Gnat Executable Size

Gnat Executable Size

[Robert B. Love]

Over in the Modula-2 newsgroup somebody was asking about free compilers.
Another person pointed out GNAT as suitable for his task and free.
Ada vs. M2 wasn't the issue.  The original poster said he'd tried GNAT
on a simple "hello world" program and was surprised to see a 200K 
executable.  Others said their favorite M2 compiler yielded about 10K
executables for this.

I tried a simple Put_Line program myself with GNAT under Linux and
got an 84K program.  I've played with Pragma Optimze(Space) and
Pragma Supress (All_Checks) but it stays about  84K.  

I then run strip on the GNAT generated executable and get it down to
about 47K.

How small can GNAT get "Hello World"?  What compiler options and pragmas
would you use for this?



----------------------------------------------------------------
 Bob Love                                   MIME & NeXT Mail OK
 rlove@neosoft.com                            
----------------------------------------------------------------

Re: Gnat Executable Size

[Niklas Holsti]

Robert B. Love wrote:
> 
> Over in the Modula-2 newsgroup somebody was asking about free compilers.
> Another person pointed out GNAT as suitable for his task and free.
> Ada vs. M2 wasn't the issue.  The original poster said he'd tried GNAT
> on a simple "hello world" program and was surprised to see a 200K
> executable.  Others said their favorite M2 compiler yielded about 10K
> executables for this.
> 
> I tried a simple Put_Line program myself with GNAT under Linux and
> got an 84K program.  I've played with Pragma Optimze(Space) and
> Pragma Supress (All_Checks) but it stays about  84K.
> 
> I then run strip on the GNAT generated executable and get it down to
> about 47K.
> 
> How small can GNAT get "Hello World"?  What compiler options and pragmas
> would you use for this?

On my i486 Slackware Linux 2.0.0, GNAT 3.05 generates an executable
for Put_Line("Hello world!") of 22842 bytes (not stripped) or 10400
bytes (stripped). Compiler options were "-c -g". Seems comparable
to the M2 sizes. Source code below.

- Niklas

with Text_IO;
procedure HELLO is
begin
   Text_IO.Put_Line ("Hello world!");
end HELLO;

Re: Gnat Executable Size

[Bryce Bardin]

Robert B. Love wrote:
> 
> Over in the Modula-2 newsgroup somebody was asking about free compilers.
> Another person pointed out GNAT as suitable for his task and free.
> Ada vs. M2 wasn't the issue.  The original poster said he'd tried GNAT
> on a simple "hello world" program and was surprised to see a 200K
> executable.  Others said their favorite M2 compiler yielded about 10K
> executables for this.

This is a naive view of program size.  The size of executables depend on
several factors, including:

  1) the target machine architecture;
  2) the size and complexity of the program source code;
  3) the amount of debugging information included, if any;
  4) the degree and kinds of optimization performed by the compilation
     system and linker; and
  5) the size of any libraries used (which must be bound into the
     image).

Let us assume that you want to make a "fair" comparison between two
languages, comparing "oranges" with "oranges" and "apples" with
"apples".  
Then you need to control all of the above factors, and make
sure that the programs really are doing the same work:

  1) use the same target (this is easy); 
  2) make sure the program has as nearly as possible the same
semantic    
     effect in both languages (this is *much* harder than it appears at
     first sight);
  3) make sure that debugging information is stripped out (easy);
  4) compile and build using the options that lead to the smallest
     executable size (which will, in general, be *slower* to execute
     and may change the semantics of the program!) (this is
non-trivial);
     and
  5) use libraries with identical or very similar capabilities 
     (again hard to do between languages).

In comparing languages with differing semantics, the issue of program
behavior under stress must be accounted for.  For example, in Ada,
many checks that help to improve robustness are performed, both at
compile-time and at run-time.  If the comparison language doesn't
perform similar checks, then the comparison may be unfair (and there 
ain't no such thing as a free lunch, or in this case, a free check).  
Since most languages don't perform similar checks automatically, 
usually checks are turned off in the Ada executable to improve the 
comparison.  This defeats much of the ability for a well-designed Ada
program to perform reliably, but brings the semantics closer to the
norm for most other languages. 

Making sure that no debugging information is included is often 
overlooked.  Making sure that the "best" size optimizations are
performed 
can be surprisingly difficult.  You need to understand how both 
compilation systems work and all of the options quite well in order to 
do this correctly.
  
But the most difficult (and at the same time the easiest) parameter to 
control is the size of the libraries bound into the image.  The
capabilities of Ada's Text_Io package are much more sophisticated
than those of other languages' I/O libraries, so the Ada library size
will inevitably be larger.  On the other hand, an Ada program using I/O 
can be made more reliable.  

If the only measure of size that you use is the size of the "standard"
"hello world" program, Ada will always lose!

But there is also a simple way of making the comparison *very* fair: 
  use the same library!  
In Ada, import the other languge's I/O library and call it the
same way the other languges does.  Then the executable sizes will
be *very* similar (and in the case of gcc and gnat, the executables
may, in fact, actually be identical).

In the end, however, the best measure of executable size will not be a 
trivial program like "hello world", because most programs are not that 
small and simple. 

This is just one of the reasons why both hardware and software vendors 
and sophisticated developers find both hardware and application and 
development software benchmark testing to be a difficult and frustrating 
task.

 
(snip)

> 
> ----------------------------------------------------------------
>  Bob Love                                   MIME & NeXT Mail OK
>  rlove@neosoft.com
> ----------------------------------------------------------------

Re: Gnat Executable Size

[Corey Minyard]

Niklas Holsti <nholsti@icon.fi> writes:

> Robert B. Love wrote:
> > 
> > Over in the Modula-2 newsgroup somebody was asking about free compilers.
> > Another person pointed out GNAT as suitable for his task and free.
> > Ada vs. M2 wasn't the issue.  The original poster said he'd tried GNAT
> > on a simple "hello world" program and was surprised to see a 200K
> > executable.  Others said their favorite M2 compiler yielded about 10K
> > executables for this.
> > 
> > I tried a simple Put_Line program myself with GNAT under Linux and
> > got an 84K program.  I've played with Pragma Optimze(Space) and
> > Pragma Supress (All_Checks) but it stays about  84K.
> > 
> > I then run strip on the GNAT generated executable and get it down to
> > about 47K.
> > 
> > How small can GNAT get "Hello World"?  What compiler options and pragmas
> > would you use for this?
> 
> On my i486 Slackware Linux 2.0.0, GNAT 3.05 generates an executable
> for Put_Line("Hello world!") of 22842 bytes (not stripped) or 10400
> bytes (stripped). Compiler options were "-c -g". Seems comparable
> to the M2 sizes. Source code below.
> 

With 3.11b I get 57832 after stripping no matter how I compile it
(optimiation at any level, with or without runtime checks) on my Linux
box.  Which makes sense because there's nothing to check or optimize.

This seems like a silly thing to be worried about.  All it measures is
the efficiency of including a minimal set of I/O libraries for a
minimal program.

Note that the .o files for a C and Ada version of "Hello World" are
within 16 bytes in size for gcc and gnat, and about that difference in
actual executable code and data.

-- 
Corey Minyard               Internet:  minyard@acm.org
  Work: minyard@nortel.ca       UUCP:  minyard@wf-rch.cirr.com

Re: Gnat Executable Size

[Lieven Marchand]

rlove@antispam.neosoft.com (Robert B. Love ) writes:

> [snipped discussion of size of Hello world program]
> How small can GNAT get "Hello World"?  What compiler options and pragmas
> would you use for this?
> 

These kind of discussions seem to come up about every 3 months. I
never understood why the size of this particular program is of any
real importance. In some of the languages I really like like Modula-3
or Common Lisp such an executable could very well be a few
megabytes. So what? These kind of languages come with a large and
usefull standard environment that in the end saves you a lot of
trouble. This also means that the size of the executable grows very
modestly with increased functionality and complexity.

Unless you have special requirements like restricted space in an
embedded application or something that needs to start up instantly a
large number of times per second it doesn't seem a very usefull
question to ask.

-- 
Lieven Marchand <mal@bewoner.dma.be> 
------------------------------------------------------------------------------
Few people have a talent for constructive laziness. -- Lazarus Long

Re: Gnat Executable Size

[Donald Duck]

As far as I understand, the comparison was _not_ between Ada and M2, but
between GNAT and any other compiler. What is true is, that the GNAT code
is bigger than code from other (even other Ada) compilers. For your
comparison you have to add the size of the EMX runtime also.
 

Bryce Bardin wrote:
> 
> Robert B. Love wrote:
> >
> > Over in the Modula-2 newsgroup somebody was asking about free compilers.
> > Another person pointed out GNAT as suitable for his task and free.
> > Ada vs. M2 wasn't the issue.  The original poster said he'd tried GNAT
> > on a simple "hello world" program and was surprised to see a 200K
> > executable.  Others said their favorite M2 compiler yielded about 10K
> > executables for this.
> 
> This is a naive view of program size.  The size of executables depend on

Re: Gnat Executable Size

[dennison]

In article <6volj0$250$1@uuneo.neosoft.com>,
  rlove@antispam.neosoft.com (Robert B. Love ) wrote:

> How small can GNAT get "Hello World"?  What compiler options and pragmas
> would you use for this?

To save myself some typing, I will repost the response I gave to basicly the
same question 4 days ago in the intel-objectAda mailing list:
-------------
{Name Deleted} wrote:

>         I'm glad to see that the ObjectAda (7.1.1) compiler produces smaller
>         executables.  But even with that I still get ridiculed by the C/C++
>         hacks that I know about the size of "hello world".  Under 7.1.1,
>         "hello world" is almost a meg (~775K) for the debug version, and
>         just over 200K for the release version with "Suppress Checks All".
>         The C/C++ exe's they show me are about 20-30K.

There was a quite long thread about this on c.l.a. a couple of years back. If
you are interested, go to http://www.dejanews.com and search comp.lang.ada on
"Executable File Size". There is a certain fixed-size overhead to Ada
(exceptions, tasks, Text_IO on files enums and floats, etc) Basicly the size
of "Hello World" is a useless metric unless you are actually trying to write
an embedded "Hello World" app. Even in that case, the C user would be
compelled to include the size of all his runtime libraries, which would
lessen the difference greatly.

If you are instead looking for a flip answer, mine would be that Ada has a
fair bit of overhead because it is assuming you will be doing "real work".
The less work your program does, the better C is for the job. :-)

-------------

This is the second time in a week this silly question has come up (if you
count once on the intel-objectada mailing list). Why all the sudden interest
in writing useless programs? Did the DoD just put out an RFC for an embedded
"Hello World" program?

Someone please, please put this in the FAQ.

--
T.E.D.

