**** BEGIN LOGGING AT Fri Sep 14 02:59:56 2007 Sep 14 03:04:26 great, the DVD recorder crash managed to somehow corrupt the DVD-RW, and now the recorder thinks it's empty and asks if we want to format... Sep 14 03:11:47 cesarb: did you see this one -> http://xkcd.com/117/ Sep 14 03:37:12 Im starting to prefer the Poko UI over the OM version Sep 14 03:46:37 not Poko I meant Poky Sep 14 03:47:05 http://www.pokylinux.org Sep 14 04:47:03 moin Sep 14 04:47:24 evenin' Sep 14 04:47:45 * rwhitby seems to be in stable build land now - all incremental builds started this morning have worked. Sep 14 04:48:56 Nice Sep 14 04:49:22 rwhitby: The problem I had with emotion seems to be a libpng error on my 64 bit: http://bugs.openembedded.org/show_bug.cgi?id=2992 Sep 14 04:49:27 and free42 is in the official feed now Sep 14 04:50:11 * CM was used to the TI82 and such.. Sep 14 04:52:23 And yes, there should be some celebration after those two earlier commits.. ;) Sep 14 04:57:52 wtf, i broken my svn Sep 14 04:59:10 do anyone of you knows how to change the revision number to some files? Sep 14 04:59:26 s/do// Sep 14 04:59:26 bedboi meant: anyone of you knows how to change the revision number to some files? Sep 14 05:00:38 damn Sep 14 05:07:42 rwhitby: this morning it came right for me too Sep 14 05:07:50 rwhitby: well, builds anyhow Sep 14 05:07:53 gsmd is still cancerous Sep 14 05:08:12 console remove and restart in matchbox session still works for me Sep 14 05:10:20 you mean to avoid the serial/uart hang? Sep 14 05:10:26 i have a work around for that Sep 14 05:10:37 but gsmd still needs to be killed and restarted for me Sep 14 05:10:40 which means restarting x Sep 14 05:10:47 (becaues otherwise the dialer wont talk to it) Sep 14 05:11:19 cw666: gmsd restart in /etc/matchbox/session once at boot works for me Sep 14 05:12:21 rwhitby: ah right, sorta a hack though Sep 14 05:12:34 i started looking at gsmd and am not sure how best to even attack it Sep 14 05:40:22 apt: mv oldfile newfile Sep 14 05:41:30 lol @ dialup Sep 14 05:44:06 actually I'm on DSL, but my stupid ISP allocates me to the dialup IP range for some reason Sep 14 06:01:04 could anyone enlighten me as to what the lowlevel foo .... bin files are? Sep 14 06:03:51 they contain initialization code so you can boot from RAM, without a bootloader. see http://wiki.openmoko.org/wiki/Bootloader#Using_JTAG_to_boot_from_RAM Sep 14 06:04:36 tuukkah: thanks Sep 14 06:10:47 to be fair.. lowlevel-foo is actually named very exact after what it does ;) Sep 14 06:12:17 so... starting a clean build... Sep 14 06:15:45 * CM starts a dirty build :P Sep 14 07:07:21 With the dialer, is there any way yet to dial whilst in a call - i.e., interacting with voice system , needing to send a "1", etc. Sep 14 08:02:11 i love the feedreader Sep 14 08:23:39 good morning Sep 14 08:30:45 goooood morning openmoko! Sep 14 08:31:35 * hhf423 does ipkg update, upgrade, reboot Sep 14 08:36:14 doh, 200MHz is just not enough Sep 14 08:36:39 hhf423: 640k and 14Mhz is enough for anyone, the big man said so Sep 14 08:37:11 hrhrhr Sep 14 08:37:26 this should be a good weekend for openmoko builds Sep 14 08:37:30 yes, but back then we were coding in assembler Sep 14 08:37:38 rwhitby: any particular reason? Sep 14 08:38:07 things should be stable now - no more floating svn versions Sep 14 08:40:20 sound good Sep 14 08:40:36 any news/rumors on gsmd btw? Sep 14 08:40:37 is anyone else getting '+CME ERROR: 32' from gsmd? Sep 14 08:40:39 gah, I shall have to find new ways to rock the boat :_) Sep 14 08:41:07 what were the old rumors? Sep 14 08:41:37 DukeOfURL: 32 could mean you are trying to roam but not allowed Sep 14 08:42:00 can't register with T-Mobile Sep 14 08:42:16 DukeOfURL: I have only seen 13 Sep 14 08:42:25 DukeOfURL: wrong pin? Sep 14 08:42:40 no pin. it's a prepaid card. Sep 14 08:43:08 DukeOfURL: the old rumors on gsmd were that openmoko/fic wanted to put someone on it. but no date, nothing Sep 14 08:43:15 DukeOfURL: did you try at-commands yourself? Sep 14 08:43:23 DukeOfURL: well, here prepaid SIMs also have PINs Sep 14 08:43:37 yes, at commands fail--libgsmd-tool hangs Sep 14 08:44:07 the SIM in a Samsung phone doesn't ask for a pin Sep 14 08:44:18 DukeOfURL: did you do the change to boot parameters? bootargs_base? Sep 14 08:44:25 yes Sep 14 08:44:42 DukeOfURL: what does at+cpin? say? Sep 14 08:44:54 morning Sep 14 08:45:48 DukeOfURL: AT+CMEE=2 enable better error messages Sep 14 08:46:13 my CME ERROR table end on 31 ;( Sep 14 08:47:13 at+cpin? says '+CPIN: READY' Sep 14 08:47:44 all errorcodes here: http://www.activexperts.com/activsms/sms/gsmerrorcodes/ Sep 14 08:48:14 DukeOfURL: maybe you need to specifically register with your operator? Sep 14 08:49:23 I have been able to make calls in the past (with libgsmd-tool) Sep 14 08:50:26 has anyone seen this when starting mediaplayer2? http://scap.linuxtogo.org/files/6ee9a47478e2ee6aaea5b91b88e5a933.png Sep 14 08:51:35 how do I start screen? Sep 14 08:52:19 ahh Sep 14 09:03:11 morning all Sep 14 09:04:48 morning Sep 14 09:06:38 I had a problem with flashing qemu-local, failing at nand write.jffs2 0x31000000 rootfs 0x2794000, but increasing time wait from 120 to 300, just made it work. Anybody have seen it before? Sep 14 09:07:09 in flash.sh Sep 14 09:12:28 Any_Key have a slow computer? Sep 14 09:13:47 p4 2.40GHz Sep 14 09:13:57 hm.. should do Sep 14 09:14:24 1266M of RAM Sep 14 09:14:30 strange Sep 14 09:22:21 roh, until it continues to work, little hacks acceptable Sep 14 09:29:10 is there any info on sim card compatability? Sep 14 09:32:44 I was able to make one call--the call was disconnected after a few seconds. Sep 14 09:33:04 Now I'm not getting any audio, not even from key presses. Sep 14 09:35:54 alsactl -f /etc/alsa/stereoout.state restore Sep 14 09:36:26 XorA: did you played FullHD movies already? (on desktop) Sep 14 09:36:41 hrw: FullHD? Sep 14 09:37:29 1920x1080 Sep 14 09:37:35 hrw: yeah Sep 14 09:37:50 hrw: although my monitor is only 1024x768 Sep 14 09:39:07 if I remeber right I played my first HD movie on an Athlon 1Ghz at approx 2fps Sep 14 09:39:15 mplayer chokes here - output to 1680x1050 Sep 14 09:39:26 hrw: which codec? Sep 14 09:39:30 xine player (kaffeine) lack sound Sep 14 09:39:34 XorA: wmv9 ;( Sep 14 09:39:46 hrw: yeah, that one doesnt work well Sep 14 09:40:02 hrw: h264 plays nice if your gfx card can handle it Sep 14 09:40:14 none of that ATI junk Sep 14 09:40:22 VIDEO: [WMV3] 1920x1080 24bpp 1000.000 fps 8384.0 kbps (1023.4 kbyte/s) Sep 14 09:40:51 ops 3 not 9 Sep 14 09:44:36 VIDEO: [WMV3] 1920x1080 24bpp 23.976 fps 0.0 kbps ( 0.0 kbyte/s) plays here Sep 14 09:44:51 Selected video codec: [ffwmv3] vfm: ffmpeg (FFmpeg M$ WMV3/WMV9) Sep 14 09:47:13 hhf423: I get the same gstreamer error as you (Resource busy or not available) on starting mediaplayer. Sep 14 09:50:00 hrw: nice trailers site for testing :-) http://www.drfoster.f2s.com/trailers_hd-dvd.shtml Sep 14 09:51:26 kayon: Try killing pulseaudio Sep 14 09:54:59 now this is what the OM keypad should look like http://www.blogsmithmedia.com/www.engadget.com/media/2007/09/%5B1065%5Dsvsip.png Sep 14 09:55:29 abraxa_: That fixed the error. Music is apparently playing, but no sound. Playing with alsactl to see if that's the problem, although ogg123 plays the same file just fine. Sep 14 09:55:52 are we missing gstreamer sink for pulseaudio or is mediaplayer just not using it? Sep 14 09:56:24 thats the new VoIP setup hacked onto the Nitendo DS by a user Sep 14 09:56:54 there also seems to be something to make alsa apps go through pulseaudio Sep 14 09:57:16 hi all Sep 14 09:57:31 hi all, from me too Sep 14 09:58:14 i have problems building uboot for my x86 Sep 14 09:58:37 error message is: hello_world.c:0: error: bad value (armv4) for -march= switch Sep 14 09:58:46 and similar... Sep 14 09:59:09 using mokomakefile and the wiki page for xephyr Sep 14 09:59:39 any ideas? Sep 14 10:01:48 kayon: Try restarting pulseaudio then? Sep 14 10:02:34 tuukkah: The media player is currently tied to gstreamer's alsasink in an effort to try reducing CPU usage but the real culprit is something else - it'll get fixed asap :) Sep 14 10:02:45 shahn_ shouldn't that value be armv4t ? Sep 14 10:03:24 honestly, i have no idea. Sep 14 10:03:52 abraxa_, but pulseaudio is reserving the alsa sound device? Sep 14 10:03:56 i just blindly followed the instructions Sep 14 10:04:22 shahn_, also why do you need uboot - a bootloader - on your desktop computer? Sep 14 10:04:32 tuukkah: Yeah, lemme revert the change and commit to make the media player cooperate better while the real fix gets worked on Sep 14 10:04:46 yes, that came to my mind too Sep 14 10:05:05 abraxa_, what would the real fix look like? Sep 14 10:05:06 so how can i disable uboot? Sep 14 10:05:39 openmoko: 03abraxa * r2966 10/trunk/src/target/OM-2007.2/applications/openmoko-mediaplayer2/src/playback.c: PulseAudio isn't the culprit so let's not force alsasink anymore - real fix comes later Sep 14 10:06:39 shahn_, i don't think openmoko-devel-image contains ubuut Sep 14 10:06:43 s/uu/oo/ Sep 14 10:06:45 tuukkah meant: shahn_, i don't think openmoko-devel-image contains uboot Sep 14 10:07:21 then again, i don't know about openmoko on x86 Sep 14 10:07:23 tuukkah: libgstmad outputs the samples 32-bit int while PulseAudio only accepts 16-bit int, so conversion/resampling takes place that bogs down the CPU Sep 14 10:07:57 well, it tries to compile "packages/uboot/uboot-openmoko_1.2.0+gitf3402..." Sep 14 10:08:11 with make openmoko-devel-image Sep 14 10:11:02 abraxa_, so either libgstmad would learn to output with precision more suitable for an embedded device or the extra bits need to be ignored at some point? Sep 14 10:11:11 shahn_ try posting the last output lines on pastebin Sep 14 10:11:31 tuukkah: Exactly, and patching libgstmad seems the easiest way so I'm looking into how to do that Sep 14 10:11:50 ok Sep 14 10:12:30 abraxa_, cool Sep 14 10:14:10 i'm starting to see this dbus stuff could really work in making python programs more feasible on the phone Sep 14 10:15:00 for example, if the media player can be controlled via dbus Sep 14 10:15:26 Ghiottone|work: http://pastebin.com/d80f0fc1 Sep 14 10:15:43 although for video, we'd also need a way to plug the video back to the pygtk window Sep 14 10:16:23 i never learned to use the plug system in gtk Sep 14 10:17:07 "Together with gtk.Socket, gtk.Plug provides the ability to embed widgets from one process into another process in a fashion that is transparent to the user." Sep 14 10:17:16 tuukkah: Considering that gta01 can't even get "good" video performance with pure mplayer I doubt getting gtk/python involved will make it any better ;x Sep 14 10:18:33 abraxa_, it should happen like browser plugins work so that the video player talks directly to X and the gui is there just to call the player to start and stop Sep 14 10:19:01 so i hope it wouldn't make it any worse either Sep 14 10:19:48 Hmm, I'm not sure if that's feasible either cause that would mean we can only switch between "gui+no video" and "fullscreen" Sep 14 10:21:52 abraxa_, no no. you know how plugging makes the video appear as a part of a web page in firefox? the same thing: a window of one program becomes a widget of another Sep 14 10:22:54 shahn_ try revising the local.conf file. I've seen a i486 but maybe there should be i686. Apart from this, I don't have any other advice as I have never tried to build the image for x86, sorry Sep 14 10:23:17 Ah, I misunderstood you then - I thought you meant "the video player" as in "application XYZ that runs in a separate process and gets commanded through pipes or d-bus" Sep 14 10:23:24 i should try to do it on my laptop first with pygtk+mplayer and if it works like i think it works then it'll work on neo too. and then to get a nicer programming interface we would use mediaplayer dbus service instead of mplayer Sep 14 10:23:24 yeah, thanks Sep 14 10:23:50 * XorA would suggest not using mplayer on Neo Sep 14 10:24:10 abraxa_, i thought that's how mplayer plugin for firefox works Sep 14 10:24:17 well, local.conf says BUILD_ARCH = "i686" Sep 14 10:24:25 XorA, just as a simple start Sep 14 10:24:26 thanks anyway... Sep 14 10:24:42 what's the uickest simplest way to rebuild libgsmd-tool ... ? Sep 14 10:25:05 * ScaredyCat seriously hates bitbake build system Sep 14 10:25:08 tuukkah: do it right from the start, if gstreamer is used there is at least a chance we will get wmv support :-) Sep 14 10:25:32 tuukkah: and mplayer is a nice player, but a horrible monolithic lump for embedded systems Sep 14 10:25:36 ScaredyCat: make rebuild-package-libgsmd-tool if you use the makefile Sep 14 10:25:53 scaredycat . /setup-env; cd build; bitbake -i Sep 14 10:26:21 XorA, just a sec, i hope :-) Sep 14 10:26:32 ScaredyCat: this will keep you in bitbake interactive shell. launch rebuild libgsmd-tool every time you need Sep 14 10:26:36 XorA: Unfortunately no one has tested video playback performance using gstreamer yet, so it seems mplayer is the reference... Sep 14 10:26:57 abraxa_: with gst-ffmpeg it would be using the same codecs as mplayer anyway Sep 14 10:27:01 ScaredyCat: this will skip the reparse time after the first rebuild Sep 14 10:27:06 the problem is bitbake doesn't know about it Sep 14 10:27:07 morning guys Sep 14 10:27:12 Hiho Mickey :) Sep 14 10:27:21 hi mickey Sep 14 10:27:29 where are the gllin experts? Sep 14 10:27:37 yo mickeyl Sep 14 10:27:39 XorA: It does? Awesome - I hope the ffmpeg plugin lets us define frame-dropping modes, too Sep 14 10:27:52 No providers of build target libgsmd-tool Sep 14 10:28:30 Elrond: ping Sep 14 10:29:50 ScaredyCat: what about libgsm? Sep 14 10:29:59 SpeedEvil: wakey wakey, talk to mickeyl Sep 14 10:30:21 i need some guidance what we need re. eabi gps Sep 14 10:30:31 i have a bunch of files here, but no idea how they connect :D Sep 14 10:30:59 mickeyl: check that is doesnt SUID root and follow symlinks otherwise gllin is a secuity flaw :-) Sep 14 10:31:04 that doesn't appear to build libgsmd-tool ... Sep 14 10:31:35 libgsmd-tool is a subpackage, not a providee Sep 14 10:31:56 so you cna't build it on its onw Sep 14 10:31:58 own? Sep 14 10:32:01 right Sep 14 10:32:03 bitbake libgsmd Sep 14 10:32:07 will build both Sep 14 10:32:49 * ScaredyCat starts sobbing Sep 14 10:33:05 * XorA hands ScaredyCat a tissue Sep 14 10:35:10 did someone look ad the bootchart I put on the wiki? Sep 14 10:37:21 hmm, k, i'll try to get some gps guys later Sep 14 10:37:25 * ScaredyCat eats the tissue from XorA Sep 14 10:52:16 http://www.theinquirer.net/?article=42293 Sep 14 10:56:06 Ghiottone|work: url? Sep 14 10:56:14 ScaredyCat: I saw that one. :) Looks like something every bitbake user could need. ;) Sep 14 10:56:22 :) Sep 14 10:57:07 rwhitby: quick question re mokomakefile: we svn co all svn.openmoko.org - how does this get used by openembedded/bitbake? Sep 14 10:57:24 i am seeing bitbake check out individual applications from openmoko.org when it needs to ... Sep 14 10:57:29 jpcass: it doesn't Sep 14 10:57:42 some of it is used by the qemu targets Sep 14 10:57:42 rwhitby: ah. so why are we checking it out then? Sep 14 10:58:01 cause that's the way it was in the original build instructions that mokomakefile automated Sep 14 10:58:12 and I haven't got around to improving it yet Sep 14 10:58:29 it's also nice to peek inside the developer directories ;-) Sep 14 10:58:48 rwhitby: ok, good to know - thanks! (agreed ;-) ) Sep 14 10:59:06 rwhitby: bottom of my user page http://wiki.openmoko.org/wiki/User:Alessandro Sep 14 11:01:00 so ldconfig and pulseaudio are the main offenders Sep 14 11:01:07 rwhitby: I have been thinking about the changes to mokomakefile to checkout trunk instead of all the repo. And I think that the best thing should be to clobber openmoko and rerun setup if an incorrect checkout is detected Sep 14 11:01:33 Ghiottone|work: yeah, that's probably the only way - we'll need to clobber build/qemu too Sep 14 11:04:06 rwhitby: ok. what should be the best place to put that check? Sep 14 11:10:09 Ghiottone|work: in check-generation Sep 14 11:19:54 mickey|bbiab i'll meet telmich - the author of cinit - this weekend. do you think a different init-system would make it into OE? Sep 14 11:22:11 rwhitby, have you thought of having in MokoMakefile targets like make build PACKAGE=SMSTool? Sep 14 11:22:47 we have make (build|rebuild|clean)-package-foo Sep 14 11:23:01 kristian-m: 100% sure on that. the OE team would love to support multiple init systems. OM is currently thinking upstart is the best candidate, but I'd welcome more inputl. Sep 14 11:23:30 upstart is the ubuntu init, isnt it? seems like a good one, event-based... Sep 14 11:23:35 rwhitby, oh ok, they are there like that Sep 14 11:24:23 tuukkah: I do it that way so you can use them as dependencies and build more than one with a single make invocation Sep 14 11:24:36 rwhitby, i