**** BEGIN LOGGING AT Sat Jan 08 02:59:57 2011 Jan 08 07:31:16 pespin: yeah, feel free to do some cleaning :-) Jan 08 07:31:56 pespin: and ping me if you're level of trac admin is not sufficient Jan 08 11:59:32 does anyone know where I could download lowlevel_foo-gta01bv4-1.3.0+git20071214+svnr3648-r1.bin and u-boot-gta01bv4-1.3.0+git20071214+svnr3648-r1.bin ? Jan 08 11:59:47 trying to follow http://wiki.openmoko.org/wiki/Neo1973_Debug_Board_v2/Unbricking but can't get the files :) Jan 08 12:01:20 heyho Jan 08 12:03:22 morning morning Jan 08 12:03:40 mickeyl: moin Jan 08 12:03:46 mrmoku: one of these days it might be handy if I had access to the SHR git Jan 08 12:03:56 any idea on PDP? Jan 08 12:04:00 onen|openBmap: ! Jan 08 12:04:07 onen|openBmap: happy new year, my friend. all the best for 2011. Jan 08 12:04:08 mickeyl: send me ypur key then Jan 08 12:04:10 onen|openBmap: how are things? Jan 08 12:04:22 mickeyl: frohe neues Jahr du Alter ;-) Jan 08 12:04:25 hehehe Jan 08 12:04:35 mrmoku: will do Jan 08 12:05:06 and as our git has access per repo a list of repos Jan 08 12:05:19 mickeyl: I wish to you and your lady all the best (health being the first) for the new year! Jan 08 12:05:45 onen|openBmap: thanks a lot. greetings back from Sabine Jan 08 12:05:52 mickeyl: well I switch job a year ago, from research to software development (java) Jan 08 12:06:01 oh! Jan 08 12:06:03 mickeyl: so I think I have learned quite a lot already Jan 08 12:06:21 that sounds good Jan 08 12:06:24 is it fun? Jan 08 12:06:31 mrmoku: ah, i see. which server are you using? Jan 08 12:06:47 gitosis Jan 08 12:06:57 mickeyl: what makes me look at my old code with some kind of "curiosity" ;-) Jan 08 12:07:12 mickeyl: it is yes. started with existing code maintenance + small changes Jan 08 12:07:19 mickeyl: pushed some rework and refactoring Jan 08 12:07:48 mickeyl: start a new project, everything from scratch, with a good spirit team, willing to do things right, and with modern methods Jan 08 12:08:30 onen|openBmap: excellent. that sounds like a good environment to work with. did you have to move for that or are still in the vincinity of Paris? Jan 08 12:08:39 mickeyl: and I try to work on a more regular basis on my obm client. not far from a release Jan 08 12:09:06 mickeyl: does moving to the building 200 m far from where I did work count as a "move" ? ;-) Jan 08 12:09:20 mickeyl: still the same company Jan 08 12:09:30 mickeyl: another entity of the group Jan 08 12:09:44 hehe, right. cooll Jan 08 12:10:01 mickeyl: yes, big companies do not have only bad sides Jan 08 12:10:08 mickeyl: also advantages Jan 08 12:10:19 mickeyl: and you how are things going? Jan 08 12:10:38 mickeyl: working as freelance? Jan 08 12:10:50 mickeyl: hired by nokia to work on ofono? Jan 08 12:12:03 fairly well in general. i had a very busy 2010 as me and my colleague have been building up our mobile application development company (focusing on Apple iOS, but also targeting Android with freelancers). I could spend very little time with FSO. This has now changed a bit and i try to sneak on more time, especially we have some promising progress on new platforms (n900, palm pre). Jan 08 12:12:33 so FSO has stalled a bit in 2010, but will gain more speed in 2011 Jan 08 12:13:09 and some random private notice... Sabine is pregnant in the 5th month... which makes for an adventurous future as well ;) Jan 08 12:31:58 freesmartphone.org: 03mickey 07cornucopia * re24133e8834e 10/fsotdld/src/plugins/provider_location_gpsd/libgps.vapi: fsotdld: bind more of libgps Jan 08 12:34:06 mickeyl: so... any idea about PDP ? :-) Jan 08 12:34:27 did