-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own    

Re: Gnat Executable Size

[Andi Kleen]

Corey Minyard <minyard@acm.org> writes:

> Niklas Holsti <nholsti@icon.fi> writes:
> 
> > Robert B. Love wrote:
> > > 
> > > Over in the Modula-2 newsgroup somebody was asking about free compilers.
> > > Another person pointed out GNAT as suitable for his task and free.
> > > Ada vs. M2 wasn't the issue.  The original poster said he'd tried GNAT
> > > on a simple "hello world" program and was surprised to see a 200K
> > > executable.  Others said their favorite M2 compiler yielded about 10K
> > > executables for this.
> > > 
> > > I tried a simple Put_Line program myself with GNAT under Linux and
> > > got an 84K program.  I've played with Pragma Optimze(Space) and
> > > Pragma Supress (All_Checks) but it stays about  84K.
> > > 
> > > I then run strip on the GNAT generated executable and get it down to
> > > about 47K.
> > > 
> > > How small can GNAT get "Hello World"?  What compiler options and pragmas
> > > would you use for this?
> > 
> > On my i486 Slackware Linux 2.0.0, GNAT 3.05 generates an executable
> > for Put_Line("Hello world!") of 22842 bytes (not stripped) or 10400
> > bytes (stripped). Compiler options were "-c -g". Seems comparable
> > to the M2 sizes. Source code below.
> > 
> 
> With 3.11b I get 57832 after stripping no matter how I compile it
> (optimiation at any level, with or without runtime checks) on my Linux
> box.  Which makes sense because there's nothing to check or optimize.

The standard GNAT binary distribution on Linux seems to generate statically
linked libraries, while other distributions generate shared linked libraries
(to a libgnat.so). This can explain the size difference.

-Andi 

Re: Gnat Executable Size

[Dave Wood]

While I agree with T.E.D. that "hello" is truly a 
useless metric in general, it can have meaning in the 
context of the "first impression", where the uninformed 
or uninitiated can fall prey to the default notion that 
"Ada is too big".  First impressions can hurt, and in
the Ada business we can't afford to be too cavalier
about first impressions (or second impressions, for
that matter).  In general, the world seems much more
forgiving of the myriad calamaties lurking beneath
the surface of C/C++, waiting to pounce upon the 
unsuspecting Pandora unlucky enough to use it for
serious business.  Such is our lot in life.

In that regard, it might be worth consideration that the 
same program, when built with ObjectAda's "runtime in a dll"
option, is about 25K, which is basically in line with an
equivalent C program.  Of course, the runtime size does
not grow linearly with program complexity, so for programs
of any significance, the runtime size is a non-issue.  To 
my knowledge, this hasn't been a serious issue for any of 
our customers.

Of course, I'm talking here about "native" targets, like
Windows.  For embedded targets, it's a different story.
For example, ObjectAda targeting our RAVEN kernel for
PowerPC, 68K, or Intel will result in minimal footprint
size on the order of a few K.  A C program bound with
Microsoft's Windows CE (which MS promotes as an RTOS),
is going to result in something on the order of several
hundred K minimum footprint, way too big for a lot of
embedded apps, substantially bigger than our closest 
equivalent product, ObjectAda for Intel/ETS (both ETS 
and WinCE are Win32 compliant.)  I grant you, in some 
ways this is an apples-to-oranges comparison, but then 
again so is pretending that "hello world" is a rational 
metric for comparing Ada and C.  It's all a matter of
what it is you are trying to do.

-- Dave Wood, Aonix
-- Product Manager, ObjectAda for Windows
-- http://www.aonix.com


dennison@telepath.com wrote:
> 
> In article <6volj0$250$1@uuneo.neosoft.com>,
>   rlove@antispam.neosoft.com (Robert B. Love ) wrote:
> 
> > How small can GNAT get "Hello World"?  What compiler options and pragmas
> > would you use for this?
> 
> To save myself some typing, I will repost the response I gave to basicly the
> same question 4 days ago in the intel-objectAda mailing list:
> -------------
> {Name Deleted} wrote:
> 
> >         I'm glad to see that the ObjectAda (7.1.1) compiler produces smaller
> >         executables.  But even with that I still get ridiculed by the C/C++
> >         hacks that I know about the size of "hello world".  Under 7.1.1,
> >         "hello world" is almost a meg (~775K) for the debug version, and
> >         just over 200K for the release version with "Suppress Checks All".
> >         The C/C++ exe's they show me are about 20-30K.
> 
> There was a quite long thread about this on c.l.a. a couple of years back. If
> you are interested, go to http://www.dejanews.com and search comp.lang.ada on
> "Executable File Size". There is a certain fixed-size overhead to Ada
> (exceptions, tasks, Text_IO on files enums and floats, etc) Basicly the size
> of "Hello World" is a useless metric unless you are actually trying to write
> an embedded "Hello World" app. Even in that case, the C user would be
> compelled to include the size of all his runtime libraries, which would
> lessen the difference greatly.
> 
> If you are instead looking for a flip answer, mine would be that Ada has a
> fair bit of overhead because it is assuming you will be doing "real work".
> The less work your program does, the better C is for the job. :-)
> 
> -------------
> 
> This is the second time in a week this silly question has come up (if you
> count once on the intel-objectada mailing list). Why all the sudden interest
> in writing useless programs? Did the DoD just put out an RFC for an embedded
> "Hello World" program?
> 
> Someone please, please put this in the FAQ.
> 
> --
> T.E.D.
> 
> -----------== Posted via Deja News, The Discussion Network ==----------
> http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own

Re: Gnat Executable Size

[Robert B. Love]

In <6vtdfc$r50$2@xenon.inbe.net> Lieven Marchand wrote:
> rlove@antispam.neosoft.com (Robert B. Love ) writes:
> 
> > [snipped discussion of size of Hello world program]
> > How small can GNAT get "Hello World"?  What compiler options and 
pragmas
> > would you use for this?
> > 
> 
> These kind of discussions seem to come up about every 3 months. I
> never understood why the size of this particular program is of any
> real importance. In some of the languages I really like like Modula-3

My original intent when I first asked the question was to go back
to the M2 newsgroup and respond to the guy who didn't want GNAT 
because the executable was too large.  I wasn't worried about it
myself.  Thanks to everybody who responded.


----------------------------------------------------------------
 Bob Love                                   MIME & NeXT Mail OK
 rlove@neosoft.com                            
----------------------------------------------------------------

Re: Gnat Executable Size

[geertbosch]

In article <3621E42C.2920@Entenhausen.net>,
  Donald.Duck@Entenhausen.net wrote:
> As far as I understand, the comparison was _not_ between Ada and M2, but
> between GNAT and any other compiler. What is true is, that the GNAT code
> is bigger than code from other (even other Ada) compilers. For your
> comparison you have to add the size of the EMX runtime also.

When you use dynamic linking for the EMX and GNAT run-time, you
can get very small executable sizes indeed. (For OS/2, you can shrink
the executables a lot using the lxlite program, which comes with full
sources using GNU licensing.) When doing all this it is possible
to create executables that are less than 5 kB with GNAT on OS/2.

-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own    

Re: Gnat Executable Size

[dewar]

In article <3621E42C.2920@Entenhausen.net>,
  Donald.Duck@Entenhausen.net wrote:
> As far as I understand, the comparison was _not_ between Ada and M2, but
> between GNAT and any other compiler. What is true is, that the GNAT code
> is bigger than code from other (even other Ada) compilers. For your
> comparison you have to add the size of the EMX runtime also.

First, the EMX runtime is relevant only to the OS/2 version among the
officially supported versions of GNAT, so that seems a bit of an odd
thing to mention.

Second, we occasionally get notes of this kind, and in 100% of cases, they
have turned out to be based on apples-vs-oranges comparisons. The differences
come from many bogus sources:

  unoptimized vs optimized code
  debugging information not stripped
  non-comparable programs
  statically vs dynamically linked libraries
  etc.

After a while it is hardly worth following up each case!

Robert Dewar
Ada Core Technologies

-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own    

Re: Gnat Executable Size

[Donald Duck]

dewar@gnat.com wrote:
>   Donald.Duck@Entenhausen.net wrote:
    ^^^^^^^^^^^ only to stop the flood of spam mails I get since my
first comment in COMP.LANG.* with my real name.
  
> First, the EMX runtime is relevant only to the OS/2 version among the

That's interessting for me. I still do all my work in OS/2 environment.
I wrote a few apps and when I first tried to run them on an other PC it
didn't work. So when I switch to WinNT in the future, the EXE-files will
run on every maschine without extra stuff ?

Alfred

Gnat Free ?

[Donald Duck]

I have one question to GNAT. Is it still free of charge ? The last
version I had was 3.07 (from cs.nyu.edu), but I heard that GNAT is not
longer available free, and any trial to connect to the server above
failed. I've got nearly all jind of error messages: server not
available, connection refused, unknown domain, ...

Alfred

Re: Gnat Free ?

[Pascal Obry]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 553 bytes --]


Donald Duck a �crit dans le message <3627196D.720A@Entenhausen.net>...
>I have one question to GNAT. Is it still free of charge ? The last
>version I had was 3.07 (from cs.nyu.edu), but I heard that GNAT is not
>longer available free, and any trial to connect to the server above
>failed. I've got nearly all jind of error messages: server not
>available, connection refused, unknown domain, ...
>
>Alfred

Of course it is. You have, since more than 1 year, a 3.10p version available
for many platforms !

ftp://ftp.cs.nyu.edu/pub/gnat

Pascal.

Re: Gnat Free ?

[Donald Duck]

Pascal Obry wrote:
> Of course it is. You have, since more than 1 year, a 3.10p version available
> for many platforms !

Many thanks. I've heard that it would be no longer free as there is
commercial support now.

Alfred

Re: Gnat Free ?

[dennison]

In article <3627196D.720A@Entenhausen.net>,
  Donald.Duck@Entenhausen.net wrote:
> I have one question to GNAT. Is it still free of charge ? The last
> version I had was 3.07 (from cs.nyu.edu), but I heard that GNAT is not
> longer available free, ...

You heard wrong. Well, I suppose it could be considered not free, depending
on your definition of free, but lets not get into that. Gnat is under the
GPL, and thus can *never* be made proprietary.

>                        ...and any trial to connect to the server above
> failed. I've got nearly all jind of error messages: server not
> available, connection refused, unknown domain, ...

I just tried ftp://cs.nyu.edu/pub/gnat/ . It worked for me. I downloaded the
emacs-ada mode from there yesterday (10/16) too. If you continue to have
problems, contact your system admin. Our FAQ at
http://www.adahome.com/FAQ/comp-lang-ada.html#gnat-id lists some alternate
download sites as well.


--
T.E.D.

-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own    

Re: Gnat Free ?

[Ronald Cole]

dennison@telepath.com writes:
> You heard wrong. Well, I suppose it could be considered not free, depending
> on your definition of free, but lets not get into that. Gnat is under the
> GPL, and thus can *never* be made proprietary.

