**** BEGIN LOGGING AT Sun Sep 23 02:59:56 2007 Sep 23 03:07:43 there seems to be a prebuilt chroot for gllin.. Sep 23 03:38:19 and it seems to work. waiting 15 minutes to get a signal... Sep 23 04:06:09 i don't see any satellites.. but i'm getting a nmea stream. Sep 23 04:07:38 oh. there it goes. Sep 23 04:52:04 re Sep 23 04:53:39 re Sep 23 04:53:51 i've just reinstalled openmoko Sep 23 04:54:05 and ipkg update && ipkg ugrade Sep 23 04:54:24 it seems that i can't run applications for some reason Sep 23 04:54:59 http://ipkg.nslu2-linux.org/feeds/openmoko/images/?C=M;O=D <-- i used this Sep 23 04:55:17 ops.. http://ipkg.nslu2-linux.org/feeds/openmoko/images/openmoko-devel-image-fic-gta01.jffs2 Sep 23 05:03:48 matchbox-window 94% cpu Sep 23 05:08:15 Starting system message bus: Unknown group "netdev" in message bus configuration file Sep 23 05:08:15 dbus. Sep 23 05:13:07 bedboi: been like that for weeks Sep 23 05:13:22 avahi Sep 23 05:15:10 ah and it crashes Sep 23 05:15:21 i don't know which is the "desktop" application Sep 23 05:15:28 it crashes if i try to run the dialer Sep 23 05:16:46 juri_: The media player does work - what problems did you face? And no, can't delete playlist items just yet, there are some audio system problems that are more urgent than that atm Sep 23 05:17:24 oh man this image is so messed up Sep 23 05:19:23 feedreader, mediaplayer, dates...they all segfault Sep 23 05:19:51 What's the date on the image, do you know? Sep 23 05:20:02 bedboi: Yeah it is and libmokoui is the culprit Sep 23 05:20:08 There was a library problem not too long ago Sep 23 05:20:11 mwester: it's today Sep 23 05:20:36 I'm working with an image from 12 hours ago, and all is well. Sep 23 05:20:46 mwester: url? Sep 23 05:20:46 (at least, no worse than usual) Sep 23 05:20:55 I built it myself :) Sep 23 05:21:11 I can probably put it up for a download though. Sep 23 05:24:01 mwester: thansk Sep 23 05:24:03 thanks Sep 23 05:25:20 It'll take a bit; my link is slow tonight. Either weather-related (it's a wireless broadband link), or everyone on this tower is downloading movies or porn or something... :( Sep 23 05:26:54 ehhe Sep 23 05:27:06 i guess porn Sep 23 05:29:43 abraxa: hitting the play button does nothing. and no song length shows on the length meter. Sep 23 05:30:04 right now, i'm trying to get GPS output to my nokia n770. Sep 23 05:30:16 i've got them communicating via bluetooth... Sep 23 05:30:40 bu its nokia->neo only. neo->nokia appears to not get there... Sep 23 06:17:10 (script) wiki RecentChanges: [[P1_Owners]] [[OpenMoko_Local_Groups:_North_Texas]] Sep 23 06:18:11 bedboi: http://www.mwester.net/moko Sep 23 06:18:37 Make sure you check the md5 sums; I had to restart the transfer a few times, so there's a chance it got damaged. Sep 23 06:19:37 why do i need the gsmd? Sep 23 06:19:50 Note also that this is a patched kernel - use the supplied /etc/init.d/gsmd script and gsmd should start reliably (the dialer will complain about modem power, and the little status bar will often vanish, but those are known issues) Sep 23 06:20:47 You don't have to use the gsmd, the standard hacked one should work, but it often doesn't and has to be restarted in the /etc/matchbox/session file. Sep 23 06:26:27 mwester: thanks i have all the files Sep 23 06:32:06 bedboi: md5sums ok? Sep 23 06:39:08 yep Sep 23 06:39:12 reflashing Sep 23 06:39:26 i'll say in two mins Sep 23 06:41:35 booting Sep 23 06:43:13 damn that boot sound is so scary Sep 23 06:43:38 hehe - edit /etc/pulse/session (or something) and remove the line containing "startup" Sep 23 06:43:45 damn is so bad Sep 23 06:43:59 btw it's the same Sep 23 06:44:00 it dies Sep 23 06:44:12 I also rename /etc/init.d/psplash to no-psplash or something, so I can see what's going on. Sep 23 06:44:35 maybe i should use your gsmd Sep 23 06:44:38 And I cannot figure out what use the avahi daemon is on the device, and I can't stand waiting for it to timeout, so I rename that startup file too Sep 23 06:46:11 Oh - and I edit /etc/fstab -- comment out the entries for /tmp and /media/ram -- don't need them, and the first one causes extraneous error messages at startup Sep 23 06:47:48 I didn't say everything works, but I did make a test phone call with the dialer this afternoon, so at least that much works. :) Sep 23 06:48:03 it crashes Sep 23 06:48:10 let's see with your gsmd script Sep 23 06:48:42 and i get the "openmoko" splashscreen Sep 23 06:48:53 the one that says "please wait ..." Sep 23 06:49:50 The one with the progress bar, or the one that tells you to wait for the gui? Sep 23 06:50:33 no wait for the gui Sep 23 06:51:01 i'm trying to open the dialer Sep 23 06:51:09 but nothing happens Sep 23 06:51:26 Did you get a dialog box complaining that it couldn't "set antenna power"? Sep 23 06:51:33 no Sep 23 06:51:48 i've power it up Sep 23 06:51:53 Ok, very different from what I'm experiencing. Sep 23 06:52:13 ok now it's again dead Sep 23 06:52:25 but this time the panel is stil there Sep 23 06:52:33 i'd say it's QUITE unstable Sep 23 06:52:52 The panel problem I get as well. Sep 23 06:53:01 no the panel is there:) Sep 23 06:53:01 Yes, it's *very* unstable. Sep 23 06:53:03 the desktop Sep 23 06:53:11 the desktop has gone Sep 23 06:53:22 Nope, not seen that then. Sep 23 06:54:06 there must be some curse then Sep 23 06:54:16 I wonder if I didn't flash the image I thought I did, then... :( Sep 23 06:54:39 mwester: i'm not so good at voodoo. Sep 23 06:54:39 btw Sep 23 06:55:00 i would like to give gtk+ direct framebuffer a try Sep 23 06:55:18 i was trying to compile it but it needs some patches Sep 23 06:55:27 that are not in files/ Sep 23 07:08:40 actually no apps in the application menu start Sep 23 07:08:40 Ok, something is broken badly in the latest builds. I'll find that last working image and get that uploaded, but it'll have to wait until morning (it's 2AM here, gotta get some sleep sometime). Good luck. Sep 23 07:08:59 Yeah, they all complain about icons or something. Sep 23 07:09:07 hmm Sep 23 07:09:10 it could be Sep 23 07:10:39 You can try to find a copy of Sep 23 07:10:39 libmokoui2-0_0.1.0+svnr2868-r2_armv4t.ipk in the feeds, and force ipkg to downgrade that library (I think that's the culprit) Sep 23 07:10:51 ok Sep 23 07:12:06 If you can't bludgeon ipkg into behaving, just use ar to extract the contents, then manually put the contents of the data.tar.gz into place... can't break it any worse. Sep 23 07:23:02 http://buildhost.automated.it/OM2007.2/packages/armv4t/libmokoui2-0_0.1.0+svnr2868-r2_armv4t.ipk Sep 23 07:24:46 ScaredyCat: how to take a screenshot? Sep 23 07:24:51 cat /dev/fb0 ? Sep 23 07:26:19 ipkg install ./libmokoui2-0_0.1.0+svnr2868-r2_armv4t.ipk -force-downgrade ? Sep 23 07:27:33 is the screenshot app installed? Sep 23 07:27:52 if not cat /dev/fb0 >myscreenshot.raw Sep 23 07:27:58 then use fb2png Sep 23 07:28:03 yep Sep 23 07:28:04 ok Sep 23 07:28:18 or fbgrab ? Sep 23 07:28:19 what's the name of the screenshot app? Sep 23 07:28:21 is that there? Sep 23 07:28:30 fbgrab is there Sep 23 07:28:35 use that Sep 23 07:28:41 iirc it converts it too Sep 23 07:28:47 yep Sep 23 07:35:12 ScaredyCat: i installed it, let's see what happens Sep 23 07:38:22 ok it's enough for me. it can place calls Sep 23 08:06:54 :) Sep 23 08:07:38 i'll report success here too. i placed a call earlier. there was a ugly hum, and massive lag.. but it worked. Sep 23 08:08:32 hum like GSM buzz? or some other kind of hum? Sep 23 08:08:52 something that sounded like.. arround 120hz. Sep 23 08:09:16 how can i use openmoko-keyboard instead of the matchbox one? Sep 23 08:12:41 hi, could somebody help me setting up the openmoko environment, I've got a compilation error in the uboot package... Sep 23 08:13:31 kirikiri: Mind posting your error to a pastebin? Sep 23 08:13:40 ok Sep 23 08:15:26 http://pastebin.com/m4fddd53d Sep 23 08:15:37 * CM has quite a hangover Sep 23 08:20:24 I've followed the instructions on http://wiki.openmoko.org/wiki/Getting_OpenMoko_working_on_host_with_Xephyr Sep 23 08:24:42 kirikiri: Can I see the contents of your build/conf/local.conf, please? Sep 23 08:24:45 is there any good reason for defaulting to "No pim" Sep 23 08:25:32 woohoo! payment received! Sep 23 08:26:23 hey guys any thoughts on this effecting the neo1973, or maybe the next version. i must say it's a brilliant step forward: http://news.zdnet.co.uk/communications/0,1000000085,39289524,00.htm Sep 23 08:26:25 heh Sep 23 08:28:48 abraxa: yes, http://pastebin.com/d26aa3928 Sep 23 08:30:43 * ScaredyCat sighs Sep 23 08:31:08 " Sep 23 08:31:08 Preferable this webshop should not be written in PHP. Either Perl, Sep 23 08:31:08 Python or Ruby would be fine by us. Sep 23 08:31:10 " Sep 23 08:33:09 Man, this power-loss thing is getting ridiculous Sep 23 08:33:33 Again the nobreak stops working a few seconds before the power comes back Sep 23 08:33:51 5-minute-long power losses on a 5-minute-battery nobreak doesn't work so well... Sep 23 08:37:08 kirikiri: Hmm. For some reason it's not using the cross-compiler Sep 23 08:38:27 yes, I changed my local.conf to compile for PC (first I was compiling for qemu) Sep 23 08:39:34 kirikiri: That might get OE confused unless you clean out the build/tmp directory Sep 23 08:39:58 I've deleted the complete build first Sep 23 08:40:18 everything below ~/openmoko Sep 23 08:40:30 Ah, hm. Unfortunately I'm out of ideas then atm, did you check the bugzilla? Sep 23 08:40:52 not yet, I'll give it a look Sep 23 08:44:00 kirikiri: btw, you might want to add ENABLE_BINARY_LOCALE_GENERATION = "0" to your local.conf to avoid tedious locale building for gcc Sep 23 08:46:11 ok I'll add that, nothing interesting found for uboot on bugzilla... Sep 23 08:45:43 Bummer. Sep 23 08:53:11 I don't know the building procedure in openmoko, but I guess 'i486-angstrom-linux-gcc' builds uboot. How can I find out to which executable it points, so I could try to compile hello_world.c manually to hunt down the problem... Sep 23 08:56:04 kirikiri: Have a look at build/tmp/work/arm.../u-boot-openmoko.../temp/run.