you manage to get a BT? Jan 08 12:34:47 well Jan 08 12:34:52 before you do that i can try on my cinterion Jan 08 12:35:05 which is pretty close to the calypso Jan 08 12:35:31 let me finish the vapi first Jan 08 12:35:36 mickeyl: (bt) yeah, even posted it here :) Jan 08 12:35:58 it smells very much like a problem in glib though :/ Jan 08 12:36:20 let me scroll up Jan 08 12:36:47 http://shr.pastebin.com/SJ6ukBJm Jan 08 12:37:09 hmm :/ Jan 08 12:37:14 ok Jan 08 12:37:23 i will take a look asap Jan 08 12:37:26 ok Jan 08 12:37:56 and trying on something different than armv4t might indeed be interesting Jan 08 12:38:05 to see if it's a glib armv4t only problem Jan 08 12:38:42 yes Jan 08 12:38:49 my cinterion is connected to the PC Jan 08 12:38:58 so i'll try on x86 first Jan 08 12:40:40 good Jan 08 12:40:51 Hi all Jan 08 12:41:01 morning GarthPS Jan 08 12:42:31 congrats and good luck with your forthcomming april fish mickeyl! :p Jan 08 12:43:16 ;) thanks. schedule is mid-may though, expect me to have _very_ few time for FSO in the first months ;) Jan 08 12:43:35 that's why i plan to move it forward by a huge bit in the next months Jan 08 12:44:23 mickeyl: miscount.. yeah I will be not that surprised :) Jan 08 12:44:59 anyone used libgps before? Jan 08 12:45:16 which huge firward? gdbus fix/improvement? what else ? Jan 08 12:45:43 some gps mouvement in fso too? cool :) i am interested :p Jan 08 12:46:45 well, i want to sort fsotdld out (starting atm. by integrating gpsd plugin, then finishing the overall logic of how to communicate location information to clients), i want to improve the resource API a bit to make it even more failure proof against crashes/restarts of individual components, and then there's the big gap for settings and events, which i want to tackle both Jan 08 12:47:25 that plus some little port work (n900, palm pre) should keep me busy until summer Jan 08 12:49:51 hum ok . iam here for the pre/pre2 :). fsotdld will replace the actuall gps api in fso right? is there something to read about it actualy? Jan 08 12:52:43 yes, fsotdld will offer a new dbus-based GPS API that clients can use. there are no docs yet, but the API will be the first incarnation of our agent-based APIs, i.e. you register as a consumer of location data, and then you have to implement a small API that will be called with subsequent reports, depending on the accuray you have requested. you will receive subsequent reports with increasing accuracy, depending on available sources. Jan 08 12:53:07 the sources will be completely transparent Jan 08 12:53:45 i.e. fsotdld will provide you with cell-tower-based, access point-based, and gps-based, depending on available hardware, networking or not, and implemented location providers Jan 08 12:54:05 small API, very high level Jan 08 12:54:26 if you want more details, you can still use gpsd, which will be underlying Jan 08 12:54:46 but most apps should not have to Jan 08 12:54:59 mickeyl: more details like what ? Jan 08 12:55:42 well, at least for GPS, there's a ton of obscure information which you usually don't need Jan 08 12:55:48 health of the sattelites Jan 08 12:55:54 so fso will take care of reserving the gps ressource including waking gps up etc Jan 08 12:55:55 detailed status Jan 08 12:56:17 yes, if you use the high level API, based on the accuracy that you requested, gps will be woken up or no Jan 08 12:56:18 t Jan 08 12:56:40 if everything you want is the city you are in, we don't need to fire up GPS Jan 08 12:56:42 oh yeah right :) I know some about that, my last payed work was around GPS stuff :) Jan 08 12:57:01 right ok Jan 08 12:57:04 understood