**** BEGIN LOGGING AT Fri Aug 20 02:59:56 2010
Aug 20 04:24:28 <zhenhua1> denkenz: i run simple-agent with phonesim. after choose '8. Web Browser', it tells me 'Agent registered for session'. and nothing come back from phonesim. what should i do?
Aug 20 04:25:12 <denkenz> we don't support launch browser, ignore it
Aug 20 04:25:29 <zhenhua1> i also choose 0.News. the result is the same
Aug 20 04:25:38 <zhenhua1> which one is workable?
Aug 20 04:26:09 <zhenhua1> i can see 8 sub items from main menu now
Aug 20 04:26:27 <zhenhua1> sorry, 0~8, totally 9
Aug 20 04:27:24 <denkenz> dunno, stick to the other ones
Aug 20 04:28:00 <zhenhua1> okay, will explorer more deeper
Aug 20 04:46:55 <zhenhua1> denkenz: http://pastebin.com/YfJ3EsPc
Aug 20 04:47:52 <zhenhua1> denkenz: use mbm f3607gw and found a crash in mbm_stk_terminal_response().
Aug 20 04:54:07 <denkenz> balrog-k2n: ^^^^^
Aug 20 04:55:40 <zhenhua1> using no matter which sim card, i cann't register simple-agent still
Aug 20 04:55:58 <zhenhua1> IdleModeText is always NULL
Aug 20 04:56:13 <zhenhua1> but i am pretty sure i have some stk apps there
Aug 20 05:46:53 <kvalo> holtmann: I don't recall seeing anything related STIN
Aug 20 09:27:20 <holtmann> ofonod[14676]: Modem: < \r\n*ESTKSMS: ages/Error.png,0,0\r\n
Aug 20 09:27:20 <holtmann> ofonod[14676]: Data: < \r\n*ESTKSMS: ages/Error.png,0,0\r\n
Aug 20 09:27:33 <holtmann> Maybe me T-Mobile SIM has images ;)
Aug 20 12:28:21 <balrog-k2n> zhenhua1: IdleModeText is not used often, even if your card has STK applications
Aug 20 12:28:40 * balrog-k2n looks at the crash
Aug 20 12:50:11 <balrog-k2n> so your card re-sends the same Display Text command three times at the beginning and then a Send SMS, completely unasked
Aug 20 12:57:43 <balrog-k2n> possibly triggered by ofono's answering Terminal Not Capable to the Display Text commands
Aug 20 13:02:35 <holtmann> Title: (Extras)
Aug 20 13:02:35 <holtmann> 0. Banking
Aug 20 13:02:35 <holtmann> 1. Menü-Update
Aug 20 13:02:35 <holtmann> 2. Info
Aug 20 13:02:37 <holtmann> Default: -1
Aug 20 13:02:37 <holtmann> Enter Selection (t, b):2
Aug 20 13:02:39 <holtmann> DisplayText ((c) T-Mobile Version 2.00, 1)
Aug 20 13:02:39 <holtmann> Release
Aug 20 13:02:42 <holtmann> DisplayText works for me ;)
Aug 20 13:03:49 <holtmann> ofonod[21905]: Modem: < \r\n*STKI:"D0258103012181820281028D1A0428632920542D4D6F62696C652056657273696F6E20322E3030"\r\n
Aug 20 13:03:49 <holtmann> ofonod[21905]: drivers/mbmmodem/stk.c:stki_notify() 
Aug 20 13:03:49 <holtmann> ofonod[21905]: src/stk.c:handle_command_display_text() Immediate Response: 0
Aug 20 13:03:49 <holtmann> ofonod[21905]: src/stk.c:send_simple_response() result 0
Aug 20 13:03:49 <holtmann> ofonod[21905]: src/stk.c:stk_respond() 
Aug 20 13:03:50 <holtmann> ofonod[21905]: drivers/mbmmodem/stk.c:mbm_stk_terminal_response() 
Aug 20 13:03:52 <holtmann> ofonod[21905]: Modem: > AT*STKR="810301218182028281830100"\r
Aug 20 13:03:54 <holtmann> ofonod[21905]: Data: < \r\n*ESTKDISP: 1,1,0,(c) T-Mobile Version 2.00\r\n