Oh yeah?  Just try to talk Dewar out of 3.11b without crossing his
palm with some cash...  He's even threatened to consider not doing
business with any one who starts giving out wavefront ("non-p")
releases (i.e., exercises their right under the GPL).  There is little
difference in effect between "proprietary" and "hoarded".  The GPL
does not prevent hoarding.  Stallman has stated that Dewar is
perfectly free to decide who he gives any particular version of GNAT
to, and rejected the notion that the "act of distributing" under the
GPL requires one to become a distributor to every Tom, Dick, and Harry
and subsequently asks for a copy.

Of course, I've always secretly wondered if Dewar would shutdown his
business completely if every one of his clients started giving out
wavefront releases...

-- 
Forte International, P.O. Box 1412, Ridgecrest, CA  93556-1412
Ronald Cole <ronald@forte-intl.com>      Phone: (760) 499-9142
President, CEO                             Fax: (760) 499-9152
My PGP fingerprint: 15 6E C7 91 5F AF 17 C4  24 93 CB 6B EB 38 B5 E5

Re: Gnat Free ?

[Andi Kleen]

Ronald Cole <ronald@forte-intl.com> writes:
> 
> Of course, I've always secretly wondered if Dewar would shutdown his
> business completely if every one of his clients started giving out
> wavefront releases...

Unlikely. If someone is willing to pay big $$ he is usually is seriously
interested in support. And the others wouldn't give money anyways. I doubt ACT 
would lose much sales if at all when they were a bit more open with their 
releases. The advantage ACT would receive instead is more testing
which gives better software which translates into paying customer 
satisfaction which generates more money in the long run.

-Andi

Re: Gnat Free ?

[dewar]

In article <3627196D.720A@Entenhausen.net>,
  Donald.Duck@Entenhausen.net wrote:
> I have one question to GNAT. Is it still free of charge ? The last
> version I had was 3.07 (from cs.nyu.edu), but I heard that GNAT is not
> longer available free, and any trial to connect to the server above
> failed. I've got nearly all jind of error messages: server not
> available, connection refused, unknown domain, ...
>
> Alfred


Yes, the public version of GNAT is available at no charge, from multiple
download sites, including cs.nyu.edu. If you are having trouble with one
you should try another. We have no significant reports of trouble accessing
cs.nyu.edu, but of course you may have a particularly troublesome route.

The current public version of GNAT is 3.10p. This is available for most
targets, including both those created by ACT, and some volunteer ports,
including NetBSD, FreeBSD, and recently DOS.

We are hoping to release 3.11p in the very near future, we are doing some
final testing on the NT version, and we will also issue versions of ASIS and
GLADE that are compatible with this version at the same time.

Ada Core Technologies provides a fully supported commercial version, GNAT
Professional, that is available directly from us, including fully validated
versions (GNAT is still the only 100% validated Ada 95 technology, including
the core and all optional annexes). The GNAT Professional technology is
typically one version more advanced than the public version, since we do
not make versions public until they have already been extensively used.
GNAT Professional also comes with full support.

But if you can use the slightly older version, and are happy to use
unsupported software (which is typically perfectly reasonable for student use
or academic research use), then the public version is definitely still
available for your use. Note however that Ada Core Technologies does not
provide any kind of support for the public version, and cannot make any
warranties or statements of any kind concerning the public version. In
particular, we have no way of knowing that the version you obtained is the
same as what we originally put out there, and in the case of the volunteer
ports, it is of course the case that we have no knowledge of these products.

Robert Dewar
Ada Core Technologies
generally similar to the public

-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own    

Re: Gnat Free ?

[dewar]

In article <708040$4h4$1@nnrp1.dejanews.com>,
  dennison@telepath.com wrote:
> In article <3627196D.720A@Entenhausen.net>,
>   Donald.Duck@Entenhausen.net wrote:
> > I have one question to GNAT. Is it still free of charge ? The last
> > version I had was 3.07 (from cs.nyu.edu), but I heard that GNAT is not
> > longer available free, ...
>
> You heard wrong. Well, I suppose it could be considered not free, depending
> on your definition of free, but lets not get into that. Gnat is under the
> GPL, and thus can *never* be made proprietary.

That might be a bit misleading. The original poster was not asking whether
GNAT is proprietary, but whether it was available free of charge.

The GPL has nothing at all to do with the
amount that is charged for software (you can have non-GPL'ed software that
costs nothing, and GPL'ed software that has non-zero cost). The GPL has to
do with redistribution rights, and rights to get hold of the sources.

That's an important distinction, because we often encounter people who think
that the GPL requires us to make public releases of GNAT. This is not at all
the case. Ada Core Technologies is certainly not required to make public
releases. We choose to do so periodically, updating the public version to
the latest and greatest GNAT technology, because we feel it is helpful for
the Ada community to have a version that is accessible for downloading at
no charge, particularly for the academic community.

Robert Dewar
Ada Core Technologies

-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own    

Re: Gnat Free ?

[dewar]

In article <36275DBF.1E96@Entenhausen.net>,
  Donald.Duck@Entenhausen.net wrote:
> Pascal Obry wrote:
> > Of course it is. You have, since more than 1 year, a 3.10p version
available
> > for many platforms !
>
> Many thanks. I've heard that it would be no longer free as there is
> commercial support now.


Well there has been commercial support for GNAT for quite a while now (four
years), but public versions have also been made available periodically for
this period. During the original contract, the DoD required public releases.
After ACT took over, we continued the tradition of making periodic public
releases, and will continue to do so.

In the free software model, there is nothing incompatible about having a
freely available public version, and a commercial version with commercial
support. Different users have different needs. A student wanting to learn
Ada can perfectly well use the public version, and does not need commercial
support.

A large company using Ada for a critical project will almost certainly NOT
want to use unsupported software (certainly we avoid the use of unsupported
software here at ACT itself). And for them the commercial support is what
makes it possible for them to take advantage of the technical capabilities
of GNAT, while still getting good support.

Furthermore, the fees that the commercial customers pay help to support the
continued development and availability of the public version, which in turn
helps support the continued use of Ada, which is to everyone's advantage.

That's the free software model, and it is working very effectively to support
the continued development of GNAT. As those of you who are using 3.11
technology now know, the 3.11 version is a significant advance on 3.10, and
you will all get to see this shortly when we release 3.10p

Robert Dewar
Ada Core Technologies

-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own    

Re: Gnat Free ?

[David C. Hoos, Sr.]

Donald Duck wrote in message <3627196D.720A@Entenhausen.net>...
>I have one question to GNAT. Is it still free of charge ? The last
>version I had was 3.07 (from cs.nyu.edu), but I heard that GNAT is not
>longer available free, and any trial to connect to the server above
>failed. I've got nearly all jind of error messages: server not
>available, connection refused, unknown domain, ...
>
Here is a reasonably coplete list of gnat mirror sites:

US sites:
ftp://ftp.cdrom.com/pub/languages/ada/compiler/gnat/distrib
ftp://wuarchive.wustl.edu/languages/ada/compiler/gnat/distrib

European sites:
ftp://ftp.cs.kuleuven.ac.be/pub/Ada-Belgium/mirrors/gnu-ada
ftp://ftp.dit.upm.es/mirror/cs.nyu.edu/pub/gnat
ftp://ftp.cs.tu-berlin.de/pub/gnu/gnat
ftp://ftp.cci.de/pub/Ada/gnat
ftp://ftp.informatik.rwth-aachen.de/pub/mirror/cs.nyu.edu/pub/gnat
ftp://ftp.ibp.fr/pub/gnat
ftp://ftp.cnam.fr/pub/Ada/PAL/compiler/gnat/distrib
ftp://ftp.eunet.no/pub/gnu/gnatftp://ftp.fsz.bme.hu/pub/ada/gnat
ftp://snowwhite.it.brighton.ac.uk/gnat

Australia site:
ftp://ftp.canberra.edu.au/pub/ise/gnat

Re: Gnat Free ?

[Jerry van Dijk]

Ronald Cole (ronald@forte-intl.com) wrote:

: Of course, I've always secretly wondered if Dewar would shutdown his
: business completely if every one of his clients started giving out
: wavefront releases...

If I understand you correctly you want ACT to follow current practices
by prematurely releasing possibly unreliable software ?

Jerry.
-- 
-- Jerry van Dijk  | email: jdijk@acm.org
-- Leiden, Holland | member Team-Ada
-- Ada & Win32: http://stad.dsl.nl/~jvandyk

Re: Gnat Free ?

[Jean-Pierre Rosen]

David C. Hoos, Sr. a �crit dans le message ...
>Here is a reasonably coplete list of gnat mirror sites:
>[...]
>ftp://ftp.ibp.fr/pub/gnat


Please note that this URL has changed to ftp://ftp.lip6.fr/pub/gnat

The old "ibp" still works as an alias, but can be removed at any time.
Please update your links accordingly!
----------------------------------------------------------------------------
                  J-P. Rosen (Rosen.Adalog@wanadoo.fr)
      Visit Adalog's web site at http://perso.wanadoo.fr/adalog

Re: Gnat Free ?

[dewarr]

In article <87pvbs6zb3.fsf@yakisoba.forte-intl.com>,
  Ronald Cole <ronald@forte-intl.com> wrote:
> dennison@telepath.com writes:
> > You heard wrong. Well, I suppose it could be considered not free, depending
> > on your definition of free, but lets not get into that. Gnat is under the
> > GPL, and thus can *never* be made proprietary.
>
> Oh yeah?  Just try to talk Dewar out of 3.11b without crossing his
> palm with some cash...  He's even threatened to consider not doing
> business with any one who starts giving out wavefront ("non-p")
> releases (i.e., exercises their right under the GPL).  There is little
> difference in effect between "proprietary" and "hoarded".  The GPL
> does not prevent hoarding.  Stallman has stated that Dewar is
> perfectly free to decide who he gives any particular version of GNAT
> to, and rejected the notion that the "act of distributing" under the
> GPL requires one to become a distributor to every Tom, Dick, and Harry
> and subsequently asks for a copy.
>
> Of course, I've always secretly wondered if Dewar would shutdown his
> business completely if every one of his clients started giving out
> wavefront releases...


Here is how things work at ACT. We are constantly updating the technology.
At any particular time, we have an internal development version, which we
call the wavefront version. This is pretty well tested internally, but
not very well tested externally, since we do not expect our customers to
change compilers every day (a complete new set of compilers is built and
tested every day).

If customers need new functionality or,  as happens sometimes (much less
frequently than it used to), runs into a bug for which there is no work
around, then we give out a copy of the wavefront. Usually the wavefront is in
reasonable shape, because of our internal testing, but quite often there are
problems that need fixing. It works fine for our customers to use the wave
front compilers most of the time, not in the least because they have full
support from us so we can fix all problems.