Jan 08 12:57:30 i will add elementary's map view to zhone2 Jan 08 12:57:39 to make it a demo for fsotdld Jan 08 12:58:06 mickeyl: it will not be a waste of time If I implement gpsd in my app in addition to fsotdld. like this i can begin gpsd coding pending fsotdld to be out Jan 08 12:58:39 zhone2 I forgot about it :) Jan 08 12:58:50 for sure. if you abstract how the location data comes into your app, then you can change the underlying agent later Jan 08 12:59:02 ya, zhone2 is to be forgotten Jan 08 12:59:19 one of these days I will work on a real good demo client Jan 08 12:59:37 but that will have to wait until any FOSS UI reaches the level of abstraction i like Jan 08 12:59:48 yop , like this i will gain the support of more linux based distros (even not linux based but that an other story) Jan 08 13:00:47 ok thx for all these informations mickeyl! Jan 08 13:01:45 -- of course, until all linux distros/phones will support fso :) Jan 08 13:01:46 np, sorry for not getting back to you earlier via email, but i just didn't know about the plan on my own Jan 08 13:01:52 hehe, yeah Jan 08 13:01:57 world domination has been postponed ;) Jan 08 13:02:08 mickeyl: yeah np I don't doubt about it Jan 08 13:02:19 ;) Jan 08 13:03:39 better late than never ;) Jan 08 13:03:48 right Jan 08 13:08:52 ok so back to wirlddomination i.e fso-installer for the moment :) Jan 08 13:10:20 yay Jan 08 13:23:26 mrmoku, hi! did you read the pair of trac bugs I passed you yesterday night? Jan 08 13:36:39 mrmoku: PDP looks good here. although i could not connect on IPCP layer (i need to buy a new SIM card, this one has no flat rate), setting the credentials and initiating the context worked ok. i could see talking through ppp and nothing (except mdbus2 -sl *sigh*) crashed Jan 08 13:38:25 mrmoku: could you test with a neo via forwarding? Jan 08 13:38:32 i have already 2.27.x here Jan 08 13:38:35 (glib) Jan 08 13:38:57 mickeyl: Herzliche Glückwunsch!!!!!! Jan 08 13:39:28 danke :) Jan 08 13:39:48 mickeyl: this means you will not have *any* time left for FSO in 2012 ;-) Jan 08 13:40:50 mickeyl: is your company selling top shelf software. or are you developing for others more as a service company? Jan 08 13:41:12 onen|openBmap: we'll see... i will take a couple of months off from work in early 2012 to fully concentrate on the little one, so once and then i might be able to code a bit Jan 08 13:41:43 onen|openBmap: we have a handful of apps in the store, but this is just for fun and reference. developing for others is where we really earn money. Jan 08 13:42:36 developing and selling own apps is a lottery game for a small publisher Jan 08 13:44:09 mickeyl: indeed Jan 08 13:44:32 mickeyl: I cross read the conversation above, about your new fso service Jan 08 13:44:39 mickeyl: you talked about location service Jan 08 13:44:59 mickeyl: after the next release of my client, i plan working on location service, based on obm data Jan 08 13:45:15 mickeyl: what I have in mind, is a plugin for geoclue Jan 08 13:45:32 mickeyl: how related are geoclue and your project? Jan 08 13:46:16 i never got the hang of geoclue. while the idea is nice, it always striked me as being half-finished. plus, geoclue does not care at all about the problem of activating/deactivating resources Jan 08 13:46:17 mickeyl: ok, will try that then Jan 08 13:47:27 in essence, it's our own geoclue, but with integration into the FSO resource system and integrated w/ fsogsmd and wpa_supplicant for AP / Cell Towers Jan 08 13:47:47 onen|openBmap: for FSO it'd be best if we had a web service Jan 08 13:48:22 mickeyl: we do have: GSM and WiFi Jan 08 13:49:21 