Make C++ via Ada95/Gnat ?

Make C++ via Ada95/Gnat ?

[Reinert Korsnes]

Hi,

Is it possible/practical to produce C++ code by first making Ada95 code
and then from this produce C++ (via Gnat) ?  Just a bad idea ?

(I would prefer to make/test Ada95 code rather than C++)

reinert

-- 
http://home.chello.no/~rkorsnes

Re: Make C++ via Ada95/Gnat ?

[Larry Kilgallen]

In article <9lg0gs$9dv$1@snipp.uninett.no>, Reinert Korsnes <Reinert.Korsnes@ffi.no> writes:

> Is it possible/practical to produce C++ code by first making Ada95 code
> and then from this produce C++ (via Gnat) ?  Just a bad idea ?

It should not be possible via GNAT, since GNAT and G++ are two
different front-ends to the same back end.  The fact that GNAT
can emit Java byte code is an anomoly to support the Java byte
code engine as a "target".  GNAT generally emits either machine
code (in ASCII) or assembler code (I don't know which).

> (I would prefer to make/test Ada95 code rather than C++)

Averstar can sell you an Ada compiler that emits ANSI C, which
might be a subset of C++ (for the subset of ANSI C they use).
But since you wanted C++ you probably wanted it to use features
of C++ that are not in C.

Re: Make C++ via Ada95/Gnat ?

[sk]

> GNAT generally emits either machine code (in ASCII) 

Machine code in ASCII ???  ... how limiting :-)

> or assembler code (I don't know which).

gnatmake -S <filename> 

generates an assembly listing (-S goes to gcc rather
than gnatmake).

Re: Make C++ via Ada95/Gnat ?

[Larry Kilgallen]

In article <mailman.997958311.11304.comp.lang.ada@ada.eu.org>, sk <sknipe@ktc.com> writes:
>> GNAT generally emits either machine code (in ASCII) 
> 
> Machine code in ASCII ???  ... how limiting :-)

Not to computers, just to humans.

	06435834
	96754834
	23706421
	75325326

The only purpose of a symbolic representation is to coddle humans :-)

Re: Make C++ via Ada95/Gnat ?

[Marc A. Criley]

Larry Kilgallen wrote:
> 
> In article <9lg0gs$9dv$1@snipp.uninett.no>, Reinert Korsnes <Reinert.Korsnes@ffi.no> writes:
> 
> > Is it possible/practical to produce C++ code by first making Ada95 code
> > and then from this produce C++ (via Gnat) ?  Just a bad idea ?
> 
> It should not be possible via GNAT, since GNAT and G++ are two
> different front-ends to the same back end.  The fact that GNAT
> can emit Java byte code is an anomoly to support the Java byte
> code engine as a "target".  GNAT generally emits either machine
> code (in ASCII) or assembler code (I don't know which).

_JGNAT_, not GNAT, emits Java byte code.

Marc A. Criley
Senior Staff Engineer
Quadrus Corporation
www.quadruscorp.com

Re: Make C++ via Ada95/Gnat ?

[Robert Dewar]

Reinert Korsnes <Reinert.Korsnes@ffi.no> wrote in message news:<9lg0gs$9dv$1@snipp.uninett.no>...
> Hi,
> 
> Is it possible/practical to produce C++ code by first making Ada95 code
> and then from this produce C++ (via Gnat) ?  Just a bad idea ?
> 
> (I would prefer to make/test Ada95 code rather than C++)
> 
> reinert




No, it is not possible to translate Ada to C++ using GNAT, since
in no sense does GNAT translate Ada to either C or C++ during
compilation.

There is no technology I am aware of that can translate Ada code to
high level C++ code, nor is it easy to conceive of anything that
would achieve a 100% translation that resulted in useful and
readable C++, so the conception here is basically flawed. If you
want C++ you must write in C++ :-(