Aug 20 13:03:56 <holtmann> ofonod[21905]: Modem: < \r\nOK\r\n
Aug 20 13:03:58 <holtmann> ofonod[21905]: drivers/mbmmodem/stk.c:mbm_stkr_cb() 
Aug 20 13:04:00 <holtmann> ofonod[21905]: src/stk.c:stk_command_cb() TERMINAL RESPONSE to a command reported no errors
Aug 20 13:04:02 <holtmann> ofonod[21905]: Modem: < \r\n*ESTKDISP: 1,1,0,(c) T-Mobile Version 2.00\r\n
Aug 20 13:04:04 <holtmann> ofonod[21905]: Modem: < \r\n*STKEND\r\n
Aug 20 13:07:50 <balrog-k2n> zhenhua1: i sent a patch to fix that crash, thanks for the backtrace
Aug 20 13:08:16 <balrog-k2n> the message the SIM sends on startup is 尊敬的全球通客户：感谢您选择中国移动通信！全球通助您成功
Aug 20 13:08:41 <balrog-k2n> (google) Dear GSM customers: Thank you for choosing China Mobile! Global links help you succeed
Aug 20 13:09:32 <holtmann> balrog-k2n: Nice.
Aug 20 13:09:48 <balrog-k2n> i wonder if the UCS to UTF8 conversion in ofono works
Aug 20 13:09:59 <holtmann> Any idea on how to get the STK init done without hardreset. E2REST doesn't work for me at all.
Aug 20 13:10:52 <balrog-k2n> holtmann: i've tried everything that was in the docs but didn't find a way to reset the sim without resetting the modem
Aug 20 13:11:24 <balrog-k2n> for example +CFUN=1,1 resets the entire modem so it gets reenumerated on USB
Aug 20 13:55:55 <denkenz> balrog-k2n: It should, since you wrote a ton of unit tests for it
Aug 20 13:56:53 <holtmann> denkenz: I have been tested netreg and gprs for the whole day so far.
Aug 20 13:56:56 <holtmann> Seems to be pretty fine.
Aug 20 13:57:09 <holtmann> I do have issues with my new micro SIM from T-Mobile.
Aug 20 13:57:16 <holtmann> Sometimes I get "denied" :(
Aug 20 14:01:58 <denkenz> yeah, that is common
Aug 20 14:02:12 <denkenz> Probably something on the backend gets funny
Aug 20 14:02:31 <denkenz> I get denied while roaming quite often
Aug 20 14:07:13 <denkenz> And damn, your cops fix for nokia is nasty
Aug 20 14:07:22 <denkenz> But then that modem should be tossed
Aug 20 14:07:40 <holtmann> I told you that that one is nasty.
Aug 20 14:07:52 <holtmann> I am at home. And T-Mobile is my home network :(
Aug 20 14:08:38 <denkenz> Btw, list-operators scan already does what scan-operators seems to do
Aug 20 14:09:09 <holtmann> list-operators and get-operators are mostly the same. I realized that after I added it.
Aug 20 14:12:03 <denkenz> Rest looks fair enough
Aug 20 14:16:28 <holtmann> ofonod[23197]: Modem: < \r\n+CIEV: 2,4\r\n
Aug 20 14:16:28 <holtmann> ofonod[23197]: Modem: < \r\n+C
Aug 20 14:16:28 <holtmann> ofonod[23197]: Modem: < IEV: 2,5\r\n
Aug 20 14:16:28 <holtmann> ofonod[23197]: Modem: < \r\n+CIEV: 2,4\r\n
Aug 20 14:16:45 <holtmann> Hah. These ones are a serious problem when not using a ringbuffer and only regex ;)
Aug 20 14:31:11 <holtmann> Even if an EF can be uniquely identified by its <fileid> alone, if the correct directory has not been selected by *ECDF then the file will not be accessible.
Aug 20 14:31:19 <holtmann> denkenz: You read this one?
Aug 20 14:31:28 <holtmann> Maybe that is why the icon stuff fails.