mickeyl: my idea is to develop a local db use location service. no network connection service needed Jan 08 13:49:47 onen|openBmap: ok, that would augment the services we currently use over the network Jan 08 13:50:09 mickeyl: let me know if you see this fitting to FSO. We could then discuss how to. For now I understand I can go for geoclue. Right? Jan 08 13:50:44 mickeyl: http://openbmap.org/api/openbmap_api.php5 Jan 08 13:50:49 onen|openBmap: fore sure, there might still be folks that use geoclue. if you have a C api, we can easily cook an fsotdld plugin out of it Jan 08 13:50:53 +1 for no network phone tower database Jan 08 13:51:01 ooh, nice Jan 08 13:51:11 you already have the webservice Jan 08 13:51:13 didn't know that Jan 08 13:51:54 mickeyl: gsm web api has been available for more than a year, easily. ;-) in which world are you living for not knowing that! :P Jan 08 13:52:13 hehe Jan 08 13:52:25 well, as i've said, i couldn't pay much attention to non-iOS things in 2010 :) Jan 08 13:52:54 ssam: proof of concept exist for months (unusable, but just the basis of what it could be). I plan to start working back on this. Jan 08 13:53:34 :-) Jan 08 13:54:37 mickeyl: well I have to ask to the right people, but it is not clear to me what language would fit to develop a geoclue plugin (C obviously, but I am thinking more of java/vala/python). It would be ideal if a simple layer would allow that future service to run on android... Jan 08 13:55:35 mickeyl: maybe java specific implementation for android, and python or vala for real Linux systems implementations will have to be Jan 08 13:56:08 mickeyl: on android, most of people are more than happy to use Google services anyway. so it might be difficult to find interested people in that community Jan 08 13:56:20 mickeyl: works with fsogsmd on my laptop forwarding to the calypso Jan 08 13:56:33 mrmoku: ok, which version of glib are you running on the laptop? Jan 08 13:56:50 error: unable to chain up to base constructor requiring arguments Jan 08 13:56:50 public class LabelBox : Elm.Box { Jan 08 13:56:50 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ Jan 08 13:56:53 onen|openBmap: yeah, the android folks are so used to permanent network connection it might be toughh Jan 08 13:57:01 mickeyl, can you tell me what I'm doing wrong please? Jan 08 13:57:11 Version : 2.26.1-1 Jan 08 13:57:25 mrmoku: ok, that's the version we have on arm as well... hmm Jan 08 13:57:26 mickeyl: so it's not 2.27 fixing it Jan 08 13:57:51 pespin: if there is no constructor without arguments in the base class, you need to add a custom constructor to your LabelBox Jan 08 13:58:01 mickeyl: for me the local approach is two folds: 1. efficiency of battery usage (and availability of network connection) 2. privacy Jan 08 13:58:32 mickeyl, ah ok, and from there call base(params) right? Jan 08 13:58:36 pespin: yep Jan 08 13:59:04 onen|openBmap: for sure. however, at the end of the day most folks are using the location data to query more data, which -- usually -- comes from the network Jan 08 13:59:18 like maps or services Jan 08 13:59:18 mickeyl: fsogsmd rev on FR is a567dc74696fb342fd5abc8388d1dd6064e62219 Jan 08 13:59:26 nothing interesting fsogsmd related after that Jan 08 13:59:43 so that boils down to glib on armv4t being subtly broken Jan 08 13:59:46 fun :/ Jan 08 13:59:49 *sigh* Jan 08 14:00:16 hmm Jan 08 14:00:24 which vala version do we have in OE? Jan 08 14:00:32 there is one fix that came later then my last snapshot Jan 08 14:00:40 i should cut a new release just to make sure it's not that Jan 08 14:00:44 let me do that right now Jan 08 14:00:51 