GNAT GPL 2005 Edition is now available

GNAT GPL 2005 Edition is now available

[Jamie Ayre]

AdaCore announces the immediate availability of the GNAT GPL 2005 Edition.

AdaCore is pleased to announce the release of the GNAT GPL 2005 Edition 
to provide Free Software developers, that is developers that distribute 
their work under the GPL (GNU General Public License), the latest and 
most advanced Ada 2005 software development environment.

As many of you know, the Ada programming language is undergoing a 
revision, called Ada 2005. Many of the new features in the revision are 
available in the GNAT GPL 2005 Edition, notably:

New language level features:
      - Abstract interface types to provide multiple inheritance
      - Task, protected and synchronized interfaces
      - Limited-with and Private-with clauses
      - Prevention of accidental overloading when overriding
      - Object.operation notation
      - General use of anonymous access subtypes
      - Limited aggregates
      - Access to constant parameters and null-excluding access
          subtypes
      - Unchecked_Unions for interfacing with C
      - Nested type extensions
      - Support for 16-bit and 32-bit characters

New standard libraries:
      - Container library
      - Complete definition of string subprograms (fixed, bounded,
           unbounded)
      - Directory operations

The GNAT GPL 2005 Edition, which is available free of charge from 
http://libre.adacore.com/, is licensed for Free Software development 
under the terms and conditions of the GNU General Public License (GPL). 
Implementation of the new Ada 2005 features is also available in GNAT 
Pro, which is licensed for all types of software development.

For more information visit the following links:

  * Ada 2005: http://www.adacore.com/ada_2005.php/
  * GNAT GPL 2005 Edition: http://libre.adacore.com/
  * GNAT Pro: http://www.adacore.com/gnatpro_summary.php/

Re: GNAT GPL 2005 Edition is now available

[Jacob Sparre Andersen]

Jamie Ayre wrote:

> AdaCore announces the immediate availability of the GNAT GPL 2005
> Edition.

Unfortunately it appears that I am not - in my professional capacity -
among the intended users (and you have to declare your purpose, before
you get to download the compiler). :-(

I will stick to the GNU Ada compiler supplied by Debian, so I don't
have to use different compilers at work and at home.

Jacob
-- 
Growing older is compulsory. Growing up isn't.

Re: GNAT GPL 2005 Edition is now available

[Ludovic Brenta]

Jacob Sparre Andersen a écrit :
> Jamie Ayre wrote:
>
> > AdaCore announces the immediate availability of the GNAT GPL 2005
> > Edition.
>
> Unfortunately it appears that I am not - in my professional capacity
> - among the intended users (and you have to declare your purpose,
> before you get to download the compiler). :-(
>
> I will stick to the GNU Ada compiler supplied by Debian, so I don't
> have to use different compilers at work and at home.

Out of curiosity, which one of the GNATs supplied by Debian do you
use?  Sarge has gnat (3.15p), gnat-3.3 (3.3.6) and gnat-3.4 (3.4.3).
All of them use the GMGPL for the run-time library.  Etch (currently
in testing and in a state of flux) has gnat (3.15p), gnat-3.4 (3.4.4)
and gnat-4.0 (4.0.1).

I've downloaded GNAT GPL 2005 Edition and will review it as time
permits.  If I execute on my previously published Debian Policy for
Ada, this GPL Edition will become the next "gnat" package in Debian
Etch.  The fact that the run-time library uses the GPL is not a
problem for me, nor is it for Debian (for example, Debian already
supplies Qt which is also GPL).  However, if sufficiently many people
object to this, I will reconsider.

As of now, there are several possible choices for the next default Ada
compiler in Debian:

* GNAT GPL 2005 Edition, with libgnat under GPL (from
  libre.adacore.com)

* gnat-3.4, with libgnat-3.4 under GMGPL (from gcc.gnu.org)

* gnat-4.0, with libgnat-4.0 under GMGPL (from gcc.gnu.org)

* gnat-3.4 from gcc.gnu.org, with patches merged from GNAT GPL 2005
  Edition, retaining the libgnat-3.4 under GMGPL.  This means much
  more work for me, and I invite contributions.  This also means that
  this compiler will be different from compilers in all other
  distributions.

Please take into account that recent versions of ASIS and GLADE are
only available under GPL, not GMGPL.  The same is true for AWS,
GtkAda, XML/Ada etc. but is less of a concern because these libraries
are not tightly coupled with the compiler.

Please cast a vote on this forum.  This is the time to influence the
next version of Debian.

-- 
Ludovic Brenta.

Re: GNAT GPL 2005 Edition is now available

[Martin Dowie]

Ludovic Brenta wrote:
> Please cast a vote on this forum.  This is the time to influence the
> next version of Debian.

Which versions will have ASIS available for them?

-- Martin

Re: GNAT GPL 2005 Edition is now available

[Ludovic Brenta]

Martin Dowie a écrit :
> Ludovic Brenta wrote:
> > Please cast a vote on this forum.  This is the time to influence the
> > next version of Debian.
>
> Which versions will have ASIS available for them?

ASIS will be available for GNAT GPL 2005 Edition and gnat-3.4; note
however that the ASIS for gnat-3.4 may be the one from SourceForge
rather than the one from AdaCore, depending on which works best.
There may be efforts to port it to gnat-4.0, but this cannot be
promised.

GLADE will be available for GNAT GPL 2005 Edition, and *may* be
ported to gnat-3.4.  I don't think I can port it to gnat-4.0, at
least not by myself.

-- 
Ludovic Brenta.

Re: GNAT GPL 2005 Edition is now available

[Brian May]

>>>>> "Ludovic" == Ludovic Brenta <ludovic@ludovic-brenta.org> writes:

    Ludovic> As of now, there are several possible choices for the
    Ludovic> next default Ada compiler in Debian:

    Ludovic> * GNAT GPL 2005 Edition, with libgnat under GPL (from
    Ludovic> libre.adacore.com)

    Ludovic> * gnat-4.0, with libgnat-4.0 under GMGPL (from gcc.gnu.org)

I would suggest one of the above two. I would tend to prefer gnat-4.0
if anything, but I suspect GNAT GPL 2005 might have better Ada2005
support (true/false?).

The licensing issue is not an issue for me. Would it be
possible/practical to develop the library packages in such a way they
can be recompiled against another compiler if desired?

As a wishlist, I would also ask for a Debian package of GNAT GPL 2005
that will install (or at least compile) on sarge... ;-)
-- 
Brian May <bam@snoopy.apana.org.au>

Re: GNAT GPL 2005 Edition is now available

[Ludovic Brenta]

Brian May a écrit :

> >>>>> "Ludovic" == Ludovic Brenta <ludovic@ludovic-brenta.org> writes:
>
>     Ludovic> As of now, there are several possible choices for the
>     Ludovic> next default Ada compiler in Debian:
>
>     Ludovic> * GNAT GPL 2005 Edition, with libgnat under GPL (from
>     Ludovic> libre.adacore.com)
>
>     Ludovic> * gnat-4.0, with libgnat-4.0 under GMGPL (from gcc.gnu.org)
>
> I would suggest one of the above two. I would tend to prefer gnat-4.0
> if anything, but I suspect GNAT GPL 2005 might have better Ada2005
> support (true/false?).

As I said, I will review GNAT GPL 2005 Edition as time permits, and
this
will be one area of investigation.  I anticipate that GNAT GPL 2005
Edition will be on par, or even better than gnat-4.0 WRT Ada 2005.
However, I also anticipate that ASIS and GLADE will be difficult to
port to gnat-4.0, but this needs to be investigated.

> The licensing issue is not an issue for me. Would it be
> possible/practical to develop the library packages in such a way they
> can be recompiled against another compiler if desired?

Well, I wouldn't think this is necessary; instead, you'd want to
compile with GNAT GPL 2005 Edition but link against libgnat-3.4 or
libgnat-4.0 (GMGPL).  I would think you can use -lgnat-x.y to achieve
this, without the need to patch the compiler.

> As a wishlist, I would also ask for a Debian package of GNAT GPL 2005
> that will install (or at least compile) on sarge... ;-)

This should be easy to achieve, *if* I package GNAT GPL 2005 Edition.
This discussion will decide that.

OTOH, I don't see much value in doing this: you can just download
AdaCore's binaries.

-- 
Ludovic Brenta.

Re: GNAT GPL 2005 Edition is now available

[Brian May]

>>>>> "Ludovic" == Ludovic Brenta <ludovic@ludovic-brenta.org> writes:

    Ludovic> Well, I wouldn't think this is necessary; instead, you'd
    Ludovic> want to compile with GNAT GPL 2005 Edition but link
    Ludovic> against libgnat-3.4 or libgnat-4.0 (GMGPL).  I would
    Ludovic> think you can use -lgnat-x.y to achieve this, without the
    Ludovic> need to patch the compiler.

If you can do that, what is the problem? Doesn't this mean you are now
using the GMGPL run time library instead of the GPL run time library?
If you are using the GMGPL run time library, what license issues
exist?

Why is "Package GNAT GPL 2005 but make libgnat-4.0 the default" not an
option?


Also, the relevant part of the GPL, I believe is:

--- cut ---
    b) You must cause any work that you distribute or publish, that in
    whole or in part contains or is derived from the Program or any
    part thereof, to be licensed as a whole at no charge to all third
    parties under the terms of this License.
--- cut ---

and

--- cut ---
These requirements apply to the modified work as a whole.  If
identifiable sections of that work are not derived from the Program,
and can be reasonably considered independent and separate works in
themselves, then this License, and its terms, do not apply to those
sections when you distribute them as separate works.  But when you
distribute the same sections as part of a whole which is a work based
on the Program, the distribution of the whole must be on the terms of
this License, whose permissions for other licensees extend to the
entire whole, and thus to each and every part regardless of who wrote it.

Thus, it is not the intent of this section to claim rights or contest
your rights to work written entirely by you; rather, the intent is to
exercise the right to control the distribution of derivative or
collective works based on the Program.

In addition, mere aggregation of another work not based on the Program
with the Program (or with a work based on the Program) on a volume of
a storage or distribution medium does not bring the other work under
the scope of this License.
--- cut ---

My interpretation (yes, I realize debian-legal may disagree) if a
program A uses library B, this does not make program A a derivative of
the library B. Also see the sentence "Thus, it is not the intent of
this section to claim rights or contest your rights to work written
entirely by you;".

This is particularly the case here, because the program A source
wasn't created from the run-time library in any form, not even the
documentation (if any) for the library - instead the source uses a
public standard known as "Ada 95" or Ada "2005". Hence I don't see how
you can argue that A is a derivative of B.

(for another similar situation, see OpenAFS - OpenAFS was developed
independently of Linux, and as such has a non-GPL open source license,
but apparently some developers see this as non-free, because that
could be seen as incompatible with the GPL license Linux uses; Others
argue that openafs was developed independently, and as such cannot be
a derivative work of Linux, so the GPL requirements don't apply. Also
Linus' interpretation of the GPL in that it still allows closed kernel
modules is seen to be wrong by some developers. I admit I don't
understand all the issues involved here; it gets complicated...)
-- 
Brian May <bam@snoopy.apana.org.au>

Re: GNAT GPL 2005 Edition is now available

[Jacob Sparre Andersen]

Ludovic Brenta wrote:

> Out of curiosity, which one of the GNATs supplied by Debian do you
> use?  Sarge has gnat (3.15p), gnat-3.3 (3.3.6) and gnat-3.4 (3.4.3).
> All of them use the GMGPL for the run-time library.  Etch (currently
> in testing and in a state of flux) has gnat (3.15p), gnat-3.4
> (3.4.4) and gnat-4.0 (4.0.1).

On my Debian/unstable workstation (iBook G4):

ii  gnat           3.15p-13       The GNU Ada 95 compiler

On my Debian/stable server (Soekris net4801):

ii  gnat           3.15p-12       The GNU Ada 95 compiler

> I've downloaded GNAT GPL 2005 Edition and will review it as time
> permits.  If I execute on my previously published Debian Policy for
> Ada, this GPL Edition will become the next "gnat" package in Debian
> Etch.  The fact that the run-time library uses the GPL is not a
> problem for me, nor is it for Debian (for example, Debian already
> supplies Qt which is also GPL).  However, if sufficiently many
> people object to this, I will reconsider.

I want to have some freedom in choosing which license I distribute my
software under.  With "libgnat" under GPL that will not be possible
using Debian, and I will have to find a different platform or a
different language for my future work.

Since it is practically impossible to write an Ada program which
doesn't use "libgnat" when compiled with GNAT, I find that comparing
"libgnat" with Qt is unreasonable.

> As of now, there are several possible choices for the next default
> Ada compiler in Debian:
> 
> * GNAT GPL 2005 Edition, with libgnat under GPL (from
>   libre.adacore.com)

This one is out of the question for me.

> * gnat-3.4, with libgnat-3.4 under GMGPL (from gcc.gnu.org)
> 
> * gnat-4.0, with libgnat-4.0 under GMGPL (from gcc.gnu.org)

I haven't got any specific preference among these two, but I tend to
hope that the newer version is somewhat better.

> * gnat-3.4 from gcc.gnu.org, with patches merged from GNAT GPL 2005
>   Edition, retaining the libgnat-3.4 under GMGPL.  This means much
>   more work for me, and I invite contributions.  This also means that
>   this compiler will be different from compilers in all other
>   distributions.

This sounds like too much work to be practical.  Then I will rather
spend some time learning the ins and outs of the gnat-4.0 and
libgnat-4.0 source code, so I can start contributing.

> Please take into account that recent versions of ASIS and GLADE are
> only available under GPL, not GMGPL.

I hope these recent versions haven't been copied to gcc.gnu.org, since
that clearly would be against GCC policy (as I understand it).

> The same is true for AWS, GtkAda, XML/Ada etc. but is less of a
> concern because these libraries are not tightly coupled with the
> compiler.

Exactly.

> Please cast a vote on this forum.  This is the time to influence the
> next version of Debian.

 1) gnat-4.0
 2) gnat-3.4
 3) gnat-3.4 + patches from GNAT GPL 2005
 -) GNAT GPL 2005

Greetings,

Jacob
-- 
"Any newsgroup where software developers hang out is
 an Emacs newsgroup."

Re: GNAT GPL 2005 Edition is now available

[Samuel Tardieu]

>>>>> "Ludovic" == Ludovic Brenta <ludovic@ludovic-brenta.org> writes:

Ludovic> I've downloaded GNAT GPL 2005 Edition and will review it as
Ludovic> time permits.  If I execute on my previously published Debian
Ludovic> Policy for Ada, this GPL Edition will become the next "gnat"
Ludovic> package in Debian Etch.

Please don't. It would mean that no software using a Free Software
licence other than the GPL (BSD, MIT, Apache, ...) could then be
bundled with Debian if it is written in Ada and compiled with this
compiler.

  Sam
-- 
Samuel Tardieu -- sam@rfc1149.net -- http://www.rfc1149.net/sam

Re: GNAT GPL 2005 Edition is now available

[Stephane Riviere]

> Please don't. It would mean that no software using a Free Software
> licence other than the GPL (BSD, MIT, Apache, ...) could then be
> bundled with Debian if it is written in Ada and compiled with this
> compiler.

I second that.

Jacob choices are mine too.

  1) gnat-4.0
  2) gnat-3.4
  3) gnat-3.4 + patches from GNAT GPL 2005
  -) GNAT GPL 2005

-- 
Stephane Riviere
Oleron Island - France
http://stephane.rochebrune.org
OpenPgp Key <5fd6a1e6> available on the web site above

GtkAda and GPL/GMGPL, was: Re: GNAT GPL 2005 Edition is

[Simon Clubley]

In article <1126773856.876636.265130@g49g2000cwa.googlegroups.com>, "Ludovic Brenta" <ludovic@ludovic-brenta.org> writes:
> 
> Please take into account that recent versions of ASIS and GLADE are
> only available under GPL, not GMGPL.  The same is true for AWS,
> GtkAda, XML/Ada etc. but is less of a concern because these libraries
> are not tightly coupled with the compiler.
> 

Talking only about GtkAda, a quick check of the source using cvsweb shows
that it still has the special exception comments at the top of the modules.
This one taken from gtk-dnd.adb, modified 4 weeks ago:

-----------------------------------------------------------------------
--              GtkAda - Ada95 binding for Gtk+/Gnome                --
--                                                                   --
--                     Copyright (C) 2001-2005                       --
--                         AdaCore                                   --
--                                                                   --
-- This library is free software; you can redistribute it and/or     --
-- modify it under the terms of the GNU General Public               --
-- License as published by the Free Software Foundation; either      --
-- version 2 of the License, or (at your option) any later version.  --
--                                                                   --

[snip]

--                                                                   --
-- As a special exception, if other files instantiate generics from  --
-- this unit, or you link this unit with other files to produce an   --
-- executable, this  unit  does not  by itself cause  the resulting  --
-- executable to be covered by the GNU General Public License. This  --
-- exception does not however invalidate any other reasons why the   --
-- executable file  might be covered by the  GNU Public License.     --
-----------------------------------------------------------------------

Simon.

-- 
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP       
Microsoft: The Standard Oil Company of the 21st century

Re: GtkAda and GPL/GMGPL, was: Re: GNAT GPL 2005 Edition is

[Ludovic Brenta]

Simon Clubley a écrit :
> Talking only about GtkAda, a quick check of the source using cvsweb shows
> that it still has the special exception comments at the top of the modules.
> This one taken from gtk-dnd.adb, modified 4 weeks ago:

Yes, and that is also the case for the GtkAda 2.4.1 which is included
in gtkada-gps-3.0.0-src.tgz.

-- 
Ludovic Brenta.

Re: GNAT GPL 2005 Edition is now available

[Samuel Tardieu]

>>>>> "Ludovic" == Ludovic Brenta <ludovic@ludovic-brenta.org> writes:

Ludovic> Please take into account that recent versions of ASIS and
Ludovic> GLADE are only available under GPL, not GMGPL.

The case of GLADE at least needs to be cleared. To the best of my
knowledge, all of its Ada sources belong to the Free Software
Foundation and none to AdaCore. It is not clear to me whether the FSF
(the only entity allowed to do it) has really changed the licence or
if it is simply a mistake from AdaCore at packaging time.

  Sam
-- 
Samuel Tardieu -- sam@rfc1149.net -- http://www.rfc1149.net/sam

Re: GNAT GPL 2005 Edition is now available

[Ludovic Brenta]

Samuel Tardieu a écrit :

> >>>>> "Ludovic" == Ludovic Brenta <ludovic@ludovic-brenta.org> writes:
>
> Ludovic> Please take into account that recent versions of ASIS and
> Ludovic> GLADE are only available under GPL, not GMGPL.
>
> The case of GLADE at least needs to be cleared. To the best of my
> knowledge, all of its Ada sources belong to the Free Software
> Foundation and none to AdaCore. It is not clear to me whether the FSF
> (the only entity allowed to do it) has really changed the licence or
> if it is simply a mistake from AdaCore at packaging time.

Interesting.  I was saying that because of the file names on
on libre.adacore.com: asis-gpl-2005-src.tgz and glade-gpl-2005-src.tgz.
Has someone had the time to check the actual license in the files?

-- 
Ludovic Brenta.

Re: GNAT GPL 2005 Edition is now available

[Samuel Tardieu]

Ludovic> Please take into account that recent versions of ASIS and
Ludovic> GLADE are only available under GPL, not GMGPL.

Sam>  The case of GLADE at least needs to be cleared. To the best of my
Sam> knowledge, all of its Ada sources belong to the Free Software
Sam> Foundation and none to AdaCore. It is not clear to me whether the
Sam> FSF (the only entity allowed to do it) has really changed the
Sam> licence or if it is simply a mistake from AdaCore at packaging
Sam> time.

Ludovic> Interesting.  I was saying that because of the file names on
Ludovic> on libre.adacore.com: asis-gpl-2005-src.tgz and
Ludovic> glade-gpl-2005-src.tgz.  Has someone had the time to check
Ludovic> the actual license in the files?

I did. That's why I am waiting for this clarification.

  Sam
-- 
Samuel Tardieu -- sam@rfc1149.net -- http://www.rfc1149.net/sam

Re: GNAT GPL 2005 Edition is now available

[Samuel Tardieu]

>>>>> "Sam" == Samuel Tardieu <sam@rfc1149.net> writes:

>>>>> "Ludovic" == Ludovic Brenta <ludovic@ludovic-brenta.org> writes:
Ludovic> Please take into account that recent versions of ASIS and
Ludovic> GLADE are only available under GPL, not GMGPL.

Sam> The case of GLADE at least needs to be cleared. To the best of my
Sam> knowledge, all of its Ada sources belong to the Free Software
Sam> Foundation and none to AdaCore. It is not clear to me whether the
Sam> FSF (the only entity allowed to do it) has really changed the
Sam> licence or if it is simply a mistake from AdaCore at packaging
Sam> time.

I have checked with AdaCore, and it appears that they are entitled to
do this redistribution with a licence change. When an exception is
applied to the GPL, anyone can choose to redistribute a derivative (or
verbatim) version either under the GPL+exception or under the GPL
alone.

This situation is described at
http://www.gnu.org/licenses/gpl-faq.html#GPLIncompatibleLibs:

 "Note that people who make modified versions of ABC are not obligated
  to grant this special exception for their modified versions; it is
  their choice whether to do so. The GNU General Public License gives
  permission to release a modified version without this exception; this
  exception also makes it possible to release a modified version which
  carries forward this exception."

So I stand corrected on this issue, AdaCore is allowed to do such a
release even if they do not own the copyright on the source
files. Anyone would be entitled to do the same thing.

 Sam
-- 
Samuel Tardieu -- sam@rfc1149.net -- http://www.rfc1149.net/sam

Re: GNAT GPL 2005 Edition is now available

[Florian Weimer]

* Samuel Tardieu:

> The case of GLADE at least needs to be cleared. To the best of my
> knowledge, all of its Ada sources belong to the Free Software
> Foundation and none to AdaCore. It is not clear to me whether the FSF
> (the only entity allowed to do it) has really changed the licence or
> if it is simply a mistake from AdaCore at packaging time.

I think it's FSF policy that anyone can remove the special exception
when redistributing the code.  There are variants of the exception
which make this explicit.

Re: GNAT GPL 2005 Edition is now available

[Dr. Adrian Wrigley]

On Thu, 15 Sep 2005 01:44:16 -0700, Ludovic Brenta wrote:

...
> Please take into account that recent versions of ASIS and GLADE are
> only available under GPL, not GMGPL.  The same is true for AWS,
> GtkAda, XML/Ada etc. but is less of a concern because these libraries
> are not tightly coupled with the compiler.
> 
> Please cast a vote on this forum.  This is the time to influence the
> next version of Debian.

The versions of GLADE and GtkAda I use are GMGPL.
GLADE is (C) 1996-2001 Free Software Foundation, Inc
GtkAda is (C) 1998-1999 Briot, Brobecker and Charlet 

I am concerned that tools like this will no longer
be maintained for unrestricted use.

Does the FSF really want to change their code so that
you can't use GLADE in closed source applications?
Perhaps the recent versions are not derived works of the
FSF version?

I oppose Debian moving to any Ada compiler which generates
code that has serious license restrictions.

Doesn't the gcc project prohibit inclusion of sources that place
restrictions on the generated code?  Surely this is no
different to code fragments from any of the gcc back-ends - authors
could easily claim copyright on those sequences and that
compiled code is a "derived work" that they assert rights over.
But this would not be compatible with the aims of gcc.
Why should GNAT be any different?

We seem to be losing coherence rapidly with Ada compilers
and libraries.  ACT should be unifying and strengthening
Ada as a viable language for modern programming, not
forcing code forks, fragmentation and FUD.  If the only
complete, freely licensable tool chain is only available
in obsolete/obsolescent versions, surely their market
will dwindle, not expand?  These latest developments seem
to be some of the worst possible things they could do
to harm Ada's image and uptake.

I think I should have sent the following rant to this thread,
not the other...

<rant>
Now I'm really confused!

If we want to supply Ada programs compiled with GNAT (without source), we
need to beg for a copy of GNAT Pro from someone? Or what?

GNAT Pro users can supply anyone they want with the GNAT Pro tools
licensed under GMGPL and GPL?

Can users of the new GNAT GPL edition merge in old GMGPL
library code before they can supply their binaries to others
under a license of their choice?

This amplifies the kind of Fear/Uncertainty/Doubt that Ada already
suffers from.

Are there any precedents of library licenses being downgraded like
this in other languages?  I can't image gcc users writing in C++
putting up with this kind of change!

Will there be *any* compilers available for Ada 2005 suitable
for developers of free, closed source code?  Surely lack of
suitable (free or inexpensive) compilers for Ada projects
was *exactly* the problem GNAT was developed to solve?
Even open source, GPL-incompatible licenses can't be used
with GNAT GPL, except for internal projects :(

Is ACT serious about enforcing these new restrictions?
Do they have the lawyers ready to harrass transgressors?
It seems like a most unfriendly thing to do...

I think it is a serious trap for people hoping to use
gcc for Ada projects to find they cannot use Ada unless they
switch to a suitable license.

I hope ACT will give a detailed explanation of their rationale
for the change in terms, and explain the *all* options available
to software developers who are caught up by this change.

Does this change really bring ACT more benefit than the
negativity towards them and towards Ada that will result?
</rant>
-- 
Adrian

Re: GNAT GPL 2005 Edition is now available

[Martin Dowie]

Dr. Adrian Wrigley wrote:
[snip]
> Will there be *any* compilers available for Ada 2005 suitable
> for developers of free, closed source code?  Surely lack of
> suitable (free or inexpensive) compilers for Ada projects
> was *exactly* the problem GNAT was developed to solve?
> Even open source, GPL-incompatible licenses can't be used
> with GNAT GPL, except for internal projects :(

Perhaps Randy will step up to the plate and Janus/Ada2005 will surprise 
us all by being a 'modified' GPL open source compiler (+ tools).

GNAT doesn't have to be the only GPL-game in town! :-)

Cheers

-- Martin

Re: GNAT GPL 2005 Edition is now available

[Larry Kilgallen]

In article <dgc4ka$ahn$1@nwrdmz03.dmz.ncs.ea.ibs-infra.bt.com>, Martin Dowie <martin.dowie@btopenworld.com> writes:
> Dr. Adrian Wrigley wrote:
> [snip]
>> Will there be *any* compilers available for Ada 2005 suitable
>> for developers of free, closed source code?  Surely lack of
>> suitable (free or inexpensive) compilers for Ada projects
>> was *exactly* the problem GNAT was developed to solve?
>> Even open source, GPL-incompatible licenses can't be used
>> with GNAT GPL, except for internal projects :(
> 
> Perhaps Randy will step up to the plate and Janus/Ada2005 will surprise 
> us all by being a 'modified' GPL open source compiler (+ tools).
> 
> GNAT doesn't have to be the only GPL-game in town! :-)

There is no reason why Randy would have to do anything GPL to offer
a compiler sufficiently inexpensive to be used on GPL projects.

Re: GNAT GPL 2005 Edition is now available

[Dr. Adrian Wrigley]

On Thu, 15 Sep 2005 13:01:30 +0000, Dr. Adrian Wrigley wrote:

> The versions of GLADE and GtkAda I use are GMGPL.
> GLADE is (C) 1996-2001 Free Software Foundation, Inc
> GtkAda is (C) 1998-1999 Briot, Brobecker and Charlet 

Oh ****!   I've just noticed that the COPYING file in GtkAda
was switched to GPL from LGPL *years* ago, although the
individual source files weren't.

I had been looking at the copyright notices in the
individual files (GMGPL).  But it looks like the
intention since 1999 was that GtkAda and GLADE is
not covered GMGPL/LGPL :(  (was this discussed
here a while back?)

One would hope that publicly available Ada bindings and
corresponding publicly available libraries would use the
license...

Can we supply other people with our (closed) source code
for Gtk applications for them to compile and link with
GtkAda?  The source code is not a derived work of GtkAda,
and the recipient is not distributing it, so there is
no breach of the GPL on the binding.
-- 
Adrian

Re: GNAT GPL 2005 Edition is now available

[Jeff Creem]

Dr. Adrian Wrigley wrote:
> On Thu, 15 Sep 2005 13:01:30 +0000, Dr. Adrian Wrigley wrote:
> 
> 
>>The versions of GLADE and GtkAda I use are GMGPL.
>>GLADE is (C) 1996-2001 Free Software Foundation, Inc
>>GtkAda is (C) 1998-1999 Briot, Brobecker and Charlet 
> 
> 
> Oh ****!   I've just noticed that the COPYING file in GtkAda
> was switched to GPL from LGPL *years* ago, although the
> individual source files weren't.
> 
> I had been looking at the copyright notices in the
> individual files (GMGPL).  But it looks like the
> intention since 1999 was that GtkAda and GLADE is
> not covered GMGPL/LGPL :(  (was this discussed
> here a while back?)
> 
> One would hope that publicly available Ada bindings and
> corresponding publicly available libraries would use the
> license...
> 
> Can we supply other people with our (closed) source code
> for Gtk applications for them to compile and link with
> GtkAda?  The source code is not a derived work of GtkAda,
> and the recipient is not distributing it, so there is
> no breach of the GPL on the binding.

Again no one on these groups is qualified to answer....but I will give 
my unqualified response.

It is not a bad thing that Glade and GtkAda are GPL + Exception instead 
of LGPL. LGPL would be bad for most Ada libraries. If you want to 
distribute binaries only it is very very difficult with Ada Generics or 
C++ templates to meet the requirements of section 6 of the LGPL..Thus 
the GPL+Exception gives you (the developer) more leeway.


Note you could even give other people the binary only for your program 
(just give them source for GtkAda and Gtk) with the exception clause.

Re: GNAT GPL 2005 Edition is now available

[Dr. Adrian Wrigley]

On Thu, 15 Sep 2005 19:26:16 -0400, Jeff Creem wrote:
...
> It is not a bad thing that Glade and GtkAda are GPL + Exception instead 
> of LGPL.

That's the point... GLADE and GtkAda are *not* GPL + Exception,
according to the COPYING file.
-- 
Adrian

Re: GNAT GPL 2005 Edition is now available

[Jeff Creem]

Dr. Adrian Wrigley wrote:
> On Thu, 15 Sep 2005 19:26:16 -0400, Jeff Creem wrote:
> ...
> 
>>It is not a bad thing that Glade and GtkAda are GPL + Exception instead 
>>of LGPL.
> 
> 
> That's the point... GLADE and GtkAda are *not* GPL + Exception,
> according to the COPYING file.

But that is the way GMGPL and all GPL + exception stuff is always marked 
  (when done correctly)

The GPL has right at the beginning:

Everyone is permitted to copy and distribute verbatim copies
of this license document, but changing it is not allowed.


Take a look at

http://www.gnu.org/licenses/license-list.html

Scroll down to the portion where it talks about free licenses that are 
not compatible with the GPL and how to handle them. in particular, look 
at the way they tell you to deal with the Q Public License.

