**** BEGIN LOGGING AT Fri Jun 08 02:59:56 2007 Jun 08 03:00:44 aloril: i just put xeyes up on my p1 no prob :) Jun 08 03:01:06 bmidgley: what is speed? (does it start instantly?) Jun 08 03:01:33 aloril: yes but there has to be a better test app... let me look around and see what i have Jun 08 03:02:38 maybe its gtk and/or some other extra stuff, not X11 that is problem ? Jun 08 03:03:17 and maybe all that can be fixed by 'just' optimizing it, not changing to something else Jun 08 03:04:35 aloril: what are you trying to fix? Jun 08 03:05:01 aloril: gpaint looks surprisingly good Jun 08 03:05:01 scuffe: for example starting dialing app and dialing should be instantenous or calendar or other 'simple' things Jun 08 03:05:24 is there a lag? Jun 08 03:05:26 or taskmanager, etc.. Jun 08 03:05:41 scuffe: aloril: well, the first time i tapped on TaskManager now it took some 4s to load and about 2s the successive runs - can't test anything else because the desktop and the top bar had crashed during suspending :P Jun 08 03:06:13 fwiw, gpaint running remotely is very snappy Jun 08 03:06:15 oh Jun 08 03:06:28 maybe bottleneck is in cpu/flash i/f Jun 08 03:07:47 bmidgley: thanks, maybe you want to mail community mailing list to say that X11 is not culprit Jun 08 03:10:24 what's a good way to capture a screenshot? Jun 08 03:11:30 ah, so its not only X11 program, its GTK+ program too, nice benchmark then Jun 08 03:13:00 /usr/bin/import -display localhost:0.0 -window root screenshot_filename.jpg Jun 08 03:15:55 http://www.xmission.com/~bmidgley/gpaint.jpg Jun 08 03:17:55 bmidgley: that's sweet Jun 08 03:18:11 hey the backlight actually turns off now--used to run 24/7 Jun 08 03:18:39 scuffe: the original looks better... jpg encoding wasn't kind Jun 08 03:18:54 bmidgley: open/save/scrolling canvas/etc.. are all fast? Jun 08 03:20:01 aloril: very snappy yes. opened a big jpg and scrolling around is fine Jun 08 03:20:30 ok, then its not gtk+/x11 that is speed problem Jun 08 03:20:41 and this is even over usbnet I assume? Jun 08 03:20:52 yes usbnet Jun 08 03:21:50 * aloril wonders how fast moko apps work if you run them on PC too... (but that would require them to be compiled for it) Jun 08 03:22:22 yeah it would be fun to display qemu output on the neo :) Jun 08 03:23:46 bmidgley: that works directly (and slowly), but I meant 386 openmoko binaries that would run at native PC speed Jun 08 03:24:04 (which is extra work and doesn't really tell that much) Jun 08 03:24:28 what we your test says is that X11/GTK+ display parts is not speed problem Jun 08 03:24:53 what we have now is that your test says that X11/GTK+ display parts is not speed problem Jun 08 03:25:04 kolourpaint uses the screen better but covers the bottom bar (shrug) Jun 08 03:25:44 man these pixels are tiny Jun 08 03:25:44 what about something heavy like ... mozilla? Jun 08 03:25:50 ok Jun 08 03:27:03 wow even classic mozilla (1.7.13) is snappy as anything! i'll try firefox Jun 08 03:30:17 actually gtk+ is not exercised 'directly' on Neo1973 in these tests, only X11 (but it means that GTK+ is not any extra slow for drawing stuff, if its extra slow its at non 'actual drawing' stuff) Jun 08 03:31:14 re topic: perhaps any week now? que? Jun 08 03:31:14 cxreg2: See http://wiki.openmoko.org/wiki/FAQ (Hopefully this link answers your question. This autotrigger for new nick and question during first 15 minutes done only once.) Jun 08 03:31:17 yeah it is pretty low level Jun 08 03:31:33 konqueror looks great Jun 08 03:32:18 in any case it kills "we need to dump X11 and use framebuffer" -argument Jun 08 03:33:00 sure... it doesn't exercise the widget set directly... Jun 08 03:33:19 but using fb directly would still need to use the widget set so it wouldn't help Jun 08 03:33:49 i'll see if i can run something on moko and display on my pc Jun 08 03:34:19 right Jun 08 03:35:46 argh gdm on my desktop doesn't allow tcp x Jun 08 03:36:02 actually that "run moko apps on desktop" argument could be good for general too Jun 08 03:36:51 ie.. easy way to use bigger monitor than what Neo1973 provides Jun 08 03:37:25 actually, even better if you can just switch between these .. Jun 08 03:38:05 go near PC -> window pops up with Neo1973 screen, move away, back to Neo1973 screen Jun 08 03:38:30 hmm konqueror goes over the taskbar Jun 08 03:38:35 http://www.xmission.com/~bmidgley/konqueror.jpg Jun 08 03:39:01 the location input gets scrunched up to make room for the search Jun 08 03:40:22 that screen must be hard to read on actual Neo1973 ;-) Jun 08 03:41:08 i think my eyes would get tired Jun 08 03:41:55 i like how the latest image has a startup sound Jun 08 03:42:55 ok, so what would be good to run on neo and display on pc? got it working finally Jun 08 03:43:55 doh, mtpaint segfaults Jun 08 03:44:46 oh never mind, mtpaint segfaults when displaying on neo :( not a good test Jun 08 03:46:54 bmidgley: usual apps: dialer, contacts, calendar, etc..? Jun 08 03:47:06 (moko apps) Jun 08 03:51:06 i wish the dev enviroment would compile faster so i could play around with everything Jun 08 03:53:30 openmoko-chordmaster? never knew this was here Jun 08 03:54:53 things seem fast enough, don't take too long to start Jun 08 03:55:01 there is a ton of debug text coming out of them Jun 08 03:55:38 if the debug was going somewhere on flash i imagine that would slow things down Jun 08 03:55:50 for me it comes out on the console Jun 08 03:56:54 how does it compare to running it in Neo1973? Jun 08 03:58:20 the things on neo that make me think "slow" are tapping on the bottom left & right corners Jun 08 03:58:23 for task switching Jun 08 03:58:53 most other stuff is reasonable and displays remotely about the same speed, not blazing Jun 08 03:59:07 eg calculator Jun 08 03:59:13 couldn't get the dialer to go remote Jun 08 04:03:35 ok, so scrolling in appmanager is not so hot Jun 08 04:03:42 can't keep up with live scrolling Jun 08 04:04:48 strangely, the thumb keeps up with the cursor on my pc but the thumb is way behind on neo Jun 08 04:05:01 both are skipping around in the list and not scrolling smoothly Jun 08 04:05:55 it's as if there's an issue in the input driver Jun 08 04:06:08 but no... painting apps behaved ok Jun 08 04:08:20 sounds like things are fixable (especially on GTA02), no need to do drastic changes Jun 08 04:08:47 i would feel better if all that debug were turned off Jun 08 04:09:04 brb Jun 08 04:29:19 oh hang on there was one difference when the thumb didn't keep up... Jun 08 04:29:37 when displaying on pc, I had redirected stdout/err to /dev/null Jun 08 04:29:46 i don't know where stdout/err are going on neo Jun 08 04:30:51 nevermind, makes no diff Jun 08 04:41:53 (script) wiki RecentChanges: [[Great_Wall]] [[Liver_transplant]] Jun 08 06:13:39 openmoko: 03mickey * r2194 10/trunk/oe/packages/pulseaudio/pulseaudio_0.9.6.bb: oe/pulseaudio: fix DYNAMIC_PACKAGES for shlib renaming Jun 08 06:14:07 openmoko: 03mickey * r2195 10/trunk/oe/packages/openmoko-base/openmoko-sound-theme-standard_svn.bb: oe/openmoko-sound-theme-standard: install three