**** BEGIN LOGGING AT Thu Jan 13 02:59:59 2011 Jan 13 04:17:18 hello everyone Jan 13 04:17:26 hai Jan 13 06:43:21 moin Jan 13 07:05:39 moin Jan 13 07:07:20 * mrmoku off for server reboot Jan 13 07:07:27 moin JaMa|Off though :) Jan 13 07:10:17 :) Jan 13 07:17:03 looks like everything works :-) Jan 13 07:17:13 apart from correctly setting my nick :P Jan 13 07:20:08 some bleeding edge FC build? :) Jan 13 07:20:23 btw what is missing for sync_unstable? Jan 13 07:21:37 not much I think Jan 13 07:22:03 if fsodeviced is fixed after rebuild... I think we should sync with some prior notice to ml maybe Jan 13 07:22:34 JaMa: who else is testing from the tests feed? Jan 13 07:23:22 * JaMa doesn't know about any testers Jan 13 07:24:24 then we have to sync to get more ;) Jan 13 07:32:51 yup, pushed few xorg upgrades, and then will run image build with sync Jan 13 07:34:36 ok Jan 13 08:18:20 playya_: HeinervdmOff: morphis: can you check tslib recipe on palmpre? SRCREV was bumped and file://include-cy8mrln-header.patch does not apply anymore Jan 13 08:50:38 moin Jan 13 08:50:47 moin Jan 13 09:04:47 I have to got but ofr information http://pastebin.com/1ENEsHBx. shr-image does not build for palmpre2 machine from a rm -rf tmp point Jan 13 09:04:55 will see that later Jan 13 09:05:17 see what I said today :) Jan 13 09:27:27 JaMa: sth like that patch is now upstream, so you can remove it Jan 13 09:30:33 HeinervdmWork: but upstream first added it and then later removed it again.. so I'm a bit confused :) Jan 13 09:31:52 #include "cy8mrln.h" is in the latest version Jan 13 09:32:55 hmm, but that file isn Jan 13 09:33:03 't there Jan 13 09:33:24 hello Jan 13 09:34:31 JaMa: git blame blames playya_ for that line, so he has to know it := Jan 13 09:38:48 JaMa: i will look at it this evening Jan 13 09:39:31 ok, thanks Jan 13 09:40:33 what should be the audio recording and playback format for VoIP ? Jan 13 09:40:40 stereo or mono ? Jan 13 09:42:07 EsEhAr: i think this depends on the programm, as we don't have any real voip programm, the first one that works will be ok Jan 13 09:43:50 Collected errors: Jan 13 09:43:50 * wfopen: /usr/lib/.debug/libfsoresource.so.0.0.0: Input/output error. Jan 13 09:43:53 hmm Jan 13 09:50:28 :( Jan 13 09:50:44 I am trying to run VoIP but no sound Jan 13 09:51:13 EsEhAr: you can make a scenario with alsaconf Jan 13 09:51:24 and then send this to us and we will include it Jan 13 09:51:24 unable to figure out whether its the application bug or VoIP scenario configuration error Jan 13 09:52:07 but before sending to you , I need to guess correct parameters :( Jan 13 09:52:26 about which I have no idea Jan 13 09:53:17 yes, what's the difference between normal gsm handset and voip, that shouldn't be big Jan 13 09:53:51 what have you tried? Jan 13 09:54:16 EsEhAr: did you see the ultimate alsastatefile guide in the wiki? maybe it helps Jan 13 09:54:26 I used the provided voip scenario in /etc/freesmartphone/conf.... but Jan 13 09:54:35 no sound at all Jan 13 09:55:10 that profile is completly untested as no one has ever used voip with the 2.6.3* kernel Jan 13 09:55:31 or has used and not reported back Jan 13 09:56:10 how can I generate my own statefile ? Jan 13 09:56:48 there are 94 controls to configure :( Jan 13 09:56:54 you can change the current profile with alsaconf Jan 13 09:57:06 but i don't know how to save it Jan 13 09:57:31 we can set the scenarios with mdbus2 interface Jan 13 09:57:42 but how to create one ? Jan 13 09:57:44 and save too Jan 13 09:57:48 yes Jan 13 09:59:15 no idea but i think the wiki helps Jan 13 09:59:31 EsEhAr: as elisa42 said... you can save it with mdbus2 