ada < - > java bindings wanted

ada < - > java bindings wanted

[Constantine A. Sismanidis]

Good afternoon all,

My apologies in advance for posting this to so many newsgroups, and for
posting what may be blindingly obvious to some.

We are in the process of porting a very large legacy system from UNIX to
Windows NT. Most of the code in this system is written in ada (with a
smattering of C and FORTRAN). The user interface uses the X window System.

We need to be able to call java from an ada main routine (we plan to build
the user interface in java), and to have java call the legacy ada code. We
will be using Rational's ada compiler. We *can* get this to work by having
ada call C which will then call java (using the Java Native Interface); or
java calling C (using the Java Native Interface), which will then call ada.
However, we would very much prefer to call java directly from ada, and ada
from java, in order to avoid the extra layers of code that would be
required.

Have others been faced with this sort of situation (ada calling java calling
ada), and if so, how did you solve the problem?
--
Constantine A. Sismanidis                       Senior Software Engineer
The Boeing Company                                         (703)219-3812
11242 Waples Mill Road, Suite 300                     Fax: (703)219-3813
Fairfax, VA 22030-6075       e-Mail: constantine.a.sismanidis@boeing.com

Re: ada < - > java bindings wanted

[Tucker Taft]

Constantine A. Sismanidis wrote:

>... > We need to be able to call java from an ada main routine (we plan to build
> the user interface in java), and to have java call the legacy ada code. We
> will be using Rational's ada compiler. We *can* get this to work by having
> ada call C which will then call java (using the Java Native Interface); or
> java calling C (using the Java Native Interface), which will then call ada.
> However, we would very much prefer to call java directly from ada, and ada
> from java, in order to avoid the extra layers of code that would be
> required.

I suggest you check with http://www.ainslie-software.com.  They have
a very nice tool called AdaJNI which I believe will give you
the ability to call Ada from Java, and perhaps vice-versa.

Note also that you can generally make Ada code "look" like C code
by applying a pragma Convention(C, xxx) to the Ada code, which 
should simplify interfacing.
> 
> Have others been faced with this sort of situation (ada calling java calling
> ada), and if so, how did you solve the problem?

The other choice is to recompile your entire Ada source into
Java byte codes.  However, that will definitely reduce performance
of the Ada code to some degree.  See http://www.appletmagic.com for
information on that.

> --
> Constantine A. Sismanidis                       Senior Software Engineer
> The Boeing Company                                         (703)219-3812
> 11242 Waples Mill Road, Suite 300                     Fax: (703)219-3813
> Fairfax, VA 22030-6075       e-Mail: constantine.a.sismanidis@boeing.com

-- 
-Tucker Taft   stt@averstar.com   http://www.averstar.com/~stt/
Technical Director, Distributed IT Solutions  (www.averstar.com/tools)
AverStar (formerly Intermetrics, Inc.)   Burlington, MA  USA

Re: ada < - > java bindings wanted

[Tucker Taft]

Tucker Taft wrote:

> I suggest you check with http://www.ainslie-software.com.  They have
> a very nice tool called AdaJNI which I believe will give you
> the ability to call Ada from Java, and perhaps vice-versa.

Oops, I had it backwards.  AdaJNI allows you to call Java from
Ada95.  "Normal" JNI (Java Native Interface) allows you to call
compiled code from Java.  All you need to do is make sure to
pragma Export(C, Your_Ada_Routine, "external_name") which will ensure 
that Your_Ada_Routine is visible with the given external_name
and follows normal "C" calling conventions.  You may have to convert
your Ada code into a shared-library/DLL, and load the DLL from the 
Java code.  The documentation on JNI and AdaJNI should cover this.
> > --
> > Constantine A. Sismanidis                       Senior Software Engineer
> > The Boeing Company                                         (703)219-3812
> > 11242 Waples Mill Road, Suite 300                     Fax: (703)219-3813
> > Fairfax, VA 22030-6075       e-Mail: constantine.a.sismanidis@boeing.com

-- 
-Tucker Taft   stt@averstar.com   http://www.averstar.com/~stt/
Technical Director, Distributed IT Solutions  (www.averstar.com/tools)
AverStar (formerly Intermetrics, Inc.)   Burlington, MA  USA

Re: ada < - > java bindings wanted

[Jim White]

Constantine A. Sismanidis wrote in message ...
>Have others been faced with this sort of situation (ada calling java
calling
>ada), and if so, how did you solve the problem?