Aug 20 14:34:24 <denkenz> whats ECDF?
Aug 20 14:34:30 <denkenz> And where is this from?
Aug 20 14:34:38 <holtmann> MBM manual.
Aug 20 14:35:23 <denkenz> There's no syntax of ECDF tho
Aug 20 14:37:29 <denkenz> And on my F3507 AT*ECDF=? returns error
Aug 20 14:39:38 <denkenz> However, the icons are definitely there on SIMs
Aug 20 14:39:56 <denkenz> I think AT&T even ships a carrier logo on theirs
Aug 20 15:07:23 <holtmann> The F36xx that I have also doesn't support AT*ECDF
Aug 20 15:11:16 <denkenz> It might be an STE command
Aug 20 15:11:37 <denkenz> I noticed a few of these in MBm manuals and vice versa in STE
Aug 20 16:44:27 <mikeleib> denkenz: http://www.amazon.com/Option-iCON-452-USB-Modem/dp/tech-data/B002Y2THNS/ref=de_a_smtd
Aug 20 16:44:30 <mikeleib> ought to work, right?
Aug 20 16:44:54 <denkenz> yah
Aug 20 16:45:08 * mikeleib sends email to boss asking for a slew of them
Aug 20 16:45:17 <mikeleib> what is best way to get SIM cards?
Aug 20 16:47:25 <denkenz> go to T-Mobile or AT&T store and buy a bunch of prepaid numbers
Aug 20 16:47:49 <mikeleib> don't they ship you to gitmo for that sort of thing now?
Aug 20 16:48:19 <denkenz> Dunno, maybe
Aug 20 16:48:53 <denkenz> Alternatively have your manager get a corporate plan
Aug 20 16:49:42 <denkenz> I buy prepaid SIMs from every carrier in every country I visit
Aug 20 16:54:46 * mikeleib will send travel request to visit countries and then buy sims ;)
Aug 20 16:55:07 <denkenz> Tell me how that works out
Aug 20 16:55:18 <denkenz> I'll do the same if you're successful
Aug 20 17:48:22 <denkenz> holtmann: So looks like the next gen mbm is the C3607w
Aug 20 17:48:30 <denkenz> I can't find any devices with it though
Aug 20 19:05:42 <alt-route> Hi, i have a beginner question about dial modifiers, such as "g" or "i". Do those have to come at the end of the dial string (as seen in the phonesim callmanager.cpp)? Or can it be added in the beginning, like ATDG123454 ?
Aug 20 19:10:34 <denkenz> see 3GPP 27.007 for a very exhaustive answer.  They're given at the end
Aug 20 19:11:00 <alt-route> ah great. Thanks!
Aug 20 19:22:54 <holtmann> denkenz: So your ZTE dongle should work fine now.
Aug 20 19:23:06 <holtmann> Besides the stupid cutting of USSD messages ;)
Aug 20 19:23:11 <holtmann> Just pushed its GPRS bits.
Aug 20 19:24:07 <denkenz> nice
Aug 20 19:24:32 <holtmann> My Nokia one doesn't wanna play with my new microSIMs.
Aug 20 19:26:01 <holtmann> ofonod[30201]: Control:> AT+CMGL=4\r
Aug 20 19:26:01 <holtmann> ofonod[30201]: Control:< \r\n+CMGL: 1,3,,76\r\n0001FF008100004ECD72DA5D06B9CBF532885A6697CD6FB7BBDE6E97E5A0F49C0E82C56EB6D9CD868BD970381728394683CCF272BD0C6AA7C768905F5C9683CA697719E40C8FD1F2F4184D7701\r\n\r\nOK\r\n
Aug 20 19:26:12 <holtmann> Btw. what is this and why don't we decode it?
Aug 20 19:26:50 <denkenz> we decode it, it is an SMS
Aug 20 19:28:51 <holtmann> I see nothing in the logs.
Aug 20 19:29:14 <holtmann> Or is this the Internet settings binary message.