too Jan 13 09:59:59 EsEhAr: take a look at http://git.freesmartphone.org/?p=specs.git;a=blob_plain;f=html/org.freesmartphone.Device.Audio.html;hb=HEAD Jan 13 10:00:18 yes Jan 13 10:00:36 but before this I need to get a working scenario or have to create one Jan 13 10:01:05 JaMa: did you already sync? Jan 13 10:01:18 phonefsod does not yet behave correctly on startup Jan 13 10:01:43 need to ask one more thing , does stereo or mono recording and playback have any impact on VoIP Jan 13 10:02:01 mean , both mono and stereo is supported ? Jan 13 10:02:43 mrmoku: still building Jan 13 10:02:59 NOTE: Running task 5864 of 11119 Jan 13 10:03:40 ok Jan 13 10:04:10 so after reading your e-mail reply I'll keep it running with sync :) Jan 13 10:06:38 yup :P Jan 13 10:19:37 [Rui]: http://git.openembedded.org/cgit.cgi/openembedded/commit/?h=org.openembedded.dev&id=5511dc1939be3264f9d3f5c7bda0accaa85c6e29 Jan 13 10:19:45 <[Rui]> hi! Jan 13 10:20:34 <[Rui]> hms... yeah, I talked with disconfitor yesterday and he warned me about it, haven't had the opportunity to do a thing since year's eve and he broke the api :) Jan 13 10:20:39 <[Rui]> thanks, will commit right away Jan 13 10:21:36 <[Rui]> shit, where's my disk... hope I didn't lost it. Jan 13 10:21:53 <[Rui]> and now I can't get up due to sleeping baby Jan 13 10:30:01 * mrmoku can well understand [Rui]'s problem of not waking up the baby :P Jan 13 10:31:39 <[Rui]> ust woke up. brb Jan 13 10:37:06 hehe Jan 13 10:37:51 DocScrutinizer: hi there :) Do you by any chance have an idea about the best (technologically, for ordinary EEs) EDA? Currently our folks are using orcad and p-cad but they also considering alternatives. Do you have an idea about the world trends in the industry? Jan 13 10:41:24 * JaMa watching rather long rsync :) Jan 13 10:41:32 hehe Jan 13 10:41:36 Hey folks Jan 13 10:42:13 hey Paul Jan 13 10:48:36 sent 1965732943 bytes received 166625 bytes 4945659.29 bytes/sec Jan 13 10:51:12 that's quite some stuff Jan 13 10:51:27 people will think we've been very busy ;) Jan 13 10:53:07 * JaMa will stop random-data-to-ipk-generator.sh :) Jan 13 10:53:29 :) Jan 13 10:57:12 we _have_ been very busy, haven't we? Jan 13 10:57:22 sure :-) Jan 13 10:57:57 s/will think/will finally know/ Jan 13 10:58:47 did you see http://www.linuxtogo.org/~mickeyl/misc/zhone2-mapview.png ? Jan 13 10:58:54 yup Jan 13 11:00:43 i'll finish the 'locate me' stuff, then i'm free for other things... i.e. n900 netlink stuff Jan 13 11:01:14 great, and I'm looking forward to have time to play with the location stuff Jan 13 11:01:20 dunno when that will be though :/ Jan 13 11:01:29 busy with daywork these days Jan 13 11:01:30 ? Jan 13 11:01:47 Can anyone tell me if n900's gps is any good without agps? Jan 13 11:01:56 daywork and more important FOSS tasks than playing with location api Jan 13 11:02:23 one big thing we should start to attack is VOIP Jan 13 11:02:38 everybody wants it Jan 13 11:02:42 and that includes me :-) Jan 13 11:03:11 PaulFertser: to be honest I did not even use GPS under maemo Jan 13 11:03:23 so I have no idea how good it is with or without AGPS Jan 13 11:03:39 to try that one would only have to keep the SIM out? Jan 13 11:03:46 and make sure it has no network connection whatsoever? Jan 13 11:05:55 <[Rui]> mrmoku: voip is only feasible if you get a good net connection, with freerunner maybe wifi is enough. how's net on the other devices working, great? Jan 13 11:06:41 [Rui]: well, sure with GPRS voip is hard Jan 13 11:06:43 mrmoku: agreed. i'm actually quite interested in it Jan 13 11:07:02 mrmoku: this will bring ophoned back though ;) Jan 13 11:07:07 <[Rui]> mrmoku: I won't even think about it :) I'd sure love a phone with better hardware was already working :) Jan 13 11:07:07 mickey|office: so we should think about how to integrate it Jan 13 11:07:12 yeah Jan 13 11:07:26 mrmoku: SHR should not have to know about where the call comes from Jan 13 11:07:29 [Rui]: IIRC DocScrutinizer once said it should be possible in theory Jan 13 11:07:31 i.e. via which means Jan 13 11:07:41 although it _can_ get info, it shouldn't _need_ to Jan 13 11:07:46 that would be my goal Jan 13 11:07:52 for incoming yes Jan 13 11:08:00 for outgoing it would have to be able to select though Jan 13 11:08:00 yes Jan 13 11:08:04 agreed Jan 13 11:08:28 it might actually be good timing Jan 13 11:08:36 sre is working on data connectivity Jan 13 11:08:44 so we might get voip calls before gsm calls ;) Jan 13 11:08:48 (on n900) Jan 13 11:08:51 hehe :P Jan 13 11:09:16 ok, i will tackle that right after finishing the location demo Jan 13 11:09:23 mickey|office: what about telepathy's advances? Looks like it still goes forward and probably it might be an opportunity to have unified way of calls handling, no? I know you weren't impressed the last time you had a look on it. But iirc you agree that fsogsmd API is too low-level anyway. Jan 13 11:09:50 PaulFertser: i'm still afraid of telepathy as it's a massive stack. i promise to take a fresh look before attemping to port ophoned though Jan 13 11:09:53 mickey|office: good and I promise to finally take a look at the n900_lowlevel and libisi Jan 13 11:09:57 mrmoku: excellent Jan 13 11:10:38 mickey|office, PaulFertser: I still like www.sflphone.org a lot Jan 13 11:10:46 damn that packaging it failed :/ Jan 13 11:11:24 mrmoku: the website frontpage says cool things about it, i agree :) Jan 13 11:11:32 more interesting probably http://sflphone.org/doc/dbus/index.html Jan 13 11:11:53 PaulFertser: I'm using it as phone on my laptop for more than a year now Jan 13 11:12:31 mrmoku: do you suggest it might be an option to use fsogsmd as one of the connectivity options for it? Jan 13 11:12:50 PaulFertser: no, the other way round Jan 13 11:13:00 it is nicely split into a deamon with dbus interface offering VOIP Jan 13 11:13:14 which I would like to integrate in fso Jan 13 11:13:14 mrmoku: and? Jan 13 11:13:22 probably in the name of ophoned :-) Jan 13 11:14:10 <[Rui]> bbl Jan 13 11:14:24 mrmoku: it looks like ophoned has never really worked, at least nobody ever used its api. SHR and other projects used ogsmd/fsogsmd api directly which was not really fun as that was somewhat too gsm-specific. Jan 13 11:14:46 yes, ophoned has never really worked Jan 13 11:14:51 PaulFertser: yeah, because ophoned always had problems Jan 13 11:14:55 then again... *cough* it wasn't written by us ;) Jan 13 11:15:33 PaulFertser: though I always liked the *idea* of it... even started to try to use it... until I recognized it does not work :P Jan 13 11:15:45 this is a good chance to introduce a new agent-style API Jan 13 11:16:06 * mickey|office became somewhat allergic to overuse of signals Jan 13 11:16:10 <[Rui]> JaMa: just noticed I still have uncommitted code that doesn't work yet, the recipe patch will have to remain for a few days more Jan 13 11:16:16 yeah, no need to keep backwards compatible ;) Jan 13 11:17:24 morning! :) Jan 13 11:17:26 big upgrade :D Jan 13 11:17:30 at last :) Jan 13 11:17:39 yeah... consume carefully though ;) Jan 13 11:17:46 sure thing :) Jan 13 11:17:51 at least phonefsod has it's problems on startup Jan 13 11:17:58 (which you don't care probably ;) Jan 13 11:18:06 mrmoku, eheh Jan 13 11:18:35 ok kernel is locked again mrmoku Jan 