🚀 go-pugleaf

Here's a problem some of you might recognize and hopefully tell me what is 
happening. 

Back in the distant past I started writing a package.  This package, fully 
asynchronous and multitasking, eventually was to be used as a process control 
package.  To do this, I needed to write a serial port handler that talked to 
serial.device.  I did this.  To test it, I connected two Amigas via a null 
modem, and shipped messages back and forth, thus verifying that my 
communications worked and my protocol was correct.

Time passed, and we got to now.  My code was sufficiently mature, so I hooke 
up to a piece of hardware that I really intend to control in the real world.  
This box communicates over a serial port using 7-O-1.  Welllll..... I did my 
original testing at 8-N-1.  No problem,  just change the flags and come up as 
7-O-1.  Easy as pie.

Except... at 7-O-1, instead of every byte of my message being sent out the 
serial port, every other byte is being sent out.  Here is the relevant code:

char Dataout[100]

main()

(a bunch of setup code)

Write_Req->IOSer.io_Data = (APTR)Dataout;
Write_Req->io_SerFlags = SERF_XDISABLED|SERF_EOFMODE|SERF_PARTY_ODD|
                          SERF_PARTY_ON;

At the present, I am working around this problem by pushing my outgoing 
message into every other byte of Dataout, and that seems to work reliably.  
BUT...clearly something is wrong.

My question is this:  What set of conditions would cause serial.device to send 
every byte at 8-N-1, but every other byte at 7-O-1?

Jim Locker   <puzzled but pushing on>


--  
------------------------------------------------------------------------
AFIT Amiga Users BBS/UFGateway   |Jim Locker - via FidoNet node 1:110/300
    1:110/300 Dayton, Ohio       |UUCP: uunet!dayvb!afitamy!Jim.Locker
        (513)-252-7681           |ARPA: Jim.Locker@afitamy.fidonet.org
------------------------------------------------------------------------