* Sep 23 08:58:29 kirikiri: Come to think of it... you want to run the apps natively on your PC or what do you try to do? Sep 23 08:59:11 Heya Zecke :) Sep 23 08:59:24 abraxa: yes, natively on PC with Xephyr Sep 23 09:00:25 kirikiri: Ah, hm Sep 23 09:01:33 kirikiri: In that case you might want to just build the OM libs/apps by hand but then again I don't know how much experience you have... Sep 23 09:01:34 no device available here and qemu is way too slow ;) Sep 23 09:04:23 ok, I thought it wouldn't be too difficult (at least the wiki mentioned just some changes in local.conf to correctly build the environment) Sep 23 09:04:12 Unfortunately, succeeding an OE build is like trying to walk on egg shells - you can succeed but more often than not, things will break Sep 23 09:06:43 :), yeah I've seen that compiling for qemu, 1 of 10 attempts succeeded. Now manually changing something in the build system seems a bit of a overhead for me, want just try out some apps... Sep 23 09:24:02 abraxa: thanks for helping... Sep 23 09:24:30 kirikiri: Wish I could've done more ;p Sep 23 10:10:48 what about the libmokoui2 issue? does somebody have a fix? r3032 is still broken, it seems. Sep 23 10:12:22 pH5: Nope, and I assume that it won't be fixed till monday morning UTC (which is when the guys at openedhand start working) Sep 23 10:14:44 (script) wiki RecentChanges: [[Main_Page/it]] Sep 23 10:14:49 abraxa_: thanks, that's a pity. Sep 23 10:15:04 Unfortunately, yeah Sep 23 10:15:47 ...it really does make me cherish the core team :) Sep 23 10:26:17 hi Sep 23 10:28:15 cu in two weeks Sep 23 10:28:16 take care Sep 23 10:28:54 mickey|portugal: have fun :) Sep 23 10:29:17 * pH5 waves goodbye Sep 23 10:30:01 mickey|portugal: have a good time ;) Sep 23 10:30:32 * ScaredyCat hides in mickey|portugal's suitcase Sep 23 10:31:16 oops my neo is again deeply discharged .. Sep 23 10:31:37 abraxa_: how many 'guys' is 'guys at openedhand start working' Sep 23 10:31:38 is there a way to prevent this ? Sep 23 10:31:58 after you power it down, take the battery out for a minute or two Sep 23 10:32:18 assuming you are actually powering it down Sep 23 10:33:22 ScaredyCat, is see .. it should revive ins some hours right ? Sep 23 10:35:25 oh yeah... Sep 23 10:36:05 30 mins - depends on if your port can deliver 500ma ... Sep 23 11:37:54 can the feedreader cache a website? That would be a good idea to read the feeds offline => before you go to work cache all linked websites Sep 23 11:38:05 in the rss feed Sep 23 11:38:32 http://wiki.openmoko.org/wiki/Server:Proxy would be nice too Sep 23 11:39:34 with wlan on the gta02 is the connection to the interet not the problem i think :) Sep 23 11:40:07 SpeedEvil: i look :) Sep 23 11:40:08 I should expand that as to 'cache on demand' section Sep 23 11:41:14 SpeedEvil: webproxy is a great idea Sep 23 11:41:22 Yeah. Sep 23 11:41:32 Tehre were a number of projects to do this. Sep 23 11:41:39 But they all died when dialup died. Sep 23 11:41:58 SpeedEvil: on the go is this a super idea :) Sep 23 11:42:24 SpeedEvil: with wlan at home we need a cache to read offline :) Sep 23 11:42:36 The underlying software I was thinking of using - wwwoffle - www offline explorer - also supports grabbing entire websites Sep 23 11:42:55 and grabbing pages that you click on that are not cached next time you get to a network Sep 23 11:43:09 OperaMini is using a proxy Sep 23 11:43:23 http://www.gedanken.demon.co.uk/wwwoffle/ Sep 23 11:43:36 wwwoffle + feedreader would be a good combination Sep 23 11:43:43 But adding the differential cache option would _drastically_ reduce bandwidth Sep 23 11:44:02 I've used wwwoffle for a while Sep 23 11:44:13 Another usefull app is to do all your browsing through it Sep 23 11:44:13 who writes the feedreader? Sep 23 11:44:23 perhaps we can collect a few ideas Sep 23 11:44:27 then you can run a search engine onto it Sep 23 11:44:33 PBeck: It's zecke Sep 23 11:44:43 nice ok :) Sep 23 11:44:44 searching your cache tends to be much more productive for stuff. Sep 23 11:44:46 thanks Sep 23 11:44:51 if you know you've seen it before Sep 23 11:45:06 SpeedEvil: yes Sep 23 11:46:16 i've wanted that feature so many time Sep 23 11:46:18 *times Sep 23 11:47:00 cjb_ie: :) Sep 23 11:47:16 feedreader + browser with a good cache Sep 23 11:47:28 for the proxy solution we need a extra server Sep 23 11:47:48 anybody musst host it self? Sep 23 11:48:10 would be ok for me :) Sep 23 11:48:36 i havent a server :> Sep 23 11:48:41 maybe this could be an optional (commercial) service from OpenMoko Inc. Sep 23 11:49:00 yes that will be a option Sep 23 11:49:25 but with a good cache in browser + feedreader and wlan at home is the problem not so big for me :) Sep 23 11:49:29 It would need a modest amount of CPU, storage, per user. Sep 23 11:49:43 50-500M or so. Sep 23 11:49:59 SpeedEvil: server oder cache in the mobilephone? Sep 23 11:50:04 a opera-like service would be really cool Sep 23 11:50:15 The cache needs to be the same size on each end Sep 23 11:50:50 that's the point. If you reload ebay.com, it sends the differences between the one on the server, and the one on the local proxy Sep 23 11:50:52 SpeedEvil: ? Sep 23 11:51:09 you have two identical caches Sep 23 11:51:24 Neo Proxy Sep 23 11:51:42 rsync is very inefficient, compared to this. Sep 23 11:52:11 SpeedEvil: ah yes i understand Sep 23 11:52:15 finding differences between two files over the network is much harder than transmitting updates to known identical files Sep 23 11:52:16 only an example Sep 23 11:52:32 there was a rsync-http protocol idea. Sep 23 11:52:39 which would add it into http Sep 23 11:52:48 erm, gentoo does that already afaik Sep 23 11:53:01 it's called web-rsync Sep 23 11:53:11 but that basically has problems. Sep 23 11:53:23 no - the idea was that every website would support it Sep 23 11:53:36 ah Sep 23 11:53:39 SpeedEvil: i ask zecke for the cache integration in the feedreader ;) Sep 23 11:53:41 ok, that makes it a bit harder :) Sep 23 11:54:05 then you connect to a website, say you can do rsync, and the website and you work out the changes in the page between its version, and your local stored version Sep 23 11:54:28 PBeck: it's pretty much a transparent http proxy Sep 23 11:54:31 wwwoffle Sep 23 11:55:23 you just tell whatever web-thingy that you're using that as a proxy, and the (cached) bit of the web just works offline Sep 23 11:55:47 wait wwwoffle also does this differential download bit? Sep 23 11:56:02 no Sep 23 11:56:04 or was that just something we'd need added Sep 23 11:56:05 right Sep 23 11:56:07 that was a patch Sep 23 11:56:12 (to be written) Sep 23 11:56:17 SpeedEvil: that means we need no direct integration in the application Sep 23 11:56:44 PBeck: integration is nice, but not required Sep 23 11:56:59 SpeedEvil: all internet traffic goes throuh wwwoffle Sep 23 11:58:09 why we need then a integration? Sep 23 11:58:35 It's nice to be able to trigger downloads in a clean way Sep 23 11:58:59 otherwise when you fall off the edge of the cache, you get 'this page has not been downloaded, click here to request' Sep 23 11:59:01 hum a "clean way" Sep 23 11:59:22 hum yes Sep 23 11:59:23 And index pages of downloads Sep 23 11:59:38 but for this problem we need a button in the feedreader to cache alle websites Sep 23 11:59:46 *all Sep 23 12:00:00 SpeedEvil, hmm. The default should rather be that it's downloaded transparently on the proxy end, IMO Sep 23 12:00:09 this is in 'offline' mode Sep 23 12:00:27 ah, so it is a mode question then Sep 23 12:00:31 in online mode, it just downloads and caches Sep 23 12:00:43 hm, linux-openmoko builds a moko11-r2 and linux-openmoko-devel a moko11-r1 image? Sep 23 12:00:52 * mjr may be a bit slow, going through withdrawal cause ran accidentally out of meds Sep 23 12:01:29 (corrected already, but effect takes a while) Sep 23 12:03:40 SpeedEvil: the following scenario => i am at home with wlan and connect through the feedreader to the internet and download all websites to the local cache (automaticly, perhaps with a simple button). Now i am on the go and can offline read all feeds Sep 23 12:04:41 There is - on the stock wwwoffle - a 'download from this page n deep' Sep 23 12:05:54 wwwoffle certainly seems like good stuff to start with Sep 23 12:06:38 SpeedEvil: yes, but with a normal rss feed the first site would be enough for all news Sep 23 12:09:42 http://zecke.blogspot.com/2007/05/rss-reader-updates.html <= SpeedEvil Sep 23 12:10:22 SpeedEvil: i am not the first with this idea ;) Sep 23 12:15:04 (script) wiki RecentChanges: [[FIC/de]] [[Booting_from_SD]] Sep 23 13:00:16 hrm, anybody know how mature the OpenMokoFramework is? Sep 23 13:07:56 rtyler: openmoko is in a early development status Sep 23 13:08:15 rtyler: not for