*** Begin reference to GNU License List Text

"Since the QPL is incompatible with the GNU GPL, you cannot take a 
GPL-covered program and QPL-covered program and link them together, no 
matter how.

However, if you have written a program that uses QPL-covered library 
(called FOO), and you want to release your program under the GNU GPL, 
you can easily do that. You can resolve the conflict for your program by 
adding a notice like this to it:

   As a special exception, you have permission to link this program
   with the FOO library and distribute executables, as long as you
   follow the requirements of the GNU GPL in regard to all of the
   software in the executable aside from FOO.

You can do this, legally, if you are the copyright holder for the 
program. Add it in the source files, after the notice that says the 
program is covered by the GNU GPL."

*** End reference

(Hopefully that quote falls under fair use since copying that part of 
the FSF webpage is not a right granted by the webpage copyright )

As you can see, the way the recommend it is exactly the way GtkAda 
handles the similar issue.

...

Re: GNAT GPL 2005 Edition is now available

[Jeffrey Carter]

Jeff Creem wrote:
> 
> It is not a bad thing that Glade and GtkAda are GPL + Exception instead 
> of LGPL. LGPL would be bad for most Ada libraries. If you want to 
> distribute binaries only it is very very difficult with Ada Generics or 
> C++ templates to meet the requirements of section 6 of the LGPL..Thus 
> the GPL+Exception gives you (the developer) more leeway.
> 
> 
> Note you could even give other people the binary only for your program 
> (just give them source for GtkAda and Gtk) with the exception clause.

With a GMGPL version of GtkAda, such as the ones I have, one need not 
give anyone the source to anything. If my program does not otherwise 
fall under the GPL, using the GMGPL GtkAda does not make it fall under 
the GPL. If I choose not to license my program under the GPL, then my 
program is not covered by the GPL. If my program is not covered by the 
GPL, then I am not obliged to distribute any source, even though I 
distribute executables of my program.

-- 
Jeffrey Carter
"Now go away or I shall taunt you a second time."
Monty Python and the Holy Grail
E-mail: jeffrey_r_carter-nr [commercial-at]
         raytheon [period | full stop] com

Re: GNAT GPL 2005 Edition is now available

[Ludovic Brenta]

OK, here are the votes so far.  There are four contenders for the
title of "next Ada compiler for Debian": GNAT GPL 2005 Edition
(gnat-gpl below), gnat-3.4, gnat-4.0 and gnat-3.4+patches backported
from gnat-gpl.

I'm counting two votes when someone expressed strong preference for
one of the contenders, one vote for the second-best, and one negative
vote for a contender that is outright rejected.  I'm not voting myself
because, as I already said, I don't really have a preference one way
or another.

I propose to declare a winner next Tuesday evening (CET).  This gives
time for people who have remained silent to speak up, or for others to
research the issue for themselves.  Also, if I got something wrong in
the matrix below, feel free to amend it.  The more votes we get, the
more the easier the final decision will be to justify.

Voter                  gnat-gpl  gnat-3.4  gnat-4.0  gnat-3.4+patches
Jacob Sparre Andersen     -1        1         2
Brian May                  1                  1
Samuel Tardieu            -1
Stéphane Rivière          -1        1         2
Adrian Wrigley            -1

PS. If you voice your opinion, I'd like to hear about your current
distribution and version of GNAT.  I'd also like to hear your reasons
for using, or not using, Debian.  Just curious.

-- 
Ludovic Brenta.

Re: GNAT GPL 2005 Edition is now available

[Andreas Schwarz]

Ludovic Brenta wrote:
> OK, here are the votes so far.  There are four contenders for the
> title of "next Ada compiler for Debian": GNAT GPL 2005 Edition
> (gnat-gpl below), gnat-3.4, gnat-4.0 and gnat-3.4+patches backported
> from gnat-gpl.

I'm not using the Debian gnat packages, and I have only started using 
Ada a few weeks ago. In my opinion this "GPL fork" makes Ada a lot less 
attractive. I wouldn't have any use for gnat-gpl, and for me Ada has not 
/that/ many advantages that I would pay $$$ for it.

Re: GNAT GPL 2005 Edition is now available

[Georg Bauhaus]

Andreas Schwarz wrote:
 

> I wouldn't have any use for gnat-gpl, and for me Ada has not 
> /that/ many advantages that I would pay $$$ for it.

$$$ ist quite the usual price for commercial use of compilers
+ tools in (really) small business, I think, whatever the language.
GCC etc. being an exceptionally inexpensive option only when you
don't need timely support. (And the GMGPL is still in the latest
sources in GCC.)

The price for compilers, libraries, and tools also doesn't seem
to depend on whether they are delivered with sources or not.

By small business I'm referring to normal-people-sums of money,
a small number of developers, starting at 1, and perhaps fairly
limited project duration. That is, neither Airbus  components,
nor subway controls, nor banking systems, nor mediaphone broadcasting,
nor car electronics, nor assembly line robotics.

GNAT Pro users have indicated a cost of $$$$ or more, not just
$$$. I think that's at least the same price range as Aonix,
Greenhills, or IBM. (insert standard disclaimer)

I'm not aware of Ada toolsets, other than two for Windows
(RR Software, Aonix), that cost < $$$$; does anyone know of any
other? For GNU/Linux?

I think that the prices are not different from what you pay for compilers
and tools for other languges. Last time I have looked, commercial Eiffel
tools started at $$$, Lisp tools at 1$$$, full(!) C++ tools at 1$$$,
Fortran at $$$, and so on.
There are some offerings of toolsets in the range $$-$$$,
but most of them can only be used for "console applications",
unless you build a host of tools, libraries and OS bindings yourself,
or get them elsewhere, with or without sources.

 Digging the archives will probably reveal comments made by/about
ACT, indicating that supporting less than 3 or 5 "seats" per
contract isn't the preferred basis for a support contract. (insert
standard disclaimer) But I'm only speculating here, things
might even change, since obviously free riding isn't possible any
longer by using the GNAT GPL edition, unless someone pays you for
producing software that is GPLed.

In fact, someone is paying ACT for producing GPLed software.
Part of the GNAT library is used by the competition, that's good
for everyone.

Perhaps we can dream of a GNAT Std Edition, on a par with
what you get from Borland, Microsoft, or Intel, with limited support,
but with GMGPL?

Re: GNAT GPL 2005 Edition is now available

[Niklas Holsti]

Georg Bauhaus wrote:

> GNAT Pro users have indicated a cost of $$$$ or more, not just
> $$$.

Last time I asked, GNAT Pro with support was more like mid-range 
$$_$$$.

> Perhaps we can dream of a GNAT Std Edition, on a par with
> what you get from Borland, Microsoft, or Intel, with limited support,
> but with GMGPL?

That would be OK for me. I have hoped for a long time that 
something like that would appear -- a way for small users to pay 
for using GNAT. But of course there are overheads in collecting 
such small sums from a limited number of users.

-- 
Niklas Holsti
Tidorum Ltd
niklas holsti tidorum fi
       .      @       .

Re: GNAT GPL 2005 Edition is now available

[Adrien Plisson]

Georg Bauhaus wrote:
> Andreas Schwarz wrote:
> 
>> I wouldn't have any use for gnat-gpl, and for me Ada has not /that/ 
>> many advantages that I would pay $$$ for it.
> 
> 
> $$$ ist quite the usual price for commercial use of compilers
> + tools in (really) small business, I think, whatever the language.
> GCC etc. being an exceptionally inexpensive option only when you
> don't need timely support. (And the GMGPL is still in the latest
> sources in GCC.)

for C/C++ on Windows, digital mars is also free. it does not come with 
source, it is not GPL, it is not endorsed by the FSF, it is simply 
free (see http://www.digitalmars.com/).

free development tools are not so uncommon...

-- 
rien

Re: GNAT GPL 2005 Edition is now available

[Georg Bauhaus]

Adrien Plisson wrote:

> for C/C++ on Windows, digital mars is also free.

I know. And as I said, it offers great C++ but otherwise very limited
possibilities. E.g., nothing but bare Win32. Debugger and IDDE only
on $$ CD-Version.

I also invite you to read the Digital Mars License Agreement.

BTW, the D frontend by Digitalmars is open source.
From the readme.txt:
"many people have expressed a strong interested
in producing a D compiler with the GNU compiler sources.
This release should enable that."


> it does not come with 
> source, it is not GPL,

The D frontend by Digitalmars is either GPL or Artistic L, at your
option.

> it 
> is not endorsed by the FSF, it is simply free 

And with no investment you get what you have paid for, a good
compiler and nothing else.

The Comeau C++ compiler is also only $$. An indispensible tool,
outstanding error messages and language support, also for $$.
But it is also just a compiler.


> free development tools are not so uncommon...

I don't think so. Development is not just a compiler.

almost free development tools (Was: GNAT GPL 2005 Edition is now available)

[Adrien Plisson]

Georg Bauhaus wrote:
> Adrien Plisson wrote:
> 
>> for C/C++ on Windows, digital mars is also free.
> 
> I know. And as I said, it offers great C++ but otherwise very limited
> possibilities. E.g., nothing but bare Win32. Debugger and IDDE only
> on $$ CD-Version.

bare win32 + DOS 32 + DOS 16 (nice when playing with older computers).

having to pay for the IDDE and the debugger may be restrictive, but 
you can find these tools for free elsewhere, although paying gives 
some support to the developper. i also don't think their IDDE and 
debugger is cutting edge.

> I also invite you to read the Digital Mars License Agreement.

well, i don't write mission critical softwares, and i surely would not 
do that in C++. for my need, their license agreement is no problem.

> BTW, the D frontend by Digitalmars is open source.
>  From the readme.txt:
> "many people have expressed a strong interested
> in producing a D compiler with the GNU compiler sources.
> This release should enable that."

i wasn't aware of this fact. i never used D, nor took the time to read 
more about it. good to know !

>> it does not come with source, it is not GPL,
>  
> The D frontend by Digitalmars is either GPL or Artistic L, at your
> option.

(see above)

>> it is not endorsed by the FSF, it is simply free 
> 
> And with no investment you get what you have paid for, a good
> compiler and nothing else.

that should be sufficient. at least for me.

> The Comeau C++ compiler is also only $$. An indispensible tool,
> outstanding error messages and language support, also for $$.
> But it is also just a compiler.

i heard a lot about it, but never used it. one more cheap compiler in 
the list.

>> free development tools are not so uncommon...
> 
> I don't think so. Development is not just a compiler.

the compiler is just a part of the problem. you can get some nice 
tools other than compilers for free.

editor: jEdit, vim, emacs, notepad (hum, well, when paying a windows 
license)...
IDE: eclipse, GPS...
for simpler tools, if you are a software developper, you should not be 
affraid of rolling your own.

also, note that i will happily pay $$ for a good compiler or any other 
good software that suits my needs. but in no way i can afford a $$$$$ 
software.

-- 
rien

Re: GNAT GPL 2005 Edition is now available

[Martin Krischik]

Georg Bauhaus wrote:

> , unless someone pays you for
> producing software that is GPLed.

Why not? You only have to give the sources to the customer and no one else
and only if the customer actually demands the sources.

Martin
-- 
mailto://krischik@users.sourceforge.net
Ada programming at: http://ada.krischik.com

Re: GNAT GPL 2005 Edition is now available

[Florian Weimer]

* Georg Bauhaus:

> GNAT Pro users have indicated a cost of $$$$ or more, not just
> $$$. I think that's at least the same price range as Aonix,
> Greenhills, or IBM. (insert standard disclaimer)

$$$$ is not a real problem, either, at least in countries which have a
high per-job overhead. 8-)

The problem is that below a certain number of developers, you can't
get support for GNAT from AdaCore.  The other companies around GCC
aren't able to offer support, either.  I don't know how much it costs
to get one of your developers involved in GNAT development.  It would
probably take a couple of months, something you might not be able to
afford.

I still think Ada is a fine language, and GNAT is a great compiler,
but right now, circumstances seem to restrict me to x86 on Linux,
probably uniprocessor only.  This makes me feel uneasy.

Re: GNAT GPL 2005 Edition is now available

[Martin Krischik]

Andreas Schwarz wrote:

> Ludovic Brenta wrote:
>> OK, here are the votes so far.  There are four contenders for the
>> title of "next Ada compiler for Debian": GNAT GPL 2005 Edition
>> (gnat-gpl below), gnat-3.4, gnat-4.0 and gnat-3.4+patches backported
>> from gnat-gpl.
> 
> I'm not using the Debian gnat packages, and I have only started using
> Ada a few weeks ago. In my opinion this "GPL fork" makes Ada a lot less
> attractive. I wouldn't have any use for gnat-gpl, and for me Ada has not
> /that/ many advantages that I would pay $$$ for it.

Where is the problem? The GNAT/GPL does not cost any money.

Martin
-- 
mailto://krischik@users.sourceforge.net
Ada programming at: http://ada.krischik.com

Re: GNAT GPL 2005 Edition is now available

[Andreas Schwarz]

Martin Krischik schrieb:
> Andreas Schwarz wrote:
> 
> 
>>Ludovic Brenta wrote:
>>
>>>OK, here are the votes so far.  There are four contenders for the
>>>title of "next Ada compiler for Debian": GNAT GPL 2005 Edition
>>>(gnat-gpl below), gnat-3.4, gnat-4.0 and gnat-3.4+patches backported
>>>from gnat-gpl.
>>
>>I'm not using the Debian gnat packages, and I have only started using
>>Ada a few weeks ago. In my opinion this "GPL fork" makes Ada a lot less
>>attractive. I wouldn't have any use for gnat-gpl, and for me Ada has not
>>/that/ many advantages that I would pay $$$ for it.
> 
> 
> Where is the problem? The GNAT/GPL does not cost any money.

But it is limited to producing GPL software.

Re: GNAT GPL 2005 Edition is now available

[Simon Clubley]

In article <1126794444.071827.191320@o13g2000cwo.googlegroups.com>, "Ludovic Brenta" <ludovic@ludovic-brenta.org> writes:
> OK, here are the votes so far.  There are four contenders for the
> title of "next Ada compiler for Debian": GNAT GPL 2005 Edition
> (gnat-gpl below), gnat-3.4, gnat-4.0 and gnat-3.4+patches backported
> from gnat-gpl.
> 
> I'm counting two votes when someone expressed strong preference for
> one of the contenders, one vote for the second-best, and one negative
> vote for a contender that is outright rejected.  I'm not voting myself
> because, as I already said, I don't really have a preference one way
> or another.
> 

(Although I don't use Debian, I will vote because I don't want other
distributions to think it's ok for them to include gnat-gpl if Debian
do so.)

A strong negative vote for gnat-gpl. A compiler that restricts what you
can do with it has no place as a primary compiler in a Linux system when
other options are available. If gnat-4.0 is stable enough, use that; if
not, use gnat-3.4.

> PS. If you voice your opinion, I'd like to hear about your current
> distribution and version of GNAT.  I'd also like to hear your reasons
> for using, or not using, Debian.  Just curious.
> 

I'm using GCC-3.4 for native code. For RTEMS based code, I'm using GCC-3.2.3
because that version is the one supported by RTEMS. (BTW, before anyone
points out that a new RTEMS version has just come out which supports
GCC 4.x, I know, but I'm in the middle of a project...).

As for OS, I use older Redhat versions, Fedora and Scientific Linux (which
is Redhat based). I'm about to dump Fedora in favour of Scientific Linux.
I don't have a reason for not using Debian, it's just that I started out
using Redhat based distributions, and haven't had a reason to change.

Simon.

-- 
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP       
Microsoft: The Standard Oil Company of the 21st century

Re: GNAT GPL 2005 Edition is now available

[chris]

Ludovic Brenta wrote:

> PS. If you voice your opinion, I'd like to hear about your current
> distribution and version of GNAT.  I'd also like to hear your reasons
> for using, or not using, Debian.  Just curious.

I have gnat 4.0 on Ubuntu Breezy.  Gnat GPL is useless to me.  Since 
there doesn't seem to be a reasonably priced Ada distribution for solo 
developers, this change is bad news.

Can Adacore change the license in the FSF tree?  If they do that, those 
who want a gmgpl runtime may have to port the current one themselves. 
That would be really bad news.

Re: GNAT GPL 2005 Edition is now available

[Ludovic Brenta]

"Ludovic Brenta" <ludovic@ludovic-brenta.org> writes:
> OK, here are the votes so far.

With the large number of responses, here is an update:

Voter                gnat-gpl  gnat-3.4  gnat-4.0  gnat-3.4+patches
---------------------------------------------------------------------
Adrian Wrigley          -1              
Andreas Schwartz        -1              
Brian May                1                  1
Chris Danx              -1                  1
Jacob Sparre Andersen   -1        1         2
Jeffrey Carter          -1          
Marc A. Criley          -1        1         2
Niklas Holsti           -1
Samuel Tardieu          -1        1         2
Simon Clubley           -1        1         2
Stéphane Rivière        -1        1         2
                        
TOTALS:                 -9        5        12             0

As before, please feel free to add to this matrix or amend your vote
as necessary.  In the mean time, this looks like a landslide, I hope
AdaCore is listening :)

-- 
Ludovic Brenta.

Re: GNAT GPL 2005 Edition is now available

[Georg Bauhaus]

Ludovic Brenta wrote:

> As before, please feel free to add to this matrix or amend your vote
> as necessary.  In the mean time, this looks like a landslide, I hope
> AdaCore is listening :)

Can I extent the Debian GNAT horizon a bit? The Ada part of the 4.0.x
compiler in the FSF tree hasn't changed for some time, and probably
won't change any more. If the C part does change, then this doesn not
seem to be true for the FSF Ada part.
GCC 4.0.x is really a tad old:

- The gcc/ada ChangeLog of the latest snapshot of the 4.0.x distribution
stops on April 24, 2005.

Starting end of April, when GCC 4.0.x Ada changes stopped,
there are 3_225 lines of ChangeLog entries documenting lots
of corrections, and improvements introducing Ada 2005 features.

- There are a number of does-not-work, internal compiler errors,
outdated Ada.Containers, o_o.op(...) notation not working properly,
etc. etc. in 4.0.x

Are we talking about GCC 4.0.x as is, with only some support for
Ada 2005, part of it not working?

If "The next Debian system compiler is going to be...!(bang)", then
does this require that the Ada part is based on sources that
have not changed since 2005-04-24? Is there an ABI issue?

Re: GNAT GPL 2005 Edition is now available

[Georg Bauhaus]

Georg Bauhaus wrote:

> Starting end of April, when GCC 4.0.x Ada changes stopped,
> there are 3_225 lines of ChangeLog entries documenting lots
> of corrections, and improvements introducing Ada 2005 features.

I should add that these changes are present in mainline, i.e.
GCC 4.1.x.

Re: GNAT GPL 2005 Edition is now available

[Florian Weimer]

* Georg Bauhaus:

> Georg Bauhaus wrote:
>
>> Starting end of April, when GCC 4.0.x Ada changes stopped,
>> there are 3_225 lines of ChangeLog entries documenting lots
>> of corrections, and improvements introducing Ada 2005 features.
>
> I should add that these changes are present in mainline, i.e.
> GCC 4.1.x.

Are you sure?  I don't remember a commit implementing abstract
interfaces.

Re: GNAT GPL 2005 Edition is now available

[Georg Bauhaus]

Florian Weimer wrote:

>>I should add that these changes are present in mainline, i.e.
>>GCC 4.1.x.
> 
> 
> Are you sure?  I don't remember a commit implementing abstract
> interfaces.

There are some references to abstract interfaces in ChangLog,
though interface does not seem to be supported yet, in any GNAT.

I hope I can run "make check" again in the near future (there
is a problem with fixincludes and string.h, for some users at
least).

Re: GNAT GPL 2005 Edition is now available

[Pascal Obry]

Georg,

> There are some references to abstract interfaces in ChangLog,
> though interface does not seem to be supported yet, in any GNAT.

Not in any GNAT. GNAT Pro, GNAT GAP and GPL Editions have support for
interfaces. That's at least 3 GNAT :)

Pascal.

-- 

--|------------------------------------------------------
--| Pascal Obry                           Team-Ada Member
--| 45, rue Gabriel Peri - 78114 Magny Les Hameaux FRANCE
--|------------------------------------------------------
--|              http://www.obry.net
--| "The best way to travel is by means of imagination"
--|
--| gpg --keyserver wwwkeys.pgp.net --recv-key C1082595

Re: GNAT GPL 2005 Edition is now available

[Georg Bauhaus]

Pascal Obry wrote:
> Georg,
> 
> 
>>There are some references to abstract interfaces in ChangLog,
>>though interface does not seem to be supported yet, in any GNAT.
> 
> 
> Not in any GNAT. GNAT Pro, GNAT GAP and GPL Editions have support for
> interfaces. That's at least 3 GNAT :)

Yes, sorry, misinformation due to forgetting -gnat05. Grr.

Re: GNAT GPL 2005 Edition is now available

[Tapio Kelloniemi]

Georg Bauhaus <bauhaus@futureapps.de> wrote:
>Ludovic Brenta wrote:
>
>> As before, please feel free to add to this matrix or amend your vote
>> as necessary.  In the mean time, this looks like a landslide, I hope
>> AdaCore is listening :)
>
>Can I extent the Debian GNAT horizon a bit? The Ada part of the 4.0.x
>compiler in the FSF tree hasn't changed for some time, and probably
>won't change any more. If the C part does change, then this doesn not
>seem to be true for the FSF Ada part.
>GCC 4.0.x is really a tad old:

Yes since AdaCore only makes development at the head ov CVS unless
there are critical bugs. Please take a look at this ChangeLog which is
only 10 days old:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ada/ChangeLog?rev=1.683&content-type=text/x-cvsweb-markup

Of course this is not GCC 4.0, but perhaps the newer Ada frontend can
be used in GCC 4.0 with only minor modifications and who knows when
4.1 is released (and how stable it will be...).

Perhaps I'll install the newest bleeding edge GCC tomorrow and see
how it {doesn't work,works}.

-- 
Tapio

Re: GNAT GPL 2005 Edition is now available

[Tapio Kelloniemi]

Tapio Kelloniemi <invalid@localdomain.com> wrote:
>Georg Bauhaus <bauhaus@futureapps.de> wrote:
>>Ludovic Brenta wrote:
>>
>>> As before, please feel free to add to this matrix or amend your vote
>>> as necessary.  In the mean time, this looks like a landslide, I hope
>>> AdaCore is listening :)
>>
>>Can I extent the Debian GNAT horizon a bit? The Ada part of the 4.0.x
>>compiler in the FSF tree hasn't changed for some time, and probably
>>won't change any more. If the C part does change, then this doesn not
>>seem to be true for the FSF Ada part.
>>GCC 4.0.x is really a tad old:
>
>Yes since AdaCore only makes development at the head ov CVS unless
>there are critical bugs.

I also should add that mainline still has special exception clauses.

-- 
Tapio

Re: GNAT GPL 2005 Edition is now available

[Ludovic Brenta]

Yes, AdaCore only work on the main line of development in GCC.  Yes,
this means that very few updates are being made to the Ada part of GCC
4.0.  And yes, it is true that the FSF considers Ada not to be a
release criterion for any release of GCC.

That said, I have a history of backporting fixes from the head of GCC
into "The Debian Ada compiler".  I have backported about 20 such fixes
into gnat 3.15p, and I intend to continue doing that for the next
"Debian Ada compiler", whichever that is.  Of course, I do this only
on my spare time, and because I volunteer to do so.  And, of course,
I am only willing to support *one* branch this way; I have not
backported any fixes into either gnat-3.4 or gnat-4.0 yet for this
reason.

So, if GCC 4.0 should be chosen as the next Ada compiler for Debian,
chances are that I'll be able to backport Ada fixes from 4.1 into it.
I make no promises, but at least I'll try if my time allows.

I am not a fan of "the latest version at all costs".  I prefer to
choose a stable branch such as 3.4 or 4.0, and selectively backport
fixes in order to improve *stability* and *correctness*.

Anyone is free to join the GCC team and backport Ada fixes from HEAD
to the active release branches (currently 3.4 and 4.0).  There are
criteria for which patches are acceptable in a release branch.  Such
patches are automatically picked up by Debian, as Debian uses the
weekly snapshots of the release branches.

Anyone is free to join Debian and contribute patches.  Debian has
many patches applied to its compilers (and kernels, and glibc, and
binutils for that matter).

Debian has chosen GCC 4.0 as its system C and C++ compiler for Etch.
I think it would be perverse to choose GCC 4.1 as the Ada system
compiler.  That is why I did not consider 4.1 as a contender, even
though 4.1 may be released before Etch.

-- 
Ludovic Brenta.

Re: GNAT GPL 2005 Edition is now available

[Tapio Kelloniemi]

"Ludovic Brenta" <ludovic@ludovic-brenta.org> wrote:
>That said, I have a history of backporting fixes from the head of GCC
>into "The Debian Ada compiler".  I have backported about 20 such fixes
>into gnat 3.15p, and I intend to continue doing that for the next
>"Debian Ada compiler", whichever that is.  Of course, I do this only
>on my spare time, and because I volunteer to do so.  And, of course,
>I am only willing to support *one* branch this way; I have not
>backported any fixes into either gnat-3.4 or gnat-4.0 yet for this
>reason.

I would suggest making GNAT GPL one of those "unsupported" branches
(since almost nobody seems to like it). Then dropping gnat-3.4 could
be done and GNAT-4.0 could get fixes/features backported from 4.1, if
someone does that. I don't much like the idea of multiple branches
(in Debian), since all Ada libraries are compiled using what is choosen to
be the main Ada compiler. This leaves those who don't want to use the main
compiler without precompiled binaries.

>So, if GCC 4.0 should be chosen as the next Ada compiler for Debian,
>chances are that I'll be able to backport Ada fixes from 4.1 into it.
>I make no promises, but at least I'll try if my time allows.

In that case I would suggest doing those backports to upstream as well,
since the effort is the same anyway.

>I am not a fan of "the latest version at all costs".  I prefer to
>choose a stable branch such as 3.4 or 4.0, and selectively backport
>fixes in order to improve *stability* and *correctness*.

I suggest this as well and as Debian has already choosen 4.0 as the host
compiler, I encourage choosing 4.0 for Ada as well.

-- 
Tapio

Re: GNAT GPL 2005 Edition is now available

[Florian Weimer]

* Georg Bauhaus:

> - The gcc/ada ChangeLog of the latest snapshot of the 4.0.x distribution
> stops on April 24, 2005.

GCC 4.0 is a release branch.  The Ada part is not really maintained on
the release branches, and the Ada 2005 changes are not really suitable
for a release branch, either.

We could create an Ada-specific GCC 3.4 branch in the FSF tree,
though.  This might also help to improve the licensing situation.
(Without a GIGI rewrite, I strongly feel that GCC 4 is totally out of
reach anyway.)

Re: GNAT GPL 2005 Edition is now available - A Vote and some other comments

[Jeff Creem]

Ludovic Brenta wrote:
> "Ludovic Brenta" <ludovic@ludovic-brenta.org> writes:
> 
>>OK, here are the votes so far.
> 
> 
> With the large number of responses, here is an update:
> 
> Voter                gnat-gpl  gnat-3.4  gnat-4.0  gnat-3.4+patches
> ---------------------------------------------------------------------
> Adrian Wrigley          -1              
> Andreas Schwartz        -1              
> Brian May                1                  1
> Chris Danx              -1                  1
> Jacob Sparre Andersen   -1        1         2
> Jeffrey Carter          -1          
> Marc A. Criley          -1        1         2
> Niklas Holsti           -1
> Samuel Tardieu          -1        1         2
> Simon Clubley           -1        1         2
> Stéphane Rivière        -1        1         2
>                         
> TOTALS:                 -9        5        12             0
> 
> As before, please feel free to add to this matrix or amend your vote
> as necessary.  In the mean time, this looks like a landslide, I hope
> AdaCore is listening :)
> 

I vote against gnat-gpl. Any other the others seem fine and may be based 
on what sort of a quality product can be built.

Note I don't really have that big of a problem with the AdaCore 
approach. It limits the applicability of their free distribution but we 
all seem to crave an AdaCore blessed release which implies we feel they 
add value. They want to encourage people that can pay for support to pay 
for support. People generally are happy with the FSF version of GCC for 
their C,C++ work and don't complain that Wind River charges for their 
blessed version (and offers no free GPL only version or any other 
version for public download).

The biggest problems with the FSF tree are:

1) Ada quality is not considered as part of the release criterea.
2) Only real activity is in head (mostly because the majority of the 
activity is from AdaCore)
3) Lack of (or at least difficult to configure) some of the supporting 
tools such as ASIS.


To answer the other question that went along with this poll.

I currently use WindowsXP with mingw, recently added the gnat-gpl-2005 
from adacore for testing purposes, also use CentOS 4 for my 2 Linux 
boxes with the distributions version of gcc/ada.

As for debian. Many years ago I used debian. I was relatively happy with 
it. At some point when moving to new hardware I looked around again and 
went redhat 7 then redhat 9 on the following upgrade based upon what 
seemed like some more recent libaries for Gtk/GNOME. After a short stint 
at fedora core 1 I settled on CentOS given that many of the bigger 
companies only claim redhat enterprise support (this making Redhat based 
distributions the preferred version at work). So, the logical choice at 
home was then Centos.

Re: GNAT GPL 2005 Edition is now available

[Florian Weimer]

* Ludovic Brenta:

> Voter                gnat-gpl  gnat-3.4  gnat-4.0  gnat-3.4+patches

> TOTALS:                 -9        5        12             0

There's something funny going on here.  I can't believe that so many
people trust the Ada compiler on GCC 4. 8->

Re: GNAT GPL 2005 Edition is now available

[Jacob Sparre Andersen]

Florian Weimer wrote:

> There's something funny going on here.  I can't believe that so many
> people trust the Ada compiler on GCC 4. 8->

I think it is partially based on ignorance.  I did at least try to
make that clear, when I put in my vote.  If somebody can explain me
that gnat-3.4 or gnat-3.15p in practice solves my problems better,
then I'm very open to move my vote.

My priorities are (roughly in order):

 + No license restrictions on compiled programs
 + Ada95
 + Tasking (on Linux/PPC, Linux/IA32-SMP and Linux/AMD64-SMP)
 + Efficient numeric code (on Linux/IA32-SMP and Linux/AMD64-SMP)
 + Ada2005
 + ASIS (on Linux/PPC)

Jacob
-- 
"Experience" is what we have that enables us to recognize a
mistake when we make it again.

Re: GNAT GPL 2005 Edition is now available

[Tapio Kelloniemi]

"Ludovic Brenta" <ludovic@ludovic-brenta.org> wrote:
>OK, here are the votes so far.  There are four contenders for the
>title of "next Ada compiler for Debian": GNAT GPL 2005 Edition
>(gnat-gpl below), gnat-3.4, gnat-4.0 and gnat-3.4+patches backported
>from gnat-gpl.