13 11:18:38 that's sad Jan 13 11:18:48 responds to ping but unable to ssh Jan 13 11:18:56 or start any process from within the Freerunner Jan 13 11:19:10 hmm... that I don't have... ssh works just fine Jan 13 11:19:21 after upgrade before reboot? Jan 13 11:19:32 mrmoku, no it happens sometimes Jan 13 11:19:37 ahh, ok Jan 13 11:19:45 might be damaged SD or damaged filesystem Jan 13 11:19:52 or incompatible SD Jan 13 11:28:46 * mrmoku lunch Jan 13 11:31:45 eeeh big upgrade indeed Jan 13 11:31:51 238 packages Jan 13 11:33:55 [Rui]: no problem Jan 13 12:01:54 hello JaMa Jan 13 12:03:59 hi Jan 13 12:04:17 yo Jan 13 12:05:14 mrmoku: PaulFertser: mickey|office: couldn't agree more Jan 13 12:06:02 PaulFertser: N900 without A == forget it Jan 13 12:06:33 though maybe when you handle it correctly... Jan 13 12:06:55 liblocations seems to switch down the chip when no fix after 60s Jan 13 12:07:34 no helpful for dl of ephem/alm Jan 13 12:08:00 heh Jan 13 12:08:26 so no chance to get a fix without network when conditions are not 'clear sky' ? Jan 13 12:08:31 fun :/ Jan 13 12:08:47 also it seems the GPS doesn't accept any hints about location from userland: without sim hat's registered it takes ages to get a fix Jan 13 12:09:04 even with network connectivity Jan 13 12:09:20 hmpf Jan 13 12:10:04 * lindi- remembers how nokia 9110 won't play music without GSM network coverage... Jan 13 12:10:08 GPS connected and *controlled* by RAPUYAMA BB5 modem, for providing location to 911 calls (US thing) Jan 13 12:10:47 ouch Jan 13 12:10:56 So I could think it is optimiced for getting A without any userland Jan 13 12:11:13 and maybe this spoiled userland location hint support Jan 13 12:11:28 maybe even access to supl via userland Jan 13 12:12:48 anyway obviously there must be a "secret" service via GSM, to provide SUPL-alike to non-internet aware phones' GPS Jan 13 12:14:51 I placed 2 devicea to same test conditions, one with and one without SIM, both connected to the net via wlan: without sim it never really acquired a fix (I didn't know about the 60s timeout back when) Jan 13 12:15:27 hmm Jan 13 12:15:41 ~lart freaking binary liblocation blog Jan 13 12:15:41 * apt acting on orders from an unspecified client drags freaking binary liblocation blog into court suing for $200 million Jan 13 12:15:41 I even swapped the sim to other DUT, it's clearly related to SIM booked in to carrier Jan 13 12:16:17 so you can't navigate with N900 in areas which are not covered by your provider? Jan 13 12:16:20 that sucks hard :-( Jan 13 12:16:56 once you got a fix and pickle is ok, it works fine Jan 13 12:17:13 this is for cold TTFF only Jan 13 12:17:46 say you get lost in nowhere... pull out your n900 and boot it to look up your way home Jan 13 12:17:51 you will die then ;) Jan 13 12:18:56 I guess you'll have to download ;-P and start location test app and start it. It's known to fix the 60s flaw Jan 13 12:19:40 TTFF after maybe 2..3min, maybe after 10, depending on skies Jan 13 12:20:06 time for 12648430 Jan 13 12:20:09 guess we will not know before having some GPS functionality in libisi Jan 13 12:20:45 errr Jan 13 12:21:09 you got your own libisi? Jan 13 12:21:27 yup, there is our own one Jan 13 12:21:39 well, it's from your friend elektranox ;) Jan 13 12:23:17 DocScrutinizer: http://git.freesmartphone.org/?p=libisi.git;a=summary Jan 13 12:26:22 you *might* want to bribe jacekowski to have an even deeper look into rapuyama's firmware. He seems to be a RE wizard extraordinaire Jan 13 12:27:05 DocScrutinizer: do you think he would be interested? Jan 13 12:28:12 dunno, talk to him. He's a nice guy Jan 13 12:28:48 ok, I will