a productive system Sep 23 13:08:34 rtyler: but here can you find good informationen to the functions they work mom Sep 23 13:09:10 http://wiki.openmoko.org/wiki/Phase_1_Software_Testing Sep 23 13:09:16 rtyler: look at the end of the page Sep 23 13:11:34 do I have to set the bootable flag for the partition on my SD I want to boot from? Sep 23 13:11:38 PBeck: http://flickr.com/photos/agentdero/1053119243/in/set-72157600934925721/ Sep 23 13:11:42 that's me, trust me, I know ;) Sep 23 13:11:52 ~seen zdanek Sep 23 13:11:53 zdanek was last seen on IRC in channel #openmoko, 1d 15h 10m 16s ago, saying: 'cu'. Sep 23 13:11:57 thomasg - probably not. Sep 23 13:12:09 Elrond: unfortunately it wont boot :( Sep 23 13:12:13 rtyler: ? Sep 23 13:12:36 I have 2 partitions, on the first (ext2) is the system including the kernel with ext2 support built in (fix) Sep 23 13:12:43 thomasg - You use the AUX button's menu for sd-booting? Sep 23 13:12:47 yes Sep 23 13:12:54 wheres a turing test when you need it, I can't tell if PBeck is a bot or not :P Sep 23 13:13:09 rtyler: why you have ask then for the mature of the framework? Sep 23 13:13:10 thomasg - Get on to u-boot console and boot it up by hand. :-) Sep 23 13:13:24 I didn't ask *if*, I asked *how much* Sep 23 13:14:03 rtyler: and the wikipage not answer your question? Sep 23 13:14:07 Elrond: in the u-boot console I do a mmcinit first Sep 23 13:14:16 it detectes the sd card, but I get an error Sep 23 13:14:26 mmc_block_read: err SDIDSTA=0x00000020 Sep 23 13:14:36 thomasg - Hmm, so maybe u-boot doesn't support your sd? Sep 23 13:14:42 * rtyler sighs Sep 23 13:14:43 Elrond: might be :( Sep 23 13:14:45 forget it >_> Sep 23 13:15:10 thomasg - I assume the supplied sd works with u-boot? Sep 23 13:15:12 would it be possible to flash the kernel only and let the rootfs on the partition? Sep 23 13:15:20 Elrond: didn't test it :) Sep 23 13:15:27 (shame on me) Sep 23 13:15:55 thomasg - Yes, putting kernel in flash, giving it the needed bootparam <-- should work. Sep 23 13:16:06 Assuming the kernel can talk to the sd. Sep 23 13:16:12 it should Sep 23 13:16:37 rtyler: hum the framework .. Sep 23 13:17:16 rtyler: look at the source :P Sep 23 13:19:29 thomasg - (assuming you have a debugboard and know to use it) Of couse you upgraded u-boot? ;o) Sep 23 13:19:51 I have a upgraded u-boot Sep 23 13:19:57 even if I don't have a debugboard :) Sep 23 13:19:59 PBeck: been there done that, looks like just barely libmokogsm and libmokoui are there :( Sep 23 13:20:19 its uboot-1.2.0-moko9-r0 Sep 23 13:21:03 but I guess I'm the first one with a 6gb microSDHC, so this probably won't work. in the wiki there are no informations about u-boot and SDHCs at all Sep 23 13:21:28 rtyler: http://wiki.openmoko.org/wiki/OpenMokoFramework perhaps that help you (when you not know this site already) Sep 23 13:21:49 hum last change 30 July ... Sep 23 13:21:55 ok we can forget this site ... Sep 23 13:22:10 thomasg - Ahh. You might ask mickey, if it's worth to file a u-boot bug on this. Sep 23 13:22:42 Elrond: k, I'll do that Sep 23 13:22:57 * thomasg asks mickey|portugal :) Sep 23 13:27:25 PBeck: I'm well familiar with the wiki -_- Sep 23 13:27:30 and yes, that page is never updated Sep 23 13:27:56 rtyler: super, than i can shut up ... Sep 23 13:28:45 :) Sep 23 14:01:48 how long does it take to compile the whole build using the mokomakefile on a 2ghz pc? Sep 23 14:03:01 IfChinsCouldKill: 3+ hours Sep 23 14:03:31 thanks! Sep 23 14:04:19 IfChinsCouldKill: 2GHz pc gives rather fuzzy idea of the speed :) one core of 1.6GHz core2duo is about as fast as 3GHz Pentium4 :) Sep 23 14:05:55 True. I started the makefile a couple hours ago an a P4 2Ghz notebook... you should hear the vent noise... hehe Sep 23 14:06:10 I guess I'll go for a walk or something... Sep 23 14:06:23 somebody here tested booting from a microSDHC? Sep 23 14:08:30 P4 in a notebook... I guess the running time on battery is not too hot :) Sep 23 14:09:38 I just started a parallel build on core2quad with 8 threads and i'm timing it ;) I'q curious to see how the build time compares to the sequential one ;) Sep 23 14:09:41 IfChinsCouldKill: be sure to also factor in the download time Sep 23 14:14:17 thomasg: I'm trying to boot from SD right now. No luck so far. Sep 23 14:15:35 sagacis: same for me :( Sep 23 14:15:43 (script) wiki RecentChanges: [[Wishlist:Unlikely]] [[Buying_Interest_List]] [[Template:CURRENTMMDDhhmmYYYY]] [[Template:CURRENTDATE]] Sep 23 14:18:06 thomasg: I"m getting a kernel panic on boot. I've tried to use ext3 and ext2 on the sd card to no avail. I don't even know what's killing the init. Sep 23 14:19:25 I can't barely see the error, as the screen clears so fast :( but it is kind of "bad block" Sep 23 14:47:44 Elrond: btw - u-boot says "MMC found." but the descrition seems to be completely initialized with 0, so there's no information about the card. seems it simply isn't detected by uboot Sep 23 14:49:05 maybe I'll flash a new uboot version to try again - then I'll do a bug report Sep 23 15:25:43 * ScaredyCat is being spammed to death :( Sep 23 15:32:21 PH5: Ping Sep 23 15:33:55 mwester: 64 bytes from pH5: icmp_seq=1 ttl=248 time=6150.5 ms Sep 23 15:34:02 hehe Sep 23 15:34:59 We can revert a change to see if we can get the build working again, but I'd like for someone besides myself to check that the change actually fixes the most recent problem... Sep 23 15:35:47 Are you in a position to be able to do such a test? Sep 23 15:36:26 mwester: what problem, what change? Sep 23 15:37:46 " what about the libmokoui2 issue? does somebody have a fix? r3032 is still broken, it seems." Sep 23 15:38:04 mwester: ok, sure. Sep 23 15:38:58 Ok, I'll try to revert locally and post the change here :) Sep 23 15:54:27 moin Sep 23 15:57:12 pH5: When you get a chance, you can try this: Sep 23 15:57:29 rm -rf build/tmp/work/armv4t-angstrom-linux-gnueabi/libmokoui2-0.1.0+svnr* Sep 23 15:57:37 rm -f build/tmp/stamps/armv4t-angstrom-linux-gnueabi/libmokoui2-0.1.0+svnr* Sep 23 15:57:50 vi openembedded/conf/distro/include/sane-srcrevs.inc Sep 23 15:58:37 locate the line for libmokoui2, and change it to SRCREV_pn-libmokoui2 ?= "2868" Sep 23 15:59:11 heh, yes that works Sep 23 15:59:16 Then build the image again; this time you should see it build the older svn version, and the image should function again (well, at least as well as it did yesterday!) Sep 23 15:59:19 (it's what I've done locally) Sep 23 16:00:16 just be sure to move the 3032 packages from the deploy directory Sep 23 16:00:40 It should end up in the "morgue" directory automatically, I thought? Sep 23 16:00:43 * mwester goes to check Sep 23 16:01:18 I thought the older versions always end up in the morgue. Sep 23 16:01:59 You are correct, the wrong one remains. I'll try bumping the PR and see if that fixes that problem. Sep 23 16:02:37 no that won't do it Sep 23 16:02:46 as 0.1-r9000 is still < 0.2-r0 Sep 23 16:03:23 Well that's wrong then - we should have the r# and svn# in the other order then. Sep 23 16:03:36 mwester: no. Sep 23 16:04:21 It seems wrong to bump the PV, though. Sep 23 16:04:40 what is wrong is that 3032 should have never went into sane-srcrevs without testing. other than that it works exactly like it's supposed to. Sep 23 16:05:02 the right way is to fix svn, which will automatically bump srcrev. Sep 23 16:05:08 I agree with that -- a process was put in place so that this wouldn't happen. Sep 23 16:05:43 But now we uncover another problem which is that it *DID* happen, and we have no easy sane way to revert in OE. Sep 23 16:06:32 I can roll the sane-srcrevs back, but apparently that doesn't really fix it, as the wrong ipk remains in the feed (although the image should be correct if the PR is bumped) Sep 23 16:06:55 I guess this will have to wait for someone with svn access to fix. Sep 23 16:07:10 mwester: if you want to fix it in oe, you have to apply a patch and bump PR on r3032 Sep 23 16:07:33 I'm back and now it's gerarating locale ca_ES ... ooof Sep 23 16:07:34 the mechanism for downgrading PV in OE are epochs (PE), but those should be used very carefully. Sep 23 16:08:10 generating even Sep 23 16:08:53 Ok. So it's not easy to fix this in OE. Let's find out who did it, then, and boil them in oil. :p Sep 23 16:10:46 you're too kind Sep 23 16:15:44 (script) wiki RecentChanges: [[Wishlist:Games]] [[Talk:Supported_microSD_cards]] [[Supported_microSD_cards]] Sep 23 16:20:39 hi @ all, I have two questions... how good is the audio quality with earplugs at the Neo 1973 and what do you guys say to this article http://www.roughlydrafted.com/RD/RDM.Tech.Q3.07/B10AE668-EAD3-46DC-A042-5EF3461D63EF.html Sep 23 16:21:02 IMO it's "okay", but I'm no hifist Sep 23 16:22:02 it makes a valid point that some tout openmoko as if it was finished, but is clueless on many points Sep 23 16:23:28 hi all Sep 23 16:25:00 mjr: ahh come on. I had a good laugh about the "openmoko is developed as a windows smartphone by chinese companies"... Sep 23 16:25:29 eheheh Sep 23 16:26:01 mjr: I don't need a cell phone to be as good as a hifi system but good enough to play some music with normal-good quality ;) Sep 23 16:26:10 don't you think that the font used in contact list is too small? Sep 23 16:26:44 it's difficult to select the right contact with the finger Sep 23 16:29:38 that article is so old now, and so inaccurate... Sep 23 16:29:59 have a look on the ml for responses