Aug 20 19:29:41 <denkenz> Dunno, check if the parser is confused
Aug 20 19:29:52 <denkenz> But we decode all CMGLs on other modems
Aug 20 19:29:59 <denkenz> it gets notified as a simple pdu
Aug 20 19:30:52 <holtmann> Normally I see the text output from the history plugin in the logs.
Aug 20 19:31:16 <denkenz> Yeah, but it could be just garbage
Aug 20 19:31:58 <denkenz> Next GSOC we get a student to write a bunch of tools for us
Aug 20 19:32:09 <denkenz> Like a proper hexpdu dumper, etc
Aug 20 19:37:12 <holtmann> www.nobbi.de
Aug 20 19:37:35 <holtmann> www.nobbi.com
Aug 20 19:38:54 <denkenz> Yah but that requires wincrap
Aug 20 19:39:23 <holtmann> PDUspy, but yes, we need a Linux version.
Aug 20 19:52:04 <holtmann> denkenz: What is SMS status 3?
Aug 20 19:52:11 <holtmann> Why do we ignore these?
Aug 20 19:52:37 <holtmann> Maybe this wasn't even mine since I bought that dongle on ebay ;)
Aug 20 19:52:59 <denkenz> status 3?
Aug 20 19:54:19 <denkenz> I've no idea what you're talking about :)
Aug 20 19:57:15 <holtmann> Nevermind. I just deleted that message.
Aug 20 19:59:42 <denkenz> There are Class 3 SMS which are meant for the Modem firmware
Aug 20 19:59:51 <denkenz> If we fail to decode them, we just toss
Aug 20 20:00:21 <holtmann> Anyhow, I am done with coding for today.
Aug 20 20:00:32 <holtmann> I pushed everything that I had pending from yesterday.
Aug 20 20:06:01 <denkenz> Hm, wish I could say the same :P
Aug 20 20:10:01 <holtmann> Only Sierra is missing. And mine only works in the US on AT&T. So I can't look into that one.
Aug 20 20:53:33 <zhenhua1> balrog-k2n: thanks for the fixing so quickly. google translate is good. i often use it to translate Egnlish idioms into chinese. ;-)
Aug 20 20:55:08 <zhenhua1> like pipe dream or cross that bridge:P
Aug 20 21:17:01 <denkenz> holtmann: Ok, last of API changes are done
Aug 20 21:17:13 <denkenz> They are valgrind clean and seem to work as intended
Aug 20 21:17:20 <holtmann> Nice.
Aug 20 21:18:08 <denkenz> The only other thing on my immediate todo is an implementation for ConnectionManager.DeactivateAll
Aug 20 21:18:12 <denkenz> but that can wait at this point
Aug 20 21:18:21 <holtmann>   CC     src/voicecall.o
Aug 20 21:18:21 <holtmann> cc1: warnings being treated as errors
Aug 20 21:18:21 <holtmann> src/voicecall.c: In function ‘manager_get_properties’:
Aug 20 21:18:21 <holtmann> src/voicecall.c:863: error: unused variable ‘callobj_list’
Aug 20 21:18:55 <denkenz> ah crap :P
Aug 20 21:18:56 <holtmann> Pushed a fix for it :)
Aug 20 21:19:34 <holtmann> ConnMan.DeactivateAll sounds like a good idea.
Aug 20 21:19:46 <denkenz> ok, I actually had a sigsegv during this round that I had to debug ;)
Aug 20 21:20:03 <holtmann> How do you like the backtrace feature ;)
Aug 20 21:20:33 <denkenz> Haven't played with it honestly, need to look deeper
Aug 20 21:20:49 <holtmann> If you had a segfault, you should have seen it already.