Periodically, we gear up for a new release. The wavefront becomes the
tentative release candidate, and we encourage customers to beta test the
new release, and thus catch any new errors and regressions, and any other
undesirable behavior to be corrected before the final release.

Then when the final release is made, we make a public version of this
release. This is the process we are currently following for version
3.11b. The status is as follows. Our customers have tested 3.11b releases
for most machines. We have found a few errors, and have corrected them, and
in particular, the NT version of the debugger had some glitches which
necessitated a second beta version.

Things are looking pretty good now for the final 3.11b release, which will be
called 3.11b2, and at that point we will release an equivalent version with
the version number 3.11p for the public release.

Some people would like us to post our wavefronts day by day, but we decline,
not because we are hoarding the software, but because we don't think it is
useful to have unsupported public versions around that are going to cause
people trouble. It is one thing for a customer to take a new beta version and
adapt it for their use with full support from ACT, it is quite another for
a student to pick up a public version without any kind of support.

We think it is in the best interests of the Ada community and the GNAT project
to ensure that the public versions of GNAT that are distributed are as free
as possible from glitches that will cause unsupported users to run into
trouble. We think that it is better to wait, and perhaps be a little behind
the latest-and-greatest, rather than deal with glitches.

Now if you are an expert hacker (in the good sense of the word) like Ronald,
you could probably indeed get by at least some of the problems on your own,
but we do not distribute the public version for the Ronald Coles of the
world, but rather for students learning Ada for the first time, a very
different market.

As for our customers releasing the wavefront, it would be unfortunate, since
it would mean that there were insufficiently tested versions publicly
floating around. If these versions had significant problems, it could reflect
negatively on Ada in general, the GNAT project, and free software in general,
and we think it better to avoid this. Our customers share this viewpoint, and
understand that it is in the best interests of the GNAT project for them not
to publish the wavefront versions.

We certainly have never "threatened to consider not doing
business with any one who starts giving out wavefront". That is
pure fantasy on the part of Ronald Cole. If any customer ever asked
us about our views here, we would simply explain why we felt it was
in the best interests of the GNAT project that wavefront preleases
not be generally released, but the decision would of course be up
to them. Interestingly, we have never been asked the question, let
alone given the answer that Mr. Cole fantasizes.

It certainly would not be a disaster that would cause us to close down
operations here at ACT, that is a very odd view of the world. Our customers
do not pay for support contracts to get hold of the beta releases of our new
technology earlier than otherwise, but to get support. Of course in some
cases, this support does mean that they can get their hands on special
versions that incorporate useful enhancements done for them earlier than
otherwise, but this is a natural and desirable consequence of the way we
operate.

The distance between the current wavefront and the public version varies of
course with the time span. In a couple of weeks, when 3.11p is out, the
public version will be close to synchronized with the commercially supported
version.

Then of course the cycle will start again, and we already have some exciting
new enhancements underway for an eventual 3.12 release. GNAT development
continues at a fast pace, and this will benefit both users of the supported
commercial version, and the publicly available version.

Robert Dewar
Ada Core Technologies

-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own    

Re: Gnat Free ?

[dewarr]

In article <m3af2wtd4e.fsf@fred.muc.de>,
  Andi Kleen <ak@muc.de> wrote:
> Ronald Cole <ronald@forte-intl.com> writes:
> >
> > Of course, I've always secretly wondered if Dewar would shutdown his
> > business completely if every one of his clients started giving out
> > wavefront releases...
>
> Unlikely. If someone is willing to pay big $$ he is usually is seriously
> interested in support. And the others wouldn't give money anyways. I doubt
ACT
> would lose much sales if at all when they were a bit more open with their
> releases. The advantage ACT would receive instead is more testing
> which gives better software which translates into paying customer
> satisfaction which generates more money in the long run.


That's absolutely right, I don't think it would affect sales one way or
another, but as I explained in my previous note, that is not the reason for
our hesitation to prematurely do public releases of insufficiently tested
versions of GNAT.

We are perfectly aware that there are enthusiasts who would like to get their
hands on the latest technology, and don't care if it has problems. But that is
not the public we are addressing in making public versions of GNAT available.

Robert Dewar
Ada Core Technologies

-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own    

Re: Gnat Free ?

[The Ludwig Family]

dewarr@my-dejanews.com wrote: 

<snip>
 
> We think it is in the best interests of the Ada community and the GNAT project
> to ensure that the public versions of GNAT that are distributed are as free
> as possible from glitches that will cause unsupported users to run into
> trouble. We think that it is better to wait, and perhaps be a little behind
> the latest-and-greatest, rather than deal with glitches.

I wholeheartedly agree.  One of the challenges Ada is facing now is
overcoming a bad reputation earned in the mid 1980s because many compilers
were being released long before they were mature:  full of bugs, poorly
optimized code, slow, ...  We can thank the so-called "Ada mandate" for
this.  I appreciate the desire of ACT to put out a quality product at no
charge for many users--a tremendously positive contrast to the typical vendor
behavior in the 1980s giving Ada a bad reputation.
 
> Now if you are an expert hacker (in the good sense of the word) like Ronald,
> you could probably indeed get by at least some of the problems on your own,
> but we do not distribute the public version for the Ronald Coles of the
> world, but rather for students learning Ada for the first time, a very
> different market.

Since students are key recipients of the public version of GNAT and I am
often involved in teaching Ada in both an educational and an industrial
environment with training operating on a typical academic calendar, it surely
would be helpful if new public releases could come out early August if the
intent is to have them be annual.  Then instructors and system administrators
could be set up and ready to go with the latest public release at the start
of the academic year.  Switching versions in mid-term does not make sense, at
least in my environment.  I know that various events can preclude good
intentions, and the emphasis on quality needs to dominate; also, your paying
customers have far more right than the non-paying academic world to direct
what the contents of the next release should be and when it is needed.  Given
those caveats, it would be nice, and most appreciated by the academic
community, if possible, to aim for early August.
 
> As for our customers releasing the wavefront, it would be unfortunate, since
> it would mean that there were insufficiently tested versions publicly
> floating around. If these versions had significant problems, it could reflect
> negatively on Ada in general, the GNAT project, and free software in general,
> and we think it better to avoid this. Our customers share this viewpoint, and
> understand that it is in the best interests of the GNAT project for them not
> to publish the wavefront versions.

At least your customers seem to have some wits about them, unlike some on
this newsgroup. 
 
> Robert Dewar
> Ada Core Technologies

Thank you for your hard work and support of the Ada community.  Keep it up!

Howard W. LUDWIG

Re: Gnat Free ?

[Jonathan Guthrie]

dewarr@my-dejanews.com wrote:
> We are perfectly aware that there are enthusiasts who would like to get their
> hands on the latest technology, and don't care if it has problems. But that is
> not the public we are addressing in making public versions of GNAT available.

Actually, the biggest pain in the neck with GNAT is the fact that ACT
doesn't update the GCC patch files between public releases. 

A while back I was doing well, cruising through the Lovelace tutorial with
my very own copy of GNAT.  I had figured out how to get GNAT (based on GCC
2.7.1 or some such) to coexist with G77 (a version based on GCC 2.7.2.1)
and all was right with the world.

Then, I needed to run software that could only be built with GCC 2.8.1, so
I installed it.  I spent six weeks (or so) trying to get GNAT to build
with GCC 2.8.1.  I even looked in deja news to see if anyone else had
tried this, run into the same problem, and gotten an explanation for it.
(They had:  Robert Dewar said that "you are starting with the wrong
version of GCC."  Like the old Microsoft joke, that answer is absolutely
true and completely useless.  I judged that to ask for more information
would probably solicit a request that I pay for GNAT.  That would be
reasonable, but not productive.)

Now, I MUST run GCC 2.8.1.  That is not negotiable, but the only thing I
need GNAT for is learning Ada.  I continued to attempt to find the
solution, but the enthusiasm for the search, which had been waning before
I checked DejaNews, was almost completely gone.  Eventually, I concluded
that the only approach that was going to work was to wait for ACT to
release the next version of GNAT.  So I've waited and waited and waited.  
I've been patiently waiting for the release of GNAT 3.11P for six months.
(I've never made a peep about this before now.  I'm only saying this now
because it came up and because the end of the wait is in sight.  
Sometimes I wonder why I continue to wait.  Ada can NOT be a great enough
language to make all this waiting worthwhile.)

While I don't think there's anything wrong with only making a release of
the free compiler every once in a while, (I so rarely update compilers
without being forced to that I'd never notice,) I think that when you base
your compiler on somebody else's, you should accept the burden of making
sure that the available free release of your compiler works with any new
releases of the base compiler soon after the release of the new version of
that compiler.

-- 
Jonathan Guthrie (jguthrie@brokersys.com)
Information Broker Systems   +281-895-8101   http://www.brokersys.com/
12703 Veterans Memorial #106, Houston, TX  77014, USA

We sell Internet access and commercial Web space.  We also are general
network consultants in the greater Houston area.

Re: Gnat Free ?

[dewar]

In article <36280492.55B6D77C@bellsouth.net>,
  The Ludwig Family <ludwigs@bellsouth.net> wrote:
> Since students are key recipients of the public version of GNAT and I am
> often involved in teaching Ada in both an educational and an industrial
> environment with training operating on a typical academic calendar, it surely
> would be helpful if new public releases could come out early August if the
> intent is to have them be annual.  Then instructors and system administrators
> could be set up and ready to go with the latest public release at the start
> of the academic year.  Switching versions in mid-term does not make sense, at
> least in my environment.  I know that various events can preclude good
> intentions, and the emphasis on quality needs to dominate; also, your paying
> customers have far more right than the non-paying academic world to direct
> what the contents of the next release should be and when it is needed.  Given
> those caveats, it would be nice, and most appreciated by the academic
> community, if possible, to aim for early August.

First, thanks for your nice note :-)

Yes, we fully realize this, and our intention was indeed that 3.11p be
released in time for the fall semester. But sometimes the "best laid plans of
mice and men gang aft agley" [a Burns quote I learned in school and never
forgot :-) Note that in this case it was not a matter of paid customers
coming first, since the real issue was getting out 3.11b for customers (3.11p
will follow almost immediately). We do have 3.11b out for some targets, but
not all, and we are working furiously to get both the final 3.11b and 3.11p
releases out for Tri^H^H^HSig-Ada). The delay has been in GDB and other
aspects for NT, which now look to be in very good shape, we are in final
testing stages now.

I know this has been a bit of a frustrating wait, especially for those wanting
to use ASIS with the public version (since there was no public release of
ASIS for 3.10, it just wasn't ready on time). But pretty soon the wait should
be over, and we think you will find that the wait was worth it. GNAT 3.11
technology has a lot of improvements, and several very nice new tools, as well
as a much improved debugger.