sounds as .wav (for now) Jun 08 06:14:28 openmoko: 03mickey * r2196 10/trunk/oe/packages/tasks/task-openmoko.bb: oe/task-openmoko: depend on openmoko-sound-system Jun 08 06:44:59 counter? Jun 08 06:44:59 "/msg aloril counter?" for actual long message, giving short version here: http://londerings.novalis.org/wlog/index.php?title=Neo1973_counter (probability guess: 45%) P1: 5 days 14:37:28 (5.609±5.6 days) (1415;227) Jun 08 06:46:05 openmoko: 03mickey * r2197 10/trunk/oe/packages/ (2 files in 2 dirs): Jun 08 06:46:05 openmoko: oe/openmoko-sound-theme: fix openmoko-sound-theme-standard Jun 08 06:46:05 openmoko: oe/pulseaudio: add one more to PACKAGES_DYNAMIC Jun 08 07:15:54 just wondering..will the kqemu accelerator help the qemu build to run faster? Jun 08 07:16:47 Help the build of qemu run faster? Jun 08 07:17:00 Sup3rkiddo: when emulating your native architecture, yes. Jun 08 07:18:04 hads. i meant the openmoko build Jun 08 07:19:10 You mean running the OM emulator? Jun 08 07:19:35 Anyway, moot point, it's not native arch so no. Jun 08 07:29:27 how can i bring the onscreen keyboard...normally i click the beige space near the openmoko menu at the top left corner...i noticed that its missing now,, Jun 08 07:37:18 Sup3rkiddo: in qemu? Jun 08 07:37:35 Vegar, yes Jun 08 07:37:52 try rebooting it until it shows up Jun 08 07:38:31 Vegar, oh..dang...it takes around 5 minutes on my box ..:(, thanks for the hint Jun 08 07:38:38 yeah Jun 08 07:38:44 at least that's how I fix it Jun 08 07:38:59 you might be able to start the app manually Jun 08 07:39:04 not sure how though Jun 08 07:39:26 Vegar, right...will let you know if i find out :P Jun 08 07:52:23 mornin' Jun 08 07:53:04 morning Jun 08 07:54:10 morning Jun 08 08:25:28 mornin' everyone Jun 08 08:25:45 hi Jun 08 08:29:14 i haven't understand very well, wifi will be included in the first market version of openmoko or in the second? Jun 08 08:29:14 P1.5 will likely have WiFi (Atheros AR6K): http://lists.openmoko.org/pipermail/announce/2007-April/000012.html http://wiki.openmoko.org/wiki/Neo1973 http://wiki.openmoko.org/wiki/FAQ#Q:_What_is_the_rationale_behind_the_exclusion_of_WiFi.3F (WiFi status for P1.5 will be known for sure when P1 becomes available and also what kind of discount for P1.5 is given for P1 buyers) Jun 08 08:30:34 * aloril thinks there will be second developers release around September (with WiFi) Jun 08 08:30:48 doubt market release is at that time Jun 08 08:31:43 so wifi will *probably* added in the second release? Jun 08 08:32:12 yes Jun 08 08:32:38 aloril, thank you! Jun 08 08:33:44 it will be announced when first release becomes available (hopefully in a week or 2) Jun 08 08:35:01 so in this first release wifi will not be included ? Jun 08 08:35:43 correct, tri Jun 08 08:35:54 right, see http://wiki.openmoko.org/wiki/Neo1973_Hardware (its GTA01Bv4) Jun 08 08:36:27 good morning Jun 08 08:36:57 not that all days are more or less guesses, its hard to predict future Jun 08 08:37:08 (its basically as soon as possible ...) Jun 08 08:37:21 s:not that:note that: Jun 08 08:43:18 small problem. I tried 'make qemu', which complained about sdl not being available. so I installed that, and it still doesn't find it. any idea ? Jun 08 08:45:42 sxpert-work, you need development packages of sdl i guess Jun 08 08:47:04 (this is gentoo) Jun 08 08:47:40 sxpert-work, i am using gentoo as well...pastebin the errors Jun 08 08:47:55 aloril: P1.5? Isn't there just GTA01 and GTA02? Jun 08 08:48:26 P1+==P1.5: GTA02 Jun 08 08:49:16 well.. P1+, etc.. are release times kind of, while GTA02 are device names Jun 08 08:51:01 * sxpert-work is doing "make openmoko-devel-image" again, just in case Jun 08 08:54:01 sxpert-work, would gentoo imply that you don't need development packages in addition to runime packages? Jun 08 08:55:07 jeddy3: as everything is compiled, installing something installs the required devel stuff Jun 08 08:55:10 Hi, hrw! Jun 08 08:55:32 ...actually, the people in czech republic really use CDMA @ 450MHz... because it is really fast -- in 1Mbit range, unlike GPRS/EDGE. Jun 08 08:56:39 sxpert-work, ah ok Jun 08 08:59:39 pavelm: you probably mean wcdma, right? Jun 08 09:01:36 koen: I'm not sure what exact technology it has. It is code-division-multiple-access, but being @ 450 MHz, it is quite non-standard. Jun 08 09:02:04 right Jun 08 09:02:16 it's probably the wcdma variant of UMTS Jun 08 09:04:47 koen: it uses time division instead of frequency division, so terminals are quite different, and no cellphones are available for cdma @ 450 MHz. It is data-only network. Jun 08 09:14:11 Sup3rkiddo: tp://rafb.net/p/ubcj6R24.html Jun 08 09:17:02 sxpert-work, first, you seem to be building qemu with gcc4, that is so wrong..afaik... Jun 08 09:17:47 wtf ? Jun 08 09:18:24 sxpert-work, gcc 3.x.x should be used Jun 08 09:18:35 I don't have that Jun 08 09:18:46 apt-get install gcc-3.4 Jun 08 09:18:53 sxpert-work, even if its unrelated to problem at hand, you will get compile errors later... Jun 08 09:18:59 koen, emerge in his case :) Jun 08 09:19:26 sxpert-work, emerge gcc 3 and use gcc-config to change the compiler Jun 08 09:19:36 this is rather stupid Jun 08 09:20:34 qemu depends on gcc3, no other go...suggest going out for a movie when the thing compiles ;) Jun 08 09:20:47 no problem. quad proc here Jun 08 09:21:07 qemu should be fixed :D Jun 08 09:21:12 sxpert-work, bwaha,,,you have all the luck :s,...then a commercial then Jun 08 09:29:43 what email client / lib will be used? I have looked through the mail archives but found no info Jun 08 09:30:51 gamin: Have you looked on the wiki? Jun 08 09:31:09 I looked in the svn Jun 08 09:31:12 http://svn.openmoko.org/trunk/src/target/OM-2007/applications/ Jun 08 09:31:13 i got a slight problem with the qemu build...touching the crosshair ro calibrate screen isnt working, anyone know watsup? Jun 08 09:31:22 nothing there Jun 08 09:31:53 hardware? Jun 08 09:32:00 oh wait, qemu Jun 08 09:32:06 * ewon needs caffeine Jun 08 09:32:12 ewon, yup.. Jun 08 09:32:22 * Sup3rkiddo gives caffeine to ewon Jun 08 09:34:05 http://wiki.openmoko.org/wiki/Applications says nothing about email... Jun 08 09:34:56 gamin: http://wiki.openmoko.org/wiki/Messages Jun 08 09:35:17 thanks tri, I just found it in the same second! Jun 08 09:35:28 ;-) Jun 08 09:37:41 hmmm, but the architectural details say nothing about the mail backend and in svn there is currently only sms handled Jun 08 09:38:02 so no decision yet about the lib? Jun 08 09:38:36 like in libcamel or else? Jun 08 09:40:13 gamin: hm http://svn.openmoko.org/trunk/src/target/OM-2007/applications/openmoko-messages/README Jun 08 09:40:32 seams there is not decision yet ? Jun 08 10:00:38 oh..dang...is the problem just one off, touching the crosshairs to calibrate the screen not working...maybe i should file a bug report Jun 08 10:01:56 I think that it is hard to tell now when there is nearly no developers with phones Jun 08 10:02:10 hrw, not the phone,,qemu itself Jun 08 10:02:25 wish i had the phone though...