see. nice Sep 14 11:25:02 mickey|bbiab great i'll keep you informed then Sep 14 11:28:38 jpcass japp. just not very usable documented yet Sep 14 11:29:53 yay... Sep 14 11:29:54 real 197m25.465s Sep 14 11:29:54 user 291m23.985s Sep 14 11:29:54 sys 66m1.068s Sep 14 11:30:11 it built though.. in record time Sep 14 11:37:33 hi Sep 14 11:40:47 * hrw wants car kit holder... Sep 14 11:41:10 * PBeck wants a gta02 ;> Sep 14 11:41:37 * rwhitby wants candy Sep 14 11:41:43 PBeck: I do not want gta02 currently Sep 14 11:42:05 hrw: why? Sep 14 11:42:35 PBeck: lack of time for it and gta02 is not yet finished Sep 14 11:42:59 so what use I will make of gta02v2 when final-send-to-devs will be gta02v5? Sep 14 11:43:02 for example Sep 14 11:43:13 hrw: of course when it is finished :P Sep 14 11:43:23 hrw: i dont want a gta02 with bugs :P Sep 14 11:43:59 * blindcoder has candy Sep 14 11:44:14 hrw: we will see, when gta02 is ready Sep 14 11:44:21 and I agree, a bugged phone is Bad :-) Sep 14 11:46:27 anyone got a nice example video file for playing on the neo? Sep 14 11:47:21 hrw the same i will do with my gta01bv02 when i find the time.. build a waterproof device for playing mp3 when i'm showering Sep 14 11:47:32 tuukkah: http://www.drfoster.f2s.com/trailers_hd-dvd.shtml - grab something Sep 14 11:47:56 ha ha =) Sep 14 11:48:51 the play button in out gtk theme is localized for rtl cultures =) Sep 14 11:51:03 well guys - I finally completed my openmoko-develop-image Sep 14 11:51:19 now i cant find the flash image Sep 14 11:51:32 could you tell me where it is? Sep 14 11:53:16 find OpenMokoDirectory -name *.jffs2 -print Sep 14 11:53:44 build/tmp/deploy/glibc/images/fic-gta01 Sep 14 11:54:04 someone here use 'Full PM' mode? Sep 14 11:54:39 the above was in response to RalphEichelberge Sep 14 11:55:08 thx cindy Sep 14 11:56:20 rwhitby, BUILD_ALL_DEPS doesn't seem to exist anymore: bitbake doesn't know about it and it's not used anywhere Sep 14 12:02:27 it seems that after a sleep, neo can't receive and make any call ... does that happen to someone ? Sep 14 12:03:33 tuukkah: looks like do_build[recrdeptask] = "do_build" in meta.bbclass replaces it Sep 14 12:04:06 RalphEichelberge: what was the missing step which enabled you to complete the build after so many days? Sep 14 12:04:31 I just did a make update again Sep 14 12:04:36 rwhitby, if i wanted to do that when i run "bitbake mplayer", could i somehow include that in the command line? Sep 14 12:04:56 tuukkah: I have no idea, you would need to ask a bitbake expert. Sep 14 12:05:17 thanks, on to #oe then Sep 14 12:05:19 but flashing dont work Sep 14 12:05:28 i now tried t Sep 14 12:05:30 to Sep 14 12:05:33 tuukkah: isn't mplayer already in the feed? Sep 14 12:05:44 just rm * all stuff in the image dir Sep 14 12:06:17 with make openmoko-devel-image Sep 14 12:06:25 rwhitby, i haven't built a feed because when i asked about interesting targets to learn no-one remembered to mention it :-) Sep 14 12:07:35 tuukkah: I mean the official feeds Sep 14 12:07:49 haven't been using them Sep 14 12:08:11 http://buildhost.openmoko.org/OM2007.2/tmp/deploy/glibc/ipk/armv4t/mplayer_0.0+1.0rc1-r14_armv4t.ipk Sep 14 12:08:23 back Sep 14 12:08:38 rwhitby: great. also I think that there is another issue. in setup-openmoko, there is the svn co. if that checkout is interrupted someway you have to clobber openmoko Sep 14 12:08:48 i have mplayer built but then problem is when i try to install it ipkg complains about all the missing dependencies Sep 14 12:09:14 tuukkah: all the deps are in that feed. what do you have your feeds set to on the neo? Sep 14 12:09:20 rwhitby: as you can't svn co again on the same dir. doing svn up or svn switch however will work also after interrupt Sep 14 12:09:45 rwhitby, my own builds. like i said, i haven't been using the official feeds Sep 14 12:09:58 tuukkah: then you need to build our own feed then Sep 14 12:10:32 i suppose that could be the easy solution although i'd like to learn how to do this per package Sep 14 12:10:49 tuukkah: asking me is not going to get you that answer, I'm afraid. Sep 14 12:11:37 no problem, i'll see if anyone in #oe can tell something about it Sep 14 12:11:57 tuukkah: after you build a new package you have to do a make build-package-package-index Sep 14 12:12:11 tuukkah: for ipkg update to get it Sep 14 12:12:14 Ghiottone|work, done that Sep 14 12:13:07 tuukkah: Ok I think I got the point. A few time I built a package that required some other package to be built Sep 14 12:13:26 who wrote MediaPlayer? it is unusable. Sep 14 12:13:35 At the moment. Sep 14 12:13:39 tuukkah: bitbake did them all, then I did package-index, then ipkg update and install Sep 14 12:13:43 I think it's listed on [[MediaPlayer]] Sep 14 12:13:48 rwhitby, yesterday i tried to get some help on #oe wrt the wiki but no-one was available. someone would beed to rename http://www.openembedded.org/wiki/bbfile to not start with a lower-case letter so it could be linked to from the frontpage Sep 14 12:14:07 hrw, it's not meant to be ready yet Sep 14 12:14:10 tuukkah: install complained on missing packages. running make build-package- for each missing package I solved the issue Sep 14 12:14:58 tuukkah: I think this happens because packages are listed in DEPENDS and not in RDEPENDS Sep 14 12:14:58 Ghiottone|work, i know. i'd hate to do that manually, i've done that work enough already Sep 14 12:15:03 tuukkah: one way is to create a new .bb file, inherit meta.bbclass, and make it depend on mplayer - see meta/openmoko-feed.bb for an example Sep 14 12:15:23 Ghiottone|work, oh so there're bugs about these still... Sep 14 12:15:41 tuukkah: then adjust RDEPENDS in that package recipe to your needs Sep 14 12:16:03 Ghiottone|work: no it happens cause that's the way that bitbake is designed to work. You only get packages for an image or a feed if you use that package in an image or a feed. Sep 14 12:16:22 tuukkah: this is not a bug Sep 14 12:18:20 tuukkah: I see. what about a for i in $DEPENDS; do make build-package-$i; done ? Sep 14 12:29:00 rwhitby: or do 'bitbake package1 package2 package3' Sep 14 12:29:19 if you know all the dependencies, yes. Sep 14 12:30:38 is there a way to see what are the dependencies? Sep 14 12:30:53 bitbake -g will tell you that Sep 14 12:30:58 apart from generating all the graph that I found to be unmanageable Sep 14 12:31:04 yeah that Sep 14 12:32:08 oh, don't think of it as graphical output - think of it as something to feed to grep and less Sep 14 12:33:15 could be a good idea to turn MokoMakefile into a full bitbake wrapper for openmoko Sep 14 12:34:08 no, better for people who need to build distros to use bitbake directly, and for application developers to use the still-coming-in-the-future SDK which hopefully won't be based on bitbake commands Sep 14 12:34:57 well, would the SDK still be implemented using bitbake? Sep 14 12:36:04 I hope not Sep 14 12:36:08 imho running debian proper is also an interesting path to try Sep 14 12:36:39 bitbake is not designed to be an IDE. it is an embedded distro build tool. Sep 14 12:37:11 (well, bitbake is general, but when it is used with OE it's an embedded distro build tool) Sep 14 12:41:44 rwhitby: do you think that the processor graph in the bootchart has sense? it is always at 100% Sep 14 12:42:00 Ghiottone|work: no idea Sep 14 12:42:13 Ghiottone|work, so we would be fully cpu bound. why not Sep 14 12:42:42 there are parts with sleeps though, what runs during those times? Sep 14 12:45:45 mickey|bbiab - pong Sep 14 12:51:09 it seems that little cpu is taken by jffs2 Sep 14 12:52:22 well, i did it! Sep 14 12:52:31 i have a self build rootfs Sep 14 12:52:57 can anyone tell me how to make my selfbuild kernel? Sep 14 12:55:55 openmoko: 03thomas * r2967 10/trunk/src/target/OM-2007.2/artwork/icons/openmoko-standard/ (29 files in 2 dirs): Sep 14 12:55:55 openmoko: * Add new stock icons from designer Sep 14 12:55:55 openmoko: * Replace yes/no/ok icons with new images from designer Sep 14 12:57:36 RalphEichelberge: if you are using the MokoMakefile then make openmoko-devel-image should built the rootfs, the kernel and the uboot image Sep 14 12:57:43 om has a designer? Sep 14 12:57:56 be nice Sep 14 12:58:29 i didn't know Sep 14 12:58:32 i wasn't being sarchastic Sep 14 12:58:43 i thoguht there were mokup designs done once Sep 14 12:58:44 raster jap.. or do you believe the icons are done by software developers? Sep 14 12:58:46 and that was it Sep 14 12:59:08 roh: depends. some coders can also draw Sep 14 12:59:08 :) Sep 14 12:59:29 all the ones i've seen were in the original mockups Sep 14 12:59:34 so i thoguht they came from that Sep 14 12:59:42 raster heh.. you seem to know strange people ;) Sep 14 13:00:01 maybe om needs more of them :) Sep 14 13:00:51 so om actualyl does have a designer? Sep 14 13:01:07 ie a guy you can go to saying "draw me up some foo for X" Sep 14 13:01:18 raster ack Sep 14 13:01:22 cool Sep 14 13:01:31 u need one of those Sep 14 13:01:40 it gets easier as sharper the description gets Sep 14 13:01:48 svaes coders making a dogs breakfast of their ui's :) Sep 14 13:02:09 Elrond: nevermind. i already found out. the new gllin eabi binary doesn't work *sigh* Sep 14 13:02:17 getting a SIGSEGV pretty early in startup Sep 14 13:02:25 roh: of course Sep 14 13:02:28 ~lart not having sourceode Sep 14 13:02:28 * apt teaches not having sourceode that M$ Access is a database. No, really, a database. A real live multi-user... well, ok, not multi-user, but a database. Yeah, that sounds right. Sep 14 13:02:30 this sucks big time Sep 14 13:02:32 i got only the rootfs Sep 14 13:02:34 e has its own icon guru these days Sep 14 13:02:41 no uboot no kernel Sep 14 13:02:41 it's not me Sep 14 13:02:50 i'm entirely out of time to do it Sep 14 13:03:44 mickeyl does the guy with the compiler have a device from us? Sep 14 13:03:52 roh: yes Sep 14 13:04:26 mickeyl hm. then he should propably be able to test it with a build from us Sep 14 13:04:27 and i refuse to do more work based on non-working binaries Sep 14 13:04:40 this already eat 8 hours of my precious time Sep 14 13:04:48 *sigh* Sep 14 13:05:11 RalphEichelberge: they should all be in build/tmp/deploy/glibc/images/fic-gta01. Does make flash-qemu-local work? Sep 14 13:05:48 ill ty Sep 14 13:05:52 Ghiottone|work, how did you get the data for the graph? Sep 14 13:06:34 openmoko: 03thomas * r2968 10/trunk/src/target/OM-2007.2/applications/openmoko-today2/ (ChangeLog src/today-main.c): * src/today-main.c: (today_create_home_page): Fix Dates icon (bug 830) Sep 14 13:06:39 bootchartd Sep 14 13:07:42 no Sep 14 13:07:48 only lots of trouble Sep 14 13:07:50 :) Sep 14 13:08:11 U-boot fails to build Sep 14 13:08:37 RalphEichelberge: that's probably why you don't have a uboot image then... :-/ Sep 14 13:08:57 Please wait, programming the NAND flash... Sep 14 13:08:57 sleep: ungültiges Zeitintervall „„setenv““ Sep 14 13:08:57 sleep: ungültiges Zeitintervall „„dontask““ Sep 14 13:08:57 sleep: ungültiges Zeitintervall „„y““ Sep 14 13:08:57 ..... Sep 14 13:09:15 sleep: ungültiges Zeitintervall „„saveenv““ Sep 14 13:09:15 „sleep --help“ gibt weitere Informationen. Sep 14 13:09:15 openmoko/flash.sh: line 88: 13926 Terminated $qemu -mtdblock "$script_dir/$flash_base" -kernel "$script_dir/$1" -serial stdio -nographic -usb -monitor null 0<&0 Sep 14 13:09:24 sorry for foating Sep 14 13:09:55 RalphEichelberge: that output is usually symptomatic of missing one of the "holy trinity" of rootfs, kernel and uboot image. Which makes sense if your kernel and uboot image haven't built Sep 14 13:10:05 should gsm work allready? Sep 14 13:10:23 Not sure why `make openmoko-devel-image` would succeed without building them though Sep 14 13:10:36 aha Sep 14 13:10:43 well how should I :)# Sep 14 13:11:51 RalphEichelberge: if you've had loads of trouble so far it may be worth clobbering everything and starting again from scratch, the build has gotten much more stable over the last while Sep 14 13:12:43 does make update let me start from scratch or do I have to make clean or such? Sep 14 13:13:15 `make clobber` will clean out *everything*, it may take a while to run Sep 14 13:13:41 then you need to `make setup`, `make update`, `make openmoko-devel-image` Sep 14 13:14:17 * raster wonders if om will build this time... Sep 14 13:14:21 well i will let it run over the weekend thus Sep 14 13:15:07 how about the gsm Sep 14 13:15:17 can we make phone calls allready? Sep 14 13:16:24 tuukkah: yes, bootchartd Sep 14 13:16:34 RalphEichelberge, if you can get gsmd to behave, then yes Sep 14 13:16:46 i see Sep 14 13:16:56 RalphEichelberge: i managed to do it after a hack yesterday Sep 14 13:17:10 hack? Sep 14 13:17:19 what did you do? Sep 14 13:17:49 RalphEichelberge: patched /etc/init.d/gsmd Sep 14 13:18:28 so it just didnt start? or with the wrong params? Sep 14 13:18:30 http://bugzilla.openmoko.org/cgi-bin/bugzilla/show_bug.cgi?id=788 Sep 14 13:19:59 is the mailinglist broken again? Sep 14 13:20:53 http://bugzilla.openmoko.org/cgi-bin/bugzilla/buglist.cgi?bug_status=NEW&email1=hrw%40openedhand.com&emailtype1=exact&emailassigned_to1=1&emailreporter1=1 Sep 14 13:20:54 and then i added Sep 14 13:20:54 /etc/init.d/gsmd stop Sep 14 13:20:54 /etc/init.d/gsmd start Sep 14 13:20:54 before dialer in /etc/matchbox/session Sep 14 13:21:15 don't do that in session Sep 14 13:21:15 sunix: dialer do same if cant connect to gsmd Sep 14 13:21:23 yeah Sep 14 13:21:31 sunix: http://bugzilla.openmoko.org/cgi-bin/bugzilla/show_bug.cgi?id=831 Sep 14 13:22:45 oh fark Sep 14 13:22:49 mh Sep 14 13:22:50 there goes pcmcia utils again Sep 14 13:22:54 infinite loop Sep 14 13:22:57 filling up my disk Sep 14 13:23:02 thx sunix Sep 14 13:23:13 * sunix is reading :) Sep 14 13:23:24 Ghiottone|work, so the README says it uses /proc/*/stat and /proc/stat and /proc/diskstats. could be as simple as information missing in them. at least mtdblock* shows 0 Sep 14 13:23:29 Elrond: ping Sep 14 13:23:58 ML dead again? Sep 14 13:24:32 hi alexzhang Sep 14 13:24:49 hi doctor! Sep 14 13:25:03 ML dead again? Sep 14 13:25:28 tuukkah: that was I initially thought. but I checked /proc/diskstats and information seemed to be there Sep 14 13:25:53 alphaone - *pong* Sep 14 13:25:56 tuukkah: some data was in the resulting diskstats log Sep 14 13:25:59 Ghiottone|work, i see other that 0 only for mmcblk0* Sep 14 13:26:40 Elrond: Shoragan and I are at "Gambacher Kreuz" now Sep 14 13:26:57 tuukkah: yeah. I also added a |grep mmcblk to the logger function but results did not change. maybe the graph generation routine does not support that device Sep 14 13:27:02 alphaone - Ahh, so you're there soon. :) Sep 14 13:27:07 mickeyl - gllin-eabi? Huh? Sep 14 13:27:10 hrw: so what should i do ? :p Sep 14 13:27:20 Elrond: SIGSEGV Sep 14 13:27:35 Elrond: yeah Sep 14 13:27:41 mickeyl - Is it available anywhere? Sep 14 13:27:47 Elrond: no Sep 14 13:27:53 i'll make it available once it works Sep 14 13:27:54 Ghiottone|work, mmcblk0 is the sd card and i wouldn't expect it to be used significantly during the boot Sep 14 13:27:57 and the license is cleaned Sep 14 13:28:11 hrw: don't understand a thing to the bug 831 Sep 14 13:29:03 tuukkah: I know. me too. but I am starting to wonder if bootchart ignores that device for some reason Sep 14 13:29:21 i wonder if its om building with make -j3 on pcmciautils Sep 14 13:29:28 and that somehow getting make into an infinite loop Sep 14 13:29:29 hi, mickeyl, what about QVGA theme now? Sep 14 13:29:34 alexzhang: I just subscribed to a bunch of lists and have not received any response Sep 14 13:29:40 sunix: now gsmd is restarted up to 2 times during x11 startup Sep 14 13:29:42 mickeyl - Ahh, good. Sep 14 13:29:44 alexzhang: almost 100% done Sep 14 13:29:51 sunix: so do not add another one Sep 14 13:29:51 raster: http://gstaedtner.net/bilder/neo1973/e17_triple.jpg Sep 14 13:29:59 alexzhang: works already fine on my a780 Sep 14 13:30:00 mickeyl - So finally the output-data from gllin isn't anymore owned by GL? ;o) Sep 14 13:30:20 oh! Cool!!! Thank you so much!! Sep 14 13:30:35 alexzhang: thank you for all your work! Sep 14 13:30:37 32G log.do_compile.67 <- bad .... mmmkay Sep 14 13:30:39 really .. ok :) Sep 14 13:30:42 alexzhang: we need to continue on a1200 :) Sep 14 13:30:51 thomasg: insane, but cool Sep 14 13:30:51 raster: what you compile? Sep 14 13:30:54 cbrake: thanks for your information Sep 14 13:31:11 mickeyl: I also have a video if you want :) Sep 14 13:31:16 pcmciautils-014-r1 Sep 14 13:31:18 mickeyl: OK! I will try my best. it's so busy these days... ;) Sep 14 13:31:34 mickeyl: why to concentrate on a1200 when there is still lot of work to do on the Neo? (just to ask, not to polemize) Sep 14 13:31:45 I can even use it with my fingers, but I'm well trained with touchscreens, so this is definately not was everyone would like :) Sep 14 13:31:47 raster: hmm.. Sep 14 13:32:12 Ghiottone|work: well, that