In my opinion, GNAT GPL has only one problem: it is based on GCC 3.4
which is not what my host system compiler is (4.0.1) and I don't want
to have two GCC's installed, nor do I want to compile some packages
with one and others with another. If I used Debian, I would use the
most recent GCC version (the same which is used for compiling programs
in other languages as well) and hope that it has Ada2005 support.

So:
  2 pts: GCC-4.0 (or preferably 4.1)
  1 pts: GNAT GPL (but ONLY if it HAS support for C, C++, Objective-C and
         Java). If it does not support these languages: 0 pts for all
         the rest, since GCC 4.0 has better Ada2005 support (and is
         otherwise more featureful) than the older versions (obsolescent
         form my point of view).

Actually patching the most recent GCC to get the most recent Ada2005
support would be my choice, but it won't most likely be available for
a while.

>PS. If you voice your opinion, I'd like to hear about your current
>distribution and version of GNAT.  I'd also like to hear your reasons
>for using, or not using, Debian.  Just curious.

I don't use Debian, andmost likely will never use, since I compile all
my installed software from source code (http://www.linuxfromscratch.org/).
If I used a distribution, I would probably use Debian, since it strictly
separates free and non-free software packages from each other and because
it calls itself a GNU/Linux distro.

-- 
Tapio

Re: GNAT GPL 2005 Edition is now available

[xavier]

> PS. If you voice your opinion, I'd like to hear about your current
> distribution and version of GNAT.  I'd also like to hear your reasons
> for using, or not using, Debian.  Just curious.
> 


Hi,

I'm currently using debian with 3.15p-13 for ppc release with some patch 
to have a fully asis + xmlada + aws + glade for power pc architecture
and 3.15p-12 for x86 architecture.

I have recompiled glade (cvs tree) with gnat 4.1 recently without 
problem on x86 architecture but I have probleme with aws reconstruction 
(sources from debian repository).

I was away of compiler consideration for a while and I have problem to 
understand fully the discution actually undergowing, but you can 
consider I vote for a compiler that permit to a full coherent package 
list libaws,xmlada,glade and so on... If for licensing reason the future 
gnat compiler choosen doesn't have all this software available, I'll try 
to help to compile the missing packages both for ppc/x86 architecture.

xavier

Re: GNAT GPL 2005 Edition is now available

[Björn Persson]

Ludovic Brenta wrote:
> Please cast a vote on this forum.  This is the time to influence the
> next version of Debian.

Well, I don't really use Debian, at least not for programming, so it's 
up to you whether you want to count this post as a vote.

I want to encourage people to set their software free, but I also want 
to encourage people to write in Ada. Free software is doing rather well 
these days, but Ada is burdened with an undeserved bad reputation. Of 
course a lot of non-free software is still around, but my impression is 
that Ada is in greater need of support than free software is. Therefore, 
when I have to choose between promoting software freedom or promoting 
Ada, I choose Ada. Consequently I would avoid the GPL-only edition. I 
think it will drive people away from Ada rather than driving them 
towards software freedom.

> I'd also like to hear your reasons for using, or not using, Debian.

I have been interested in Debian and installed it twice to try it out, 
but that's about it.

In case my boss wanted to change OSes on our server farm and wanted a 
really stable and reliable distribution, only one thing would keep me 
from suggesting Debian: I haven't been able to find a way to 
automatically verify, by cryptographic signatures, that downloaded 
packages are genuine. That makes automated updating and installation 
seem a bit too risky for my taste.

On my desktop box at home, stability is a tad less important and I often 
want the new features in recent versions of various programs, so there I 
have a second reason.

-- 
Bj�rn Persson                              PGP key A88682FD
                    omb jor ers @sv ge.
                    r o.b n.p son eri nu

[OT] Reasons for using Debian (Was: GNAT GPL 2005 Edition is now available)

[Jacob Sparre Andersen]

Bj�rn Persson wrote:
> Ludovic Brenta wrote:

> > I'd also like to hear your reasons for using, or not using,
> > Debian.

> In case my boss wanted to change OSes on our server farm and wanted
> a really stable and reliable distribution, only one thing would keep
> me from suggesting Debian: I haven't been able to find a way to
> automatically verify, by cryptographic signatures, that downloaded
> packages are genuine. That makes automated updating and installation
> seem a bit too risky for my taste.

It is my impression that the current Debian/stable (and the newer
versions) uses cryptographic signatures on all packages, and that
Aptitude requires operator interaction for upgrading of unsigned
packages.

> On my desktop box at home, stability is a tad less important and I
> often want the new features in recent versions of various programs,
> so there I have a second reason.

For this purpose Debian/unstable (with "apt-listbugs" installed) is a
pretty good solution - at least on my workstation.

Jacob

PS: I apologise for posting this to the newsgroup, but I couldn't
    remember Bj�rn's e-mail address.
-- 
�Alts� h�r nu, jeg ved snart ikke... Jeg var ikke s� vild
 med sidste gang p� Jorden... Korset, s�mmene, tornene og
 alt det der...�       -- S�nnen (ifgl. M�zi�res & Christin)

Re: [OT] Reasons for using Debian (Was: GNAT GPL 2005 Edition is now available)

[Björn Persson]

I'm moving this to email. If anyone else wants to discuss signatures on 
Debian packages, please send email to the address below. It should be 
readable in a fixed-width font. (Sorry about the trouble but it's the 
spammers' fault.)

-- 
Bj�rn Persson                              PGP key A88682FD
                    omb jor ers @sv ge.
                    r o.b n.p son eri nu

Re: [OT] Reasons for using Debian

[Florian Weimer]

* Jacob Sparre Andersen:

> It is my impression that the current Debian/stable (and the newer
> versions) uses cryptographic signatures on all packages, and that
> Aptitude requires operator interaction for upgrading of unsigned
> packages.

This should read "Debian/unstable", of course.

The current stable version of Debian (code-named "sarge") still does
not very the origin of any software you install.

Re: GNAT GPL 2005 Edition is now available

[Florian Weimer]

* Ludovic Brenta:

> OK, here are the votes so far.  There are four contenders for the
> title of "next Ada compiler for Debian": GNAT GPL 2005 Edition
> (gnat-gpl below), gnat-3.4, gnat-4.0 and gnat-3.4+patches backported
> from gnat-gpl.

How does the new GNAT compiler behave from a technical point of view?
Is it better than the GCC 3.4 one?  Comparable to GNAT 3.15p?  Will
there be regular releases from now on, incorporating bug fixes?

I don't think I'll have answers to those questions by next
Tuesday. 8-(

Re: GNAT GPL 2005 Edition is now available

[Ludovic Brenta]

Florian Weimer <fw@deneb.enyo.de> writes:

> * Ludovic Brenta:
>
>> OK, here are the votes so far.  There are four contenders for the
>> title of "next Ada compiler for Debian": GNAT GPL 2005 Edition
>> (gnat-gpl below), gnat-3.4, gnat-4.0 and gnat-3.4+patches backported
>> from gnat-gpl.
>
> How does the new GNAT compiler behave from a technical point of
> view?  Is it better than the GCC 3.4 one?  Comparable to GNAT 3.15p?
> Will there be regular releases from now on, incorporating bug fixes?
>
> I don't think I'll have answers to those questions by next
> Tuesday. 8-(

Good point.  I have a more new data items from Matthias Klose, the
main GCC maintainer in Debian.

* The target release date for Debian Etch is 18 months after Sarge,
  i.e. December 2006.

* It is likely that Debian Etch will switch again from GCC 4.0 to GCC
  4.1, after 4.1 is released, and if 4.1 proves good enough.  The C++
  ABI will not (should not) change, so the transition should be much
  smoother than the current transition from 3.4 to 4.0 (which is already
  quite smooth).

* If Debian Etch moves to GCC 4.1, then GCC 4.0 will be dropped.  GCC
  3.4 will be kept, because the upstream maintainers of Fortran
  recommend shipping g77 (a Fortran 77 front-end and run-time library),
  which has been replaced in 4.0 by gfortran (Fortran 95), which is
  new and has incompatibilities.

* Whichever compiler is chosen for Etch (be it 4.0 or 4.1) will be
  frozen one month before the rest of Debian.  The plan is to freeze
  Debian for a couple of months before release.  This means that the
  compiler (and binutils and glibc) will probably be frozen in
  September 2006, and the rest of Debian in October 2006.

Nothing is 100% certain at this point, except that there is no hurry
to decide on a definitive Ada compiler for Debian Etch.  However, I
sense that some people are in a hurry to move away from gnat 3.15p.
Am I correct in this?

I encourage people to investigate the relative technical merits of the
contenders.  Most votes so far have ignored this issue altogether,
which is bad (IMHO).  If some people need more time, then I won't
close the vote on tuesday, but OTOH I'd like to reach some kind of
decision and then stick to it.

BTW, how many people here use gnat-3.4 or gnat-4.0 regularly?  Any
success or horror stories you might want to share?

And here is more food for thought:

If gnat-3.4 is chosen, its future availability in Etch is guaranteed,
so I can start a transition of the existing Debian packages soon
(e.g. October-November 2005).  No matter what, this transition will
take time and effort on my part.

If gnat-4.0 is chosen, then I'll probably wait until Matthias and the
Debian GCC maintainers decide on whether or not to move to 4.1, and
only then will I start the transition.  This places the transition in
the May-June 2006 timeframe.

None of the above affects Sarge users; only users of Etch or Sid, the
unstable, bleeding-edge branch of Debian.

-- 
Ludovic Brenta.

Re: GNAT GPL 2005 Edition is now available

[Florian Weimer]

* Ludovic Brenta:

> Good point.  I have a more new data items from Matthias Klose, the
> main GCC maintainer in Debian.

GCC 4.1 makes things worse because GCC optimizers are better and rely
on more data provided the by the language front-end.  This means that
the demands on GIGI with regard to exact output increase even
more. 8-(

> Nothing is 100% certain at this point, except that there is no hurry
> to decide on a definitive Ada compiler for Debian Etch.  However, I
> sense that some people are in a hurry to move away from gnat 3.15p.
> Am I correct in this?

GNAT 3.15p has no AMD64 support, and the tasking run-time is broken on
multiprocessor systems on x86.  (Could be a kernel bug on this
particular machine, I don't have many SMP systems to test, obviously.)

> I encourage people to investigate the relative technical merits of the
> contenders.  Most votes so far have ignored this issue altogether,
> which is bad (IMHO).  If some people need more time, then I won't
> close the vote on tuesday, but OTOH I'd like to reach some kind of
> decision and then stick to it.

Do you have a download link for the "GNAT GPL Edition"?  I find it
strange that I have to register before I can download software
licensed under the GPL.

> BTW, how many people here use gnat-3.4 or gnat-4.0 regularly?  Any
> success or horror stories you might want to share?

GCC 4.0 has a wrong-code bug in enumeration types with holes.  I'm
reluctant to invest in a rewrite of my (quite valid!) Ada code to work
around this bug.  There appear to be some other issues with array
range checks, but these are more subtle, and I haven't investigated
them further.

GCC 3.4 simply ICEs on my (valid) code.  Not investigated further.

> If gnat-3.4 is chosen, its future availability in Etch is guaranteed,
> so I can start a transition of the existing Debian packages soon
> (e.g. October-November 2005).  No matter what, this transition will
> take time and effort on my part.

This will be painful for me because I cannot use your GNAT packages
anymore.  But unless there's a perspective for de-facto support of
GNAT (with regular releases, and new platforms such as AMD64), I
expect that I'll have to rewrite the software anyway. 8-(

I don't really care about the GPL restriction because it is a
non-issue for me.  Look how AdaCore handles the situation; their GPLed
customer releases don't leak on the net, either.

Re: GNAT GPL 2005 Edition is now available

[Ludovic Brenta]

Florian Weimer <fw@deneb.enyo.de> writes:
> GNAT 3.15p has no AMD64 support, and the tasking run-time is broken
> on multiprocessor systems on x86.  (Could be a kernel bug on this
> particular machine, I don't have many SMP systems to test,
> obviously.)

I have a dual Pentium II @300 MHz, where Ada tasking programs utilise
both processors without a problem (Debian Sarge, Linux 2.4.27, gnat
3.15p-12).

> GCC 4.0 has a wrong-code bug in enumeration types with holes.  I'm
> reluctant to invest in a rewrite of my (quite valid!) Ada code to work
> around this bug.  There appear to be some other issues with array
> range checks, but these are more subtle, and I haven't investigated
> them further.

Yes, this bug was reported recently on this newsgroup: ada/23836 on
GCC's bugzilla database.

> GCC 3.4 simply ICEs on my (valid) code.  Not investigated further.

gnat-3.4 still has many of the bugs from gnat 3.15p; solves a few,
adds a few.

> This will be painful for me because I cannot use your GNAT packages
> anymore.

You can use my packages for as long as you use Sarge :)

More seriously, your desire for support for amd64 is one good reason
to move forward.  GNAT GPL 2005 Edition is based on GCC 3.4.4 (with
several patches).  It may or may not ICE on your code like GCC 3.4
does.  I suggest you try it and decide for yourself which one suits
you best.

> But unless there's a perspective for de-facto support of GNAT (with
> regular releases, and new platforms such as AMD64), I expect that
> I'll have to rewrite the software anyway. 8-(

I will try to support whichever version of GNAT is chosen for Debian.
Both 3.4 and 4.0 support amd64, AFAIK.

-- 
Ludovic Brenta.

Re: GNAT GPL 2005 Edition is now available

[Florian Weimer]

* Ludovic Brenta:

> Florian Weimer <fw@deneb.enyo.de> writes:
>> GNAT 3.15p has no AMD64 support, and the tasking run-time is broken
>> on multiprocessor systems on x86.  (Could be a kernel bug on this
>> particular machine, I don't have many SMP systems to test,
>> obviously.)
>
> I have a dual Pentium II @300 MHz, where Ada tasking programs utilise
> both processors without a problem (Debian Sarge, Linux 2.4.27, gnat
> 3.15p-12).

Ah, I saw the problems on a quad-Xeon machine (with hyperthreading),
Linux 2.6.0-test9 (slightly patched, don't ask) and NTPL.  The latter
is probably the culprit.

>> This will be painful for me because I cannot use your GNAT packages
>> anymore.
>
> You can use my packages for as long as you use Sarge :)

I use unstable because I feel I need security support. 8-)

> More seriously, your desire for support for amd64 is one good reason
> to move forward. 

To some extent, yes.  But if AdaCore doesn't manage to base a stable
compiler on current 4.x releases (which is a real possibility we
should consider), 3.4 won't buy us that much, especially if AdaCore
won't make any further GNAT GPL releases.

AMD64 shows that even in today's marketplace, it's possible to
introduce a new instruction set and gain at least some relevance.
This is quite surprising.  Otherwise, I would have thought that x86
and SPARC support would have sufficed for years to come (with problems
in the area of tasking, sure, but this is a different issue).

> GNAT GPL 2005 Edition is based on GCC 3.4.4 (with
> several patches).  It may or may not ICE on your code like GCC 3.4
> does.  I suggest you try it and decide for yourself which one suits
> you best.

Okay, downloaded it and ran it through my testsuite.  It passes, on
x86 at least.  However, there are two quite unsettling warnings: one
about an initialized variable of type Exception_Occurrence (the RM
guarantees that objects of this type are initialized to
Null_Occurrence), and a claim that a certain expression is always
false, which doesn't seem to match reality.  (Both issues are also
present in the GCC 4 codebase.)

There's also a new restriction on pragma Inline_Always: it can't be
applied to dispatching subprograms.  This doesn't make much sense
because often, the tag is known at compile time, so cross-compilation
unit inlining would still be possible.

>> But unless there's a perspective for de-facto support of GNAT (with
>> regular releases, and new platforms such as AMD64), I expect that
>> I'll have to rewrite the software anyway. 8-(
>
> I will try to support whichever version of GNAT is chosen for
> Debian.

And I'm grateful for your work. 8-)

Re: GNAT GPL 2005 Edition is now available

[Ludovic Brenta]

Florian Weimer a écrit :

[tasking on all processors on SMP systems]
> Ah, I saw the problems on a quad-Xeon machine (with hyperthreading),
> Linux 2.6.0-test9 (slightly patched, don't ask) and NTPL.  The latter
> is probably the culprit.

I don't think NPTL is the culprit, as I backported support for it into
gnat 3.15p-10 from gnat-3.4.  And I use native threads (as opposed to
FSU threads) in gnat.  It could be something stupid like a
uniprocessor kernel?  Don't laugh, I know from experience that it
happens :)

Maybe I should try it again on a 2.6 kernel on an SMP system.  I think
I did that a while ago, and it worked.

> * Ludovic Brenta:
>> GNAT GPL 2005 Edition is based on GCC 3.4.4 (with several patches).
>> It may or may not ICE on your code like GCC 3.4 does.  I suggest
>> you try it and decide for yourself which one suits you best.
>
> Okay, downloaded it and ran it through my testsuite.  It passes, on
> x86 at least.  However, there are two quite unsettling warnings: one
> about an initialized variable of type Exception_Occurrence (the RM
> guarantees that objects of this type are initialized to
> Null_Occurrence), and a claim that a certain expression is always
> false, which doesn't seem to match reality.  (Both issues are also
> present in the GCC 4 codebase.)

Could you now summarise your results?  I'm curious to know how the
four compare: gnat, gnat-3.4, gnat-4.0 and GNAT GPL.  If you could
post a "quality assessment matrix" with i386 and amd64, that would be
very helpful for everyone here.

> There's also a new restriction on pragma Inline_Always: it can't be
> applied to dispatching subprograms.  This doesn't make much sense
> because often, the tag is known at compile time, so cross-compilation
> unit inlining would still be possible.

Inline_Always was nonstandard, but if I remember correctly it is being
introduced into Ada 200y.  In any case, you cannot rely on it in Ada
95 programs.  But you are correct that inlining non-dispatching calls
should be feasible.

-- 
Ludovic Brenta.

Re: GNAT GPL 2005 Edition is now available

[Martin Krischik]

Florian Weimer wrote:

> Do you have a download link for the "GNAT GPL Edition"?  I find it
> strange that I have to register before I can download software
> licensed under the GPL.

They use the same AWS based software as used for the GNAT-Tracker. This
software used a dynamic packaging system. You select say 10 packages to
download from the list of packages available and the download manager packs
them up for you in one big tar file.

To the original question: No - because there is no single download link -
every user gets it own.

Martin
-- 
mailto://krischik@users.sourceforge.net
Ada programming at: http://ada.krischik.com

Re: GNAT GPL 2005 Edition is now available

[Jacob Sparre Andersen]

Ludovic Brenta wrote:

> Nothing is 100% certain at this point, except that there is no hurry
> to decide on a definitive Ada compiler for Debian Etch.  However, I
> sense that some people are in a hurry to move away from gnat 3.15p.
> Am I correct in this?

Not really.  I miss ASIS and have the impression that tasking doesn't
work quite as it should on my Debian/PPC workstation, but things do
basically seem to work.

> BTW, how many people here use gnat-3.4 or gnat-4.0 regularly?

I normally don't use either, but I could give them a shot with some
critical (for my work) code, to see how they perform.

> If gnat-3.4 is chosen, its future availability in Etch is
> guaranteed, so I can start a transition of the existing Debian
> packages soon (e.g. October-November 2005).  No matter what, this
> transition will take time and effort on my part.
>
> If gnat-4.0 is chosen, then I'll probably wait until Matthias and
> the Debian GCC maintainers decide on whether or not to move to 4.1,
> and only then will I start the transition.  This places the
> transition in the May-June 2006 timeframe.

It sounds like it will make life easier for you (and those who get
around to help you), if we go for gnat-3.4 rather than gnat-4.0.  If
everything else balances, this should also be taken into account.

Jacob
-- 
�When Roman engineers built a bridge, they had to stand under it while
 the first legion marched across. If programmers today worked under
 similar ground rules, they might well find themselves getting much more
 interested in Ada!�                                     -- Robert Dewar

Re: GNAT GPL 2005 Edition is now available

[Ludovic Brenta]

Jacob Sparre Andersen <sparre@nbi.dk> writes:

> Ludovic Brenta wrote:
>
>> Nothing is 100% certain at this point, except that there is no hurry
>> to decide on a definitive Ada compiler for Debian Etch.  However, I
>> sense that some people are in a hurry to move away from gnat 3.15p.
>> Am I correct in this?
>
> Not really.  I miss ASIS and have the impression that tasking doesn't
> work quite as it should on my Debian/PPC workstation, but things do
> basically seem to work.

Good news thanks to Xavier Grave: ASIS 3.15p is now available on
Debian Etch on powerpc, too!  See
http://packages.qa.debian.org/a/asis/news/1.html

>> BTW, how many people here use gnat-3.4 or gnat-4.0 regularly?
>
> I normally don't use either, but I could give them a shot with some
> critical (for my work) code, to see how they perform.

Excellent.

>> If gnat-3.4 is chosen, its future availability in Etch is
>> guaranteed, so I can start a transition of the existing Debian
>> packages soon (e.g. October-November 2005).  No matter what, this
>> transition will take time and effort on my part.
>>
>> If gnat-4.0 is chosen, then I'll probably wait until Matthias and
>> the Debian GCC maintainers decide on whether or not to move to 4.1,
>> and only then will I start the transition.  This places the
>> transition in the May-June 2006 timeframe.
>
> It sounds like it will make life easier for you (and those who get
> around to help you), if we go for gnat-3.4 rather than gnat-4.0.  If
> everything else balances, this should also be taken into account.

Correct.

-- 
Ludovic Brenta.

Re: GNAT GPL 2005 Edition is now available

[Jacob Sparre Andersen]

Ludovic Brenta <ludovic@ludovic-brenta.org> writes:

> Good news thanks to Xavier Grave: ASIS 3.15p is now available on
> Debian Etch on powerpc, too!  See
> http://packages.qa.debian.org/a/asis/news/1.html

Thanks!  I'm installing this now.

Jacob
-- 
�When in Rome; burn it�                        -- Diziet Sma

Re: GNAT GPL 2005 Edition is now available

[Niklas Holsti]

Ludovic Brenta wrote:
>  ...
> As of now, there are several possible choices for the next default Ada
> compiler in Debian:
> 
> * GNAT GPL 2005 Edition, with libgnat under GPL (from
>   libre.adacore.com)
> 
> * gnat-3.4, with libgnat-3.4 under GMGPL (from gcc.gnu.org)
> 
> * gnat-4.0, with libgnat-4.0 under GMGPL (from gcc.gnu.org)
> 
> * gnat-3.4 from gcc.gnu.org, with patches merged from GNAT GPL 2005
>   Edition, retaining the libgnat-3.4 under GMGPL.  This means much
>   more work for me, and I invite contributions.  This also means that
>   this compiler will be different from compilers in all other
>   distributions.
> ...
> Please cast a vote on this forum.  This is the time to influence the
> next version of Debian.

I use GNAT 3.15p (20020523) on Debian "sarge" to create what is 
planned to be a non-GPL commercial program. Therefore, the GMGPL 
is important to me (also on Microsoft systems, of course).

As I have not experimented with the post-3.15 GNAT versions, I 
have no favourite among 3.4 or 4.0, but vote for anything with the 
GMGPL rather than GNAT GPL 2005. My application uses GtkAda but 
does not use ASIS nor GLADE nor any of the other Ada libraries 
mentioned in this thread.

Ludovic Brenta wrote:
 >
 > PS. ... I'd also like to
 > hear your reasons for using, or not using, Debian.

I chose Debian mainly because it seemed the purest "Linux-spirit" 
distribution. It has worked well for me and I feel no need to change.

-- 
Niklas Holsti
Tidorum Ltd
niklas holsti tidorum fi
       .      @       .

Re: GNAT GPL 2005 Edition is now available

[Jeffrey Carter]

Ludovic Brenta wrote:
> 
> I've downloaded GNAT GPL 2005 Edition and will review it as time
> permits.  If I execute on my previously published Debian Policy for
> Ada, this GPL Edition will become the next "gnat" package in Debian
> Etch.  The fact that the run-time library uses the GPL is not a
> problem for me, nor is it for Debian (for example, Debian already
> supplies Qt which is also GPL).  However, if sufficiently many people
> object to this, I will reconsider.

Is it Debian's policy to provide tools that prevent the creation of 
non-GPL SW? Even libre SW covered by another license, such as the BSD 
license?

> Please take into account that recent versions of ASIS and GLADE are
> only available under GPL, not GMGPL.  The same is true for AWS,
> GtkAda, XML/Ada etc. but is less of a concern because these libraries
> are not tightly coupled with the compiler.
> 
> Please cast a vote on this forum.  This is the time to influence the
> next version of Debian.

I have no problem with programs being released under the GPL. I have 
more difficulty with the output of those programs having to be covered 
by the GPL. How many users of GNU-Emacs would be happy if every output 
from it were covered by the GPL?

What AdaCore has done is force a fork in the GNAT sources. The GCC/FSF 
version currently allows the creation of non-GPL programs. If they 
change that for one language, they will certainly have to change it for 
all. Luckily, people can and do have their own copies of the sources 
with the exceptions, so there should always be versions that allow the 
creation of non-GPL programs.

My vote is for a version, including GLADE, that allows the creation of 
non-GPL programs. Libraries such as AWS, GtkAda, and so on, should also 
have an exception if possible.

-- 
Jeffrey Carter
"Now go away or I shall taunt you a second time."
Monty Python and the Holy Grail
E-mail: jeffrey_r_carter-nr [commercial-at]
         raytheon [period | full stop] com

Re: GNAT GPL 2005 Edition is now available

[Georg Bauhaus]

Jeffrey Carter wrote:

> What AdaCore has done is force a fork in the GNAT sources. The GCC/FSF 
> version currently allows the creation of non-GPL programs. If they 
> change that for one language, they will certainly have to change it for 
> all. Luckily, people can and do have their own copies of the sources 
> with the exceptions, so there should always be versions that allow the 
> creation of non-GPL programs.

Some companies have invested in GCC, not all of them producing
GPLed software. I find it hard to imagine all of them switching
compilers and libraries because at the same time they decide to
change GCC licensing.

Re: GNAT GPL 2005 Edition is now available

[Ludovic Brenta]

Jeffrey Carter <spam@spam.com> writes:
> Is it Debian's policy to provide tools that prevent the creation of
> non-GPL SW? Even libre SW covered by another license, such as the
> BSD license?

No; it is Debian policy to respect upstream authors' wishes.  It is up
to the users to decide for themselves whether or not the license suits
them.  Every package in Debian is required to have the license in the
file /usr/share/doc/<package>/copyright, so that users know where to
look.

It is also Debian policy not to ship non-free software; GPL software
is certainly free.

However it is also Debian policy to listen to our users; that's what
I'm doing right now.

> My vote is for a version, including GLADE, that allows the creation
> of non-GPL programs. Libraries such as AWS, GtkAda, and so on,
> should also have an exception if possible.

OK, I'll count that as (another) -1 for gnat-gpl, and no other votes.

-- 
Ludovic Brenta.

Re: GNAT GPL 2005 Edition is now available

[David Trudgett]

"Ludovic Brenta" <ludovic@ludovic-brenta.org> writes:

> I've downloaded GNAT GPL 2005 Edition and will review it as time
> permits.  If I execute on my previously published Debian Policy for
> Ada, this GPL Edition will become the next "gnat" package in Debian
> Etch.  The fact that the run-time library uses the GPL is not a
> problem for me, nor is it for Debian (for example, Debian already
> supplies Qt which is also GPL).  However, if sufficiently many people
> object to this, I will reconsider.

Hi Ludovic!

Please put me down as against GNAT GPL being used as the new Debian
compiler. As far as I'm concerned, it is *less* free, not more free,
notwithstanding its GPL status. I'm not a fan of closed source,
proprietary software (obviously, if you look at my web page), but I am
also not a fan of taking away people's freedoms. At the moment a GNAT
developer on the Debian platform has complete freedom to choose how to
license his/her software to meet his own special requirements. If that
means supplying closed source to a few friends (for example), then
it's no one else's business to question why. Giving the developer this
freedom in no way jeopardizes the Free status of the
compiler/run-time. Taking away the freedom has no benefit (except
corporate profit?).

I'm not in a position to make an informed vote about which of the
other choices should be used. I'll trust that to your technical
judgement. Neither ASIS nor GLADE are critical for me, I think, though
both would be nice: I'd like to experiment with GLADE at some stage;
and ASIS is required for gnatelim, isn't it? which would be nice to have.

You were also interested in why people use Debian. I have a few reasons:

   o Emphasis on Free Software. Other distribution mix in non-Free but
     gratis software.

   o Community developed and supported. I was a Red Hat user before
     they turned their back on the Free Software community. (Actually,
     I still have a Red Hat installed, but am about to switch it to
     Debian.)

   o Great support for software development. Ada is one example
     (thanks to you, Ludovic). Common Lisp is another.

   o Don't have to re-install every six months when a new release
     comes out.

   o apt-get! and friends :-)



> Please cast a vote on this forum.  This is the time to influence the
> next version of Debian.

Please don't use GNAT GPL 2005! Maybe AdaCore will get the message.



Cheers,

David



-- 

David Trudgett
http://www.zeta.org.au/~wpower/

We come here upon what, in a large proportion of cases, forms the
source of the grossest errors of mankind. Men on a lower level of
understanding, when brought into contact with phenomena of a higher
order, instead of making efforts to understand them, to raise
themselves up to the point of view from which they must look at the
subject, judge it from their lower standpoint, and the less they
understand what they are talking about, the more confidently and
unhesitatingly they pass judgment on it.

    -- Leo Tolstoy, "The Kingdom of God is Within You"

Re: GNAT GPL 2005 Edition is now available

[Ludovic Brenta]

Ludovic Brenta a écrit :
> Please cast a vote on this forum.  This is the time to influence the
> next version of Debian.

I have received the following reply from someone who wishes to
remain anonymous.

If you look in the src directory of the GNAT GPL distribution, there is
a
patch file of more than 700kB. One can assume that these patches mostly
increase the compiler's quality and stability. Thus, my vote goes to
the
GNAT GPL 2005 version.

As there are so many strong votes against that version, I probably
can't
change anything. My second vote than would be gcc-3.4 strongly over
4.0. I
had seen a lot of regressions in the 4.0 prereleases which stopped me
from
further using it. Either 3.4 (strongly prefered) or the upcoming 4.1.

Voter   gnat-gpl gnat-3.4 gnat-4.0 gnat-3.4+patches
---------------------------------------------------------------------
Unnamed    +2        +1       -1          0

-- 
Ludovic Brenta, on behalf of an anonymous contributor.

Re: GNAT GPL 2005 Edition is now available

[Ludovic Brenta]

"Ludovic Brenta" <ludovic@ludovic-brenta.org> writes:
> Please cast a vote on this forum.  This is the time to influence the
> next version of Debian.

Here is another (and extensive) contribution from an anonymous
contributor, posted on their behalf.  Here goes:

From the reactions in the comp.lang.ada forum it is clear that the
positive news about the much anticipated "official" public GNAT
distribution is completely drowned by the strong negative feelings
triggered by the limitation that it can only be used to develop GPL
software.