good :) Jan 08 14:01:16 interesting would be to test it on a OE built thing Jan 08 14:01:21 with different arch Jan 08 14:01:26 yeah Jan 08 14:01:27 pity gnutoo is not around Jan 08 14:01:32 he could test on the dream Jan 08 14:01:36 yep Jan 08 14:01:59 mickeyl, uhm Jan 08 14:01:59 error: invocation not supported in this context Jan 08 14:01:59 base(parent); Jan 08 14:02:29 let me try Jan 08 14:02:54 mickeyl: correct. nevertheless: point 1 might still be useful for some usage (OSM maps can also be embedded); point 2 is about dividing what information knows every service you use Jan 08 14:02:57 maybe because Elm.Box is not coming from Glib.Object? Jan 08 14:04:25 mickeyl: if you use google maps for all your searches, google knows all the geographical points you show interest to, even if everytime you talk to a different service afterwards for every query Jan 08 14:05:12 pespin: this works fine here: Jan 08 14:05:21 public class MyBox : Elm.Box Jan 08 14:05:21 { Jan 08 14:05:21 public MyBox() Jan 08 14:05:21 { Jan 08 14:05:21 base( null ); Jan 08 14:05:23 } Jan 08 14:05:24 mickeyl: well of course this makes sense more for a big company like google, which knows a lot more things about you (emails, documents, books, firends, etc.) Jan 08 14:05:25 } Jan 08 14:06:04 mickeyl: if you replace google maps by a web service from another company doing just that, this is maybe not that much different as a local location service Jan 08 14:06:10 onen|openBmap: sure. i don't doubt the privacy aspect, so querying an offline database can definitely make sense even if you _have_ network Jan 08 14:07:13 my biggest issue if the update of the local database, which I would like to be incremental. I have understood debian package system has tools for that. not sure how to do that otherwise Jan 08 14:07:52 hmm Jan 08 14:08:05 mickeyl, ops, I had a remaining "void" in the constructor. But I get another error: Jan 08 14:08:08 error: derived compact classes may not have instance fields Jan 08 14:08:08 public class LabelBox : Elm.Box { Jan 08 14:08:24 Does someone can help me on this : (hopping it is enought explicit) Jan 08 14:08:25 if it's an sql database you can always add migration scripts in postinstall instead of overwriting the whole db Jan 08 14:08:48 I'm using this: Jan 08 14:08:48 public LabelBox(Elm.Object parent, string label, string Value) { Jan 08 14:08:49 Jan 08 14:08:49 base(parent); Jan 08 14:08:59 = >http://pastebin.com/rxurxqVm Jan 08 14:09:23 about checking a variable's value in a list Jan 08 14:09:54 pespin: compact classes have little value beyond binding in vapi. try removing all [Compact] statements from the elm.vapi Jan 08 14:10:49 GarthPS: hmm, my shell foo is little, but would an if-else-cascade be too handish? Jan 08 14:11:34 GNUtoo|laptop: ahh... hi :-) Jan 08 14:11:45 GNUtoo|laptop: we would have some small important test for you to do Jan 08 14:11:46 mickeyl: for me if_else means in a target and i want it checked at the begining Jan 08 14:11:55 mrmoku, hi Jan 08 14:12:03 * GNUtoo|laptop wants to test Jan 08 14:12:15 I could do a test target but I wanted to try it Jan 08 14:12:20 GNUtoo|laptop: does org.freesmartphone.GSM.PDP.SetPdpCredentials work for you on the dream? Jan 08 14:12:31 that's 3g? Jan 08 14:12:40 I've no 3g suscription Jan 08 14:12:46 mrmoku: 0.11.3.1 up on fso Jan 08 14:12:48 no connection needed Jan 08 14:12:51 ah ok Jan 08 14:12:53 then I try Jan 08 14:13:00 just sets apn, user and password Jan 08 14:13:02 and won't connect Jan 08 14:13:06 do I need a sim card? Jan 08 14:13:13 mickeyl: ^^? Jan 08 14:13:15 because I started to boot without sim? Jan 08 14:13:18 s/?