**** BEGIN LOGGING AT Mon Aug 30 02:59:57 2010
Aug 30 14:49:13 <Spider-Pork> denkenz you was waiting some documents from me?
Aug 30 14:49:22 <Spider-Pork> *was you
Aug 30 14:57:41 <denkenz> padovan: pong
Aug 30 14:57:55 <denkenz> Spider-Pork: About the weird .1 ms after sending a command?
Aug 30 15:07:09 <Spider-Pork> denkenz yes
Aug 30 15:07:16 <Spider-Pork> and the time is 0.1s
Aug 30 15:07:41 <denkenz> sure, give it to me
Aug 30 15:07:51 <Spider-Pork> give me a link where i can send to you cinterion doc
Aug 30 15:08:08 <Spider-Pork> cinterion = siemens
Aug 30 15:08:11 <denkenz> Just use email it to me
Aug 30 15:08:27 <Spider-Pork> and email is?
Aug 30 15:08:35 <Spider-Pork> the address i mean
Aug 30 15:08:37 <denkenz> denkenz@gmail.com
Aug 30 15:11:53 <Spider-Pork> ok i'm uploading it to my gmail
Aug 30 15:14:30 <Spider-Pork> denkenz sent
Aug 30 15:14:44 <denkenz> ok thanks
Aug 30 15:15:30 <Spider-Pork> no problem
Aug 30 15:16:55 <Spider-Pork> it says too that 100ms is standard delay, some commands need more time
Aug 30 15:20:45 <denkenz> lol, that's helpful
Aug 30 15:22:11 <Spider-Pork> good
Aug 30 15:23:08 <Spider-Pork> I think you'll have something to read this night :)
Aug 30 15:27:22 <denkenz> yes, nothing like curling up with an at command manual before bed
Aug 30 15:29:01 <Spider-Pork> of course
Aug 30 15:33:46 <Spider-Pork> maybe it will help you to solve some problems
Aug 30 15:34:56 <denkenz> akiniemi: Is the crash happening during cleanup?
Aug 30 15:55:09 <akiniemi> denkenz: yup
Aug 30 15:59:29 <denkenz> akiniemi: should be fixed btw
Aug 30 15:59:52 <akiniemi> denkenz: problem's gone, thanks
Aug 30 16:15:45 <denkenz> akiniemi: So I'm actually fine moving to an Agent API for Messaging
Aug 30 16:16:43 <denkenz> akiniemi: However, I don't like 3 Register functions and and a single Agent interface
Aug 30 16:16:47 <denkenz> seems unintuitive
Aug 30 16:21:05 <akiniemi> denkenz: yeah, those should be separate interfaces
Aug 30 16:21:34 <denkenz> Another question, how do we get the WAP Application ID
Aug 30 16:21:54 <akiniemi> denkenz: you mean from the SMS?
Aug 30 16:21:56 <denkenz> Don't we have to parse the WSP for that?
Aug 30 16:22:01 <denkenz> yep
Aug 30 16:22:11 <akiniemi> denkenz: yes, I was planning to write a WSP parser
Aug 30 16:22:52 <denkenz> So if we're investing in that, why do we use array{byte} in the agent?
Aug 30 16:23:02 <denkenz> Is that the XMLized WSP or just the raw pdu?
Aug 30 16:23:25 <akiniemi> That would be the raw PDU
Aug 30 16:23:35 <akiniemi> Parsing that would be left to the agent to do
Aug 30 16:23:50 <denkenz> So we're investing in WSP parser just to extract the id?
Aug 30 16:24:32 <akiniemi> Yes, I believe there's a restricted subset of WSP we need only for this purpose
Aug 30 16:25:50 <denkenz> I looked at WSP before, its not terrible
Aug 30 16:26:06 <denkenz> However, I'd rather go all in or all out
Aug 30 16:26:23 <denkenz> e.g. treat Pushes just as another app src/dest id
Aug 30 16:26:34 <denkenz> Or generate XML / dict from the WSP
Aug 30 16:30:14 <akiniemi> Hmm...
Aug 30 16:32:00 <akiniemi> I thought it would be neat to be able to register only using the WAP application ID
Aug 30 16:32:22 <akiniemi> But since you probably need to dish out the WSP header along with the body to the agent anyway...
Aug 30 16:33:16 <denkenz> And since the agent knows how to talk WSP anyway...
Aug 30 16:33:47 <denkenz> In theory we could dump the WSP decoder into oFono, but then you guys have WAP over IP
Aug 30 16:34:32 <akiniemi> denkenz: WAP over IP?
Aug 30 16:34:54 <denkenz> Wasn't there some way of receiving WAP PUSH over other transports than SMS?