to/frm the author Sep 23 16:35:57 * hhf423 yawns Sep 23 16:36:36 * CM too Sep 23 16:39:04 tin ting tiring Sep 23 16:48:56 hi all :) Sep 23 16:53:15 hi sunix Sep 23 16:54:12 unique: imho the sound quality with the included earplugs is not that good. I've heard much worse plugs (included with mp3 players and so on), but that's not what I'd call hifi. unfortunatately I do not have a 2.5->3.5mm adapter, so I can't test my good earphones Sep 23 16:56:04 (script) openmoko-community: Sean Moss-Pultz Re: Help Request for our Webshop Sep 23 16:56:41 thomasg: I don't mean the quality of the earplugs but the sound-quality in general Sep 23 16:57:36 good sound is in hardware Sep 23 16:59:56 the speakers are ok (in fact the included plugs are hardly better), but like I said - I had no chance to test the real quality because I don't own the needed adapter Sep 23 17:00:22 for a mobile phone it's quite good Sep 23 17:00:47 fyi, for those who wish 2.5 -> 3.5 adaptors, I think the ones commonly sold for the Treo will work. Sep 23 17:01:22 (except that the shipping costs from online retailers usually exceed the item's value :( ) Sep 23 17:02:00 thomasg: if the quality is ok ... it's ok for me :-) Sep 23 17:11:00 how long does it last to boot the neo 1973 with openmoko ? Sep 23 17:11:07 -last +take Sep 23 17:12:39 unique, IIRC about 1min 30sec or so Sep 23 17:12:46 there are bootcharts on the wiki Sep 23 17:13:15 cb22: thx Sep 23 17:13:30 cb22: does the neo use standard sysVinit? Sep 23 17:13:48 thomasg, no clue unfortunatly Sep 23 17:14:16 http://wiki.openmoko.org/wiki/Bootcharts Sep 23 17:14:24 * thomasg boots the neo :) Sep 23 17:14:29 thomasg, do you mean like /etc/init.d/S??something ? Sep 23 17:14:35 in that case, yes Sep 23 17:15:41 no, that's not what I mean Sep 23 17:16:01 e.g. gentoo uses a advanced sysVinit with parallel-bootup options and so on Sep 23 17:16:30 * cb22 scurries back behind freebsd's rcNG system Sep 23 17:29:09 unique: The boot is far from optimal - there are huge opportunities for speedup in almost every area. Sep 23 17:30:11 mwester: that's good to hear... I'm confident this will be done in future Sep 23 17:31:51 The current boot time is sort of an upper bound then... Sep 23 17:31:54 Yep - future work. We need to get it all working first. But there are a few really simple things we can do right now that would help a lot. e.g. there's no need to start avahi-daemon at all. Sep 23 17:32:47 Removing the time delays in /etc/init.d/gsmd will also get back a few free seconds, and the delays only make matters worse with gsmd, as it turns out. Sep 23 17:33:20 mwester: well, one of the advantages of the whole stack being open is that, if someone feels strongly about the boot time, they can fix it themselves Sep 23 17:33:31 mwester: not everybody can hack on gsmd ;-) Sep 23 17:33:36 We can also defer the mounting of the SD card (unless some flag indicates that a dependency has been introduced). Sep 23 17:35:43 will the neo be fast enough for scumm-vm or other games? does it have a 3d co-cpu ? Sep 23 17:35:53 No. Sep 23 17:35:56 this version Sep 23 17:36:01 yes, next Sep 23 17:36:33 people have been running scummvm on the current gta01 as well, though; I don't know how fast it is Sep 23 17:36:46 but scummvm shouldn't really require that much Sep 23 17:37:18 mjr: I mean the gta02v3 or how the mass market version is called Sep 23 17:38:05 It says faster cpu (Samsung 2442 SoC) but nothing specific Sep 23 17:38:13 read again Sep 23 17:38:27 there is a 3d accel chip Sep 23 17:39:05 Is the mass market version still on track for an October release? Sep 23 17:39:22 SpeedEvil: that is really nice Sep 23 17:39:30 No. Sep 23 17:39:34 December Sep 23 17:40:14 cant wait to see the final version of the Neo and Openmoko and read the first tests about it... I hope there will be a huge community to contribute with programs and stuff Sep 23 17:40:51 there will be 2 communities at least Sep 23 17:41:14 the qtopia and the openmoko/gtk crowd Sep 23 17:41:45 buz: there's a significant overlap between the two Sep 23 17:41:54 buz: the lower layer is common Sep 23 17:42:01 yeah, not saying it's entirely different Sep 23 17:42:17 sort of like how kubuntu and ubuntu works Sep 23 17:42:30 buz: (in fact, it was the qtopia guys who made it that way... they only changed the UI and phone layers, and kept the base framework) Sep 23 17:42:42 trolltech is not stupid Sep 23 17:42:55 buz: well, it's a bit more separate, since AFAIK qtopia uses the framebuffer, while OM uses X Sep 23 17:42:56 they know it saves em work AND gets em street cred Sep 23 17:43:02 yeah thats true Sep 23 17:43:36 making a X server for qtopia would help bridge both sites quite nicely Sep 23 17:43:48 I thought there was one already? Sep 23 17:43:51 such a beast exists i think Sep 23 17:44:05 and there's always qvfb which does the other way Sep 23 17:45:21 To buy a openmoko now or to buy it later Sep 23 17:45:31 Hum, its more fun if it got alot of broken stuff. Sep 23 17:45:41 qvfb, frankly, seems the saner solution, but it's good there's interop options Sep 23 17:46:14 it would depend on trolltech getting accellerated drawing layer as well i guess Sep 23 17:46:15 mjr: depends on which you want to be the "window manager" Sep 23 17:46:25 mjr: qvfb would make more sense with the OM window manager Sep 23 17:46:43 mjr: a X server for qtopia would make more sense with qtopia doing the "window" management Sep 23 17:46:57 or maybe someone will eventually port qtopia phone to x Sep 23 17:47:10 * cesarb wonders if qtopia can run over X instead of over FB Sep 23 17:47:12 i'm told its not a simple recompile against qt but it might not be that hard either Sep 23 17:47:38 if the smedia driver is any good, you'd want it to run against x Sep 23 17:47:45 It could run as a fullscreen window... should be easy (famous last words) Sep 23 17:48:20 isnt that what qvfb does Sep 23 17:50:01 cesarb, I don't count qtopia as doing the window management in a mixed environment as sane :] Sep 23 17:51:18 hey gyus, haven't been keeping up with the project much lately, just curious how the development is coming overall. I'm interested in getting the second hardware revision Sep 23 17:51:48 Hm, the qtopia recipes seem to be unable to use ccache Sep 23 17:51:57 peepsalot: december Sep 23 17:52:10 * cesarb trying to compile the qtopia tech preview with 2007.2 Sep 23 17:52:13 i'm not sure i'd even speak of windows with fullscreen only thingy Sep 23 17:52:31 any summaries of the state of the software? Sep 23 17:52:36 buz: think of it as a set of maximized windows Sep 23 17:52:47 peesalot: pre-alpha Sep 23 17:52:59 doh, still? Sep 23 17:53:03 i do. i'm just not sure whether i wouldnt think of it as separate desktops rather Sep 23 17:53:03 pre-pre-alpha really Sep 23 17:53:13 yes. Sep 23 17:53:13 hey it does receive calls now ;) Sep 23 17:53:20 (i've seen it ;) Sep 23 17:53:27 hehe, nice buz Sep 23 17:53:29 buz: when gsmd doesn't seize up... Sep 23 17:53:42 it didnt do so twice in a row Sep 23 17:54:57 wow i've never seen xp boot as quickly as a reboot in kvm (obviously all data was cached in ram) Sep 23 17:57:09 what is the preferred input method for the Neo ? stylus or finger? what is better ? Sep 23 17:57:35 unique: depends on what you want to do Sep 23 17:57:54 unique: for instance, dialing and basic interface is geared towards being easy to use with the finger Sep 23 17:57:56 cesarb: *gg* phone-call, write a sms Sep 23 17:58:19 unique: a drawing application, on the other hand, would be hard to use without the stylus Sep 23 17:58:53 Depends. Sep 23 17:59:00 For some things it's quite adequate Sep 23 17:59:07 unique: writing SMS depends on which onscreen keyboard you are using... so far the ones I've seen can be used with good sight and alignment abilities and neatly trimmed fingernails Sep 23 17:59:13 cesarb: of course, I meant for general, daily use. phone call, sms, select an ogg song for listening.. Sep 23 17:59:32 cesarb: that wasnt my impression, not for the standard OM anyway Sep 23 17:59:52 cesarb: thx Sep 23 18:01:04 when did GTA02 slip from october to december? Sep 23 18:01:18 just looked at wiki and seems it has changed Sep 23 18:01:31 observe the history! Sep 23 18:01:39 however - a week ago officially. Sep 23 18:01:53 the real slip must of course have happened earlier and not been reported Sep 23 18:02:18 I wonder if part of it happened when smoke was coming out from mickyl's gta02 Sep 23 18:02:38 So that they had to make another hw revision Sep 23 18:03:01 if taking orders ~dec 20th, then i guess can't expect a christmas present Sep 23 18:03:25 manyoso: that's an estimated timeline. Sep 23 18:03:47 manyoso: hte only date that's officially been given is 'Dcembter" Sep 23 18:04:09 * SpeedEvil stabs the invetor of the qweryt keyboard. Sep 23 18:04:37 yep. which means it is entirely likely we won't be seeing the device in hand before new year Sep 23 18:05:02 unless happen to get one of first devices shipped Sep 23 18:05:13 SpeedEvil, i'd stab some guy too if they invented a qweryt keyboard Sep 23 18:09:36 * hhf423 does not believe to have the gta02 for chrismas Sep 23 18:09:53 hope dies last Sep 23 18:10:04 I mean, the have the specs for the shop ready, but not the shop Sep 23 18:10:27 i dont see why they arent outsourcing that whole distribution stuff Sep 23 18:10:30 say the hardware is ready, but the software is not, will they still ship? Sep 