AdaCore obviously is fully entitled to distribute the GNAT GPL 2005
Edition as they did, but this license change is unfortunate for the
promotion of Ada.

GNAT has been and still is one of the most important tools for Ada
advocacy, and Ada 2005 offers many opportunities to get new people
interested.  Although I am a strong supporter of open source software,
I am an even stronger supporter of Ada and would like to see its usage
increase.

It is a fact that the Ada community is in much more need for support
and advocacy than the open source community is.

A freely available Ada toolset, without limitation on how the produced
executables can be distributed, is hugely more important for Ada's
progress than a GPL enforcing Ada toolset might be for the open source
community.

The former can attract new developers and one-man-shops to Ada, help
them build new products, and increase the "market share" of Ada.
The latter will most probably have a negligible effect on the strength
of the open source movement, while it might chase away many who are
genuinely interested in Ada, because they have other alternatives where
this GPL "contamination" is not an issue (GNU C++ for example).

Therefore, I feel it is very important that on a platform such as
Debian a free "no strings attached" Ada development environment stays
available.  Ideally, this should be a recent GNAT release (both for
bug fixes and Ada 2005 features), close to the "normal" Debian GCC
distribution, and with an as complete toolset as possible (ASIS, GLADE,
etc.).  Developers of GPL software would then still have the option of
using the GNAT GPL 2005 Edition directly from libre.adacore.com, should
they so desire.

Hence, my vote is:

Voter             gnat-gpl gnat-3.4  gnat-4.0 gnat-3.4+patches WAIT
-------------------------------------------------------------------
Anonymous #3          -1                2

-- 
Ludovic Brenta, on behalf of an anonymous poster.

Re: GNAT GPL 2005 Edition is now available

[Florian Weimer]

* Ludovic Brenta cites anonymous:

> From the reactions in the comp.lang.ada forum it is clear that the
> positive news about the much anticipated "official" public GNAT
> distribution is completely drowned by the strong negative feelings
> triggered by the limitation that it can only be used to develop GPL
> software.

I find this rather strange as well.  My admiration for AdaCore as a
company has declined over the years, and I do feel somewhat neglected,
but if this paves the way for continued support of free GNAT versions
(as both in beer and freedom), it should be really good news.

Re: GNAT GPL 2005 Edition is now available

[Marc A. Criley]

Ludovic Brenta wrote:
> Here is another (and extensive) contribution from an anonymous
> contributor, posted on their behalf.  Here goes:
> 
> AdaCore obviously is fully entitled to distribute the GNAT GPL 2005
> Edition as they did, but this license change is unfortunate for the
> promotion of Ada.

<snipped>

Heh, I'm not asking, but I'm wondering if any of the anonymous 
submissions are coming from AdaCore employees who disagree with this 
approach.  Is there a minority there going "I told you so!"?

-- Marc A. Criley
-- McKae Technologies
-- www.mckae.com
-- DTraq - XPath In Ada - XML EZ Out

Re: GNAT GPL 2005 Edition is now available

[Ludovic Brenta]

"Marc A. Criley" <mcNOSPAM@mckae.com> writes:

> Ludovic Brenta wrote:
>> Here is another (and extensive) contribution from an anonymous
>> contributor, posted on their behalf.  Here goes:
>> AdaCore obviously is fully entitled to distribute the GNAT GPL 2005
>> Edition as they did, but this license change is unfortunate for the
>> promotion of Ada.
>
> <snipped>
>
> Heh, I'm not asking, but I'm wondering if any of the anonymous 
> submissions are coming from AdaCore employees who disagree with this 
> approach.  Is there a minority there going "I told you so!"?

No, so far no anonymous contributor was an AdaCore employee.

-- 
Ludovic Brenta.

Re: GNAT GPL 2005 Edition is now available

[Niklas Holsti]

Ludovic Brenta wrote:
> 
> Please take into account that recent versions of ASIS and GLADE are
> only available under GPL, not GMGPL.  The same is true for AWS,
> GtkAda, XML/Ada etc. but is less of a concern because these libraries
> are not tightly coupled with the compiler.
> 
> Please cast a vote on this forum.  This is the time to influence the
> next version of Debian.

I applaud Ludovic for arranging this vote (and I voted already).

However, I think that Ludovic should not feel *bound* to choose 
the alternative that gets the most votes; at least I did not mean 
my vote to have that effect. I understood -- perhaps selfishly -- 
the question to be "what sort of GNAT compiler do you, as a Debian 
Ada user, need for what you are doing" and I voted from that 
perspective.

But perhaps the choice of the next Debian Ada compiler should 
depend more on attracting *new* Ada users. It seems to me that 
this would give more importance to good support of Ada 2005 and 
ASIS and GLADE, to make the full, modern Ada technology available 
on Debian. Therefore, I urge Ludovic to override my vote for a 
GMGPL GNAT if it conflicts with good support of Ada 2005, ASIS and 
GLADE.

Practically speaking, it is probably a lot easier for me to 
install a GMGPL version of GNAT (perhaps an older one) without 
ASIS and GLADE (which I don't need)  than it is for Ludovic and 
his team to make ASIS and GLADE work with a GNAT distribution that 
does not come from AdaCore.

I would also like to ask the group's opinion on a licensing issue 
related to the above. Suppose that GNAT GPL 2005 is chosen as the 
new Debian Ada compiler, but I want to install a GNAT with a GMGPL 
library. Supppose further that I get the sources for such a GNAT 
and compile it (compiler and libraries) with GNAT GPL 2005. The 
resulting compiler is under the strict GPL since it contains the 
strict GPL library. However, as I understand it, application 
programs compiled with the resulting compiler and linked with its 
GMGPL library are not GPL, but can be distributed as binary only. 
Do you agree? I know that no-one (?) here is a lawyer, but I would 
value your opinion on this.

-- 
Niklas Holsti
Tidorum Ltd
niklas holsti tidorum fi
       .      @       .

Re: GNAT GPL 2005 Edition is now available

[Ludovic Brenta]

Niklas Holsti <nobody@nowhere.fi> writes:

> Ludovic Brenta wrote:
>> Please take into account that recent versions of ASIS and GLADE are
>> only available under GPL, not GMGPL.  The same is true for AWS,
>> GtkAda, XML/Ada etc. but is less of a concern because these libraries
>> are not tightly coupled with the compiler.
>> Please cast a vote on this forum.  This is the time to influence the
>> next version of Debian.
>
> I applaud Ludovic for arranging this vote (and I voted already).

Thanks.

> However, I think that Ludovic should not feel *bound* to choose the
> alternative that gets the most votes; at least I did not mean my
> vote to have that effect. I understood -- perhaps selfishly -- the
> question to be "what sort of GNAT compiler do you, as a Debian Ada
> user, need for what you are doing" and I voted from that
> perspective.

And you were right.  I am just standing by Debian's Social Contract,
clause 4: "We will be guided by the needs of our users and the free
software community".

> But perhaps the choice of the next Debian Ada compiler should depend
> more on attracting *new* Ada users. It seems to me that this would
> give more importance to good support of Ada 2005 and ASIS and GLADE,
> to make the full, modern Ada technology available on
> Debian. Therefore, I urge Ludovic to override my vote for a GMGPL
> GNAT if it conflicts with good support of Ada 2005, ASIS and GLADE.

Understood, but I wouldn't want to spend time and effort doing
packages that nobody wants to use.  Therefore I urge everyone here to
try and consider all aspects - philosophical, technical, practical and
timing-related - in their votes.

[...]

> I would also like to ask the group's opinion on a licensing issue
> related to the above. Suppose that GNAT GPL 2005 is chosen as the
> new Debian Ada compiler, but I want to install a GNAT with a GMGPL
> library. Supppose further that I get the sources for such a GNAT and
> compile it (compiler and libraries) with GNAT GPL 2005. The
> resulting compiler is under the strict GPL since it contains the
> strict GPL library. However, as I understand it, application
> programs compiled with the resulting compiler and linked with its
> GMGPL library are not GPL, but can be distributed as binary only.
> Do you agree? I know that no-one (?) here is a lawyer, but I would
> value your opinion on this.

The license of the compiler has no effect on executables; the license
of the run-time library has.  Therefore, you can legally use GNAT GPL
2005 Edition to compile a GMGPL version of libgnat (from the FSF), and
link your programs with that libgnat.

In fact, you could install GNAT GPL and libgnat GMGPL separately, and
use them together.  This may require some tweaking, but it is
feasible.  I think that the tweaking would be minimal if you mix GNAT
GPL 2005 Edition with libgnat-3.4.

-- 
Ludovic Brenta.

Re: GNAT GPL 2005 Edition is now available

[Brian May]

>>>>> "Ludovic" == Ludovic Brenta <ludovic@ludovic-brenta.org> writes:

    Ludovic> In fact, you could install GNAT GPL and libgnat GMGPL
    Ludovic> separately, and use them together.  This may require some
    Ludovic> tweaking, but it is feasible.  I think that the tweaking
    Ludovic> would be minimal if you mix GNAT GPL 2005 Edition with
    Ludovic> libgnat-3.4.

That gets my vote. At least I think we should seriously consider it as
an option.
-- 
Brian May <bam@snoopy.apana.org.au>

Re: GNAT GPL 2005 Edition is now available

[Samuel Tardieu]

>>>>> "Ludovic" == Ludovic Brenta <ludovic@ludovic-brenta.org> writes:

Ludovic> In fact, you could install GNAT GPL and libgnat GMGPL
Ludovic> separately, and use them together.  This may require some
Ludovic> tweaking, but it is feasible.  I think that the tweaking
Ludovic> would be minimal if you mix GNAT GPL 2005 Edition with
Ludovic> libgnat-3.4.

This would be highly risky. For those who do not know the compilation
process of GNAT, here is a short overview.

Complex Ada constructs cannot be translated into a *short* sequence of
machine code. While basic operations such as

  A + B * C

are translated into a few arithmetic operations (I'm purposely
skipping range checking considerations here) by the GNAT compiler,
task creation or task priority manipulation require complex
operations.

The GNAT compiler doesn't try to directly generate machine code for
those complex operations. It replaces them by calls to the runtime
library, which will perform its duties, such as allocating memory for
tasks (task control block, stack space, ...) or inserting the task
at the right place into the scheduler.

While replacing the standalone library packages (such as Ada.Text_IO)
with other versions is almost straightforward and is not likely to
lead to bogus behaviour, replacing the runtime tasking library by
another version which may not *exactly* match the output produced by
the compiler is very unsafe and may cause hard-to-catch errors.

  Sam
-- 
Samuel Tardieu -- sam@rfc1149.net -- http://www.rfc1149.net/sam

Re: GNAT GPL 2005 Edition is now available

[David Trudgett]

Jamie Ayre <ayre@adacore.com> writes:

> AdaCore announces the immediate availability of the GNAT GPL 2005
> Edition.
>
> AdaCore is pleased to announce the release of the GNAT GPL 2005
> Edition to provide Free Software developers, that is developers that
> distribute their work under the GPL (GNU General Public License), the
> latest and most advanced Ada 2005 software development environment.

Hello, Jamie,

Your website says: 

    The GNAT GPL Edition is licensed for Free Software development and
    is subject to the terms and conditions of the GNU General Public
    License (GPL). Read the terms carefully before downloading.

(It then provides a link to the GPL licence at gnu.org.)

The GCC compiler is also licensed under the GPL, which allows the
development and distribution of proprietary software using that
compiler. Could you please explain here on comp.lang.ada how the same
licence can have two different effects, one for GCC, and another for
GNAT?

Thank you.

David



-- 

David Trudgett
http://www.zeta.org.au/~wpower/

It is seldom that any liberty is lost all at once.

    -- David Hume

Re: GNAT GPL 2005 Edition is now available

[Ludovic Brenta]

David Trudgett a écrit :

> Jamie Ayre <ayre@adacore.com> writes:
>
> > AdaCore announces the immediate availability of the GNAT GPL 2005
> > Edition.
> >
> > AdaCore is pleased to announce the release of the GNAT GPL 2005
> > Edition to provide Free Software developers, that is developers that
> > distribute their work under the GPL (GNU General Public License), the
> > latest and most advanced Ada 2005 software development environment.
>
> Hello, Jamie,
>
> Your website says:
>
>     The GNAT GPL Edition is licensed for Free Software development and
>     is subject to the terms and conditions of the GNU General Public
>     License (GPL). Read the terms carefully before downloading.
>
> (It then provides a link to the GPL licence at gnu.org.)
>
> The GCC compiler is also licensed under the GPL, which allows the
> development and distribution of proprietary software using that
> compiler. Could you please explain here on comp.lang.ada how the same
> licence can have two different effects, one for GCC, and another for
> GNAT?
>
> Thank you.
>
> David

It's the run-time library.

The compiler has always been GPL, and this does not affect programs
compiled with it.  In contrast, parts of the run-time library are
linked into programs compiled with GCC or GNAT (unless the zero
run-time option is chosen).  The GNAT run-time library used to have
a slightly different license, the "GNAT-Modified GPL" (or GMGPL as
we know it here), granting special permission to link this library
into proprietary programs.  GNAT GPL 2005 Edition revokes that
special permission by using the pure GPL.

Now what is confusing is that it is still possible to obtain a
GMGPL run-time library from gcc.gnu.org.  But this may change in
the future, at the FSF's option.

-- 
Ludovic Brenta.

Re: GNAT GPL 2005 Edition is now available

[wojtek]

> Now what is confusing is that it is still possible to obtain
> a GMGPL run-time library from gcc.gnu.org.  But this may
> change in the future, at the FSF's option.

If libstdc++ will be eligible for use in commercial projects, and GNAT
runtime won't, it is not going to be so great.

Regards,
Wojtek Narczynski

Re: GNAT GPL 2005 Edition is now available

[Martin Krischik]

wojtek@power.com.pl wrote:

>> Now what is confusing is that it is still possible to obtain
>> a GMGPL run-time library from gcc.gnu.org.  But this may
>> change in the future, at the FSF's option.
> 
> If libstdc++ will be eligible for use in commercial projects, and GNAT
> runtime won't, it is not going to be so great.

Well GPL does not mean "hobby only" - just give your customer the sources
and all is well. Where is the problem?

Martin

-- 
mailto://krischik@users.sourceforge.net
Ada programming at: http://ada.krischik.com

Re: GNAT GPL 2005 Edition is now available

[Samuel Tardieu]

>>>>> "Martin" == Martin Krischik <krischik@users.sourceforge.net> writes:

Martin> Well GPL does not mean "hobby only" - just give your customer
Martin> the sources and all is well. Where is the problem?

Giving your customer the sources is not enough. You have to place them
under a GPL compatible licence[1]. Not *any* Free Software licence you
want.

Let's take an example: you choose to write an Ada program, under the
2-clauses GPL-compatible BSD licence, using AWS to offer a HTTPS
interface. Even though your program is distributed under a
GPL-compatible licence, even though AWS allows you to link with
whatever you want thanks to the GMGPL, you will not be able to
distribute a binary of your application because OpenSSL, used to
implementing HTTPS in AWS, uses a non-GPL-compatible Free Software
licence.

It means that such an application cannot be included as a binary
package in GNU/Linux distributions, even though all the sources are
Free Software, and even if you distribute all of them.

If the GNAT runtime was still using the GMGPL as it was before, there
would be no problem at all. Linking bits of GMGPL, 2-clauses BSD
licence and the OpenSSL licence is perfectly possible. Adding GPL bits
(the new GNAT runtime licence) in the middle makes it impossible.

[this post does not judge the value of respective Free Software
licences but shows that the new GPL-only licence adds new restrictions
to Free Software development in Ada when using the compiler provided
by AdaCore]

  Sam

[1] I think that AdaCore is wrong in saying that you can only
    develop GPL software with GNAT GPL 2005. My understanding is that
    if you distribute a binary compiled with GNAT GPL 2005, your
    program licence has to be *compatible* with the GPL. See
    http://www.gnu.org/philosophy/license-list.html
-- 
Samuel Tardieu -- sam@rfc1149.net -- http://www.rfc1149.net/sam

Re: GNAT GPL 2005 Edition is now available

[Björn Persson]

Samuel Tardieu wrote:
> [1] I think that AdaCore is wrong in saying that you can only
>     develop GPL software with GNAT GPL 2005. My understanding is that
>     if you distribute a binary compiled with GNAT GPL 2005, your
>     program licence has to be *compatible* with the GPL. See
>     http://www.gnu.org/philosophy/license-list.html

You *can* take source code with a GPL-compatible license, such as the 
X11 license, link it with a GPL library, such as those of the GNAT GPL 
2005 Edition, and distribute the resulting binary, *but* the binary you 
distribute will be GPL-licensed. This is because the X11 license is 
GPL-compatible but the GPL isn't X11-compatible.

-- 
Bj�rn Persson                              PGP key A88682FD
                    omb jor ers @sv ge.
                    r o.b n.p son eri nu

Re: GNAT GPL 2005 Edition is now available

[Samuel Tardieu]

>>>>> "Bj�rn" == Bj�rn Persson <spam-away@nowhere.nil> writes:

Bj�rn> You *can* take source code with a GPL-compatible license, such
Bj�rn> as the X11 license, link it with a GPL library, such as those
Bj�rn> of the GNAT GPL 2005 Edition, and distribute the resulting
Bj�rn> binary, *but* the binary you distribute will be
Bj�rn> GPL-licensed. This is because the X11 license is GPL-compatible
Bj�rn> but the GPL isn't X11-compatible.

You cannot say that licence A is compatible with licence B and licence
B is not compatible with licence A :)

GPL and X11 licence are compatible. One of them (the GPL) forces the
combined work to be GPL, the other one (the X11 licence) doesn't care
under what licence the combined work is being distributed. That's why
they are compatible with each other.

  Sam
-- 
Samuel Tardieu -- sam@rfc1149.net -- http://www.rfc1149.net/sam

Re: GNAT GPL 2005 Edition is now available

[Florian Weimer]

* Samuel Tardieu:

>>>>>> "Martin" == Martin Krischik <krischik@users.sourceforge.net> writes:
>
> Martin> Well GPL does not mean "hobby only" - just give your customer
> Martin> the sources and all is well. Where is the problem?
>
> Giving your customer the sources is not enough. You have to place them
> under a GPL compatible licence[1]. Not *any* Free Software licence you
> want.

You don't even have to use a Free Software license if you don't
distribute binaries and don't use the innards of the GNAT run-time
library.  The GPL isn't *that* viral.

Whether compiling the application on-site is fine, I don't know.

Re: GNAT GPL 2005 Edition is now available

[Samuel Tardieu]

>>>>> "Florian" == Florian Weimer <fw@deneb.enyo.de> writes:

Sam> Giving your customer the sources is not enough. You have to place
Sam> them under a GPL compatible licence[1]. Not *any* Free Software
Sam> licence you want.

Florian> You don't even have to use a Free Software license if you
Florian> don't distribute binaries and don't use the innards of the
Florian> GNAT run-time library.  The GPL isn't *that* viral.

I was precisely talking about distributing binaries :) And for not
using the GNAT runtime library, please read the other message I just
posted. You can hardly afford not to use it in real programs except if
you use the "No_Runtime" mode, in which case you are probably writing
mission-critical software and are likely to need a support contract to
be safe.

Florian> Whether compiling the application on-site is fine, I don't
Florian> know.

It is an interesting point: the GPL forces you to release your
changes under the GPL regardless of the kind of redistribution you
choose (source or binary). However, in this case, your program is
clearly not a derivative work of the GNAT runtime library, which would
mean that you are not bound by the GPL contamination brought by the
GNAT RTL if you redistribute your program in source form. Why? Because
it could for example be compiled by another Ada compiler.

However, if your program uses GNAT-specific library packages, the
question may be raised again, as it may be considered a derivative
work unless you provide another implementation of those packages.

  Sam
-- 
Samuel Tardieu -- sam@rfc1149.net -- http://www.rfc1149.net/sam

Re: GNAT GPL 2005 Edition is now available

[Ole-Hjalmar Kristensen]

So in principle, I could package my non-GPL-compatible (possibly
encrypted) program source with the GNAT GPL 2005 and run an automated
install script at the user's computer which would install GNAT, then
compile and link the program, effectively installing a binary version?
Maybe even throw away the source (and GNAT) afterwards, since
technically, the user has compiled the program from sources himself?
The mind boggles.

-- 
   C++: The power, elegance and simplicity of a hand grenade.

Re: GNAT GPL 2005 Edition is now available

[Björn Persson]

Ole-Hjalmar Kristensen wrote:
> So in principle, I could package my non-GPL-compatible (possibly
> encrypted) program source with the GNAT GPL 2005 and run an automated
> install script at the user's computer which would install GNAT, then
> compile and link the program, effectively installing a binary version?
> Maybe even throw away the source (and GNAT) afterwards, since
> technically, the user has compiled the program from sources himself?

No. Then you would have created an installer based on Gnat. I'm fairly 
sure the package as a whole would be considered a work based in part on 
Gnat, so that those copies of your code that were distributed in that 
package would have to be placed under the GPL.

-- 
Bj�rn Persson                              PGP key A88682FD
                    omb jor ers @sv ge.
                    r o.b n.p son eri nu

Re: GNAT GPL 2005 Edition is now available

[Björn Persson]

Florian Weimer wrote:
> Whether compiling the application on-site is fine, I don't know.

I asked this question to David Axmark of MySQL AB some time ago. They've 
changed the license of the MySQL client library from LGPL to GPL, but 
you can get another license if you pay for it. I asked as follows: If a 
program that uses the MySQL library is distributed as source code � but 
with a non-free license � and compiled on the customer's computer, does 
that require buying a license?

He said that if I were to ask Richard Stallman or Eben Moglen this, I 
would get a long exposition but no clear answer, and that's basically 
what I got from David too. He said that he considered it a matter of 
intent, and talked a little about how lawyers can pick a text apart 
until you don't recognize it. As I understood it he thought it was 
acceptable if you'd distribute as source code for technical reasons but 
not if you'd do it to circumvent the need to buy a license, but it seems 
rather unclear what the GPL actually says about this.

-- 
Bj�rn Persson                              PGP key A88682FD
                    omb jor ers @sv ge.
                    r o.b n.p son eri nu

Re: GNAT GPL 2005 Edition is now available

[wojtek]

Try to convince some venture capitalists, or high level managers, that
it's better to code in Ada, even thought you'll have to supply your
sources to the customers (practically also competitors) under GPL.
Please get back to me if you succeed.

Re: GNAT GPL 2005 Edition is now available

[Larry Kilgallen]

In article <1127171105.496123.178650@g44g2000cwa.googlegroups.com>, wojtek@power.com.pl writes:

> Try to convince some venture capitalists, or high level managers, that
> it's better to code in Ada, even thought you'll have to supply your
> sources to the customers (practically also competitors) under GPL.

Presumably any company with "venture capitalists" or "high level managers"
has no hesitation about paying for compiler support with GNAT Professional.
Thus the issue does not arrive in those environments.

Re: GNAT GPL 2005 Edition is now available

[Preben Randhol]

In article <6cFjLy+0$XDn@eisner.encompasserve.org>, Larry Kilgallen wrote:
>In article <1127171105.496123.178650@g44g2000cwa.googlegroups.com>, wojtek@power.com.pl writes:
>
>> Try to convince some venture capitalists, or high level managers, that
>> it's better to code in Ada, even thought you'll have to supply your
>> sources to the customers (practically also competitors) under GPL.
>
>Presumably any company with "venture capitalists" or "high level managers"
>has no hesitation about paying for compiler support with GNAT Professional.
>Thus the issue does not arrive in those environments.

However if you are a tiny company with 1-5 employees and starting up it
would be difficult to afford an expensive support which you probably do
not need. However, there is one solution and that is to make the program or
interface GPL while the data it uses is non-GPL. Scientific programs
usually don't costs so much, but as you start buying database to do
calculations the costs increase rapidly. Another way is to ditch Ada and
use Java. I fear the last is the most probable... :-(

Re: GNAT GPL 2005 Edition is now available

[Larry Kilgallen]

In article <slrndj2282.oqn.randhol@bacchus.pvv.ntnu.no>, randhol@bacchus.pvv.ntnu.no (Preben Randhol) writes:
> In article <6cFjLy+0$XDn@eisner.encompasserve.org>, Larry Kilgallen wrote:
>>In article <1127171105.496123.178650@g44g2000cwa.googlegroups.com>, wojtek@power.com.pl writes:
>>
>>> Try to convince some venture capitalists, or high level managers, that
>>> it's better to code in Ada, even thought you'll have to supply your
>>> sources to the customers (practically also competitors) under GPL.
>>
>>Presumably any company with "venture capitalists" or "high level managers"
>>has no hesitation about paying for compiler support with GNAT Professional.
>>Thus the issue does not arrive in those environments.
> 
> However if you are a tiny company with 1-5 employees and starting up it
> would be difficult to afford an expensive support which you probably do
> not need.

But then you would not have a need to "convince some venture capitalists,
or high level managers", which is the remark to which I responded.

Re: GNAT GPL 2005 Edition is now available

[Preben Randhol]

In article <Usovzqyupy7k@eisner.encompasserve.org>, Larry Kilgallen wrote:
>But then you would not have a need to "convince some venture capitalists,
>or high level managers", which is the remark to which I responded.

Sure. My point was that a small firm might not need support, but only
GMGPL compiler. However, this does not seem to be for sale at an
affordable price. I mean why not just buy a language X compiler off the
shelf to make your software in stead of paying serious amounts of money
for an Ada compiler WITH support for a year? Especially when the latter
would topple the whole economy of the project...

Preben

Re: GNAT GPL 2005 Edition is now available

[Steve]

<wojtek@power.com.pl> wrote in message 
news:1127171105.496123.178650@g44g2000cwa.googlegroups.com...
> Try to convince some venture capitalists, or high level managers, that
> it's better to code in Ada, even thought you'll have to supply your
> sources to the customers (practically also competitors) under GPL.
> Please get back to me if you succeed.
>

A strong argument to support making source code available is competition.
  You can purchase software from company A or company B.
  The software has similar price and function.
  Company A supplies the source code licensed under the GPL.
  Company B does not supply source code.

All other things the same, which would you choose?

On the other hand...

I really don't like AdaCore releasing a GPL-only edition of the GNAT runtime 
libraries.

My objection:
  Over the years I have contributed ideas, reported bugs, and submitted 
patches to fix bugs in libraries that were originally provided by AdaCore. 
I have been more than glad to give AdaCore feedback and corrections in 
return for the distributions they have provided.  I never expected support, 
for that I expect to pay for.
  Now some of the ideas and bug fixes I have submitted are released under 
the GPL, which is a more restrictive license than I contributed to.  I feel 
somehow betrayed.
  I haven't produced any commercial software using GNAT that would conflict 
with the GPL, but it still makes me uneasy on how easy it was to change the 
rules.

Also, a silly question:
  If AdaCore can release a distribution of GNAT that is more restrictive, 
what's to prevent me from releasing a distribution that is closed source

Steve
(The Duck)

Re: GNAT GPL 2005 Edition is now available

[Ludovic Brenta]

"Steve" <nospam_steved94@comcast.net> writes:
> I really don't like AdaCore releasing a GPL-only edition of the GNAT
> runtime libraries.

Could you be more explicit about your vote, if any?  I am in fact more
interested in knowing which version you would recommend, than which
one you don't like :)

-- 
Ludovic Brenta.

Re: GNAT GPL 2005 Edition is now available

[Larry Kilgallen]

In article <878xxr1uvo.fsf@ludovic-brenta.org>, Ludovic Brenta <ludovic@ludovic-brenta.org> writes:
> "Steve" <nospam_steved94@comcast.net> writes:
>> I really don't like AdaCore releasing a GPL-only edition of the GNAT
>> runtime libraries.
> 
> Could you be more explicit about your vote, if any?  I am in fact more
> interested in knowing which version you would recommend, than which
> one you don't like :)

I did not see anything in Steve's post regarding your Linux vote,
and this thread is about the general release action, not just Linux.

Things would have been cleaner with a separate Linux vote thread.

Re: GNAT GPL 2005 Edition is now available

[Ludovic Brenta]

Kilgallen@SpamCop.net (Larry Kilgallen) writes:
> I did not see anything in Steve's post regarding your Linux vote,
> and this thread is about the general release action, not just Linux.
>
> Things would have been cleaner with a separate Linux vote thread.

Your feeble attempt at stealing "my thread" is futile as I've already
posted the results :) All your votes are belong to me :) Now you may
go back and argue, you are dismissed. :) :) :) :)

Are you going to start a vote for VMS now?

-- 
Ludovic Brenta.

Re: GNAT GPL 2005 Edition is now available

[Steve]

"Ludovic Brenta" <ludovic@ludovic-brenta.org> wrote in message 
news:878xxr1uvo.fsf@ludovic-brenta.org...
> "Steve" <nospam_steved94@comcast.net> writes:
>> I really don't like AdaCore releasing a GPL-only edition of the GNAT
>> runtime libraries.
>
> Could you be more explicit about your vote, if any?  I am in fact more
> interested in knowing which version you would recommend, than which
> one you don't like :)

Sorry, I was not responding to your poll.
If I were to vote, it would be for one of the versions that has the runtime 
under the GMGPL.  I haven't tried the different versions, so I can't give a 
recommendation.

Steve
(The Duck)

>
> -- 
> Ludovic Brenta. 

Re: GNAT GPL 2005 Edition is now available

[Steve Whalen]

Ludovic,

First, let me thank you for the work you do keeping GNAT on Debian.
Since I converted to Debian a few years ago, I've used the version of
GNAT you've packaged and really appreciate the work you and the other
maintainers do to keep Debian going.

I wish I could provide you some of the input you seek on the GNAT2005
vs. 3.4 vs. 4.0 vs. etc. debate with respect to quality.  Unfortunatly
I've stuck with the 3.15 you package on Debian, and then I just use the
most compatible version of the same "generation" (aka 3.15) on Windows.

I vote AGAINST any verion of GNAT with the runtime libraries under GPL
instead of the GMGPL.

I vote FOR whichever of the 3 GMGPL runtime library other alternatives
is:

1) most stable and easiest for you to package and maintain (my exposure
to them is so limited as to be of no use to you, sorry).

2) if it's a tie after looking at #1 then whichever version has the
most stable SMP support is my next priority (I'm not sure it's even the
GNAT library that's causing some of the reported SMP problems).

3) if you get some solid feedback from people who have had time to do
some systematic testing between the alternatives, you can convert my
vote to match your judgement of the consensus on which is the most
stable version, and your judgement of which would be easist for you to
package and maintain.

In other words, I'd be willing to give up a little stability or a few
more minor bugs if a version was a LOT easier for you to maintain: if
two choices are about the same amount of work for you, then I'd go for
the most stable.

