Re: Ada & JGNAT

[Laurent GUERBY <laurent at guerby dot net>]

Sorry I wasn't providing more information, here are two
links of interest:

http://www.cs.aston.ac.uk/~barnesa/jgnat_ug.html
http://www.usafa.af.mil/dfcs/bios/mcc_html/a_sharp.html

JGNAT indeed emits JVM bytecodes, taking the place
of the "gnat to gnu" phase. I just know the project
through early design while I was working at ACT,
I don't know of the specifics of the latest developments.

Code has been written by ACT employees just like GNAT.

A# takes JGNAT output and run it through a Microsoft
JVM bytecode to the MS IL converter.

Both JGNAT and A# come with binding generators
that take native interface and export Ada
specification.

A# was written by an employee of the US government
as part of his regular work (teaching at USAF) so I believe the code is
in the public domain, I don't know if the FSF had to handle such
cases before, but that comes after JGNAT.

I agree about your comments about the political issues
involved, but I believe this list is not the place to discuss this
particular topic.

Laurent

On Wed, 2004-07-07 at 01:41, Jim Wilson wrote:
> Laurent GUERBY wrote:
> > Is it possible to add JGNAT sources to the GCC CVS repository?
> > Even if they don't compile, that would be a possible good starting
> > point for reviving development (for A# too).
> 
> It would be nice to have some info about JGNAT.  You included a link to 
> a tarball in a later message, but there don't seem to be any doc files 
> in there.  I noticed that the java files have a comment that says this 
> is Ada for the JVM.  Is this an Ada front end that produces Java byte codes?
> 
> I see this has ACT copyrights.  This would have to be formally 
> contributed to the FSF.  You would have to make sure you can change the 
> copyright, and that all contributors have valid gcc assignments.
> 
> Will this be a separate front end or a modification to the existing 
> front end?
> 
> rms has expressed concerns to the GCC SC about java byte code support. 
> He is concerned that it could be used to subvert the GPL.  A proprietary 
> compiler could emit java byte codes for instance, and then feed it into 
> gcc to avoid linking with gcc code.  The GCC SC convinced rms that java 
> byte codes were not general enough to be used in this fashion.  An Ada 
> front end that can emit java byte codes weakens that argument.
> 
> And .NET support is an entirely different matter.  A# is presumarly an 
> Ada front end that emits .NET IL (I forget what it is called), which is 
> a much more general IL than Java byte codes.  We won't be able to 
> convince rms that the .NET IL is safe.  This is an FSF policy that needs 
> to change anyways, but meanwhile, this is something to be wary of.