23 18:10:32 better a later launch and stable than an early launch and quality problems Sep 23 18:10:38 manyoso: they will Sep 23 18:10:38 manyoso: yes Sep 23 18:10:50 i'd much rather have the hardware even if software isn't ready Sep 23 18:11:01 since i'll just put qtopia on it anyway Sep 23 18:12:46 manyoso: qtopia is not a completely different software Sep 23 18:13:04 manyoso: for instance, the kernel is the same, together with the associated power management needs Sep 23 18:13:27 manyoso: (and so far, the power management can be best described as "turn off the backlight") Sep 23 18:13:36 hehe Sep 23 18:13:43 cesarb: what about frequency scaling? Sep 23 18:13:49 heh Sep 23 18:14:00 cesarb: is there any info about that? Sep 23 18:14:15 actually i still didn't study the hardware of neo Sep 23 18:14:23 so maybe i'm talking bullshit Sep 23 18:14:30 bedboi: google s3c2410 cpufreq Sep 23 18:14:50 bedboi: SpeedEvil has been studying the frequency scaling, but AFAIK the software doesn't do it Sep 23 18:15:11 afaiu partial patches exist, but support is not complete (every device driver that uses a clock that changes needs to be cpufreq-aware) Sep 23 18:15:15 bedboi: the software doesn't even disable the sound amplifier when no sound is playing! (yet) Sep 23 18:15:23 omg Sep 23 18:15:24 that's bad Sep 23 18:15:44 cesarb: is it a userspace issue? Sep 23 18:15:45 bedboi: that's why the battery life so far is more of a lower bound than an average ;-) Sep 23 18:15:47 (script) wiki RecentChanges: [[Talk:Video_Player]] Sep 23 18:15:49 or the kernel code is not there Sep 23 18:15:53 bedboi: mostly kernel space Sep 23 18:15:56 ok Sep 23 18:16:00 gpio stuff Sep 23 18:16:04 bedboi: the userspace will probably have to cooperate Sep 23 18:16:56 of course Sep 23 18:16:56 bedboi: you can currently disable by hand the sound amplifier (and in fact must do so before powering off if you want it not to completely drain the battery, unless the patch to fix that has already been applied) Sep 23 18:17:11 cesarb: so the kernel code is there. Sep 23 18:17:15 if you can do it manually :) Sep 23 18:17:25 bedboi: you can also turn off by hand the GSM modem (nothing else does; and if nothing does, it also drains down the battery) Sep 23 18:17:45 cesarb: but it does not make any sense Sep 23 18:17:48 bedboi: the kernel code for turning it off is there; the kernel code for turning it off when nothing is playing isn't Sep 23 18:17:52 turning the gsm modem off Sep 23 18:18:01 cesarb: ok. Sep 23 18:18:24 i mean, this is a phone, you are supposed to use the gsm modem Sep 23 18:18:24 :) Sep 23 18:18:34 bedboi: if you want to completely power off the phone, you have to turn off the amplifier and GSM modem, since they are not connected via the power management chip, else the battery will drain completely Sep 23 18:18:59 cesarb, i read on the wiki about a backup battery... know anything about this? Sep 23 18:19:11 cb22_: the backup battery is the PMU's RTC battery Sep 23 18:19:17 ok, but i mean, you can't count on that for power management when the phone is on Sep 23 18:19:23 ah, ok Sep 23 18:19:23 cb22_: it keeps the PMU alive enough to keep the time and some of its settings Sep 23 18:19:32 cesarb: the point is: you have to leave the gsm modem on Sep 23 18:19:37 that's the lower bound Sep 23 18:19:47 but, you should lower the cpu draining rate Sep 23 18:20:01 by throttling the cpu speed Sep 23 18:20:29 bedboi: when the phone is on, you can AFAIK turn off part of the AT command interpreter, and must send a sequence of commands to make it proprerly wake back up (and wake the CPU back up) when needed; there's no code for that so far Sep 23 18:20:31 yep of course you have to turn off the screen Sep 23 18:21:24 it's a pity the smedia chip doesnt have audio support Sep 23 18:21:28 cesarb: btw what's the most battery hungry part? (not counting the screen) Sep 23 18:21:35 bedboi: and you have to turn off the sound amplifier when suspending the CPU, UNLESS you want to suspend the CPU but use the sound with it suspended (in the middle of a call is a common case) Sep 23 18:21:46 having that do mp3 decoding while the cpu is in slow mode or similiar would be nice Sep 23 18:22:07 bedboi: SpeedEvil will probably know the answer, but I'd guess the most power hungry part would be the CPU itself Sep 23 18:22:16 I'm unsure. Sep 23 18:22:38 The power usage seems way over what would be expected by the datasheet. Sep 23 18:22:50 eheh why's that? Sep 23 18:23:10 handwarmer feature :-) Sep 23 18:23:16 for us northern climes... Sep 23 18:23:24 Because what I measure is greater than what is expected by the sum of all the datasheet consumptions. Sep 23 18:23:26 | /usr/bin/ld: skipping incompatible /scratch/build/cesarb/openmoko/build/tmp/staging/arm-angstrom-linux-gnueabi/lib/libXtst.so when searching for -lXtst Sep 23 18:23:42 urgh, this OE recipe is not using the cross compiler properly... Sep 23 18:24:14 SpeedEvil, arn't data sheets always like that... Sep 23 18:24:33 cb22_: to an extent, yes. Sep 23 18:25:12 what have you measured the neo as using? Sep 23 18:25:24 SpeedEvil: don't some chips have a feature *cough* where if you put a pull-up when you had to pull-down or vice versa they use a very high amount of power? Sep 23 18:25:46 cb22_: a wide variety of figures Sep 23 18:25:50 SpeedEvil: and speaking of pull-up/pull-down, isn't it possible you are also measuring one of these? Sep 23 18:26:03 I'm measueing current out of the battery Sep 23 18:26:40 well the gta02 battery is 40% or so bigger Sep 23 18:26:52 indeed. Sep 23 18:27:02 but i wonder what the increase is like for the 2442 Sep 23 18:27:15 I neee to crack open my neo, and do some breakdown measurementd Sep 23 18:27:28 i suppose it would be able to run with the same consumption at clocked down + undervolted settings Sep 23 18:27:41 undervolting isn't really much practical. Sep 23 18:27:50 the PMU is limited to .3V steps Sep 23 18:27:51 Because what I measure is greater than what is expected by the sum of all the datasheet consumptions. <-- This is not uncommon Sep 23 18:27:58 the datasheets are always cheating Sep 23 18:28:05 bedboi: no, it's not. The question is where are the lies. Sep 23 18:28:11 And how can they be reduced. Sep 23 18:28:25 according to the power page on the wiki it seems as if ram is a real killer too Sep 23 18:28:31 Perhaps. Sep 23 18:28:55 I wrote most of the numbers in there - I need to update them a bit. Sep 23 18:29:00 And see if I can break out more Sep 23 18:29:38 Which probably means cracking open the PMU module Sep 23 18:29:45 good luck Sep 23 18:31:37 I started off being an electrical engineer when I was in college, and after seeing a friend of mine fried by 13,600 volts and 400 amps, which was not a pretty sight, I decided to switch over to the software thing, where I figured 5 volts and a milli-amp would do it for me. Sep 23 18:31:38 AHAHHA Sep 23 18:31:46 I love Maddog Sep 23 18:32:33 lol bedboi Sep 23 18:33:40 it's so funny Sep 23 18:36:12 bedboi: link? Sep 23 18:36:30 wait Sep 23 18:36:37 http://www.linuxtoday.com/news_story.php3?ltsn=1999-04-19-005-05-NW-LF Sep 23 18:37:19 THe whole interview is good :) Sep 23 18:38:00 yep man Sep 23 18:38:10 "Most of my best moments have something to do with Susa, he thought. With Susan, or with a computer. Never with Susan and computers. These where the worst moments." Douglas Adams Sep 23 18:38:28 this is from "Dirk Gentley, the Olistic Investigator" Sep 23 18:38:39 *detective Sep 23 18:39:01 Yeah, that sofa in the stairs was really troublesome :) Sep 23 18:39:08 ahaha yep Sep 23 18:39:12 ahah and the horse Sep 23 18:39:16 Hehe Sep 23 18:39:27 wtf that man was really amazing Sep 23 18:40:06 i love Douglas Adams. Sep 23 18:40:14 he is so smart. Sep 23 18:40:19 damn "was" Sep 23 18:40:27 "shoot" he says Sep 23 18:40:36 so the monk believed him Sep 23 18:40:37 I once saw Maddog speak at São Paulo... What I found most striking was the way he inserted pauses after every sentence, to give time for the translator to catch up Sep 23 18:40:41 and shot him Sep 23 18:40:43 hhf423: AHAHAH Sep 23 18:41:03 hhf423: that part is too hilarious.... i love the monk Sep 23 18:41:17 bedboi: I believe you Sep 23 18:41:18 hhf423: Must have been 15 years ago I read that book. Should re-read it Sep 23 18:41:42 I wound the note on compatibility most amusing Sep 23 18:41:42 hhf423: yep i believe in me and in the openmoko stuff Sep 23 18:42:16 "Never hook an electronic monk up to your VCR -- it will try to believe all the ads and burn out" Sep 23 18:42:56 ahahaha Sep 23 18:42:58 but I forgot where the monk had the gun from Sep 23 18:43:19 hhf423: he believed he had one. Sep 23 18:43:26 :-) Sep 23 18:44:25 * hhf423 makes a note never to take play-doh to an airport Sep 23 18:45:00 hhf423: Not if you have a circuit board on your shirt ;) Sep 23 18:45:20 hhf423: do you have electronic copy of dirk? Sep 23 18:45:31 damn i have the italian version i want to read the english one Sep 23 18:45:37 no, I have the book in my shelf :-) Sep 23 18:45:44 damn. Sep 23 18:45:55 and plus, my book is in italy Sep 23 18:45:58 CM: well, I am white, so maybe I'd have less issues Sep 23 18:46:26 bedboi: I'd be very surprised if you weren't able to find all of Adams' books in a torrent Sep 23 18:46:41 CM: i'm not so used to torrents Sep 23 18:46:52 actually never used that torrent stuff Sep 23 18:46:54 did anyone read the new Terry Pratchett