comment was made with my wearing the OpenEZX hat. I'm still commited to do some work on this project in my spare time. that has nothing do to with not concentrating on the Neo Sep 14 13:32:13 indeed Sep 14 13:32:27 Ghiottone: I would say OpenMoko is not only for Neo, but also for other phones Sep 14 13:32:31 thomasg: the video is the important thing ;) Sep 14 13:32:31 :) Sep 14 13:32:43 Ghiottone|work, if you have bootloader that doesn't crash when booting from sd, you could graph a boot from it Sep 14 13:33:18 hopefully someone from motorola notices OpenMoko on their phones Sep 14 13:34:08 thomasg: video! yes, please :) Sep 14 13:34:16 mickeyl: oh I see :) Sep 14 13:35:02 tuukkah: :-O damn. I just only realized that you said SD card, not the flash card! now I see why it is always zero! Sep 14 13:35:22 tuukkah: later when I'm at home I'll try that Sep 14 13:35:37 XorA: I think they (motorola guys) have noted that. Sep 14 13:36:01 alexzhang: I wonder if they will consider replacing their horrible UI Sep 14 13:36:12 alexzhang: and chucking OM some money for consultancy Sep 14 13:36:26 alexzhang: yeah I know that. but thought that mickey was working on the neo and not thinking at his spare time Sep 14 13:36:58 thomasg: how responsive is e17? Sep 14 13:37:13 XorA: I have given an OpenMoko presentation in a Motorola facility. Sep 14 13:37:21 guaqua: look: http://gstaedtner.net/e17_on_Neo.avi Sep 14 13:37:34 sorry, missed it, thanks Sep 14 13:37:47 thomasg: crazy! :) Sep 14 13:37:49 notice, that I used standard software-x11, not xrender and not even the 16-bit software-renderer Sep 14 13:37:57 I've shown the UI to someone at a Motorola office as well. :) Sep 14 13:38:12 raster: maybe crazy, but definitely extremely cool :) Sep 14 13:38:22 thomasg: i need to work on the layout stuff Sep 14 13:38:23 mwester: this was an hour long talk to about 30 Motorola engineers. Sep 14 13:38:24 amazing! Sep 14 13:38:25 it's _FAST_ Sep 14 13:38:31 so e can manage windows nicely on such small screens Sep 14 13:38:35 some of the basics is there Sep 14 13:38:47 mickeyl: yeah, it's really ok - I expect real improvements with software-16-x11 Sep 14 13:39:07 thomasg: *nod* Sep 14 13:39:16 raster: I tested layout and basically it works - but it's not really usable right now Sep 14 13:39:36 one of these days i need to find out what that delay is in drawing borders Sep 14 13:39:39 mickeyl: Thanks for committing the /etc/init.d/gmsd workaround. Would a kernel patch to ensure the flow control is set correctly when one does the "echo '0' >.../power_on" operation be a better solution? Sep 14 13:39:52 thomasg: yeah - layotu is just a start on a proof of concept Sep 14 13:40:03 that it *IS* possible to just have a moduel define window layout policies Sep 14 13:40:05 mwester: yes Sep 14 13:40:19 *droool* Sep 14 13:40:23 and i need to put in all the hooks Sep 14 13:40:23 thomasg: can you upload your image? I'd like to showcase it on OpenExpo Sep 14 13:40:23 :) Ok, I'll work on that then. Sep 14 13:40:32 mwester: cool, thanks Sep 14 13:40:40 mickeyl: you mean the jffs2 image? Sep 14 13:40:44 or the picture? Sep 14 13:40:49 i can get speedusp in normal software_x11 if i cacfhe buffers and xshmimages Sep 14 13:40:49 jffs2 Sep 14 13:40:54 to avoid re-alloc per frame update Sep 14 13:41:07 mickeyl: of course, but it's more a evil hack than a real image :) Sep 14 13:41:19 thomasg: sure. i don't expect otherwise :) Sep 14 13:41:20 but well, it's not that bad Sep 14 13:41:32 if you have any OE patches, just upload to bugzilla ;) Sep 14 13:41:33 I can upload the jffs2 image plus the needed ipkg packages Sep 14 13:41:40 cool Sep 14 13:41:49 you will have to install the ipkgs on the device, otherwise e17 will crash at start Sep 14 13:42:07 didn't have the time to build a real image yet, I'm just glad to have it working now :) Sep 14 13:51:08 openmoko: 03thomas * r2969 10/trunk/src/target/OM-2007.2/applications/openmoko-today2/ (ChangeLog src/today-pim-journal.c): Sep 14 13:51:08 openmoko: * src/today-pim-journal.c: (today_pim_journal_box_new): Don't try to Sep 14 13:51:08 openmoko: load the journal if opening it failed. Prevents segfault if d-bus is Sep 14 13:51:08 openmoko: not available. Sep 14 13:54:01 I'm getting the following with dfu-util: No such Alternate Setting: "kernel" Sep 14 13:54:09 any suggestions? Sep 14 13:55:48 ahh, dfu-util -a 3 -R -D works -- perhaps I need to update my BL Sep 14 13:57:51 cbrake, depends what version of dfu-util/u-boot you have Sep 14 13:58:27 thos: BL version says 1.20-moko9_r0, dfu-util was built from a very recent OE build Sep 14 13:59:31 could the names come from uboot env variables? Sep 14 14:00:14 at least there we have mtdparts=mtdparts=neo1973-nand:0x00040000(u-boot),0x00004000(u-boot_env),0x00200000(kernel),0x000a0000(splash),0x03d1c000(rootfs) Sep 14 14:07:31 is there any working mediaplayer for neo other then mpg321/mpg123/madplay type? Sep 14 14:08:48 hrw - abraxa_ was/is working on one. Sep 14 14:10:20 that would be openmoko-mediaplayer Sep 14 14:10:35 but it's not quite there yet Sep 14 14:11:21 hrw: gst-launch :D Sep 14 14:11:27 mplayer works somewhat Sep 14 14:11:37 mickeyl: s.... ... ;D Sep 14 14:11:45 haha Sep 14 14:12:04 !ombug 866 Sep 14 14:12:04 mickeyl: qry Sep 14 14:12:10 also mpd / mpc works Sep 14 14:12:20 mickeyl: http://bugzilla.openmoko.org/cgi-bin/bugzilla/show_bug.cgi?id=866 is for you Sep 14 14:14:08 ├@°␋␌-±├▒01:·$ DISPLAY=:0 ␍␉┤-┌▒┤┼␌␤ ␊┼└┐-├␍▒≤ Sep 14 14:17:22 * mwester tries to read hrw's last comment there, and decides perhaps he's missing a character set... Sep 14 14:17:43 looks more like an encoding problem to me Sep 14 14:17:55 or just pasting in some binary Sep 14 14:17:56 mwester: I do not understand that too - terminal fuckup on neo gave me this Sep 14 14:18:16 * XorA was ready to blame cat typing Sep 14 14:18:20 it is 'root@fic-gta01:~$ DISPLAY=:0 dbus-launch openmoko-today" Sep 14 14:19:34 I managed to flash a BL from my current OE build, but it says 1.3.0-rc1-moko11_r0 on the boot screen but, u-boot-gta01bv2-1.2.0+gitf34024... in the images directory -- strange Sep 14 14:20:01 and "dfu-util -a kernel" still does not work Sep 14 14:21:01 cbrake gta01bv2 ? Sep 14 14:21:11 dfu-util -l still lists all the names as "UNDEFINED" Sep 14 14:21:18 you have a really old device, not a p1? Sep 14 14:21:38 roh: no gta01bv4 Sep 14 14:22:05 cbrake, did you check what printenv says in uboot? Sep 14 14:22:10 roh: I did flash the bv4 image, I just copied the wrong text when pasting into ric Sep 14 14:22:16 *irc Sep 14 14:22:20 ;) Sep 14 14:22:48 tuukkah: let me get connected so I can do that ... Sep 14 14:26:29 tuukkah: hmm, my u-boot boot menu only has two items: Boot & Factory Reset Sep 14 14:26:55 cbrake oh.. then your uboot env is propably empty.. thats the hardcoded minimum-default Sep 14 14:27:04 cbrake, you have to go to the uboot console via serial port or usbtty Sep 14 14:27:11 use ttyacm if you can to set the mtd up again Sep 14 14:27:18 cbrake, cu -l /dev/ttyACM0 Sep 14 14:27:44 (on the computer) Sep 14 14:28:42 openmoko: 03thomas * r2970 10/trunk/src/target/OM-2007.2/applications/openmoko-feedreader2/ (ChangeLog data/openmoko-feedreader2.desktop): * data/openmoko-feedreader2.desktop: Enable single instance Sep 14 14:30:04 tuukkah: cu -l /dev/ttyACM0 returns Connected, but nothing else -- no prompt, etc Sep 14 14:30:23 wonder if the console is defaulted to serial Sep 14 14:30:52 cbrake: no debug board? Sep 14 14:30:59 my great python interface to video players =) http://tuukka.iki.fi/tmp/plugtest.py.png Sep 14 14:31:03 XorA: yes, I can get that out if needed Sep 14 14:31:16 cbrake: run the devirginator to get stuff back sane then upgrade Sep 14 14:31:34 cbrake, it doesn't help if you press enter when cu connects? Sep 14 14:31:48 tuukkah: let me try again Sep 14 14:32:19 but perhaps you really need the menu entry for "use usb console" so that you can use that to active usbtty Sep 14 14:32:20 tuukkah: nothing Sep 14 14:32:28 tuukkah: seems likely at this point Sep 14 14:32:35 * cbrake reaches for the debug board ... Sep 14 14:32:38 Elrond: We're here now at Schloßgartenstraße 9 and trying to find the mrmcd Sep 14 14:32:46 So see you there in a bit. Sep 14 14:33:05 openmoko: 03thomas * r2971 10/trunk/src/target/OM-2007.2/applications/openmoko-terminal2/ (ChangeLog data/openmoko-terminal.desktop): * data/openmoko-terminal.desktop: Enable single instance Sep 14 15:16:04 hi Sep 14 15:22:50 openmoko: 03mickey * r2972 10/trunk/src/target/OM-2007.2/applications/openmoko-terminal2/ (ChangeLog src/moko-terminal.c): openmoko-terminal2: set default font to "LiberationMono 5" Sep 14 15:27:13 hi Elrond, we have now arrived at mrmcd Sep 14 15:27:55 Shoragan - Ahh, fine. :-) Sep 14 15:28:20 we're waiting for you Sep 14 15:28:35 Shoragan - See query. ;) Sep 14 15:55:02 Ghiottone|work, bootchart running alone takes 90% of cpu... Sep 14 15:57:29 but running that way, the resulting cpu graph has some non-100% areas :-) Sep 14 15:58:12 tuukkah: how did you get that value? Sep 14 15:58:35 tuukkah: this makes it almost impossible to benchmark boot process with it Sep 14 15:59:07 tuukkah: so the graph results have sense... Sep 14 15:59:16 ran bootchartd start sleep 60 and looked in top at the same time Sep 14 15:59:39 i'll try with a longer sample interval Sep 14 16:01:27 tuukkah: any tip about I/O ? Sep 14 16:02:08 i would try with sd card if i wasn't afraid to brick the phone in a uboot downgrade Sep 14 16:03:03 downgrade? why is that needed? Sep 14 16:03:36 the 2007.8 snapshot uboot doesn't boot from sd, for me at least Sep 14 16:03:49 what shipped with the phone did boot Sep 14 16:04:31 argh. I have that version (19 august isn't that?) too. and I'm without debug board Sep 14 16:06:03 Ghiottone|work, indeed Sep 14 16:09:41 zdanek: poke :-) Sep 14 16:12:20 Ghiottone|work, changing the interval to 1 second instead of 0.1 second, the idle cpu graph is like 25%. but the boot graph is still 100% all the time :-( Sep 14 16:13:58 anyone having problems building uicmoc4-native, anyone please confirm in the latest svn head.. Sep 14 16:14:18 Sup3rkiddo: built just fine for me a few hours ago. Sep 14 16:15:10 mwester, hm...why am i the only one to get screwed up, have been bulidng without luck for the last 2 weeks Sep 14 16:15:51 You may need to delete the build/tmp directory and start over -- I do that about one per week just to avoid problems... Sep 14 16:16:12 mwester, i did that today morning, =| Sep 14 16:16:31 Can you pastebin the log? Sep 14 16:16:41 wanke5r Sep 14 16:16:44 jus a sec Sep 14 16:16:48 If its a -native package, it might e something on your host... Sep 14 16:16:59 yeoooooooooooooooooooooooooooooo Sep 14 16:17:06 use all are gay Sep 14 16:17:16 mwester, yes it is a native package...am on gentoo Sep 14 16:17:25 StreetSurfer3655, buy a phone Sep 14 16:17:39 04wat Sep 14 16:18:00 04 piss off Sep 14 16:18:02 mwester, known issues eh? Sep 14 16:18:28 04do you wana fight fuck face Sep 14 16:18:59 Not, not specifically - just in general I've found that the native packages have more dependencies on missing/broken host libraries and things like that. Sep 14 16:19:39 04wankers Sep 14 16:19:45 is there a mokomakefile target for just uboot Sep 14 16:19:46 ? Sep 14 16:19:51 Ghiottone|work, adding a sleep 60 to the middle of the boot process made a nice gap so we just happen to be cpu bound Sep 14 16:19:57 04pisss off Sep 14 16:19:59 can someone plz kick that sucker/bot? Sep 14 16:20:10 i'm already on it Sep 14 16:20:13 ignore dude Sep 14 16:20:16 ignore Sep 14 16:20:26 DPThough1, sorry for feeding the troll :-( Sep 14 16:21:37 mwester, http://pastebin.ca/697684 have snipped the messages since it was tooo big... Sep 14 16:21:59 *udpmegaphonekill* Sep 14 16:22:16 Ghiottone|work, which is easy to imagine when simple tasks like bootchart eat almost all cpu Sep 14 16:22:42 cw666: should be something like make rebuild-package-uboot Sep 14 16:31:53 ah, one less troll. ;) Sep 14 17:01:43 raster: any idea why gtk/gnome UI stuff tends to have quite high latency on the neo? Sep 14 17:03:23 cw666: gtk isn't really designed with slow processor & low mem in mind... Sep 14 17:04:48 emdete: well, memory isn't that big of a deal in some ways, we have plenty, cache might be Sep 14 17:04:49 is there some kind of timeline online somewhere? Sep 14 17:04:57 i'd like to know some dates Sep 14 17:05:01 i tried looking on the wiki but no luck :( Sep 14 17:05:09 emdete, does the mobile gnome initiative help change this? Sep 14 17:05:46 x itself seems pretty reasonable, so i'm wondering where the delays come from, inside gtk, inside gnome or some pipeline elsewhere? Sep 14 17:06:25 is the theme pixmap? Sep 14 17:06:28 or does it use cairo? Sep 14 17:07:07 if it's cairo, Xrender might be being slow Sep 14 17:07:23 alex-weej, here: http://wiki.openmoko.org/wiki/Neo1973#.22Phase_2.22_.28GTA02.2C_.22Mass_Market.22.29 Sep 14 17:07:34 thanks Sep 14 17:07:43 christmas! Sep 14 17:08:04 yeah so i'd guess the graphics accelerator will have good EXA Sep 14 17:08:11 speeding up XRender, Cairo, and GTK by proxy :) Sep 14 17:08:14 alex-weej: The original software releases had no cairo at all and a pixmap-based theme. Sep 14 17:08:30 yeah, it might be theat, except it *feels* like there are delays *before* anything is rendered Sep 14 17:08:34 alex-weej: Newer releases include cairo, but I don't know if the theme has changed yet. Sep 14 17:08:39 ah ok Sep 14 17:08:47 cw666: does sysprof run on it? :) Sep 14 17:08:52 hm, i don't know, if the mobile gnome initiative will help - for me it seems gnome always waited for faster processors... Sep 14 17:08:57 so unless it's doing off-screen rendering and blitting it on, i dont think the perf issues are rendering Sep 14 17:09:00 alex-weej: And that Christmas timeline appears to have just been a guess made by a wiki editor. Sep 14 17:09:38 "December is a random date I picked that may be corrected by someone with a clue. --Speedevil" Sep 14 17:10:06 It's an educated guess. Sep 14 17:10:14 alex-weej: no sysprof im sure, and no oprofile (yet) Sep 14 17:10:20 Based on comments made in channel, and previous information. Sep 14 17:10:24 SpeedEvil, i was just about to use the same words :-) Sep 14 17:10:36 cw666: i can postulate many a theory on the slowness. Sep 14 17:11:36 cw666: i think the current gtk on the neo is using the pixmap engine Sep 14 17:11:47 these days that uses gdpixbuf to do all rendering basically Sep 14 17:11:52 beyond that i dont know details Sep 14 17:12:07 but that means u can take cairo and xrender out of the picture Sep 14 17:12:16 (i think xrender at any rate) Sep 14 17:12:31 so you slowness will be: Sep 14 17:12:37 1. gdkpixbuf Sep 14 17:12:53 2. gtk itself (and its rendering pipeline, double buffering etc.) Sep 14 17:13:08 in comparison evas uses the cpu to do rendering too Sep 14 17:13:11 like gdkpixbuf Sep 14 17:13:19 but it can manage it significantly more snappily Sep 14 17:13:27 a few things might factor in there Sep 14 17:13:38 gdkpixbuf i THINK uses RGB buffers for solid images Sep 14 17:13:46 and ARGB32 for those with alpha Sep 14 17:13:52 RGB are 33% smaller Sep 14 17:13:58 but are nto dword (32bit) aligned Sep 14 17:14:05 u have to access R, G and B separately Sep 14 17:14:09 1 byte at a time Sep 14 17:14:16 arm's will not do spectacularly at that Sep 14 17:14:24 they are much better 32bits at a time Sep 14 17:14:32 well, any word access unaligned will trap in the kernel too, that's very slow Sep 14 17:14:35 evas keeps everyting 32bit aligned Sep 14 17:14:36 i guess that's worth checking Sep 14 17:14:38 alpha or not Sep 14 17:14:39 but again Sep 14 17:14:45 i'm digging into deep memory of gdkpixbuf Sep 14 17:14:51 from many a year ago Sep 14 17:14:54 i could be wrong now Sep 14 17:14:58 and no Sep 14 17:15:04 not unaligned Sep 14 17:15:26 just dword (aligned) vs 3 byte (aligned) accesses Sep 14 17:15:49 probably the scaling algorithms are faster in evas Sep 14 17:15:51 actually it probably is aligned, i did see printks for some unaligned accesses but not from what looked like gui related things Sep 14 17:15:52 by a fair margin Sep 14 17:15:53 etc. Sep 14 17:16:00 alpha blending too Sep 14 17:16:05 as w3ell as blitting Sep 14 17:16:08 what about using a fixed pallete? dropping 24-bit? Sep 14 17:16:19 u still have 16bpp Sep 14 17:16:26 so u'd have to drop native rendering to all 16bpp Sep 14 17:16:29 that complicates life Sep 14 17:16:33 trust me Sep 14 17:16:37 evas has a 16bpp engine now Sep 14 17:16:42 it has nicely complicated things Sep 14 17:17:02 try doing filtered (super or sub-sampled/interploated) scaling in 16bpp :) Sep 14 17:17:09 (and keep it fast) Sep 14 17:17:14 no, i can see that would suck Sep 14 17:17:26 fixed palette then would internally end up doing that? Sep 14 17:17:30 but note Sep 14 17:17:33 i am speculating here Sep 14 17:17:34 the hw has a 256-color mode option Sep 14 17:17:42 i think it