Robert Dewar
Ada Core Technologies

P.S. The 3.11p releases will include a version for Redhat 5.0 (thus
eliminating one FAQ from this group), and also a version for Solaris x86.

P.P.S. Another reason for the delay was that we have also been working on
validation issues. As you know, GNAT was the first compiler to achieve full
100% validation with ACVC 2.1 (core and all annexes), on VMS, and we are
planning to extend the list of ACVC 2.1 validations very shortly.

-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own    

Re: Gnat Free ?

[Andi Kleen]

jerry@jvdsys.nextjk.stuyts.nl (Jerry van Dijk) writes:

> Ronald Cole (ronald@forte-intl.com) wrote:
> 
> : Of course, I've always secretly wondered if Dewar would shutdown his
> : business completely if every one of his clients started giving out
> : wavefront releases...
> 
> If I understand you correctly you want ACT to follow current practices
> by prematurely releasing possibly unreliable software ?

If it is clearly marked as such that is OK. The method is called the
"bazaar model" compared to the "cathedral style" ACT is using. It works
very well for lots of projects, for example the EGCS Compiler, the GIMP
or the Linux kernel.



-Andi

Re: Gnat Free ?

[Ronald Cole]

Jonathan Guthrie <jguthrie@weck.brokersys.com> writes:
> Then, I needed to run software that could only be built with GCC 2.8.1, so
> I installed it.  I spent six weeks (or so) trying to get GNAT to build
> with GCC 2.8.1.

GNAT built and ran without incident (for me anyway), without needing any
patches, with egcs-1.0.3c.

> Now, I MUST run GCC 2.8.1.  That is not negotiable, but the only thing I
> need GNAT for is learning Ada.

gcc-2.8.1 has quite a few more bugs than egcs-1.x.  Is there any reason
your software won't build with either egcs-1.0.3c or egcs-1.1b?

A patch is required to build gnat-3.10p with egcs-1.1b.  I can make it
available to anyone who wants it...

-- 
Forte International, P.O. Box 1412, Ridgecrest, CA  93556-1412
Ronald Cole <ronald@forte-intl.com>      Phone: (760) 499-9142
President, CEO                             Fax: (760) 499-9152
My PGP fingerprint: 15 6E C7 91 5F AF 17 C4  24 93 CB 6B EB 38 B5 E5

Re: Gnat Free ?

[Ronald Cole]

jerry@jvdsys.nextjk.stuyts.nl (Jerry van Dijk) writes:
> If I understand you correctly you want ACT to follow current practices
> by prematurely releasing possibly unreliable software ?

You don't understand how the Linux development model has proven
successful, do you?  Dewar's arguments for keeping Ada compiler
development in the "Cathedral" are unconvincing.

-- 
Forte International, P.O. Box 1412, Ridgecrest, CA  93556-1412
Ronald Cole <ronald@forte-intl.com>      Phone: (760) 499-9142
President, CEO                             Fax: (760) 499-9152
My PGP fingerprint: 15 6E C7 91 5F AF 17 C4  24 93 CB 6B EB 38 B5 E5

Re: Gnat Free ?

[Ronald Cole]

dewarr@my-dejanews.com writes:
> We think it is in the best interests of the Ada community and the GNAT project
> to ensure that the public versions of GNAT that are distributed are as free
> as possible from glitches that will cause unsupported users to run into
> trouble. We think that it is better to wait, and perhaps be a little behind
> the latest-and-greatest, rather than deal with glitches.

Just judging from the HPPA optimizer bugs in 3.09 (and the fact that they
went unfixed in 3.10p), you've been having problems meeting your own goal.

> We certainly have never "threatened to consider not doing
> business with any one who starts giving out wavefront". That is
> pure fantasy on the part of Ronald Cole.
> -----------== Posted via Deja News, The Discussion Network ==----------
> http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own    

It's certinly ironic that the post you claim is "pure fantasy" is
archived at dejanews...  Took me all of 5 minutes to find, too.

-- 
Forte International, P.O. Box 1412, Ridgecrest, CA  93556-1412
Ronald Cole <ronald@forte-intl.com>      Phone: (760) 499-9142
President, CEO                             Fax: (760) 499-9152
My PGP fingerprint: 15 6E C7 91 5F AF 17 C4  24 93 CB 6B EB 38 B5 E5

Re: Gnat Free ?

[Brian Rogoff]

On 18 Oct 1998, Jonathan Guthrie wrote:
> dewarr@my-dejanews.com wrote:
> > We are perfectly aware that there are enthusiasts who would like to get their
> > hands on the latest technology, and don't care if it has problems. But that is
> > not the public we are addressing in making public versions of GNAT available.
> 
> Actually, the biggest pain in the neck with GNAT is the fact that ACT
> doesn't update the GCC patch files between public releases. 

This is a legitimate gripe. However, in your case, as someone who is
"just" using GNAT to learn (no disrespect intended!) the solution is to
install a standalone version of GNAT which doesn't interfere with the GCC
you are using to do real work. Disk space is cheap, no?

Still, even though I've done that myself, it isn't an entirely
satisfactory solution, since part of your learning experience is to use 
GNAT for real work. There have been rumblings about a "bazaar" development
model Ada compiler spun off from the next relase of GNAT. That may be 
helpful to you if (when?) it happens, but I suspect that some amount of
stability will be lost. 

> Now, I MUST run GCC 2.8.1.  That is not negotiable, but the only thing I
> need GNAT for is learning Ada.  I continued to attempt to find the
> solution, but the enthusiasm for the search, which had been waning before
> I checked DejaNews, was almost completely gone.  Eventually, I concluded
> that the only approach that was going to work was to wait for ACT to
> release the next version of GNAT.  So I've waited and waited and waited.  
> I've been patiently waiting for the release of GNAT 3.11P for six months.
> (I've never made a peep about this before now.  I'm only saying this now
> because it came up and because the end of the wait is in sight.  
> Sometimes I wonder why I continue to wait.  Ada can NOT be a great enough
> language to make all this waiting worthwhile.)

That's right. No language is worth waiting for when you need to get work
done now. However, if your work is learning, the stand alone install
should serve you just fine. 

> While I don't think there's anything wrong with only making a release of
> the free compiler every once in a while, (I so rarely update compilers
> without being forced to that I'd never notice,) I think that when you base
> your compiler on somebody else's, you should accept the burden of making
> sure that the available free release of your compiler works with any new
> releases of the base compiler soon after the release of the new version of
> that compiler.

I roughly agree, with the caveat that I'd qualify the last sentence to
only include stable new releases of the base compiler, since there have
been quick fixes after new GCC releases, and I'd rather ACT be focused on 
improving GNAT rather than tracking each miniscule patch to GCC. 

-- Brian 

Re: Gnat Free ?

[Brian Rogoff]

On 19 Oct 1998, Jonathan Guthrie wrote:
> Brian Rogoff <bpr@shell5.ba.best.com> wrote:
> > On 18 Oct 1998, Jonathan Guthrie wrote:
> >> Actually, the biggest pain in the neck with GNAT is the fact that ACT
> >> doesn't update the GCC patch files between public releases. 
> 
> > This is a legitimate gripe. However, in your case, as someone who is
> > "just" using GNAT to learn (no disrespect intended!) the solution is to
> > install a standalone version of GNAT which doesn't interfere with the GCC
> > you are using to do real work. Disk space is cheap, no?
> 
> "Standalone version"?  You speak a strange dialect, sir.  What do you mean?
> It's been too many months for me to remember why replacing gnatmake (gnatbuild?
> that's gone with the mists of time, too) with a version that calls gnatgcc
> instead of gcc (the fix I applied to get g77 and gnat to coexist) didn't work
> in this case.  Is that similar to what you're talking about?

My apologies; sometimes I post "from the hip", just like I code :-)

What I meant was to install your version of GNAT in such a way that it
does not interfere with whatever version of gcc you are using to do "real" 
work. When I was in graduate school, I didn't have the option of installing 
GNAT in the normal place, so I had to install it under my home directory
and rely on a little shell hackery to make it easier to use both
gcc-<whatever>, which was our main development tool, and GNAT, which I was
using just to learn Ada. I believe that the configure script for that
ancient version of GNAT supported this option directly. This is the brute
force approach to coexistence, maintaining two entirely separate
compilers, but it worked for learning.

Obviously, this approach fails when you actually want to use GNAT for
something more than learning, and your "work" version of gcc is not the
same as the one used to build GNAT. 

Perhaps this discussion ought to move over to gnat-chat or gnatlist?
If you posted there before, I missed it.

-- Brian

Re: Gnat Free ?

[Jonathan Guthrie]

Finally!  While I appreciate all the people who have made suggestions directed
toward helping me get a copy of this working, what I REALLY want to talk about
is the consequences of basing your compiler on someone else's compiler that is
being actively developed.  For those of you who have made suggestions, I have
already tried some of them, I may try others, later.

Brian Rogoff <bpr@shell5.ba.best.com> wrote:
> On 18 Oct 1998, Jonathan Guthrie wrote:
>> dewarr@my-dejanews.com wrote:
>> > We are perfectly aware that there are enthusiasts who would like to get their
>> > hands on the latest technology, and don't care if it has problems. But that is
>> > not the public we are addressing in making public versions of GNAT available.

>> Actually, the biggest pain in the neck with GNAT is the fact that ACT
>> doesn't update the GCC patch files between public releases. 

> This is a legitimate gripe. However, in your case, as someone who is
> "just" using GNAT to learn (no disrespect intended!) the solution is to
> install a standalone version of GNAT which doesn't interfere with the GCC
> you are using to do real work. Disk space is cheap, no?

"Standalone version"?  You speak a strange dialect, sir.  What do you mean?
It's been too many months for me to remember why replacing gnatmake (gnatbuild?
that's gone with the mists of time, too) with a version that calls gnatgcc
instead of gcc (the fix I applied to get g77 and gnat to coexist) didn't work
in this case.  Is that similar to what you're talking about?

>> While I don't think there's anything wrong with only making a release of
>> the free compiler every once in a while, (I so rarely update compilers
>> without being forced to that I'd never notice,) I think that when you base
>> your compiler on somebody else's, you should accept the burden of making
>> sure that the available free release of your compiler works with any new
>> releases of the base compiler soon after the release of the new version of
>> that compiler.

> I roughly agree, with the caveat that I'd qualify the last sentence to
> only include stable new releases of the base compiler, since there have
> been quick fixes after new GCC releases, and I'd rather ACT be focused on 
> improving GNAT rather than tracking each miniscule patch to GCC. 

When make the choice to use the Gnu C back end, you gain tremendous
advantages.  There is no need to write a code generator or to port that
code generator to new platforms.  Since Gnu C is already popular and has
been ported to a wide variety of platforms, you get easy (well, easier)
porting to those platforms and since there are numbers of people improving
parts of the compiler, including the optimizations and code generation,
you have a realistic hope of getting good code out of that back end.