yet? "Making Money"? Sep 23 18:46:55 * CM waves a black flag with some white stuff on it, standing in a bay Sep 23 18:46:56 :D Sep 23 18:47:10 ahaha Sep 23 18:47:18 hhf423: nope. Sep 23 18:47:22 what's that? Sep 23 18:47:40 whats what? Sep 23 18:47:46 http://en.wikipedia.org/wiki/Making_Money Sep 23 18:48:42 with moist von lipwig -- its going to be hillarious, like "going postal" Sep 23 18:49:54 yeah, going postal was fun to read Sep 23 18:50:13 * mwester still thinks "Interesting Times" was the most fun. Sep 23 18:51:05 I also liked "The Truth" a lot Sep 23 18:52:22 ah, making money is not out yet Sep 23 18:52:29 That was almost too close to reality, what with Mr Pin and all. I think Mr. Pin was born in one of Chicago's suburbs... Sep 23 18:52:41 hhf423: thanks for that. i'll read it Sep 23 18:53:21 hhf423: it is out - http://www.amazon.co.uk/Making-Money-Terry-Pratchett/dp/0385611013/ref=pd_bbs_sr_1/203-1421806-6653506?ie=UTF8&s=books&qid=1190573569&sr=8-1 Sep 23 18:53:38 oh Sep 23 18:53:49 oooh, and a new book from spethen donaldson Sep 23 18:53:53 stephen Sep 23 18:55:12 grrreat Sep 23 18:55:50 I wonder why amazon can not just send me mail when one of the authors I name publishes a new book Sep 23 18:56:08 there should be nothing easier than that Sep 23 18:56:14 I liked "The gap" series Sep 23 18:56:15 Real life rss. Sep 23 18:56:53 Have anyone tried evince on the neo? Or xpdf or something? Sep 23 18:57:57 I've ran evince and xpdf remotely as a tentative pdf test, was okay Sep 23 18:58:09 CM: if you liked the gap, then you're sick :-) Sep 23 18:58:36 hhf423: Whaat? :P Sep 23 18:58:55 Angus Thermophyle or what ever his name was was discusting, sure.. Sep 23 18:59:28 CM: but I agree, one has to read them all after one started with one Sep 23 18:59:31 making money: http://blog.alism.com/fake-one-pound-coins-part-one/ Sep 23 18:59:32 But I remembered that I liked them. Read those ones about the leper too, but that was even longar ago Sep 23 19:00:09 * CM is reading Tad Williams right now Sep 23 19:00:15 CM: did you read the "The Man Who..." series? Sep 23 19:00:21 Btw, the Otherworld series by him is really great Sep 23 19:00:23 which tad williams? Sep 23 19:00:46 hhf423: http://en.wikipedia.org/wiki/Tad_Williams Sep 23 19:00:55 CM: doh, I read Otherworld, all. hard cover. and then I sold them real quick. Sep 23 19:01:12 has anyone managed to switch the screen to 320x240 yet? Sep 23 19:01:20 Why? Sep 23 19:01:22 I had to find out how its end, but the story is slow and we wrote it as if he got paid by the word Sep 23 19:01:49 hhf423: all his books are slow, the dragon bone chair too, but I don't mind that. :) Sep 23 19:01:53 and I found no new idea or concept in that story Sep 23 19:02:07 kristian-m: Haven't heard anyone doing that yet Sep 23 19:02:22 CM hmm it wasn't working for me Sep 23 19:02:40 CM could be useful for video Sep 23 19:02:45 hhf423: Read Peter F. Hamilton though? Sep 23 19:03:05 kristian-m: For sure, full screen mediaplayer :) Sep 23 19:03:50 is it some hello world application that shows how openmoko UI framework works? Sep 23 19:04:01 CM: I think I read one or two of his books, can't really remember, though Sep 23 19:05:06 s/is it/is there/ Sep 23 19:05:06 bedboi meant: is there some hello world application that shows how openmoko UI framework works? Sep 23 19:05:19 CM: which did you like best? Sep 23 19:06:07 i know i'm off-topic Sep 23 19:06:08 :) Sep 23 19:06:54 bedboi not really afaik, but take a look at the calculator for example Sep 23 19:07:08 ok Sep 23 19:07:16 i'll get all the tree now Sep 23 19:09:10 CM: of Peter F Hamilton? Sep 23 19:09:17 hhf423: Pandora's Star and Judas Unchained were good, but I think I read some oter ones too Sep 23 19:09:42 I see Sep 23 19:10:26 I read Judas Unchained in one read, 11h straight on a plane from japan Sep 23 19:10:34 Felt totally weird :D Sep 23 19:11:23 where is the svn? Sep 23 19:11:32 :-) Sep 23 19:11:42 I'm guessing projects.openmoko.org or? Sep 23 19:12:54 yep the svn of the openmoko libraries and stuff Sep 23 19:15:07 (script) planet: Holger "zecke" Freyther: Traveling, Planning, Failing http://zecke.blogspot.com/2007/09/traveling-planning-failing.html Sep 23 19:16:12 oh stupid me... it was in the openembedded tree Sep 23 19:17:48 in calculator's soruces there's no particular reference to openmoko ui Sep 23 19:17:58 so it's plain gtk Sep 23 19:18:54 uh.. sorry Sep 23 19:19:11 roh: btw it's ok i guess Sep 23 19:19:17 calc seems to work Sep 23 19:19:25 i want some simple example Sep 23 19:19:26 hm... http://folks.o-hand.com/thomas/libmokoui2/ Sep 23 19:23:19 roh: thanks. Sep 23 19:23:39 so mainly a gtk application should just work Sep 23 19:25:59 bedboi of course Sep 23 19:27:48 cool Sep 23 19:29:33 is there any way to use the openmoko keyboard instead of the matchbox one? Sep 23 19:29:49 the matchbox keyb is way to difficult to use (small keys) Sep 23 19:56:41 someone needs to update the openmoko.org portal site picture it's showing the old interface Sep 23 20:05:41 anyone know if there are plans to have a micro usb instead of mini? Sep 23 20:07:20 and will there be a stage 1.5? with stage 2 hardware? Sep 23 20:08:44 webjames no. why should we change to micro-usb? the connector is total crap Sep 23 20:08:59 mini-usb isn't usb anymore ;) Sep 23 20:09:16 because it's just become an industry standard, and it's not Sep 23 20:09:18 drath it is. even specified by the usb consortium Sep 23 20:09:35 hi all Sep 23 20:09:37 Deprecation of the Mini-A and Mini-AB Connectors Sep 23 20:09:53 roh http://uk.gizmodo.com/2007/09/23/one_mobile_phone_charger_to_ru.html Sep 23 20:10:07 http://www.usb.org/developers/Deprecation_Announcement_052507.pdf Sep 23 20:10:36 webjames well.. not everyone needs to jump on the train running against a solit concrete wall Sep 23 20:11:34 i'm not so sure that's true Sep 23 20:11:40 why do you say that roh? Sep 23 20:11:50 webjames everyone i know has mini-usb in his pockets and atleast one or 2 devices with that already. micro-usb is much more fragile, does not survive dust, dirt and reasonable charging currents. do i will do what i can to protect us against it Sep 23 20:12:42 webjames i say that out of the practical experience as someone who uses equipment every day and lives with people who need atleaqst one new notebook keyboard per year due to wear-out. Sep 23 20:12:59 guess lack of certification won't be a problem for the current neo, because its usb wont be certified anyway, but i suppose mass-market devices without a usb logo might be a problem Sep 23 20:13:01 roh: surely the manufacturers would have similar concerns. How then did this come to be, if it's as bad as all that? Sep 23 20:13:14 and yes i do use my equipment outside of clean-rooms. e.g. while camping. Sep 23 20:13:36 i mean, it's been tested and performs better than mini Sep 23 20:14:03 You're not supposed to use the USB cable for stringing fish or hold up tent poles. Sep 23 20:14:24 mwester-laptop just look at the socket and tell me how it should survive even minimal mechanical stress, some sand or even 500mA current. don't think about one amp for changing with a external pocket-charger Sep 23 20:15:05 mwester-laptop i don't care what people in suits tell me what to do. i see what happens when you use stuff and are not ready to force stupid limitations on the users Sep 23 20:15:20 My observational skills are limited, I rely on engineers to test things like that :)P Sep 23 20:15:27 (script) wiki RecentChanges: [[Talk:Supported_microSD_cards]] [[User:Roh]] Sep 23 20:15:51 china has just accepted it as well: http://english.peopledaily.com.cn/200612/19/eng20061219_334047.html Sep 23 20:16:02 mwester-laptop i have an electric bike which runs of lead-gel batterys.. believe me.. current is evil for connectors ;) Sep 23 20:16:16 500ma is plenty for phones. Sep 23 20:16:45 I don't like my phone to double as a handwarmer just so I can fully charge it in 10 minutes. Sep 23 20:16:47 mwester-laptop its the minimum for reasonable chargetimes. but a lot of these devices can also use 1amp with a fitting charger Sep 23 20:17:25 webjames china has specified usb, but not the explicit socket. atleast ive not seen a reasonable spec yet Sep 23 20:18:51 roh: I see you feel strongly - that's a good thing. I travel for business a lot, and anything that can reduce the tangle of chargers and cables I am forced to carry by manufacturers who all try to tell me that its for my own good, well, it's a good thing. Sep 23 20:19:13 I can carry a micro-mini adaptor, as long as I can carry only a single charging device. Sep 23 20:19:35 Or better yet, plug into one of the many USB chargers that are appearing in business centers and airports. Sep 23 20:20:03 Long live standardization on things (like chargers and cables) that should be commodities! Sep 23 20:20:39 mwester-laptop: that's what it's meant to stop, you have one lead. you could even take spares roh Sep 23 20:21:11 mwester-laptop i see it pragmatic. we could change stuff, but not for gta02. if we want to is still not discussed. but my standpoint is and will be to use the most accepted, most usable and mechanically most solid one. that is currently mini-usb. eventually we would go the mini-ab way when we really implement otg (not decided or discussed yet also) Sep 23 20:22:07 besides that. even motorola got the message, long before the chinese call for usb-charging and also uses