Sorry I can't give more positive feedback, but it's kind of a chicken
and egg thing: I'm on GNAT 3.15 as long as that's the version in
Debian: as soon as you package something else, I'll try it and if it
handles everything I care about, I'll most likely switch to it.

I'd like to get some of the features of Ada 2005, but not if the price
is dropping the GMGPL on the runtime libraries or going to a less
stable compiler.

Steve

Re: GNAT GPL 2005 Edition is now available

[chris]

wojtek@power.com.pl wrote:
> Try to convince some venture capitalists, or high level managers, that
> it's better to code in Ada, even thought you'll have to supply your
> sources to the customers (practically also competitors) under GPL.
> Please get back to me if you succeed.

Why would you tell them what language it's in?  Surely all they need to 
know is it's not BrainF**k or Whitespace ;)

Re: GNAT GPL 2005 Edition is now available

[Ole-Hjalmar Kristensen]

>>>>> "c" == chris  <spamoff.danx@ntlworld.com> writes:

    c> wojtek@power.com.pl wrote:
    >> Try to convince some venture capitalists, or high level managers, that
    >> it's better to code in Ada, even thought you'll have to supply your
    >> sources to the customers (practically also competitors) under GPL.
    >> Please get back to me if you succeed.

    c> Why would you  tell them what language it's in?  Surely all they need
    c> to know is it's not BrainF**k or Whitespace ;)

Actually, both potential customers and investors may care. About 1992,
in a previous company, I managed to convince the rest of the project
that Ada would be a good language to implement a soft real time, fault
tolerant distributed RDBMS in. We got quite a bit of negative comments
from technically minded customers, and also potential investors,
mostly because C++ was seen as the coming language, and Ada was
perceived as un-cool. They figured that being left out of the
mainstream was not a good idea. What eventually shot the idea down,
was ridiculous prices for compilers, and non-availability for certain
platforms (which we ended up not porting to anyway).  The system is
now C, C++, and Java.

-- 
   C++: The power, elegance and simplicity of a hand grenade.

Re: GNAT GPL 2005 Edition is now available

[Martin Krischik]

Ludovic Brenta wrote:

> Now what is confusing is that it is still possible to obtain a
> GMGPL run-time library from gcc.gnu.org.  But this may change in
> the future, at the FSF's option.

Not that confusing: If you are capable enough to compile you own compiler
you get GMGPL as bonus.

And I can see the point: The release of GNAT 2005 GPL probably cost AdaCore
a lot of money. And they did not want to speed that money on thouse who are
neither OpenSource developers nor AdaCore customers.

And remember: As an OpenSource developer you only have to give the sources
to thouse you have given the Binaries and to no one else. And thats fair
enough.

Martin
-- 
mailto://krischik@users.sourceforge.net
Ada programming at: http://ada.krischik.com

Re: GNAT GPL 2005 Edition is now available

[Martin Dowie]

Martin Krischik wrote:
> And remember: As an OpenSource developer you only have to give the sources
> to thouse you have given the Binaries and to no one else. And thats fair
> enough.

And only if they ask for them, yes?

Cheers

-- Martin

Re: GNAT GPL 2005 Edition is now available

[Samuel Tardieu]

>>>>> "Martin" == Martin Krischik <krischik@users.sourceforge.net> writes:

Martin> And remember: As an OpenSource developer you only have to give
Martin> the sources to thouse you have given the Binaries and to no
Martin> one else. And thats fair enough.

It depends on the Free Software licence you have chosen and what kind
of distribution you have used. For example, if you choose to
distribute your program under section 3.b) of the GPL (that is provide
sources on demand only instead of spontaneously), you have to provide
*any third party* with a copy of the sources.

  Sam
-- 
Samuel Tardieu -- sam@rfc1149.net -- http://www.rfc1149.net/sam

Re: GNAT GPL 2005 Edition is now available

[Preben Randhol]

In article <87vf1284wu.fsf@willow.rfc1149.net>, Samuel Tardieu wrote:
>>>>>> "Martin" == Martin Krischik <krischik@users.sourceforge.net> writes:
>
>Martin> And remember: As an OpenSource developer you only have to give
>Martin> the sources to thouse you have given the Binaries and to no
>Martin> one else. And thats fair enough.
>
>It depends on the Free Software licence you have chosen and what kind
>of distribution you have used. For example, if you choose to
>distribute your program under section 3.b) of the GPL (that is provide
>sources on demand only instead of spontaneously), you have to provide
>*any third party* with a copy of the sources.

I don't understand. I can make a GPL program and keep it for myself or
give it to one or two friends. I don't need to give it to the rest of
the world even if it is GPL.

So, could one in fact not just make a program which one gives to company
A with the source (or at least that it is available to them)? One didn't
need to give it to company B?

And could one also draw up and agreement with Company A not to spread
the source code and binaries or it would not be possible?

I mean if Company A can spread the code or binary to any company you
would not sell more than one copy...

Preben

Re: GNAT GPL 2005 Edition is now available

[Samuel Tardieu]

Sam>  It depends on the Free Software licence you have chosen and what
Sam> kind of distribution you have used. For example, if you choose to
Sam> distribute your program under section 3.b) of the GPL (that is
Sam> provide sources on demand only instead of spontaneously), you have
Sam> to provide *any third party* with a copy of the sources.

Preben> I don't understand. I can make a GPL program and keep it for
Preben> myself or give it to one or two friends. I don't need to give
Preben> it to the rest of the world even if it is GPL.

If you distribute it in binary form only to your friends under the GPL
license and choose to use the 3.b) section of the GPL instead of
giving them the source code, then you have to provide *anyone*
requesting it a copy of the source code at a reasonable fee. This is
meant to encourage you to distribute the source code spontaneously in
a first place. Excerpt:

|   3. You may copy and distribute the Program (or a work based on it,
| under Section 2) in object code or executable form under the terms of
| Sections 1 and 2 above provided that you also do one of the following:
| [...]
|     b) Accompany it with a written offer, valid for at least three
|     years, to give any third party, for a charge no more than your
|     cost of physically performing source distribution, a complete
|     machine-readable copy of the corresponding source code, to be
|     distributed under the terms of Sections 1 and 2 above on a medium
|     customarily used for software interchange;

Note the *any third party* in the text above.

  Sam
-- 
Samuel Tardieu -- sam@rfc1149.net -- http://www.rfc1149.net/sam

Re: GNAT GPL 2005 Edition is now available

[Brian May]

>>>>> "Preben" == Preben Randhol <randhol@bacchus.pvv.ntnu.no> writes:

    Preben> And could one also draw up and agreement with Company A
    Preben> not to spread the source code and binaries or it would not
    Preben> be possible?

Then it wouldn't be GPL anymore. You would be replacing the GPL with
another more restrictive license.
-- 
Brian May <bam@snoopy.apana.org.au>

Re: GNAT GPL 2005 Edition is now available

[Jeff Creem]

David Trudgett wrote:
> Jamie Ayre <ayre@adacore.com> writes:
> 
> 
>>AdaCore announces the immediate availability of the GNAT GPL 2005
>>Edition.
>>
>>AdaCore is pleased to announce the release of the GNAT GPL 2005
>>Edition to provide Free Software developers, that is developers that
>>distribute their work under the GPL (GNU General Public License), the
>>latest and most advanced Ada 2005 software development environment.
> 
> 
> Hello, Jamie,
> 
> Your website says: 
> 
>     The GNAT GPL Edition is licensed for Free Software development and
>     is subject to the terms and conditions of the GNU General Public
>     License (GPL). Read the terms carefully before downloading.
> 
> (It then provides a link to the GPL licence at gnu.org.)
> 
> The GCC compiler is also licensed under the GPL, which allows the
> development and distribution of proprietary software using that
> compiler. Could you please explain here on comp.lang.ada how the same
> licence can have two different effects, one for GCC, and another for
> GNAT?
> 
> Thank you.
> 
> David
> 
> 
> 

We will have to wait and see if they respond but in general questions 
like this to AdaCore usually come back (correctly) with "consult your 
own qualified legal council for advise"...


But the short ** non-qualified ** answer is that whether programming in 
C,C++ or Ada, there is generally some run-time library that one links 
against for any non-trivial programs (ok..there are probably non-trivial 
programs that can be 100% run-time free.)...

In any case, with GCC for C one usually links against the the OS libc 
which is non-GPL (or at least has that option)

For C++, the GCC runtime library has a "special exception" clause like 
the GNAT-3.15 runtime used to have.

For Ada, the runtime (and supporting libraries) which are extensive used 
to be GPL with the special exception that prevented having to apply the 
GPL to programs that "make use of" the library.

In this AdaCore version, they have removed the exception. One can only 
assume that they negotiated this with the FSF (who is the copyright 
holder on most of this even though it was mostly developed by AdaCore). 
FSF probably would not have a problem with it anyway since the more 
"Free" it is the more they like it. Note I am only speculating that this 
requires FSF approval. There is a copyright assignment process that one 
goes through when submitting code to the FSF GCC tree but it is not 
clear to me what unique rights the original copyright holder might retain.



The runtime files in the FSF tree still have the special exception so I 
would guess AdaCore has written some script that they run when packaging 
the GPL version that strips out the special exception portion of the 
header and again has cleared this with the FSF.

This does muddy the water a little even for free SW developers that want 
  to use this compiler version since it makes it (a little) more 
difficult to release the code under a non GPL but still free license. 
However, it really has no effect for people creating libraries that they 
distribute as source code under the GMGPL (with the old exception clause).

We will have to wait and see how this ends up working out. In the end

Re: GNAT GPL 2005 Edition is now available

[David Trudgett]

Jeff Creem <jcreem@yahoo.com> writes:

> David Trudgett wrote:
>> The GCC compiler is also licensed under the GPL, which allows the
>> development and distribution of proprietary software using that
>> compiler. Could you please explain here on comp.lang.ada how the same
>> licence can have two different effects, one for GCC, and another for
>> GNAT?
>
> We will have to wait and see if they respond but in general questions
> like this to AdaCore usually come back (correctly) with "consult your
> own qualified legal council for advise"...

It has been little more than 24 hours since I made my reasonable
request for a public explanation, and I expect that they are still
working on a suitable reply. "Consult legal counsel for advice," would
not be a reasonable reply, because I am not asking for legal advice, I
am asking them to publicly justify their action. Being unable or
unwilling to publicly justify what they have done would, I speculate,
have a negative impact upon their image as, for example, it might
justifiably be interpreted as a snub to the Ada community (or at least
that section which frequents c.l.a).

If AdaCore were to respond citing only a pure technical reason for the
difference between GCC and GNAT, along the lines that the C and C++
run-time licenses give freedom to developers, but the GNAT run-time
has had this freedom revoked, then AdaCore may find themselves in a
little spot of bother. This is because in that case it would be quite
apparent they are unwilling to explain their motives, and people will
be left to draw conclusions about that which will probably be
negative.

If AdaCore responds with silence, the result will be somewhat similar,
though less decisive. So, if I were AdaCore and had something to hide
about my motivations, then I would definitely fail to reply.

The explanation AdaCore comes up with must explain how restricting the
freedom of developers to license their own work in the way most
suitable to them, does not prove contempt of the very principles of
freedom that the GPL is supposed to be defending. Maybe they have such
an explanation, and (for all of us) to find out what it is, is the
purpose of my enquiry.


>
>
> But the short ** non-qualified ** answer is that whether programming
> in C,C++ or Ada, there is generally some run-time library that one
> links against for any non-trivial programs (ok..there are probably
> non-trivial programs that can be 100% run-time free.)...
>
> In any case, with GCC for C one usually links against the the OS libc
> which is non-GPL (or at least has that option)
>
> For C++, the GCC runtime library has a "special exception" clause like
> the GNAT-3.15 runtime used to have.
>
> For Ada, the runtime (and supporting libraries) which are extensive
> used to be GPL with the special exception that prevented having to
> apply the GPL to programs that "make use of" the library.
>
> In this AdaCore version, they have removed the exception. One can only
> assume 

Avoiding assumptions is a good reason for my question to Jamie.


> that they negotiated this with the FSF (who is the copyright
> holder on most of this even though it was mostly developed by
> AdaCore). FSF probably would not have a problem with it anyway since
> the more "Free" it is the more they like it. 

On the contrary, it is demonstrably less free. Previously, there were
no restrictions on what a GNAT developer could produce. Now, a GNAT
GPL developer will be restricted to producing GPL products. Too bad if
the developer wants to license under some other Free licence (for
instance). Anyone can easily see that we are going from more free to
less free in this sequence of events, and, furthermore, that this
reduction in developer freedom does not increase the protection of the
compiler or run-time against those who would like to make proprietary
versions of them.

It seems to me (correct me if I'm wrong) that libgnat is pretty
useless in practice without the GNAT compiler. It further seems to be,
therefore, that libgnat and GNAT are effectively, for all practical
purposes, parts of the same product. Although there is no problem in
licensing different parts of the same product differently (should that
be useful, and create more freedoms), by licensing libgnat under the
GPL (without the exception that is under discussion), AdaCore is,
ironically, trying to use the GPL to violate the intent of the GPL. In
particular (section 2 of the GPL, version 2):

    Thus, it is not the intent of this section to claim rights or
    contest your rights to work written entirely by you; rather, the
    intent is to exercise the right to control the distribution of
    derivative or collective works based on the Program.

At the beginning of the same section:

    You may modify your copy or copies of the Program or any portion
    of it, _*thus*forming*a*work*based*on*the*Program*_, and copy and
    distribute such modifications or work under the terms of Section 1
    above, provided that [... etc.]

    (my emphasis)


Linking one's program into a compiler's run-time library *should* not
and *does* not in itself constitute a derivative work, nor a
collective work. I am not modifying my copy of the compiler or the
run-time. I am not extending or enhancing the library, and nor am I
bundling the library within a collection. I am certainly not doing any
of this with respect to the compiler itself. When I write a program
and compile it with a compiler, and I use no libraries besides the
run-time library that comes with the compiler, then I am (and I claim
that any programmer can plainly see it) producing a work entirely by
myself. This is what everyone understands by this phrase. Otherwise,
the only works one would be able to produce entirely by oneself would
be machine code programs. No one's programs run from source code
(interpreters effectively compile to machine code on the fly).

AdaCore, in fact, seems to tacitly admit this point when they use the
word 'guarantees' in the following sentence:

    The GMGPL guarantees that executables generated by GNAT Pro can be
    distributed under customer-specific terms and conditions.

    -- https://libre2.adacore.com/dynamic/gnat_faq.html#licensing

This wording would seem to mean that the GMGPL merely provides a
_guarantee_ over a GPL that could otherwise perhaps be misinterpreted.
It would be unnecessary to _guarantee_ it unless the right might
already exist nevertheless under the plain GPL.

Of course, further down, they claim that the GPL "guarantees" that
executables compiled by GNAT GPL will be Free Software. So they could
just be using the word 'guarantees' in a sloppy way. What they really
mean by this is that they intend to *force* (this is not freedom)
developers who use GNAT GPL to licence their own software under the
GPL. This goes way beyond protecting the Free status of one's own
software and enhancements to it, and firmly steps across the line into
gratuitous control of the actions of others by restricting their
freedom.

Needless to say, using the GPL not to enhance, but to gratuitously
restrict the freedom of others, is quite distasteful, if not
hypocritical.

Notice that these points remain valid even if lawyers determine that
the GPL technically allows them to get away with it.


> Note I am only
> speculating that this requires FSF approval. There is a copyright
> assignment process that one goes through when submitting code to the
> FSF GCC tree but it is not clear to me what unique rights the original
> copyright holder might retain.
>
>
>
> The runtime files in the FSF tree still have the special exception so
> I would guess AdaCore has written some script that they run when
> packaging the GPL version that strips out the special exception
> portion of the header and again has cleared this with the FSF.

I don't believe that it is necessary to clear that with the FSF. It is
my understanding that exceptions to a GPL licence can be removed by
any distributor if they wish. [1] A distributor, however, is prohibited
from adding additional restrictions. One could say that telling
developers they must license their own programs under the GPL is an
additional restriction.


    [1] When other people modify the program, they do not have to make
    the same exception for their code--it is their choice whether to
    do so.

    -- http://www.gnu.org/licenses/gpl-faq.html



>
> This does muddy the water a little even for free SW developers that
> want to use this compiler version since it makes it (a little) more
> difficult to release the code under a non GPL but still free
> license. 

GPL software, or modifications to it, can only be released under the
GPL: i.e., not BSD, not X11, not LGPL. How is "impossible" the same as
"a little more difficult"? :-)



> However, it really has no effect for people creating
> libraries that they distribute as source code under the GMGPL (with
> the old exception clause).

GMGPL is not the GPL. GPL licensed programs can't be distributed under
the GMGPL (unless the original copyright holder says they can).




As a final comment, I would add that I am not against the GPL in any
way except where it is used to gratuitously limited other people's
freedoms. The purpose of the GPL is to protect the freedom of
software, not to force other people to produce free software. A
compiler is a special class of program that allows the GPL to be
abused, because the *purpose* of a compiler is to facilitate the
production of programs that are *not* derivative works, but original
works, yet parts of a compiler, especially but not exclusively, the
"run-time", must necessarily form part of the output of the compiler.

GPL'ed compilers, by their very nature and purpose, contain and embody
by implication an exception to a literal interpretation of the terms
of the GPL that do not allow GPL'ed code to be incorporated into
non-GPL'ed code. Can a licence really negate the raison d'etre of the
software it is supposed to be protecting? And can it do this while
negating its own raison d'etre of promoting freedom? A more perverse
situation is hard to imagine.


David


-- 

David Trudgett
http://www.zeta.org.au/~wpower/

For in a Republic, who is "the country?" Is it the Government which is
for the moment in the saddle?  Why, the Government is merely a servant
- merely a temporary servant; it cannot be its prerogative to
determine what is right and what is wrong, and decide who is a patriot
and who isn't.  Its function is to obey orders, not originate them.

    -- Mark Twain

Re: GNAT GPL 2005 Edition is now available

[Larry Kilgallen]

In article <m3mzmdh0jx.fsf@rr.trudgett>, David Trudgett <wpower@zeta.org.au.nospamplease>  writes:

> The explanation AdaCore comes up with must explain how restricting the
> freedom of developers to license their own work in the way most
> suitable to them, does not prove contempt of the very principles of
> freedom that the GPL is supposed to be defending.

The GPL effect is designed to be viral -- AdaCore is now adhering
to it more closely.  They are quite in tune with the FSF manifesto.

The fact that AdaCore motives are likely to be promoting their own
supported product is obvious, as is the fallacy of those who had
viewed GNAT as the answer to all problems in provision of compilers
for Ada.

Computing (like life) has no "magic bullet".

Re: GNAT GPL 2005 Edition is now available

[Ludovic Brenta]

Larry Kilgallen a écrit :
> David Trudgett writes:
>
>> The explanation AdaCore comes up with must explain how restricting
>> the freedom of developers to license their own work in the way most
>> suitable to them, does not prove contempt of the very principles of
>> freedom that the GPL is supposed to be defending.
>
> The GPL effect is designed to be viral -- AdaCore is now adhering to
> it more closely.  They are quite in tune with the FSF manifesto.
>
> The fact that AdaCore motives are likely to be promoting their own
> supported product is obvious, as is the fallacy of those who had
> viewed GNAT as the answer to all problems in provision of compilers
> for Ada.
>
> Computing (like life) has no "magic bullet".

I agree.  As I have stated before, I personally do not have any
objection to libgnat being GPL for non-customers.  There seems to be
quite strong feelings the opposite way, and even some frustration.
These feelings and frustration do not have to be.

AdaCore does not serve amateurs or hobbyists; AdaCore does not serve
SMEs or individual professionals.  AdaCore concentrates on serving
large businesses.  This is not new.

I asked for votes about the next Ada compiler in Debian only because I
want to do what users want.  AdaCore has no obligation to do what
non-customers want, and no obligation to explain their decisions.

Because GNAT is (very) inexpensive and (was) legally unencumbered,
amateurs, hobbyists, SMEs and individual professionals have come to
depend on AdaCore for continued updates to GNAT.  Now, many feel
betrayed.  But there is no betrayal since there was no prior
agreement.

If this dependence on AdaCore is really a problem, the solution is
independence from AdaCore, not complaints.  This independence exists
today in the form of the FSF tree.  Also, AdaCore's own CVS server
(containing AWS, GLADE, GtkAda, GPS, etc.) is still accessible and
contains sources with the special permission to link these libraries
with non-GPL programs.  If someone wants independence from AdaCore,
they can always take a snapshot of that repository and start
maintaining the thing themselves.  This is much better than being a
captive customer.  And, as a convenience, many binary distributions of
the GMGPL libraries still exist at no cost.

Also, several people here have stated their willingness to pay for
support, but that they couldn't afford AdaCore's astronomical rates.
This seems to indicate that there is a market opportunity to form a
small business to provide "low-end" support contracts.  This market is
free for the taking, as is the software itself.

Since I don't personally take issue with libgnat being GPLed, I'm
not going to be the one doing a fork.  But if this is enough of an
itch to somebody, then by all means, scratch it.

-- 
Ludovic Brenta.

Re: GNAT GPL 2005 Edition is now available

[Samuel Tardieu]

>>>>> "Ludovic" == Ludovic Brenta <ludovic@ludovic-brenta.org> writes:

Ludovic> I agree.  As I have stated before, I personally do not have
Ludovic> any objection to libgnat being GPL for non-customers.  There
Ludovic> seems to be quite strong feelings the opposite way, and even
Ludovic> some frustration.  These feelings and frustration do not have
Ludovic> to be.

Do you think that the DoD (the funder of the early GNAT versions)
intent was to restrict the use of GNAT to programs using the GPL
license and to AdaCore (which didn't exist) customers?

The real problem is not technical. It is still possible to build a
"clean" compiler which can be used on any kind of sources. The problem
is political. AdaCore's move causes a lot of confusion in the
community and may make companies unsure of what they can and can't do.

I think it would have been much better if AdaCore had kept the GMGPL
version, or if they had not complicated everything by providing
GPL-only packages. The situation was simple (be an AdaCore customer or
get your compiler from another source such as a GNU/Linux distribution
or build it from FSF sources), it is now ugly (in some cases, you can
get a GNAT compiler which doesn't allow you to use a GPL-incompatible
license for your source code if you intend to redistribute the
result).

I still fail to understand this tactical move.

  Sam
-- 
Samuel Tardieu -- sam@rfc1149.net -- http://www.rfc1149.net/sam

Re: GNAT GPL 2005 Edition is now available

[Ludovic Brenta]

Samuel Tardieu a écrit :
> Do you think that the DoD (the funder of the early GNAT versions)
> intent was to restrict the use of GNAT to programs using the GPL
> license and to AdaCore (which didn't exist) customers?

I think the DoD didn't care, because it had the money to pay for
any license it required.  Or maybe they'd have insisted on GPL
for everyone except the DoD, for fear that the Russians would build
military software with GNAT :)

By the same token, I don't think that current paying customers of
AdaCore care.  I think that the only ones who are really hurt by
AdaCore's decision are the SMEs and individuals who want to write
commercial software in Ada, but don't have the money to pay for
AdaCore's (or other companies') support contracts.  Like I said,
these non-customers of AdaCore's are a potential market.

> The real problem is not technical. It is still possible to build a
> "clean" compiler which can be used on any kind of sources. The problem
> is political. AdaCore's move causes a lot of confusion in the
> community and may make companies unsure of what they can and can't do.

I agree that there is confusion.  A lot of companies would like "free
software" to be free for companies; it isn't.  It is "free" for end
users.  It gives freedoms to end users and imposes requirements on
companies or producers of software.  Additional confusion is fueled
by "open source" (business- friendly attitude) and by the multiplicity
of licenses.

> I think it would have been much better if AdaCore had kept the GMGPL
> version, or if they had not complicated everything by providing
> GPL-only packages. The situation was simple (be an AdaCore customer or
> get your compiler from another source such as a GNU/Linux distribution
> or build it from FSF sources), it is now ugly (in some cases, you can
> get a GNAT compiler which doesn't allow you to use a GPL-incompatible
> license for your source code if you intend to redistribute the
> result).

Yes, I agree it is ugly to have multiple sources for the same software
under different licenses.   I hope that once the vote is over (next
Tuesday), the ugliness will be dealt with for good, at least as far as
Debian is concerned.  BTW, Tapio Kelloniemi suggested providing both
GNAT GPL 2005 Edition and GCC 4.0 in Debian.  I will not support two
different GNATs in Debian, especially if they have different licenses.

> I still fail to understand this tactical move.

Me too, but as I said, AdaCore is under no obligation to explain.  If
they care to explain, I'm all ears, of course.

My message is that nobody has a right to *complain* about AdaCore's
decision, or even demand an explanation; but people should decide for
themselves what to do about it.  The vote is designed to do just that.

I anticipate that (1) this public vote will show that a majority of
software developers reject the GPL for libgnat, (2) Debian will not
include GNAT GPL 2005 Edition, (3) no other distribution will, and
(4) AdaCore will probably take that into account when they think about
a new public release.

-- 
Ludovic Brenta.

Re: GNAT GPL 2005 Edition is now available

[Niklas Holsti]

Ludovic Brenta wrote:
> ...
> I think that the only ones who are really hurt by
> AdaCore's decision are the SMEs and individuals who want to write
> commercial software in Ada, but don't have the money to pay for
> AdaCore's (or other companies') support contracts.

I am in exactly this position. Of course AdaCore have the right to 
define the licenses for what they own and distribute. I think GNAT 
GPL 2005 Edition, as such, will be a good thing for Ada users who 
accept the GPL.

However, I worry about what this move by AdaCore implies for the 
future existence of a GMGPL GNAT (the FSF one, I guess). Does it 
make it more likely that AdaCore will reduce their support of such 
a GNAT? I hope that AdaCore will say something about that, in 
response to this discussion on c.l.a.

I always understood that the rationale for a free Ada compiler was 
to increase the respect for Ada and the usage of Ada overall, and 
so build a market for commercial compilers for those who need 
them. I believe GNAT has succeeded very well in this. The question 
is if AdaCore now have a different perception, or think that the 
middle ground (GMGPL) is not important in this market-building 
process.

My guess is that good and successful commercial software products 
built with Ada are better advertisements for Ada than GPL'ed Ada 
programs are, from the point of view of potential buyers of Ada 
compilers, because such products show that using Ada can be 
profitable. It is understandable and acceptable that AdaCore want 
some of that profit, but the current entry price is too high at 
least for me. Chicken and egg, again...

-- 
Niklas Holsti
Tidorum Ltd
niklas holsti tidorum fi
       .      @       .

Re: GNAT GPL 2005 Edition is now available

[Georg Bauhaus]

Ludovic Brenta wrote:

> I anticipate that (1) this public vote will show that a majority of
> software developers reject the GPL for libgnat, 


I take it that this vote does show that individuals who

(a) read c.l.ada (frequently)
(b) are software professionals, mostly
(c) have known commercial/licensing needs

will prefer a compiler + libraries supporting their current
or unstated project needs. I'm a Debian fan and want to
continue using Ada 2005 features which have facilitated
and simplified programming for me. The GNAT GPL 2005 edition is
easy to install und unintrusive, leaving room for a Debian
system compiler beside it.

But my vote is from the film "The Mouse that Roared" (18).
"I think we should wait."

Re: GNAT GPL 2005 Edition is now available

[Ludovic Brenta]

Georg Bauhaus <bauhaus@futureapps.de> writes:
> Ludovic Brenta wrote:
>
>> I anticipate that (1) this public vote will show that a majority of
>> software developers reject the GPL for libgnat,
>
>
> I take it that this vote does show that individuals who
>
> (a) read c.l.ada (frequently)
> (b) are software professionals, mostly
> (c) have known commercial/licensing needs
>
> will prefer a compiler + libraries supporting their current
> or unstated project needs. I'm a Debian fan and want to
> continue using Ada 2005 features which have facilitated
> and simplified programming for me. The GNAT GPL 2005 edition is
> easy to install und unintrusive, leaving room for a Debian
> system compiler beside it.
>
> But my vote is from the film "The Mouse that Roared" (18).
> "I think we should wait."

Wait for what?

Of course we are in no hurry to make a decision; the release date for
Etch has not yet been defined.  OTOH, I have received so many comments
to the effect that gnat 3.15p is too old that Debian should move on
that I take the release of GNAT GPL 2005 Edition as an opportunity to
do that.

-- 
Ludovic Brenta.

Re: GNAT GPL 2005 Edition is now available

[Tapio Kelloniemi]

"Ludovic Brenta" <ludovic@ludovic-brenta.org> wrote:
>Samuel Tardieu a =E9crit :
>By the same token, I don't think that current paying customers of
>AdaCore care.  I think that the only ones who are really hurt by
>AdaCore's decision are the SMEs and individuals who want to write
>commercial software in Ada, but don't have the money to pay for
>AdaCore's (or other companies') support contracts. 

Plus those who develop free software, but need to publish their work under
another license so that other free software can link against it. These
situations are, however, quite rare.

>> The real problem is not technical. It is still possible to build a
>> "clean" compiler which can be used on any kind of sources. The problem
>> is political. AdaCore's move causes a lot of confusion in the
>> community and may make companies unsure of what they can and can't do.
>
>I agree that there is confusion.  A lot of companies would like "free
>software" to be free for companies; it isn't.

And it must not be, and neither it must be free for any other closed-source
developers, who think that they can get money from something (or using
something) that others give for free.

>> I think it would have been much better if AdaCore had kept the GMGPL
>> version, or if they had not complicated everything by providing
>> GPL-only packages. The situation was simple (be an AdaCore customer or
>> get your compiler from another source such as a GNU/Linux distribution
>> or build it from FSF sources), it is now ugly (in some cases, you can
>> get a GNAT compiler which doesn't allow you to use a GPL-incompatible
>> license for your source code if you intend to redistribute the
>> result).
>
>Yes, I agree it is ugly to have multiple sources for the same software
>under different licenses.   I hope that once the vote is over (next
>Tuesday), the ugliness will be dealt with for good, at least as far as
>Debian is concerned.  BTW, Tapio Kelloniemi suggested providing both
>GNAT GPL 2005 Edition and GCC 4.0 in Debian.  I will not support two
>different GNATs in Debian, especially if they have different licenses.

I did not recommend supporting two editions, but rather including two
editions (one super-stable for free software development) and another
because it comes with the same GCC as the other system's compilers. Currently
Debian includes three(?) GNAT versions (3.15, 3.4 and 4.0). Including
GNAT-4.0 just because it does not have those special exceptions has no
sense. Free software is always better than non-free, and if GNAT-4.0 is
really buggy and no one has time to backport fixes and features from 4.1,
I recommend including the super-stable and complete GNAT GPL.

>> I still fail to understand this tactical move.
>
>Me too, but as I said, AdaCore is under no obligation to explain.  If
>they care to explain, I'm all ears, of course.

If I were AdaCore I had done the same (but long time ago), so that SMEs
and greedy individuals would not use something written by me to collect
money for themselves. All of my software is GPL to prevent someone from
using it in non-free projects. Free software is about giving without
losing, non-free software has very different goals and it should not be
built using free components, because non-free software developers don't
want to help their neighbours, as free software developers do.