Aug 30 16:35:22 <akiniemi> Ah, yeah. I think that was only needed for wimax
Aug 30 16:35:36 <akiniemi> Wouldn't worry about that in oFono
Aug 30 16:35:58 <denkenz> Nod, if SMS was the only transport, then we could do WSP in oFono
Aug 30 16:36:05 <denkenz> Then have apps just deal with XML/dicts
Aug 30 16:36:26 <denkenz> But if there are other transports, then probably easier to just keep WSP out of oFono completely
Aug 30 16:36:45 <akiniemi> Agree. Simplifies the task a bit too ;)
Aug 30 16:37:11 <denkenz> yep :)
Aug 30 16:37:30 <akiniemi> I'll go forward with a pair of Register methods and interfaces, one for text, one for applications (using ports)
Aug 30 16:37:43 <denkenz> Also, what is the usecase behind regexp for the Text agent?
Aug 30 16:37:57 <denkenz> It seems we can get away with just one...
Aug 30 16:38:09 <denkenz> no need for multiple text agents
Aug 30 16:39:52 <akiniemi> Ah, that reminds me, some agents may need to "consume" messages, e.g., so that an SMS with certain confirmation token don't end up in the UI
Aug 30 16:40:33 <akiniemi> That would mean a run order between agents, and return value for the method...
Aug 30 16:40:50 <akiniemi> Like propagating a signal through handlers
Aug 30 16:41:18 * akiniemi goes offline for a bit...
Aug 30 16:50:31 <pessi> sounds like you are going to re-invent Telepathy ;)
Aug 30 16:56:18 <denkenz> lets hope we don't go down that path, it leads to the dark side
Aug 30 19:35:19 <kristenc> balrog-k1n: denkenz: is it possible to use phonesim test stk, or do I need to have a real modem with stk capability?
Aug 30 19:35:33 <denkenz> Its possible
Aug 30 19:35:51 <denkenz> start phonesim, ./enable-modem, ./online-modem
Aug 30 19:36:01 <denkenz> Then in the SIM tab of phonesim, hit Start Application
Aug 30 19:36:08 <denkenz> Your main menu should go live
Aug 30 19:36:19 <denkenz> test-stk-menu is the reference agent
Aug 30 19:37:01 <kristenc> denkenz: what main menu?
Aug 30 19:37:13 <kristenc> when I hit Start Application, nothing happens.
Aug 30 19:37:31 <denkenz> Now run test/test-stk-menu
Aug 30 19:37:55 <denkenz> There's a set of menu items which give icons
Aug 30 19:38:34 <kristenc> denkenz: ok, I have the menu items, but if I select one of them, again nothing happens.  Is that expected?
Aug 30 19:39:07 <denkenz> what do you mean nothing happens?
Aug 30 19:40:28 <denkenz> Enter Selection: 5
Aug 30 19:40:29 <denkenz> Agent registered for session
Aug 30 19:40:29 <denkenz> Title: (Icons)
Aug 30 19:40:29 <denkenz> 0. Basic Icon
Aug 30 19:40:30 <denkenz> 1. Color Icon
Aug 30 19:40:30 <denkenz> 2. Bad Icon
Aug 30 19:40:30 <denkenz> Default: -1
Aug 30 19:40:30 <denkenz> Enter Selection (t, b):
Aug 30 19:40:38 <kristenc> by nothing I mean, something happens :), but it isn't anything visible on my terminal.
