**** BEGIN LOGGING AT Tue Dec 21 02:59:59 2010 Dec 21 07:05:37 balrog-k1n: ping Dec 21 14:54:35 holtmann: Has make distcheck been fixed yet? Dec 21 14:54:49 yes, but not pushed yet. Dec 21 14:56:11 ok Dec 21 15:11:19 denkenz: holtmann : Hi. I want to take the "long phone numbers" task, so I sent a patch for that Dec 21 15:38:53 Jeevaka: pong Dec 21 15:39:28 Jeevaka: yes, one sim i have has a "Send E-mail" application and the Get Input comes with not timeout defined Dec 21 15:39:59 but also as far as i remember ts102223 also mentions two manufacturer-specific timeouts Dec 21 15:40:08 normal and short timeout Dec 21 15:47:05 ok incase of GetInput, GetInkey we are using normal timeout which is 10minutes when duration object is not present Dec 21 15:48:09 in that case, few conformance test cases might fail right? Dec 21 15:54:02 Jeevaka: which test do you have in mind? Dec 21 15:55:13 i didn't find the information regarding normal and short delay in TS 102 223. If you know the section, let me know Dec 21 15:56:30 Test cases: 27.22.4.2.2, 27.22.4.3.2 and 27.22.4.9.8 Dec 21 15:59:02 TS 102 223 6.4.1 (Display Text) talks about the short delay in a couple of places Dec 21 15:59:14 Display Text is the only place the short delay is used i think Dec 21 16:03:54 Display Text and Display Multimedia message are the places where the short delay is used Dec 21 16:05:58 * balrog-k1n tries to find the Table A.2/2 in the TS 102 384 and can't find it :) Dec 21 16:07:02 ah there it is Dec 21 16:08:07 but there is no actual timeout value for that test given, is there? Dec 21 16:08:49 as far as i understand the manufacturer-specific timeout should be adjusted before running the test Dec 21 16:09:19 also if there was a defined length of time, it wouldn't be manufacturer-specific anymore Dec 21 16:12:46 test doesn't give any actual time value as it is manufacturer specific. agree with you but the current timer is quite high Dec 21 16:14:23 maybe it should be shortened, but i don't know what value would be good... anything below 3 minutes seems too short for the email app Dec 21 16:15:42 it would be nice if we have a timeout value of 1min or even less than that. so, we can just run the conformance by taking the release images Dec 21 16:15:48 ;) Dec 21 16:15:51 3 minutes starting from the last interaction would be good, but we don't have such input in ofono Dec 21 16:20:25 I'm ok with anything below 5min Dec 21 16:23:14 let's set it to 3 mins then Dec 21 16:24:07 now, i have some questions on immediate response Dec 21 16:24:11 i hate to write 90% of a message and have the battery run out or something similar happen :) Dec 21 16:28:25 question regarding imme.response is why are we removing the agent? Dec 21 16:31:37 terminal response is already sent if the pcmd is received with immediate response is set, aren't we supposed to just discard any message from agent? Dec 21 16:34:54 Jeevaka: i think it was Denis who made the assumption that immediate response will only ever by used by the sim at the end of a session Dec 21 16:35:09 (it's documented in the comment above that piece of code) Dec 21 16:37:06 i think this assumption was made to make the code simpler.. i'm not sure, the previous version didn't terminate the session after a DisplayText with immediate response unless the card terminated it Dec 21 16:42:22 denkenz, balrog-k1n: was there any issue seen with the previous version of Display Text with immediate response, I believe card will terminate the session for sure incase of Display Text with immediate response Dec 21 16:46:39 by "previous version" i mean the code that was first merged into ofono, the main issue was probably style. since it was refactored immediately after it was merged it's hard to say if there were other issues :) Dec 21 16:49:12 is there a problem or are you being paranoid again? Dec 21 16:49:25 (not that there's anything wrong with that) Dec 21 16:54:15 I have added few conformance cases to phonesim and tried to run all the cases in a single session Dec 21 16:55:44 when i ran the immediate response related test case, session got released, so i just wanted