Re: Modern copy of GNAT for OpenVMS?

["Marin David Condic" <mcondic.auntie.spam@acm.org>]

Jeffrey Carter <jrcarter@acm.org> wrote in message
news:3DEF8BDB.7050009@acm.org...
>
> Most VAX/DEC/Compaq Ada code can be recompiled unchanged with GNAT,
> which is available for Linux or Windows on PCs, as well as for VMS on
> Alphas.
>
I have noted some slight differences in the past, but usually, portability
between DEC Ada and Gnat Ada was pretty good. The *real* problem is how
seriously hooked into the OS and other platform specific features is the
application at hand. The OP didn't say anything about that, and that's where
all the serious issues will lie.

I've ported code that did nothing beyond a command line interface & a bunch
of computations and had it compile/execute right out of the box. When I've
thrown in access to files of a non-text nature, I've had problems with byte
sex and other binary representation issues. (Yes, you can design for this,
but often portability is not an issue when the program is first built.) Once
the application starts getting into OS features like a GUI interface,
network services, etc., you can forget portability. At that point the amount
of trouble one will have depends on how well isolated the accesses to the OS
were built.

Aside: Wouldn't it be nice to have a *standard* way of getting at various
"normal" OS services like networking and a GUI? Vast amounts of software
these days take advantage of that sort of thing, but then must become bound
to the OS that provides it. Portability starts becoming a pipe dream.

MDC
--
======================================================================
Marin David Condic
I work for: http://www.belcan.com/
My project is: http://www.jast.mil/

Send Replies To: m c o n d i c @ a c m . o r g

    "I'd trade it all for just a little more"
        --  Charles Montgomery Burns, [4F10]
======================================================================