>My message is that nobody has a right to *complain* about AdaCore's
>decision, or even demand an explanation; but people should decide for
>themselves what to do about it.  The vote is designed to do just that.

Many people have said that what AdaCore has done is not good for the image
of Ada. I ask, isn't that bad for Ada's image that although new Ada2005
features are available in GNAT GPL, they are not provided for Free
Software developers (in Debian or other distros). Generally people are
lazy (or unable) to install software themselves (even from third-party
provided binaries) and when searching for a language, they just choose C++
instead, since the most featureful (and stable) Ada compiler is not
available. Those who have voted against GNAT GPL seem not to be mostly
free software developers, but I think that Debian should consider
free software developers as the primary target of their distribution, since
its free.

-- 
Tapio

Re: GNAT GPL 2005 Edition is now available

[Marc A. Criley]

Tapio Kelloniemi wrote:

> Many people have said that what AdaCore has done is not good for the image
> of Ada. I ask, isn't that bad for Ada's image that although new Ada2005
> features are available in GNAT GPL, they are not provided for Free
> Software developers (in Debian or other distros). Generally people are
> lazy (or unable) to install software themselves (even from third-party
> provided binaries) and when searching for a language, they just choose C++
> instead, since the most featureful (and stable) Ada compiler is not
> available. Those who have voted against GNAT GPL seem not to be mostly
> free software developers, but I think that Debian should consider
> free software developers as the primary target of their distribution, since
> its free.

I think there are two motivations at work here.  One is the Stallman/FSF 
philosophy of Free (libre) software, as licensed by the GPL.  The other, 
here in this newsgroup anyway, is Ada advocacy.

Now I'm all for the GPL, I wish all the software in the world was under 
the GPL.  But it isn't, and truth be told, of the multi-billion dollar 
software industry very little of that software is GPLed.  (I don't want 
to haggle over the percentage, but think about this:  what is the nature 
of the licensing of most of the software that most businesses, your 
friends, and family are using?  And where are they getting it?  I doubt 
it's GPL, and I doubt they're downloading it from CVS repositories.)

So if you want to market a software product, you have to go to where the 
money is, and that means customers with proprietary concerns.

Trying to sell them a software tool or utility, or give it to them and 
sell support, that would force them to GPL their own code is just going 
to be a non-starter in an overwhelming number of instances.

So if you want to sell them an Ada tool, to enhance or encourage a 
customer company's use of Ada, the last thing you want to do is tell 
them they have to change or set the licensing of their products to be 
GPL compatible.  They're not going to do that, so they're not going to 
gain the advantages of your product, and so why bother with Ada anyway 
any more, since Visual Studio C++ or C# or Java doesn't force them to do 
anything like that?

GPL-GNAT appears to be promoting the Free agenda over the Ada advocacy 
agenda, which is the problem the OMSs (One-Man-Shops) here are having 
with it.  It's hard enough selling just Ada, without also having to sell 
the GPL to commercial, proprietary companies.

-- Marc A. Criley
-- McKae Technologies
-- www.mckae.com
-- DTraq - XPath In Ada - XML EZ Out

Re: GNAT GPL 2005 Edition is now available

[Martin Dowie]

Marc A. Criley wrote:
[snip]
> GPL-GNAT appears to be promoting the Free agenda over the Ada advocacy 
> agenda, which is the problem the OMSs (One-Man-Shops) here are having 
> with it.  It's hard enough selling just Ada, without also having to sell 
> the GPL to commercial, proprietary companies.

RR Software's Janus/Ada95 and Aonix's ObjectAda are within the price 
range of a one-man-shop.

Cheers

-- Martin

Re: GNAT GPL 2005 Edition is now available

[Jacob Sparre Andersen]

Martin Dowie wrote:

> RR Software's Janus/Ada95 and Aonix's ObjectAda are within the price
> range of a one-man-shop.

I don't doubt that (I've actually spent money on an ObjectAda license
once), but it appears that neither of them is available for the
platforms I work on today. :(

I will have to manage with GNAT 3.15p until I (or somebody else) can
find time to get my hands dirty on the GNAT source code.

Greetings,

Jacob
-- 
"Banning open source would have immediate, broad, and
 strongly negative impacts on the ability of many sensitive
 and security-focused DOD groups to protect themselves
 against cyberattacks"                        -- Mitre Corp.

Re: GNAT GPL 2005 Edition is now available

[Georg Bauhaus]

Jacob Sparre Andersen wrote:
> Martin Dowie wrote:
> 
> 
>>RR Software's Janus/Ada95 and Aonix's ObjectAda are within the price
>>range of a one-man-shop.
> 
> 
> I don't doubt that (I've actually spent money on an ObjectAda license
> once), but it appears that neither of them is available for the
> platforms I work on today. :(

If your platform happens to be running Linux, there is a Linux
edition of ObjectAda. But I don't know whether the Linux edition
is a non-Enterprice edition, like the Windows professional edition.

Re: GNAT GPL 2005 Edition is now available

[Jacob Sparre Andersen]

Georg Bauhaus wrote:

> If your platform happens to be running Linux, there is a Linux
> edition of ObjectAda. But I don't know whether the Linux edition is
> a non-Enterprice edition, like the Windows professional edition.

My workstation does happen to run Linux (Debian/unstable on PPC) and
so do my primary target platforms (various Linux distributions on PPC
and IA32).  I'll take a look at the Linux edition of ObjectAda.

Thanks.

Jacob
-- 
"Those who will not reason, are bigots,
 those who cannot, are fools, and
 those who dare not, are slaves."

Re: GNAT GPL 2005 Edition is now available

[Martin Krischik]

Jacob Sparre Andersen wrote:

> I will have to manage with GNAT 3.15p until I (or somebody else) can
> find time to get my hands dirty on the GNAT source code.

Actually it is not that difficult:

http://ada.krischik.com/index.php/Articles/CompileGcc

just time and space consuming.

Martin

-- 
mailto://krischik@users.sourceforge.net
Ada programming at: http://ada.krischik.com

Re: GNAT GPL 2005 Edition is now available

[Jacob Sparre Andersen]

Martin Krischik wrote:
> Jacob Sparre Andersen wrote:

>> I will have to manage with GNAT 3.15p until I (or somebody else)
>> can find time to get my hands dirty on the GNAT source code.
>
> Actually it is not that difficult:
>
> http://ada.krischik.com/index.php/Articles/CompileGcc
>
> just time and space consuming.

It was not compiling, but fixing/improving GNAT, I was worried about.
Ludovic Brenta does a great job of compiling GNAT (and I even tried it
myself once).  What I am worried about is fixing errors in GNAT and
extending GNAT to handle Ada2005.  I am seriously afraid that _that_
is difficult.  So far I've only sketched fixes for some errors in the
GNAT implementation of the standard library.  The actual compiler
source code is huge enough to scare me away.

Jacob
-- 
"Nobody writes jokes in base 13."
 Douglas Adams

Re: GNAT GPL 2005 Edition is now available

[Ludovic Brenta]

Jacob Sparre Andersen <sparre@nbi.dk> writes:
> It was not compiling, but fixing/improving GNAT, I was worried
> about.  Ludovic Brenta does a great job of compiling GNAT (and I
> even tried it myself once).  What I am worried about is fixing
> errors in GNAT and extending GNAT to handle Ada2005.  I am seriously
> afraid that _that_ is difficult.  So far I've only sketched fixes
> for some errors in the GNAT implementation of the standard library.
> The actual compiler source code is huge enough to scare me away.

Yes, fixing and extending GNAT is difficult and requires huge amounts
of time to build up the expertise.  Fixing and extending libgnat
requires less time and expertise.  If you are contemplating a fork,
you can restrict the fork to just libgnat, since the compiler's
license poses no problem to anyone.

BTW, thanks for the compliments.

-- 
Ludovic Brenta.

Re: GNAT GPL 2005 Edition is now available

[Ludovic Brenta]

Ludovic Brenta <ludovic@ludovic-brenta.org> writes:

> Jacob Sparre Andersen <sparre@nbi.dk> writes:
>> It was not compiling, but fixing/improving GNAT, I was worried
>> about.  Ludovic Brenta does a great job of compiling GNAT (and I
>> even tried it myself once).  What I am worried about is fixing
>> errors in GNAT and extending GNAT to handle Ada2005.  I am seriously
>> afraid that _that_ is difficult.  So far I've only sketched fixes
>> for some errors in the GNAT implementation of the standard library.
>> The actual compiler source code is huge enough to scare me away.
>
> Yes, fixing and extending GNAT is difficult and requires huge amounts
> of time to build up the expertise.  Fixing and extending libgnat
> requires less time and expertise.  If you are contemplating a fork,
> you can restrict the fork to just libgnat, since the compiler's
> license poses no problem to anyone.
>
> BTW, thanks for the compliments.

I'd like to add this for the record: I do not encourage anyone to
consider a fork.  For one thing, I don't feel inclined to encourage
people to write non-free software using free libraries like libgnat
(but this is debatable and is being debated right now).  For another
thing, a fork is not currently required anyway because the FSF tree is
still GMGPL.

My comments above were purely technical.

-- 
Ludovic Brenta.

Re: GNAT GPL 2005 Edition is now available

[Björn Persson]

Ludovic Brenta wrote:
> Yes, fixing and extending GNAT is difficult and requires huge amounts
> of time to build up the expertise.  Fixing and extending libgnat
> requires less time and expertise.  If you are contemplating a fork,
> you can restrict the fork to just libgnat, since the compiler's
> license poses no problem to anyone.

The standard library packages may not be so hairy, but what about the 
runtime? I would have thought it would be difficult to understand and 
modify. I also expect it to be tightly coupled to the compiler so that 
it would be hard to update it without looking at AdaCore's GPL code.

(Does libgnat contain both the library packages and the runtime?)

-- 
Bj�rn Persson                              PGP key A88682FD
                    omb jor ers @sv ge.
                    r o.b n.p son eri nu

Re: GNAT GPL 2005 Edition is now available

[Ludovic Brenta]

Björn Persson <spam-away@nowhere.nil> writes:
> Ludovic Brenta wrote:
>> Yes, fixing and extending GNAT is difficult and requires huge amounts
>> of time to build up the expertise.  Fixing and extending libgnat
>> requires less time and expertise.  If you are contemplating a fork,
>> you can restrict the fork to just libgnat, since the compiler's
>> license poses no problem to anyone.
>
> The standard library packages may not be so hairy, but what about
> the runtime? I would have thought it would be difficult to
> understand and modify. I also expect it to be tightly coupled to the
> compiler so that it would be hard to update it without looking at
> AdaCore's GPL code.

The coupling with the compiler is no problem from a legal standpoint,
because the compiler's license has no effect on the executables
produced with it.  Porting libgnat to a new platform is difficult, but
not quite as difficult as maintaining the compiler.  In terms of
technical difficulty, and broadly speaking:

standard packages < run-time < compiler

> (Does libgnat contain both the library packages and the runtime?)

Yes, libgnat contains the System.* and Ada.* packages, the GNAT.*
packages and any supporting run-time packages necessary for the target
platform; in particular, support for interrupts, tasking, and
exceptions.

-- 
Ludovic Brenta.

Re: GNAT GPL 2005 Edition is now available

[Björn Persson]

Ludovic Brenta wrote:
> The coupling with the compiler is no problem from a legal standpoint,
> because the compiler's license has no effect on the executables
> produced with it.

As you probably know, if you read someone else's source code and then 
write down what you've read and distribute it with a different license, 
then you infringe on the original author's copyright. Therefore, someone 
who's maintaining a GMGPL fork of the Gnat runtime should be careful to 
avoid looking at any GPL version of the runtime. Because of the tight 
coupling, I assume that the fork would have to be updated to exactly 
match each version of the compiler. Because it would have to match 
exactly, I think this would be difficult to do without looking at 
AdaCore's changes. I'm not sure whether it would be legally safe to read 
the sources of the compiler to figure out what would need to be changed 
in the runtime. A lawyer might argue that you've transferred 
"intellectual property" from the compiler into the runtime. If one 
couldn't look at the compiler sources, it would be even more difficult 
to update the runtime. That's why I thought the tight coupling would be 
a problem.

On the other hand, I've never quite understood what makes a runtime 
library different from any other library. Maybe a runtime isn't quite as 
magic as I think.

Anyway, let's hope we'll never have to deal with these problems in practice.

-- 
Björn Persson                              PGP key A88682FD
                    omb jor ers @sv ge.
                    r o.b n.p son eri nu

Re: GNAT GPL 2005 Edition is now available

[tmoran]

> As you probably know, if you read someone else's source code and then
> write down what you've read and distribute it with a different license,
> then you infringe on the original author's copyright. Therefore, someone
> who's maintaining a GMGPL fork of the Gnat runtime should be careful to
> avoid looking at any GPL version of the runtime. Because of the tight
   From "The Software Developer's and Marketer's Legal Companion"
"Software copyright protection does not cover functional elements that
are inherent in the idea of a particular application."
   So if you look at their patches and see that one consists of
initializing a previously uninitialized variable, I would expect it would
be completely legitimate for you to say "ah, the bug was the uninitialized
variable so I will insert code, not copied from them, whose function is to
initialize it."  If one person analyzed their code to figure out what
function it performs, then told a coder to implement that function, that
would be OK.  But then IANAL.

Re: GNAT GPL 2005 Edition is now available

[Brian May]

>>>>> "tmoran" == tmoran  <tmoran@acm.org> writes:

    tmoran>    From "The Software Developer's and Marketer's Legal
    tmoran> Companion" "Software copyright protection does not cover
    tmoran> functional elements that are inherent in the idea of a
    tmoran> particular application."  So if you look at their patches
    tmoran> and see that one consists of initializing a previously
    tmoran> uninitialized variable, I would expect it would be
    tmoran> completely legitimate for you to say "ah, the bug was the
    tmoran> uninitialized variable so I will insert code, not copied
    tmoran> from them, whose function is to initialize it."  If one
    tmoran> person analyzed their code to figure out what function it
    tmoran> performs, then told a coder to implement that function,
    tmoran> that would be OK.  But then IANAL.

Would it be OK to run the test suite (or whatever it is called) from
the GPL version on your new non-GPL version to ensure it is "correct"?
-- 
Brian May <bam@snoopy.apana.org.au>

Re: GNAT GPL 2005 Edition is now available

[Florian Weimer]

* Jacob Sparre Andersen:

> What I am worried about is fixing errors in GNAT and extending GNAT
> to handle Ada2005.  I am seriously afraid that _that_ is difficult.

The real problem we face in GCC 4.0 is that GIGI, the converter from
Ada's intermediate language to what is nowadays GNEERIC or GIMPLE on
the GCC side, hasn't been updated since 1997 or so.  This means that a
lot of things work by accident only.

Re: GNAT GPL 2005 Edition is now available

[Marc A. Criley]

Martin Dowie wrote:

> RR Software's Janus/Ada95 and Aonix's ObjectAda are within the price 
> range of a one-man-shop.

Janus Ada certainly is at $295, but doesn't include ASIS, and doesn't come
right out and say it supports Linux.  The Janus/Ada FAQ says: "Janus/Ada 95
for Unix conforms to the Intel ABI for Unix. Systems that are known to
conform are SCO UNIX and older versions of Sun Solaris for Intel
processors. Other systems claim to support the ABI, including some of the
free systems."

ObjectAda is a bit pricier:  Windows is $1495, then $1495/yr maintenance;
Linux is $5000 upfront, then $1000/yr maintenance.  ObjectAda is also
available on a variety of Unix platforms, but I don't have pricing for those.

RR is certainly in the range of most developers, but you probably gotta be
a serious OMS to justify ObjectAda.

-- Marc A. Criley
-- McKae Technologies
-- www.mckae.com
-- DTraq - XPath In Ada - XML EZ Out

Re: GNAT GPL 2005 Edition is now available

[Martin Dowie]

Marc A. Criley wrote:
> Martin Dowie wrote:
> 
>> RR Software's Janus/Ada95 and Aonix's ObjectAda are within the price 
>> range of a one-man-shop.
> 
> 
> Janus Ada certainly is at $295, but doesn't include ASIS, and doesn't come
> right out and say it supports Linux.  The Janus/Ada FAQ says: "Janus/Ada 95
> for Unix conforms to the Intel ABI for Unix. Systems that are known to
> conform are SCO UNIX and older versions of Sun Solaris for Intel
> processors. Other systems claim to support the ABI, including some of the
> free systems."

Yeah, ASIS is _always_ nice to have, no matter what the compiler... I 
wish compiler vendors would just package it - it has to be in their own 
long term interest to do so...


> ObjectAda is a bit pricier:  Windows is $1495, then $1495/yr maintenance;
> Linux is $5000 upfront, then $1000/yr maintenance.  ObjectAda is also
> available on a variety of Unix platforms, but I don't have pricing for 
> those.

That sounds pricey, I have an ObjectAda/Windows Pro license and this 
seems to be the first case _ever_ where the price in the UK is less than 
in the US!


> RR is certainly in the range of most developers, but you probably gotta be
> a serious OMS to justify ObjectAda.

Or very keen about keeping Ada alive! :-)

Cheers

-- Martin

Re: GNAT GPL 2005 Edition is now available

[Martin Dowie]

Marc A. Criley wrote:
[snip]
> Janus Ada certainly is at $295, but doesn't include ASIS, and doesn't come
[snip]

Just noticed that Janus/Ada95 for WinNT/95 is $195 - the $295 version 
includes Claw (well worth the money! :-)

Cheers

-- Martin

Re: GNAT GPL 2005 Edition is now available

[Marc A. Criley]

Marc A. Criley wrote:
> 
> Janus Ada certainly is at $295, but doesn't include ASIS, and doesn't come
> right out and say it supports Linux.  The Janus/Ada FAQ says: "Janus/Ada 95
> for Unix conforms to the Intel ABI for Unix. Systems that are known to
> conform are SCO UNIX and older versions of Sun Solaris for Intel
> processors. Other systems claim to support the ABI, including some of the
> free systems."
> 
> ObjectAda is a bit pricier:  Windows is $1495, then $1495/yr maintenance;
> Linux is $5000 upfront, then $1000/yr maintenance.  ObjectAda is also
> available on a variety of Unix platforms, but I don't have pricing for 
> those.

Finally got some reportedly recent AdaCore pricing for GNAT Pro.  If 
you're a "5-Man Shop" it's not tooooo bad:  "An annual subscription for 
a five-user team starts at $14,000."

(http://www.elecdesign.com/Articles/Index.cfm?AD=1&AD=1&ArticleID=10992)

-- Marc A. Criley
-- McKae Technologies
-- www.mckae.com
-- DTraq - XPath In Ada - XML EZ Out

Re: GNAT GPL 2005 Edition is now available

[Preben Randhol]

In article Martin Dowie wrote:
>Marc A. Criley wrote:
>[snip]
>> GPL-GNAT appears to be promoting the Free agenda over the Ada advocacy 
>> agenda, which is the problem the OMSs (One-Man-Shops) here are having 
>> with it.  It's hard enough selling just Ada, without also having to sell 
>> the GPL to commercial, proprietary companies.
>
>RR Software's Janus/Ada95 and Aonix's ObjectAda are within the price 
>range of a one-man-shop.

Do they run on Linux?

Preben

Re: GNAT GPL 2005 Edition is now available

[Niklas Holsti]

Preben Randhol wrote:
> In article Martin Dowie wrote:
> 
>>Marc A. Criley wrote:
>>[snip]
>>
>>RR Software's Janus/Ada95 and Aonix's ObjectAda are within the price 
>>range of a one-man-shop.
> 
> 
> Do they run on Linux?

According to very recent information from RR Software, Janus/Ada95 
does not (now) run on Linux.

-- 
Niklas Holsti
Tidorum Ltd
niklas holsti tidorum fi
       .      @       .

Re: GNAT GPL 2005 Edition is now available

[Martin Dowie]

Preben Randhol wrote:
> In article Martin Dowie wrote:
>> Marc A. Criley wrote:
>> [snip]
>>> GPL-GNAT appears to be promoting the Free agenda over the Ada
>>> advocacy agenda, which is the problem the OMSs (One-Man-Shops) here
>>> are having with it.  It's hard enough selling just Ada, without
>>> also having to sell the GPL to commercial, proprietary companies.
>>
>> RR Software's Janus/Ada95 and Aonix's ObjectAda are within the price
>> range of a one-man-shop.
>
> Do they run on Linux?

ObjectAda has a linux version - don't know the price though.

See http://www.aonix.com/pr_09.12.05c.html

Re: GNAT GPL 2005 Edition is now available

[Niklas Holsti]

Martin Dowie wrote:
> Preben Randhol wrote:
> 
>>In article Martin Dowie wrote:
>>
>>>Marc A. Criley wrote:
>>>[snip]
>>>
>>>>GPL-GNAT appears to be promoting the Free agenda over the Ada
>>>>advocacy agenda, which is the problem the OMSs (One-Man-Shops) here
>>>>are having with it.  It's hard enough selling just Ada, without
>>>>also having to sell the GPL to commercial, proprietary companies.
>>>
>>>RR Software's Janus/Ada95 and Aonix's ObjectAda are within the price
>>>range of a one-man-shop.
>>
>>Do they run on Linux?
> 
> 
> ObjectAda has a linux version - don't know the price though.
> 
> See http://www.aonix.com/pr_09.12.05c.html

 From a recent press release: "ObjectAda for Linux is available 
immediately for Red Hat Enterprise Version 4 and Fedora Core 
Version 4 and starts at $5,000".

-- 
Niklas Holsti
Tidorum Ltd
niklas holsti tidorum fi
       .      @       .

Re: GNAT GPL 2005 Edition is now available

[Martin Dowie]

Niklas Holsti wrote:
>  From a recent press release: "ObjectAda for Linux is available
> immediately for Red Hat Enterprise Version 4 and Fedora Core
> Version 4 and starts at $5,000".

Shame they don't do a 'Pro' edition for $$$ as they do for Windows.

I'm surprised there isn't a 'zero' support, dirty cheap version of these
compilers - it's not a huge market, but they could just send out the
same CD but filter on license key as to whether, if you ask for it, you
are entitled to any support.

Cheers

-- Martin

Re: GNAT GPL 2005 Edition is now available

[Alex R. Mosteo]

Martin Dowie wrote:
> Niklas Holsti wrote:
> 
>> From a recent press release: "ObjectAda for Linux is available
>>immediately for Red Hat Enterprise Version 4 and Fedora Core
>>Version 4 and starts at $5,000".
> 
> 
> Shame they don't do a 'Pro' edition for $$$ as they do for Windows.
> 
> I'm surprised there isn't a 'zero' support, dirty cheap version of these
> compilers - it's not a huge market, but they could just send out the
> same CD but filter on license key as to whether, if you ask for it, you
> are entitled to any support.

And they would get more betatesters.

Re: GNAT GPL 2005 Edition is now available

[Jeffrey Carter]

Martin Dowie wrote:
> I'm surprised there isn't a 'zero' support, dirty cheap version of these
> compilers - it's not a huge market, but they could just send out the
> same CD but filter on license key as to whether, if you ask for it, you
> are entitled to any support.

There's a free version of ObjectAda for Windows. It has limits on the 
number of units in a project.

-- 
Jeffrey Carter
"Now go away or I shall taunt you a second time."
Monty Python and the Holy Grail
E-mail: jeffrey_r_carter-nr [commercial-at]
         raytheon [period | full stop] com

Re: GNAT GPL 2005 Edition is now available

[Martin Dowie]

Jeffrey Carter wrote:
> Martin Dowie wrote:
> 
>> I'm surprised there isn't a 'zero' support, dirty cheap version of these
>> compilers - it's not a huge market, but they could just send out the
>> same CD but filter on license key as to whether, if you ask for it, you
>> are entitled to any support.
> 
> 
> There's a free version of ObjectAda for Windows. It has limits on the 
> number of units in a project.

Yeah, I know, but it's limits make it unusable for anything other than 
learning Ada with small examples. :-(

-- Martin

Re: GNAT GPL 2005 Edition is now available

[Randy Brukardt]

"Martin Dowie" <martin.dowie@baesystems.com> wrote in message
news:43313bf2$1_1@glkas0286.greenlnk.net...
> Niklas Holsti wrote:
> >  From a recent press release: "ObjectAda for Linux is available
> > immediately for Red Hat Enterprise Version 4 and Fedora Core
> > Version 4 and starts at $5,000".
>
> Shame they don't do a 'Pro' edition for $$$ as they do for Windows.
>
> I'm surprised there isn't a 'zero' support, dirty cheap version of these
> compilers - it's not a huge market, but they could just send out the
> same CD but filter on license key as to whether, if you ask for it, you
> are entitled to any support.

Well, speaking for RRS, there are a number of issues:
(1) There really isn't any such thing as a 'zero' support product. Problems
that people have using your products always reflect back on you, and people
rarely if ever prefix rants with "well, I got it for free, but...".
Moreover, if there is a security problem or the like, you'll pretty much
*have* to support it.
(2) Janus/Ada's packaging needs a complete overhaul; it was designed in the
mid 90's for floppy disk/printed manual distribution, and those days are
gone. Installation takes too much hand-holding right now. (And I personally
have been spending way too much time on Ada 200Y to address this; that will
end soon, of course.)
(3) The other issue is simply making enough money from the work to justify
the effort. Our experience (admittedly mostly from the 1980's) is that
compiler demand is rather inelastic. We have relatively low prices because
we believe in the small developer. But if we wanted to maximize our revenue,
pricing more like Aonix's would probably work better. Our current prices are
about as low as we can go, I think.

                         Randy Brukardt, R.R. Software, Inc.

Re: GNAT GPL 2005 Edition is now available

[Marc A. Criley]

Preben Randhol wrote:
> In article Martin Dowie wrote:
> 
>>RR Software's Janus/Ada95 and Aonix's ObjectAda are within the price 
>>range of a one-man-shop.
> 
> 
> Do they run on Linux?

I posted this previously in this huge thread we've got going here, so 
it's easy to miss :-)


Marc A. Criley wrote:

 >
 > Janus Ada certainly is at $295, but doesn't include ASIS, and doesn't 
come
 > right out and say it supports Linux.  The Janus/Ada FAQ says: 
"Janus/Ada 95
 > for Unix conforms to the Intel ABI for Unix. Systems that are known to
 > conform are SCO UNIX and older versions of Sun Solaris for Intel
 > processors. Other systems claim to support the ABI, including some of the
 > free systems."
 >
 > ObjectAda is a bit pricier:  Windows is $1495, then $1495/yr maintenance;
 > Linux is $5000 upfront, then $1000/yr maintenance.  ObjectAda is also
 > available on a variety of Unix platforms, but I don't have pricing 
for those.


Finally got some reportedly recent AdaCore pricing for GNAT Pro.  If 
you're a "5-Man Shop" it's not tooooo bad:  "An annual subscription for 
a five-user team starts at $14,000."

(http://www.elecdesign.com/Articles/Index.cfm?AD=1&AD=1&ArticleID=10992)

-- Marc A. Criley
-- McKae Technologies
-- www.mckae.com
-- DTraq - XPath In Ada - XML EZ Out

Re: GNAT GPL 2005 Edition is now available

[chrismiller677]

>
> Trying to sell them a software tool or utility, or give it to them and
> sell support, that would force them to GPL their own code is just going
> to be a non-starter in an overwhelming number of instances.
>

I agree with this 100% but believe the problem is with the Free
Software Foundation, not AdaCore.

The situation where there are multiple licences, created by multiple
companies or organisations is an unstainable mess. What, I believe,
AdaCore have done is to say that they are not in the business of
creating the "free software" rules and that they will simply release
Gnat under what ever rules the FSF comes up with. I believe this is a
correct long term decision.

That said, the situation whereby that if you compile your code with a
FSF compiler that you must inturn release the code under their licence
is absurd. But it is up to the FSF to fix it, not AdaCore. It's not
hard to do, just change the licence rules.

Chris Miller

Re: GNAT GPL 2005 Edition is now available

[Brian May]

>>>>> "chrismiller677" == chrismiller677  <chrismiller677@hotmail.com> writes:

    chrismiller677> The situation where there are multiple licences,
    chrismiller677> created by multiple companies or organisations is
    chrismiller677> an unstainable mess. What, I believe, AdaCore have
    chrismiller677> done is to say that they are not in the business
    chrismiller677> of creating the "free software" rules and that
    chrismiller677> they will simply release Gnat under what ever
    chrismiller677> rules the FSF comes up with. I believe this is a
    chrismiller677> correct long term decision.

No, the FSF have two licenses that a relevant, the LGPL and the GPL.

The FSF license the C runtime libraries using the LGPL. I think
something similar has been done for the C++ runtime libraries too.

I fail to see why the Ada runtime library needs to be different.

So I don't think this is an issue, as the FSF has already found a
solution... Otherwise we would be having the same debate about gcc and
g++ on other forums.

Maybe the LGPL is "non-preferred", however I think this is a special
case situation that justifies it - not for the compiler, just the
run-time library.

The situation is also stupid, because I can write my own open source
software with the BSD license. I can distribute the source code under
the BSD license. However, as soon as I compile it using the compiler
in question, according to some interpretations of the GPL, I can not
distribute the resultant binary file. Instead I am forced to
distribute only the source code. This means end users must work out
how to compile my code. This in turn could give Ada a bad reputation;
eg.  "No don't use Brian's package - you have to install that stupid
Ada compiler to compile it - use this other package instead - it may
not be as good but at least you don't have to compile it."

Does restricting my from distributing binary files help AdaCore in
anyway?

I don't think so. It is still open source. I don't think the fact
somebody could later turn it into a closed source project at a latter
date is an issue (actually, I have been told by a lawyer it is legally
possible to completely revoke a license, including the GPL, on
distributed no-cost software, given notice, so this difference between
BSD and GPL may not actually exist).

Does it help promote the GPL in anyway?

No, the GPL is already my preferred license for projects I create, if
I use another license it would be because I had no option (e.g. I am
not the copyright owner).
-- 
Brian May <bam@snoopy.apana.org.au>

Re: GNAT GPL 2005 Edition is now available

[Martin Dowie]

Brian May wrote:
> The FSF license the C runtime libraries using the LGPL. I think
> something similar has been done for the C++ runtime libraries too.
>
> I fail to see why the Ada runtime library needs to be different.
>
> So I don't think this is an issue, as the FSF has already found a
> solution... Otherwise we would be having the same debate about gcc and
> g++ on other forums.
>
> Maybe the LGPL is "non-preferred", however I think this is a special
> case situation that justifies it - not for the compiler, just the
> run-time library.

Indeed this is what GNU suggest in the COPYING.txt file that comes with
GNAT:

   "This General Public License does not permit incorporating your program
into
   proprietary programs.  If your program is a subroutine library, you may
   consider it more useful to permit linking proprietary applications with
the
   library.  If this is what you want to do, use the GNU Library General
   Public License instead of this License."

