Brian White writes:
>Am I missing something, or shouldn't be fairly easy to port HP's C++ compiler
>for Unix to MPE, now that the hierarchical file system exists under MPE/iX
>5.0??
 
The problem is not one of "how difficult" but "how many would be sold."
Having recently "de-ported" two libraries (a general-purpose lexer and
a PCL parser) from C++ to C in order to use them on the HP3000, I would
LOVE to have C++ on MPE iX. But for some silly reason, HP engineering
insists on doing things the right way, so the port is harder than
it looks.
 
First, the ported compiler has to have the "look and feel" of any
other HP3000 compiler. That means its defaults may need to be different
for the two platforms; it needs to be trained to read MPE-style files
(though that problem partially goes away under 5.0) and the MPE
directory structure. It needs to support the same #pragma intrinsic
mechanism that the C compiler supports, which means changing the
preprocessor and mucking around with the symbol table logic. It needs
to support MPE iX procedure calling conventions, including those
with variable-length argument lists, which means mucking around with
the code generator. It may be necessary to modify the
linker to deal with mangled function names, depending on how the
mangling is done. Finally, the iostream library may need modification to
handle MPEisms in the file system. Oh, and xdb and the system debugger
would need to be modified to de-mangle function names (else they'd be
pretty useless -- I've tried.)
 
A "quick-and-dirty" port, not including any of these items, would probably
be easy. It would also be difficult to sell and nearly impossible to
support. Given the engineering effort necessary to do the job right,
it's unlikely to happen.
 
<sniff> Hurry up, Mark!
 
-- Bruce