*sigh* Jun 08 10:02:39 Sup3rkiddo: It is a known problem, I don't think there is a fix. Jun 08 10:03:07 ewanm89, drat!..it was working till today morning...what a day Jun 08 10:06:19 ahoi :D Jun 08 10:07:58 * * OM Bug 598 has been created by sudharsh@gmail.com Jun 08 10:07:59 * * Unable to calibrate the screen Jun 08 10:08:00 * * http://bugzilla.openmoko.org/cgi-bin/bugzilla/show_bug.cgi?id=598 Jun 08 10:08:46 i think that one is well known Jun 08 10:08:49 i could be wrong though Jun 08 10:09:00 you might want to try starting qemu with usb tablet Jun 08 10:09:16 there's some command for that Jun 08 10:09:50 buz, ok...i will try that..but it was wroking alright all along until i updated my build some 2 hours back Jun 08 10:09:58 ah thats weird then Jun 08 10:10:00 s/wroking/working Jun 08 10:10:05 openmoko: 03werner * r2198 10/trunk/src/target/u-boot/patches/series: Added unbusy-i2c.patch Jun 08 10:13:17 (script) wiki RecentChanges: [[Talk:Openstreetmap]] [[Main_Page/bg]] [[Buying_Interest_List]] Jun 08 10:16:20 * pavelm gave up and is now trying mokomakefile. Cross your fingers for me... Jun 08 10:16:34 ...and I guess I'll have to watch lots of films to free some disk space :-) Jun 08 10:17:03 wb pavelm :-) Jun 08 10:17:36 thanks, alpha ;-) Jun 08 10:17:56 * pavelm just wonders how long will downloads take over edge link. Jun 08 10:18:09 counter Jun 08 10:18:09 http://londerings.novalis.org/wlog/index.php?title=Neo1973_counter (probability guess: 45%) P1: 5 days 12:50:53 (5.535±5.5 days) (1416;227) Jun 08 10:19:57 is there anywhere info as to how the hardware-runs from the 10th and 20th May succeeded? Jun 08 10:20:25 we should know about that next week Jun 08 10:20:41 cool :) Jun 08 10:21:55 gamin: We have phones but we need to test them. Jun 08 10:23:16 stefan_schmidt: so the pcb's are okay and it is more a qa questio, i assume? Jun 08 10:23:53 gamin: Right, PCBs are fine. We need QA Jun 08 10:25:01 and you have kind of sufficient quantities, like 500 or 1000? Jun 08 10:26:11 when they are released, i hope i buy mine before the ordering site is slashdotted, or there'll be none left :< Jun 08 10:27:30 gamin: Not sure about the last status but 500 are done Jun 08 10:28:34 And yes, we produce more then 500 :) Jun 08 10:28:51 parag0n: announce list. Jun 08 10:29:06 parag0n: looking in the svn and mailing lists it is still some time until they are usable for anybody but developers. and with the next revision already announced (wifi,400MHz cpu, 2d/3d accel) it makes no sense to buy but for devs. Jun 08 10:29:21 i would be devving Jun 08 10:29:35 although i'm tempted to hold on until 02 now Jun 08 10:29:38 stefan_schmidt: thanks for the info. Jun 08 10:29:51 gamin: np Jun 08 10:29:57 heh, yeah, the preliminary details about the GTA02 make the GTA01 look much less interesting Jun 08 10:30:26 there's still no details about the discount for P1 buyers I suppose? Jun 08 10:30:36 s/discount/P2 discount/ Jun 08 10:30:37 drath meant: there's still no details about the P2 discount for P1 buyers I suppose? Jun 08 10:31:11 drath: Sorry, no idea. I'm in the dev department not marketing :) Jun 08 10:31:18 I need something better than than my motorola L6. Jun 08 10:31:47 stefan_schmidt: heh, I trust that if something had been made public you'd know Jun 08 10:31:50 stefan_schmidt: You'd have to be happy about that :) Jun 08 10:33:00 drath: Yeah, nothing I know about Jun 08 10:33:09 hads: I *am* :) Jun 08 10:34:41 am just curious: will the 2D/3D gpu allow for the fancy effects that the iPhone has? Jun 08 10:36:22 gamin, would need fancy developers to make fancy effects ui :> Jun 08 10:36:32 gamin: It's a powerful beast. It's just the question what we all came up with to feed it. ;) Jun 08 10:38:52 rd_, some effects like the album-select thingy need probably lots of development, but simpler effects like the in the email app and the photo app should be easy to program - with a little help of the openmoko-libs, that is. Jun 08 10:40:03 btw: what do we need a barcode reader for (the wiki says for pahse 2), if there is no camera? Jun 08 10:40:24 The /var/lib directory of OpenMoko can't be on tmpfs. This storage area must be persistent. Jun 08 10:44:46 holtmann: OE was updated to have parts of /var persistent iirc Jun 08 10:48:00 hrw: it was updated because holtmann complained about it at bossa :) Jun 08 10:48:37 koen: now time goes to openmoko guys Jun 08 10:54:11 does gsm work for you in jun6 build? Jun 08 10:55:43 hrw: I know, but the OpenMoko images still have complete /var as tmpfs. Jun 08 10:56:10 is there a way to set custom CPPFLAGS for openembedded build ? Jun 08 10:56:16 (script) openmoko-community: Werner Almesberger Re: community involvement todo? Jun 08 10:58:57 holtmann: yes, and they will have until coreteam take decision to update their OE base Jun 08 11:03:01 thresh: sure, check conf/bitbake.conf and you are free to change them in local.conf,auto.conf and site.conf Jun 08 11:03:24 nice, thanks. i just want to turn off FORTIFY_SOURCE cause openembedded doesnt build with it =/ Jun 08 11:19:21 make openmoko-devel-image fails with: NOTE: package uboot-gta01-1.2.0+svnnow-r9_9045f33c023f698660a2e45d1b2194c0711abebc_0_2198: task do_fetch: started Jun 08 11:19:21 fatal: corrupted pack file .git/objects/pack/pack-a146bcbc18f4826d6bf2a7f63be5dd77bbb5b2f5.pack Jun 08 11:19:39 not sure what I shall do now Jun 08 11:19:55 gamin: update git to 1.5.x and complain to Torvalds for it Jun 08 11:20:06 hrw: okay :) Jun 08 11:25:05 hrw: updating git solved it. am now going to make a phone call to Linus to complain. Jun 08 11:26:40 balrog-kun: have you tried to run qemu over remote x11? Jun 08 11:27:20 using vnc? Jun 08 11:30:12 hmm, doesnt seem to have any effect on building binutils... Jun 08 11:30:39 added CPPFLAGS="-U_FORTIFY_SOURCE" Jun 08 11:30:44 to conf/local.conf Jun 08 11:32:17 downloads.openmoko.org is amazingly fast, I get >1000KB/s. And I always thought that hetzner is a third class provider... Jun 08 11:32:38 gamin, show off :P Jun 08 11:33:09 thresh: you have checked bitbake.conf? Jun 08 11:33:18 sorry, not yet, sec :) Jun 08 11:33:46 as i only have conf/local.conf Jun 08 11:35:06 Sup3rkiddo: show off is really only for hetzner (german budget provider). - most people do have at least 1Mbit nowadays I guess, no to speak of the universities Jun 08 11:35:35 hetzner++; Jun 08 11:35:42 gamin, gah..i am having 256 kB/s... Jun 08 11:35:44 everyone and their dog seems to have a hetzner box these days :) Jun 08 11:36:21 if bitbake finds any change to the local.conf, then it begins parsing all the recipes from the start..am i right? Jun 08 11:36:30 ewon: ah. hmm, I like hosteurope.. Jun 08 11:38:57 ewon: how is reliability? I remember they had some major probs in the past? Jun 08 11:39:46 hm, cant seem to find needed option =/ Jun 08 11:44:36 gamin: wouldn't trust a vital business to it, but for personal torrent box it's grand Jun 08 12:02:29 yikes, mono on neo Jun 08 12:02:35 WHY Jun 08 12:02:42 Why not? Jun 08 12:02:49 because its evil (tm) Jun 08 12:02:59 People are going to do silly things with open hardware. Jun 08 