// Jan 08 14:13:18 GNUtoo|laptop meant: because I started to boot without sim Jan 08 14:13:28 good question Jan 08 14:13:33 never tried that command without one Jan 08 14:13:40 worth a try anyways Jan 08 14:13:42 ok I'll add one Jan 08 14:15:51 mickeyl, uhm I sed all "[Compact]" but I still get same error :S Jan 08 14:16:00 phonefsod not running, hmmm Jan 08 14:16:01 maybe something inheriting from evas.vapi which is compact? Jan 08 14:16:10 I'll build Jan 08 14:16:27 pespin: possible Jan 08 14:17:13 ok I'll sed evas.vapi too Jan 08 14:17:58 GNUtoo|laptop: phonefsod not needed Jan 08 14:18:01 just use mdbus2 Jan 08 14:18:02 ok Jan 08 14:18:13 request the GSM resource Jan 08 14:18:21 mickeyl: yeah, I have though about that. This is an sqllite db. And you would have to manage building a script from the version you run- the latest version of the db, to send to the client, to update its db Jan 08 14:18:22 call SetFunctionality Jan 08 14:18:29 mickeyl: not a big deal for now Jan 08 14:18:35 SetFunctionality full 1 '' Jan 08 14:18:40 if you have no PIN Jan 08 14:18:47 mickeyl: I'll see this point later Jan 08 14:19:02 give it a bit of time to startup and call PdpSetCredentials then Jan 08 14:19:06 onen|openBmap: yep Jan 08 14:19:24 mickeyl, I get a C error with Compact removed in evas.vapi -> http://paste.pocoo.org/show/317100/ Jan 08 14:20:12 here's the little code I'm using -> http://paste.pocoo.org/show/317101/ Jan 08 14:20:17 root@htcdream ~ # mdbus2 -s org.freesmartphone.ogsmd /org/freesmartphone/GSM/Device org.freesmartphone.GSM.PDP.SetCredentials apn username pass => () Jan 08 14:20:32 shit Jan 08 14:20:34 works then Jan 08 14:20:40 root@htcdream ~ # mdbus2 -s org.freesmartphone.ogsmd /org/freesmartphone/GSM/Device org.freesmartphone.GSM.PDP.GetCredentials => ("apn","username","pass") Jan 08 14:20:45 with phonefsod started Jan 08 14:20:48 and without sim Jan 08 14:21:04 and GetContextStatus ? Jan 08 14:21:37 mdbus2 -s org.freesmartphone.ogsmd /org/freesmartphone/GSM/Device org.freesmartphone.GSM.PDP.GetContextStatus => org.freedesktop.DBus.Error.NoReply: Message did not receive a reply (timeout by message bus) Jan 08 14:21:38 pespin: hmm, bummer. this means we can basically not derive our own classes from E classes, since the type models are not compatible. you need to solve this with aggregating Jan 08 14:21:52 GNUtoo|laptop: hmm Jan 08 14:21:54 mrmoku, note that I've an old gdbus build Jan 08 14:22:02 I'm building the lastest right now Jan 08 14:22:11 I think the build I get was from yesterday night Jan 08 14:22:11 ok, maybe retry when the build finished? Jan 08 14:22:20 should be new enough though Jan 08 14:22:26 ahh... the one you get now? Jan 08 14:22:54 ouch the electricity will go off Jan 08 14:22:57 maybe Jan 08 14:23:07 ENEL dislikes you? Jan 08 14:23:17 lol maybe Jan 08 14:23:57 fwiw Jan 08 14:24:08 mdbus2 seems to crash - still or again - on empty a{sv} Jan 08 14:24:09 *sigh* Jan 08 14:24:15 mickeyl, you mean doing it like public class LabelBox { private Elm.Box box; blablabla } ? Jan 08 14:24:35 pespin: yes. i don't see another way currently. might ask in #vala though Jan 08 14:24:54 ok Jan 08 14:25:36 mrmoku: didn't we (rather you) fix that once? Jan 08 14:28:44 hopefully the laptop didn't shut down this time Jan 08 14:28:54 anyway I bet it's some kind of truffa as always Jan 08 14:29:17 mickeyl do you think it's possible to write sth like efl-gobject? Jan 08 14:29:18 if you usee too much electricity at the same time it goes down Jan 08 14:29:24 so, I'm building Jan 08 14:29:50 mickeyl, btw AKM8976 got decoded Jan 08 14:29:58 accelerometers of nexusone