 20:01:53 i hadn't thought of that Jun 16 20:01:57 it might be nice to see Jun 16 20:02:15 well for one, you dont NEED xfree/xorg to run it Jun 16 20:02:24 emte: sorry - office - peolpe at desk asking me stuff Jun 16 20:02:31 providing enough peopel write native apps Jun 16 20:02:40 raster, np Jun 16 20:03:01 i am trying to finish a C lab i skipped Jun 16 20:03:14 simple, stupid lab at that Jun 16 20:03:29 jnc: u havent played with e17 then i guess :) Jun 16 20:03:36 emte: eew.! labs! Jun 16 20:03:40 lol Jun 16 20:03:57 well i have PIC asm, VHDL, and C all at once Jun 16 20:04:04 nice mix of languages Jun 16 20:04:28 a tomato Jun 16 20:04:40 a fruit that thinks it's a vegetable? Jun 16 20:04:43 fuck vhdl Jun 16 20:04:46 lol Jun 16 20:04:56 not your favorite language i take it? Jun 16 20:04:57 but asm and c go togther like bacon and eggs Jun 16 20:04:58 yum Jun 16 20:04:59 :) Jun 16 20:05:02 lol Jun 16 20:05:10 actually VHDL isnt bad Jun 16 20:05:16 quite a lot of uses Jun 16 20:05:33 i am not sure i'd like to try and write a state machine in C Jun 16 20:05:52 its easy Jun 16 20:05:57 u have boejcts they have states Jun 16 20:06:03 at imes u evaluate a new state Jun 16 20:06:04 lol Jun 16 20:06:08 so u keep revioous and current state Jun 16 20:06:17 on evaluation current state gets copied to previous (if ti changed) Jun 16 20:06:29 well you will have to show me a C state machine example one day Jun 16 20:06:33 u build an uodate set Jun 16 20:06:37 i'd be interested Jun 16 20:06:42 update set (what to do to go from prev to cur state) Jun 16 20:06:43 then do them Jun 16 20:06:49 evas Jun 16 20:06:50 :) Jun 16 20:07:02 lol Jun 16 20:07:10 okay how about a simple example? Jun 16 20:07:27 maybe two states or something Jun 16 20:12:21 anyway i named the embedded wm e17biohazard , since that will probably be what it is by the time i get done mangling it Jun 16 20:17:02 hahaha Jun 16 20:17:08 2 states Jun 16 20:17:12 well simple Jun 16 20:17:17 a box Jun 16 20:17:21 is at x, y Jun 16 20:17:29 nbow u move it to x2,y2 Jun 16 20:17:37 (change its state) Jun 16 20:17:46 then u move it again to x3, y3 Jun 16 20:17:50 now u evaluate state Jun 16 20:18:10 it currently is at x,,y but must go TO x3,y3 Jun 16 20:18:20 operation -> Jun 16 20:18:28 new_x = x3; Jun 16 20:18:31 new_y = y3 Jun 16 20:18:33 done. Jun 16 20:18:35 realyl simple Jun 16 20:19:06 how do you address cases where it "idles" or goes "back" ? if statements? Jun 16 20:19:50 at that point it seems to sound like you need an evil goto in the case of going back Jun 16 20:21:39 hmm Jun 16 20:21:46 actually it doesnt sound too bad ... Jun 16 20:25:24 eh? Jun 16 20:25:31 idles? Jun 16 20:25:33 goes back? Jun 16 20:25:36 if u go back Jun 16 20:25:39 its another state change Jun 16 20:25:48 simply to tbe previous state Jun 16 20:25:53 idling - well it does nothing. Jun 16 20:26:00 nothnig to do. Jun 16 20:27:06 yup Jun 16 20:27:14 was thinking about that after i asked Jun 16 20:27:38 now i'll have to think of something i need a statemachine for ... maybe the menu :P Jun 16 20:43:37 What You Dels? Jun 16 21:08:34 dels? Jun 16 21:08:48 devs ? developers? **** ENDING LOGGING AT Thu Jun 16 23:59:57 2005