David, I've been out of town for a couple of weeks, and then
extraordinarily busy for several more, so I've made little progress on my
Named Pipes problem.  As I sent you last time, the Rexx procedure acts
exactly like Keith's program.  Here is his response to your query:

 >> Keith,
 >>         It's been a while since I heard from you, but that's OK, I've
 >> been out-and-about myself.  I've had some feedback from some fellows in
 >> the OS2PROGRAMMING echo, and they tend to want to look at problems with
 >> the code.  I think it's something at this end.  What I would like is
 >> permission to distribute some of our test code to them for testing
 >> purposes only, extracting a promise not to use the software for their
 >> own purposes.  I'd like to send a copy of the executable to three
 >> fellows who have been working with me.  Would this be OK with you?
 >
 >Feel free to pass it around to whomever may help you out.
 >
 >His code:
 >
 >> #define Buffer_size 4096UL
 >>
 >>    APIRET rc;
 >>    HPIPE Pipe;
 >>    static const unsigned short Pipe_instances = 1;
 >>
 >>    /* Create the named pipe. */
 >>    rc = DosCreateNPipe("\\PIPE\\DWN_pipe", &Pipe,
 >>         NP_INHERIT|NP_ACCESS_OUTBOUND,
 >>         NP_WAIT|NP_TYPE_BYTE|NP_READMODE_BYTE|Pipe_instances,
 >>         Buffer_size, 0UL, 0UL);
 >
 >My code:
 >
 >   //
 >   // create the status pipe
 >   //
 >   rc = DosCreateNPipe ("\\PIPE\\DMPSTAT", &hStatusPipe,
 >      NP_ACCESS_OUTBOUND | NP_NOWRITEBEHIND,
 >      NP_NOWAIT | NP_TYPE_BYTE | NP_UNLIMITED_INSTANCES,
 >      0, 255, 0);
 >
 >Differences:
 >
 >NP_INHERIT and NP_READMODE_BYTE are default values, so the only real
 >differences are that I specify NP_NOWAIT and NP_NOWRITEBEHIND in order to
 >push the data through as quickly as possible.
 >
 >The other difference of the last three function parameters is his error I
 >believe.
 >
 >I'd check this with him and see if this could be the problem.

        
        

        What do you think?

        I've also gotten permission from him to distribute the beta code
for testing, if necessary.  In addition, he is going to pretty much abandon
the project, and is willing to give me the source code if it would help. 
It's written in CSet/2, but I have Borland C++/OS2.  I'm not sure how
different the two compilers are, but in any case I suspect the problem
isn't with his code.  If it is, I may need a bit of assistance in porting
the code, and/or possibly compiling it on someone else's machine.  Do you
or anyone else on the echo have the resources you might be able to lend me
should the need arise?  I of course expect to pay for any time spent
actually doing any programming, but I expect in the unlilkely event
anything is needed, it will only be advice and/or use of a compiler to
compile any changes I make in the source code.

                                                        Les

e-mail: lrhorer{at}fibrcom.com

--- EZPoint V2.2