Jan 08 14:30:06 htc dream has nearly the same Jan 08 14:30:12 but you don't care about htc dream anymore Jan 08 14:30:16 playya_: possible, for sure, might be quite a task though Jan 08 14:30:36 GNUtoo|laptop: nice. yes, the nightmare with wifi and gps pretty much ruined the experience for me Jan 08 14:30:51 on nexusone wifi is WEXT Jan 08 14:30:59 gps is free software Jan 08 14:31:11 it uses AT commands Jan 08 14:31:26 the only issue ...cough ....cough....is alsa Jan 08 14:32:06 and wifi is not too hard to make work on htcdream Jan 08 14:32:15 basically it needs something like Jan 08 14:32:29 .irq = MSM_GPIO_TO_INT(29) Jan 08 14:32:34 and to pass that to the driver Jan 08 14:32:45 the only problem is that when I tried it blocks the phone Jan 08 14:32:53 but I don't know nothing about that sutff..... Jan 08 14:33:01 I asked multiple time to kvalo Jan 08 14:33:05 I have to send a mail Jan 08 14:33:28 if that's fixed, the only remaining tasks are integration + better alsa Jan 08 14:33:40 that is to say intone+mplayer works really fine Jan 08 14:33:49 but not VOIP stuff etc... Jan 08 14:34:17 is anything of the nexusone integrated in OE? Jan 08 14:34:24 no Jan 08 14:34:49 as there is no working alsa yet I didn't integrate into oe Jan 08 14:34:56 there is a hack tough Jan 08 14:35:05 http://wiki.freesmartphone.org/index.php/HardwareComparison Jan 08 14:35:23 you can use sdl file to write to the DSP Jan 08 14:35:26 no routing etc... Jan 08 14:36:15 hmm Jan 08 14:36:25 i'm afraid that all rules it out as... uninteresting Jan 08 14:36:27 for me at least Jan 08 14:36:27 I started to write an alsa driver Jan 08 14:36:33 don't need more 80% complete ports Jan 08 14:36:38 we already have so many :/ Jan 08 14:36:39 but it played 4x faster Jan 08 14:36:40 yes Jan 08 14:36:58 the only blocking point is alsa Jan 08 14:37:07 on the nokia900 the blocking point is the modem Jan 08 14:37:48 yeah, where are the libisi folks Jan 08 14:37:50 I sent a mail about what kernel to use for n900 on shr-dev ml Jan 08 14:38:18 mrmoku, is busy with other bugs Jan 08 14:38:22 yeah, seen it Jan 08 14:39:07 freesmartphone.org: 03mickey 07cornucopia * r69255441f533 10/fsotdld/src/plugins/provider_location_gpsd/ (libgps.vapi plugin.vala): fsotdld: provider_location_gpsd: first version working Jan 08 14:39:23 nice for the commit Jan 08 14:39:32 if gpsd can be used it's so nice Jan 08 14:39:44 yeah, paul triggered that Jan 08 14:39:59 re-use of stuff makes things easier Jan 08 14:40:15 folks can chose whether to use gpsd or fso API then. and we don't have to implement all the gps protocols Jan 08 14:40:52 indeed Jan 08 14:41:07 how n900 case will be handled? Jan 08 14:41:12 and palm-pre case too Jan 08 14:41:19 as they don't speak NMEA Jan 08 14:41:42 well, i write plugins for fsotdld Jan 08 14:41:48 if someone writes plugins for gpsd, good Jan 08 14:41:50 BTW, i somewhat negotiated that if we (i) come with a justified enough approach and implementation for turning on/off hooks, it'll get accepted. It seems that almost everything needed for proper power management and agps is there. Almost but not all. Gotta try to do that if i'll feel sane enough today/tomorrow. Jan 08 14:41:51 but not me :) Jan 08 14:41:52 yes but will they export gpsd? Jan 08 14:42:05 no Jan 08 14:42:06 (negotiated with gpsd folks) Jan 08 14:42:10 ok Jan 08 14:42:12 PaulFertser: nice! Jan 08 14:42:44 PaulFertser: libgps is actually quite straightforward these days. no need to use the dbus interface (which is disabled by default anyways) Jan 08 14:43:11 mickeyl: (nice) if i manage to it might have a chance to be nice... Jan 08 14:43:12 so