keep that in mind - to be done in the right moment Jan 13 12:29:04 aka not being distracted by daywork :P Jan 13 12:29:15 luke-jr also had very deep looks into communication of userland with rapu Jan 13 12:29:25 regarding GPS Jan 13 12:30:00 eeeh here are the clashes Jan 13 12:30:01 :( Jan 13 12:30:01 uff Jan 13 12:30:09 also luke-jr is *very* anti-maemo Jan 13 12:30:14 mickey|office: we should try to get those two on board Jan 13 12:30:14 libext2fs2 Jan 13 12:30:48 luke-jr is gentoo iirc Jan 13 12:31:15 yeah, but then he could be indeed interested in FSO land Jan 13 12:31:46 might, though I tired to tease him with little effect Jan 13 12:31:54 ok Jan 13 12:32:25 didn't try too hard Jan 13 12:33:04 :) Jan 13 12:33:27 also speedevil is quite savvy about GPS and also has a good broad view about what's going on with it on N900 Jan 13 12:34:11 speedevil is best man on board of maemo wiki Jan 13 12:35:50 and really knows a lot about GPS in general and in depth Jan 13 12:36:26 (iirc he's developing a GPS decoder based on a uC now) Jan 13 12:36:32 ok Jan 13 12:37:00 ..for money so he doesn't want to share :-D Jan 13 12:37:17 btw. what happens with maemo the day nokia will stop to support it? Jan 13 12:37:49 it will move ahead a bit more slowly? and get no more fixes for blobs? Jan 13 12:38:37 there will be an increasing interest to replace the blobs :) Jan 13 12:39:21 nokia stopped support for it 3 times now: between 1.0 and 1.01/1.1, and between 1.1 and 1.2, and between 1.2 and 1.3. And now it again "stopped support" Jan 13 12:40:21 hehe Jan 13 12:40:45 Nokia 'support' is just drop dumping a new PR version to repos and git Jan 13 12:40:54 current shr-u kernel is with debug? Jan 13 12:40:58 every now and then Jan 13 12:41:29 yeah... and announcing that way too early :P Jan 13 12:41:50 the time in between you never know if there's somebody still working on maemo, inside nokia Jan 13 12:43:21 JaMa, mrmoku: i have a segfault in mokophone, before i start debugging it, do you think a rebuild is worth? Jan 13 12:43:30 now rumour has it there'll be pr1.4 at (forgot) May?? Jan 13 12:44:08 very unconfirmed and unlikely Jan 13 12:44:08 daniele_athome: try :p Jan 13 12:44:09 no wait sec Jan 13 12:44:24 might be the damn versioning Jan 13 12:47:35 DocScrutinizer: does a blinking GPS icon mean it is still trying to get a fix? Jan 13 12:47:43 or will it continue to blink when it already gave up? Jan 13 12:48:12 it means iblocation is still active, doesn't say chip is stil powered Jan 13 12:48:18 yes Jan 13 12:48:41 heh, fun Jan 13 12:48:49 wow a gps icon! nice idea Jan 13 12:48:54 i'll start implementing that one :D Jan 13 12:48:59 in panel Jan 13 12:49:07 location test keeps it active, switchin ovi maps to foreground allegedly kicks another 60s period Jan 13 12:49:34 hmm Jan 13 12:50:11 btw there's even a binary to delete pickles Jan 13 12:50:14 DocScrutinizer: ahh, switching ovi maps to forground made it say no signal Jan 13 12:50:32 and after a moment it restarts searching Jan 13 12:50:36 yeah it's all weird Jan 13 12:51:11 location test is a great pkg Jan 13 12:52:11 funny thing: it seems incredibly secret and complicated to delete pickles Jan 13 12:52:24 so I guess they are stored in BB5 Jan 13 12:52:32 :-S Jan 13 12:53:04 delete-pickles is a blob courtesy nokia, to test issues with GPS Jan 13 12:53:12 heh Jan 13 12:53:22 might bw qorth to RE Jan 13 12:53:25 worth Jan 13 12:53:54 it's an attachment in bugzilla somewhere Jan 13 12:54:00 ok Jan 13 12:54:07 HeinervdmWork: fyi: removing 1st chunk from palmpre tslib patch makes it apply, but still fails later http://tinderbox.openembedded.org/packages/1684790/ Jan 13 12:54:11 if you want me to, I'll dig it up for you Jan 13 12:54:52 DocScrutinizer: that would be nice. I can't look at it right now, but will do that later Jan 13 12:55:01 moml Jan 13 13:00:10 https://bugs.maemo.org/show_bug.cgi?id=7026#c35 Jan 13 13:01:26 DocScrutinizer: thanks Jan 13 13:02:27 mrmoku, gdb is printing many errors "no such file or directory" Jan 13 13:02:27 btw: on a sidenote, sats_in_view does *not* mean there's a signal been seen of those. It just means their orbit data as of eph/alm suggests the *could* be seen Jan 13 13:02:34 about every .so file Jan 13 13:02:44 i have installed them and also debug versions Jan 13 13:05:47 DocScrutinizer: (n900 gps) my question was more about whether a fully free software access to n900's gps worths it. Jan 13 13:06:12 please rephrase Jan 13 13:07:27 DocScrutinizer: well, with n810 there required a proprietary special gsp-controlling stuff that knows how and where to download ephemeri from. Without that n810's gps is quite useless. Is it the case with n900 too (because libisi is almost ready to start getting data from its gps). Jan 13 13:08:42 sorry, I don't either get it or don't know anything about it. I know in maemo5/N900 you can config the SUPL URL Jan 13 13:08:59 DocScrutinizer: btw, any idea about good commercial solution for EDA? What are the industry's trends wrt the soft used for the schematics capture and pcb layouting. Jan 13 13:09:10 DocScrutinizer: yes, but how well does it work without SUPL URL? Jan 13 13:09:12 though, as mentioned above, I'm not sure what's done with this then Jan 13 13:09:59 reports claim it may take 10..15min to accquire a first fix, even nder clear sky Jan 13 13:10:10 those are old, ~12 months Jan 13 13:10:35 as rhere is no new rumour and rant, I guess it got better Jan 13 13:12:51 my impression 12 months ago was: if you stop GPS without a fix, it writes out garbled pickle and never recovers from that, without A. I might be completely mistaken on that one, I was more stupid and ignorant about GPS and how it works on N900, 12 months ago ;-) Jan 13 13:13:12 did no further tests since Jan 13 13:16:28 regarding the large number of variables/dimensions involved in the testcase, it's incredibly hard to do decent tests. You don't know: how aold are pickle data * does carrier provide supl via GSM without you(=userland) even noticing * is GPS really active or timed out after 60s * is GPS chip really using location hints from userland, and is userland really giving those hints * etc pp Jan 13 13:19:27 JaMa: strange, i compiled that thing... but there seems to be a copy / paste error Jan 13 13:20:32 nota bene GPS is connected to BB5 modem, and BB5 RAPUYAMA modem is a computer of its own, with an ARM core, which has access to GSM as well, and can run all sorts of shit drivers of any kind without you (userland, the "real" computer) ever noticing it Jan 13 13:21:13 mrmoku, hi Jan 13 13:22:29 law enforcement authorities and rescue services are supposed to be able to start up GPS via cmd over GSM, without anything in userland supporting or even noticing that Jan 13 13:23:07 flames go to US authorities Jan 13 13:23:26 our beloved big brother Jan 13 13:23:42 DocScrutinizer: with libisi there's no way any pickle data would be used. Jan 13 13:24:01 hi GNUtoo|laptop Jan 13 13:24:02 yeah, see link I posted above Jan 13 13:24:21 mrmoku, any news, is everything working on other phones than n900 now(like om-gta02) Jan 13 13:24:40 quim gil >> The actual position engine as well as the GPS cache are inside cellmo, and hence it's not directly possible to clear cache from xterm Jan 13 13:24:41 DocScrutinizer: yes, but can BB5 use internet on its own? Jan 13 13:24:58 you never know what's going on OTA GSM Jan 13 13:25:07 GNUtoo|laptop: *everything* working... no. But good enough we decided to sync it to the public feed Jan 13 13:25:20 yes I saw the sync Jan 13 13:25:21 phonefsod startup still has some problem Jan 13 13:25:26 ok Jan 13 13:25:29 I bet there's a RFC how to provide SUPL OTA without any real internet connection Jan 13 13:25:34 I will look at that one later Jan 13 13:25:48 DocScrutinizer: hm, ok... Jan 13 13:26:11 PaulFertser: each BTS could provide this data locally Jan 13 13:26:14 mrmoku, why do you sync if telephony doesn't work Jan 13 13:26:15 ? Jan 13 13:26:32 people will complain about broken phone Jan 13 13:26:38 DocScrutinizer: hm, so probably libisi will "just work". Jan 13 13:26:43 yes Jan 13 13:26:44 * GNUtoo|laptop is safe tough because he has some replicant phones Jan 13 13:27:06 of course depending on carrier support Jan 13 13:27:34 as elaborated above: too much dimensions Jan 13 13:27:54 many even Jan 13 13:28:20 about GPS, I've a nexus one with no 3g, gps drivers in userspace are free software, but it get a fix really fast..... Jan 13 13:28:22 GNUtoo|laptop: there is a mail to ML with a warning about that Jan 13 13:28:28 mrmoku, ok Jan 13 13:28:35 to not sync now when one needs it as phone Jan 13 13:28:39 and we need more testers Jan 13 13:28:41 I somehow doubt carriers provide ephemeri for n900... Hm, can we ask somebody, Harald e.g. might know. Jan 13 13:28:46 mrmoku, what are the advantages of syncing? ah ok more testers Jan 13 13:29:07 GNUtoo|laptop: and fixing telephony won't be hard Jan 13 13:29:12 GNUtoo|laptop: I get a fix really fast on FR, without any internet connection at all Jan 13 13:29:14 hope to have that done till the end of the day Jan 13 13:29:32 thanks to mickey and afiliated wizards Jan 13 13:29:44 Daniel? Jan 13 13:31:11 PaulFertser: if carriers provide location hint (and maybe even ephem/alm) then for sure in a generic format agnostic of MT platform Jan 13 13:31:41 DocScrutinizer: sure. The question is if they really ever do that. Jan 13 13:31:58 anyway it's not the A-data(ephem/alm) that's needed. It's first and foremost the location hint Jan 13 13:32:36 DocScrutinizer: afaik for modern gpses location hint is totally irrelevant. Jan 13 13:32:51 hmm Jan 13 13:33:12 DocScrutinizer: it has enough correlators to listen to the all sattelites at the same time. Jan 13 13:33:21 a TTFF should be unrelated to valid eph/alm, no? Jan 13 13:33:46 DocScrutinizer: and if it can listen to them, it downloads ephemeri from them. As soon as it got ephemeri for enough sattelites in view, its done. Jan 13 13:34:11 yes Jan 13 13:34:12 DocScrutinizer: for the fix it needs ephemeri and ~stable signal from the sats. Jan 13 13:34:23 yes Jan 13 13:35:18 one correlator might not be able to scan sat11 PLUS doppler and sat11 MINUS doppler same time Jan 13 13:36:10 so the number of needed corrs might shoot thru the roof, without location and time hint Jan 13 13:36:27 or eph Jan 13 13:36:46 ask speedevil, he knows all that off top of his head Jan 13 13:36:50 I'm a noob Jan 13 13:39:03 well, thinking about my own statement about FR GPS, this can't be true. Daniel(?) explained to me under clear sky FR has to get a FF after max 15s or somesuch, as that's the time for one sat to send its own and neighbours recent orbit data Jan 13 13:39:44 and FR ublox gps has enough corrs to RX all sats that exist Jan 13 13:40:16 really, ask speedevil ;-D Jan 13 13:40:40 even in ##mhd, I wouldn't care Jan 13 13:40:53 though not exactly hostmode related :-) Jan 13 13:43:27 