However, those benefits have a price.  The Gnu C people can easily screw you
up, even if only by accident.  To avoid that, there needs to be coordination
between the you and the Gnu C maintainers.  I would expect that this
coordination, which I have reason to believe IS going on in the case of GNAT,
would result in two phenomena:  First, the number and significance of patches
to be made to the existing Gnu C back end should decrease with time as they
are integrated into the base Gnu C source set.  Second, the number of large
changes to the function calls that constitute the front end/back end interface
should also decrease with time.

The consequences of the decreasing number of patches to the back end are, I
hope, self-evident.  The lack of changes to the calling conventions should mean
that the front end source does not have to be patched in the event of any likely
back end changes.

That means that, ideally, there is NO effort to "port" GNAT to a new version
of Gnu C.  Until ACT reaches that ideal, some kind of judgment has to be made
about how much effort goes in to keeping the distributed source for the front
end up with the changes someone else keeps making to the back end.  I agree
that it probably isn't worthwhile to make a patch file for every release, that
it probably isn't worthwhile to update the GNAT source at all between GNAT
releases because of changes to Gnu C, and that the changes to the Gnu C patch
file would necessarily be released some time after the version of Gnu C which
they patch, but Gnu C V2.8.1 has been out for seven months and (to my
knowledge---I quit looking in May) no patch file for has been made for it.

-- 
Jonathan Guthrie (jguthrie@brokersys.com)
Information Broker Systems   +281-895-8101   http://www.brokersys.com/
12703 Veterans Memorial #106, Houston, TX  77014, USA

We sell Internet access and commercial Web space.  We also are general
network consultants in the greater Houston area.

Re: Gnat Free ?

[Pascal Obry]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 404 bytes --]


dewar@gnat.com a �crit dans le message <708ln4$52b$1@nnrp1.dejanews.com>...
>That's the free software model, and it is working very effectively to
support
>the continued development of GNAT. As those of you who are using 3.11
>technology now know, the 3.11 version is a significant advance on 3.10, and
>you will all get to see this shortly when we release 3.10p

Robert meant 3.11p of course !

Re: Gnat Free ?

[dewar]

In article <Pine.BSF.4.05.9810181759320.11365-100000@shell5.ba.best.com>,
  Brian Rogoff <bpr@shell5.ba.best.com> wrote:

> What I meant was to install your version of GNAT in such a way that it
> does not interfere with whatever version of gcc you are using to do "real"
> work. When I was in graduate school, I didn't have the option of installing
> GNAT in the normal place, so I had to install it under my home directory
> and rely on a little shell hackery to make it easier to use both
> gcc-<whatever>, which was our main development tool, and GNAT, which I was
> using just to learn Ada. I believe that the configure script for that
> ancient version of GNAT supported this option directly. This is the brute
> force approach to coexistence, maintaining two entirely separate
> compilers, but it worked for learning.


This approach actually works quite well for production use if you set things
up properly.

One of the problems with gcc at the moment is that there are
multiple source bases (e.g. Apple (Rhapsody), Wind River (VxWorks),
Softway Systems (Interix), Cygnus (EGCS)) all differ significantly from
the FSF GCC version that is used by GNAT.

There is no easy way to distribute GNAT in a manner that is intimately
compatible with other compilers using all these possible source bases, so
indeed the safest thing if you are using other gcc compilers with a different
source base is to install an entirely separate version of GNAT.

Note that all GCC patches developed for GNAT are immediately incorporated
into the FSF GCC sources (e.g. GCC 2.8 contains all the GCC 2.7 patches
developed for GNAT, and the snapshots contain the patches as they are
developed). This is why for example people have been successfully able
to build GNAT with EGCS (though that requires some additional patches,
since EGCS has developed some incompatible interfaces from GCC), since
EGCS generally picks up the FSF GCC patches.

It is definitely the case that using the binary version of GNAT as distributed
by us is the least likely to get you into to trouble. It does require some
shell script work to integrate with other gcc compilers around, but this
is straightforward, and several of our customers have adopted this approach
successfully.

Incidentally, one thing that is not possible for us is to distribute updated
versions of the gcc patch file that work with a current public release, and
some newer or different version of gcc. That is because we simply don't
develop such files. Once a version of GNAT is released, we maintain it at
the same GCC base level for its entire life time.

Developing such patch files is certainly useful, and if the volunteer
community can do this, as for example they have for GNAT 3.10p + EGCS, this
is certainly most welcome by the user community of the public version. We are
happy to arrange for posting of any such patch files in the contrib
directories at NYU. Note that we don't control this site, but they will put
up anything we give them! Note also in developing such patch files that, as I
mentioned above, the GNAT patches do appear in the GCC source snap shots.

I would suggest that future more detailed discussion of this issue might
better be done on chat@gnat.com where many serious GNAT users participate
who do not have time to participate in comp.lang.ada.

Robert Dewar
Ada Core Technologies

-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own    

Re: Gnat Free ?

[dewar]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 752 bytes --]

In article <70esdr$gkf$1@cf01.edf.fr>,
  "Pascal Obry" <p.obry@der.edf.fr> wrote:
>
> dewar@gnat.com a �crit dans le message <708ln4$52b$1@nnrp1.dejanews.com>...
> >That's the free software model, and it is working very effectively to
> support
> >the continued development of GNAT. As those of you who are using 3.11
> >technology now know, the 3.11 version is a significant advance on 3.10, and
> >you will all get to see this shortly when we release 3.10p
>
> Robert meant 3.11p of course !


Indeed I did, thankyou Pascal for noticing this unfortunate typo!

Robert Dewar
Ada Core Technologies

-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own    

Re: Gnat Free ?

[Andi Kleen]

dewar@gnat.com writes:
> 
> I would suggest that future more detailed discussion of this issue might
> better be done on chat@gnat.com where many serious GNAT users participate
> who do not have time to participate in comp.lang.ada.

BTW, the list would be easier to follow for those not subscribed if the WWW
archive on www.gnat.com was splitted (e.g. into months). With the current 
size of the index page netscape requires >15s just to render the HTML from
the cache, which makes switching between messages and index unpleasant. Also
the page takes a very long time to load.

-Andi (who hopes that 3.11p will be released soon so that he can play with
the 'no runtime system' option, e.g. for linux kernel modules) 

Re: Gnat Free ?

[Ronald Cole]

Ronald Cole <ronald@forte-intl.com> writes:
> dewarr@my-dejanews.com writes:
> > We certainly have never "threatened to consider not doing
> > business with any one who starts giving out wavefront". That is
> > pure fantasy on the part of Ronald Cole.
> 
> It's certinly ironic that the post you claim is "pure fantasy" is
> archived at dejanews...  Took me all of 5 minutes to find, too.

Go to "power search".  Select "English" for the Language,
"comp.lang.ada" as the Forum, "dewar" for the Author, and "May 1 1997"
for the Date.  Then "wavefront" for Power Search and click on "Find".
The first article found says:

  "If people started distributing wavefront versions freely, then w[e]
  would probably reluctantly decide to stop making them available,
  since it would be clear that their distribution was harmful. That
  would be too bad for the cases where they really solve a problem. ...

  "So if we did decide to curtail the distribution of wavefront
  release because of problems with uncontrolled releases, then this
  would be nothing like "refusing to do business with people" or
  anything like that. it would just be a matter of balancing the needs
  of customers in certain situations with other needs, something we
  have to do all the time."

and

  "There is, as Richard Kenner, points out a huge difference between
  illegal and impolite or uncooperative. This distinction seems to be
  lost on Ronald, but I think the rest of the community understands it."

Two points, here:
  - it is certainly *not clear* that wide distribution of unsupported,
    development releases (clearly marked as such) would be harmful.
And,
  - being "impolite or uncooperative" (your words) to a client as a
    result of their exercising their right under the GPL would,
    presumably, be a breach of a support contract if there were no
    other way to meet your guarantee of "immediate problem
    resolution".

-- 
Forte International, P.O. Box 1412, Ridgecrest, CA  93556-1412
Ronald Cole <ronald@forte-intl.com>      Phone: (760) 499-9142
President, CEO                             Fax: (760) 499-9152
My PGP fingerprint: 15 6E C7 91 5F AF 17 C4  24 93 CB 6B EB 38 B5 E5

Re: Gnat Free ?

[Dale Pontius]

In article <70cuat$9bo$1@nnrp1.dejanews.com>,
        dewar@gnat.com writes:
>
> P.S. The 3.11p releases will include a version for Redhat 5.0 (thus
> eliminating one FAQ from this group), and also a version for Solaris x86.
>
I presume by RedHat 5.0 you really mean glibc. But of course RedHat
5.1 has been out for some time now, and its default compiler is
egcs instead of gcc. I doubt you could or should try to match the
RedHat update cycle. I also would imagine that guaranteeing/forcing
the presence of gcc would be no problem. You just might want to
see if any installation notes are in order for RH5.1.

Dale Pontius
(NOT speaking for IBM)

Re: Gnat Free ?

[dewar]

In article <70e04a$7bc$1@news.hal-pc.org>,
  Jonathan Guthrie <jguthrie@weck.brokersys.com> wrote:
> Finally!  While I appreciate all the people who have made
> suggestions directed
> toward helping me get a copy of this working, what I REALLY want
> to talk about
> is the consequences of basing your compiler on someone else's compiler that
> is being actively developed.  For those of you who have made suggestions
> I have already tried some of them, I may try others, later.

Just to be clear, the someone else here is us! Richard Kenner maintains the
version of GCC we work on, and is one of the founders of Ada Core
Technologies. The development of GNAT and GCC are kept very well synchronized
as a result. That is why for example it is the case that all GNAT patches
appear immediately in the GCC snapshots. There are no GCC patches in GNAT
that are not reflected in the GCC sources.

What we do for any given version of GNAT is to baseline on a particular GCC
version, and then apply just those patches that are needed by GNAT. That
results in the patch file that you see. We could get the same effect by basing
the release on a particular set of snapshots, but it is cleaner to key the
release to an established gcc release.

Robert Dewar
Ada Core Technologies

-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own    

Re: Gnat Free ?

[dewar]

> it probably isn't worthwhile to update the GNAT source at all between GNAT
> releases because of changes to Gnu C, and that the changes to the Gnu C patch
> file would necessarily be released some time after the version of Gnu C which
> they patch, but Gnu C V2.8.1 has been out for seven months and (to my
> knowledge---I quit looking in May) no patch file for has been made for it.