Aug 30 19:41:04 <kristenc> [kristen@kcaccard-MOBL3 test]$ ./test-stk-menu 
Aug 30 19:41:05 <kristenc> Main Menu:
Aug 30 19:41:05 <kristenc> Items:
Aug 30 19:41:05 <kristenc> 0. News
Aug 30 19:41:05 <kristenc> 1. Sports
Aug 30 19:41:05 <kristenc> 2. Time
Aug 30 19:41:07 <kristenc> 3. Sticks Game
Aug 30 19:41:09 <kristenc> 4. Tones
Aug 30 19:41:11 <kristenc> 5. Icons (not self-explanatory)
Aug 30 19:41:13 <kristenc> 6. Icons (self-explanatory)
Aug 30 19:41:15 <kristenc> 7. Finance
Aug 30 19:41:19 <kristenc> 8. Web Browser
Aug 30 19:41:21 <kristenc> Enter Selection: 5
Aug 30 19:41:23 <kristenc> Agent registered for session
Aug 30 19:41:26 <kristenc> it just sits there after this and does nothing.  Output from ofonod is this:
Aug 30 19:41:41 <kristenc> ofonod[10972]: src/stk.c:stk_select_item() 
Aug 30 19:41:41 <kristenc> ofonod[10972]: src/stk.c:stk_select_item() 
Aug 30 19:41:41 <kristenc> ofonod[10972]: src/stk.c:stk_send_envelope() 
Aug 30 19:41:41 <kristenc> ofonod[10972]: drivers/atmodem/stk.c:at_csim_envelope_cb() csim_envelope_cb: 2
Aug 30 19:41:41 <kristenc> ofonod[10972]: src/stk.c:envelope_cb() length 0
Aug 30 19:41:42 <kristenc> ofonod[10972]: src/stk.c:menu_selection_envelope_cb() 
Aug 30 19:41:44 <kristenc> ofonod[10972]: src/stk.c:menu_selection_envelope_cb() Menu Selection envelope submission gave no error
Aug 30 19:41:48 <kristenc> ofonod[10972]: drivers/atmodem/stk.c:csim_fetch_cb() csim_fetch_cb: 63
Aug 30 19:41:50 <kristenc> ofonod[10972]: src/stk.c:stk_menu_create() 
Aug 30 19:41:52 <kristenc> ofonod[10972]: SimToolkitAgent /test/agent replied with error org.freedesktop.DBus.Error.AccessDenied, Rejected send message, 1 matched rules; type="method_call", sender=":1.793" (uid=0 pid=10972 comm="./src/ofonod) interface="org.ofono.SimToolkitAgent" member="RequestSelection" error name="(unset)" requested_reply=0 destination=":1.802" (uid=500 pid=11106 comm="/usr/bin/python))
Aug 30 19:41:57 <kristenc> ofonod[10972]: src/stk.c:session_agent_notify() Session Agent removed
Aug 30 19:41:57 <denkenz> Ah yes
Aug 30 19:41:59 <kristenc> ofonod[10972]: src/stk.c:session_agent_notify() Sending Terminate response for session agent
Aug 30 19:42:01 <kristenc> ofonod[10972]: src/stk.c:send_simple_response() result 16
Aug 30 19:42:03 <kristenc> ofonod[10972]: src/stk.c:stk_respond() 
Aug 30 19:42:05 <kristenc> ofonod[10972]: drivers/atmodem/stk.c:at_csim_terminal_response_cb() csim_terminal_response_cb: 2
Aug 30 19:42:07 <kristenc> ofonod[10972]: src/stk.c:stk_command_cb() TERMINAL RESPONSE to a command reported no errors
Aug 30 19:42:10 <denkenz> You need the dbus policy hole
Aug 30 19:42:47 <denkenz> Copy the latest src/ofono.conf to your dbus system.d directory
Aug 30 19:42:57 <denkenz> Then change the user from root to whatever
Aug 30 19:43:01 <denkenz> If you're using it as non-root
Aug 30 19:43:51 <denkenz>   <policy user="denkenz">
Aug 30 19:43:51 <denkenz>     <allow own="org.ofono"/>
Aug 30 19:43:51 <denkenz>     <allow send_destination="org.ofono"/>
Aug 30 19:43:51 <denkenz>     <allow send_interface="org.ofono.SimToolkitAgent"/>
Aug 30 19:43:52 <denkenz>   </policy>
Aug 30 19:51:21 <balrog-k1n> kristenc: currently the Icon IDs sent by the card/phonesim are not provided on D-bus
Aug 30 19:52:17 <balrog-k1n> you will "0" everywhere as the icon id, it needs to be implemented
Aug 30 19:57:21 <denkenz> The icon id information is there, we just never call the agent with it
Aug 30 19:57:32 <denkenz> Its a trivial change if you wanna play with icons
Aug 30 20:23:25 <kristenc> hum, it still seems like I'm not doing something right.  Output changed on the command line, but should there be more for me to see than that?  I still don't see a demo sim application anywhere
Aug 30 20:26:36 <denkenz> You have to start the demo app every time you boot phonesim
Aug 30 20:27:19 <denkenz> Otherwise its just those 4 steps, enable, online, start sim app, test-stk-menu
Aug 30 20:45:06 <kristenc> maybe I'm just expecting something that doesn't exist.
Aug 30 20:48:15 <denkenz> what are you expecting?
Aug 30 20:56:14 <kristenc> I guess I was expecting some kind of indication on the gui that I did something.
Aug 30 20:56:27 <kristenc> or maybe some sort of submenu.
Aug 30 21:04:07 <denkenz> kristenc: In phonesim itself?