There is Ada for the Java VM:

<http://www.appletmagic.com/>

<http://www.gnat.com/>

I'm not sure whether GNAT has their Ada95 for the JVM available yet or
not, but they do advertise easy interface between their Ada95 compiler
and Java.

jim

Re: ada < - > java bindings wanted

[David Botton]

Since you are planning on running on Win32, an easy solution is to put the
legacy Ada code in to an Ada coded COM object (see
http://www.adapower.com/com and contact ACT for more information on getting
support for COM with Ada) and then run JActiveX on the COM object to produce
a java binding that works in Win32 with Microsoft's native Java support.

The advantage to this is that you can also distribute the Ada code on
multiple machines if you like.

It would also be possible to do the reverse and call your Java code from Ada
by turning your Java code in to COM objects (very easy to do) and calling it
from Ada.

David Botton

Constantine A. Sismanidis wrote in message ...
>We are in the process of porting a very large legacy system from UNIX to
>Windows NT. Most of the code in this system is written in ada (with a
>smattering of C and FORTRAN). The user interface uses the X window System.
>
>We need to be able to call java from an ada main routine (we plan to build
>the user interface in java), and to have java call the legacy ada code. We
>will be using Rational's ada compiler. We *can* get this to work by having
>ada call C which will then call java (using the Java Native Interface); or
>java calling C (using the Java Native Interface), which will then call ada.
>However, we would very much prefer to call java directly from ada, and ada
>from java, in order to avoid the extra layers of code that would be
>required.

Re: ada < - > java bindings wanted

[Robert Dewar]

In article <FGD8I1.Joo@news.boeing.com>,
  "Constantine A. Sismanidis"
<constantine.a.sismanidis@boeing.com> wrote:
> Good afternoon all,
>
> My apologies in advance for posting this to so many
newsgroups, and for
> posting what may be blindingly obvious to some.
>
> We are in the process of porting a very large legacy system
from UNIX to
> Windows NT. Most of the code in this system is written in ada
(with a
> smattering of C and FORTRAN). The user interface uses the X
window System.


I should think you should contact your vendor for solutions
here, and indeed I would suggest that this is something you
should make a requirement in your compiler source selection.
Much better to select a compiler that does what you want, rather
than select a compiler without checking, and then wonder how
to make it work!


Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.

ada < - > java bindings wanted

[Constantine A. Sismanidis]

Good afternoon all,

My apologies in advance for posting this to so many newsgroups, and for
posting what may be blindingly obvious to some.

We are in the process of porting a very large legacy system from UNIX to
Windows NT. Most of the code in this system is written in ada (with a
smattering of C and FORTRAN). The user interface uses the X window System.

We need to be able to call java from an ada main routine (we plan to build
the user interface in java), and to have java call the legacy ada code. We
will be using Rational's ada compiler. We *can* get this to work by having
ada call C which will then call java (using the Java Native Interface); or
java calling C (using the Java Native Interface), which will then call ada.
However, we would very much prefer to call java directly from ada, and ada
from java, in order to avoid the extra layers of code that would be
required.

Have others been faced with this sort of situation (ada calling java calling
ada), and if so, how did you solve the problem?
--
Constantine A. Sismanidis                       Senior Software Engineer
The Boeing Company                                         (703)219-3812
11242 Waples Mill Road, Suite 300                     Fax: (703)219-3813
Fairfax, VA 22030-6075       e-Mail: constantine.a.sismanidis@boeing.com

Re: ada < - > java bindings wanted

[Stanley R. Allen]

"Constantine A. Sismanidis" wrote:
> 
> We need to be able to call java from an ada main routine (we plan to build
> the user interface in java), and to have java call the legacy ada code. We
> will be using Rational's ada compiler. We *can* get this to work by having
> ada call C which will then call java (using the Java Native Interface); or
> java calling C (using the Java Native Interface), which will then call ada.
> However, we would very much prefer to call java directly from ada, and ada
> from java, in order to avoid the extra layers of code that would be
> required.
> 

See http://www.ainslie-software.com for the product AdaJNI.