Right, to our knowledge, no one has created a patch file for version 3.10 and
GCC 2.8.1. At least at ACT, we have never built version 3.10 with GCC 2.8.1,
because 3.10 technology was baselined at GCC 2.7. Yes, it would be possible
to create a new version of GNAT by combining GNAT 3.10 with GCC 2.8, but this
is work we simply do not do, because we prefer to put our efforts into
developing the subsequent release. We would far rather see 3.11 out sooner,
than make updated versions of 3.10 (for most people the gain in going from
GCC 2.7 to GCC 2.8 with GNAT 3.10 is insigificant).

Note that to a first approximation, no patches should be needed for GNAT 3.10
under GCC 2.8. That is because all GNAT patches are, as I previously noted,
included in GCC immediately. However, there are lots of other patches and
fixes in GCC 2.8 that may introduce regressions etc in GNAT 3.10, and we
simply have never tested this combination.

Basically our strategy is, as I mentioned before, to baseline each new release
of GNAT on the current release of GCC, with an appropriate minimal patch file
to correct any known problems. For example, GNAT version 3.11 is based on GCC
version 2.8 with an appropriate patch file.

Robert Dewar
Ada Core Technologies

-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own    

Re: Gnat Free ?

[dewar]

In article <87vhlhbikr.fsf@yakisoba.forte-intl.com>,
  Ronald Cole <ronald@forte-intl.com> wrote:
> Jonathan Guthrie <jguthrie@weck.brokersys.com> writes:
> > Then, I needed to run software that could only be built with GCC 2.8.1, so
> > I installed it.  I spent six weeks (or so) trying to get GNAT to build
> > with GCC 2.8.1.
>
> GNAT built and ran without incident (for me anyway), without needing any
> patches, with egcs-1.0.3c.
>
> > Now, I MUST run GCC 2.8.1.  That is not negotiable, but the only thing I
> > need GNAT for is learning Ada.
>
> gcc-2.8.1 has quite a few more bugs than egcs-1.x.  Is there any reason
> your software won't build with either egcs-1.0.3c or egcs-1.1b?
>
> A patch is required to build gnat-3.10p with egcs-1.1b.  I can make it
> available to anyone who wants it...


This is very much appreciated. Ronald, if you like we would be happy to
arrange for this patch to be available on the standard GNAT mirror sites.
What will of course be even more interesting is a similar patch for GNAT
3.11. I think this will be fairly easy to adapt from your current patch.

Robert Dewar
Ada Core Technologies

-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own    

Re: Gnat Free ?

[dewar]

In article <87vhlhbikr.fsf@yakisoba.forte-intl.com>,
  Ronald Cole <ronald@forte-intl.com> wrote:
> Jonathan Guthrie <jguthrie@weck.brokersys.com> writes:
> > Then, I needed to run software that could only be built with GCC 2.8.1, so
> > I installed it.  I spent six weeks (or so) trying to get GNAT to build
> > with GCC 2.8.1.
>
> GNAT built and ran without incident (for me anyway), without needing any
> patches, with egcs-1.0.3c.

Note that this reflects the fact that egcs generally picks up the patches
from gcc, and gcc reflects all the patches that have been found necessary for
use by GNAT. Occasionally there is a discussion about a particular patch, and
when this happens we try to make sure that things are resolved in a
consistent manner.

Robert Dewar
Ada Core Technologies

-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own    

Re: Gnat Free ?

[dewarr]

In article <87soglbhwu.fsf@yakisoba.forte-intl.com>,
  Ronald Cole <ronald@forte-intl.com> wrote:
> dewarr@my-dejanews.com writes:
> > We think it is in the best interests of the Ada community and the GNAT
project
> > to ensure that the public versions of GNAT that are distributed are as free
> > as possible from glitches that will cause unsupported users to run into
> > trouble. We think that it is better to wait, and perhaps be a little behind
> > the latest-and-greatest, rather than deal with glitches.
>
> Just judging from the HPPA optimizer bugs in 3.09 (and the fact that they
> went unfixed in 3.10p), you've been having problems meeting your own goal.

Indeed, it is not always easy to make the balace between when to release and
when to delay. We try to get the best balance possible. If we waited till we
were 100% sure that there were no bugs at all, you would never see a new
release! It is pretty difficult to satisfy everyone when it comes to the
public release timing. So far we have been pretty successful in achieving the
right balance, and we are trying to do the same for 3.11p.

As for the optimizer bugs that Ronald refers to, they were not causing trouble
for our customers, so fixing them was not a priority, if in fact they existed.
We do look at bug reports from non-supported users of GNAT, and many problems
have been reported by non-supported users of GNAT, and subsequently fixed, but
of necessity these reports have lower priority than customer problems.

> > We certainly have never "threatened to consider not doing
> > business with any one who starts giving out wavefront". That is
> > pure fantasy on the part of Ronald Cole.
> > -----------== Posted via Deja News, The Discussion Network ==----------
> > http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own
>
> It's certinly ironic that the post you claim is "pure fantasy" is
> archived at dejanews...  Took me all of 5 minutes to find, too.

Nope, no such event is archived anywhere, since it has never happened, not
once, not ever, never! There was a discussion on CLA that pointed out that
it was perfectly within the spirit and the letter of the GPL to ask people not
to distribute particular versions of a GPL'ed product on the basis that the
versions were not yet in appropriate form for public release, and then to
refuse to give them future versions if they did not acceed to reasonable
requests. This is indeed true. However, for RC to extrapolate this theoretical
possibility into a claim of actual action by ACT, is entirely unjustified.
Ronald, you have no evidence of this ever having happened, and it is
inappropriate for you to be making this damaging claim in the absence of such
evidence!

Robert Dewar
Ada Core Technologies

-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own    

Re: Gnat Free ?

[Jerry van Dijk]

Ronald Cole (ronald@forte-intl.com) wrote:

: jerry@jvdsys.nextjk.stuyts.nl (Jerry van Dijk) writes:
: > If I understand you correctly you want ACT to follow current practices
: > by prematurely releasing possibly unreliable software ?

: You don't understand how the Linux development model has proven
: successful, do you?  Dewar's arguments for keeping Ada compiler
: development in the "Cathedral" are unconvincing.

I am familiar with the 'bazaar' vs 'cathedral' styles of development.

For now, ACT has -for reasons they deem sufficient- choosen to follow 
the cathedral style of development. 

Since GNAT is (what nowadays is called) Open Source, there is no reason
why another group could not develop GNAT on a paralel track using bazaar
style development.

In fact, a group of Linux users is trying just that. However, the
appearent lack of progress (waiting for 3.11 ? :-) and AFAIK lack of other
initiatives seem to justify ACT's decision.

However there is no reason why this couldn't change in the future. I have not
seen ACT taking any steps in preventing GNAT development other than
limiting the distribution of their own interim versions.

Jerry.

-- 
-- Jerry van Dijk  | email: jdijk@acm.org
-- Leiden, Holland | member Team-Ada
-- Ada & Win32: http://stad.dsl.nl/~jvandyk

Re: Gnat Free ?

[dewar]

In article <70g2ph$dek$1@mdnews.btv.ibm.com>,
  pontius@btv.MBI.com (Dale Pontius) wrote:
> In article <70cuat$9bo$1@nnrp1.dejanews.com>,
>         dewar@gnat.com writes:
> >
> > P.S. The 3.11p releases will include a version for Redhat 5.0 (thus
> > eliminating one FAQ from this group), and also a version for Solaris x86.
> >
> I presume by RedHat 5.0 you really mean glibc. But of course RedHat
> 5.1 has been out for some time now, and its default compiler is
> egcs instead of gcc. I doubt you could or should try to match the
> RedHat update cycle. I also would imagine that guaranteeing/forcing
> the presence of gcc would be no problem. You just might want to
> see if any installation notes are in order for RH5.1.


In fact as you note, the primary issue is the glibc version, and in fact the
current version of GNAT 3.11 does in fact run under Redhat 5.1, that is what
we have installed. We are not making any attempt at the current time to build
with EGCS (in fact we found that several of our customers using Linux actually
preferred to use GCC 2.8 in any case).

Robert Dewar
Ada Core Technologies

-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own    

Re: Gnat Free ?

[dewarr]

In article <87vhlgflfg.fsf@yakisoba.forte-intl.com>,
  Ronald Cole <ronald@forte-intl.com> wrote:
> Ronald Cole <ronald@forte-intl.com> writes:
> > dewarr@my-dejanews.com writes:
> > > We certainly have never "threatened to consider not doing
> > > business with any one who starts giving out wavefront". That is
> > > pure fantasy on the part of Ronald Cole.
> >
> > It's certinly ironic that the post you claim is "pure fantasy" is
> > archived at dejanews...  Took me all of 5 minutes to find, too.
>
> Go to "power search".  Select "English" for the Language,
> "comp.lang.ada" as the Forum, "dewar" for the Author, and "May 1 1997"
> for the Date.  Then "wavefront" for Power Search and click on "Find".
> The first article found says:
>
>   "If people started distributing wavefront versions freely, then w[e]
>   would probably reluctantly decide to stop making them available,
>   since it would be clear that their distribution was harmful. That
>   would be too bad for the cases where they really solve a problem. ...

Ronald, note the *if* here, a predicate discussing how we might change our
policies if this were to occur. We have made no such decisions, since the
situation has never arisen. Certainly it would not be a good thing for our
customers to have wavefronts floating around freely in our judgment. In fact
we are quite selective in giving out wavefronts, we encourage customers to
pick up wavefronts only if it is absolutely necessary to get around a
particular bug (since it is not a good idea to update to a new version of
the compiler too frequently). At the current time we do not even *ask* our
customers not to distribute wavefronts, let alone "threaten" them if they
do so.

The thing we want to avoid is a proliferation of wavefront versions that are
slightly different, and which cannot, of necessity, be as well tested as
releases. One of the troubles with public releases is that once something is
out there, it tends to stick around (we often get bug reports from people
using amazingly out of date versions of the public release).

Our preference therefore is for there to be only a limited set of official
releases publicly available. We are quite aware that some people would like
a much freer release policy, and we are also aware that a freer release policy
works well for other free software products like Linux. But GNAT is different
for a number of reasons, and we are convinced that our policy is in the best
interests of the Ada community.

>   "So if we did decide to curtail the distribution of wavefront
>   release because of problems with uncontrolled releases, then this
>   would be nothing like "refusing to do business with people" or
>   anything like that. it would just be a matter of balancing the needs
>   of customers in certain situations with other needs, something we
>   have to do all the time."


>   "There is, as Richard Kenner, points out a huge difference between
>   illegal and impolite or uncooperative. This distinction seems to be
>   lost on Ronald, but I think the rest of the community understands it."
>
> Two points, here:
>   - it is certainly *not clear* that wide distribution of unsupported,
>     development releases (clearly marked as such) would be harmful.
> And,
>   - being "impolite or uncooperative" (your words) to a client as a
>     result of their exercising their right under the GPL would,
>     presumably, be a breach of a support contract if there were no
>     other way to meet your guarantee of "immediate problem
>     resolution".