Aug 30 21:04:09 <denkenz> nope
Aug 30 21:05:18 <kristenc> I get nothing beyond what is in the terminal where I ran test-stk-menu.  And that output looks like this:
Aug 30 21:05:34 <kristenc> Items:
Aug 30 21:05:34 <kristenc> 0. News
Aug 30 21:05:34 <kristenc> 1. Sports
Aug 30 21:05:34 <kristenc> 2. Time
Aug 30 21:05:34 <kristenc> 3. Sticks Game
Aug 30 21:05:35 <kristenc> 4. Tones
Aug 30 21:05:37 <kristenc> 5. Icons (not self-explanatory)
Aug 30 21:05:39 <kristenc> 6. Icons (self-explanatory)
Aug 30 21:05:43 <kristenc> 7. Finance
Aug 30 21:05:43 <kristenc> 8. Web Browser
Aug 30 21:05:44 <kristenc> Enter Selection: 5
Aug 30 21:05:48 <kristenc> Agent registered for session
Aug 30 21:05:50 <kristenc> Title: (Icons)
Aug 30 21:05:52 <kristenc> 0. Basic Icon
Aug 30 21:05:54 <kristenc> 1. Color Icon
Aug 30 21:05:57 <kristenc> 2. Bad Icon
Aug 30 21:05:58 <kristenc> Default: -1
Aug 30 21:06:01 <kristenc> Enter Selection (t, b):t
Aug 30 21:06:03 <kristenc> Release
Aug 30 21:06:04 <kristenc> what does the t, b mean anyway?
Aug 30 21:09:27 <denkenz> t -> Terminate
Aug 30 21:09:29 <denkenz> b -> Back
Aug 30 21:09:48 <denkenz> kristenc: So I'm not sure what you're expecting
Aug 30 21:11:01 <kristenc> denkenz: I guess I thought I could select 0, 1, or 2.  And then have it display something.
Aug 30 21:11:49 <kristenc> but I guess all I can do is terminate or go back.  That part works :).
Aug 30 21:12:06 <denkenz> Whatever you pick sends you to the main menu
Aug 30 21:12:39 <kristenc> well, selecting 'b' actually doesn't go back, it terminates too.
Aug 30 21:12:57 <denkenz> Yep, that's just how the application is written
Aug 30 21:13:08 <denkenz> see phonesim/src/simapplication.cpp
Aug 30 21:13:28 <denkenz> The main test there is that you display the icons in the SelectItem menu
Aug 30 21:13:29 <kristenc> and selecting 0, 1, 2 doesn't send me to the main menu, it just re-displays the icon menu.
Aug 30 21:13:44 <denkenz> yep, fair enough
Aug 30 21:14:15 <denkenz> void DemoSimApplication::iconMenu( const QSimTerminalResponse& resp )
Aug 30 21:14:15 <denkenz> {
Aug 30 21:14:16 <denkenz>     if ( resp.result() == QSimTerminalResponse::Success )
Aug 30 21:14:16 <denkenz>         sendIconMenu();
Aug 30 21:14:16 <denkenz>     else
Aug 30 21:14:17 <denkenz>         endSession();
Aug 30 21:14:17 <denkenz> }
Aug 30 21:15:39 <kristenc> ok, well it seems to be behaving as expected
Aug 31 00:05:25 <padovan> denkenz: ping
Aug 31 00:06:35 <denkenz> padovan: pong
Aug 31 00:08:38 <padovan> denkenz: should ofono_gprs_add_context go into DUN post_sim()? I'm having problems because post_sim is not getting called.
Aug 31 00:09:19 <denkenz> for you it doesn't matter
Aug 31 00:09:30 <denkenz> Try latest upstream though, that part should be fixed
Aug 31 00:23:23 <holtmann> denkenz: What are you doing with the D-Bus policy change here. You don't need to change anything.
Aug 31 00:44:36 <denkenz> holtmann: What are you referring to?
Aug 31 00:44:52 <holtmann> The ofono.rules that you proposed to Kristen.
Aug 31 00:44:58 <holtmann> What are you doing there?
Aug 31 00:45:26 <denkenz> err, pasting?
Aug 31 00:46:02 <holtmann> <denkenz> 19:19:31 You need the dbus policy hole
Aug 31 00:46:02 <holtmann> <denkenz> 19:20:08 Copy the latest src/ofono.conf to your dbus system.d directory
Aug 31 00:46:03 <holtmann> <denkenz> 19:20:18 Then change the user from root to whatever
Aug 31 00:46:03 <holtmann> <denkenz> 19:20:22 If you're using it as non-root
Aug 31 00:46:05 <holtmann> <denkenz> 19:21:12   <policy user="denkenz">
Aug 31 00:46:05 <holtmann> <denkenz> 19:21:12     <allow own="org.ofono"/>
Aug 31 00:46:07 <holtmann> <denkenz> 19:21:12     <allow send_destination="org.ofono"/>
Aug 31 00:46:07 <holtmann> <denkenz> 19:21:13     <allow send_interface="org.ofono.SimToolkitAgent"/>
Aug 31 00:46:09 <holtmann> <denkenz> 19:21:13   </policy>
Aug 31 00:46:52 <zhenhua> what's wrong?