the same connector as we do Sep 23 20:22:59 I'm all for pragmatic, and I certainly hope that we don't slow the gta02 for that. But a choice based on sound engineering makes sense, and I can't reject the micro until there's more field experience or engineering data. Sep 23 20:23:09 nokia on the contrary still is totally braindead. they have also a mini-usb on the new communicator, but can only charge via their own (now also changed) stuff which means it discharges, even while connected to a computer Sep 23 20:23:39 roh: yeah, thats brainddea. gf ha s aNokia 5300 with a mini-usb that doesnt charge Sep 23 20:23:41 roh, that is daft Sep 23 20:23:42 mwester-laptop gta02 external case design is fixed. so i can assure you the connector will not change Sep 23 20:23:49 Good Sep 23 20:23:59 what is the ipkg install command for webkit? Sep 23 20:24:25 i'm sure with a soldering iron you could change it Sep 23 20:24:35 I've done 'make build-package-webkit-gtk' Sep 23 20:24:38 BTW, one of my motorola phones shipped with a right-angle mini-B adaptor -- I cannot find that anywhere, and it would reduce the stress on my tethered gta01 -- anyone know where to find such an item? Sep 23 20:24:39 webjames nope. will not fit mechanical Sep 23 20:24:40 should i inherit from openmoko if i use the kinetic scrolling? Sep 23 20:25:49 ive not seen angled ones. but i have an moto charger here which i got with my razr v3i. thats the sort of 'standardized mobile phone power cord' which i really would like to see more Sep 23 20:26:15 mwester-laptop: I can get you the code number for you to search Sep 23 20:26:28 cesarb: that would be good. Sep 23 20:28:05 mwester-laptop: SKN6182A Sep 23 20:28:14 Thanks Sep 23 20:28:15 hm.. htc seems also to use mini-usb Sep 23 20:28:38 need to get some chargers from all these manufs and find out the idpin resistors.. Sep 23 20:28:44 mwester-laptop: just had to get my razr box and check all the small plastic bags on google until I found the right one ;-) Sep 23 20:29:00 (each one has a code for the piece which came on it) Sep 23 20:30:51 roh: no need to play with id pin resistors, just check for the short between D+ and D- Sep 23 20:31:03 roh: see the new usb battery charging specification for how Sep 23 20:31:37 cesarb ignore that spec. its new and not implemented everywhere. besides that we will not change the hardware to be able to detect the short Sep 23 20:31:44 atleast not for gta02 Sep 23 20:32:11 roh: if I read that specification correctly, doing the short is very common... (I know moto does it) Sep 23 20:32:19 cesarb i know that motorola has another resistor than we do, and than htc has. i just need to measure them Sep 23 20:32:51 cesarb moto does it longer than the spec exists, so i assume theyve written it Sep 23 20:34:06 roh: the problem is that the resistor on ID pin trick is also used for other things (like headsets or car adapters), and if the value used for chargers by one manufacturer is the value used for headsets on other manufacturer, you cannot use the ID pin as a means of detection... Sep 23 20:34:07 That's the smart thing to do, if you can -- do it first, and write the spec... Sep 23 20:35:18 cesarb yes. but its not a problem since you can detect external connected voltage and the value. there are a lot different ones. and they are usually 1% tolerance specified Sep 23 20:35:38 What is the purpose of the resistor then? Is it for more than just to determine the 100ma vs 500ma thing? Sep 23 20:36:01 mwester-laptop yes. basically. Sep 23 20:36:25 it also is used with different values fo the other 10-12 modes the socket can do (on motos) Sep 23 20:37:00 Ah, ok - so that's how the razr tells if the earplugs are inserted vs the charger. Sep 23 20:37:26 they also have some high-voltage programming mode, a serial, audio for headset, audio for stereo headset and other stuff on there. all differenciated by short on d+ d- and/or resistor values Sep 23 20:38:02 ive not found a sane spec about that but you can find some service.manuals/schems with the info on google *couch* Sep 23 20:38:14 * mwester-laptop wonders if it might not be wise to just "re-use" the moto scheme on the >gta02... Sep 23 20:38:23 dunno where 3rd party equipment manufs get their specs Sep 23 20:38:45 mwester: nope. the scheme is only partial the right thing to do. Sep 23 20:39:52 But if there are no existing standards, doesn't it make sense to use the same as one of the larger players, so as to capitalize on the third-party accessories? Sep 23 20:40:32 There's also CEA-936-A, which manages to do even more bizarre things (serial over USB cable...) Sep 23 20:40:52 mwester: as i said it does not fit Sep 23 20:41:39 would it cause additional delays to switch the connector? or is there some other reason? Sep 23 20:43:21 roh: can GTA01/02 read the precise ID pin voltage, and add a pull-up to measure the resistence? Sep 23 20:43:35 roh: that's why I mentioned > gta02... Sep 23 20:45:11 Of course, I would prefer the 2.5mm audio jack in addition to the USB, rather than a single "does-everything-but-only-one-at-a-time" USB+ connector... Sep 23 20:50:09 btw.. how many bluetooth devices GTA01B_v04 can do at same time? Sep 23 20:51:51 mwester: i don't see any reson right now to change to a moto scheme, which does need a lot of parts and thus is not cheap Sep 23 20:52:14 and ack. i want to have a 3.5mm jack in addition to the usb, not shared ones Sep 23 20:52:51 what about 2 usb sockets? Sep 23 20:52:55 cesarb i wouldnt say precise and ive not tested it yet, but atleast gta02 should be able to measure the idpin Sep 23 20:53:04 aloril: too big. and whatfor Sep 23 20:53:07 roh, was that 'a 3.5mm jack' a leak that we are going to have 3.5mm jacks for gta02 or just a slip of the toung :) Sep 23 20:53:29 cb22 no. gta02 will have no casechanges as i said. so 2.5mm jack. Sep 23 20:54:03 cb22 but lateron we hopefully can get space for a 3.5mm jack. its more stable (mechanically) and you can get more high-quality headphones Sep 23 20:54:12 yeah Sep 23 20:54:29 cb22: besides that.. nothing is decided yet for post gta02 Sep 23 20:54:30 Hm, it's quite interesting that they deprecated mini-A and mini-AB so easily? Must it be because nobody actually used them? And, if so, what makes they think people will use micro-A and micro-AB? Sep 23 20:54:32 roh: well.. 2.5mm/3.5mm audio jacket is probably better than second usb jacket, but then if there was 2, you could use 2 different usb devices (just like PC) Sep 23 20:54:35 do you think that the gta02's will be ready for christmas? Sep 23 20:55:16 cesarb thats bullshit. motorola hast them in every phone. all the usb equipment i have here which could not get a usb-b socket due to size has mini-usb Sep 23 20:55:33 roh: I said -A and -AB, not -B Sep 23 20:55:46 aloril nope. would not make sense much. besides that usb-host is short time only Sep 23 20:56:02 aloril at 100ma 5V the battery drains rather quickly Sep 23 20:56:06 aloril: there's no need at all for two A-side USB plugs... since you can simply use a hub. Sep 23 20:56:13 cb22 dunno. i hope so Sep 23 20:56:27 cesarb: yeah Sep 23 20:57:24 aloril: and in fact most of the use-cases of the host USB on the neo can also be used via bluetooth, except if you want power Sep 23 20:57:39 aloril: headset? bluetooth. keyboard? bluetooth. mouse? bluetooth Sep 23 20:57:41 roh: I would say: audio jacket + usb > 2*usb > 1*usb only ;-) Sep 23 20:58:22 aloril: I'd think that, if you didn't have to recharge the neo (and use uboot), it could all be done via bluetooth... Sep 23 20:59:01 (I'm playing a bit of devil's advocate here; my keyboards/mouse are all wired USB) Sep 23 21:00:35 There's fortune and fame waiting for the person who comes up with the energy source these devices really need - small, light, super-high capacity, capable of fast charge, and non-explosive. Sep 23 21:01:01 Then bluetooth becomes a much better solution for many options. Sep 23 21:03:41 mwester-laptop: "non-explosive" being the particularly hard bit :D Sep 23 21:04:24 mwester-laptop: the standard answer is "supercapacitors". It has been the standard answer for several years so far. Sep 23 21:05:44 mwester-laptop: fast charge is a false requirement; just make it hot-swappable, and one can let it charging while the device is being used with another one Sep 23 21:15:07 hm.. in http://uk.gizmodo.com/2007/09/23/one_mobile_phone_charger_to_ru.html they reference the OMTP .. i cannot find them Sep 23 21:15:27 also they write "Like mini-USB, micro USB also allows mobile phones to be charged from PCs but is a lot more durable.