Aug 20 21:21:24 <denkenz> Nope, it just segfaulted like it always does
Aug 20 21:21:36 <denkenz> But I was running release
Aug 20 21:22:12 <denkenz> DeactivateAll was meant as a feature for multi-context
Aug 20 21:22:29 <denkenz> Need to get it working anyway though
Aug 20 21:23:14 <holtmann> ofonod[11466]: Aborting (signal 11)
Aug 20 21:23:14 <holtmann> ofonod[11466]: ++++++++ backtrace ++++++++
Aug 20 21:23:14 <holtmann> ofonod[11466]: [0]: /lib64/libpthread.so.0() [0x31a980f440]
Aug 20 21:23:14 <holtmann> ofonod[11466]: [1]: ./src/ofonod() [0x46a0f2]
Aug 20 21:23:16 <holtmann> ofonod[11466]: [2]: ./src/ofonod() [0x40c023]
Aug 20 21:23:16 <holtmann> ofonod[11466]: [3]: /lib64/libdbus-1.so.3() [0x31af01cd3e]
Aug 20 21:23:18 <holtmann> ofonod[11466]: [4]: /lib64/libdbus-1.so.3(dbus_connection_dispatch+0x37c) [0x31af010a1c]
Aug 20 21:23:18 <holtmann> ofonod[11466]: [5]: ./src/ofonod() [0x40b6f8]
Aug 20 21:23:20 <holtmann> ofonod[11466]: [6]: /lib64/libglib-2.0.so.0() [0x31aa83c4cb]
Aug 20 21:23:22 <holtmann> ofonod[11466]: [7]: /lib64/libglib-2.0.so.0(g_main_context_dispatch+0x1f2) [0x31aa83bd02]
Aug 20 21:23:24 <holtmann> ofonod[11466]: [8]: /lib64/libglib-2.0.so.0() [0x31aa83fae8]
Aug 20 21:23:26 <holtmann> ofonod[11466]: [9]: /lib64/libglib-2.0.so.0(g_main_loop_run+0x195) [0x31aa83fff5]
Aug 20 21:23:28 <holtmann> ofonod[11466]: [10]: ./src/ofonod() [0x44afb7]
Aug 20 21:23:30 <holtmann> ofonod[11466]: [11]: /lib64/libc.so.6(__libc_start_main+0xfd) [0x31a901ec5d]
Aug 20 21:23:34 <holtmann> ofonod[11466]: [12]: ./src/ofonod() [0x40b019]
Aug 20 21:23:36 <holtmann> ofonod[11466]: +++++++++++++++++++++++++++
Aug 20 21:24:21 <holtmann> And if you run it through Samuel's backtrace script:
Aug 20 21:24:22 <holtmann> -------- backtrace --------
Aug 20 21:24:22 <holtmann> [0]: stk_get_properties() [stk.c:420]
Aug 20 21:24:22 <holtmann> [1]: generic_message() [object.c:262]
Aug 20 21:24:22 <holtmann> [2]: message_dispatch() [mainloop.c:80]
Aug 20 21:24:25 <holtmann> [3]: example_history_call_ended() [example_history.c:60]
Aug 20 21:24:25 <holtmann> [4]: _start() [stkagent.c:0]
Aug 20 21:24:27 <holtmann> [5]: stk_select_item() [stk.c:626]
Aug 20 21:24:28 <holtmann> [6]: generic_message() [object.c:262]
Aug 20 21:24:30 <holtmann> [7]: message_dispatch() [mainloop.c:80]
Aug 20 21:24:34 <holtmann> [8]: main() [main.c:229]
Aug 20 21:24:36 <holtmann> [9]: _start() [stkagent.c:0]
Aug 20 21:24:38 <holtmann> ---------------------------
Aug 20 21:24:41 <denkenz> sweet
Aug 20 21:25:10 <holtmann> Pretty neat.
Aug 20 21:25:29 <zhenhua1> where can i get the script?
Aug 20 21:25:41 <holtmann> All in the source code. test/backtrace.
Aug 20 21:25:45 <holtmann> Same for ConnMan btw.
Aug 20 21:26:09 <zhenhua1> we should copy it to ofono
Aug 20 21:26:17 <zhenhua1> looks so good
Aug 20 21:26:18 <holtmann> It is there.
Aug 20 21:26:48 <sameo> holtmann: sorry for not building an ofono patch btw...
Aug 20 21:26:58 <holtmann> No problem. I just copied your work.
Aug 20 21:27:12 <holtmann> Took me just a few minutes.
Aug 20 21:27:24 <sameo> good.
