JGNAT and JRE

JGNAT and JRE

[fabio de francesco]

Hello,

Some time ago I wrote:

>
> I am going to write a few pages article about Ada for an Italian
Linux
> and Open Source devoted printed magazine. It'll be just an
introduction article
> to the language main features and great capabilities, it won't be a
> very technical detailed one.
>
> [cut some more lines]
>

Now a first draft has been approved by the editor who however asked me
to provide more informations on Ada executing on Java Virtual Machine.
That was because I cited JGNAT in the article as a bytecode compiler
for JVM. He thinks that the capability to compile once and execute
everywhere there is run-time support could "convert" many software
developers (especially Java ones) to Ada. All things considered I must
agree with him, even when I'have already structured the article with
respect to the language basic features.

The problem is that I didn't know much more than just the existence of
this capability and I'd never used it. Anyway I have been able to
downloaded JRE and JGNAT, then compiling and executing some sample
code. My intention was to understand something more about these tools
in order to be able to provide better informations and some practical
examples of this subject in the article.

JGNAT seemed working fine with some simple programs until I compiled a
new one that uses tasks without errors. But when I tried to execute it
I got "Exception in thread "main"
java.lang.IllegalMonitorStateException". What is this exception?

Above all, what really matters is whether or not I can write (1) JGNAT
is reliable for production use, (2) if it is actively maintained and
updated, (3) if there exist any kind of practical restrictions in using
it. Any answers and comments on either these issues or everything
related would be very appreciated and useful in writing a good Ada
introduction (I imagine it may matter many of you).

Another related subject is A#, intended as another tool to write once
and execute everything run-time support is provided. Some time ago I
looked at M.Carlisle site where I knew it is actively maintained, but
my interest run out when it seemed working only on MS-Win. Anyway
someone recently told me A# should also execute in Linux with Mono. Is
it true? Can someone provide much precise information about this issue?

What above all interests is (1) whether or not it works with Mono, (2)
if someone willing to use it must learn a different syntax from plain
old Ada95, (3) if there are any issues restricting a "normal" serious
production.

Thanks in advance to any of you who will reply.

Ciao,

fabio de francesco

Re: JGNAT and JRE

[Ed Falis]

Fabio,

On Thu, 28 Apr 2005 18:24:50 -0400, fabio de francesco <fmdf@tiscali.it>  
wrote:

> JGNAT seemed working fine with some simple programs until I compiled a
> new one that uses tasks without errors. But when I tried to execute it
> I got "Exception in thread "main"
> java.lang.IllegalMonitorStateException". What is this exception?

JGNAT never made it past compatibility with JDK 1.3, due to lack of  
commercial interest.

> Above all, what really matters is whether or not I can write (1) JGNAT
> is reliable for production use, (2) if it is actively maintained and
> updated, (3) if there exist any kind of practical restrictions in using
> it. Any answers and comments on either these issues or everything
> related would be very appreciated and useful in writing a good Ada
> introduction (I imagine it may matter many of you).

We (AdaCore) dropped active support for this product a few years ago  
because there was no significant commercial interest in Ada running on a  
JVM (and we did give it a fair amount of time and marketing effort).  You  
could also look into AppletMagic from SoftTool or Aonix as an alternative.

JGNAT largely works fine with JDK 1.3 or less.  With 1.4, Sun added some  
additional byte code validation that causes the exception you're seeing.

Unfortunately, what interest we saw in JGNAT was largely confined to  
jokers who thought it would be a way for them to migrate their Ada code  
base to Java. :-P

On the other hand, it is free software, and open to maintenance by others.

- Ed

Re: JGNAT and JRE

[Jeffrey Carter]

fabio de francesco wrote:

> Now a first draft has been approved by the editor who however asked me
> to provide more informations on Ada executing on Java Virtual Machine.
> That was because I cited JGNAT in the article as a bytecode compiler
> for JVM. He thinks that the capability to compile once and execute
> everywhere there is run-time support could "convert" many software
> developers (especially Java ones) to Ada. All things considered I must
> agree with him, even when I'have already structured the article with
> respect to the language basic features.

What I've always found more impressive is the way a lot of well-designed 
Ada can be compiled and run on different platforms without change.

-- 
Jeff Carter
"Beyond 100,000 lines of code you
should probably be coding in Ada."
P. J. Plauger
26