is porbably a combination of multiple things Sep 14 17:17:46 that add up together Sep 14 17:17:52 and yes it does Sep 14 17:17:58 your lifw ill SUCK in 256color mode Sep 14 17:18:02 you do NOT want to go there Sep 14 17:18:27 the suck comes from where? internally matching/futzing with a limited pallete? Sep 14 17:18:39 a SHAREd display between multiepl apps Sep 14 17:18:43 all trying to steal colors Sep 14 17:18:45 (i dont really know anything about x tbqh) Sep 14 17:18:47 having to flip/change colors Sep 14 17:18:53 looow quality display Sep 14 17:18:57 this isnt x Sep 14 17:19:05 this is just 8bpp vs truecolor 15bpp+ displays Sep 14 17:19:16 the only sane way to deal with 8bpp is colorcubes Sep 14 17:19:29 eg 332 (or 6x6x6 "websafe" colors) Sep 14 17:19:45 i was thinking the web-palette actually Sep 14 17:19:56 http://www.rasterman.com/files/e17-in-8bpp.png Sep 14 17:20:00 http://www.rasterman.com/files/e17-in-15bpp.png Sep 14 17:20:03 http://www.rasterman.com/files/e17-in-16bpp.png Sep 14 17:20:06 as a comparison Sep 14 17:20:21 tyhe 8bpp one is websafe Sep 14 17:20:24 raster: right, but see, i want a phone that's hackable with limited GUI if that's the cost Sep 14 17:20:45 you are possibly better running 320x240 qvga Sep 14 17:20:52 you have 1/4 the # of pixels to fill Sep 14 17:20:59 technically - a 4x speedup Sep 14 17:21:16 i wasn't aware the LCD controller did that Sep 14 17:21:21 (not really as u have otheroverheads - but more bang for your buck than anything else you can do simply) Sep 14 17:21:28 i believe it can Sep 14 17:21:40 personally i will just use EFL Sep 14 17:21:50 as it performas just nicely Sep 14 17:21:52 EFL? Sep 14 17:22:02 the same libraries enlightenment is built on Sep 14 17:22:10 cw666: There's no solid evidence that there's any fundamental performance limitation needing a huge drop in output quality like 8bpp would mean. Sep 14 17:22:13 ecore, evas, edje, ewl/etk etc. Sep 14 17:22:16 cworth: true Sep 14 17:22:24 raster: is that what 'feh' uses too? Sep 14 17:22:29 no Sep 14 17:22:32 feh uses imlib2 Sep 14 17:22:36 thats much older Sep 14 17:22:41 basically obsolete Sep 14 17:23:47 raster: so ELF would use used instead of cairo, gdkpixbuf, etc? Sep 14 17:23:53 oop, EFL :-) Sep 14 17:24:00 evas took over much of imlib2 and is intended for realtime display Sep 14 17:24:06 not just image manipulation routines Sep 14 17:24:20 well not for gtk Sep 14 17:24:20 no Sep 14 17:24:25 i wouldnt use gtk Sep 14 17:24:32 etk or ewl are widget sets Sep 14 17:24:40 they use evas/edje/ecore etc. Sep 14 17:25:03 but you can use gtk Sep 14 17:25:06 it is a good widget set Sep 14 17:25:15 it is very complete and full featured Sep 14 17:25:38 i personally care a lot about speed, slickness and visual appearance and flexibility Sep 14 17:25:41 thus i go the EFL way Sep 14 17:25:42 the thing with gtk is that people have already done a lot of GUI things for it, there is a lot of code to borrow & reuse Sep 14 17:25:47 also i wrote a chunk of it Sep 14 17:25:52 so its my own backyard Sep 14 17:25:55 thus i'm biased Sep 14 17:25:56 :) Sep 14 17:26:37 you take your pick Sep 14 17:26:47 i know what matters to me Sep 14 17:26:50 * polz is just reading the openembedded docs Sep 14 17:26:59 and again, I feel like crying Sep 14 17:27:03 and i am making mu choice accordingly Sep 14 17:27:07 my Sep 14 17:27:28 I just want to get a hello world package working... and then I get to this:http://www.openembedded.org/user-manual&dpage=commonuse_new_package Sep 14 17:39:56 polz: the best way to create a custom application is: 1) set up your app just like you would any other app and check it into SVN. Sep 14 17:40:12 2) use autotools to build your app -- this makes it very easy to croscompile with OE. Sep 14 17:40:38 3) write a OE receipe that will build your app directly from SVN. Sep 14 17:41:05 this method allows you to easily build it on your host system, and in OE for your embedded target Sep 14 17:43:17 polz: here is an example: http://pastebin.ca/697771 Sep 14 17:44:59 cbrake: so far, I haven't even been able to get bitbake to find my .bb files :( Sep 14 17:45:26 polz: what does echo $BBPATH tell you? Sep 14 17:45:30 cbrake: I've followed the app. devel. crash course in the wiki, but for some reason, it doesn't work for me Sep 14 17:45:42 after I source setup-env ? Sep 14 17:46:30 export BBPATH="${OMDIR}/local:${OMDIR}/build:${OMDIR}/openembedded" Sep 14 17:46:55 however, up until a few minutes ago, I had my local dir at the middle of the path Sep 14 17:47:00 at -> in Sep 14 17:47:23 between the build and openembedded dirs Sep 14 17:47:36 polz: location should not matter for a bb file Sep 14 17:47:55 polz: it does for conf files as they define global variables, and the last one wins Sep 14 17:48:06 cbrake: I think the order in which the various setup.conf files are read might matter Sep 14 17:48:22 polz: very possible Sep 14 17:48:51 so... what do I have to do to get my conf file to win over the others ? Sep 14 17:49:16 polz: my projects are not set up exactly like openmoko, so give me a min ... Sep 14 17:49:25 btw, I'm doing a make setup update openmoko-build-image with the new path... Sep 14 17:50:12 and grepping through all the files for BBFILES Sep 14 17:50:24 so... trying anything is going to take a while Sep 14 17:51:41 polz: bitbake -i is also handy sometimes Sep 14 17:52:17 polz: echo $BBPATH returns /build/openmoko/oe/build:/build/openmoko/oe/openembedded in my setup Sep 14 17:52:21 is there a bitbake user-manual in .pdf? Sep 14 17:52:31 mbuf: yes ... Sep 14 17:53:12 cbrake: and I'm guessing all your apps are in openmoko/oe/openembedded, right ? Sep 14 17:53:23 mbuf: http://www.angstrom-distribution.org/openembedded-user-manual-available-pdf Sep 14 17:53:39 polz: I've not built any custom apps for openmoko yet ... Sep 14 17:53:55 polz: but I'm wondering if ${OMDIR} is getting expanded correctly Sep 14 17:53:57 cbrake: you seem to be avoiding my problem by not placing your custom files in a separate dir... Sep 14 17:54:07 heh, I seem to have found the problem Sep 14 17:54:19 or at least _a_ problem Sep 14 17:54:21 polz: I've used bbcollections on many other projects, and it works great Sep 14 17:54:32 the instructions in the wiki are a bit wrong Sep 14 17:55:43 polz: do you have OMDIR set in setup-env? Sep 14 17:56:04 yep. Sep 14 17:56:10 cbrake, excellent thanks! Sep 14 17:56:25 cbrake: the make just finished. It was the order in which the dirs are specified in BBPATH Sep 14 17:56:39 the local dir should come before build Sep 14 17:57:48 I've changed the instructions in the wiki accordingly... Sep 14 17:58:03 polz: yes, I always put my custom overlay directory first. Sep 14 17:58:04 and just hope that doesn't break anything for the others Sep 14 17:58:09 * mbuf wished he owned a brewery so he could just hand over a pitcher everytime someone helped him with OpenMoko Sep 14 17:58:18 polz: bitbake must stop at the first conf file it finds Sep 14 17:59:04 mbuf: buy a couple of crates at a discount store - if you can afford to spend 250+ euros on a phone, you can also afford to buy a crate of beer :) Sep 14 17:59:44 cbrake: oh ? Are you sure ? If you are, I'll add that piece of info to the wiki page Sep 14 18:00:13 polz: no, just an emperical observation. Its how my builds are set up, and its what is working for you. Sep 14 18:00:23 polz: so i'd say it is a pretty safe bet Sep 14 18:00:32 * cbrake groks the bb code quickly ... Sep 14 18:01:29 polz: you have two site.conf files, so as long as the one still existed in build, it was winning Sep 14 18:01:37 polz, will remember that, thanks Sep 14 18:02:10 polz: the image recipe is wrong too ... Sep 14 18:02:29 cbrake: yep. Too bad noone told me this last week :) Sep 14 18:02:50 cbrake: oh ? What's wrong there ? Sep 14 18:03:13 this is the first time I've been able to get bitbake to at least find the recipees Sep 14 18:03:44 so I don't know what other problems one can run into when actually trying to _use_ the recipees. Sep 14 18:04:27 polz: PACKAGE_INSTALL has been replaced by IMAGE_INSTALL Sep 14 18:04:32 * cbrake fixes ... Sep 14 18:10:50 polz: wiki updated Sep 14 18:11:01 When booting, what does Bad Data CRC mean? Reflashing doesn't help. Sep 14 18:11:26 DukeOfURL: if it is a jffs2 message, it is probably benign -- does it still boot? Sep 14 18:11:32 cbrake: we, the incompetent of the world, salute you :) Sep 14 18:11:54 polz: OE is constant change :-) Sep 14 18:12:06 polz: lots of new good stuff, but hard to keep up Sep 14 18:12:52 polz: also, mucking with DISTRO_EXTRA_RDEPENDS is frowned upon by OE core devs -- consider: http://bec-systems.com/web/content/view/79/9/ Sep 14 18:14:49 cbrake: the message occurs at the beginning of the boot, after it has printed "Image Name", "Vreated", etc. It then goes back to the uboot menu. Sep 14 18:15:05 s/Vreated/Created/ Sep 14 18:15:05 DukeOfURL meant: cbrake: the message occurs at the beginning of the boot, after it has printed "Image Name", "Created", etc. It then goes back to the uboot menu. Sep 14 18:15:26 DukeOfURL: ahh, that is likely not a jffs2 message then Sep 14 18:15:50 no, it hasn't unpacked the kernel yet Sep 14 18:16:14 cbrake: is there an official wiki one should refer to ? I mean, if the system is in constant flux, at least the docs should also be up-to-date Sep 14 18:16:53 polz: no, unfortunately not -- you best bet ATM is ask on IRC when things don't work, and look at other recipies Sep 14 18:17:27 polz: the OE wiki is in a sorry state Sep 14 18:18:35 cbrake: isn't there a bitbake wiki the OE wiki could link to ? Sep 14 18:19:02 polz: not that I know of. Bitbake is only used by the OE project as far as I know. Sep 14 18:19:42 polz: the openembedded manual is pretty decent, so I'll try to get some of this stuff in there when I get a chance. Sep 14 18:24:24 cbrake: the kernel image was corrupt Sep 14 18:25:01 it reboots after the fourth reflash Sep 14 18:26:39 DukeOfURL: interesting Sep 14 19:03:20 * cbrake is pretty impressed with the openmoko factory install script -- little fussy, but works remarkably well Sep 14 19:03:33 and glad to have my unit working again with latest sw Sep 14 19:05:29 hello all Sep 14 19:05:54 hi Ghiottone! i finally got a chart with an io graph Sep 14 19:11:32 so look Sep 14 19:11:42 will evo syncing be working on the openmoko at release time? Sep 14 19:11:53 the neo1973 I mean Sep 14 19:14:02 tuukkah, great! what was the problem? Sep 14 19:15:21 i remembered i can use a kernel from the nand and rootfs from the sd card. then in the output i needed to replace mmcblk0 with sda Sep 14 19:16:10 wow. can I see the graph? Sep 14 19:17:27 sure: http://tuukka.iki.fi/tmp/bootchart-sd.png Sep 14 19:18:48 tuukkah, awesome express! Sep 14 19:20:00 tuukkah, have you put that in the wiki ? Sep 14 19:20:19 nope, i don't know if it's good enough yet Sep 14 19:20:26 but we could start a page for bootcharts Sep 14 19:20:37 one for each release Sep 14 19:23:16 does anybody has problems with the mailing lists ? Sep 14 19:23:19 i think it's telling how in this sd chart mount takes the most cpu, and in the nand chart ldconfig takes the most cpu. here it takes the most io Sep 14 19:24:02 so is nand really so cpu-intensive to use or could there be a bug in the driver? Sep 14 19:24:04 does the mailinglist still work? Sep 14 19:24:59 * daltona commented out ldconfig call cause he thinks that it's only usefull when new libs are added Sep 14 19:25:08 * daltona may be wrong though Sep 14 19:25:14 tuukkah, that was for sd, do you have a similar one for the internal memory card ? Sep 14 19:25:26 * daltona gains 1min of boot time Sep 14 19:25:26 * SpeedEvil agrees with daltona Sep 14 19:25:49 xkr47, we don't get io stats for the nand Sep 14 19:25:58 If any new files in dirs mentioned in /etc/ld.so.conf - then run it Sep 14 19:26:00 actually Sep 14 19:26:08 * daltona never been as motivated as since it has the moko in his hands... Sep 14 19:26:11 directory dates should work fine Sep 14 19:27:09 SpeedEvil, it doesn't make sense to run in boot time since you don't install libraries while the system is off :-) Sep 14 19:27:10 SpeedEvil, don't you think it can be implemented by checking the cache directory date vs the libs dirs dates ? Sep 14 19:27:33 ipkg takes care of ldconfig when new libraries are installed Sep 14 19:27:37 with dpkg install ldconfig is run while installing, not at boot I think Sep 14 19:27:53 so ldconfig should not be called by init then... Sep 14 19:28:36 daltona, debian etch doesn't seem to call anymore Sep 14 19:28:41 Something like if find `cat /etc/ld.so.conf` -maxdepth 0 -cnewer /etc/ld.so.cache ; then ldconfig; fi Sep 14 19:28:56 if find on the neo supported cnewer Sep 14 19:29:25 Actually - ignore that Sep 14 19:29:37 finds return value means nothing Sep 14 19:30:56 SpeedEvil, you can do it with -exec Sep 14 19:30:59 i think it's the same thing as with depmod. it used to run during each boot but now only runs when a new kernel or modules are installed Sep 14 19:31:06 if it finds something Sep 14 19:31:12 True Sep 14 19:31:26 we don't want ldconfig run several times! Sep 14 19:31:27 The right way is to run it at lib install time, but that's fragile Sep 14 19:31:40 * santaclaus WinAmp (Stopped) 5 *** Winamp 5.35 00:00/00:00 Sep 14 19:31:43 * santaclaus WinAmp (Stopped) 5 *** Winamp 5.35 00:00/00:00 Sep 14 19:31:47 * daltona going to cook steaks Sep 14 19:32:03 how debian can be so radical as to do the fragile thing? Sep 14 19:32:12 daltona, niice Sep 14 19:32:39 -exec touch updateld Sep 14 19:32:49 works of course - then run ldconfig if the file exists Sep 14 19:32:56 but find doesn't support cnewer Sep 14 19:33:19 you can do a while loop in shell with test Sep 14 19:33:20 tuukkah: I saw your photo of plugtest.py - what exactly does your proof-of-concept... prove? :) Sep 14 19:33:23 true Sep 14 19:34:01 abraxa_, that i can have an external c app do all the video rendering directly to my python app window Sep 14 19:34:42 tuukkah: Ohh, nice. Mind if I check it out? Sep 14 19:34:49 abraxa_, the script is http://tuukka.iki.fi/tmp/plugtest.py Sep 14 19:35:37 tuukkah: Thanks :) Sep 14 19:35:54 I think that busybox can be configured to support it... Sep 14 19:35:59 i read somewhere this is *not* the way browser plugins work but how they *should* work. and actually they do work like that in konqueror Sep 14 19:36:21 I\ll have a look Sep 14 19:36:24 tuukkah: Hehe, never saw GTK in Python before... nice Sep 14 19:37:37 abraxa_, you can mostly look at it like GTK in C, it's just much less typing (in both senses of the word ;-) Sep 14 19:38:20 tuukkah: Yeah... looks nice. And I assume the -wid parameter tells mplayer to use the window ID of the pyGTK window for video display? Sep 14 19:38:48 abraxa_, yes, window id of the gtk.Socket widget, actually Sep 14 19:39:00 tuukkah, :) Sep 14 19:39:32 tuukkah: Yeah. So I can have it display in pretty much any widget... very neat. Thank you for this hint, muchly appreciated =) Sep 14 19:39:54 back finally Sep 14 19:39:57 xkr47, ok here's the nand boot chart for comparison: http://tuukka.iki.fi/tmp/bootchart.png Sep 14 19:40:03 tuukkah nice graph indeed Sep 14 19:40:40 some 5 secs omre Sep 14 19:40:50 the nand one is similar to the one I made Sep 14 19:41:12 tuukkah: wow, how do you do that> Sep 14 19:41:20 what FS type have you got on the sd? Sep 14 19:41:25 emdete, with bootchart Sep 14 19:41:37 * daltona is eating his steak Sep 14 19:41:38 abraxa_, what i hope is that openmoko-mediaplayer could be used as a dbus service: you create a Socket, send its id to the service, the service creates a Plug and connects it Sep 14 19:41:40 what does bootchart need to run? Sep 14 19:42:06 emdete, bash and another package. all are available in open embedded Sep 14 19:42:39 emdete, in the top of the charts you can see the kernel command line has additional init=/sbin/bootchartd Sep 14 19:43:26 (i am asking because i would like to see that for the greenphone - have some trouble with that bootup ;) ) Sep 14 19:43:34 tuukkah: What kind of app would connect to that plug? Sep 14 19:43:38 Ghiottone, the fs is ext3 Sep 14 19:44:08 SpeedEvil, In the neo config, it seems that -cnewer does not exists but -newer does. It compares mtime Sep 14 19:44:14 abraxa_, a mediaplayer gui :-) or a browser plugin Sep 14 19:44:31 daltona: ah Sep 14 19:44:35 daltona: that'd do Sep 14 19:46:28 but in the neo, there are no /etc/ld.so.cache wonder where it is... Sep 14 19:46:36 abraxa_, i wonder if it would also save memory when one dbus service process can take care of video playing/html rendering/other complex stuff of each running application Sep 14 19:47:01 tuukkah: Ahh, I see. The media player incorporate gstreamer's A/V capabilities and show them in its own window - letting apps specify a different window through d-bus seems a little complex - writing a simpler "video host" for other apps to incorporate might be better Sep 14 