Aug 20 21:27:44 <holtmann> Pretty nice idea for getting backtrace from a real life system.
Aug 20 21:27:56 <holtmann> Since all this abrt or corewatcher stuff is not really working out.
Aug 20 21:28:26 <holtmann> If you could compile the code so that it resolves static functions as well, that would be sweet.
Aug 20 21:28:27 <sameo> QA crash reports inspired me: "ConnMan crashes sometimes."
Aug 20 21:28:50 <sameo> holtmann: yes, that would be nice, but I really think you can't
Aug 20 21:29:19 <holtmann> In theory if you disable all optimization and uninline everything and maybe some other magic.
Aug 20 21:29:26 <holtmann> But then the binary gets pretty crappy.
Aug 20 21:29:41 <holtmann> And we loose all the compiler warnings from -O2.
Aug 20 21:30:14 <sameo> I tried that, it seems there's no way to tell the linker to keep static functions name around. At least I haven't found it.
Aug 20 21:30:43 <sameo> the backtrace post processing sounded like a good compromise.
Aug 20 21:31:22 <holtmann> In theory it should be possible if addr2line can do. So should be glibc.
Aug 20 21:31:34 <denkenz> 206 commits so far this release
Aug 20 21:31:38 <denkenz> That's just nuts
Aug 20 21:31:52 <holtmann> Essentially you have a copy of your binary in memory.
Aug 20 21:33:11 <zhenhua1> even O2, the static function symbol should still there, if you add -g2
Aug 20 21:33:11 <sameo> oh, I'm not saying it's impossible, but backtrace() from glibc doesnt support it. You'd have to start manually parsing the ELF tables to do what addr2line does.
Aug 20 21:33:43 <zhenhua1> you can use readelf or objdump to get fun name
Aug 20 21:34:01 <zhenhua1> what's the problem? maybe i missed something
Aug 20 21:34:54 <sameo> zhenhua1: we'd like to be able to do it from the binary itself
Aug 20 21:35:38 <zhenhua1> that's addr2line's job as in your script
Aug 20 21:36:00 <zhenhua1> gcc has -fno-inline option
Aug 20 21:36:28 <sameo> exactly. What we'd like to do is have the oFono's signal handler actually resolving the static function names, and not having to call a post processing script
Aug 20 21:38:14 <zhenhua1> that' means ofono itself integrate your backstrace?
Aug 20 21:38:59 <sameo> not integrating, but would be able to do what addr2line does
Aug 20 21:39:31 <zhenhua1> it could catch SIGxxx in main event loop and do something
Aug 20 21:39:53 <sameo> that's what it does :)
Aug 20 21:40:20 <zhenhua1> dunno how to do that yet.  using C?;-)
Aug 20 21:40:32 <sameo> yes
Aug 20 21:40:52 <zhenhua1> i don't see any prog do that yet
Aug 20 21:41:31 <zhenhua1> let me read APUE if have time later
Aug 20 22:18:24 <holtmann> sameo: I think the only way is to open the binary from /proc/<pid>/exe points to at program loading time. And then create our own lookup table for static functions.
Aug 20 22:19:55 <sameo> holtmann: maybe not the _only_ way, but that would be a simple one.
Aug 20 22:20:37 <holtmann> Nasty part is that backtrace doesn't give us the addresses as addresses, they come as strings.
Aug 20 22:21:21 <sameo> holtmann: no, you can get them as addresses as well, iirc.
Aug 20 22:22:24 <sameo> dladdr() should do that for you.
Aug 20 22:29:17 <zhenhua1> so could dladdr resolve static function address?
Aug 21 00:05:51 <luke-jr> zhenhua1: static functions don't *have* addresses
Aug 21 00:05:58 <luke-jr> wait, I'm thinking inline XD
Aug 21 00:06:09 <luke-jr> probably not
Aug 21 00:06:14 <luke-jr> maybe if it has debug info
Aug 21 00:56:05 <zhenhua1> luke-jr: static functions have addresses unless it is inlined
Aug 21 00:56:49 <zhenhua1> otherwise you can't jump
**** ENDING LOGGING AT Sat Aug 21 02:59:57 2010