Re: GNAT GPL 2005 Edition is now available

[Martin Dowie]

Martin Dowie wrote:
> Indeed this is what GNU suggest in the COPYING.txt file that comes
> with GNAT:
>
>    "This General Public License does not permit incorporating your
> program into
>    proprietary programs.  If your program is a subroutine library,
>    you may consider it more useful to permit linking proprietary
> applications with the
>    library.  If this is what you want to do, use the GNU Library
>    General Public License instead of this License."

More over, from the GNAT RM, section on "pragma License":

"Modified_GPL This is used for a unit licensed under the GNAT modified GPL
that includes a special exception paragraph that specifically permits the
inclusion of the unit in programs without requiring the entire program to be
released under the GPL. This is the license used for the GNAT run-time which
ensures that the run-time can be used freely in any program without GPL
concerns."

file:///C:/GNAT/doc/gnat/html/gnat_rm_2.html#SEC53

So, which one is it?!?!?! :-)

Cheers

-- Martin

Re: GNAT GPL 2005 Edition is now available

[Dirk Craeynest]

In article <sa4br2puwoo.fsf@snoopy.microcomaustralia.com.au>,
Brian May  <bam@snoopy.apana.org.au> wrote:
[...]
>The FSF license the C runtime libraries using the LGPL. I think
>something similar has been done for the C++ runtime libraries too.

This is incorrect.  The C++ runtime library is not distributed under
the LGPL, but under conditions identical to those that were used for
the GNAT 3.15p runtime library.

See <http://gcc.gnu.org/onlinedocs/libstdc++/17_intro/license.html>.

>I fail to see why the Ada runtime library needs to be different.

The situation became different with the GNAT GPL 2005 Edition.
Hence this thread...

Dirk

Re: GNAT GPL 2005 Edition is now available

[Georg Bauhaus]

Dirk Craeynest wrote:

>  The C++ runtime library is not distributed under
> the LGPL, but under conditions identical to those that were used for
> the GNAT 3.15p runtime library.
> 
> See <http://gcc.gnu.org/onlinedocs/libstdc++/17_intro/license.html>.
> 
> 
>>I fail to see why the Ada runtime library needs to be different.
> 
> 
> The situation became different with the GNAT GPL 2005 Edition.
> Hence this thread...

I don't think that the license used in the GNAT GPL 2005 Edition
has affected the license used in FSF GNAT.

Also, there is clause 10 of the GPL. Thus, if you have an
Open Source(tm) Ada program that is not GPLed, you could ask AdaCore
for permission to distribute your program using that other license.

Re: GNAT GPL 2005 Edition is now available

[Björn Persson]

Dirk Craeynest wrote:
> Brian May  <bam@snoopy.apana.org.au> wrote:
>>The FSF license the C runtime libraries using the LGPL. I think
>>something similar has been done for the C++ runtime libraries too.
> 
> This is incorrect.  The C++ runtime library is not distributed under
> the LGPL, but under conditions identical to those that were used for
> the GNAT 3.15p runtime library.

Which are *similar* to the LGPL, just like Brian said.

-- 
Bj�rn Persson                              PGP key A88682FD
                    omb jor ers @sv ge.
                    r o.b n.p son eri nu

Re: GNAT GPL 2005 Edition is now available

[Georg Bauhaus]

chrismiller677@hotmail.com wrote:

> That said, the situation whereby that if you compile your code with a
> FSF compiler that you must inturn release the code under their licence
> is absurd. But it is up to the FSF to fix it, not AdaCore.

In the case of FSF GNAT, there is nothing to fix because compiler
and run-time do not require releases under whatever license.
I think this is true for the other compilers in FSF GCC, too.

Re: GNAT GPL 2005 Edition is now available

[Preben Randhol]

In article <3467d$432b0af1$49956f8$22115@ALLTEL.NET>, Marc A. Criley wrote:
>I think there are two motivations at work here.  One is the Stallman/FSF 
>philosophy of Free (libre) software, as licensed by the GPL.  The other, 
>here in this newsgroup anyway, is Ada advocacy.
>
>Now I'm all for the GPL, I wish all the software in the world was under 
>the GPL.  But it isn't, and truth be told, of the multi-billion dollar 
>software industry very little of that software is GPLed.  (I don't want 
>to haggle over the percentage, but think about this:  what is the nature 
>of the licensing of most of the software that most businesses, your 
>friends, and family are using?  And where are they getting it?  I doubt 
>it's GPL, and I doubt they're downloading it from CVS repositories.)
>
>So if you want to market a software product, you have to go to where the 
>money is, and that means customers with proprietary concerns.
>
>Trying to sell them a software tool or utility, or give it to them and 
>sell support, that would force them to GPL their own code is just going 
>to be a non-starter in an overwhelming number of instances.
>
>So if you want to sell them an Ada tool, to enhance or encourage a 
>customer company's use of Ada, the last thing you want to do is tell 
>them they have to change or set the licensing of their products to be 
>GPL compatible.  They're not going to do that, so they're not going to 
>gain the advantages of your product, and so why bother with Ada anyway 
>any more, since Visual Studio C++ or C# or Java doesn't force them to do 
>anything like that?
>
>GPL-GNAT appears to be promoting the Free agenda over the Ada advocacy 
>agenda, which is the problem the OMSs (One-Man-Shops) here are having 
>with it.  It's hard enough selling just Ada, without also having to sell 
>the GPL to commercial, proprietary companies.

Yes, I agree 100%.

By the way if one has contributed patches for GtkAda does one have to be
consulted before a license change or can FSF/Adacore just change the
license from GMGPL to GPL?

Preben who has not been able to check comp.lang.ada lately and is
possibly a bit late in the debate...

Re: GNAT GPL 2005 Edition is now available

[Samuel Tardieu]

>>>>> "Preben" == Preben Randhol <randhol@bacchus.pvv.ntnu.no> writes:

Preben> By the way if one has contributed patches for GtkAda does one
Preben> have to be consulted before a license change or can
Preben> FSF/Adacore just change the license from GMGPL to GPL?

Anyone is free to redistribute GMGPL (or LGPL) code under the GPL
license. If you get the code through such a source, you are not
allowed to redistribute it under any license but the GPL.

  Sam
-- 
Samuel Tardieu -- sam@rfc1149.net -- http://www.rfc1149.net/sam

Re: GNAT GPL 2005 Edition is now available

[Preben Randhol]

On 2005-09-21, Samuel Tardieu <sam@rfc1149.net> wrote:
>>>>>> "Preben" == Preben Randhol <randhol@bacchus.pvv.ntnu.no> writes:
>
>Preben> By the way if one has contributed patches for GtkAda does one
>Preben> have to be consulted before a license change or can
>Preben> FSF/Adacore just change the license from GMGPL to GPL?
>
> Anyone is free to redistribute GMGPL (or LGPL) code under the GPL
> license. If you get the code through such a source, you are not
> allowed to redistribute it under any license but the GPL.

So if I bought support from Adacore (not going to as I personally only
make GMGPL stuff) I could in fact redistribute the GMGPL compiler? Or is
there some other contract stopping me from doing it?

I mean can Adacore have a contract saying that if the compiler is
redistributed they will not offer support anymore after the period you
have paid for expires?

I'm just trying to understand how one can make a business out of GPLed
stuff...

Well anyway I guess I need to learn Java or C## as it probably will be
the languages used in the future... 

Preben

Re: GNAT GPL 2005 Edition is now available

[Preben Randhol]

On 2005-09-21, Preben Randhol <randhol@pvv.ntnu.no> wrote:
> So if I bought support from Adacore (not going to as I personally only
> make GMGPL stuff) I could in fact redistribute the GMGPL compiler? Or is
> there some other contract stopping me from doing it?
>
> I mean can Adacore have a contract saying that if the compiler is
> redistributed they will not offer support anymore after the period you
> have paid for expires?

Ok that didn't come out right. I mean you won't be able to buy support
again for a new time period...

Re: GNAT GPL 2005 Edition is now available

[Jacob Sparre Andersen]

[ I am not a lawyer ]

Preben Randhol wrote:

> So if I bought support from Adacore (not going to as I personally
> only make GMGPL stuff) I could in fact redistribute the GMGPL
> compiler?

Yes.

> Or is there some other contract stopping me from doing it?

That contract would mean that AdaCore broke their contract with the
copyright holders (i.e. GNU GPL/GMGPL), which basically would either:

 + invalidate the contract with the copyright holders - and thus
   prevent AdaCore from using and distributing the code, or

 + not be valid.

So the answer is no.

> I mean can Adacore have a contract saying that if the compiler is
> redistributed they will not offer support anymore after the period
> you have paid for expires?

That might be possible, but I suspect that FSF would be happy to pay
the cost of disputing such a contract in court.

> I'm just trying to understand how one can make a business out of
> GPLed stuff...

By knowing the code better than your competitors, so people come to
you, when they need help with adjusting the code to a specific
purpose.

Another angle is that whenever I do consulting work, one of the
conditions is that the code can be published under an Open Source
license (which one is generally up to the costumer).  As long as you
give your costumers a chance to get first to the market, it seems like
they don't find it problematic.

Jacob

PS: I am not a lawyer - for those who didn't notice it at first.
-- 
�I'm perfectly happy with my current delusional system.�
                                           -- Mirabel Tanner

Re: GNAT GPL 2005 Edition is now available

[Hyman Rosen]

Jacob Sparre Andersen wrote:
> > Or is there some other contract stopping me from doing it?
>
> That contract would mean that AdaCore broke their contract with the
> copyright holders (i.e. GNU GPL/GMGPL)

AdaCore could grant the GMGPL exemption only to their paid customers,
and not permit that exemption to be transferred. There's nothing wrong
with that under the GPL. So any paid customer could redistribute the
compiler as widely as they wish, but recipients could use it only to
produce GPLed software, as they would not be granted the GMGPL right
to distribute the runtime as part of a non-GPLed program.

Re: GNAT GPL 2005 Edition is now available

[Samuel Tardieu]

>>>>> "Hyman" == Hyman Rosen <hyman.rosen@gmail.com> writes:

Hyman> AdaCore could grant the GMGPL exemption only to their paid
Hyman> customers, and not permit that exemption to be
Hyman> transferred.

They could do that only on files they own totally. It excludes files
whose copyright belongs, fully or in part, to:

  - the Free Software Foundation
  - the Florida State University
  - the European Space Agency
  - the Universidad Politecnica de Madrid
  - the University of Toronto

as AdaCore would need an agreement from the other copyright holders to
do the change, and I doubt any of them would agree.

That lets us with 49 runtime files owned fully by AdaCore, of which 31
belong to the GNAT hierarchy and 10 are plateform-specific files.

If GNAT.* is not to be used, only 1 package would be problematic. It
is System.Case_Util, whose function To_Upper is used only once in
System.Val_Util. This unique call could easily be thrown away and
replaced by an ad-hoc function.

As of today, there is no risk as loosing the GMGPL in the FSF
repository for the GNAT runtime (except the GNAT.* hierarchy and a
few system-specific files) unless one of the other copyright owners
also decide to do so.

  Sam
-- 
Samuel Tardieu -- sam@rfc1149.net -- http://www.rfc1149.net/sam

Re: GNAT GPL 2005 Edition is now available

[Preben Randhol]

In article <m2fyryy61c.fsf@hugin.crs4.it>, Jacob Sparre Andersen wrote:
>That might be possible, but I suspect that FSF would be happy to pay
>the cost of disputing such a contract in court.

Why? I don't see how a (GM)GPL software license influences the
conditions for qualifying to get access to a paid-support service?

>Another angle is that whenever I do consulting work, one of the
>conditions is that the code can be published under an Open Source
>license (which one is generally up to the costumer).  As long as you
>give your costumers a chance to get first to the market, it seems like
>they don't find it problematic.

The problem is if you have spent 1-10M euro on research and have come up
with results you don't want the competition to take from you by knowing
how the database is constructed. 

Re: GNAT GPL 2005 Edition is now available

[Jacob Sparre Andersen]

Preben Randhol wrote:
> Jacob Sparre Andersen wrote:

>>That might be possible, but I suspect that FSF would be happy to pay
>>the cost of disputing such a contract in court.
>
> Why? I don't see how a (GM)GPL software license influences the
> conditions for qualifying to get access to a paid-support service?

That condition can be seen as an attempt to do something the (GM)GPL
clearly says is not allowed.  I am pretty sure I would be able to
convince a Danish court that such a clause should be considered a
breach of the agreement in the (GM)GPL.  Whether it would help me much
with getting support afterwards is a different matter.

>>Another angle is that whenever I do consulting work, one of the
>>conditions is that the code can be published under an Open Source
>>license (which one is generally up to the costumer).  As long as you
>>give your costumers a chance to get first to the market, it seems
>>like they don't find it problematic.
>
> The problem is if you have spent 1-10M euro on research and have
> come up with results you don't want the competition to take from you
> by knowing how the database is constructed.

If you distribute the database and it really is 1-10M euro better, you
can be sure that people will find out anyway.  The trick is to stay in
front of the pack.

I don't know any software companies doing research on quite that
scale, but the one chemical processing company I know does spend that
kind of money on developing new technologies - and then publish it [1].
By the time the articles are reviewed and published, they're already
working on the next improvement anyway.

Greetings,

Jacob

[1] Occassionally as patents, but they claim (and seem) to try to
    avoid taking out patents, since they mostly are a waste of money
    on lawyers.
-- 
"This page inadvertently left blank."

Re: GNAT GPL 2005 Edition is now available

[Tapio Kelloniemi]

Jacob Sparre Andersen <sparre@nbi.dk> wrote:
>[ I am not a lawyer ]
>
>Preben Randhol wrote:
>
>> So if I bought support from Adacore (not going to as I personally
>> only make GMGPL stuff) I could in fact redistribute the GMGPL
>> compiler?
>
>Yes.
>
>> Or is there some other contract stopping me from doing it?
>
>That contract would mean that AdaCore broke their contract with the
>copyright holders (i.e. GNU GPL/GMGPL), which basically would either:
>
> + invalidate the contract with the copyright holders - and thus
>   prevent AdaCore from using and distributing the code, or
>
> + not be valid.
>
>So the answer is no.

At least AdaCore asks their customers not to distribute any software
they receive, but this can't be an official condition. Seemingly
capitalists honour requests of AdaCore, since we don't have a copy of GNAT
PRO. The reason is quite like "selfish", they have already paid for the
SW and A) don't want others to get it for free and B) would not
benefit from it.

-- 
Tapio

Re: GNAT GPL 2005 Edition is now available

[Jeffrey Carter]

Preben Randhol wrote:
> 
> Well anyway I guess I need to learn Java or C## as it probably will be
> the languages used in the future... 

Or use the GCC/FSF GMGPL version of GNAT, or make your own copy of the 
GMGPL GNAT and maintain it yourself, or (depending on platform) use 
Janus Ada ($200) or ObjectAda ($0 and up), or ...

-- 
Jeffrey Carter
"Now go away or I shall taunt you a second time."
Monty Python and the Holy Grail
E-mail: jeffrey_r_carter-nr [commercial-at]
         raytheon [period | full stop] com

Re: GNAT GPL 2005 Edition is now available

[Pascal Obry]

Jeffrey,

> Or use the GCC/FSF GMGPL version of GNAT, or make your own copy of the
> GMGPL GNAT and maintain it yourself, or (depending on platform) use
> Janus Ada ($200) or ObjectAda ($0 and up), or ...

Exactly. We used to be sad because there was no Ada compiler "free" or
almost "free" (as in free beer) not so long ago. Today we have:

AdaCore GNAT Pro
AdaCore GNAT GAP
AdaCore GNAT GPL
GNAT FSF
RR Software / Janus-Ada
Aonix / ObjectAda

All come with different licenses. Choose the one you like. I really
don't see the problem.

Also as this thread is a bit mixed up with the GNAT/Debian case.
Remember that no compiler is currently supporting 100% of Ada 2005. So
for all vendors there is still lot of work to do for Ada 2005 support. I
let Ludovic make the best choice for this paltform. Ludovic, thank you
for all the work you do on well supporting Ada on Debian.

Pascal.

-- 

--|------------------------------------------------------
--| Pascal Obry                           Team-Ada Member
--| 45, rue Gabriel Peri - 78114 Magny Les Hameaux FRANCE
--|------------------------------------------------------
--|              http://www.obry.net
--| "The best way to travel is by means of imagination"
--|
--| gpg --keyserver wwwkeys.pgp.net --recv-key C1082595

Re: GNAT GPL 2005 Edition is now available

[Björn Persson]

Ludovic Brenta wrote:
> I don't think that current paying customers of
> I think that the only ones who are really hurt by
> AdaCore's decision are the SMEs and individuals who want to write
> commercial software in Ada, but don't have the money to pay for
> AdaCore's (or other companies') support contracts.

And also those who distribute free libraries under the GMGPL or other 
free non-GPL licenses. But I think AdaCore may have overlooked that 
small companies sometimes grow into big companies. A company that is 
below AdaCore's radar today may be big enough to become a customer in a 
few years, but not if they have to choose C++ to get a good development 
environment.

-- 
Bj�rn Persson                              PGP key A88682FD
                    omb jor ers @sv ge.
                    r o.b n.p son eri nu

Re: GNAT GPL 2005 Edition is now available

[David Trudgett]

"Ludovic Brenta" <ludovic@ludovic-brenta.org> writes:

>> The real problem is not technical. It is still possible to build a
>> "clean" compiler which can be used on any kind of sources. The problem
>> is political. AdaCore's move causes a lot of confusion in the
>> community and may make companies unsure of what they can and can't do.
>
> I agree that there is confusion.  A lot of companies would like "free
> software" to be free for companies; it isn't.  

Yes, it is. The GPL doesn't discriminate. What you mean to say is that
a lot of companies would like to steal Free Software, make secret
changes to it (maybe), and then distribute it without source, thereby
making the software un-Free, and incidentally making a profit from
other people's generosity (this bit the GPL has no problem with, by
the way).


> It is "free" for end users.  

The end user of a compiler is the developer, so what you say is false,
unless you define freedom as the freedom to license one's own software
under the GPL.

Let's be quite clear here. No one is contesting the software
(compiler) owner's right to indicate his desire to maintain the free
status of his work (which is the compiler). (It is, however, wrong to
use violent laws to do this -- and nearly all that we know as "law",
including copyright law, is based on the violence of the state.) No
one is contesting the right of the author to have his wishes respected
in that regard. What *is* being very strongly contested is that
author's right to dictate to others what they produce with his
software. The author may wish whatever he likes, but he cannot
*demand* compliance in the actions of others, unless he has contempt
for the principle of individual freedom, and is prepared to use
violence to obtain that compliance.

Absolutely no one, I put it to you, will consider it illegitimate for
a developer to use Emacs, for example, to create whatever they wish to
create with it, which includes non-Free software. There would be a
thunderous world-wide uproar if RMS even hinted that he wanted to
restrict what people could do with Emacs.

Yet some of the same people see no problem with compiler authors doing
just this, using nothing more than a flimsy, transparent excuse to do
with the run-time library. These people are not champions of freedom,
they have contempt for it, because you cannot promote freedom by using
tools of repression. If you respect freedom, then you have to respect
it for those who wish to do things you don't like, even if they're
using "your" [1] products to do it.

I have made it quite clear that the mere fact of linking the
compiler's run-time with a non-GPL program (even a proprietary one),
in no way impinges upon the freedom of that library, nor does it
endanger its future continued freedom. That is the *sole* valid moral
excuse for preventing (or attempting to prevent) linking of library
code. Seeking to force others to do what *you* want (something that
Stallman and the FSF are guilty of) is not a valid moral excuse...
even if what *you* want, you call "freedom". There is never an excuse
for violent coercion (through copyright law, in this instance).


    [1] "your" as in "you created them".


Now, no one has to use GNAT GPL 2005. But that is not the point, of
course. Microsoft says the same about Windows. It is the intent to
force the actions of others that is wrong in both cases. That some
people will be able to find alternatives is good but beside the main
point. The main point is that software authors have no business
dictating to end users how those end users should employ their
software, including how the end users should license the product of
their own work.

What that means is that there is no such thing as "intellectual
property" if by "property" you mean the right to use violence to
coerce the actions of others. (This is the reality behind "copyright"
laws.) 

Stallman himself ridicules talking about copyright as "intellectual
property", because "property" is an incorrect analogy. One does not
"own" an essay, book or program like one might own a chair. A chair
has different characteristics entirely. You cannot give someone your
chair and still retain the use of it; yet this is the very nature of
software. In effect, we have this principle:

    Omnis enim res, quae dando non deficit, dum habetur
        et non datur, nondum habetur, quomodo habenda est.

    For if a thing is not diminished by being shared with others,
        it is not rightly owned if it is only owned and not shared.

    Book I, Chapter 1 "De doctrina christiana"
    "Corpus Christianorum", "Series latina", Vol. 32, p. 6, lines 10-11.
    Written 397 AD by Saint Augustinus


But even with a chair, I cannot sell it to someone and demand that it
only be used by black people, or that it can only be sat on between
the hours of six and nine in the morning, or that it can only be used
at home and not in a place of business. It would be morally outrageous
to even suggest it, even if you had made the chair yourself.

This is even more the case with software, which can be shared
virtually limitlessly without diminishing anyone else's use of it.
The whole commercial software industry as it exists today is based on
the use of violence to enforce artificial scarcity on a resource which
is, by its very nature, virtually limitless. This is immoral, and
certainly unchristian.



> It gives freedoms to end users and imposes requirements on
> companies or producers of software.  Additional confusion is fueled
> by "open source" (business- friendly attitude) and by the multiplicity
> of licenses.

Free Software is not anti-business, as you seem to be trying to imply
here. A business can use and distribute Free Software and make a
profit doing so. It can even create its own software based on the Free
Software and not share it with anyone. But the software has to remain
Free if it is distributed.

And on a different point, how does the GPL "impose" requirements if
not by the violence of law? So, Stallman seems to like the idea of
"Free Software" but doesn't mind using violent coercion (the
antithesis of freedom) to get it. I would say that's getting pretty
close to hypocritical, but I give him the benefit of the doubt. After
all, I've used the GPL myself, but do not intend to use violence to
enforce it (but even that policy may have to be reviewed).

On an only somewhat related note, one thing that I will definitely
*not* be doing in future is including the "or later versions" (of the
GPL) in any of my licence files. I have concluded that it is of the
utmost stupidity to effectively give one man (or even very few people)
the power to unilaterally make arbitrary changes to the licensing
terms on a whole world of software. Do people really have that much
faith in the continued integrity of one person?



> Me too, but as I said, AdaCore is under no obligation to explain.  If
> they care to explain, I'm all ears, of course.
>
> My message is that nobody has a right to *complain* about AdaCore's
> decision, or even demand an explanation; but people should decide for
> themselves what to do about it.  The vote is designed to do just that.

A couple of points need to be made about that. Everyone has the right
to free speech, regardless of whatever local laws may pretend to
say. You may call such free speech "complaint" if you like, but
everyone *does* indeed have a right to it. 

Or is there a provision in the GPL which controls what one may
complain about? ;-)

Second, no one has *demanded* anything of AdaCore. People have
requested an explanation, and they think it reasonable to not only
request it, but to receive it. I have pointed out that it will not
look good for AdaCore to refuse an explanation in this case. This
still does not make it a *demand*, it simply points out consequences
(which may or may not be correctly predicted, but that is another
matter).

As for deciding what to do about it: by all means! But this does not
preclude multi-tasking: we can voice disapproval, ask for
explanations, *and* do something about it. I think there is even
special Ada syntax for doing more than one thing at a time, isn't
there? :-)


David



-- 

David Trudgett
http://www.zeta.org.au/~wpower/

"On two occasions I have been asked [by members of Parliament!],
'Pray, Mr. Babbage, if you put into the machine wrong figures, will
the right answers come out?' I am not able rightly to apprehend the
kind of confusion of ideas that could provoke such a question."
    
    -- Charles Babbage

Re: GNAT GPL 2005 Edition is now available

[Simon Wright]

David Trudgett <wpower@zeta.org.au.nospamplease> writes:

> I have made it quite clear that the mere fact of linking the
> compiler's run-time with a non-GPL program (even a proprietary one),
> in no way impinges upon the freedom of that library, nor does it
> endanger its future continued freedom.

This is entirely your opinion, clearly the FSF (who were probably
acting on legal advice) felt constrained, probably reluctantly, to
support the LGPL to allow the linking of an OSS library into a
non-open program.

Re: GNAT GPL 2005 Edition is now available

[Ludovic Brenta]

David Trudgett <wpower@zeta.org.au.nospamplease> writes:
> The end user of a compiler is the developer, so what you say is
> false, unless you define freedom as the freedom to license one's own
> software under the GPL.

But we're not talking about the license of the compiler here.  We're
talking about the license of the run-time library, which changed from
GMGPL to pure GPL.

And whether you like it or not, if you write software that links with
libgnat, your software is a "work based on" libgnat.  This is true
even if you don't use any compiler to write that software, i.e. even
if you do

cat > a.out

and a.out uses libgnat, then a.out must be GPL.

Therefore, your comparison with emacs or compilers is besides the
point.  The point is the license for the *library*.

> I have made it quite clear that the mere fact of linking the
> compiler's run-time with a non-GPL program (even a proprietary one),
> in no way impinges upon the freedom of that library, nor does it
> endanger its future continued freedom.

You only have made it clear that you didn't read the GPL.  Also, you
are in no position to decide whose liberty has or has not been
impinged; only the authors of libgnat are in that position.  It is
their right and prerogative to allow or disallow certain ways to
modify, copy or distribute libgnat.  Take it or leave it.  We are
voting on this forum to decide whether Debian sould take it or leave
it.

> But even with a chair, I cannot sell it to someone and demand that
> it only be used by black people, or that it can only be sat on
> between the hours of six and nine in the morning, or that it can
> only be used at home and not in a place of business. It would be
> morally outrageous to even suggest it, even if you had made the
> chair yourself.

Copyright law is quite clear on this.  The copyright holder has the
right to restrict copying, modification and distribution of their
work; not usage.

> This is even more the case with software, which can be shared
> virtually limitlessly without diminishing anyone else's use of it.
> The whole commercial software industry as it exists today is based on
> the use of violence to enforce artificial scarcity on a resource which
> is, by its very nature, virtually limitless. This is immoral, and
> certainly unchristian.

Precisely.  By your own argument, you should therefore be a big
promoter of the pure GPL for all software and libraries.  One od the
things that the pure GPL disallows is linking a GPL library into a
proprietary program, the artificial scarcity of which is organised by
the program's author.  The GPL says: "thou shalt not make non-free
software".

> And on a different point, how does the GPL "impose" requirements if
> not by the violence of law? So, Stallman seems to like the idea of
> "Free Software" but doesn't mind using violent coercion (the
> antithesis of freedom) to get it. I would say that's getting pretty
> close to hypocritical, but I give him the benefit of the
> doubt. After all, I've used the GPL myself, but do not intend to use
> violence to enforce it (but even that policy may have to be
> reviewed).

Welcome to the real world.  All freedoms must be fought for and
defended with some violence, lest the law of the strongest prevail and
all freedom be banished.  Amen :)

> On an only somewhat related note, one thing that I will definitely
> *not* be doing in future is including the "or later versions" (of
> the GPL) in any of my licence files. I have concluded that it is of
> the utmost stupidity to effectively give one man (or even very few
> people) the power to unilaterally make arbitrary changes to the
> licensing terms on a whole world of software. Do people really have
> that much faith in the continued integrity of one person?

The GPL version 3 is being worked on *in public* and *everyone* is
welcome to comment on it.  Of course, as an author, it is your right
and prerogative not to allow GPL version 3.

> A couple of points need to be made about that. Everyone has the right
> to free speech, regardless of whatever local laws may pretend to
> say. You may call such free speech "complaint" if you like, but
> everyone *does* indeed have a right to it. 

Okay, I concede that, but this does not make AdaCore liable to you for
their decisions.  Complain all you want, they don't have to listen.

> Second, no one has *demanded* anything of AdaCore. People have
> requested an explanation, and they think it reasonable to not only
> request it, but to receive it. I have pointed out that it will not
> look good for AdaCore to refuse an explanation in this case. This
> still does not make it a *demand*, it simply points out consequences
> (which may or may not be correctly predicted, but that is another
> matter).

Fair enough.  I suspect that AdaCore do not care about the
consequences to non-customers, but that's just my opinion.

> As for deciding what to do about it: by all means! But this does not
> preclude multi-tasking: we can voice disapproval, ask for
> explanations, *and* do something about it. I think there is even
> special Ada syntax for doing more than one thing at a time, isn't
> there? :-)

yes :) and besides we're doing all three in this thread, and even
going into philosophical arguments :)

-- 
Ludovic Brenta.

Re: GNAT GPL 2005 Edition is now available

[Martin Krischik]

Ludovic Brenta wrote:

> Fair enough.  I suspect that AdaCore do not care about the
> consequences to non-customers, but that's just my opinion.

Thinking about that: With an GNAT Tracker account you can only download
binaries and sources for which the company you work for has purchased
support.

Does that ring a bell?

Martin.

-- 
mailto://krischik@users.sourceforge.net
Ada programming at: http://ada.krischik.com

Re: GNAT GPL 2005 Edition is now available

[Stephane Riviere]

> Do you think that the DoD (the funder of the early GNAT versions)
> intent was to restrict the use of GNAT to programs using the GPL
> license and to AdaCore (which didn't exist) customers?

True ! GNAT was funded for widely spread Ada uses...

> The real problem is not technical. It is still possible to build a
> "clean" compiler which can be used on any kind of sources. The problem
> is political. AdaCore's move causes a lot of confusion in the
> community and may make companies unsure of what they can and can't do.

I deeply agree...

> I still fail to understand this tactical move.

I can't understand too because, at the same time, there is no business 
announces from AdaCore for SMEs and individual professionals. May be it 
is a market to take... May be not (AdaCore may change its mind later).

May be it's some sad news... May be, as Ludovic told us, a new start...

-- 
Stephane Riviere
Oleron Island - France
http://stephane.rochebrune.org
OpenPgp Key <5fd6a1e6> available on the web site above

Re: GNAT GPL 2005 Edition is now available

[Jeffrey Carter]

Samuel Tardieu wrote:
> 
> Do you think that the DoD (the funder of the early GNAT versions)
> intent was to restrict the use of GNAT to programs using the GPL
> license and to AdaCore (which didn't exist) customers?

The DOD program that resulted in GNAT was created specifically to build 
a freely-available compiler that could be used to create proprietary SW. 
The issue of the run time, including parts of the run time that might be 
added to a program by instantiating a generic, are why NYU's lawyers 
came up with the GMGPL in the first place.

Until recently public GNAT releases from AdaCore filled this 
requirement. More recently, the GCC/FSF GNAT has been filling it, and 
many here seem to be happy using the GCC 3.4 and 4.0 versions of GNAT.