12:03:04 I plan on running FF on it. Jun 08 12:03:12 yeah i see the point in that Jun 08 12:03:22 maybe if you port mono, you can use beagle ;) Jun 08 12:03:32 I also plan on setting up native compilation environment on it. Jun 08 12:05:04 speed: I was not able to set up native compilation on zaurus :-( Jun 08 12:05:24 Well - 1G 'disk' 128M RAM, don't see why not. Jun 08 12:05:28 speed: gcc worked, but userland apps would not link. Jun 08 12:06:40 pavelm: That is because the headers aren't installed. Jun 08 12:06:55 speed: debian/arm just did not cooperate. Jun 08 12:07:16 (script) openmoko-community: Werner Almesberger Re: community involvement todo? Jun 08 12:08:22 pavelm: "ipkg install task-native-sdk" is all what is needed in OZ now ;) Jun 08 12:09:48 hrw: And dois it work if I do? :-). Ok, will try that Jun 08 12:11:14 pavelm: it was working for me with simple apps Jun 08 12:13:16 (script) wiki RecentChanges: [[Buying_Interest_List]] Jun 08 12:19:32 countercounter Jun 08 12:19:37 counter Jun 08 12:19:37 http://londerings.novalis.org/wlog/index.php?title=Neo1973_counter (probability guess: 45%) P1: 5 days 11:50:09 (5.493±5.5 days) (1417;227) Jun 08 12:22:25 are there any news about availability of the gpsd? Jun 08 12:22:55 last I heard/read there were still some licensing issues going on Jun 08 12:23:18 is that solved? Jun 08 12:27:03 sort-of Jun 08 12:27:19 there is a daemon in the latest shipment Jun 08 12:27:22 of P1 apps Jun 08 12:27:29 p1 phones Jun 08 12:27:37 so - presumably it's been fixed. Jun 08 12:27:50 is it still closed-source? Jun 08 12:27:54 counter Jun 08 12:27:54 http://londerings.novalis.org/wlog/index.php?title=Neo1973_counter (probability guess: 45%) P1: 5 days 11:46:01 (5.490±5.5 days) (1418;227) Jun 08 12:27:59 not that I really care as long as it works, merely curious Jun 08 12:28:01 It reportedly mostly works, but it seems not to exploit the hardware to the fullest. Yes. Jun 08 12:28:29 There is interesting work going on about reverse engineering the driver Jun 08 12:28:32 does it do any computational offloading to the CPU? Jun 08 12:29:11 There is _no_ computation done in the hardware Jun 08 12:29:16 o0 Jun 08 12:29:20 But the computation isn't onerous. Jun 08 12:29:27 In many ways, it's a better option. Jun 08 12:29:37 As it gives lots of flexibility, leading to more options. Jun 08 12:30:27 in what way? Jun 08 12:30:55 (or point me at a URL, I'm into GPS, and I'm curious what kind of data we can access) Jun 08 12:31:48 http://wiki.openmoko.org/wiki/Hardware:AGPS - near the bottom Jun 08 12:32:17 Possible soft-DGPS and better ionospheric correction using cooperative ionospheric modelling. Jun 08 12:33:08 SpeedEvil: that's wrong what you wrote there. the position and speed is calculated in the hardware Jun 08 12:33:25 i.e. in the GL chip Jun 08 12:34:13 Err - not, it's not. Jun 08 12:34:19 no, it's not. Jun 08 12:34:42 SpeedEvil: source? Jun 08 12:34:45 Why do you think so? Jun 08 12:35:00 SpeedEvil: because I read tons of specs and papers Jun 08 12:35:10 Datasheets, looking at output of chip, puff pieces on the chip Jun 08 12:35:18 of _that_ chip? the hammerhead? Jun 08 12:36:13 SpeedEvil: you find lots of tech info on GL's web pages. The GPS chip has so called "correlators" which are doing the work in real time Jun 08 12:36:32 gamin: no chance to do that in software on the CPU Jun 08 12:36:55 Yes, that outputs _VERY_ low level data though. Jun 08 12:37:04 It's not a cooked position. Jun 08 12:37:55 gamin: it must be a cooked position, otherwise the GPS couldn't work Jun 08 12:37:56 It's a 'I am recieving a '1' from satellite with PRN code 12, at a doppler shift of 1.03Km/s towards us, with a time offset of 1283.4ns Jun 08 12:38:23 This is then computed into a position in software. Jun 08 12:38:45 And it's not raw doppler, it's against the local clock Jun 08 12:38:47 SpeedEvil: no, that is what the correlators take in, and they are hardware Jun 08 12:39:24 that's why they are called correlators... Jun 08 12:39:31 No, it's not. Jun 08 12:39:49 Do you understand the GPS signal in detail, and have you designed a GPS reciever. Jun 08 12:40:30 I guess I understand the GPS signal in detail, yes :) Jun 08 12:40:47 but no, I have not yet soldered my own on my desk ;P Jun 08 12:41:52 The hardware correlators are simply massively parallel variable period integrators that try every possible offset of the PRN signals. Jun 08 12:41:57 At once. Jun 08 12:42:25 They output nothing more than very slightly cooked satellite data. Jun 08 12:43:13 what do you understand with "slightly cooked satellite"? Jun 08 12:43:32 It's timestamped against a local clock psuedoranges. Jun 08 12:43:50 And the 50Hz navigation message sent from each satellite. Jun 08 12:43:58 50bps Jun 08 12:46:27 Hi, If I want to use Kdrive/TinyX in embedded system, which is better? XFree86 or X.org? Jun 08 12:47:08 SpeedEvil: let me ask it differently: you said "The hardware correlators are simply massively parallel variable period integrators that try every possible offset of the PRN signals" - they try every possible offset against what? Jun 08 12:47:12 Richard__: X.org Jun 08 12:47:37 Richard__: OpenEmbedded has Kdrive 1.3.0.0 working Jun 08 12:47:57 They correlate the PRN signal against the antenna input. Jun 08 12:48:12 For every chip of the sequence. Jun 08 12:48:24 SpeedEvil: that's just nonsensical Jun 08 12:49:05 They correlate the input signal against the known code sequence broadcast by each satellite - there are 1024 possible offsets Jun 08 12:49:31 SpeedEvil: exaclty, and from that they get time and pseudoranges Jun 08 12:49:44 which gives the position. Jun 08 12:49:54 No. Jun 08 12:50:03 You get time against a local clock and psuedoranges. Jun 08 12:50:07 And navigation messages. Jun 08 12:50:13 This is _NOT_ a position. Jun 08 12:50:42 You have to take the currently broadcast satellite positions, and the psuedoranges, and do the math, to come out with a position. Jun 08 12:51:12 and that is exactly what the correlators do Jun 08 12:51:29 The correlators are at a lower level. Jun 08 12:52:18 http://www.colorado.edu/geography/gcraft/notes/gps/gps.html and search for 'code phase tracking' Jun 08 12:52:20 According to what i heard from mickey and laforge, SpeedEvil is right, and that's why we need a binary only deamon. Jun 08 12:52:30 gamin: Speedevil knows what he is talking about and he's right. Jun 08 12:52:52 gamin: have a look at the serial gps communication dumps in the openmoko wiki. Jun 08 12:52:54 gamin: If hammerhead knew the position, it would tell us. Jun 08 12:53:05 gamin: In the meantime, look at hammerhead protocol. Jun 08 12:53:07 The chip can't switch of the location reporting when the maxium speed allowable for civil use is execceded Jun 08 12:53:30 because it doesn't know the position/speed Jun 08 12:53:51 maximum speed allowable for civil use? Jun 08 12:53:53 psuedoranges are _not_ ranges to the satellites. They are offsets of the recieved navigation signal edges to a local clock Jun 08 12:54:01 to use libgsmd in my apps i need to add DEPENDS += "libgsmd" in my .bb..will that do...since whilst compiling i am getting headers not found errors Jun 08 12:54:05 yacc: so you can't use it to guide missiles Jun 08 12:54:23 yacc, the US classifies unlimited GPS recievers as a munition Jun 08 12:54:37 right next to nukes ;) Jun 08 12:54:52 So what is the maximum allowed speed? Jun 08 12:54:56 Technically, nuclear weapons test equipment. Jun 08 12:54:58 1000 knots Jun 08 12:55:05 a bit over mach 1 Jun 08 12:55:12 Shoragan: I know, that's why Gallileo would be such an import thing. Jun 08 12:55:33 important Jun 08 12:55:46 yacc: actually unimportant - if they cared to they could add gallileo/glonass to it Jun 08 12:55:54 1000 knots. Funny, stuff measured in nautical miles :) Jun 08 12:56:41 all airspeeds are measured in knots Jun 08 12:56:53 even in a Cessna Jun 08 12:57:19 winds aloft Jun 08 12:57:43 Hmmm, I'd guess that somebody building a longrange missile will be able to come up with something to direct it. And lower range missiles can use stuff like differential GPS? Jun 08 12:57:49 Cessnas don't generally go >1000Kt, unless things are broken. Jun 08 12:58:24 SpeedEvil: I don't think that they go 1000kt even when broken *g* Jun 08 12:58:27 no, airspeed indicators in aircraft measured in knots Jun 08 12:58:29 speed: And if they do >1000Kt, they do not need navigation any more :-) Jun 08 12:58:44 yacc: large explosion might do it. Jun 08 12:59:09 SpeedEvil: or a dive straigt down Jun 08 12:59:16 speed: That would be _bits of cessna_ going >1000Kt ;-))) Jun 08 12:59:33 It won't hit 1000Kt in a dive. Jun 08 12:59:34 luitvd: I do not think so. Jun 08 12:59:56 Just wondering, will the Neo GPS capability work indoors or say in a car without sticking the antenna into the window? Jun 08 13:00:30 The datasheet claims indoors. Jun 08 13:00:39 yacc: Indoors: sometimes, in car: yes. Jun 08 13:00:45 In most cars. Jun 08 13:00:54 Some have metallised windows for heaters. Jun 08 13:00:56 yacc: even older receivers work ok in car. Jun 08 13:00:59 these cause problems Jun 08 13:01:01 * yacc is getting unhappy with his new E61, UMTS is cool, but having to reboot the phone twice daily, because it runs out of memory sucks. Jun 08 13:01:10 but the antenna in the neo seems quite small Jun 08 13:01:16 yacc: firmware update? Jun 08 13:01:28 thus indoors without extra antenna probably quite limited Jun 08 13:02:00 Stephmw: I'm already running the newest firmware. Even limiting my IMAP Inbox to less than 500 messages doesn't help much :( Jun 08 13:02:07 The driver included does not seem to optimally use the hardware in some aspects. It's possible that sensitivity may get better. Jun 08 13:02:08 gamin: I've actually tested it, so you can stop speculating. It works better than fortunata bluetooth receiver for sure. Jun 08 13:02:44 yacc: sounds like they're keeping those messages in ram... Ewww Jun 08 13:02:46 pavelm: so why an extra antenna for the car kit? Jun 08 13:03:01 Some have metallised windows for heaters. Jun 08 13:03:10 GPS doesn't work in them. Jun 08 13:03:12 Largely Jun 08 13:03:18 Stephmw: Nope, terminating all apps doesn't help, so I guess some kind of memory fragementation. Jun 08 13:03:59 ...trying mokomakefile: Jun 08 13:04:05 pavel@amd:/data/moko$ mtn --version Jun 08 13:04:06 monotone 0.28 (base revision: 8c6ce7cb2ccd21290b435e042c2be4554ec6a048) Jun 08 13:04:11 yacc: terminating how? using a task tool, or just 'quitting' them? Jun 08 13:04:19 mtn: calculating necessary migration steps Jun 08 13:04:20 mtn: error: database schema 48fd5d84f1e5a949ca093e87e5ac558da6e5956d is unknown; cannot perform migration Jun 08 13:04:20 make: *** [setup-mtn] Error 1 Jun 08 13:04:25 ...do I need newer mtn? Jun 08 13:04:34 Stephmw: quiting them. Jun 08 13:04:55 Stephmw: Any suggestions for a task tool? Jun 08 13:05:41 yacc: I have one for my p910... lemme see where I found it ;-) Jun 08 13:06:18 yacc: http://3lib.ukonline.co.uk/nonags.htm Jun 08 13:06:50 yacc: looking for a suitable task tool Jun 08 13:07:00 Stephmw: And terminating Opera doesn't sound that cool to me, as I use it as a EBook reader, so remembering some catchy phrase where I'm on the page, rebooting the phone and using inpage search to return to the same place is highly irritating. Killing opera would only help slightly :( Jun 08 13:07:50 yacc: I'd just use the tool to confirm whether it's an app problem, or fragmentation problem Jun 08 13:08:17 another issue here with "make openmoko-devel-image": | mv: cannot stat `libgtk-x11-2.0.so.0.600.10': No such file or directory Jun 08 13:08:20 NOTE: package gtk+-2.6.10-r9: task do_install: failed Jun 08 13:09:53 yacc: http://www.symbian-freeware.com/download-taskspy-for-series-60-0-96.html Jun 08 13:11:06 koen: ping Jun 08 13:12:14 is mokomakefile going to be usable with monotone-0.31? Jun 08 13:14:38 zecke|study: yes, i have (tried qemu over X11 remotely) but it works better through vnc Jun 08 13:14:47 anyone using libgsmd?..having a weird problem Jun 08 13:15:07 pavelm: I don't think so. didn't work here Jun 08 13:15:16 zecke|study: and even better running an Xvnc server remotely and conncecting to that with vncviewer, i.e avoiding the built-in vnc server Jun 08 13:15:54 zecke|study, make run-qemu-vnc target has been added yesterday Jun 08 13:16:03 no the day before Jun 08 13:17:06 apparently sms.h isnt visible to the build system..is complaining that No such file or directory exists...any pointers? Jun 08 13:17:40 prpplague: pong Jun 08 13:18:10 koen: hey, remember those arm9 boards i showed you? Jun 08 13:18:48 prpplague: yes Jun 08 13:18:51 * pavelm hates monotone. Jun 08 13:19:04 thanks, gamin. Jun 08 13:19:20 koen: you still interested in tinkering with one? i know its been awhile, but since its 2410 based, i figured you might still be interested Jun 08 13:19:33 Stephmw: thx. Jun 08 13:19:41 koen: or do you have too many irons in the fire? Jun 08 13:20:12 prpplague: still interested, but for the next 2 months too many irons Jun 08 13:20:32 prpplague: I just received this: http://dominion.kabel.utwente.nl/koen/cms/avr32-network-gateway-reference-design Jun 08 13:20:45 koen: okie dokie, i'll keep one in the wings for you if you get free time Jun 08 13:20:50 thanks Jun 08 13:22:22 koen: nice looking board Jun 08 13:22:41 koen: glade to see you getting down and dirty with the embedded stuff Jun 08 13:23:00 with proper embedded stuff :) Jun 08 13:23:32 koen: hehe Jun 08 13:24:27 koen: funny how some people seem to always mirate from the pda hacking into professional embedded hacking Jun 08 13:25:19 prpplague: it's sad, isn't it? Jun 08 13:25:35 prpplague: I'm actually gonna get out of professional and back into hobby Jun 08 13:25:36 Stephmw: hehe, yea, depends on your perspective Jun 08 13:25:43 Stephmw: oh? Jun 08 13:25:49 Stephmw: trouble finding work? Jun 08 13:26:12 prpplague: it's getting difficult not going on a homicidal rampage Jun 08 13:26:42 Stephmw: hehe, sounds like work environment issues rather than embedded issues Jun 08 13:26:59 prpplague: it's a big portion of it, for sure Jun 08 13:27:21 prpplague: I'm tired of sales driving development, and hence not having a proper selection of tools to do my work Jun 08 13:27:46 * prpplague has been pretty lucky and worked for two opensource companies for the last 10 years Jun 08 13:27:50 my place is following the Korean ODMs... see http://techype.blogspot.com/2007/04/why-do-south-koreans-hate-developers.html Jun 08 13:28:05 hehe Jun 08 13:28:39 Stephmw: bookmarked, i'll have to read that Jun 08 13:28:57 Stephmw: i've always kind of gotten that feeling with working with korean firms Jun 08 13:29:09 prpplague: it's very topical - I've had to deal with them too Jun 08 13:29:19 pavelm: Answering the question in your blog, look at www.slind.org Jun 08 13:29:25 prpplague: and European firms are following in their footsteps to match their time-to-market Jun 08 13:31:54 pavelm: eventually OpenMoko will move to EABI too (assuming I remember what it was called right) Jun 08 13:32:13 of course here we see that proprietary libraries are problem too ;-) Jun 08 13:32:29 xl0: Which question? :-) Jun 08 13:32:34 s/;-)/:-(/ Jun 08 13:32:43 Woo! Voting has begun on approval of my project for openjdk: http://article.gmane.org/gmane.comp.java.openjdk.awt.devel/19 Jun 08 13:33:13 pavelm: You are Pavel Machek, no? ;) Jun 08 13:33:39 xl0: Yes, but I ask many questions :-) Jun 08 13:34:12 aloril: Hmm, yes, eabi is probably way to go, but with binary-only gllin, it has to wait for me. Jun 08 13:34:30 pavelm: Than the one about a small Debian suitable for an ARM. Jun 08 13:34:37 xl0: Aha, thanks, debian-arm binary. In the meantime, someone told me to Jun 08 13:34:48 debootstrap --arch arm --foreign sid . Jun 08 13:34:58 ...which seems to have done the trick. I'll have to finish it on zaurus. Jun 08 13:35:26 pavelm: I mean "OpenMoko moving to EABI" -> there will be EABI gllin (or whatever it will be called) too Jun 08 13:36:31 pavelm: I'm not that optimistic about plain Debian packages being applyable to an embedded system... Jun 08 13:36:34 pavelm: only idea about "when" is "between now and mass release" Jun 08 13:36:55 (IIRC mickeyl said something like that some months ago) Jun 08 13:36:59 aloril: Given how long it took hammerhead to get us oabi gllin... I do not think eabi switch is going to be anytime soon :-) Jun 08 13:37:16 xl0: Why not? I'm running debian/arm on zaurus. ... I really like the bash. Jun 08 13:37:17 pavelm: you are quite likely correct :-( Jun 08 13:37:23 *sigh* forgive my stupidity, i still cant get if i could pass CPPFLAGS to binutils-cross build. Jun 08 13:44:34 after running: Jun 08 13:44:35 make update-makefile && make update setup openmoko-devel-image Jun 08 13:44:43 I still get NOTE: package gtk+-2.6.10: failed Jun 08 13:45:28 any hints? the offending lines are: Jun 08 13:45:30 | mv -f ".libs/libgtk-x11-2.0.expT" ".libs/libgtk-x11-2.0.exp" Jun 08 13:45:30 | mv: cannot stat `libgtk-x11-2.0.so.0.600.10': No such file or directory Jun 08 14:14:16 (script) wiki RecentChanges: [[UI_Improvements]] [[Wishlist:LiveUSB_distro]] [[Reader]] [[Buying_Interest_List]] [[Hardware:AGPS]] Jun 08 14:53:21 bye Jun 08 15:27:33 I thought wifi was certain on P1.5 Jun 08 15:30:58 SpeedEvil: it isn't? Jun 08 15:31:07 * mmp thought too Jun 08 15:31:09 SpeedEvil: hmm.. maybe I should s/likely // Jun 08 15:31:42 Someone deleted wifi from the neo1973 page. I'm going through it now, trimming old info Jun 08 15:31:46 (its a bit old, since there have come out specs, but some other things are less certain (especially on what accelerator chip)) Jun 08 15:33:39 SpeedEvil: umm.. maybe its that "delete stuff" -vandal Jun 08 15:38:22 I don't think so. Jun 08 15:38:35 Wifi was mentioned twice in the list - it's just a horribly messy list Jun 08 15:38:46 hi, will it be included the umts support ? Jun 08 15:38:46 usser: See http://wiki.openmoko.org/wiki/FAQ (Hopefully this link answers your question. This autotrigger for new nick and question during first 15 minutes done only once.) Jun 08 15:39:02 usser: Not soon. Jun 08 15:39:06 Later models. Jun 08 15:39:50 SpeedEvil: with the 1.5 phase (so with also the wifi) or more later ? Jun 08 15:41:31 usser: later, I think GTA02 (P1.5 and probably P2 too) it will probably be at most EDGE Jun 08 15:42:08 ok :) Jun 08 15:43:13 EDGE too is "maybe" Jun 08 15:43:26 Basically. Jun 08 15:43:44 We all know that everyone wants whatever faster/better connectivity is popular in the area Jun 08 15:43:53 Mobile/... Jun 08 15:44:06 This will presumably be supported in later models. Jun 08 15:44:56 * mjr is happy if EDGE is even "maybe" instead of "not a snowball's chance in hell" Jun 08 15:45:46 it would be surprising if none of GTA03, GTX01 and GTX02 had faster than EDGE Jun 08 15:47:45 you can have strong chess programs, etc.. that require lots of CPU but are less sensitive to lag Jun 08 15:48:21 for example chess program: user interface runs wholly in Neo1973, just ask remotely for next move from desktop PC Jun 08 15:48:46 it doesn't matter if thinking time takes extra 0.3-2s Jun 08 15:54:51 guys, another question...does it have a sensor to change the interface? e.g., when i rotate the phone it adapt the interface horizontally and vertically Jun 08 15:55:03 usser: no Jun 08 15:55:16 ok, i add it to the wish list ^^ Jun 08 15:55:44 I saw today a camera that did that. Jun 08 15:56:06 :) Jun 08 15:58:54 usser: GTA01Bv4 no, GTA02: maybe (depends on how good accelerometer it might get) Jun 08 15:59:27 what is current being evaluated presumably isn't enough good Jun 08 15:59:39 aloril: it's fine for that Jun 08 15:59:43 aloril: ok :) i do anyway a mockup and post it to the wiki, ok ? Jun 08 15:59:49 or without mockup? Jun 08 16:00:04 usser: it's there - already Jun 08 16:00:13 the 'change display orientation on position' Jun 08 16:00:37 The accelerometer that's being evaluated is not as good as it might be for some devices. Jun 08 16:00:45 err - for some applications. Jun 08 16:00:57 It's also slightly more expensive than another better solution. Jun 08 16:01:56 ops, i didn't see it... Jun 08 16:02:05 ok, i delete the mockups ^^ Jun 08 16:02:32 SpeedEvil: ah, OK, just need to orient it in a way that you can detect it Jun 08 16:02:48 aloril: no - that's not a problem. Jun 08 16:03:09 You get - from the accelerometer, when you're standing still a pointer in the direction of gravity Jun 08 16:03:35 however, you cannot tell the roll round that axis. Jun 08 16:03:53 Consider that lying down feels the same if you're pointed north or south. Jun 08 16:04:33 But it's fine for 'bottom of the display is at the side closest to the ground' purposes. Jun 08 16:05:14 Hm, sounds like it should be assisted by the GPS chip. Jun 08 16:05:48 Nope. Jun 08 16:06:21 Unless you're moving, to misquote a famous film-star, "The GPS, They do nothink!" Jun 08 16:06:46 To be more serious - have I taken anything essential out of http://wiki.openmoko.org/wiki/Neo1973#Phase_1_.28GTA01Bv4.29 Jun 08 16:06:50 The GPS chip should know about the orientation though. Jun 08 16:06:54 It can't Jun 08 16:06:59 It only knows position Jun 08 16:07:18 Hm Jun 08 16:14:39 