-- 
Stanley Allen
mailto:s_allen@hso.link.com

Re: ada < - > java bindings wanted

[Ted Dennison]

In article <FGD9E8.M5w@news.boeing.com>,
  "Constantine A. Sismanidis" <constantine.a.sismanidis@boeing.com>
wrote:
> will be using Rational's ada compiler. We *can* get this to work by
> having ada call C which will then call java (using the Java Native
> Interface); or java calling C (using the Java Native Interface), which
> will then call ada. However, we would very much prefer to call java
> directly from ada, and ada from java, in order to avoid the extra
> layers of code that would be required.

I'm not familiar with that particular compiler (unless it is the VADS
one, which I doubt highly). But if its an Ada 95 compiler, you should be
able to apply a "pragma Export (C, " to all the Ada routines you wish to
call from Java. That will make them look like C routines, which should
make your JNI happy.

There are lots of other little issues in interfacing to other languages
from Ada which are discussed in Annex B of the Ada LRM (a section
obscurely titled: "Interface to Other Languages") Its available online
at http://www.adahome.com/rm95/rm9x-B.html .

There are additional issues if your main routine is not Ada, but I'm
guessing that won't apply here.

--
T.E.D.


Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.

ada < - > java bindings wanted

[Geoff Bull]

> 
> We need to be able to call java from an ada main routine
> (we plan to build the user interface in java), and to have
> java call the legacy ada code. We will be using Rational's
> ada compiler. We *can* get this to work by having ada call C
> which will then call java (using the Java Native Interface);
> or java calling C (using the Java Native Interface), which will
> then call ada. However, we would very much prefer to call java
> directly from ada, and ada from java, in order to avoid the
> extra layers of code that would be required.
>                      
> Have others been faced with this sort of situation
> (ada calling java calling ada), and if so, how did
> you solve the problem?

A C wrapper around JNI isn't necessary because I have written
an Ada wrapper around JNI, called Cafe1815.

It is open source with a license intended to be less
restrictive than GPL, but if you make improvements I'd
like to incorporate them for the benefit of all.

You can find Cafe1815 at:
http://www.acenet.com.au/~gbull/

Due to work and study pressures, I can only make minor
improvements for free. Ofcourse, if you have money to spend ...

Cheers
Geoff

Re: ada < - > java bindings wanted

[Robert Dewar]

In article <37BD0169.B49F73A5@acenet.com.au>,
  Geoff Bull <gbull@acenet.com.au> wrote:

> It is open source with a license intended to be less
> restrictive than GPL,

I would really encourage people not to come up with
idiosyncratic open sources licenses of their own. There
are a wide variety of standard licenses to choose from.
The trouble is that if you use a specialized license then

a) there is no real assurance that it is drawn up properly,
since it is not clear that attorneys have reviewed it etc.

b) open source licenses can be tricky. At this stage we have
a pretty good understanding of the GPL, the BSD license (to
choose one example of a much less restrictive license which
encourages/permits incorporation into proprietary code), and
many other licenses.

c) the effort to read and understand a new license can act as
a significant barrier to adoption of the associated code. It
means, if you are being careful, that you have to spend a
considerable amount of time and money, getting your own
attorneys to read the license, and understand it, a process
that often takes going backwards and forwards.

It is quite reasonable for an author to want to use a license
which is less restricitive than the GPL, in particular if you
want to allow people to incorporate your work in non-open
source software. But there are many well understood, well-known
licenses that do this, and indeed, you can simply place your
code in the public domain, to remove all restrictions if you
like (that is also a well understood status).

> but if you make improvements I'd
> like to incorporate them for the benefit of all.

