ACT announces first 100% validations of Ada 95 compilers

ACT announces first 100% validations of Ada 95 compilers

[Robert Dewar]

ADA CORE TECHNOLOGIES ANNOUNCES THE FIRST 100% ADA 95 VALIDATIONS

Ada Core Technologies is pleased to announce that we have received
formal notification from NIST that our recent Ada 95 validations
under ACVC 2.0.1 have been officially registered. These validations
apply to:

   GNAT 4.00 SPARCstation 5 under Solaris 2.5.1
   GNAT 4.00 SGI Power Indigo2 under IRIX 6.2
   GNAT 4.00 SGI O2 under IRIX 6.3

These are the first validations from any vendor that cover 100% of
applicable tests in not only the core language, but also in all the
special annexes. The total number of unsupported tests is zero in
all categories. GNAT is the only Ada 95 technology that fully supports
all aspects of the Ada 95 language as defined in the international
standard. 

It is of particular note that GNAT fully supports the information
systems annex of the standard, and is the only Ada 95 compiler available
which does so. The Information Systems Annex provides COBOL-like facilities
for use in financial applications and means that GNAT is an attractive
technology for reengineering of legacy applications in COBOL.

GNAT is also the only Ada 95 capability that supports the Distributed
Systems Annex of the Ada 95 standard. For the first time, this provides
distribution capability integrated into a general-purpose language,
making Ada 95, as implemented by GNAT, an attractive candidate for
generating distributed applications, including client-server systems.

GNAT is used by a large community of Ada 95 programmers throughout
the world, and has been used for major Ada 95 applications in both
commercial and government applications.

Ada Core Technologies can offer high-quality support for these and
many other versions of GNAT. For information on obtaining these
validated compilers, please contact Ada Core Technologies.

  Ada Core Technologies
  73 Fifth Ave, Suite 11B
  New York, NY 10003

  +1 (212) 620-7300 (ext 117)

  sales@gnat.com

Re: ACT announces first 100% validations of Ada 95 compilers

[Jon S Anthony]

In article <dewar.857653821@merv> dewar@merv.cs.nyu.edu (Robert Dewar) writes:

> ADA CORE TECHNOLOGIES ANNOUNCES THE FIRST 100% ADA 95 VALIDATIONS

Congrats!

/Jon
-- 
Jon Anthony
Organon Motives, Inc.
Belmont, MA 02178
617.484.3383
jsa@organon.com

Re: ACT announces first 100% validations of Ada 95 compilers

[Ole-Hjalmar Kristensen FOU.TD/DELAB]

In article <dewar.857653821@merv> dewar@merv.cs.nyu.edu (Robert Dewar) writes:

< Stuff deleted>

Congratulations!

   GNAT is also the only Ada 95 capability that supports the Distributed
   Systems Annex of the Ada 95 standard. For the first time, this provides
   distribution capability integrated into a general-purpose language,
   making Ada 95, as implemented by GNAT, an attractive candidate for
   generating distributed applications, including client-server systems.

Ada 95 is certainly not the first general-purpose language which comes
with a standard distribution capability. Chill, Modula-3 and Occam (2)
are other examples.

How attractive a language is for implementing client-server systems
also depend on other factors than just having distribution built in.

1. Can a client call a server compiled with another compiler?
2. Can it be used in a network of heterogeneous computers?
3. Does the mechanism require that both clients and servers must be
programmed in Ada95?

And I'm not talking of what would be theoretically possible here, but
about what is actually practical to do within a project.

If you consider those factors unimportant, consider for a moment why
the telecommunications community have been so (obsessively) concerned
about protocol standards.
Of course, within an embedded or very tightly coupled system, these
factors may not be important.

   GNAT is used by a large community of Ada 95 programmers throughout
   the world, and has been used for major Ada 95 applications in both
   commercial and government applications.

   Ada Core Technologies can offer high-quality support for these and
   many other versions of GNAT. For information on obtaining these
   validated compilers, please contact Ada Core Technologies.