19:47:19 tuukkah, if you start ldconfig twice in a row, the second time it is much faster Sep 14 19:47:20 will incorporate* Sep 14 19:47:47 hmm.... =) /var/run/ld.so.cache Sep 14 19:47:57 yes. Sep 14 19:48:05 so it is dropped at each boot... Sep 14 19:48:09 so we actually need to run it at each boot if it's there Sep 14 19:48:16 that's why ldconfig is called Sep 14 19:48:20 we could prevent its dropping? Sep 14 19:48:22 i wonder if that's what we want... Sep 14 19:48:35 let see whats the siwe of that file... Sep 14 19:48:50 not big... Sep 14 19:48:53 13k Sep 14 19:49:38 but as it is modified a lot, it may explain why it's not in the flash. it would cause unecessary writes Sep 14 19:49:48 but I'm still not convinced. Sep 14 19:49:58 not a lot, only these 13k when new libraries are installed Sep 14 19:50:33 btw in addition to /etc/ld.so.conf we need to search /lib and /usr/lib of course Sep 14 19:50:49 yes I noticed that Sep 14 19:50:58 Trivial way is to put those dirs in that file Sep 14 19:51:12 ld.so.cache is in /var/run which is mounted out of volatile space each boot Sep 14 19:51:31 That makes sense if you want to run ldconfig each boot Sep 14 19:51:37 I think it makes sense to move it out. Sep 14 19:52:05 where is its location specified? Sep 14 19:52:17 in the ldconfig binary Sep 14 19:52:27 just did a strings /sbin/ldconfig... Sep 14 19:52:27 is it -C? Sep 14 19:52:29 abraxa_, i think you're right, openmoko-mediaplayer is not like mplayer but just a gui over gstreamer and we can have other interfaces over gstreamer too Sep 14 19:52:42 you can tell ldconfig -C /filetocacheto Sep 14 19:52:57 at least the one I have on my box Sep 14 19:53:14 fine Sep 14 19:53:20 but then each next time it will have to be called with that -C Sep 14 19:53:28 and we can't change ipkg scripts Sep 14 19:53:34 umm doesn't ld-linux.so also need to contain the same file path? Sep 14 19:54:11 yes Sep 14 19:54:38 =) strings /lib/ld-linux.so.3|grep ld.so.cache Sep 14 19:54:59 we can still link /var/run/ld.so.cache to /etc/ld.so.cache before calling ldconfig and the trick is done ! Sep 14 19:55:06 * daltona will try that Sep 14 19:55:23 unless it uses rm on the old file Sep 14 19:56:44 tuukkah, It is much faster if you call it twice so I think it does not Sep 14 19:57:17 I'm just about to make a test Sep 14 19:57:23 Ghiottone, i mean when it notices that it needs to recreate the cache Sep 14 19:58:14 tuukkah, when there is this need than better if it does its work Sep 14 19:58:29 s/than/then it is/ Sep 14 19:58:30 Ghiottone meant: tuukkah, when there is this need then it is better if it does its work Sep 14 19:58:51 but I would just update the cache, if I was him Sep 14 19:58:56 Ghiottone: that'll be because the files are all in memory cache Sep 14 19:58:57 or her... Sep 14 19:59:15 mmm did not think about some memory cache Sep 14 19:59:43 Ghiottone, yeah but see, if it uses rm, then it'll remove the symlink and the file on flash will become stale Sep 14 20:00:06 make symlink immutable Sep 14 20:00:09 tuukkah, greate point. Sep 14 20:00:12 actually - no Sep 14 20:00:16 you can't do that can you Sep 14 20:00:24 SpeedEvil, that would cause an error Sep 14 20:00:30 and probably make ldconfig stop Sep 14 20:05:42 ah, finally found answer to my question in the faq submission page in forum. surprised it wasn't in the wiki already. (The obvious question about when gta_02 phones are expected) I want accelerometers so I can make(or modify) cool apps that use them for user input. :) Sep 14 20:05:48 guys what version of dbus are you using on openmoko? Sep 14 20:06:29 dkr: would you be interested in an accelerometer thingy for $50 for your neo? Sep 14 20:06:44 If you have one Sep 14 20:06:58 me is currently constructing Sep 14 20:06:59 I don't yet, was waiting on gta_02 Sep 14 20:07:11 bedboi, there's http://buildhost.openmoko.org/OM2007.2/tmp/deploy/glibc/ipk/armv4t/dbus-1_1.0.2-r3_armv4t.ipk Sep 14 20:07:18 k Sep 14 20:08:27 ah cool, it is 1.0.x Sep 14 20:09:44 bedboi, are you working on something with openmoko and dbus? Sep 14 20:09:57 i would, i still need to get the phone Sep 14 20:10:21 i'm working in the FIND (Future INternet Design) project (NSF founded) Sep 14 20:10:27 we are planning to use the neo too Sep 14 20:10:47 What's FIND? Sep 14 20:10:55 better protocols and stuff? Sep 14 20:10:57 yep Sep 14 20:11:02 guaranteed bandwidth, ... Sep 14 20:11:02 better architecture Sep 14 20:11:08 SpeedEvil: all that. Sep 14 20:11:17 SpeedEvil: plus getting rid of IP Sep 14 20:11:29 In some ways tcp/ip suffered from not having design review. Sep 14 20:11:36 (well for the moment it's not true) Sep 14 20:11:40 In other ways, it benefited. Sep 14 20:12:10 SpeedEvil: companies just push the networking research to the direction they want Sep 14 20:12:21 Good evening for short. :) Sep 14 20:12:23 RFC's are company driven Sep 14 20:12:34 The scary thing is if they manage to get net-aneutrality in. Sep 14 20:12:46 And there are powerful interests for it. Sep 14 20:12:48 what you mean? Sep 14 20:12:54 ah managed networks? Sep 14 20:13:00 stuff like that? Sep 14 20:13:01 has everyone else been getting OpenMoko list emails recently? specifically community and announce - I haven't received anything for about 22 hours, and the archives has posts during that period Sep 14 20:13:28 stefan_schmidt - BTW: I've met alphaone and Shoragan today :) Sep 14 20:13:30 Weiss: do you miss not receiving 20 mails per second? Sep 14 20:13:37 If - for example, ISPs can deliver preferential performance to Sep 14 20:13:48 paying hosts, bad stuff happens Sep 14 20:13:56 SpeedEvil: well, there's a lot of debate about thta Sep 14 20:13:56 Weiss: and it's down here Sep 14 20:14:20 SpeedEvil: i was at a conference (Networking 2006), there was this cisco guy talking about "walled gardens" Sep 14 20:14:21 bedboi: yes. Indeed. I'm of the view that ISPs charge the end users for all the costs. Sep 14 20:14:59 it would be like toll roads is suppose Sep 14 20:15:00 SpeedEvil: he received a lot of "questions" Sep 14 20:15:14 tuukkah: something like that. Sep 14 20:15:20 Walled gardens are wonderful - for the companies. Sep 14 20:15:28 Not so much for the users. Sep 14 20:15:50 SpeedEvil: design by a small group of clueful people almost always yields better results than design by committee Sep 14 20:15:51 Want news? 10p/headline. Sports clips? 10p/clip Sep 14 20:16:04 tcp/ip isn't anywhere nearly as bad as some more recent related nastied Sep 14 20:16:06 SpeedEvil: exactly, that was the criticism all the researchers rose Sep 14 20:16:14 s/nastied/nasties/ Sep 14 20:16:14 cw666 meant: tcp/ip isn't anywhere nearly as bad as some more recent related nasties Sep 14 20:16:25 cw666: as long as they have decent armament to beat off hte hordes of others that want to install crap in a clean protocol. Sep 14 20:16:29 SpeedEvil: btw we are heading something really different. Sep 14 20:16:38 SpeedEvil: we don't care much about that. Sep 14 20:16:58 SpeedEvil: personally, i'm strongly against walled gardens and shit like that Sep 14 20:17:04 SpeedEvil: when it comes to the internet and modern sw/IT there will always be hordes of morons wanting to shoehorn in their crap without really thinking it through Sep 14 20:17:06 Pricing protocols for guaranteed bandwidth? Sep 14 20:17:18 Is there an FAQ page? Sep 14 20:17:29 SpeedEvil: mobility Sep 14 20:17:39 our major concern is about seamless mobility Sep 14 20:17:57 That's handy, yes, rather than pretending with VPN Sep 14 20:18:03 no no Sep 14 20:18:06 vpn are shitty Sep 14 20:18:13 and they are not transparent Sep 14 20:18:17 it's just a patch to IP :) Sep 14 20:18:26 a workaround Sep 14 20:18:49 I imagine authorisation is fun. Sep 14 20:18:55 yep, "fun" Sep 14 20:19:14 it's so bad. But that's one of the most difficult part Sep 14 20:19:23 it seems that ldconfig ignores the symbolic link Sep 14 20:19:29 i mean, designing something that just scales Sep 14 20:19:32 And doing another WEP would be bad. Sep 14 20:19:45 bedboi: im missed the topic here, what is the goal? Sep 14 20:20:00 well it's a project about future internet design Sep 14 20:20:02 Ghiottone, what does ignore mean here? Sep 14 20:20:50 that I have tried to set up the link as in ln -sf just before ldconfig gets called Sep 14 20:21:00 anybody has admin access to mokoprojects site ? Sep 14 20:21:02 but at restart the file is there and it is not a link Sep 14 20:21:11 my project's CVS is not working... Sep 14 20:21:17 Ghiottone: ldconfig -C /etc/ld.so.cache Sep 14 20:21:21 to write it Sep 14 20:21:26 SpeedEvil: btw we have a prototype of just a small part of the project Sep 14 20:21:29 and then the link should work for reading Sep 14 20:21:33 SpeedEvil: the cool stuff is NO NAT! Sep 14 20:21:39 that's what i love more. Sep 14 20:21:41 NAT is rather evil. Sep 14 20:21:48 damn evil Sep 14 20:21:56 What about cryptographically authenticated packets for slow links? Sep 14 20:22:11 SpeedEvil, we couldn't use -C because ldconfig is called from ipkg etc? Sep 14 20:22:12 you don't have to authenticate all packets Sep 14 20:22:28 SpeedEvil: you just have to be sure that there's no "impersonation" attack Sep 14 20:22:40 bedboi: you do if you are paying $.1 per K Sep 14 20:22:59 what? Sep 14 20:23:00 Topic change - sorry. Not talking about impersonation Sep 14 20:23:07 ah ok Sep 14 20:23:08 but expensive mobile connections Sep 14 20:23:25 SpeedEvil, ok I'll try. but I fear that next ldconfig will overwrite the link Sep 14 20:23:32 SpeedEvil: i'm not following Sep 14 20:23:58 You have a slow expensive link - say a mobile phone.of hosts, or a Sep 14 20:24:02 yep Sep 14 20:24:02 ok Sep 14 20:24:08 daltona, neither my svn is working on projects.openmoko Sep 14 20:24:09 So we can just copy /etc/ld.so.conf to /var/run/... on boot, before ldconfig runs and then put it back to /etc on shutdown Sep 14 20:24:11 You don't want anyone to be able to send you packets Sep 14 20:24:22 as then bad stuff happens to your bill. Sep 14 20:24:50 SpeedEvil: hmm, i'm not sure i got your point Sep 14 20:25:04 Ghiottone, that concerns me because I really want my project to be in CVS in order to keep track of versions Sep 14 20:25:24 SpeedEvil: if you don't want to receive nothing why are you connected ? :) Sep 14 20:25:31 I have a local CVS, but I would really prefer a moko official one... Sep 14 20:25:39 Expensive phone connection. Pckets can be sent by anyone. Sep 14 20:25:54 bedboi: including those virus infected probing, and stuff. Sep 14 20:26:04 ok using -C has made it use the old file Sep 14 20:26:16 but I think there is no earn Sep 14 20:26:19 bedboi: it would be nice to be contactable - but only by those with a key. And to be able to revoke that key Sep 14 20:26:27 I'm now about to create a bootchart Sep 14 20:26:34 SpeedEvil: yep, that would not be difficult to do Sep 14 20:26:38 in this new architecture Sep 14 20:27:04 you can say i want to receive just from entities that I have contacted first (something like you do with iptables) Sep 14 20:27:13 Of course - in the future, bandwidth will be free! Sep 14 20:27:20 yep of course. Sep 14 20:27:23 and no taxes Sep 14 20:27:27 and free girls i guess Sep 14 20:27:27 "in the year 2000" Sep 14 20:27:40 What about free beer? Sep 14 20:27:43 i want that Sep 14 20:27:49 That's not quite what I was meaning - I want to be able to specify hosts that I may not have contacted before. Sep 14 20:27:57 for example, other mobiles. Sep 14 20:28:07 yep, you could do that Sep 14 20:28:17 of course. This architecture would be very expressive. Sep 14 20:28:22 it's not a matter of IP's Sep 14 20:28:47 it's a matter of personality. You can attach whatever you want to your identity (or just your position without any further informations) Sep 14 20:29:19 actually it opens the door to a really big research area Sep 14 20:29:25 functional networking Sep 14 20:29:27 for instance Sep 14 20:30:24 functional networking? Sep 14 20:30:40 daltona, yeah. I would like to have it working too. so that I could put my sources there and then create some bb recipes Sep 14 20:31:19 SpeedEvil: it is putting intelligence in network devices. Sep 14 20:31:37 That'll be fun at terabit speeds :) Sep 14 20:31:41 yep Sep 14 20:31:41 Ghiottone, Interrested by the new bootchart... Sep 14 20:31:43 so fun Sep 14 20:32:00 SpeedEvil: we will have infinite CPU throughput (and free) Sep 14 20:32:29 What would be good to install would be a micropayment and billing scheme. Sep 14 20:32:29 i will push the electron to higher speed than light Sep 14 20:32:36 in process. I fear it will be the same Sep 14 20:33:12 Ghiottone, I don't think as removing ldconfig from bootmisc decreased the boot time of my neo... Sep 14 20:33:18 SpeedEvil: "No porn routed to my child" Sep 14 20:33:29 it's quite sci-fi Sep 14 20:33:49 In some ways that's a problem, as you've then got to tag porn. Sep 14 20:33:55 and btw it's not my field of research i'm much more about TCP congestion control Sep 14 20:34:11 Which means that you've got to have someone that decides what porn is Sep 14 20:34:25 SpeedEvil: which is questionable. Sep 14 20:34:54 is openmoko some kind of geek porn?? Sep 14 20:35:07 SpeedEvil: the best thing is to develop a synthetic version of a porn addict in the router Sep 14 20:35:12 with an arousal level Sep 14 20:35:17 and thresholding that level Sep 14 20:35:35 Ghiottone: it is. Sep 14 20:35:37 new bootchart show slightly worst results and the link is in place Sep 14 20:35:58 it means that ldconfig uses an in memory cache Sep 14 20:36:10 hello, does anyone know if openmoko will be supporting tmobile wing / htc herald? Sep 14 20:36:13 I'll try removing from the script and see how much is the gain Sep 14 20:36:19 Ghiottone, interresting Sep 14 20:37:03 jlf: at the moment, it barely supports the Neo1973 Sep 14 20:37:27 jlf: other hardware support is going to be user-based for the forseeable future. Sep 14 20:37:38 At least, large amounts of hardware support Sep 14 20:38:13 jlf: more like OE will start to support devices, then OM can be slightly adapted Sep 14 20:39:26 XorA: sorry, what's OE? Sep 14 20:39:39 jlf: Open Embedded, what is used to build OM Sep 14 20:40:17 ah, thanks. i've been waiting for the general consumption neo1973 but there's a vg deal on the wing atm Sep 14 20:42:23 Ghiottone about openmoko beeing geep pr0n... that depends how hard you are into s&m ;) Sep 14 20:42:58 * mjr applies lubricant Sep 14 20:43:26 roh, eheheh :) Sep 14 20:44:47 you know, about the complaints that the Neo could be smaller, I *ungh* agree Sep 14 20:45:02 how do I force the build system to refetch the psplash pkg? Sep 14 20:45:24 grasu, bitbake -c rebuild psplash I think Sep 14 20:45:52 grasu, you have to remove the stamps file from build/tmp/stamps/arm*/ Sep 14 20:46:32 already removed /build/tmp/stamps/armv4t-angstrom-linux-gnueabi/psplash-0.0+svnr249-r9.do_fetch Sep 14 20:46:43 but it does not do it Sep 14 20:47:03 or make rebuild-package-psplash with mokomakefile Sep 14 20:47:22 grasu, remove the unpack as well Sep 14 20:47:29 and then do the rebuild Sep 14 20:47:37 I removed all of them Sep 14 20:47:54 oh yeah Sep 14 20:47:57 stupid I am Sep 14 20:48:06 there should be a tar.gz in source dir Sep 14 20:48:17 s/source/sources/ Sep 14 20:48:17 Ghiottone meant: there should be a tar.gz in sources dir Sep 14 20:48:23 I did that a lot of time... Sep 14 20:48:36 results for new bootchart are interesting Sep 14 20:48:52 it needs 162 seconds instead of 194 seconds Sep 14 20:50:55 I tried make rebuild Sep 14 20:51:16 but it still complains about the missing tar.gz Sep 14 20:51:42 grasu, it could have recreated the stamps Sep 14 20:51:47 Ghiottone, you can put charts here: http://wiki.openmoko.org/wiki/Bootcharts Sep 14 20:53:41 tuukkah, ok Sep 14 20:54:05 Looks like ldconfig is eating a lot of time. Sep 14 20:54:32 Ghiottone, So we have to find a way to store the cache in flash or to delay the execution Sep 14 20:55:52 daltona, we could simply change ldconfig and ld-linux.so to use the default location again :-) Sep 14 20:56:04 Isn't it possible to pre-make ld.so.cache? And if necessary, recalculate it when new software is installed. Sep 14 20:56:29 that's how it works, yes Sep 14 20:56:33 tuukkah, yes of course Sep 14 20:56:58 Check to see if /etc/default/volatiles/00_core can modified to automagically create the correct link to a place in flash to store ld.so.cache. Sep 14 20:56:58 don't know why a volatile location was chosen Sep 14 20:57:07 tuukkah, thanks for the credits:) the bootchart is on the wiki Sep 14 20:57:09 it must have a reason Sep 14 20:57:38 I don't think that saving cache will work Sep 14 20:57:43 daltona, in deed Sep 14 20:58:08 I think that ldconfig checks the in memory cache Sep 14 20:58:31 that could as well be internal kernel or buffers for .so objects Sep 14 20:58:55 if the memory cache is not there, it needs a lot of time even if it finds its cache Sep 14 20:59:11 