On the 03-19-95, Jerry Mcbride was chatting with Ian Timms:

JM> IT> As to readwrite_denynone, if the calling program doesn't open 
JM> IT> the port in such a way that 'called' programs can use the 
JM> IT> handle then there is little point in providing the handle in 
JM> IT> the first place. OS/2 just spits the dummy and causes the 
JM> IT> child program to error if it attempts to write or read from the 
JM> IT> port when it is not allowed to do so.
JM>
JM>Perhaps this is what my bbs does, MAXIMUS. I have tried to use the passed 
JM>handle with rxasync (with out calling rxasyncopen) and it 
JM>failed to work. However, I can call rxasyncopen and tell it 
JM>what portname that the bbs is on and it'll all work just 
JM>fine. 
JM>
JM>The only thing is, I have to pass portname (com1, etc...) with the call.
JM>
JM>Ian, would it be possible to add in another function? One 
JM>that would take a hot porthandle passed from the bbs and 
JM>just start using it?

Hi Jerry, 

This functionality is already there, if you have a hot port handle
then just pass it into the RxAsyncRead and Write functions, provided
the correct permissions were set when the port was opened by MAXIMUS
you should have no trouble using it.

I hope you are not using 'START' in your invocation of the REXX script,
DETACH should be okay, but START sets up a completely new session.

If you still have trouble then please talk with Pete Fitzsimmons, and 
ask him why his 'hot' port handle went cold.

 Cheers, Ian.

 Internet:itimms{at}ariel.ucs.unimelb.edu.au   CIS:100236,1404   [Team OS/2]
___
 * MR/2 2.1 #141 * I've got a New Toy   oh  ee  oh   and it's OS/2  oooh!

--- Maximus/2 2.02