to know why we release the session when the card will release the session Dec 21 17:04:20 made sense at the time ;) Dec 21 17:16:46 denkenz: any valid example for immediate response case? Dec 21 17:19:11 denkenz: hi... did you get my patches this time? Dec 21 17:21:39 while i'm all for handling even the corner cases like immesiate response in the middle of a session, i think it wouldn't be too difficult to modify the test to restart the session after that case? Dec 21 17:24:03 I don't believe immediate response within a session ever makes sense Dec 21 17:24:33 so I'm fine leaving it as is until someone shows us a real world example Dec 21 17:29:34 agree. if someone comes with a real world example, then we can think about modifying it. My case is just to run the test cases in a single session which is just for convenience :) Dec 21 17:41:33 you guys care too much about real world :) Dec 21 17:50:22 lol Dec 21 18:20:43 balrog-k1n: Still around? Dec 21 19:01:05 denkenz: yes Dec 21 19:04:53 denkenz: are you around? Dec 21 19:06:05 i have a ugly new probem with my huawei. it reacts on AT+CFUN=1 with ERROR. AT+CFUN? gives 0 Dec 21 19:07:32 i had that earlier when issuing the command manually but ofono didn't have it. now even ofono has that problem. ever heard of such behaviour? Dec 21 19:08:16 emdete: what modem? Dec 21 19:08:42 emdete: i have a huawei E220 here Dec 21 19:08:45 huawei em770 Dec 21 19:08:56 same problem... Dec 21 19:08:58 demarchi: do you see similar stuff? Dec 21 19:09:09 if you unlock the sim, it works Dec 21 19:09:13 sorry Dec 21 19:09:18 unlock? Dec 21 19:09:27 i have this behavior if there's no SIM Dec 21 19:09:57 if PIN is locked, then i need patches in order to unlock/enter pin Dec 21 19:11:07 emdete: did you see my patch in ML ? Dec 21 19:11:24 i don't follow the ml... too much MLs... :/ Dec 21 19:13:02 have you seen mine? a friend posted because i'm not allowed to write to the ml Dec 21 19:13:58 yeah... i share the same opinion as denkenz Dec 21 19:14:30 balrog-k1n: I think we need explicit sim file / sim image flushing functions Dec 21 19:14:52 balrog-k1n: Since a change to EFiidf should invalidate them all Dec 21 19:14:56 if i would make it for money, if i would only follow this one or a handful project, if... yes, i would, sure. but i don't Dec 21 19:14:58 emdete: what i did is to insert a timeout... if we don't receive a notification from modem about state change, we poll it Dec 21 19:15:53 if we are succesfful on polling, it means the SIM is inserted and waiting for pin Dec 21 19:16:01 then a notification is sent Dec 21 19:16:21 but back to topic: the sim pin is entered, i get a sim state PinRequired=none, after trying a set-prop online=true i get the error Dec 21 19:16:28 i have to update this patch in order to remove some extra parsings i'm doing Dec 21 19:16:56 emdete: humn... this error i'm not getting Dec 21 19:17:14 after PinRequired=none, it works Dec 21 19:17:17 maybe i can share my client. it's a simple py script issuing the commands one after the other Dec 21 19:17:46 and it worked some weeks ago, i don't know why it stopped working and when Dec 21 19:18:09 will try to go back in git hstory Dec 21 19:19:00 there are very few commits that could possibly break this Dec 21 19:19:09 i think a bisect is the way to go Dec 21 20:10:52 maybe your modem firmware just hates you now? Dec 21 20:11:27 denkenz: do you mean we need separate functions for image / the other files? Dec 21 20:11:38 I think so Dec 21 20:12:00 I don't believe the current 'Cache flushing functions' patch is enough Dec 21 20:12:14 or sim_fs_cache_flush could check if the file is EFiidf? Dec 21 20:13:03 So there are two possibilities: Dec 21 20:13:12 EFiidf is changed -> all of image cache needs to be flushed Dec 21 20:13:17 denkenz: that may be but it's closed source, i have to solve it outside :/ Dec 21 20:13:31 EFimgn is changed -> images on that EF needs to be flushed Dec 21 20:14:00 The current evil hack of deleting EF and image with id foo seems too ... hacky Dec 21 20:14:21 emdete: Nah, use my tried and true method: Dec 21 20:14:43 window.open(); self.throw(modem, window) Dec 