Richard Kenner made absolutely NO mention of being impolite or uncooperative
with a client, he was just discussing what the GPL does and does not permit.

As everyone should know, we consider ourselves bound by both the spirit and
the letter of the GPL, but as Richard Stallman has noted directly to Ronald
Cole, there is nothing about the GPL that discourages asking for people's
cooperation in not distributing premature copies inappropriately.

In fact of course, Ada Core Technologies goes far beyond the requirements
of the GPL in its policies. There is nothing at all in the GPL that requires
distribution of any GPL'ed software. However, ACT feels it is helpful to the
Ada community to make public versions of our technology available, and we are
willing to devote some of our resources to this goal. This policy continues
to be in effect, and we will continue to make periodic public releases of
GNAT and its tools. Note in particular that this includes ASIS and GLADE,
and the 3.11p release will be accompanied by corresponding latest releases of
both these technologies, as well as many other new useful tools.

Robert Dewar
Ada Core Technologies

-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own    

Re: Gnat Free ?

[dennison]

In article <F12Gnt.53@jvdsys.nextjk.stuyts.nl>,
  jerry@jvdsys.nextjk.stuyts.nl (Jerry van Dijk) wrote:
> I am familiar with the 'bazaar' vs 'cathedral' styles of development.

For those of you who, like myself yesterday morning, have no clue what this is
all about, the essay in question is available online at
http://earthspace.net/~esr/writings/cathedral-bazaar/ .


> In fact, a group of Linux users is trying just that. However, the
> appearent lack of progress (waiting for 3.11 ? :-) and AFAIK lack of other
> initiatives seem to justify ACT's decision.

According to the paper, the right *type* of person taking this method would
theoreticaly sweep ACT away in the tide. However, the author has a later paper
explaining why this is *not* likely to ever occur. (
http://earthspace.net/~esr/writings/homesteading/ )


--
T.E.D.

-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own    

Re: Gnat Free ?

[Ronald Cole]

dewar@gnat.com writes:
> This is very much appreciated. Ronald, if you like we would be happy to
> arrange for this patch to be available on the standard GNAT mirror sites.

Unfortunately, my patch won't have had the benefit of being run againt
your test suite.  However, it does pass the compare pass of the build.

> What will of course be even more interesting is a similar patch for GNAT
> 3.11. I think this will be fairly easy to adapt from your current patch.

I just got a copy of 3.11b from a sympathetic agent provocateur (a
real altruistic pervert!).  I'll let you know.

-- 
Forte International, P.O. Box 1412, Ridgecrest, CA  93556-1412
Ronald Cole <ronald@forte-intl.com>      Phone: (760) 499-9142
President, CEO                             Fax: (760) 499-9152
My PGP fingerprint: 15 6E C7 91 5F AF 17 C4  24 93 CB 6B EB 38 B5 E5

Re: Gnat Free ?

[Ronald Cole]

dewarr@my-dejanews.com writes:
> This is indeed true. However, for RC to extrapolate this theoretical
> possibility into a claim of actual action by ACT, is entirely unjustified.
> Ronald, you have no evidence of this ever having happened, and it is
> inappropriate for you to be making this damaging claim in the absence of such
> evidence!

The person who just sent me a copy of 3.11b was *quite* adamant that
their identity never be revealed.  I wonder what he had to fear?
Would you be willing to write an indemnity-like clause into your
"modified" GPL (i.e., something that promises no retaliation (positive
or negative) of any type against anyone who exercised their rights
with respect to *any* release of GNAT under the GPL?

-- 
Forte International, P.O. Box 1412, Ridgecrest, CA  93556-1412
Ronald Cole <ronald@forte-intl.com>      Phone: (760) 499-9142
President, CEO                             Fax: (760) 499-9152
My PGP fingerprint: 15 6E C7 91 5F AF 17 C4  24 93 CB 6B EB 38 B5 E5

Re: Gnat Free ?

[Van Snyder]

I just downloaded gnat-3.10p-i386-linux-bin from ftp.cs.nyu.edu.  I installed
it in /opt/gnat, and sourced env-vals.  It won't link any of the examples
because RedHat's path to some object files
(/usr/lib/gcc-lib/i386-redhat-linux/2.7.2.3/crtbegin.o) is different from what
gnat expects.  I couldn't find a discussion of what environment variable
(if any) to set to point to the C libraries.

Can you send it to me?

Also I've seen mention of a FAQ about using gnat in general and maybe 3.10p
specifically under RedHat Linux.  I looked at
ftp:rtfm.mit.edu/pub/usenet-by-group/comp.lang.ada but it's empty.  Where's
the FAQ list?

-- 
What fraction of Americans believe   |  Van Snyder
Wrestling is real and NASA is fake?  |  vsnyder@math.jpl.nasa.gov

Re: Gnat Free ?

[system]

-- 
Forte International, P.O. Box 1412, Ridgecrest, CA  93556-1412
Ronald Cole <ronald@forte-intl.com>      Phone: (760) 499-9142
President, CEO                             Fax: (760) 499-9152
My PGP fingerprint: 15 6E C7 91 5F AF 17 C4  24 93 CB 6B EB 38 B5 E5


Ronald Cole <ronald@forte-intl.com> writes:
>Ronald Cole <ronald@forte-intl.com> writes:
>> dewarr@my-dejanews.com writes:
>> >Ronald Cole <ronald@forte-intl.com> writes

>> > >Oh yeah?  Just try to talk Dewar out of 3.11b without crossing his
>> > >palm with some cash...  He's even threatened to consider not doing
>> > >business with any one who starts giving out wavefront ("non-p")
>> > >releases

>> > We certainly have never "threatened to consider not doing
>> > business with any one who starts giving out wavefront". That is
>> > pure fantasy on the part of Ronald Cole.
>> 
>> It's certinly ironic that the post you claim is "pure fantasy" is
>> archived at dejanews...  Took me all of 5 minutes to find, too.

Do tell?

>Go to "power search".  Select "English" for the Language,
>"comp.lang.ada" as the Forum, "dewar" for the Author, and "May 1 1997"
>for the Date.  Then "wavefront" for Power Search and click on "Find".
>The first article found says:
>
>  "If people started distributing wavefront versions freely, then w[e]
>  would probably reluctantly decide to stop making them available,
>  since it would be clear that their distribution was harmful. That
>  would be too bad for the cases where they really solve a problem. ...

Not there

>  "So if we did decide to curtail the distribution of wavefront
>  release because of problems with uncontrolled releases, then this
>  would be nothing like "refusing to do business with people" or
>  anything like that. it would just be a matter of balancing the needs
>  of customers in certain situations with other needs, something we
>  have to do all the time."

And not there

In other words Dewar has never "threatened to consider not doing
business with any one who starts giving out wavefront [] releases"

Morphis@physics.niu.edu
Real Women change tires			abuse@uu.net postmaster@uu.net
Real Men change diapers                 security@uu.net

Re: Gnat Free ?

[Bruno BEAUFILS]

dewar@gnat.com writes:
> Note that to a first approximation, no patches should be needed for GNAT 3.10
> under GCC 2.8. That is because all GNAT patches are, as I previously noted,
> included in GCC immediately. However, there are lots of other patches and
> fixes in GCC 2.8 that may introduce regressions etc in GNAT 3.10, and we
> simply have never tested this combination.

I try to compile gnat 3.10 with a gcc 2.8.1 source distribution and a gnat
3.10 binary, and it doesn't work.

I am in charge of installing GNU development tools in my lab, on a lot of
differents plateforms (SunOS 4.x, SunOS 5.5, 5.6, Linux 2.0, IRIX 6.2, 6.3,
etc.) but since I have no way to let gcc 2.8.1 and gnat 3.10 cooperates, and
knowing that gnat is absolutely needed I can't upgrade to newer version of
gcc. :-(

So I am just waiting for gnat 3.11 for a long long time now ...

Do I have a chance to see it before the end of the year ?

-- 
------------------------------------------------------------------------------
Bruno BEAUFILS                                    http://www.lifl.fr/~beaufils
Laboratoire d'Informatique Fondamentale de Lille     
Universite des Sciences et Technologies de Lille     Tel   : +33 3 20 43 69 02
UFR d'IEEA - Bat M3 - Bureau 318                     Fax   : +33 3 20 43 65 66
59655 Villeneuve d'Ascq Cedex - FRANCE               email : beaufils@lifl.fr
------------------------------------------------------------------------------

Re: Gnat Free ?

[dewar]

In article <uyur9vzrsbq.fsf@lifl.fr>,
  Bruno BEAUFILS <beaufils@lifl.fr> wrote:

> I try to compile gnat 3.10 with a gcc 2.8.1 source distribution and a gnat
> 3.10 binary, and it doesn't work.

This is possible to do, but not easy, but there is no reason for you to do it.

> I am in charge of installing GNU development tools in my lab, on a lot of
> differents plateforms (SunOS 4.x, SunOS 5.5, 5.6, Linux 2.0, IRIX 6.2, 6.3,
> etc.) but since I have no way to let gcc 2.8.1 and gnat 3.10 cooperates, and
> knowing that gnat is absolutely needed I can't upgrade to newer version of
> gcc. :-(

Many people are running gnat as distributed by us, and GCC 2.8.1 side by side
without any difficulties. The archives for the chat list can be consulted for
discussions on various ways to do this.

> So I am just waiting for gnat 3.11 for a long long time now ...
>
> Do I have a chance to see it before the end of the year ?

Yes.

Robert Dewar
Ada Core Technologies
> --
> ------------------------------------------------------------------------------
> Bruno BEAUFILS                                    http://www.lifl.fr/~beaufils
> Laboratoire d'Informatique Fondamentale de Lille
> Universite des Sciences et Technologies de Lille     Tel   : +33 3 20 43 69 02
> UFR d'IEEA - Bat M3 - Bureau 318                     Fax   : +33 3 20 43 65 66
> 59655 Villeneuve d'Ascq Cedex - FRANCE               email : beaufils@lifl.fr
> ------------------------------------------------------------------------------
>

-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own    

Re: Gnat Free ?

[Fred J. McCall]

Donald Duck <Donald.Duck@Entenhausen.net> wrote:

:Pascal Obry wrote:
:> Of course it is. You have, since more than 1 year, a 3.10p version available
:> for many platforms !
:
:Many thanks. I've heard that it would be no longer free as there is
:commercial support now.

And a free product with paid support is EXACTLY the GNU model.  If you
want extra support or special maintenance attention, you pay.
Otherwise, you take the fixes as they get released.


--
"Insisting on perfect safety is for people who don't have the balls to 
    live in the real world."   -- Mary Shafer, NASA Dryden
-----------------------------------------------------------------------
fjm@onramp.net -- I don't speak for others and they don't speak for me.