-Martin Lynch" Sep 23 21:15:54 this -Martin Lynch does not seem to be someone who knows what connectors are which.. Sep 23 21:18:53 ah.. omtp.org... hm.. does not look as 'most of the industrie' nor as if they have support in asian companies... Sep 23 21:19:18 OT: http://pidgin.im/~seanegan/webkit6.gif Sep 23 21:22:01 roh: unless there's some real advantage to micro-USB, there is no real reason to use it except on really small phones Sep 23 21:22:08 the neo is NOT a really small phone ;-) Sep 23 21:24:07 the neo could be used as a blunt weapon Sep 23 21:24:53 cesarb ive just browsed the linked document there.. and it seems they either confused micro and mini totally or have messed up their definitions Sep 23 21:25:35 cesarb besides that the consortium exists from the not-implementors of sane usb connectivity and carriers who want have more unified external equipment (carkits and so on) Sep 23 21:26:02 roh: standard have been hijacked before Sep 23 21:26:33 so.. in the end.. they are neither usb.org nor do they represent the majority of handset producing companys.. will see.. but i really have problems taking these guys seriously considered all the facts Sep 23 21:27:11 cesarb as longs as moto does mini-usb even in the razrs i thinkg small phones do not need micro ;) Sep 23 21:27:16 The real issue will be firstly will the industry take them seriously, and secondly will the consumer take them seriously. What we think is not even on the list, in the end. Sep 23 21:28:08 roh: there are thinner phones than the razr (but I'd guess they can also use mini-usb... the limiting factor is the size of the GSM antenna) Sep 23 21:29:29 I've been wondering when someone will create the "component" Sep 23 21:29:43 device - much like the component audio/video systems Sep 23 21:30:01 ok.. usb.org has revised the usb-otg spec (not that ive seen real devices using it) Sep 23 21:30:05 separate packages for the gsm modem, the audio system, the dialer, etc. Sep 23 21:30:27 and it basically says: mini-usb-b stays, mini-ab and mini-a goes and will not recieve cert. from now on. Sep 23 21:34:04 BTW, has the libmokoui2 problem been sorted yet -- do current builds generate working images again? Sep 23 22:15:47 (script) wiki RecentChanges: [[Wish_List_-_Hardware]] Sep 23 22:26:04 *sigh* .,. i should have known better... Sep 23 22:26:29 and made that 'no oscommerce please' point more prominent at start Sep 23 22:27:58 roh - *ping* Sep 23 22:28:54 pong Sep 23 22:29:42 roh - Where you able to reproduce bug #567 on any of your hw? Sep 23 22:37:43 Elrond nope.. but i also did not try... Sep 23 22:38:34 Elrond will do somewhere this week.. i also think its a bad idea to have mmc compiled in fix.. i do not boot from mmc and it takes a lot time to detect it Sep 23 22:39:55 roh - Well, mmc shouldn't hurt? Sep 23 22:40:47 mmc hurts boot time big Sep 23 22:40:53 some seconds. Sep 23 22:40:56 which is a lot Sep 23 22:41:30 same as ide or scsi does also.. detecting takes time. that could be done later, async while booting the rest Sep 23 22:42:05 Ahh. :-) Sep 23 22:42:12 Well, I don't care about boot times. :) Sep 23 22:42:45 Elrond what ive seen is that the mmc driver detects and looses the mmc all the time, about every one or 2 seconds.. but i need to recheck that with a current build Sep 23 22:43:21 that was some time after mmc was fixed (ioerrors) and sdhc was successfully tested.. could be as well be a regression Sep 23 22:43:34 eventually even combined with the move from module to compiled in Sep 23 22:44:24 BTW: Why is fat compiled in? (I think, it is) Sep 23 22:46:26 because you would need it to read the sd Sep 23 22:46:35 when its not ext Sep 23 22:52:24 Well, fat would not be able to hold a rootfs or something like that. So "boot from fat" is not an option. So you need to have your root somewhere else anyway. Which means: You could put the fat module _there_. Sep 23 22:53:04 could.. yes.. Sep 23 22:53:58 we'll see what combinations make sense in the end. thats all just 'put that way' for some reason, not genuine designed after careful surveys Sep 23 22:54:26 Yeah, no problem. Sep 23 22:54:41 As I said: I don't care a lot for boot speed or "super lite kernels". Sep 23 22:54:53 128MB RAM is a lot, I like _that_ a lot. :-) Sep 23 22:55:21 half a week ago, I used the RAM to make a backup of the rootfs. :-) Sep 23 22:55:33 dd if=/dev/mtd4 of=/tmp/rootfs.bin Sep 23 22:55:58 host$ scp -p fon:/tmp/rootfs.bin /export/backup/rootfs-$date.jffs2 Sep 23 22:56:24 (of course I had to hard-reboot after that, because the fs was not available, you know, bug above.) Sep 23 22:57:49 is it ok to go ahead and compile qemu with gcc4.0? Sep 23 22:58:20 Madnis_ - Use 4.1, if you want 4.x Sep 23 22:59:19 Madnis_ does not work. Sep 23 22:59:25 the mokomakefile is failing because qemu is asking for 3.x Sep 23 22:59:39 so bad idea to force it? Sep 23 22:59:46 Madnis_ qemu needs a 3.x gcc Sep 23 23:00:07 gotcha Sep 23 23:00:40 gcc 4 generates differend binarys and qemu depends on gcc3 layout since it chops it up and builds like 'parts' to reassemble with the dynamic recompiler afterwards on execution Sep 23 23:01:52 so.. there could be a way to use gcc4, but that involves knowledge of gcc, dyngen and some assembly/c stuff Sep 23 23:02:19 lol, little over my head Sep 23 23:03:53 ok there it goes Sep 23 23:04:05 one step closer to hello world... :) Sep 23 23:38:49 back Sep 24 00:47:00 so how much harddisk space do you need for the delete when finished method of compiling openmoko? Sep 24 00:54:50 wonko__ with rm_work my openmoko workdir (checkouts, compiled binarys, toolchain etc) takes about 2G .. while compiling it would be wise to have a few free Sep 24 00:55:36 wonko__ i think if you have about 5-8gig free space before starting it should be no problem (assumed rm_work) .. when building without rm_work its somewhere between 12 and 20gig Sep 24 00:56:14 am I the only one who thinks that is a bit ridiculous for building something that fits on a mobile phone?... Sep 24 00:56:57 Yes. Sep 24 00:57:12 consider that it's an entire linux distribution pretty much. Sep 24 00:57:52 Mek heh... youve not seen what happens elsewhere... compared to my experience om is quite sane Sep 24 00:57:56 well... but why should it need to compile lots of package it isn't going to install by default... Sep 24 00:58:04 It also builds gcc 2 times, glibc 3 times, pcmcia utils for a phone without pcmcia... Sep 24 00:58:44 cesarb thats partially clutter which comes from oe.. given somebody who cares about optimizing that will get away Sep 24 00:59:03 cesarb building gcc 2 times is neccessary for clean crosscompilers afaik Sep 24 00:59:13 glibc dito.. atleast 2 times Sep 24 00:59:23 pcmcia is true.. we do not really need that Sep 24 00:59:48 but i'm not aware of big ressource-waste there atm Sep 24 00:59:51 roh: I've seen worse, but can't recall right now Sep 24 00:59:54 for example, what is the 40MB of qtopia it was downloading in the default build? :) Sep 24 01:00:12 Mek because they do not package qmake2 as single tarball Sep 24 01:00:20 and we need that to build webkit Sep 24 01:00:21 Mek: IIRC either qtopia or qt is to get the (very small) qmake Sep 24 01:01:04 i share the sources dir here between multiple trees, and even an openwrt one.. so.. thats not really a problem (given broadband) Sep 24 01:03:38 pcmcia - there are still some phones out there with CF are there not? Sep 24 01:03:44 roh: it would be fun if the neo came with a DVD with a fully primed OE (all fixed srcrevs, everything already in the sources directory) so you'd only have to type "make" (after copying to your hd) and not download anything ;-) Sep 24 01:03:50 Or are they all too small/slow Sep 24 01:14:41 cesarb would be.. but only useful in the third world/where you have no broadband/parts of the us *duck* Sep 24 01:15:01 cesarb currently we are far from release-quality and fixed enough versions at all Sep 24 01:16:27 roh: that obviously could only work with the "final" ("frozen") release of the code in the phone to be sold to the public... Sep 24 01:16:49 roh: opposite to the current "here's the uboot" situation Sep 24 01:16:54 hd-dvd... Sep 24 01:17:08 i don't think we should ever freeze anything. we should do a release and a head trunk when appropriate Sep 24 01:17:31 release as in 'stable' .. compare that to ubuntu or debian. Sep 24 01:18:10 For some users - bugfixes only - might be nice. Sep 24 01:18:25 so you buy it and it has a stable release on it.. you update it via ipkg and the appmanager to get the latest bugfixes and securityfixes, or flash it to an 'unstable-' trunk.. your choise Sep 24 01:18:32 Once it's in a sane state of course Sep 24 01:18:38 SpeedEvil bug and security Sep 24 01:18:44 Yeah. Sep 24 01:19:16 And the users I'm talking of would look at you like you grew a third head if you said ipkg. Sep 24 01:19:36 Many - eventually - shouldn't care that it's got linux on. Sep 24 01:20:21 SpeedEvil we don't have these users in focus yet and propably will not for quite some time. this is geek, nerd and university project country as well as basis for other people with a developer in mind Sep 24 01:20:38 yeah. Sep 24 01:20:46 it's going to be a fair while Sep 24 01:55:58 Hello, If I wanted to re-partition my nand (so as to make more space available for 'kernel') is it just a matter of adjisting 'mtdparts' in uboot, and then resending kernel, splash and rootfs to their new locations? Sep 24 01:58:05 cheriff what do you want to put into the kernel which cannot work as module? Sep 24 02:02:45 roh: we (OpenKernal Labs) are using the phone as a board to demo our own kernel. Our final image has our kernel, some userspace tasks and 2x(vmlinux+minimal ramdisk) all wrapped into a single .elf Sep 24 02:03:17 *cough* Sep 24 02:03:29 flash it into the fs partition and change the bootconfig? Sep 24 02:03:43 thats much easier atleast Sep 24 02:04:00 we have a jffs2 partition there already, which it mounted post-boot Sep 24 02:04:18 unless uboot could pull the kernel out of said jffs2 partition? Sep 24 02:05:02 cheriff it can use ext from sdcard i think, but i dunno if it can do jffs2 from mtd Sep 24 02:06:17 okay. That was one option we are considering. Thanks :) **** ENDING LOGGING AT Mon Sep 24 02:59:57 2007