Re: [hylafax-users] COMREC invalid response received

[Lee Howard <faxguy@xxxxxxxxxxxxxxxx>]

Hans Klose wrote:
> i use t38modem with hylafax. From some stations i have problems to
> send to the hylafaxserver. In the log comes "COMREC invalid response
> received". When ther are 2 pages only one will be received.
> I attached a sample logfile. Is that behavior an issue from 
> t38modem or can i modify hylafax to make it work?

> Apr 16 14:13:03.46: [ 3609]: RECV recv RR (receive ready)
> Apr 16 14:13:03.46: [ 3609]: <-- [9:AT+FRS=7\r]
> Apr 16 14:13:09.27: [ 3609]: --> [2:OK]
> Apr 16 14:13:09.27: [ 3609]: RECV send MCF (message confirmation)
> Apr 16 14:13:09.27: [ 3609]: <-- [9:AT+FTH=3\r]
> Apr 16 14:13:09.27: [ 3609]: --> [7:CONNECT]
> Apr 16 14:13:09.27: [ 3609]: RECV FAX (000000057): from <UNSPECIFIED>, page 1 in 0:00:51, INF, 3.85 line/mm, 1-D MH, 14400 bit/s
> Apr 16 14:13:09.27: [ 3609]: <-- HDLC<3:FF C8 31>
> Apr 16 14:13:09.27: [ 3609]: <-- data [3]
> Apr 16 14:13:09.27: [ 3609]: <-- data [2]
> Apr 16 14:13:10.86: [ 3609]: --> [2:OK]
> Apr 16 14:13:10.86: [ 3609]: <-- [11:AT+FRM=146\r]
> Apr 16 14:13:15.90: [ 3609]: --> [8:+FCERROR]
> Apr 16 14:13:15.90: [ 3609]: <-- [11:AT+FRM=146\r]
> Apr 16 14:13:15.90: [ 3609]: --> [8:+FCERROR]
> Apr 16 14:13:15.90: [ 3609]: <-- [9:AT+FRH=3\r]
> Apr 16 14:13:15.90: [ 3609]: --> [7:CONNECT]
> Apr 16 14:13:16.86: [ 3609]: --> HDLC<5:FF C8 F6 EC 24>
> Apr 16 14:13:16.86: [ 3609]: --> [2:OK]
> Apr 16 14:13:16.86: [ 3609]: RECV recv RR (receive ready)

The sender did not hear the MCF signal.  So it retransmitted RR.

The long wait for the +FRS=7 result at 14:13:09.27 worries me that there 
is something else going wrong.  However, here's a patch to HylaFAX which 
should make it capable of continuing the protocol recovery attempt.

Thanks,

Lee.

--- hylafax.orig/faxd/Class1Recv.c++	2010-03-25 12:14:48.000000000 -0700
+++ hylafax/faxd/Class1Recv.c++	2010-04-25 21:14:24.615838544 -0700
@@ -1019,6 +1019,8 @@
 		    }
 		}
 		break;
+	    case FCF_RR:
+		// The sender did not hear our MCF signal.  Treat it like CRP.
 	    case FCF_CRP:
 		// command repeat... just repeat whatever we last sent
 		if (!useV34 && !switchingPause(emsg)) return (false);