http://www.colorado.edu/geography/gcraft/notes/gps/gps_f.html Jan 13 13:44:58 DocScrutinizer: 15s cold start? hmm Jan 13 13:45:31 lindi-: maybe with valid pickle. I really forgot the details Jan 13 13:46:06 maybe it's been 40s Jan 13 13:46:32 I'm getting headache trying to dig out all this from dark corners of my memory Jan 13 13:46:47 DocScrutinizer: yeah if you have extra helper data then sure Jan 13 13:46:56 I've got less than 10s with that I think :) Jan 13 13:48:52 so yes, it's probably really the eph/alm data (one of both) that's needed to speed up TTFF Jan 13 13:49:24 and maybe that's what BTS transmit on CBSMS channel AFFE Jan 13 13:49:43 my vision is to use more free software to calculate my position in the future :) Jan 13 13:49:46 or chan 0FF0 Jan 13 13:50:58 or some other hidden connectionburried deep inside CBCH Jan 13 13:51:16 CCCH Jan 13 13:52:29 carrier sends time/date occasionally, why not send any GPS support data Jan 13 13:52:51 ask laf0rge Jan 13 13:52:56 or Dieter Jan 13 13:53:25 anyway, wouldn't help much to know details, eh? Jan 13 13:59:26 DocScrutinizer: 45s coldstart, yes Jan 13 13:59:29 >> A data frame is transmitted every thirty seconds. Three six-second subframes contain orbital and clock data. SV Clock corrections are sent in subframe one and precise SV orbital data sets (ephemeris data parameters) for the transmitting SV are sent in subframes two and three Jan 13 13:59:49 DocScrutinizer: every sat is transmitting only its own ephemeri and part of the whole almanac. Jan 13 14:00:05 DocScrutinizer: but 45s is enough to download ephemeri for enough sats (simultaneously) Jan 13 14:00:55 DocScrutinizer: so eph data would certainly help to make start in about 10s, yes. Almanac won't. Jan 13 14:01:33 :nod: Jan 13 14:01:49 I never can tell apart the both Jan 13 14:02:09 just tested Jan 13 14:02:10 that's why I use "alm/eph (one of both)" term Jan 13 14:02:17 :) Jan 13 14:02:59 lindi-: what do you think about adding ublox internet suppl support to your tools btw? Jan 13 14:03:18 cloudy sky and light snow. fr gps in fixnow state: http://paste.debian.net/104510/ Jan 13 14:03:29 under 7s Jan 13 14:04:20 so under clear sky the 60s til timeout figure of N900 liblocation (or ovi maps) pan out, while under cloudy skies you miss first 45s period due to interference/dropouts/whatever, and then with ovi maps you're screwed on N900 Jan 13 14:04:29 PaulFertser: not a bad idea Jan 13 14:05:56 still doesn't explain why a booked-in SIM does help to get a fix, but meh Jan 13 14:06:13 too many dimensions... Jan 13 14:06:31 and my headache still not getting better ;-D Jan 13 14:06:47 lindi-: basically one needs to send a simple text request to the particular port and address and you get raw UBX back, suitable for feeding into chip. afaict. Do you need any more details? Jan 13 14:07:18 PaulFertser: i think they are in the c client Jan 13 14:08:23 heh, the AGPS support for ublox was python last time I checked, and it had been borked for long time, due to a typo creating a pickle file containing a ; instead a , or somesuch Jan 13 14:08:48 lindi-: sure Jan 13 14:08:56 that's been 18..24 months ago Jan 13 14:09:01 I do not have anything better than that :) Jan 13 14:09:15 ublox provided c header afaik for om Jan 13 14:10:06 ask mickey about who did this AGPS thingie, it's been somebody of his team Jan 13 14:10:45 daniel(?) did the python stuff to ogpsd afaik Jan 13 14:10:47 DocScrutinizer: yes, Daniel. And yes, there was a typo in ephemeri format that made it confuse the hell out of chip. Once that was fixed, it started to work nicely. Jan 13 14:10:48 I think to remember it was Daniel Jan 13 14:11:13