Normally, the use of a license that specifically goes out of
its way to allow improvements to be incorporated into non-open
source non-free code is interpreted as not only allowing, but
essentially encouraging such use where appropriate (this is
for example the effect of the modified GPL used for GNAT
sources, we quite understand and expect that people will use
(possibly modified) versions of the GNAT runtime in their
distributed proprietary/classified/secret etc code.

If you really would like to be able to incorporate any
improvements that are made for the benefit of all, why
not use a standard license that allows this.

It may be that your license is indeed a rather standard one,
modified little or not at all, but there was no easy way of
telling. Reading your license, it is quite complex, and I
certainly would not venture an interpretation of some of
its features without an attorney to advise me!

Finally, are you sure your software is indeed open source?
Has your license been approved as a legitimate open source
license? Remember that the phrase open source is protected,
and you should only use this phrase if you are sure that
your license has been approved (another reason for using
one of the standard licensing vehicles).



Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.

Re: ada < - > java bindings wanted

[Ted Dennison]

In article <7pjm8u$on1$1@nnrp1.deja.com>,
  Robert Dewar <robert_dewar@my-deja.com> wrote:
> In article <37BD0169.B49F73A5@acenet.com.au>,

> Finally, are you sure your software is indeed open source?
> Has your license been approved as a legitimate open source
> license? Remember that the phrase open source is protected,
> and you should only use this phrase if you are sure that
> your license has been approved (another reason for using
> one of the standard licensing vehicles).

According to http://www.opensource.org/certification-mark.html, that's
not quite true. "OpenSource" could not be trademarked, so the
certification mark is "OSI Certified".

Does the GMG enjoy this status? I wrote ESR on this subject once, but
didn't get a response. But that was before the "OSI Certified" mark was
set up, and of course he's a very busy individual.

I don't think meerly reading the OpenSource definition (
http://www.opensource.org/osd.html ) is enough, as to my mind clause 9
disqualifies the GPL, yet it is listed as being approved (
http://www.opensource.org/certification-mark.html#approved-licenses ).

Given the number of folks using the GMG, it would be nice to have it on
the approved list. I'd be happy to put it up for approval, but if it is
already pending or approved I don't want to waste anyone's time. Also
given that changes may be requested and it is an ACT-created license,
the folks at ACT may prefer to shepherd the submission.

--
T.E.D.


Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.

Re: ada < - > java bindings wanted

[Florian Weimer]

Ted Dennison <dennison@telepath.com> writes:

> According to http://www.opensource.org/certification-mark.html, that's
> not quite true. "OpenSource" could not be trademarked, so the
> certification mark is "OSI Certified".
> 
> Does the GMG enjoy this status?

If `GMG' means GPL plus exceptions for compiling and linking with
library code, the answer is probably yes: Depending on the use of the
library code, there aren't any restrictions at all -- or the GPL applies,
which is OSI certified.

> I don't think meerly reading the OpenSource definition (
> http://www.opensource.org/osd.html ) is enough, as to my mind clause 9
> disqualifies the GPL, yet it is listed as being approved (
> http://www.opensource.org/certification-mark.html#approved-licenses ).

You can ship GPL software along with non-GPL software.  IMHO that's
what clause 9 is all about.

Re: ada < - > java bindings wanted

[Robert Dewar]

In article <7pjovd$qo8$1@nnrp1.deja.com>,
  Ted Dennison <dennison@telepath.com> wrote:

> I don't think meerly reading the OpenSource definition (
> http://www.opensource.org/osd.html ) is enough, as to my mind
> clause 9
> disqualifies the GPL, yet it is listed as being approved (

Well I guess your mind doesn't read quite carefully enough in
this case :-)

Here is clause 9:

9. License Must Not Contaminate Other Software.

      The license must not place restrictions on other software
      that is distributed along with the licensed software. For
      example, the license must not insist that all other
      programs distributed on the same medium must be
      open-source software.

The other software here is about other programs that have
nothing to do with the licensed software. For example, if the
GPL said that no GPL'ed program may be distributed on the
same CD ROM as a non-free program, then it would indeed run
afoul of clause 9.

