**** BEGIN LOGGING AT Wed Feb 03 02:59:57 2010 Feb 03 03:16:14 denkenz: how can i test that only partial buffer was written into the channel? Feb 03 04:02:47 no way of really testing it Feb 03 04:02:56 your write will return the number of bytes written Feb 03 04:03:04 the code must be prepared for that number to be less than asked for Feb 03 04:40:34 no room check? are there libraries to simulate these things for testing? Feb 03 04:58:43 unless you can simulate a broken serial port that only accept partial data... is there such libraries or device? Feb 03 05:45:12 this is not about a serial port, any fd can act this way Feb 03 05:45:19 Just imagine the system is under stress Feb 03 05:50:55 And yes there are probably libraries for this kind of testing Feb 03 05:51:14 I'm simply too lazy to look for them, so I wrote my own debug code :) Feb 03 18:00:39 hi. is there anywhere an OE recipe for ofono? Feb 03 18:02:24 I haven't seen one Feb 03 18:04:33 ok. i tried to find one in some branches of oe, but no luck Feb 03 22:21:15 denkenz: I attached the UUIDs to DeviceCreated signal. The patch is on mailing list. It is a bit bigger since a I added some code to properly remove the watches Feb 04 00:37:31 padovan: Why do we watch DeviceRemoved? Feb 04 00:37:47 padovan: Wasn't this taken care of by Agent Release? Feb 04 00:48:08 are there any reviews of ofono out there? is it considered production material yet? Feb 04 00:49:35 padovan: So I think we can simplify this quite a bit, tell me if there are any holes in my theory: Feb 04 00:49:55 padovan: Register a single watch for UUIDs Feb 04 00:49:56 denkenz: DeviceRemoved is raised after an unsucceful pair Feb 04 00:50:17 padovan: You should be able to watch _all_ device emissions Feb 04 00:50:39 padovan: Then add those devices to the hash table which you actually register Feb 04 00:51:10 padovan: Remove them when Release gets called Feb 04 00:55:06 Jorophose: oFono is not production ready, no Feb 04 00:55:11 denkenz: tell me if I understood: device_created() will add the path to the hash table. Then the uuid watch will check if the device is on the hash table. If yes, create the modem Feb 04 00:55:52 padovan: No, I'm thinking this way, parse_uuids will add the device to the hash table assuming it has hfp uuid Feb 04 00:56:19 yes. Feb 04 00:56:22 denkenz: how long would you estimate it would take? I'm interested in using oFono with a USB GSM device dongle thing to make calls from a linux laptop. (I'm hoping it's feasable?) Feb 04 00:56:36 padovan: Then a global uuid changed watch will call parse_uuids again, if parse_uuids detects hfp and the modem is already on the hash table, ignore Feb 04 00:56:44 padovan: Otherwise add it to the hash table Feb 04 00:57:09 like, is it still in its early stages, is this a long-term sort of thing, or is it nearing usable within the year? (and I thought the Intel/LG phone used ofono?) Feb 04 00:57:13 Jorophose: Calls and SMS works today just fine Feb 04 00:57:31 denkenz: I understood it now. So we don't need to listen DeviceCreated anymore. :) Feb 04 00:57:38 padovan: Correct Feb 04 00:58:28 padovan: Also check how AdapterRemoved works, will it result in Release() getting called or not... Feb 04 00:58:40 padovan: And we might have to handle the case where bluetoothd goes down Feb 04 00:58:43 denkenz: so would it be usable for someone who's familiar with linux & filing bug reports? Feb 04 00:58:55 what Intel/LG phone? Feb 04 00:59:00 why would you not call it production ready? just because it's not 100% stable 'n stuff? Feb 04 00:59:11 LG GW990 I think? built on moorestown, runs Moblin Feb 04 00:59:19 gobi? Feb 04 00:59:20 Jorophose: Not using oFono Feb 04 00:59:37 denkenz: when bluetooth goes down it Release the agent and the modems get unregistered. Feb 04 00:59:44 Jorophose: I define production ready as has passed GCF or carrier compliance Feb 04 00:59:50 denkenz: unless bluetoothd segfaults. Feb 04 01:00:17 padovan: yeah, in the unlikely event of :) Feb 04 01:00:30 padovan: We don't have to worry about it now, just maybe something to consider for the future Feb 04 01:00:39 denkenz: oh, I see what you mean. yeah, I meant it more as just a lone user. I'm wondering if I could kill two birds with one stone and not need to use a cellphone... Feb 04 01:01:42 Jorophose: Again, calls and SMS works on the hardware we have enabled. 3G Data works on MBM and HSO devices Feb 04 01:01:55 Jorophose: Question is can you get audio out of your dongle Feb 04 01:02:00 Most don't support that Feb 04 01:03:04 dekenz: oh, wow. I didn't think about that. is there a wiki on the ofono site I can use to look it up? Feb 04 01:03:30 I was looking at the Huawei E181 Feb 04 01:04:26 Jorophose: Nothing really in terms of a wiki, I don't have much time to maintain one, everything enabled can be found in plugins/* Feb 04 01:04:52 Huawei ones are not terribly well supported right now, but they do sorta work Feb 04 01:05:25 Oh and no Huawei dongle I know supports audio Feb 04 01:05:54 hmm. alright then. what dongles would you recommend, if you know any that work? Feb 04 01:07:41 There are no USB dongles that I know of with audio support Feb 04 01:07:52 If you find one, let me know and we'll enable it ASAP ;) Feb 04 01:08:06 hahaha. really? I would have hoped it wasn't a hardware thing. Feb 04 01:08:22 hmm. the alternative was to get an n900, which I might just do. Feb 04 01:10:27 but wait, if you said calls work, is that only for full-on cellphones? Feb 04 01:10:42 and what about pairing it with a cellphone? Feb 04 01:14:44 we have it working on devices that support calls, e.g. Freerunner Feb 04 01:15:34 and you will soon be able to use oFono as a headset against any bluetooth HFP enabled mobile phone Feb 04 01:17:06 was ofono required to tether a phone and make calls? or is that part of the bluetooth/audio systems? Feb 04 01:17:19 just out of curiosity. ofono's always piqued my interest Feb 04 01:22:05 bluetooth defines a protocol for this, ofono implements it **** ENDING LOGGING AT Thu Feb 04 02:59:57 2010