21 20:14:46 window.close() Dec 21 20:15:01 i need to re-read the file because i don't understand how the images are cached and referenced Dec 21 20:15:35 but i think we can check for specific file ids inside sim_fs_cache_flush and remove all affected files Dec 21 20:15:48 denkenz: there is alot of hw around that modem, i do not want to throw that away too Dec 21 20:16:17 balrog-k1n: nope, you need to know the EFiidf info Dec 21 20:16:28 otherwise you don't know the mapping between id and EFimgn file Dec 21 20:17:01 emdete: That never stopped me ;) Dec 21 20:17:03 ah Dec 21 20:18:13 balrog-k1n: Also, I don't like the name DisplayAlphaId Dec 21 20:18:23 what about DisplayActionInformation? Dec 21 20:19:10 denkenz: you never had that hw Dec 21 20:20:13 emdete: Lol, if they put a Huawei chip in there, I don't *want* that hw ;) Dec 21 20:20:46 denkenz: you want, i'm sure. ;) Dec 21 20:21:36 denkenz: there is no alternative currently for you :D Dec 21 20:22:07 of course there is Dec 21 20:22:20 tell me Dec 21 20:22:37 (do you know what we are talking about?_ Dec 21 20:22:38 ) Dec 21 20:23:57 hm, that is not that much as an alternative, denkenz Dec 21 20:24:08 ; Dec 21 20:24:10 ;) Dec 21 20:24:13 DisplayActionInformation sounds good Dec 21 20:24:34 emdete: Yes I know you're still drooling over the wetab ;) Dec 21 20:24:54 denkenz: if you know a alternative i would be very, very happy Dec 21 20:26:17 balrog-k1n: Ok can you resend that patch? Dec 21 20:26:29 yep, will do Dec 21 20:28:00 denkenz: i'm not drooling at all, i just don't know a better platform for my stuff, so if you know one... just let me know Dec 21 20:28:09 emdete: I didn't say anything about commercially available ;) Dec 21 20:28:29 For you, I agree its the best thing right now Dec 21 20:28:35 But still a bummer about the modem hw Dec 21 20:29:02 see if you can maybe stick an mbm card in there Dec 21 20:30:43 denkenz: i could but the nice thing is that it is inside. i know huawei sux, but... other users are around with exactly that hw also Dec 21 20:31:14 denkenz: and if you have some non-commercially-available hw you may open some channel that it reaches me ; Dec 21 20:31:17 ) Dec 21 20:32:26 every time I look at the huawei driver I get depressed Dec 21 20:32:35 so maybe I'm just biased right now ;) Dec 21 20:35:57 denkenz: I have some questions on TTY, can i? Dec 21 20:36:09 yeah go ahead Dec 21 20:36:58 why are we supporting only 1 TTY mode Dec 21 20:39:26 3 modes available: TTY, VCO and HCO Dec 21 20:46:22 dunno, no one told us? Dec 21 20:53:18 ok. TTY device can be configured in one of the 3modes: TTY(type and read text on TTY device), VCO( speak in phone and read replies on TTY device), HCO(listen on phone and type on TTY device) Dec 21 20:56:57 jeez, complicated Dec 21 21:00:55 so, are we going to provide support for all the modes or only TTY mode? Dec 21 21:01:06 you tell me ;) Dec 21 21:01:29 how likely are any of the other modes to be used? Dec 21 21:02:16 I heard that the phones in US needs to have this support, so I dont have any idea on this Dec 21 21:02:55 atleast some motorola phones are providing support for all the modes Dec 21 21:05:09 I suggest you speak to the requirements definition people Dec 21 21:05:22 I really wouldn't know what modes are required Dec 21 21:06:58 balrog-k1n: Do we want to enforce the no-error return semantics for DisplayActionInformation? Dec 21 21:08:14 may i know who are the requirements definition people!! Dec 21 21:21:10 denkenz: yes, huawei is strange, i tested with manual commands and it didn't do what i wanted :/ don't know why a 'at+cpin=1' dosnt work :/ Dec 21 21:25:48 cpin=1 is invalid syntax Dec 21 21:38:56 denkenz: phonesim patches for EFcfis and EFcphs-cff record sent Dec 21 21:41:13 denkenz: i.e. terminate session if agent returns? Dec 21 21:41:58 balrog-k1n: Yeah, or at least follow the regular semantics of checking that no error is returned and that the method return has no arguments Dec 21 21:43:15 well any callback will be an error so we can save ourselves the check Dec 21 21:44:06 i can add it but i think it doesn't