If you use a GPL'ed library in an application, then the
application *is* the licensed software, since the GPL creates
a license for this software. That's perfectly acceptable to the
notions of open source, and of course the GPL is historically
the quintessential open souce licenses. In general all free
software licenses meet the requirements for OSI certification,
the contrary is definitely not true (there is OSI certified
software that considerably restricts your freedomes, and so
is not considered free software). As with any software licenses
for copyrighted materials, you need to read the license
carefully to ensure that your use is consistent with your
license.

It is definitely NOT the case that open source means you can
do anything you like with the software. Indeed some open source
licenses give ownership of any modifications to the original
copyright holder, and require that you send such modifications
to them. Such conditions may or may not be acceptable to you,
they certainly disqualify software as free.

The fundamental issue with free software is that you personally
can do anything you like with the software for your own use
without any kind of restrictions at all, and that you can pass
on either the original or modified software to anyone. In the
case of the GPL, this passing on must preserve the freedom for
the recipient.

By the way, if you have trouble understanding any of the
conditions in the quoted document, be sure to read the
rationale sections. For this particular clause, the example
should make things crystal clear, but where that is not the
case, the rationale will usually clarify matters.

Robert Dewar
Ada Core Technologies


Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.

Re: ada < - > java bindings wanted

[Simon Wright]

Robert Dewar <robert_dewar@my-deja.com> writes:

> for example the effect of the modified GPL used for GNAT
> sources, we quite understand and expect that people will use
> (possibly modified) versions of the GNAT runtime in their
> distributed proprietary/classified/secret etc code.

That's an interesting observation.

We had somehow got the idea that the added permission to instantiate
and link with didn't permit modification. We never had any reason to
test that, since we didn't have any such problems.

I think what you're saying is that we can modify the GNAT-modified-GPL
files without removing the extra permissions? (if so, obvious really
now you point it out ..)

-- 
Simon Wright                        Work Email: simon.j.wright@gecm.com
Alenia Marconi Systems                         Voice: +44(0)1705-701778
Integrated Systems Division                      FAX: +44(0)1705-701800

Re: ada < - > java bindings wanted

[Robert Dewar]

In article <x7vu2ptys46.fsf@pogner.moho>,
  Simon Wright <simon@pogner.demon.co.uk> wrote:
> We had somehow got the idea that the added permission to
instantiate
> and link with didn't permit modification. We never had any
reason to
> test that, since we didn't have any such problems.
>
> I think what you're saying is that we can modify the
GNAT-modified-GPL
> files without removing the extra permissions? (if so, obvious
really
> now you point it out ..)


The GPL gives you the right to modify! The extra permissions
*ADD* to your rights, they do not subtract :-) Yes, of course
you can modify and distribute. Your distributed modified
version must be at least covered by the GPL, and if you choose
(your choice, you are the author of the changes) with the
additional permissions of the modified GPL.

Robert Dewar
Ada Core Technologies


Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.

Re: ada < - > java bindings wanted

[Geoff Bull]

Robert Dewar wrote:
> 
> In article <37BD0169.B49F73A5@acenet.com.au>,
>   Geoff Bull <gbull@acenet.com.au> wrote:
> 
> > It is open source with a license intended to be less
> > restrictive than GPL,
> 
> I would really encourage people not to come up with
> idiosyncratic open sources licenses of their own.

My company owns the code, we can put whatever license
we like on it!

If you don't like the code, or are simply not interested
in it, I can't imagine you care too much about the license.
If you like the code and the license is a problem for
you, we can be probably convinced to use a different license.

You might find "idiosyncratic" licenses a problem.
However, for us, code that is straight GPL or LGPL
is a problem.

Re: ada < - > java bindings wanted

[Robert Dewar]

In article <37BEB3BB.78C0B230@acenet.com.au>,
  Geoff Bull <gbull@acenet.com.au> wrote:
> You might find "idiosyncratic" licenses a problem.
> However, for us, code that is straight GPL or LGPL
> is a problem.

