RE: Daily patch: gateway

> -----Original Message-----
> From: kannel [mailto:[email protected]]
> Sent: Saturday, March 09, 2002 8:30 AM
> To: [email protected]
> Subject: Daily patch: gateway

> Index: gateway/ChangeLog
> diff -u gateway/ChangeLog:1.1704 gateway/ChangeLog:1.1708
> --- gateway/ChangeLog:1.1704	Thu Mar  7 21:01:03 2002
> +++ gateway/ChangeLog	Fri Mar  8 22:28:36 2002
> @@ -1,3 +1,48 @@
> +2002-03-08 Andreas Fink <[email protected]>
> +    * gw/smsc_at2.c: fixed a bug for +CPIN which wasnt 
> waiting for OK after
> +    +CPIN answer. This was screwing up AT command answers 
> and Siemens TC35
> +    couldnt get initialized anymore.

I don't think that this solution it correct - as some modems do not send
an OK after +CPIN:READY, and AFAIR, the standard does not require it. as
the modem does not send an OK, wait_modem_command() will return a
timeout, which  init_device() will parse as an error and therby fail the
initialization. this has been discussed before and that's the reason
that the ret = 4 line was introduced. you can go around making patches
that are known to break some implemenations just because it will make
your implemenation work. 
I assert that both solutions are wrong and we need to come up with a
third one. how about calling wait_modem_command() once after recieving
+CPIN:READY if the modem is of a type that sends an OK after that
message ? (or just call it anyway and discard the return value ? at most
it will slow some modems' initialization by the call timeout).

--
Oded Arbel
m-Wise Inc.
[email protected]

Experience is what causes a person to make new mistakes instead of old
ones.