next week or so i have the first fully working fsotdld with a nice API and using gpsd underlaying Jan 08 14:43:24 mickeyl: I fixed some segfault in mdbus2... some (long) time ago... don't remember which though Jan 08 14:44:02 ok, Jan 08 14:44:20 will build a debug version for mdbus2 and have a quick look then Jan 08 14:45:04 hmm, oh Jan 08 14:45:15 hmm, perhaps i had just a too old version installed :) Jan 08 14:46:04 ya, works fine now Jan 08 14:46:19 :) Jan 08 14:46:26 that's the best errors ;) Jan 08 14:46:31 yup Jan 08 14:47:43 should i bump the vala version in OE or are you on that? Jan 08 14:48:05 I'm busy dayworking :/ Jan 08 14:48:44 ok, i'll do it then Jan 08 14:48:49 mrmoku, http://pastebin.com/rmR42Pk9 Jan 08 14:49:30 GNUtoo|laptop: and fsogsmd did segfault? Jan 08 14:49:38 timeout looks like that Jan 08 14:49:45 which would be great ;) Jan 08 14:49:50 not sure Jan 08 14:50:19 yes it did Jan 08 14:50:45 good... then let's hope the vala bump fixes it :) Jan 08 14:51:44 mickeyl: btw gpsd already implements something similar to fso resources: it keeps the device's fd open only if there's a client for this device present. I'm going to try to add a hook to enable additional scriptable action. So that when any client requests the device, it can automatically put the FSO resource to on state. Jan 08 14:54:39 good, that's handy Jan 08 14:54:56 that resembles our good old fso-gpsd plugin from FSO1 Jan 08 14:57:39 mickeyl: well, for proper operation it would need to be a plugin (i.e. linked with the gpsd executable). I doubt gpsd devs would accept hooks of that kind, that's inelegant. But forking a calling a script won't allow to use the FSO resource in the "auto" state (without some strange hacks to call fsoraw with a script that would terminate itself if gpsd dies). Jan 08 15:00:12 PaulFertser: resource handling can also be done by FSO Jan 08 15:00:27 if someone requests "GPS", we can start gpsd Jan 08 15:00:37 so either way will work Jan 08 15:00:46 just a question of which is the more natural usecase Jan 08 15:01:01 mickeyl: (start gpsd) that would be wrong imho, as it'd be unnaturally unfair to libgps-using applications. Jan 08 15:01:13 well, yes Jan 08 15:01:46 i don't care much about legacy as you know, so i won't do much towards that, but of course i welcome if all apps continue to work Jan 08 15:02:09 in the case of fsotdld, fsotdld will request the resource GPS if someone requests location data with an accuracy that goes beyond what wifi and cell towers provide Jan 08 15:02:13 mickeyl: i'd not call libgps apps legacy, i think that's still the most relevant way in the open world. Jan 08 15:02:22 yes, sure Jan 08 15:02:23 mickeyl: yeah, cool. Jan 08 15:03:41 mickeyl: and i think you know gpsd switched to JSON for their protocol and as such it became much more extensible, flexible and reasonable than the previous one. Jan 08 15:04:56 PaulFertser: yeah, i've seen that. that's quite cool, although i'm not using it yet. the fields as found in the gps_data_t and gps_fix_t structure are enough for fsotdld for now Jan 08 15:05:21 I'm glad you liked at least something about gpsd :) Jan 08 15:06:13 hehe. i still think it was the wrong decision to not use dbus properly but rather abandond it (almost), especially if they're now using json... but well... their choice :) Jan 08 15:07:03 and external process control is always annoying to me, that's why i rather use libraries that fork or spawn threads Jan 08 15:07:11 but oh well... can't have everything Jan 08 15:07:52 You can dynamically add and remove devices to gpsd via its special control socket interface. Jan 08 15:08:42