     Ada Core Technologies
     73 Fifth Ave, Suite 11B
     New York, NY 10003

     +1 (212) 620-7300 (ext 117)

     sales@gnat.com

Ole-Hj. Kristensen

Re: ACT announces first 100% validations of Ada 95 compilers

[Tom Moran]

What compiler switch settings were used for the validation?

Re: ACT announces first 100% validations of Ada 95 compilers

[Thomas Koenig]

In comp.lang.ada, dewar@merv.cs.nyu.edu (Robert Dewar) wrote:

>GNAT is also the only Ada 95 capability that supports the Distributed
>Systems Annex of the Ada 95 standard. For the first time, this provides
>distribution capability integrated into a general-purpose language,
>making Ada 95, as implemented by GNAT, an attractive candidate for
>generating distributed applications, including client-server systems.

Does this support any of the established standards for distributed
systems, e.g. RFC 1057, PVM, MPI, DCE, ...?  In other words,
what existing clients could I call from an Ada main program, and
vice versa?
-- 
74 a3 53 cc 0b 19

Re: ACT announces first 100% validations of Ada 95 compilers

[Laurent Pautet]

Ole-Hjalmar Kristensen FOU TD/DELAB <ohk@edeber.tfdt-o.nta.no> writes:

Ole-Hjalmar> How attractive a language is for implementing
Ole-Hjalmar> client-server systems also depend on other factors than
Ole-Hjalmar> just having distribution built in.

In the next release of GLADE, there will a nice example of client /
server example: a Swiss Bank. It uses plain RPCs and distributed
objects too.

Ole-Hjalmar> 1. Can a client call a server compiled with another
Ole-Hjalmar> compiler? 

Is there another compiler implementing distributed annex than GNAT ?

Ole-Hjalmar> 2. Can it be used in a network of
Ole-Hjalmar> heterogeneous computers?

Yes.

Ole-Hjalmar> 3. Does the mechanism require
Ole-Hjalmar> that both clients and servers must be programmed in
Ole-Hjalmar> Ada95?

Yes. But you can write wrappers to interface with other stubs (CORBA).

Ole-Hjalmar> And I'm not talking of what would be theoretically
Ole-Hjalmar> possible here, but about what is actually practical to do
Ole-Hjalmar> within a project.

It would be difficult to have a very general mechanism (usable with
any IDL), because some features only exist in Ada95. For instance, you
can safely abort RPCs (remote ATC) or propagate hidden exceptions or
distributed termination. Ada 95 distributed objects model is also in
some way more powerful than other models. So, it is quite difficult to
fit with simpler mechanisms.

-- 
-- Laurent

Re: ACT announces first 100% validations of Ada 95 compilers

[Chris Morgan]

In article <dewar.857653821@merv> dewar@merv.cs.nyu.edu (Robert Dewar)
writes:

> 
> ADA CORE TECHNOLOGIES ANNOUNCES THE FIRST 100% ADA 95 VALIDATIONS
> 

May I offer my sincere congratulations. In the UK, we at BAeSEMA use a
range of Ada compilers on several platforms. I can say that at present
GNAT is as good as and frequently better than any other Ada compiler I
have used and the rate of progress has been and remains astonishing.
-- 
Chris Morgan (cm@mihalis.demon.co.uk)

Re: ACT announces first 100% validations of Ada 95 compilers

[Samuel Tardieu]

>>>>> "Thomas" == Thomas Koenig <ig25@fg70.rz.uni-karlsruhe.de> writes:

Thomas> Does this support any of the established standards for
Thomas> distributed systems, e.g. RFC 1057, PVM, MPI, DCE, ...?  In
Thomas> other words, what existing clients could I call from an Ada
Thomas> main program, and vice versa?

You should read the Annex E of the Reference Manual; we are talking
here about distribtion of an Ada program, the protocol is not an issue
here.

  Sam
--
Samuel Tardieu -- sam@ada.eu.org

Re: ACT announces first 100% validations of Ada 95 compilers

[Robert Dewar]

Thomas asks

<<Does this support any of the established standards for distributed
systems, e.g. RFC 1057, PVM, MPI, DCE, ...?  In other words,
what existing clients could I call from an Ada main program, and
vice versa>>

The support in GNAT is as described in Annex E of the Ada reference manual.
The stream interface uses XDR, and the PCS that we supply (part of GLADE)
uses sockets.