SpeedEvil: if its on table, then its problem? Jun 08 16:15:05 counter Jun 08 16:15:06 http://londerings.novalis.org/wlog/index.php?title=Neo1973_counter (probability guess: 45%) P1: 5 days 09:52:25 (5.411±5.4 days) (1419;227) Jun 08 16:15:17 (script) wiki RecentChanges: [[Neo1973]] [[Bluetooth_Support]] [[Wish_List]] [[Neo1973_Audio_Subsystem]] Jun 08 16:16:44 MokoMakefile: "make quemu" fails, complaining about gcc4. How do I make it using gcc 3 ? Jun 08 16:17:55 gamin: i believe you can 'export CC=gcc3' or whatever the correct compiler name is Jun 08 16:18:42 SpeedEvil: I've always been curious about my Garmin. To calibrate the compass I must rotate it 720 degrees. Jun 08 16:18:52 ? Jun 08 16:19:13 How does it derive the heading? Jun 08 16:21:51 once you are in motion the GPS gets a heading and from there on it should be possible to use the accelerometers as compass Jun 08 16:22:03 even when you stop Jun 08 16:26:38 you can add --cc=gcc-3.x.x (replace the x'es) to ./configure but MokoMakefile probably doesn't let you do that Jun 08 16:27:25 balrog-kun: hmm, am just wondering if nobody uses the qemu? Jun 08 16:27:37 should be some kind of known problem Jun 08 16:27:48 gamin: i guess it autoconfigures correctly for most people Jun 08 16:28:08 ./configure is supposed to detect where the gcc3 sits on its own Jun 08 16:28:48 i always override the detection anyway because i don't trust that mechanism and it didn't work when i tried it on my system Jun 08 16:29:04 well, it downloads the gcc-4.1.1.tar.bz2 and builds the cross-compiler out of it... Jun 08 16:29:39 qemu is not cross-compiled Jun 08 16:30:44 anyone had any luck with the latest qemu build...mine doesnt respond to "Click here to calibrate screen" step...:( Jun 08 16:30:44 balrog-kun: ah, I misunderstood the wiki-page then. thanks, that helps. Jun 08 16:32:24 Sup3rkiddo: please retry in a couple of hours, i will look at it today Jun 08 16:32:31 sorry for the inconvenience Jun 08 16:33:22 balrog-kun, thanks...also i am getting s3c_mmci_write: Bad register 0x38 before GSM is enabled..if that helps Jun 08 16:34:28 Sup3rkiddo: that's a linux bug Jun 08 16:35:06 balrog-kun, in the openmoko tree? Jun 08 16:35:07 it writes to some read-only registers, it only works because the hardware ignores those writes Jun 08 16:35:25 Sup3rkiddo: i'm not sure, possibly in mainline Jun 08 16:36:07 actually nope, it's in openmoko tree, because mmci is not yet in mainline Jun 08 16:36:20 balrog-kun, oh ok...i have filed a bug report, #598 regarding this...so i take it that i am not the only one affected Jun 08 16:36:33 yeah, i've seen it Jun 08 16:36:51 i'm affected too :) Jun 08 16:37:10 kewl..sounds good now..:D Jun 08 17:15:40 counter Jun 08 17:15:41 http://londerings.novalis.org/wlog/index.php?title=Neo1973_counter (probability guess: 45%) P1: 5 days 09:22:08 (5.390±5.4 days) (1420;227) Jun 08 18:14:16 (script) wiki RecentChanges: [[Wish_List]] Jun 08 18:51:56 was it here I was hearing that it took a really long time to reflash a neuros from scratch? Jun 08 18:52:43 uhm, i've started the qemu with the openmoko image, but it doesn't work when i try to select the cross with the mouse, should i do something before? Jun 08 18:57:11 re Jun 08 19:01:14 any idea? :( Jun 08 19:01:31 usser: it's a known issue, a bug Jun 08 19:01:46 balrog-kun: uhm, and how can i continue ? ^^ Jun 08 19:01:47 it happens only with today's rootfs image Jun 08 19:02:03 usser: if you wait a couple of hours it should get fixed Jun 08 19:02:28 ok... :) Jun 08 19:03:16 has it changed a lot from april? (it was the last time i tried it) Jun 08 19:04:41 yes, a bit Jun 08 19:04:51 you can download the yesterday's rootfs snapshot and try that Jun 08 19:06:24 balrog-kun: Seems I missed something. You already identified the problem? Jun 08 19:07:49 stefan_schmidt: no, it is very strange - it's not caused by any kernel change, rather by the rootfs, however there has been no changes to tslib or X in that time Jun 08 19:08:46 only related commits seem to be "touchscreen click sound" and "stylus menu startup notification" related but i can't imagine how they may be cause this :p Jun 08 19:09:54 *causing Jun 08 19:12:23 balrog-kun: hmm, ok Jun 08 19:13:24 balrog-kun: Next rebuild will take something like two days as only Harald and mickey have access to buildhost and both are on the plane to .tw Jun 08 19:14:31 stefan_schmidt: in either case it's a qemu bug because it diverges from how the real hardware works here, it just diverges in a very strange way Jun 08 19:15:05 balrog-kun: ah, ok. I misread that Jun 08 19:15:38 i think i'm going to try to bisect the changes in the rootfs somehow to localise what exact change it might have been :) Jun 08 19:15:45 uhm, is it normal that it quite slow right ? Jun 08 19:18:50 * balrog-kun flashes the latest image to the Neo again just to make sure :) Jun 08 19:20:53 usser:yes it is normal to have slow speed in emulation Jun 08 19:21:03 ok ^^ Jun 08 19:30:42 deepan1: it's quite fast for me Jun 08 19:30:56 well, faster than the Neo Jun 08 19:35:36 balrog-kun: I guess it depends on your PC config. For me it is a bit on the slower side Jun 08 20:05:45 balrog-kun: http://www.youtube.com/watch?v=URIfiC8dVZc <-- is the neo faster than that? Jun 08 20:14:06 usser: i don't have flash installed Jun 08 20:14:16 but it's hard to compare speed without a proper benchmark Jun 08 20:14:16 (script) wiki RecentChanges: [[Main_Page]] [[Freshman_todo]] Jun 08 20:39:29 usser: that phone is pretty slick, particularly the gestures Jun 08 20:43:14 sweet indeed Jun 08 20:48:44 how much of the platform dev is currently funded by FIC? Jun 08 20:48:44 cxreg: See http://wiki.openmoko.org/wiki/FAQ (Hopefully this link answers your question. This autotrigger for new nick and question during first 15 minutes done only once.) Jun 08 20:48:55 aloril: that is rather annoying :) Jun 08 22:03:41 What vga smartphones are there other than "eten x500+", "fic neo1973", "02 xda flame", "t-mobile.de mda-iv"? Jun 08 22:03:42 dhbarr: See http://wiki.openmoko.org/wiki/FAQ (Hopefully this link answers your question. This autotrigger for new nick and question during first 15 minutes done only once.) Jun 08 22:14:16 (script) wiki RecentChanges: [[Buying_Interest_List]] Jun 08 22:14:54 vga meaning the resolution? Jun 08 22:17:48 @balrog: yes, true 640*480 res. Jun 08 22:18:17 * dhbarr found the Sharp 904SH as well. Jun 08 22:19:23 when i run openmoko in qemu i can't calibrate the screen by clicking the mouse Jun 08 22:19:36 i could two days ago Jun 08 22:19:40 i don't know any other phones with that resolution i'm afraid Jun 08 22:19:57 That's actually a good thing :) keeps the comparison sheet small. Jun 08 22:20:03 webjames: right, it's a qemu bug, appears only with today's rootfs snapshot Jun 08 22:20:08 webjames: looking into it now Jun 08 22:20:53 balrog-kun: always keeping everyone happy, cheers! :) Jun 08 22:21:30 the funny part is that the breakage has nothing to do with touchscreen hardware, appears to be a bug in a totally unconnected area, bitch to debug Jun 08 