Ghiottone, well you tested and the boot became much faster? Sep 14 20:59:28 as the time it took in bootchart was the same as before when I told it to use /etc/ld.so.cache Sep 14 20:59:37 tuukkah, yes, but without running ldconfig Sep 14 20:59:44 oh ok Sep 14 21:00:15 tuukkah, interestingly, you can see now programs that sleeps with I/O Sep 14 21:00:37 wow :-) Sep 14 21:00:53 probably ldconfig causes some kind of prefetching or precaching Sep 14 21:04:11 re Sep 14 21:04:55 did anyone play with gllin in chroot? Sep 14 21:05:52 I see no reason why it won't work. Sep 14 21:06:16 hey hhf423 Sep 14 21:06:31 hi Ghiottone Sep 14 21:06:44 Go vegetarian! Sep 14 21:06:45 oops Sep 14 21:06:46 I get "permission denied" for gllin Sep 14 21:07:08 where did you get it? Sep 14 21:07:21 I thought it was only present in phase 0 phones Sep 14 21:07:22 I don't get it, I checked the whole chroot and did a chown -R root.root Sep 14 21:07:25 hhf423: on 2007.2? Sep 14 21:07:38 .1 rather Sep 14 21:07:44 SpeedEvil: yes, on 2007.2 Sep 14 21:07:50 with a .1 chroot? Sep 14 21:07:53 yes Sep 14 21:08:17 Ghiottone, apparently some of the p0 list were evil and have told other people the number Sep 14 21:08:48 mjr, eh I see! Sep 14 21:09:18 (should really try that chroot thing myself..) Sep 14 21:10:08 tuukkah, I don't think bootchart is taking that much cpu. I tried running it with start and then stop for a few seconds while doing some ls, find commands. Sep 14 21:10:11 why could I get permission denied? Sep 14 21:10:23 strace Sep 14 21:10:39 tuukkah, cpu usage started when I called stop when bootchartd probably gathered the results Sep 14 21:10:50 SpeedEvil: strace and then the command? Sep 14 21:11:19 Ghiottone, did you use 0.1 or 1 second sampling? Sep 14 21:11:53 hhf423: have you tried just running ash in the chroot? Sep 14 21:11:54 hhf423, strace -e trace=file gives much less output Sep 14 21:12:09 tuukkah, 0.1 but disabled process accounting as I got an error and it refused to start with it enabled Sep 14 21:12:54 chroot: cannot execute /home/root/gps/gllin: Permission denied Sep 14 21:13:12 SpeedEvil: ash? Sep 14 21:13:22 hhf423: do you have all the libs required? Sep 14 21:13:40 Ghiottone, isn't process accounting the largest part of work bootchart does? anyway, i had it take a lot of cpu but i will be happy if there's a way around that Sep 14 21:14:10 SpeedEvil: I think so Sep 14 21:14:21 chroot ash Sep 14 21:14:42 doesn't work quite like that but you get the point. the shell Sep 14 21:15:15 my last chart has been created with this config Sep 14 21:15:41 chroot /path/to/chroot/root ash Sep 14 21:16:50 Ghiottone, ok so with "process accounting" you meant the acct stuff Sep 14 21:17:00 yes Sep 14 21:17:15 this name is used in bootchartd Sep 14 21:17:21 hmm Sep 14 21:17:29 before or after screen? Sep 14 21:17:59 no need for screen here? Sep 14 21:19:08 Ghiottone, there must be something wrong in my system then... Sep 14 21:19:24 why? Sep 14 21:19:37 that 90% cpu usage you told me this afternoon? Sep 14 21:19:42 can you put that online? Sep 14 21:20:04 the charts you mean? Sep 14 21:20:33 SpeedEvil: sorry, I don't get it. would that be chroot /home/root/chroot/ ash ? Sep 14 21:20:51 Ghiottone, didn Sep 14 21:20:56 ups Sep 14 21:21:07 tuukkah, yes Sep 14 21:21:28 well it didn't work I manually had to refetch the files Sep 14 21:21:42 thank you all anyway Sep 14 21:22:45 grasu, strange :/ Sep 14 21:23:20 did you remove only the tar.gz or also the .lock and .md5 files? Sep 14 21:23:47 I think only the tar.gz Sep 14 21:24:02 well anyway it works now Sep 14 21:24:28 http://tuukka.iki.fi/tmp/sleep-0.1.png and the middle of http://tuukka.iki.fi/tmp/sleep-1.png Sep 14 21:28:47 tuukkah, strange graphs indeed. I do not understand 0.1 at all. is that first 5 seconds? Sep 14 21:29:12 tuukkah, in the second graph I can't explain that period of lower cpu usage before the end of boot Sep 14 21:29:42 the first is taken after boot when the system was idle otherwise with bootchartd start sleep 5 Sep 14 21:29:55 strace also does not give any more cues Sep 14 21:30:23 the second is where i put sleep 60 to /etc/matchbox/session Sep 14 21:31:42 What might be interesting would be to swiitch it to 50Mhz mode if higher resolutoin was needed Sep 14 21:31:52 ah well, off to sleep then Sep 14 21:32:03 hhf423: if theroot is Sep 14 21:32:05 oh well Sep 14 21:32:39 fastest go to sleep-er ever existed! Sep 14 21:33:26 roh, should we report to you problems we experience with projects.openmoko.org Sep 14 21:33:29 roh, ? Sep 14 21:37:18 i ran "time dd if=/dev/mtdblock4 of=/dev/null" and now X is frozen and all commands over ssh result in"Input/output error" Sep 14 21:37:38 wow Sep 14 21:37:57 try adding count=some value Sep 14 21:38:26 all i wanted to do was test the speed of the memory, i didn't want to get io errors %-) Sep 14 21:38:57 yeah perhaps it went past the end of device or something Sep 14 21:40:04 Worked fine on my p1 unit (I added "bs=4096" though) Sep 14 21:41:38 Ghiottone oeh.. into bugzilla Sep 14 21:42:06 roh, ok Sep 14 21:44:08 Ghiottone actually.. i have projects.openmoko.org on my todo.. to check whats the situation and what needs to be done Sep 14 21:45:56 roh, ok so we have to wait, do we? Sep 14 21:46:08 does anyone know a free sans serif font that's not so ubiquitous as Vera (DejaVu doesn't count) but has the same quality? Sep 14 21:46:19 actually... we'll see... is it about approval of new projects? Sep 14 21:46:47 mwester, after a reset, your version worked once and hung on the second try, same thing as last time: ctrl-c doesn't work, but ctrl-z gets back to shell but all commands give io error Sep 14 21:46:51 roh, no. the project was approved. after a week or so, but it was approved Sep 14 21:47:12 roh, problem is that the subversion repository seem to be missing Sep 14 21:47:25 roh, so I cannot put the code in the project Sep 14 21:47:40 ah.. i see... Sep 14 21:48:32 roh, I think that it simply has not been created Sep 14 21:49:18 dmesg worked though probably because i had run it recently and latest messages where "<4>Erase at 0x006dc000 failed immediately: errno -5" and also in 0x006d8000, 0x006d4000, 0x006d0000 Sep 14 21:50:23 mickeyl - Hi. :-) Sep 14 21:50:27 hi elrond Sep 14 21:50:49 Ghiottone: I've been staying away from projects.openmoko.org for mokomakefile until the anon svn access situation is fixed and there is someone assigned to maintaining it (I believe the sysadmin comes on board next week?) Sep 14 21:51:44 Then I will see what the response time is for a "there is a hung transaction in my svn repo on projects.om.org" and then decide whether to move mokomakefile back there. Sep 14 21:52:04 mickeyl - alphaone and Shoragan are having fun in Darmstadt. ;) Sep 14 21:52:06 (hopefully they already use fsfs and not bdb backend for the svn server) Sep 14 21:52:26 Elrond: cool. if all goes well, I'm going to come over for a coffee tomorrow Sep 14 21:52:49 rwhitby: ack. the sysadmin will take care of keeping it working and michael and me will take care of approving Sep 14 21:53:00 mickeyl - Greet them again. I'm probably on a tour or so. Sep 14 21:53:05 k Sep 14 21:55:26 Any word on the GPS driver for P1? Sep 14 21:56:23 mwester the eabi binary is underway, but does not work yet (it did for oabi.. but that doesnt help now) Sep 14 21:56:37 roh: it will be more-or-less the same binary then? Sep 14 21:56:49 roh: i ask this because i think i found a problem with the oabi one that ideally should be sorted Sep 14 21:57:41 sorted! Sep 14 21:57:57 mwester - I'm currently updating my desktop, when that is done, I can update to eabi, after which I'll hopefully be motivated again to work on sphyrna ;-) Sep 14 21:58:07 cw666 i havent tried it yet, and still have not seen documentation about its proper usage.. so i dunno Sep 14 21:58:51 roh: anyhow, if i let the phone 'suspend' (pull PM mode as much as that's anything now) but leave the modem on, and gllin running, on resume/wake it never seems to get a fix again Sep 14 21:58:53 mickeyl - BTW: What's the difference on those three uboot versions on buildhost? (for Bv3 for example) Sep 14 21:59:04 roh: so i have to make sure everything is at full power when poking about Sep 14 21:59:10 also, gllin wakes up A LOT Sep 14 21:59:15 it would be nice if it didn't do that Sep 14 21:59:29 cw666 - It has to. Sep 14 21:59:45 Elrond: no idea offhand. no one should use an untested uboot from buildhost Sep 14 21:59:46 cw666 - And the resume thing: Just delete nvram* and restart it. Sep 14 22:00:04 Elrond: restarting means it can take some time to get a fix Sep 14 22:00:12 Elrond: also, why does it have to wake all the time? Sep 14 22:00:20 Elrond: it's not talking to the chip most of those wakes Sep 14 22:00:30 cw666 - Yes, restarting will take a time. gllin is a bad driver. Sep 14 22:00:42 chances are it's all we'll have for some time Sep 14 22:00:49 cw666 - Huh. If it's not talking to the chip, it's even worse. :-| Sep 14 22:01:01 so how's gllin supposed to work ? Sep 14 22:01:10 is it a complete substitute for gpsd? Sep 14 22:01:15 i.e. does it listen on TCP? Sep 14 22:01:27 mickeyl nope. it spits out nmea at a named pipe as far as i understood. Sep 14 22:01:34 mickeyl - No. It writes nmea to a named pipe. Nothing more. Sep 14 22:01:39 hmm Sep 14 22:01:45 mickeyl: I thought gllin was a GSM daemon Sep 14 22:01:48 oops, GPS Sep 14 22:01:49 how broken is that? Sep 14 22:01:51 mickeyl gpsd/a plugin for gpsd then gets the data and from there on its all opensource Sep 14 22:01:57 mickeyl: that's not a bad model really Sep 14 22:02:07 mickeyl: what should it do then? Sep 14 22:02:20 well, namedpipe mode crashes here Sep 14 22:02:39 however it listens on a udp socket Sep 14 22:02:41 gllin 5506 root 6u IPv4 14252 UDP *:32768 Sep 14 22:02:43 ? Sep 14 22:03:00 mickeyl: it can also spew stuff out a socket it looks like Sep 14 22:03:25 mickeyl i think there is the possibility to feed agps data from the gsm (when the modem supports that stuff and the carrier does use it) into the damon Sep 14 22:03:36 eventually the udp socket is for that Sep 14 22:04:30 the DOP values reported from gllin are also wrong as best as i can tell Sep 14 22:04:40 it's claiming accuracy that is can't have (at least for me) Sep 14 22:06:11 mickeyl - I take it that the Liberation fonts are no good for your needs. Sep 14 22:06:29 no, they're too much windows font clones Sep 14 22:06:33 cw666 - I'm not surprised, really. Sep 14 22:06:34 although the mono is very good Sep 14 22:06:42 and openmoko-terminal2 uses it now Sep 14 22:07:00 mickeyl - cool. Sep 14 22:07:09 cw666 - The HH chip is a really fun thing. The driver just feels crap. Sep 14 22:07:17 ya, finally a good font in om-terminal Sep 14 22:07:57 tuukkah, you seemed to be right about bootchart. I have tested it by waiting almost 10 minutes after booting the neo Sep 14 22:08:14 Elrond: c++ for a driver is always a bad sign of fuckhead clueless hw vendor brandamage Sep 14 22:08:19 tuukkah, and processor usage stayed at 100% even after startup completed Sep 14 22:08:47 tuukkah, the previous test I did that showed 0% usage was with bootchart started manually after boot completed Sep 14 22:09:08 hmm, what does GLLIN_WD mean? Sep 14 22:09:14 working directory? Sep 14 22:10:57 hmm oh Sep 14 22:11:09 * mickeyl goes outside trying to get a 'fix' Sep 14 22:11:10 bbiab Sep 14 22:12:05 cw666 - Yeah. Sep 14 22:12:30 cw666 - Still, the chip is really a nice thingy. Sep 14 22:13:29 Ghiottone, ok. how did you start it manually? something else than my bootchartd start sleep 60? Sep 14 22:13:36 Elrond: well, if the RE effort doesnt get far enough and we're at the mercy of the driver we've seen so far, or even the same code-base people, then im not sure i would hope for much Sep 14 22:14:13 tuukkah, yeah. simply bootchartd start Sep 14 22:14:19 tuukkah, then did some commands Sep 14 22:14:25 tuukkah, and then bootchartd stop Sep 14 22:15:26 tuukkah, mmm I have just noticed that phone was in PM mode. I'll try that test again Sep 14 22:15:26 Elrond: btw, with the chip how often does the CPU need to wake and poke about? Sep 14 22:23:28 cw666 - With good signal: Every 0.7sec or so. With bad signal: Every 1.5sec. Sep 14 22:23:36 cw666 - And then do some calculations, etc. Sep 14 22:28:39 Elrond: can you trade wake ups for accuracy? Sep 14 22:34:05 re Sep 14 22:34:25 did you see that? http://scap.linuxtogo.org/files/43091acec6b18755afc111cb45857f5c.png Sep 14 22:34:32 mickeyl: do we want openocd in moko-autorev.inc? Sep 14 22:35:12 rwhitby: yes Sep 14 22:35:34 cw666 - Not really. Good signal == integrators get filled up quicker. Sep 14 22:36:00 Elrond: i guess im dont really know how it works ... so waking every say 30s wouldn't be useful? Sep 14 22:36:08 cw666 - Hmm... thinking again about it: If you wake up earlier for bad signal, you get worse. ;o) Sep 14 22:36:40 cw666 - Every 30s: You'd probably need two or three shots to get a fix again. Sep 14 22:37:29 Elrond: how long is a shot? Sep 14 22:37:40 cw666 - The 0.7 - 1.3 sec from above. Sep 14 22:38:01 cw666 - You queue a shot on the chip, then the chip does the measurement for you. Sep 14 22:39:20 Elrond: so under 10s from waking? Sep 14 22:40:29 cw666 - So yes... One short every 30s is possible. But you shouldn't do bad things to the gps in that time, otherwise it has to say "Oh wait... you bad guy stopped my view of satelite number 13, I need that one" and will need a few (7) shots to look for other sats, which are visiable now. Sep 14 22:40:43 s/short/shot Sep 14 22:41:32 Elrond: where is all this information from? is there a URL i can get to read up on how this works in this case? Sep 14 22:42:17 can anyone tell what I see here? http://scap.linuxtogo.org/files/43091acec6b18755afc111cb45857f5c.png Sep 14 22:42:45 is that gllin based or "better"? Sep 14 22:42:57 mickeyl: one very good side effect of srcrev is now when you do a new build, you can see exactly which apps you need to test for cool new features after flashing. Sep 14 22:42:59 gllin almost certainly Sep 14 22:43:20 $PGLOR is vendor specific and gllin spews it out, i doubt any other non-hh drivers would do that Sep 14 22:43:29 well, yea Sep 14 22:43:39 it's also a dynamic format from the looks of this, unlike pretty much all over NMEA records (sigh) Sep 14 22:44:02 cw666 - Mostly based on my own research. sphyrna svn has some docs also. Sep 14 22:44:08 cw666: hmm Sep 14 22:46:40 ok, openocd is now srcrev based. I think that's the last one for things that MokoMakefile builds. Sep 14 22:47:20 cool, thanks Sep 14 22:47:28 * mickeyl goes outside to check whether the "fix" is in yet Sep 14 22:47:30 *sigh* Sep 14 22:47:43 Now to remove all date-based stamps and work dirs and see if any come back (except for Angstrom-version) Sep 14 22:47:51 any news on usb host mode btw? Sep 14 22:48:28 hhf423 nope. thats lower prio than gps and gsm. Sep 14 22:48:55 hhf423: works Sep 14 22:49:06 hmm nothing Sep 14 22:49:09 hhf423: apply patch in SVN, and the comment I made in bugzilla Sep 14 22:49:12 err Sep 14 22:49:20 mailing list - neo1973-kernel I think Sep 14 22:49:23 would you know how i can test whether gllin does anything at all? Sep 14 22:49:26 roh: is anyone working on suspend/resume? Sep 14 22:49:27 I forget who had it working earlier Sep 14 22:49:32 mickeyl: run it Sep 14 22:49:35 roh: Zeroping seems to have the necessary patch Sep 14 22:49:42 excellent, the only floating datestamp is not angstrom-version (which is intentional) Sep 14 22:49:44 SpeedEvil: well, it runs. but there is no output Sep 14 22:49:49 except two hal statements Sep 14 22:49:55 hhf423 ah.. nice.. then.. have fun playing with it Sep 14 22:49:56 see [[GPS]] in the wiki Sep 14 22:50:08 right Sep 14 22:50:15 i should consult our own wiki more often Sep 14 22:50:16 heh Sep 14 22:50:38 s/not/now/ Sep 14 22:50:38 mickeyl: let me know if you can make gllin do anything. I get permission denied Sep 14 22:50:38 rwhitby meant: excellent, the only floating datestamp is now angstrom-version (which is intentional) Sep 14 22:51:05 http://www.mauve.plus.com/dogll Sep 14 22:51:07 SpeedEvil: which bug number is usb host? Sep 14 22:51:08 is the script I use Sep 14 22:51:15 roh: btw, how do we tell the modem to send interrupts? Sep 14 22:51:19 roh: so we can have it wake the CPU? Sep 14 22:51:28 rwhi; sorry - it was not bugzilla - but mailing list Sep 14 22:51:29 roh: well, Zeroping has a patch, but its not comitted yet Sep 14 22:51:30 hang on Sep 14 22:51:39 mickeyl: are there any fresh news about shipping? I'm sure people is keep asking this silly questions Sep 14 22:52:00 no new Sep 14 22:52:01 s Sep 14 22:52:05 otherwise we would publish them Sep 14 22:52:05 heh - "cupcake" :-) Sep 14 22:52:12 and 'dogllstrace' Sep 14 22:52:29 which is similar - but straces and gzips to the media card Sep 14 22:52:50 SpeedEvil: cupcake? whats that? Sep 14 22:53:02 So, now that we have build stability, is there anyone