make much difference Dec 21 21:44:11 I think its fine to allow empty returns Dec 21 21:44:30 but the rest it seems we should be consistent Dec 21 21:44:54 or mark this method return as noreply somehow Dec 21 21:45:27 i wanted to mark it as noreply but then we would be sending Cancel() even if the method returned Dec 21 21:45:47 hm, good point Dec 21 21:46:11 but as it is now it's same as noreply, just with a callback :) Dec 21 21:46:46 I'm just thinking we should still be pedantic and make sure we perform signature checking, etc Dec 21 21:51:10 denkenz: i'll resend with the check then Dec 21 21:51:55 ok thanks, I know this doesn't make any difference, but I still want to turn away 'rogue' clients Dec 21 21:52:25 though we already have Cancel() and Release() where we don't check the signature Dec 21 21:54:59 They're sent as no reply Dec 21 22:13:57 demarchi: here? Dec 21 22:46:32 denkenz: i'm here now Dec 21 22:48:02 In your syscfg_support_cb Dec 21 22:48:48 If the parsing fails, do you really want to remove the atom? Dec 21 22:49:00 Or perhaps reset the mode to any/any/any? Dec 21 22:49:54 I'm worried about other sw messing with this one and torpedoing us Dec 21 22:54:23 denkenz: humn... the only way i can set the mode / frequency band in huawei is with ^syscfg Dec 21 22:54:45 denkenz: what do you mean by reset it to any/any/any ? Dec 21 22:55:16 if there's no support for syscfg command, i really doesn't know how to do this Dec 21 22:55:40 I meant more like SYSCFG? doesn't return an error Dec 21 22:56:12 but for instance syscfg_query_parse_mode returns FALSE Dec 21 22:56:20 because the mode is set to something strange Dec 21 22:59:50 denkenz: humn... yeah Dec 21 23:00:02 i think i have to simplify this method Dec 21 23:00:12 so, i don't do any parsing Dec 21 23:00:28 and only check for ok / !ok Dec 21 23:01:39 the way it is might be fine, if you simply reset the preferences to any/any/any and set the appropriate SYSCFG Dec 21 23:01:59 Or you can split the SYSCFG=? and SYSCFG? parsers Dec 21 23:03:56 then, what if setting to any/any/any fails? Dec 21 23:04:52 then say its not supported Dec 21 23:05:06 + if (rsd->mode == OFONO_RADIO_ACCESS_MODE_ANY || Dec 21 23:05:08 + (rsd->mode == OFONO_RADIO_ACCESS_MODE_GSM && Dec 21 23:05:08 + rsd->band_gsm != band_gsm) || Dec 21 23:05:09 + (rsd->mode == OFONO_RADIO_ACCESS_MODE_UMTS && Dec 21 23:05:09 + rsd->band_umts != band_umts)) Dec 21 23:05:09 + change_mode = TRUE; Dec 21 23:05:13 Also, what is this logic trying to accomplish? Dec 21 23:07:03 so... we only issue a syscfg command if we actually have to Dec 21 23:07:09 the core radio settings atom doesn't call the driver unless umts_band != new_umts_band Dec 21 23:07:21 and same goes for gsm_band Dec 21 23:07:28 so at least the last two checks are not needed Dec 21 23:07:37 for example: if you are configured to use only GSM Dec 21 23:07:55 then you set the umts band, you cannot call the syscfg command Dec 21 23:08:08 otherwise you'll be disconected Dec 21 23:08:35 so, i cache this new band value and whenever i use the umts band, i call the syscfg command to set it Dec 21 23:09:09 this was the problem of splitting in two properties i was telling you before doing the huawei driver Dec 21 23:09:36 But the bands are a bitfield Dec 21 23:09:49 so can't you OR the umts band and the gsm band together? Dec 21 23:11:15 humn... i have to check this... i remember to read somewhere that it's not supported Dec 21 23:12:13 At least a cursory glance at the ids for the bands I don't detect any collisions Dec 21 23:12:34 So I'd be surprised if you could set 3FFFF but not 8080 or something Dec 21 23:15:50 humn... might be Dec 21 23:16:21 i'll try some OR'd values, then i update the patch Dec 21 23:16:29 this would simplify a bit this driver Dec 21 23:17:00 that's my thinking as well Dec 21 23:23:29 denkenz: thanks, i have to go now Dec 21 23:43:26 hello room Dec 21 23:43:53 i need with with modem Dec 21 23:44:10 i have a usb wireless broadband Dec 21 23:44:20 and i need howto on it Dec 21 23:47:38 anyone Dec 21 23:47:54 i need direction **** ENDING LOGGING AT Wed Dec 22 02:59:57 2010