Aug 31 00:48:15 <holtmann> This is only needed if you try to run ofonod as non-root user.
Aug 31 00:48:36 <holtmann> What is wrong with just using "sudo env OFONO_AT_DEBUG=1 ./src/ofonod -n"
Aug 31 00:48:58 <holtmann> If you try to punch hole for user access, then it is a different one.
Aug 31 00:49:07 <holtmann>   <policy at_console="true">
Aug 31 00:49:08 <holtmann>     <allow send_destination="org.ofono"/>
Aug 31 00:49:08 <holtmann>   </policy>
Aug 31 00:49:27 <holtmann> That needs to be changed to say user="denkenz" instead of at_console="true".
Aug 31 00:49:33 <holtmann>   <policy context="default">
Aug 31 00:49:34 <holtmann>     <deny send_destination="org.ofono"/>
Aug 31 00:49:34 <holtmann>   </policy>
Aug 31 00:49:42 <holtmann> Or the deny here needs to be changed to allow.
Aug 31 00:50:08 <holtmann> denkenz: Btw. I have the GDBus security patches and PolicyKit integration almost ready.
Aug 31 00:50:42 <holtmann> That is really the way to go here.
Aug 31 00:50:52 <holtmann> D-Bus policy file security is just broken by design.
Aug 31 00:51:05 <holtmann> The only needed hole should be the own statement for the daemon.
Aug 31 00:59:50 <denkenz> holtmann: I'm lost, I'm running oFono as regular joe user
Aug 31 01:04:35 <holtmann> Hah. Okay. I am just using sudo.
Aug 31 01:51:02 <yang_office> denkenz: I have several questions: 1) Do we need to support TON/NPI when constructing a ss control string? 2) In stk's send ss, the specs says "if the command is rejected because the ME is busy on a USSD transaction, the ME shall inform the UICC using TERMINAL RESPONSE". I don't understand why send ss has interference with send ussd. Does it mean when sending ss, I have to look at ussd atom first. If it's busy, I need to send a "busy" response to SIM. 3) I
Aug 31 01:53:38 <denkenz> I suspect the reason is that generic USSDs and CCFC/CCWA/CLIP/COLP/COLR/CLIR are done using a lower level supplementary service request / response protocol
Aug 31 01:54:02 <denkenz> So it might be that they do indeed interfere
Aug 31 01:55:06 <denkenz> TON/NPI is not needed for USSD though
Aug 31 01:55:42 <yang_office> ok, I will currently ignore the TON/NPI of send ss.
Aug 31 02:00:04 <yang_office> As to the interference of ss and ussd, is it better to use a single atom to cover them all?
Aug 31 02:00:45 <denkenz> yang_office: Actually it might be needed for call-forwarding
Aug 31 02:00:45 <yang_office> Current code have some problem: If we use ussd_initiate() to send some ss, we will look at if ussd is busy or not first.
Aug 31 02:01:00 <denkenz> refer to call-forwarding driver / core
Aug 31 02:01:42 <yang_office> Thanks, I will look at the code.
Aug 31 02:02:01 <denkenz> ignore the interaction between ussd/ss
Aug 31 02:02:12 <denkenz> Since we don't know, we will simply have to find out at some point
Aug 31 02:02:43 <yang_office> ok, that will be simple now:)
Aug 31 02:03:32 <denkenz> If they indeed interfere, its easy to fix
Aug 31 02:15:48 <balrog-k1n> "Send USSD" also has a note about responding BUSY if terminal is executing ussd or ss
Aug 31 02:22:07 <denkenz> Then that's what we should do
Aug 31 02:32:40 <balrog-k1n> (same as "Sens SS", was my point) :)
Aug 31 02:33:00 <balrog-k1n> yang_office: your question got cut off at 3)
Aug 31 02:57:40 <yang_office> denkenz: balrog-k1n: Sorry, Just come back from a meeting. 
Aug 31 02:58:09 <yang_office> My third question is: 3) In ussd_initiate(), I think if (ussd->pending) and if (ussd->state != ....) should be behind the recognized_control_string(). 
Aug 31 02:58:44 <yang_office> I know the answer to this question now.
**** ENDING LOGGING AT Tue Aug 31 02:59:57 2010