Thanks for the advice; consistent with your recommendations and those of
Mr. Cleis, I will re-implement this using write-bytes rather attempting to
display a string.
My working VB code required the carriage-return linefeed character in order
to work properly ( .WriteLine("V00" & vbCrLf) ). I'll attempt various line
terminations and, failing all else, I'll check the pins with a scopemeter
to see the characters transmitted and go from there.
Thanks all,
On Fri, Dec 30, 2011 at 6:24 PM, Neil Van Dyke <n...@neilvandyke.org> wrote:
> There are a number of things that could be going wrong, due to RS232 and
> terminal device subtleties, and the OS abstractions.
>
> Assuming that your bit rate, data bits, parity, and stop bit are
> compatible, and you're using the right cable (null-modem or straight
> through, and enough of the pins wired through rather than unconnected like
> in some cables), and you're not confused by something like flow control or
> full/half duplex (and I'm assuming it's ASCII)...
>
> The very first thing I'd do from the Racket side is to send exact bytes
> rather than strings, just because bytes are the right way to do it.
>
> Then I'd be sure that I was sending whatever newline convention the device
> requires: after the #"V00" bytes, does it require a CR byte and/or an LF
> byte?
>
> The very next thing I'd look at on the Racket side is that I'd quickly
> peek at the implementation of "open-input-output-file". (From your
> description, it sounds like it is correctly not buffering the output, and
> that EOF doesn't look like it can be right if it's doing a blocking read
> with the semantics I'd expect, but I'd just make sure I knew the
> implementation, since the documentation makes this look easier than I'd
> expect it to be.)
>
> If it's not working by this point, I'd pick one of the open source
> libraries that does work with it, and see what they're doing differently.
> Maybe try to find one that's not using Win32/MFC/etc. calls, but works
> using standard C and POSIX calls. RS232 is a pain in the butt, so I think
> this is a good time to first do cargo cult engineering -- go through the
> magical incantations that seemed to work once before -- and then after it
> seems to work you can figure out the *why* so that you can have as robust
> of a solution as you need.
>
> If all else fails, you can use the Racket FFI to make example Win calls
> you find in MSDN or on some Web page, but good to try the elegant
> "open-input-output-file" first.
>
> --
> http://www.neilvandyke.org/
>
>
____________________
Racket Users list:
http://lists.racket-lang.org/users