with outstanding build problems? Sep 14 22:53:43 (cause now we can check their tyre pressure while the car is stationary) Sep 14 22:53:50 ;) Sep 14 22:53:58 It's a binary that turned up in my email at one point. Sep 14 22:54:17 rwhitby: there are bluetooth tyre pressure monitors! Sep 14 22:54:18 Where! Sep 14 22:54:19 :) Sep 14 22:54:20 SpeedEvil: is that a gllin for 2007.2? Sep 14 22:54:23 no. Sep 14 22:54:24 rwhitby: last i built i didnt get a new uboot, probably my bad though Sep 14 22:54:44 cw666: were you expecting a new one? Sep 14 22:54:59 (i.e. did it already have the latest one built?) Sep 14 22:55:09 rwhitby: i did a clobber and hoped i would get something, what im using now is old Sep 14 22:55:10 rwhitby: i got a new built on my build last night, but I suspect changes since then - do you recommend just an update now or rebuid from scratch? Sep 14 22:55:11 (latest according to OE srcrev setting) Sep 14 22:55:35 SpeedEvil: but there are bt blood pressure measurement devices, bt weight measurement devices, bt blood oxygen level measurement ... Sep 14 22:56:55 Writchie: if you rebuilt from scratch in the last couple of days an update should be all that's required. Sep 14 22:57:09 last build was scratch, tx Sep 14 22:57:14 you might want to do a build from scatch anyway to get rid of all the incrementing date work dirs Sep 14 22:57:40 (check our build/tmp/stamps/*/ dirs to see how bad that is) Sep 14 22:58:18 rwhitby: 'make clobber && make' ? Sep 14 22:58:48 SpeedEvil: so how does this work without chroot? Sep 14 22:59:08 cw666: that should do it Sep 14 22:59:21 hhf423: I'm still on 2007.1 Sep 14 22:59:23 rwhitby: tried that a few hours ago and it blew up Sep 14 22:59:25 let me try again Sep 14 22:59:34 SpeedEvil: ah, right Sep 14 22:59:47 cw666: if that blows up I'll want to know about it Sep 14 23:00:04 where is the motor? Sep 14 23:00:06 oops Sep 14 23:00:19 SpeedEvil: cupcake is what? gllin? Sep 14 23:00:35 cw666: yeah - the program that shall not be named :-) Sep 14 23:01:38 im lost, given gllin is out there who cares if you name it? Sep 14 23:02:25 well, someone who is not a P0 developer, who has access to gllin, is technically in possession of software without a valid license agreement Sep 14 23:03:00 (as are the P0 developers, but at least they didn't go out and get it, it was given to them by FIC on the device without their express knowledge) Sep 14 23:03:00 rwhitby: really? a lawyer gave this opinion? Sep 14 23:03:33 cw666: some people like to play it safe and assume that if FIC is not freely redistributing gllin, then there must be a reason for that. Sep 14 23:03:40 ymmv Sep 14 23:03:57 rwhitby: I think under UK law as they didnt issue a recall in 28 days, I own gllin and there is nothing can be done :-D Sep 14 23:03:57 rwhitby: it's not, but that doesnt mean possesion of it is necessarily wrong or criminal Sep 14 23:04:42 cw666: note that I didn't say the words "wrong" or "criminal" - I was very careful in the words I used. Sep 14 23:05:15 rwhitby: in fact, it's not even clear that if someone was to upload it to a public website that would be illegal Sep 14 23:05:22 certainly they might make a target of themselves though Sep 14 23:05:25 that judgement is up to each individual to make for themselves Sep 14 23:05:36 And perhaps the courts. Sep 14 23:05:45 SpeedEvil: which courts? Sep 14 23:05:50 no need to stir the ants nest anyway when a real version is on the way Sep 14 23:05:57 SpeedEvil: all over the world there are many different legal systems Sep 14 23:05:58 I think thats how most devs see it Sep 14 23:06:06 cw666: indeed. Sep 14 23:06:21 XorA: oh, i quite agree... i think thus far people who have it have kept it to themselves for the most part Sep 14 23:06:45 i dont even see RE efforts getting main traction if the closed driver/whatever doesnt entirely suck Sep 14 23:06:54 ah, probably it an issue of my legal system that I get permission denied in my chroot environment when I want to execute cupcake Sep 14 23:06:56 XorA: I have a real job and a family to support, and I need to travel to the US often to do so, so I personally always make the personal judgement not to flout things which could potentially cause problems at LA airport when I land :-) Sep 14 23:07:17 rwhitby: ah the threat of the rubber gloves :-D Sep 14 23:07:39 after a 13.5 hour flight in economy from australia, you just don't need hassle of any sort. Sep 14 23:08:04 rwhitby: or legal guys just recently gave the advice: US customs/immigration is not the place to discuss the constitution Sep 14 23:08:31 hhf423: rwhitby gets no rights from the constitution anyway Sep 14 23:08:38 indeed - doesn't the TSA have powers that transcend the constitution? Sep 14 23:08:41 rwhitby: if they want your laptop, you give them your laptop. you might even get a reciept Sep 14 23:08:52 rwhitby: unless you're a terrorist Sep 14 23:08:56 * XorA doesnt like Orange, only black so Ill behave Sep 14 23:09:33 XorA: that advice came from US legal and was menat for US and non US personnel Sep 14 23:09:38 Yeah, ain't that the truth. (mwester is a card-carrying member of the frustrated frequent US flier program) Sep 14 23:09:57 XorA: in front of TSA noone has any rights Sep 14 23:10:08 roh: http://lists.openmoko.org/pipermail/neo1973-hardware/2007-August/000201.html - and the one in svn reportedly works Sep 14 23:11:19 hhf423: I am well aware of the psycotic institution US people have tied themselves up with Sep 14 23:11:39 hhf423: its one of the reasons I wish to visit US as little as possible. Sep 14 23:11:57 hhf423: I dont fancy a kidnapping to cuba just because someone mistypes my name Sep 14 23:12:18 * mwester keys "XorA" into a query, and sees what names come up... Sep 14 23:12:35 yeah, or you sat next to the wrong person. or you dropped your ipod down the toilet like the guy in CA Sep 14 23:12:41 I've never had any problems visiting the US, but then I was always travelling as a Motorola or Freescale employee - big US companies :-) Sep 14 23:13:19 * XorA is off to florida in Feb Sep 14 23:13:34 Actually the worst experience I ever had was clearing customs in Toronto, as a US citizen going to visit a Canadian company... Sep 14 23:13:42 XorA: where in Florida? Sep 14 23:13:49 Orlando Sep 14 23:13:54 XorA: do you ever have problems leaving the US - being an encryption algorithm subject to export restrictions? Sep 14 23:14:04 i'm in Daytona Sep 14 23:14:07 hahaha Sep 14 23:14:22 rwhitby: well XorA is one way encryption, its impossible to retreive the results Sep 14 23:14:25 XorA: conference? Sep 14 23:14:28 one should not wear a ROT13 shirt Sep 14 23:14:38 XorA: all the more reason why it is restricted ;-) Sep 14 23:14:39 Writchie: familly holiday with wifes familly Sep 14 23:14:46 ah Mickey Sep 14 23:14:59 hhf423: Yep, good advice -- the TSA employees that actually get it do NOT appreciate that t-shirt. Sep 14 23:15:36 hhf423: And don't tattoo an encryption algorithm on yourself either... Sep 14 23:16:06 Writchie: yeah, disney sucks, I just want to see Terminator 3d and TNA Wrestling Sep 14 23:16:11 mwester: aww, so sexy encryption algo tattoo? Sep 14 23:16:24 * ckuethe is flying to the states next week ... should i wear my blowfish shirt, my rc4 shirt or my rsa shirt? Sep 14 23:16:42 doesn't matter what you wear entering ... Sep 14 23:17:29 ckuethe: they wont care, just behave and dont be a retard Sep 14 23:17:31 or leaving, for that matter - they're all (a) printed matter and (b) open source Sep 14 23:17:31 ckuethe: where you Transportation Stupidity Administration shirt. Sep 14 23:17:37 On a serious note, this is why I'm very concerned about what get's installed on the neo -- I have to cross borders, and I don't look good in orange jumpsuits. Sep 14 23:17:39 i don't have one of those... Sep 14 23:17:59 its wierd, my dads ex girlfriend flew from US to UK with a colt 45 in hand luggage as you werent allowed to stow it. How insane is that Sep 14 23:18:00 mwester: just don't set your "wallpaper" to be kiddie pr0n Sep 14 23:18:13 leave it on the dialler Sep 14 23:18:31 so long as it looks like some wanky new smartphone you should be ok Sep 14 23:18:32 how insane is more than 3.8 oz of toothpaste Sep 14 23:18:32 XorA: when was that? Sep 14 23:18:54 no one's ever bothered to snoop through my calls list or texts on my phone Sep 14 23:18:59 hhf423: 6 years ago I think Sep 14 23:19:01 and make sure the Neo does not make any ticking noises Sep 14 23:19:23 mwester: which boarder? Sep 14 23:19:27 tell them it's a new iphone Sep 14 23:19:29 current .EU rules for flying are totally insane and only serve to make Duty Free shops more money Sep 14 23:19:30 Writchie: I wonder about that -- specifically, how big a bang does 3.7 oz of explosive make on an airplane, and why is 3.8 bad, but 3.7 is ok?? Sep 14 23:19:43 what is an oz? Sep 14 23:19:48 28ml Sep 14 23:19:48 28 gramms Sep 14 23:19:55 something the british gave us Sep 14 23:19:57 Vegar: someone from australia Sep 14 23:20:03 Vegar: a place where wizards live I think Sep 14 23:20:12 Writchie: do not piss off tsa Sep 14 23:20:16 * u_l-lap <== on no fly list Sep 14 23:20:16 hhf423: no, that's an aussie Sep 14 23:20:17 cw666: Canada, mostly. Sep 14 23:20:21 http://www.designverb.com/wp-content/images/2007/06/Metric_system.jpg Sep 14 23:20:22 right.. volume or mass? Sep 14 23:20:32 mwester: US/CA are friendly, i wouldn't worry too much Sep 14 23:21:02 there are nice/asshole customs officers everywhere Sep 14 23:21:07 best of luck... Sep 14 23:21:16 u_l-lap: basically, mouting off to any security, custom, etc. guy is just plain stupid. Sep 14 23:21:30 cw666: is that a map of all the english colonies that are still stuck in the past? Sep 14 23:21:42 rwhitby: metric system vs imperial Sep 14 23:21:43 http://64.233.183.104/search?q=cache:UsyhGIsrJY8J:forums.worldofwarcraft.com/thread.html%3FtopicId%3D11211166+canada+ipod+toilet+flight&hl=de&ct=clnk&cd=1&gl=de Sep 14 23:22:02 * rwhitby forgot the smiley Sep 14 23:22:05 when I moved to Canada about 15 years ago the truck driver mouthed off at the border tha I had some computers and my household goods were impounded for weeks Sep 14 23:22:15 well, you could argue that too Sep 14 23:22:44 cw666: That's why I was surprised to spend three hours in immigration in the Toronto airport, trying to explain why I wished to enter Canada. Sep 14 23:22:55 mwester: never use the C word Sep 14 23:23:06 they don't like consultants Sep 14 23:23:18 mwester: when i had long hair and travel alone a lot if got problems, since then nobody cares Sep 14 23:23:31 getting old and having you hair fall out somehow helps Sep 14 23:23:33 the code to get across the Candadian border is be there for "Internal Sales Metting: Sep 14 23:23:37 the best place to enter the US is dublin airport, that rocked Sep 14 23:23:52 XorA: edmonton, alberta Sep 14 23:24:06 LAX tends to suck because it's very bery busy Sep 14 23:24:31 Seatle from China or Hong Kong is worst Sep 14 23:24:31 dublin rocks, because at least you can drink guiness before you queue up Sep 14 23:24:34 Writchie: wearing the wrong clothes will enrage tsa Sep 14 23:24:38 it's never taken me more than 30s to clear us customs in edmonton Sep 14 23:25:47 in hundreds of border crossing the only hassles or searches I ever had were in US or Canada Sep 14 23:26:13 suuurrre.... blame canada. Sep 14 23:26:14 ckuethe: try living in new mexico Sep 14 23:26:21 ckuethe: Ok! Sep 14 23:26:29 bedboi: i visited there a few times Sep 14 23:26:32 if you go near the border you have a lot of controls Sep 14 23:26:40 when you head north Sep 14 23:26:43 ABQ airport sucks Sep 14 23:26:55 ckuethe: eheh, i'm based in abq right now. yes it sucks Sep 14 23:27:07 u_l-lap: you going to tell us the story? Sep 14 23:27:13 ckuethe: but not that much, my hometown airport is the one that really sucks Sep 14 23:29:03 rwhitby: beard, anarchist german thrash band tshirt, steel toed boots, persian last name Sep 14 23:29:04 rwhitby: they still spray the inside of the airplanes before you land in Australia Sep 14 23:29:22 rwhitby: add a one way ticket and you have pissed off tsa people and get yourself on the no fly list Sep 14 23:29:40 :/ Sep 14 23:29:42 u_l-lap: ouch you ran out of boxes on the form Sep 14 23:29:49 now I ride amtrak ostly because they don't care Sep 14 23:29:54 Writchie: no Sep 14 23:29:58 but I can't leave the country Sep 14 23:30:13 how exactly does the output mode switch from speaker to headphones when you plug the headphones into the neo ? Sep 14 23:30:16 u_l-lap: ouch! Sep 14 23:30:35 does neod do that, or is there some other daemon involved ? Sep 14 23:30:39 u_l-lap: you guys will just have to vote for a sane president this time round Sep 14 23:30:51 * mwester wonders why one would dress that way, and attempt to cross airport security... Sep 14 23:31:16 luckilly in UK our insane PM seems to have gone and the new one is saner Sep 14 23:31:26 oh no, I'm now a known associate of u_l-lap ! Sep 14 23:31:29 although I wont breathe too easy yet Sep 14 23:31:29 XorA: At the risk of straying into politics, the problem is that we've lacked sane candidates for many many years. Sep 14 23:31:33 bring back margaret Sep 14 23:31:42 Regarding sanity, seems like SCO just filed for Chapter 11. Sep 14 23:31:46 XorA: very unlikely we haven't a sane president since ... ever Sep 14 23:31:58 XorA: the first ten or so were just a bit less obviously evil Sep 14 23:32:01 just another abuse of the legal system Sep 14 23:32:12 Chapter 11 protects them Sep 14 23:32:21 heh heh margaret was in secret meetings with new PM :-D Sep 14 23:32:23 u_l-lap: are you kidding? Sep 14 23:32:27 There's the theme: the lawyers. Sep 14 23:32:35 but I can't leave the country Sep 14 23:32:35 mwester: because I wear steel toed boots, have a beard, and like german thrash bands? Sep 14 23:32:51 What do you call 200 lawyers at the bottom of Lake Michigan? A: A good start. Sep 14 23:32:52 bedboi: I cannot board airplanes within the US Sep 14 23:32:57 sure I could take a boat Sep 14 23:32:59 mwester: heh heh Sep 14 23:33:07 u_l-lap: are you kidding (take 2) Sep 14 23:33:09 but effectively I am trapped here Sep 14 23:33:14 bedboi: nope Sep 14 23:33:18 wtf Sep 14 23:33:22 what's this? Sep 14 23:33:24 u_l-lap: walk into Russia while there's still some ice left :P Sep 14 23:33:35 i guess it is against any international laws Sep 14 23:33:36 polz: the ice is melted, see news Sep 14 23:33:41 or maybe Norway, that seems saner Sep 14 23:33:49 the full story is that I travelled between baltimore and long island a bunch with one way tickets since I didn't know my return time and I was redflagged and then placed onto the no-fly list Sep 14 23:33:50 XorA: I think there's still some during the winter Sep 14 23:33:56 Mexico has no guards on their side of the border. Sep 14 23:34:07 it turns out that the plane that flew that route started in raleigh and ended in ... california Sep 14 23:34:13 u_l-lap: thats fscking insane Sep 14 23:34:23 lol Sep 14 23:34:26 so I took one of them TERRISTS planes Sep 14 23:34:32 AHAHA Sep 14 23:34:52 u_l-lap: yep uncle sam has to know when you are leaving and when you are returning Sep 14 23:34:53 anyway its 00:33 here, time for sleep Sep 14 23:35:08 u_l-lap: why don't they put gps chips in the brain? Sep 14 23:35:18 it would be easier Sep 14 23:35:33 u_l-lap: now if you WERE a terrorist, I'm sure you could get the ACLU to represent you, win some multi-billion dollar settlement, and live happily every after. Sep 14 23:35:34 i'm glad to leave this country in less than a month Sep 14 23:35:43 bedboi: because that wouldn't give them any feedback. They want GSM so they can prevent thought-crime Sep 14 23:35:51 LOL Sep 14 23:36:04 gsm won't work - heats the brain and makes you smarter Sep 14 23:36:09 bedboi: the gps driver software wouldn't be an allowed export Sep 14 23:36:18 damn Sep 14 23:36:25 btw it is a US gov stuff Sep 14 23:36:35 different department Sep 14 23:36:37 rwhitby: they should open your brain Sep 14 23:36:39 and check Sep 14 23:36:46 bedboi: the gps chips would be ... the number of the BEAST Sep 14 23:36:58 * mwester feels generally happy - don't go out of your way to piss off the cops, TSA, or customs officials, and in general life is no different here than anywhere else... Sep 14 23:37:00 ehehe Sep 14 23:37:03 Writchie: hmm.. zigbee, then - the name reminds me of hornets buzzing around inside my scull :P Sep 14 23:37:05 and then the evangelicals would got apeshit and everyone would realize that there are in actuality a total of three of them Sep 14 23:37:19 mwester: lol Sep 14 23:37:21 followed by bloody revolution or something Sep 14 23:37:31 in other words, no freedom Sep 14 23:37:43 mwester: thats what they want you to believe Sep 14 23:37:44 rfid: is being banned atm in California I think Sep 14 23:37:55 human rfid: Sep 14 23:37:58 mwester: sucks because everything is illegal nowadays so don't piss off cop = don't