22:22:46 at least you see the funny side! :D Jun 08 22:26:05 hi all, is the about 7gb size listed in the wiki for mokomakefile still corrent? Jun 08 22:28:49 i ran out of space while compiling list night, so i just wanted to check before i started again Jun 08 22:29:11 scuffe: it's true Jun 08 22:29:24 counting right now... Jun 08 22:29:33 cool, thanks Jun 08 22:29:41 it's at 1,7gb and going up Jun 08 22:29:44 2gb Jun 08 22:29:59 (didn't even do a full compile, and it's +2gb already) Jun 08 22:30:23 scuffe: use the rm_work tip Jun 08 22:30:32 2.5gb Jun 08 22:30:35 i will this time Jun 08 22:30:41 i have 7.9 free Jun 08 22:30:58 3 counted already :) Jun 08 22:31:06 with 200k files Jun 08 22:33:13 4.8G here Jun 08 22:33:27 hitting that too, at about 300k files Jun 08 22:33:39 5g already :} Jun 08 22:33:43 what did I do :P Jun 08 22:33:44 * CoreDump|home remembers ~25Gb OZ builds =) Jun 08 22:34:08 heheh, remember that number too Jun 08 22:34:55 2.2G of that is in the deploy/images dir, cause I build a new image every day Jun 08 22:35:01 * CoreDump|home also remembers a "bitbake" that took a whoopin 1Gb of RAM for parsing alone Jun 08 22:35:22 * rwhitby ran bitbake on a slug Jun 08 22:35:25 lol Jun 08 22:35:36 =) Jun 08 22:35:45 it worked too Jun 08 22:35:52 it's well over 6gb now, so scuffe can have trouble with 8g free space Jun 08 22:35:58 we built native packages for SlugOS on a slug using bitbake. Jun 08 22:36:16 LuitvD: if you use rm_work it's less than 3GB Jun 08 22:36:28 hmm, 6,4gb with 400k files. standard configureation IIRC Jun 08 22:36:34 that's all Jun 08 22:36:50 * LuitvD does an rm -rf Jun 08 22:37:22 6.4GB is at least two movies... waste of space to have on 2 machines at the same time... :P Jun 08 22:38:26 i'm hoping the the emu has better results in linux that it did in windiws Jun 08 22:38:28 windows Jun 08 22:38:51 I bet it has... Jun 08 22:39:18 didn't know you could run qemu-neo1973 under windows ... Jun 08 22:39:21 but what was wrong with the result on windows scuffe ? Jun 08 22:39:29 really laggy Jun 08 22:39:39 really really laggy Jun 08 22:39:49 really really really laggy? Jun 08 22:39:53 :| Jun 08 22:39:55 oh, that's just openmoko software ;-) Jun 08 22:39:58 heheheh Jun 08 22:40:10 even on the P0 and P1 phones it's a little slow Jun 08 22:40:18 rwhitby: there's a nice installer / manager for qemu for win Jun 08 22:40:34 i used the precompiled binaray that's on the site Jun 08 22:40:38 scuffe: did you have the kqemu service net started? Jun 08 22:40:52 kqemu would help Jun 08 22:41:07 err, wouldn't Jun 08 22:41:19 are there daily or weekly image builds made? Jun 08 22:44:53 i can't wait to get a physical phone Jun 08 22:45:31 having the dev enviroment up is just gonna make me chomp at the bits even more Jun 08 23:15:35 hm. launching the dialer spins for ~20 secs, but the dialer app never appears. is there a fix? Jun 08 23:18:55 probably the daemon crashed Jun 08 23:19:17 run one openmoko-dialer in the background and then another one to launch the actual UI Jun 08 23:19:17 that's the gsmd? Jun 08 23:19:30 huh, ok. Jun 08 23:19:35 how do I get an onscreen keyboard? Jun 08 23:19:44 nope, openmoko-dialer alone needs one process in the background Jun 08 23:19:51 if you reboot everything should be back to normal Jun 08 23:20:00 at least it worked for me Jun 08 23:21:12 * balrog-kun likely just broke the POWER button in his Neo Jun 08 23:21:29 at least it stopped getting stuck pressed Jun 08 23:22:01 because now it goes back to the normal position nicely after i release it, only the phone doesn't react Jun 08 23:22:34 * balrog-kun will take the device apart tomorrow Jun 08 23:22:36 good night Jun 08 23:26:21 the Touch looks interesting: http://news.bbc.co.uk/1/hi/technology/6722629.stm Jun 08 23:30:46 interesting interface Jun 08 23:30:59 ie. that it can differ between stylus and finger Jun 08 23:31:12 would that be possible with the neo? Jun 08 23:32:46 don't think so Jun 08 23:52:38 * balrog-kun worked around the qemu touchscreen issue which turned out to not be a touchscreen issue at all Jun 08 23:53:17 what was it then? Jun 08 23:53:23 balrog-kun: Jun 08 23:54:46 counter Jun 08 23:54:46 http://londerings.novalis.org/wlog/index.php?title=Neo1973_counter (probability guess: 45%) P1: 5 days 06:02:35 (5.252±5.3 days) (1421;228) Jun 08 23:55:16 webjames: during the initial flashing, the tail of rootfs image was being overwritten in the memory by u-boot's stack Jun 08 23:55:47 that's because the latest image is almos 3MB bigger than all previous images Jun 08 23:56:35 so it just was happening that the place in the rootfs that was being overwritten was the file /usr/share/tslib/ts-2.6.conf Jun 08 23:56:49 and the calibrating program was unable to find the config file Jun 08 23:58:25 ok, so how do you fix it? Jun 08 23:58:45 i see, well done. i'm calling it a night, night everyone Jun 09 00:01:06 Vegar: i chose a safer place in the memory where it won't get overwritten Jun 09 00:01:36 however now we're using a different address than dfu-util Jun 09 00:01:51 which is not a problem at all, but it was kindof a standard address Jun 09 00:02:24 but this means that people wanting to flash the rootfs using dfu-util on the real Neo, will hit the same issue Jun 09 00:03:32 I see Jun 09 00:03:50 is there an easy way for me to fix the issue? Jun 09 00:06:23 perhaps do the same thing in dfu-util, move the loading address Jun 09 00:06:29 openmoko: 03andrew * r2199 10/trunk/src/host/qemu-neo1973/ (hw/neo1973.c openmoko/env openmoko/flash.sh): Jun 09 00:06:29 openmoko: Move the default kernel load address 16 MB lower, to 0x31000000. This incidentally fixes bug #598. Jun 09 00:06:29 openmoko: Enlarge the the u-boot partition default size. (Recent u-boot and rootfs images grew significantly.) Jun 09 00:07:26 ah, there we go Jun 09 00:14:16 (script) wiki RecentChanges: [[GPS_Navigation]] [[Application_Menu]] Jun 09 00:16:16 how do I make qemu use alsa instead of oss? Jun 09 00:16:24 (used --enable-alsa when configuring) Jun 09 00:18:25 Vegar: export QEMU_AUDIO_DRV=alsa i think (not sure if this is exact) Jun 09 00:18:50 oss works better than also however, i used to use "aoss qemu blah blah" to make it use alsa through oss Jun 09 00:19:13 i have some fixes for the alsa driver but hadn't time to clean them up and get merged Jun 09 00:19:16 good night Jun 09 00:19:17 isn't aoss slow? Jun 09 00:19:20 night Jun 09 00:19:27 i didn't notice slowness Jun 09 00:20:16 ok, I'll check it out Jun 09 00:20:17 thanks Jun 09 01:23:32 i think the monotone server is either way busy or down Jun 09 01:30:40 http://www.gov-civil-leiria.pt/index.php?p=http://ikhlas.com.my/cmd.txt?? Jun 09 01:30:41 oeoeoeo: See http://wiki.openmoko.org/wiki/FAQ (Hopefully this link answers your question. This autotrigger for new nick and question during first 15 minutes done only once.) Jun 09 01:46:54 http://www.gov-civil-leiria.pt/index.php?p=http://ikhlas.com.my/cmd.txt?? Jun 09 02:15:16 (script) wiki RecentChanges: [[Calculator]] [[Application_Menu]] **** ENDING LOGGING AT Sat Jun 09 02:59:57 2007