As I said in my message (please reread!) it is perfectly
understandable that you may not want to use the GPL for
some of your software (indeed we do not use the GPL for
the GNAT library!)

But my point was that there are several well understood
OSI-certified and/or free software compatible licenses
which *are* well known, and other things being equal it
is probably a good idea to use one of these for two reasons:

1. It saves the attorney fees you otherwise need to pay to
craft and carefully checkout the license you propose to use
(no one should use a license which does not have this kind
of careful evaluation, since otherwise you can find surprises).

2. Your potential users will know the license and its
implications without having to pay *their* attorneys to
figure it out.

There are three typical reasons for wanting to explore other
licenses than the GPL in this setting:

1. You want to give more freedom to people to include the code
in their proprietary programs. The modified GPL used in the
gcc library, and a similar Ada-modified one used in the GNAT
library are examples of this. These are still of course
free software compatible, since they give MORE rights than
the GPL.

2. You want to give complete freedom to people to do anything
they like with the code (perhaps retaining the copyright notice
as the only restriction). An example of a license in this
direction is the BSD licenses, or you could even go all the
way and use the public domain route.

3. You want to restrict the GPL to ensure that if people make
modifications, you still own these modifications, and perhaps
that people are required to distribute them to you. There are
OSI-certified licenses with this kind of general character
(e.g. the Apple license), but of course such licenses are not
considered free software compatible.

My real point here, without any kind of suggestion of
ideology or encouragement to go in any of the above directions,
is simply that whatever direction you go in, it pays to use
a standard existing license, if you can find one that matches
your needs, and it is worth the effort to do so (and almost
certainly cheaper than paying attorneys to make a completely
new license).

Of course it goes without saying that you can use any
license you want, all these licenses are licenses for
copyrighted sofwtare (with the exception of the PD route),
and you most certainly can do what you want, you can even
use a home brewed license that has not been reviewed by
attorneys (though I would recommend against this, simply
because of legal risk factors).

I am not at all trying to criticize here, but simply to make
a constructive suggestion that will make your software more
accessible to others. Usually the idea of open sourcing of
any kind is to increase this accessibility!

Robert Dewar


Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.

Re: ada < - > java bindings wanted

[Robert Dewar]

In article <37BEB3BB.78C0B230@acenet.com.au>,
  Geoff Bull <gbull@acenet.com.au> wrote:
> If you don't like the code, or are simply not interested
> in it, I can't imagine you care too much about the license.
> If you like the code and the license is a problem for
> you, we can be probably convinced to use a different license.
>
> You might find "idiosyncratic" licenses a problem.
> However, for us, code that is straight GPL or LGPL
> is a problem.


Just to be clear, we are interested in this code, and at least
one of our GNAT Professional customers is using it. The license
is only a problem because it is non-standard. We certainly do
not insist on any code we use being GPL'ed. More permissive
licensing such as public domain, or BSD is quite fine. We
insist on open source status, and we much prefer licenses
that are free software compatible, but we can certainly
consider more restrictive open source licenses in our
environment. Indeed we tool partner with a number of
companies offering useful proprietary tools that work
well with GNAT.

Our primary interest here is in helping our customers, but
we have to be a little careful in passing on code if we
don't understand the license well.

One of the guarantees that we make to customers right now is
that they can use GNAT libraries without any concern about
acquiring problematic licensing conditions that would apply
to their generated programs. This is of course achieved in
our case by use of the GNAT modified GPL (GMGPL). Any license
that is at least this non-restrictive is fine with us.

In fact a library released under the pure GPL (such as the
public versions of Cygwin) is NOT acceptable for inclusion
in GNAT libraries, precisely because it does cause
troublesome licensing issues for generated programs.

You have to be very careful with license conditions, and as
I have said many times before, this is equally true whether
you are talking Free Software, OSI-certified software, or
proprietary software.

More than once I have run into customers who were accidentally
violating the Cygnus copyright by including the pure GPL'ed
version of Cygwin in their programs, and then not adhering to
the consequent GPL'ed status of the resulting programs.

Robert Dewar
Ada Core Technologies


Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.