interact with a cop Sep 14 23:38:07 strange, I never noticed any humans in california myself Sep 14 23:38:19 LOL Sep 14 23:38:39 btw, seriously, this country has big problems. Sep 14 23:38:53 well italy has its problems too, but they are part of the folklore Sep 14 23:39:03 the fact that nobody wants to work, the mafia, all that stuff Sep 14 23:39:19 and that they all drive like crazy Sep 14 23:39:23 btw we do great pizza Sep 14 23:39:26 bedboi: you're telling me I haven't met someone who could think in years Sep 14 23:39:27 hhf423: yep. Sep 14 23:39:47 bedboi: bah. Even the hungarians make better pizzas :P Sep 14 23:39:51 NO! Sep 14 23:40:00 best pizza is in Chicago. Sep 14 23:40:00 pizza is italian, full stop. Sep 14 23:40:05 shut the fuck up Sep 14 23:40:05 new york Sep 14 23:40:10 no please Sep 14 23:40:16 aaargh Sep 14 23:40:24 bedboi: it might be an italian invention, but in the end, the other people perfected it. Sep 14 23:40:30 american pizza is not pizza Sep 14 23:40:32 finish reindeer pizza is better than italian pizza Sep 14 23:40:34 AAHAHHAHA Sep 14 23:40:43 this is the worst thing i ever heard Sep 14 23:40:51 and now you are telling you can cook better pasta Sep 14 23:41:03 The truth is sometimes painful... :p Sep 14 23:41:04 come on. Sep 14 23:41:11 I think the best pizza I ever had was in scottland Sep 14 23:41:20 hey let's all have a friendly emacs vs vi debate Sep 14 23:41:24 it'll probably have a lower casaulty count than best pizza Sep 14 23:41:32 ed Sep 14 23:41:34 hhf423: obviously never been in italy Sep 14 23:41:39 wordstar Sep 14 23:41:41 bedboi: I wouldn't know - don't like pasta... but that might have something to do with my mom's cooking :P Sep 14 23:41:43 mwester there is no universal truth. Sep 14 23:41:46 hhf423: otherwise you would have no taste at all Sep 14 23:41:48 u_l-lap: don't forget notepad.exe in your list. Sep 14 23:42:08 tedit Sep 14 23:42:10 and btw emacs sucks. Sep 14 23:42:15 Ooof. Sep 14 23:42:18 * bedboi hides Sep 14 23:42:22 mwester: now now only real editors are allowed in this contest Sep 14 23:42:36 * u_l-lap executes M-x murder RET bedboi RET Sep 14 23:42:39 LOL Sep 14 23:42:40 u_l-lap: OK, edit.exe, then Sep 14 23:42:42 bedboi: be warned the terminator is coming for you Sep 14 23:42:47 or was that .com ? Sep 14 23:42:49 u_l-lap: the baconator. Sep 14 23:42:57 edlin! Sep 14 23:43:01 the last wendy's invetion, it is dangerous Sep 14 23:43:06 Real programmers just use "cat". Sep 14 23:43:09 * bedboi sends a baconator to u_l-lap Sep 14 23:43:16 noooo my arteries Sep 14 23:43:20 mwester: copy con bla.txt Sep 14 23:43:22 u_l-lap: you are fucked up Sep 14 23:43:23 baconator? uiuuuuggghhhhH! Sep 14 23:43:44 hey guys i tried that, i said myself: "I have to give a try to this american stuff" Sep 14 23:43:46 mwester: but true hackers use a switch and an oscilloscope Sep 14 23:43:55 bedboi: what makes you say that Sep 14 23:44:02 u_l-lap: insanity Sep 14 23:44:10 mwester: and that's the lazy ones - the true l33t get by with three pieces of wire and a LED Sep 14 23:44:16 * mwester remembers entering the bootloader program into a Prime 400 minicomputer using 16 toggles and LEDS... Sep 14 23:44:16 u_l-lap: my gf came back to italy the day before the baconator. Sep 14 23:44:20 best editor is scissors and scotch tape for the paper tape Sep 14 23:44:34 Writchie: than ocr? Sep 14 23:44:39 it doesn't work on linux. Sep 14 23:44:47 bedboi: I was scared when I saw the signs for it Sep 14 23:44:52 yep. Sep 14 23:45:01 bedboi: I had no clue what it was since I don't watch tv ... then someone told me and I had a heart attack just hearing about it Sep 14 23:45:02 Writchie: That doesn't work; can't take scissors on an airplane. Sep 14 23:45:19 Btw i said myself: I want to be something like the governor of California Sep 14 23:45:28 (i can't write his name, nobody know how to do that) Sep 14 23:45:47 mwester: you can take that plastic scissors Sep 14 23:45:54 governator? Sep 14 23:45:58 yep Sep 14 23:46:12 the BlackNiggah ? Sep 14 23:46:34 AHAHAH Sep 14 23:46:48 yep Terminator Sep 14 23:47:11 Baconator has obvious references Sep 14 23:49:51 oh, is the vte dependency now fixed? Sep 14 23:50:01 OK, so does anyone know how the Neo switches to headphones when you plug them in ? Sep 14 23:50:16 hhf423: I think so - the bug was closed a few days ago Sep 14 23:50:26 cool Sep 14 23:50:28 polz: does it work? Sep 14 23:50:42 i did some patch for h5455 Sep 14 23:51:03 polz: you mean who tells alsa to load an other state? Sep 14 23:51:04 it was just that i had to handle an interrupt Sep 14 23:51:13 hhf423: yeah. Sep 14 23:51:15 polz there comes an event out of an /dev/evdevX (dunno which one right now) which gets then handled by neod which switches to the headphones via alsa Sep 14 23:51:32 roh: 10x. Sep 14 23:52:01 polz: neod I think Sep 14 23:52:05 I was wondering because IMHO it should also switch the input to the microphone on the headset Sep 14 23:53:38 oh, and the dialer shouldn't leave the left mixer left and right mixer right muted after finishing a conversation Sep 14 23:54:29 but that should be easy to fix now that I can finally build my own programs, wheeee :P Sep 14 23:56:31 roh: any idea on if the hold button bug on the handset is hardware or software? Sep 14 23:57:00 SpeedEvil hold-button bug? Sep 14 23:57:23 http://bugzilla.openmoko.org/cgi-bin/bugzilla/show_bug.cgi?id=695 Sep 14 23:57:35 the handset button reports no events Sep 14 23:57:43 headset Sep 14 23:58:49 hm.. need to check that.. Sep 14 23:59:18 say, are the mailinglists dead? Sep 14 23:59:23 yes Sep 14 23:59:59 Or at least to me. Sep 15 00:00:43 any bets on who uploaded that nice gps screenshot to scap? Sep 15 00:01:28 SpeedEvil: I thought that button just switched off the microphone Sep 15 00:01:59 It should - according to the kernel source at least - produce an event Sep 15 00:02:32 Wow, one of my relatives rented a DVD, and I just put it on my computer to see if DVD playback works on it... It starts with an *absolutely* *hilarious* "downloading is stealing" clip Sep 15 00:02:33 but how is it supposed to do that ? I mean... there are only four lines leading off the jack - ground, mic, left and right Sep 15 00:02:48 how could that button produce an event, then ? Sep 15 00:03:03 I assume it's connecting something across the mic Sep 15 00:03:26 polz: er... with a pin which is physically pushed back by the plug, thus touching/stopping to touch another line? Sep 15 00:03:38 polz: like every other speaker plug detection? Sep 15 00:03:47 cesarb: he's talking about the button on the headset Sep 15 00:03:55 polz: ah, sorry Sep 15 00:04:10 There is also an event on headset plug/unplug Sep 15 00:04:32 polz: I'd assume it shorts something to ground Sep 15 00:05:08 cesarb: I think it shorts the microphone... Sep 15 00:05:19 There are any number of solutions from a resistor alone, to one-wire busses. Sep 15 00:05:33 or even indeed shorting. Sep 15 00:05:34 but I'm not sure if the Neo can/does detect that Sep 15 00:05:47 According to the kernel source, it should. Sep 15 00:05:59 But it does not. Sep 15 00:08:33 the mic input seems supposed to be capacitively coupled, so shorting the mic input would indeed be the obvious solution Sep 15 00:08:48 SpeedEvil: where in the source? Sep 15 00:09:14 the mic can't be capacitively coupled - it needs to have some power. Sep 15 00:09:27 1mA at a volt or two is typical Sep 15 00:10:50 http://svn.openmoko.org/trunk/src/target/kernel/patches/gta01-inputdevice.patch Sep 15 00:10:54 search Hold Sep 15 00:11:27 SpeedEvil: correction, capacitively coupled to the codec Sep 15 00:11:40 yes, it will be. Sep 15 00:11:53 but there is also a resistor going to some power rail. Sep 15 00:12:12 probably 1-10K or so. Sep 15 00:12:59 The microphone is an electret microphone (almost certainly) which requires some power. the capacitor to the audio chip will go from the junction between the resistor and the mic Sep 15 00:13:03 SpeedEvil: hm, it should show up on /proc/interrupts... Sep 15 00:13:14 SpeedEvil: should be a IRQ-capable GPIO... Sep 15 00:13:29 hexdump -C /dev/input/event0 Sep 15 00:13:35 should reveal events Sep 15 00:13:36 counter Sep 15 00:13:37 http://londerings.novalis.org/wlog/index.php?title=Neo1973_counter Got it, get yours from openmoko.com (and read topic) (1809;252) Sep 15 00:13:47 it does for aux and jack insertion Sep 15 00:14:26 SpeedEvil: does it at least mute the microphone ? Sep 15 00:14:28 SpeedEvil: are you sure it's event0? IIRC there were three event sources... Sep 15 00:14:39 I mean... does your microphone even work ? Sep 15 00:14:48 EH??? Sep 15 00:14:51 cesarb: I tried them all Sep 15 00:14:51 what's this: We sold out our Neo 1973 now. In fact, if you place order now. The ETA will be almost October 2007. Sep 15 00:14:56 wtf is that? Sep 15 00:15:02 they told me sept 20 Sep 15 00:15:07 SpeedEvil: waaay ahead of me, as usual ;-) Sep 15 00:15:52 * polz had to open the jack and re-solder the wires because one of the earpieces and the microphone weren't working Sep 15 00:16:07 polz: ouch Sep 15 00:17:39 SpeedEvil: it works for me. Sep 15 00:17:50 polz: you get events? Sep 15 00:17:58 yep. Sep 15 00:18:06 On the headset button? Sep 15 00:18:38 yep. Sep 15 00:18:40 2007.2, or 2007.1? Sep 15 00:18:47 2007.2 Sep 15 00:18:59 ipkg upgraded a few minutes ago Sep 15 00:19:13 have you tested your microphone yet ? Sep 15 00:19:17 no. Sep 15 00:19:53 maybe you have the same problem I had Sep 15 00:20:25 hm, the button on the rubbery headset which comes with the devkit doesn't seem to increase the corresponding counter in /proc/interrupts Sep 15 00:20:35 some wires not really soldered on inside the jack, probably because they had to use a lead-free solder Sep 15 00:21:09 polz: did you report that to the neo hw guys? that kind of feedback can be really important Sep 15 00:21:28 polz: do you have a multimeter? Sep 15 00:21:59 If so, can you say which pins the hold button acts on? Sep 15 00:22:30 I'm pretty sure it's the tip, but I'll check Sep 15 00:23:53 yep, it shorts the tip to the ground Sep 15 00:25:00 * polz is going to sleep now Sep 15 00:25:08 SpeedEvil: urgh, take a look at http://pinouts.ru/connector/4_pin_3.5mm_2.5mm_plug_connector.shtml Sep 15 00:25:24 SpeedEvil: is the multimeter's voltage/current low enough to not damage the mic? Sep 15 00:25:46 * mwester doesn't see the neo listed there... Sep 15 00:26:02 Which one does it match most closely? Sep 15 00:26:09 SpeedEvil: there seems to be at least two different ones... Sep 15 00:26:26 SpeedEvil: one seems to have the mic at the tip, the other seems to have the mic at the middle ring Sep 15 00:27:11 * cesarb just found a third one... with the mic at the _other_ ring Sep 15 00:27:26 cesarb: yeah, it is. Sep 15 00:27:32 so, assuming that page is not incorrect, there are several incompatible interfaces using the exact same connector Sep 15 00:27:42 cesarb: if it isn't, I've just fried the mic in the name of science :P Sep 15 00:27:53 the only common thing being the ground farthest from the tip Sep 15 00:28:16 cesarb: and the mic is definetly connected to the tip - I got to verify that when I opened the thing Sep 15 00:28:48 * polz really has to go to sleep now Sep 15 00:29:04 SpeedEvil: good luck with the headset Sep 15 00:30:37 SpeedEvil: yeah, looks like the hold button does nothing on mine too... I have exactly 1 interrupt on the HOLD pin (probably from even before plugging it in), how about you? Sep 15 00:31:18 SpeedEvil: btw, just noticed a constantly increasing timer tick... is the kernel tickless? if not, it could bring some interesting levels of power saving (perhaps even invalidating part of your power levels research) Sep 15 00:34:09 hmm... so I'm not the only one with a faulty headset Sep 15 00:34:33 cesarb: unfortunately not - from the scope. Sep 15 00:34:41 cesarb: ticks aren't a big part of the power use. Sep 15 00:35:01 polz: well, i didn't test the mic, and I'm not in the mood today of digging into crazy alsa state files to find some way to route the mic to somewhere so I can check if it works Sep 15 00:35:07 cesarb: you can cut the connector along the seam, pry it open.. Sep 15 00:35:29 and then solder the wire that's farthest from the base Sep 15 00:35:35 SpeedEvil: but if the ticks happen often enough that the chip cannot get into a lower power setting... Sep 15 00:36:07 cesarb: Well - the problem is that the lower power states are not implemented now, so it's quite irrelevant :) Sep 15 00:36:47 and besides, it can't play music when it's in the lower states :) Sep 15 00:36:48 polz: well, if the problem is with the headset, finding a compatible headset (preferably a non-in-canal one) would be an alternative solution Sep 15 00:37:10 polz: in principle, it can play mp3 at 60Mhz or so. Sep 15 00:37:20 It certainly can play them at 100MHz Sep 15 00:37:26 cesarb: I guess you're less stingy and impatient than I Sep 15 00:37:27 I haven't tried at 60Mhz Sep 15 00:37:40 annoyingly, the undervolting possibilities are limited. Sep 15 00:37:51 as the power managment chip only does .3V increments Sep 15 00:38:10 It _almost_ works at 1.5V Sep 15 00:38:11 SpeedEvil: are your efforts documented anywhere ? Sep 15 00:38:16 Not really. Sep 15 00:38:20 I should do that. Sep 15 00:38:33 I need to recheck this, with pictures. Sep 15 00:38:38 (of scope) Sep 15 00:38:56 polz: there's no way I can be stingy if I paid double for the neo due to the crazy taxes (which I knew of before ordering) Sep 15 00:39:37 cesarb: hehe, I waited for 1 additional month for a friend to bring it from the USA personally Sep 15 00:39:49 polz: well, then you sure are stingy ;-) Sep 15 00:40:23 yep. Only bought the Neo after I checked it was cheaper than the comparable phones they sell here Sep 15 00:40:57 polz: these "comparable phones" probably can dial without having to hack /etc/init.d/gsmd ;-) Sep 15 00:41:21 cesarb: the Neo can, too... most of the time.. kinda Sep 15 00:41:33 polz: yeah, right... Sep 15 00:41:39 That makes it more fun, though... :p Sep 15 00:41:47 mwester: for sure! :D Sep 15 00:41:52 cesarb: but some of them don't have GPS... which also doesn't work properly ;) Sep 15 00:41:53 Kinda like rolling the dice with each call. Sep 15 00:42:26 the best "feature" is that the dialer sometimes "forgets" to ring on an incoming call Sep 15 00:42:47 BTW, can't we get the official pinout for the headphone/mic connector? Sep 15 00:43:15 Maybe roh would be kind enough to get that from fic and put it on the wiki. Beats cutting one open... Sep 15 00:43:24 mwester: I'm 100% sure the tip is connected to the mike Sep 15 00:43:31 and the base is ground Sep 15 00:44:08 as for the left and right - just insert it slowly while playing some music to find out which is left and which is right Sep 15 00:44:14 mwester: well, we got the official pinout for the debug connector. next to that, the official pinout for a puny analog mic-and-speaker plug would be nothing. Sep 15 00:45:43 btw, which earpiece is supposed to be L and which R ? Sep 15 00:46:18 I mean.. they could be worn either way, and I think there's a switch to invert the stereo Sep 15 00:51:02 polz: I'm actually wondering not so much about the mic or headphones as the button -- might it be that we get no events because the headset button wiring is incorrect wrt to the connector? Sep 15 00:56:29 mwester: but didn't polz get the events after fixing the conector? Sep 15 00:58:22 I don't think he explicitly said that he got events; just that the mic began working, and that the button muted the mic (which would be reasonable, as it is wired to short the mic) Sep 15 00:59:20 It seems strange that someone would use the speaker lines for an input, but there are some strange wiring diagrams on that connector wiring page that was posted here earlier... Sep 15 01:07:11 mwester: gsm firmware upgrade path? Sep 15 01:12:21 are the lists dead? Sep 15 01:12:31 we believe so Sep 15 01:13:37 bummer Sep 15 01:30:20 hmm, is the tasks icon fixed for you with the newest build? Sep 15 01:30:44 SpeedEvil: it works for me. Sep 15 01:30:44 polz: you get events? Sep 15 01:30:44 yep. Sep 15 01:30:58 ->mwester Sep 15 01:31:17 notserpe: yes, lists dead Sep 15 01:31:28 notserpe: dead as gsmd most of the time Sep 15 01:31:34 like sco Sep 15 01:31:49 sco dead? chapter 11 only, no? Sep 15 01:32:01 same diff, no? Sep 15 01:32:05 Not quite. Sep 15 01:32:13 If the company is fundamentally sound Sep 15 01:32:24 yeah, I just read about the trial being off and them getting to spend money while novell still doesn't get any Sep 15 01:32:26 but having severe cashflow problems, chap11 can be a breathing space. Sep 15 01:32:36 But, ... Sep 15 01:33:06 anyway IANAL... but I laughed when I read linuxdevice's " stick a fork" in them. Sep 15 01:33:25 aren't a number of US airlines operating under ch 11 all the time? Sep 15 01:41:29 off for bed now **** ENDING LOGGING AT Sat Sep 15 02:59:57 2007