RObert Dewar
Ada Core Technologies.

Re: ACT announces first 100% validations of Ada 95 compilers

[Richard Kenner]

In article <331FDDB7.7520@bix.com> tmoran@bix.com writes:
>What compiler switch settings were used for the validation?

-O0 -gnato -gnatv -gnatws -gnatd7

This turns off optimization, enables all checks, uses verbose listing
output format, disables warnings, and suppresses timestamps from
listings.

Re: ACT announces first 100% validations of Ada 95 compilers

[Robert Dewar]

Tom asks

<<What compiler switch settings were used for the validation>>

I would answer this (to avoid depending on memory) by going to the VSR.
You should do the same!

Re: ACT announces first 100% validations of Ada 95 compilers

[Chip Richards]

In article <dewar.857653821@merv>,
	dewar@merv.cs.nyu.edu (Robert Dewar) writes:

> ADA CORE TECHNOLOGIES ANNOUNCES THE FIRST 100% ADA 95 VALIDATIONS
> 
> Ada Core Technologies is pleased to announce that we have received
> formal notification from NIST that our recent Ada 95 validations
> under ACVC 2.0.1 have been officially registered.

Congratulations!

I'd like to say that GNAT has come as a spring breeze to one poor lonely Ada
programmer out here in the desert.  Nice to know that its respectable townie
brother has made good.

--
Chip

Re: ACT announces first 100% validations of Ada 95 compilers

[Mats Weber]

Richard Kenner wrote:
> 
> In article <331FDDB7.7520@bix.com> tmoran@bix.com writes:
> >What compiler switch settings were used for the validation?
> 
> -O0 -gnato -gnatv -gnatws -gnatd7
> 
> This turns off optimization, [...]

Did optimisation cause problems for the validation ? I'm asking this
because we have had many optimisation related problems with non-GNAT
compilers, to the point that we cannot deliver any code compiled with
optimisation.

Re: ACT announces first 100% validations of Ada 95 compilers

[Robert Dewar]

Mats asks

<<Did optimisation cause problems for the validation ? I'm asking this
because we have had many optimisation related problems with non-GNAT
compilers, to the point that we cannot deliver any code compiled with
optimisation.>>

No, optimization did not cause toruble for validation, because, as Richard
noted we chose to do the validation at -O0. You have to choose one
specific set of switches for validation, and we chose -O0 because it
makes debugging easier.

Obviously it is impractical to run the suite with every possible combination
of compile switches, but it is useful to run the suite at different
optimziation levels, and we are in fact in the process of doing that
now, and at some point will do it systematically (we run the ACVC every
night on several machines, so our plan was to use different options
each night :-)

In practice, we have known cases of bugs appearing only at -O2, and
also cases of bugs appearing only at -O0, but both events are rare,
and lots of our customers are routinely using -O2 for delivered code.
The GNAT compiler itself is tyically compiled with -O1.

Robert Dewar
Ada Core Technologies

Re: ACT announces first 100% validations of Ada 95 compilers

[Richard Kenner]

In article <332412BB.665C@elca-matrix.ch> Mats.Weber@elca-matrix.ch writes:
>Did optimisation cause problems for the validation ? I'm asking this
>because we have had many optimisation related problems with non-GNAT
>compilers, to the point that we cannot deliver any code compiled with
>optimisation.

We've rarely tried experimenting with options on ACVC.  In general,
there have been very few optimization problems with GNAT; most problems
occur with or without optimization.  GCC (and hence GNAT) used to have
the property that optimization was more reliable than no optimization
because it was more heavily used, but, at the moment I'd say both are
roughly equally reliable.

That being said, however, we became aware about a week ago about a bug
that would cause some ACVC tests to fail on Sparc with optimization
turned on.  That turned out to be a GCC bug, which I fixed this morning.