AdaCore said some time ago that they had transferred the public version 
of GNAT to the GCC/FSF effort, so I didn't expect to see any further 
public versions from them. This release simply muddies the water, 
increasing the number of types of releases by 1 (paid/supported, 
academic, public/general, and now, public/GPL-only).

-- 
Jeffrey Carter
"Now go away or I shall taunt you a second time."
Monty Python and the Holy Grail
E-mail: jeffrey_r_carter-nr [commercial-at]
         raytheon [period | full stop] com

Re: GNAT GPL 2005 Edition is now available

[Steve Whalen]

Jeffrey Carter wrote:
> Samuel Tardieu wrote:
> >
> > Do you think that the DoD (the funder of the early GNAT versions)
> > intent was to restrict the use of GNAT to programs using the GPL
> > license and to AdaCore (which didn't exist) customers?
>
> The DOD program that resulted in GNAT was created specifically to build
> a freely-available compiler that could be used to create proprietary SW.
> The issue of the run time, including parts of the run time that might be
> added to a program by instantiating a generic, are why NYU's lawyers
> came up with the GMGPL in the first place.

Thanks to you and Samuel for the posts.  I was more active in this
group from about 1994 to 1999 and have mostly been "lurking" since
then (not enough time to get involved in most discussions, though I
still care about Ada). I'd read all of this discussion and was having
a hard time figuring out why I was so mad at AdaCore, but your
reminder of some of the history of GNAT helped me figure it out.

I do believe AdaCore (or any company) has the right to choose what
they do with what they create.  But I think this move is really bad in
several ways. Bad for Ada. Bad for AdaCore. Bad for free software
(including GNU/FSF because of the licensing confusion this will cause:
only programmers will understand the significance of a run time
library and it's license, but Microsoft and others will have another
opportunity to spread confusing FUD about the GPL).

You reminded me that DOD and NYU and Robert Dewar and the others were
in many ways doing a very "good thing" by creating a free Ada compiler
to help offset the removal of the Ada "mandate" from within the DOD.

What bothers me about this move by AdaCore is that Robert and AdaCore
seem to have forgotten a lot of what they said back at the beginning
of GNAT.

10+ years ago Robert and the other AdaCore founders were making a lot
of very sensible statements of practice, principle and philosophy back
when GNAT and AdaCore were being "born", including:

1) GNAT would be good for the industry, and good for Ada. A free
compiler would encourage teaching and use of Ada in places it would
otherwise not reach.

2) There would always be demand for "validated" compilers for DOD /
safety work and for high quality support for the toolset so AdaCore
(or whatever it was called at the start) would be able to make enough
money to survive.

3) GNAT would be good advertising and not cost AdaCore any serious
money beyond the work necessary to "package" the non-validated
releases.

4) There was a lot more but that's enough for now (Anyone who wasn't
hanging around comp.lang.ada back in the mid 1990's I invite to go
back in google groups and read what Robert and the other AdaCore
founders were saying back then).

It all made sense then, and nothing has changed to invalidate it.

I've realized that I'm mad because in a way, I'm feeling betrayed.
Robert Dewar and the AdaCore people are going back on what they said
they were going to do, and why they were going to do it. (I repeat, I
agree AdaCore has the absolute right to do what they are doing, but I
think it's a really big mistake on their part and certainly NOT in
keeping with the pronouncements they made repeatedly during the first
five years of GNAT's existence ... this seems more like something I'd
expect from a Microsoft type company than from GNAT's creators).

It makes me wonder who's running AdaCore.  Have the founders been
pushed aside? Are they having so much trouble meeting their payroll
that they're getting desperate? Have they forgotten all of what they
said when GNAT was just getting started? I was glad when they founded
AdaCore (or whatever it was called at the beginning) because it seemed
appropriate that those who'd put so much work into getting GNAT off
the ground should get something back from it.  Suddenly it seems
they've forgotten all they said they believed in ten years ago.

Worse, I'm afraid it's a sign AdaCore no longer has confidence in it's
ability to attract customers to GNAT by their high quality of service
and depth of knowledge and support. That does not bode well for the
long run, for either AdaCore or Ada.

Steve

Re: GNAT GPL 2005 Edition is now available

[wojtek]

> I asked for votes about the next Ada compiler in Debian only because I
> want to do what users want.

Well, I'm one of those few crazy SME people actually trying to produce
some commercial software in Ada. No, I do not feel betrayed :-) But
GNAT GPL is not an option for us, neither are AdaCore support fees, at
least for the time being :-)

Apart from this, AdaCore has just surprised some people. Is there any
indication that this relase is not incidental and that there are going
to be follow-up releases in this line?

> If this dependence on AdaCore is really a problem, the solution is
> independence from AdaCore, not complaints.  This independence exists
> today in the form of the FSF tree.

It exists and I believe it will continue to exist. FSF is the copyright
holder, so if the "Modified" clause were removed from FSF tree, AdaCore
would be unable to provide the runtime sources to its customers under
GMGPL anymore, either.

Finally, this is not new. The same happened to MySQL and SAPDB -
drivers went from LGPL to GPL, which surprised many people. XFree86 has
been forked because some people wanted to be more famous than other.
OpenBSD has forked apache because of their new license. Now GNAT, oh
well.

Regards,
Wojtek

Re: GNAT GPL 2005 Edition is now available

[wojtek]

I wrote:

> It exists and I believe it will continue to exist. FSF is the copyright
> holder, so if the "Modified" clause were removed from FSF tree, AdaCore
> would be unable to provide the runtime sources to its customers under
> GMGPL anymore, either.

On second thought, the above is b/s. AdaCore can fork.

Maybe it would make sense to ask FSF about the future of the "Modified"
clause for GNAT runtime?

Regards,
Wojtek

Re: GNAT GPL 2005 Edition is now available

[Ole-Hjalmar Kristensen]

>>>>> "LK" == Larry Kilgallen <Kilgallen@SpamCop.net> writes:

    LK> In article <m3mzmdh0jx.fsf@rr.trudgett>, David Trudgett <wpower@zeta.org.au.nospamplease>  writes:
    >> The explanation AdaCore comes up with must explain how restricting the
    >> freedom of developers to license their own work in the way most
    >> suitable to them, does not prove contempt of the very principles of
    >> freedom that the GPL is supposed to be defending.

    LK> The GPL effect is designed to be viral -- AdaCore is now adhering
    LK> to it more closely.  They are quite in tune with the FSF manifesto.

Maybe, *if* the output of the compiler/linker is really covered by
GPL. I am not a lawyer, so this is not obvious to me.

    LK> The fact that AdaCore motives are likely to be promoting their own
    LK> supported product is obvious, as is the fallacy of those who had
    LK> viewed GNAT as the answer to all problems in provision of compilers
    LK> for Ada.

    LK> Computing (like life) has no "magic bullet".

Simple solution: use C, C++ or Java. GCC does not have the same
potential problems as GNAT.

-- 
   C++: The power, elegance and simplicity of a hand grenade.

Re: GNAT GPL 2005 Edition is now available

[Jeff Creem]

David Trudgett wrote:
> Jeff Creem <jcreem@yahoo.com> writes:
> 
> 
>>David Trudgett wrote:
>>
>>>The GCC compiler is also licensed under the GPL, which allows the
>>>development and distribution of proprietary software using that
>>>compiler. Could you please explain here on comp.lang.ada how the same
>>>licence can have two different effects, one for GCC, and another for
>>>GNAT?
>>
>>We will have to wait and see if they respond but in general questions
>>like this to AdaCore usually come back (correctly) with "consult your
>>own qualified legal council for advise"...
> 
> 
> It has been little more than 24 hours since I made my reasonable
> request for a public explanation, and I expect that they are still
> working on a suitable reply. "Consult legal counsel for advice," would
> not be a reasonable reply, because I am not asking for legal advice, I
> am asking them to publicly justify their action. Being unable or
> unwilling to publicly justify what they have done would, I speculate,
> have a negative impact upon their image as, for example, it might
> justifiably be interpreted as a snub to the Ada community (or at least
> that section which frequents c.l.a).
> 
> If AdaCore were to respond citing only a pure technical reason for the
> difference between GCC and GNAT, along the lines that the C and C++
> run-time licenses give freedom to developers, but the GNAT run-time
> has had this freedom revoked, then AdaCore may find themselves in a
> little spot of bother. This is because in that case it would be quite
> apparent they are unwilling to explain their motives, and people will
> be left to draw conclusions about that which will probably be
> negative.
> 
> If AdaCore responds with silence, the result will be somewhat similar,
> though less decisive. So, if I were AdaCore and had something to hide
> about my motivations, then I would definitely fail to reply.
> 
> The explanation AdaCore comes up with must explain how restricting the
> freedom of developers to license their own work in the way most
> suitable to them, does not prove contempt of the very principles of
> freedom that the GPL is supposed to be defending. Maybe they have such
> an explanation, and (for all of us) to find out what it is, is the
> purpose of my enquiry.
> 
> 
> 
>>
>>But the short ** non-qualified ** answer is that whether programming
>>in C,C++ or Ada, there is generally some run-time library that one
>>links against for any non-trivial programs (ok..there are probably
>>non-trivial programs that can be 100% run-time free.)...
>>
>>In any case, with GCC for C one usually links against the the OS libc
>>which is non-GPL (or at least has that option)
>>
>>For C++, the GCC runtime library has a "special exception" clause like
>>the GNAT-3.15 runtime used to have.
>>
>>For Ada, the runtime (and supporting libraries) which are extensive
>>used to be GPL with the special exception that prevented having to
>>apply the GPL to programs that "make use of" the library.
>>
>>In this AdaCore version, they have removed the exception. One can only
>>assume 
> 
> 
> Avoiding assumptions is a good reason for my question to Jamie.
> 
> 
> 
>>that they negotiated this with the FSF (who is the copyright
>>holder on most of this even though it was mostly developed by
>>AdaCore). FSF probably would not have a problem with it anyway since
>>the more "Free" it is the more they like it. 
> 
> 
> On the contrary, it is demonstrably less free. Previously, there were
> no restrictions on what a GNAT developer could produce. Now, a GNAT
> GPL developer will be restricted to producing GPL products. Too bad if
> the developer wants to license under some other Free licence (for
> instance). Anyone can easily see that we are going from more free to
> less free in this sequence of events, and, furthermore, that this
> reduction in developer freedom does not increase the protection of the
> compiler or run-time against those who would like to make proprietary
> versions of them.
> 




I think you need to go read the materials at gnu.org because 
freedom/free in their terms does not mean what you want it to mean. (Or 
at least you misunderstood what I meant when i used the quoted "Free") 
It is the stated position of the FSF that software should be "free". 
What they mean by that is not that developers should be free to do what 
they want. What the FSF wants is for end users to be free to get acess 
to source code for which they have binaries and for them to be free to 
give the source to others.  Several years ago they changed the acroynm 
for the Library GNU Public License (LGPL) to Lesser GNU Public License 
because the FSF believes it to be less "free".

It has been stated by people here (again not AdaCore) they they did not 
need the permission of the FSF (the copyright holder) in order to remove 
the special exception. I have not looked at that issue in detail but it 
certainly makes sense.


You also said

> 
> 
> As a final comment, I would add that I am not against the GPL in any
> way except where it is used to gratuitously limited other people's
> freedoms. The purpose of the GPL is to protect the freedom of
> software, not to force other people to produce free software. A
> compiler is a special class of program that allows the GPL to be
> abused, because the *purpose* of a compiler is to facilitate the

This is not correct. The purpose of the GPL IS to force people to write 
free software. There are plenty examples of libraries out there (GNU 
Scientific Library for example) that are plain old general purpose 
libraries that are distributed under the GPL. The intent is that if you 
simply link with that library, your code must be distributed under the GPL.

Note I am not saying I agree or disagree with what Stallman had in mind 
when he created the GPL..I am trying to say that AdaCore's approach is 
consistant with it.

http://www.gnu.org/licenses/why-not-lgpl.html



Finally, for anyone that does not like this new distribution approach 
(and to a large extent I include myself in this pool), just make believe 
that AdaCore has not released a new version and continue to use the one 
built into your distribuion or one built from the FSF tree...

It is hard to see how you are any worse off as a result of this.

Re: GNAT GPL 2005 Edition is now available

[Marc A. Criley]

Jeff Creem wrote:

> Finally, for anyone that does not like this new distribution approach 
> (and to a large extent I include myself in this pool), just make believe 
> that AdaCore has not released a new version and continue to use the one 
> built into your distribuion or one built from the FSF tree...
> 
> It is hard to see how you are any worse off as a result of this.

Agree, but it's the missed opportunity that I think has upset so many.

It's a fact that GNAT "owned" the Ada hobbyist and one-man-shop market, 
a group that has undeniably produced a lot of freely available Ada 
software over the years.  We've long been wishing for a comprehensive, 
integrated, out-of-the-box upgrade to GNAT, so that we could move on 
from the rather aged 3.15p.

Well, here's that comprehensive, integrated, out-of-the-box upgrade 
we've all been waiting for...but don't even think of writing commercial 
software with it, or letting your commercial customers use your software 
on open (other than full GPL) or closed licensing terms that _you_ would 
want to specify.

Other vendors, such as RR and Aonix, don't limit this, and you can still 
write GMGPL software using their compilers.

-- Marc A. Criley
-- McKae Technologies
-- www.mckae.com
-- DTraq - XPath In Ada - XML EZ Out

Re: GNAT GPL 2005 Edition is now available

[Björn Persson]

Jeff Creem wrote:
> Finally, for anyone that does not like this new distribution approach 
> (and to a large extent I include myself in this pool), just make believe 
> that AdaCore has not released a new version and continue to use the one 
> built into your distribuion or one built from the FSF tree...

The problem is that I'm not sure whether they would be allowed to change 
the license in the GCC tree too, and if so, whether they are going to do 
so, or else, whether they will stop contributing to the GCC tree, and in 
both cases, whether there will be enough people who care enough and are 
skilled enough to make a fork and successfully maintain it.

-- 
Bj�rn Persson                              PGP key A88682FD
                    omb jor ers @sv ge.
                    r o.b n.p son eri nu

Re: GNAT GPL 2005 Edition is now available

[David Trudgett]

Hi Jeff,

Jeff Creem <jcreem@yahoo.com> writes:

>
> I think you need to go read the materials at gnu.org because
> freedom/free in their terms does not mean what you want it to
> mean. 

I think you're misunderstanding what I am saying, and to some extent I
have not been very clear myself in some wording I have chosen here and
there.

I also think I know perfectly well what Stallman and the FSF
want. They want everyone to write Free Software. They would even like
to force people to write Free Software if they want to *use* Free
Software. There are several different but interrelated points to make
about that. I will enumerate some of them here:

1. There is a moral viewpoint and a legal viewpoint. I have been
   dipping into both ends (hence some confusion, perhaps), but I am
   principally speaking from the moral viewpoint, which is at a higher
   level of abstraction from the legal. Hence my use of the word
   '*should*' in at least one place that I recall.

2. What RMS and the FSF *want*, what the GPL *says*, and how it gets
   interpreted in courts of -violence- I mean law, are three different
   things. All of those three things are different from what is
   morally right. It is *never* morally right, for instance, to use
   violence to force or coerce the actions of others, no matter
   whether one's "rights" are being ignored or infringed. Copyright
   law, if taken seriously, does precisely that: it says, "I will
   steal your possessions, abuse you, throw you in prison to be abused
   by violent criminals, and probably ruin your career if you use
   something that I have created in a way I do not wish." Only a
   sociopath or a sociopathic corporation would say such a thing, yet
   that is what copyright law is all about.

3. It is one thing for software to be Free, but quite another for
   humans to be free. Software has no right or claim to be Free.
   Software is an abstraction, it is not a being, it can have no
   rights. But human beings have the right to freedom simply by virtue
   of being human. This means that there is no basis whatsoever for
   Free Software unless it is grounded in human freedom. It is the
   *human being's* right to freedom, and nothing else, that
   legitimises Free Software. Free Software that makes humans unfree
   is therefore a contradiction in terms.

4. Taking a Free Software application created by someone else and
   making it one's own exclusive property (by, for example,
   withholding source code and charging for binary licences), is
   *theft*. It is *major* theft if it is done in such a way that the
   proprietary version "kills off" the Free version by force of law or
   market. There is a moral law against theft.

5. Using a compiler/run-time to produce software unrelated to that
   compiler/run-time, is not theft. It is using the compiler/run-time
   as intended by the technical nature of the compiler/run-time. This
   is not theft, but using a tool in the intended way.

6. A software author may specify any conditions of use he/she prefers,
   but has no *right* to do so, because no one can have a right to do
   something that cannot be (morally) done. Think of your favourite
   offensive example here, such as, "Only to be used by whites." (The
   fact that there may be statutes restricting this in a particular
   time and place is irrelevant, since we are discussing things at a
   moral level here.) It is obvious that one cannot have a *right* to
   use violence against other people (including via the proxy of the
   "law"), and since the use of violence is the only way you will be
   able to prevent people from doing what you do not want them to do,
   you cannot have a "right" to thus prevent them. All you can do is
   express your wishes and expect that by and large people will follow
   them (if they see merit in them -- something which is unlikely in
   the example cited above).

7. Therefore, an author may express the desire that his run-time code
   not be included or linked into any non-Free binary. The intent of
   this wish is to force users of the compiler to licence their own
   work under a particular licence. The intent is not to protect one's
   own Free Software, but to coerce the actions of others. Thus, the
   intent is essentially violent, as it is impossible to force the
   actions of others without recourse to the use of violence.

8. The author's expressed desire in item 7 is therefore based on an
   immoral premise, and consequently there is no moral duty to respect
   the desire, the ends of which are immoral (just as there would be
   no moral duty to respect the author's wishes for only white people
   to use his compiler). Of course, one may still respect the author's
   wishes out of courtesy (especially if there are viable alternatives
   to using the author's compiler), but there is no moral compunction
   to do so. Whether or not there is a legal compunction in a
   particular time and place is irrelevant to this discussion.


> (Or at least you misunderstood what I meant when i used the
> quoted "Free") It is the stated position of the FSF that software
> should be "free". What they mean by that is not that developers should
> be free to do what they want. 

Yes, precisely what I was saying above: they want to make software
"Free" at the expense of human freedom. This is a contradiction in
terms. Software can have no rights. Only humans can.


> What the FSF wants is for end users to be free to get acess to
> source code for which they have binaries and for them to be free to
> give the source to others.  

This is a laudable desire, but, as everyone learns at mother's knee,
the ends do *not* justify the means. Actually, the ends *are* the
means.



> Several years ago they changed the acroynm for the Library GNU
> Public License (LGPL) to Lesser GNU Public License because the FSF
> believes it to be less "free".

This is pure sophistry on their part. There is no way that a GPL'ed
library can be more "Free" than an LGPL'ed library. The only
difference is what it forces other people to do with their own code.
The intent of using the GPL on libraries is *viral*, to use the common
term. GPL on non-library code (ordinary applications) is *not* viral,
and is perfectly safe for anyone to use, by the way (just for the
record -- I know you know that).

There are problems with the GPL. One is that it is not suitable for
library code because of its "viral" nature, which it needs to protect
normal applications. So, that's right: viral GPL'ed library code is
*not* morally acceptable, since the intent is *not* to protect one's
own work, but to coerce others. Now, how effective that coercion *is*
will depend on various factors, such as alternatives available, but
_coercion_ it remains.

Another problem is that there is not always a distinction between
application and library code. The compiler/run-time issue is a small
example, but a bigger example is the world of languages like Common
Lisp. Common Lisp programs work just like "libraries", so there is no
real distinction between "application" and "library" in Common
Lisp. For this reason, I should review my use of the GPL on my noughts
and crosses game in Common Lisp.


> You also said
>
>> As a final comment, I would add that I am not against the GPL in any
>> way except where it is used to gratuitously limited other people's
>> freedoms. The purpose of the GPL is to protect the freedom of
>> software, not to force other people to produce free software. A
>> compiler is a special class of program that allows the GPL to be
>> abused, because the *purpose* of a compiler is to facilitate the
>
> This is not correct. The purpose of the GPL IS to force people to
> write free software. 

Actually, *your* statement is incorrect. Sorry to be so contrary. Only
the GPL applied to a library is intended to coerce others.


David

-- 

David Trudgett
http://www.zeta.org.au/~wpower/

The whole system of domination might unravel if the idea of taking
matters into one's own hands spreads its evil tentacles.

    -- Noam Chomsky
       <http://www.chomsky.info/articles/20041217.htm>

Re: GNAT GPL 2005 Edition is now available

[Georg Bauhaus]

David Trudgett wrote:

>    no one can have a right to do
>    something that cannot be (morally) done.

It is nice to talk about Boolean morale, and due to the superficiality
of your point of view, it is also easy. It won't help in court, though, and
neither in Ada discussions I think.

(I sure will use force to steer my drunken friend (or anyone) away from the
motorway without consulting books, or web pages, on some absolute,
easy to grasp morale. And sure some friends have forgiven me being immoral
in times, others haven't, yet ...)

Re: GNAT GPL 2005 Edition is now available

[Hyman Rosen]

David Trudgett wrote:
> It is *never* morally right, for instance, to use violence to force or
> coerce the actions of others, no matter whether one's "rights" are being
> ignored or infringed.

Morality is a social construct. What is morally right is decided by the
society in which one lives, or by subgroups therein. (The tiger which
is
about to use violence to coerce your actions and then eat you, for
example,
is not acting immorally; she is values-neutral.) So your statement may
be
correct according to your moral code, but no one else need feel
themself
bound by it.

As it happens, society in America has decided that it is moral for a
creator
to retain some coercive power over the use of his creation. It is
written in
one of our founding documents which shapes our society. That may
change, and
perhaps has already changed, but for you to make your bald assertion as
if it
had the force of natural law is absurd. I could equally say that it is
moral
for the strong to take what they wish from the weak, as this is what is
the
observed state of the natural world and we should not attempt to run
counter
to it.

Re: GNAT GPL 2005 Edition is now available

[wojtek]

Jamie,

Do you intend to remove the Modified clause from all the software you
maintain? What about the sources in gcc.gnu.org?

:pserver:anoncvs@libre.adacore.com:/anoncvs is not working anymore. Is
it on purpose?

Regards,
Wojtek Narczynski

Re: GNAT GPL 2005 Edition is now available

[Thomas Quinot]

wojtek@power.com.pl wrote:

> :pserver:anoncvs@libre.adacore.com:/anoncvs is not working anymore. Is
> it on purpose?

This was a technical glitch, the anonymous CVS server is now back to
normal operation. Thank you for pointing this out!

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
                AdaCore -- Paris, France -- New York, USA

Re: GNAT GPL 2005 Edition is now available

[Brian May]

>>>>> "Jamie" == Jamie Ayre <ayre@adacore.com> writes:

    Jamie>   * Ada 2005: http://www.adacore.com/ada_2005.php/

The links (Rationale and more info) at the bottom of this page appear
to be broken - they all seem to point back at the same web page.

Where can I get more detailed information on the changes in Ada 2005?

Has the standardization process completed yet?
-- 
Brian May <bam@snoopy.apana.org.au>

Re: GNAT GPL 2005 Edition is now available

[Friess Michael]

Brian May wrote:

> The links (Rationale and more info) at the bottom of this page appear
> to be broken - they all seem to point back at the same web page.

The / at the end of the URL need to be removed. The correct links are:

   * Ada 2005: http://www.adacore.com/ada_2005.php
   * GNAT GPL 2005 Edition: http://libre.adacore.com/
   * GNAT Pro: http://www.adacore.com/gnatpro_summary.php

Michael

Re: GNAT GPL 2005 Edition is now available

[Brian May]

>>>>> "Friess" == Friess Michael <friess@adacore.com> writes:

    Friess> The / at the end of the URL need to be removed. The
    Friess> correct links are:

    Friess>    * Ada 2005: http://www.adacore.com/ada_2005.php * GNAT
    Friess> GPL 2005 Edition: http://libre.adacore.com/ * GNAT Pro:
    Friess> http://www.adacore.com/gnatpro_summary.php

I think you misunderstood me, if I go to
<URL:http://www.adacore.com/ada_2005.php>, on the right hand side (was
the bottom) of the page I see boxes with "Rationale for Ada 2005" and
"More info on Ada 2005". If I click any of the links within these
boxes, it takes me to the same page again. i.e. the links a broken.
-- 
Brian May <bam@snoopy.apana.org.au>

Re: GNAT GPL 2005 Edition is now available

[Britt Snodgrass]

Brian May wrote:
> >>>>> "Friess" == Friess Michael <friess@adacore.com> writes:
>
>     Friess> The / at the end of the URL need to be removed. The
>     Friess> correct links are:
>
>     Friess>    * Ada 2005: http://www.adacore.com/ada_2005.php * GNAT
>     Friess> GPL 2005 Edition: http://libre.adacore.com/ * GNAT Pro:
>     Friess> http://www.adacore.com/gnatpro_summary.php
>
> I think you misunderstood me, if I go to
> <URL:http://www.adacore.com/ada_2005.php>, on the right hand side (was
> the bottom) of the page I see boxes with "Rationale for Ada 2005" and
> "More info on Ada 2005". If I click any of the links within these
> boxes, it takes me to the same page again. i.e. the links a broken.
> --
> Brian May <bam@snoopy.apana.org.au>

These links to PDF files work for me (using Internet Exploder on
WinXP). However, a few days ago, I had the same problem with them that
you're experencing now because my "display PDF files in browser"
settings were damaged when I installed Adobe Reader 7.0 alongside Adobe
Acrobat 6.0.

Britt

Re: GNAT GPL 2005 Edition is now available

[Larry Kilgallen]

In article <432919be$0$10539$4d4eb98e@read.news.fr.uu.net>, Jamie Ayre <ayre@adacore.com> writes:

> The GNAT GPL 2005 Edition, which is available free of charge from 
> http://libre.adacore.com/, is licensed for Free Software development 
> under the terms and conditions of the GNU General Public License (GPL). 

Has anybody gone through their signup procedure to determine for which
platforms they offer this ?

Their last set of offerings of non-commercial compilers never seemed
to include VMS.  I am not a VMS hobbyists, but I have friends who are.

Re: GNAT GPL 2005 Edition is now available

[Ludovic Brenta]

Larry Kilgallen a écrit :

> In article <432919be$0$10539$4d4eb98e@read.news.fr.uu.net>, Jamie Ayre <ayre@adacore.com> writes:
>
> > The GNAT GPL 2005 Edition, which is available free of charge from
> > http://libre.adacore.com/, is licensed for Free Software development
> > under the terms and conditions of the GNU General Public License (GPL).
>
> Has anybody gone through their signup procedure to determine for which
> platforms they offer this ?
>
> Their last set of offerings of non-commercial compilers never seemed
> to include VMS.  I am not a VMS hobbyists, but I have friends who are.

They offer x86-linux (compiled against glibc 2.3), ppc-darwin
(includes MacOS X), and x86-windows.  No VMS, I'm afraid.  But you can
compile for VMS from the sources; GNAT supports VMS.

-- 
Ludovic Brenta.

Re: GNAT GPL 2005 Edition is now available

[Larry Kilgallen]

In article <1126797631.713038.277490@g43g2000cwa.googlegroups.com>, "Ludovic Brenta" <ludovic@ludovic-brenta.org> writes:
> 
> Larry Kilgallen a =E9crit :
> 
>> In article <432919be$0$10539$4d4eb98e@read.news.fr.uu.net>, Jamie Ayre <a=
> yre@adacore.com> writes:
>>
>> > The GNAT GPL 2005 Edition, which is available free of charge from
>> > http://libre.adacore.com/, is licensed for Free Software development
>> > under the terms and conditions of the GNU General Public License (GPL).
>>
>> Has anybody gone through their signup procedure to determine for which
>> platforms they offer this ?
>>
>> Their last set of offerings of non-commercial compilers never seemed
>> to include VMS.  I am not a VMS hobbyists, but I have friends who are.
> 
> They offer x86-linux (compiled against glibc 2.3), ppc-darwin
> (includes MacOS X), and x86-windows.  No VMS, I'm afraid.  But you can
> compile for VMS from the sources; GNAT supports VMS.

Unless you mean a cross-compiler, that would require having a GNAT
compiler for VMS in the first place, which was the flaw in their
previous cycle of free hobbyist releases.

Re: GNAT GPL 2005 Edition is now available

[Simon Clubley]

In article <1126797631.713038.277490@g43g2000cwa.googlegroups.com>, "Ludovic Brenta" <ludovic@ludovic-brenta.org> writes:
> 
> Larry Kilgallen a =E9crit :
>>
>> Has anybody gone through their signup procedure to determine for which
>> platforms they offer this ?
>>
>> Their last set of offerings of non-commercial compilers never seemed
>> to include VMS.  I am not a VMS hobbyists, but I have friends who are.
> 

You can see the list without signing up, and yes, this was the first thing
that I checked. :-)

> They offer x86-linux (compiled against glibc 2.3), ppc-darwin
> (includes MacOS X), and x86-windows.  No VMS, I'm afraid.  But you can
> compile for VMS from the sources; GNAT supports VMS.
> 

The last public version of GNAT for VMS was 3.12p.

Compiling GNAT from source on VMS is very hard without ACT support,
otherwise I would have done it by now. I have built many native (and some
cross-compiler) versions of GNAT/GCC on Linux, and solved various build
problems while doing so, but I have never been able to successfully build
GNAT from source on VMS.

The first problem is setting up a functioning build environment as ACT
appear to use private ported versions of the various Unix tools and the
public versions (in the form of the GNV tools from HP) either fail with
various errors or cause configure to hang.

Once you get past that, it is uncertain if all the necessary changes are
present in the public binutils sources. That one was pointed out to me
by someone trying to port GCC-3.3 (just C/C++, no GNAT) to VMS.

Since my interest in Ada is not work related, I just switched to using
Linux for my Ada projects.

Check the comp.os.vms newsgroup archives for details if you are interested.

Simon.

-- 
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP       
Microsoft: The Standard Oil Company of the 21st century

Re: GNAT GPL 2005 Edition is now available

[Björn Persson]

I would like to see some kind of official statement from AdaCore on the 
reasoning behind this license change. Personally I think it's a big 
mistake, but I would like to know why AdaCore think it's a good idea. I 
also very much want to know whether the license will be changed in the 
GCC source tree.

I'm very puzzled about the requirement for registration. There's not a 
word of explanation on the page that I can see. Why would I have to log 
in to download free software? It's not like I need yet another website 
account with yet another password to remember.

Finally, there's one thing that needs to get set straight:

Jamie Ayre wrote:
> AdaCore is pleased to announce the release of the GNAT GPL 2005 Edition 
> to provide Free Software developers, that is developers that distribute 
> their work under the GPL (GNU General Public License), the latest and 
> most advanced Ada 2005 software development environment.

When AdaCore explain free software on their website they typically link 
to the FSF site, but in this sentence they're not using the FSF's 
definition of free software. Free software is not at all synonymous to 
GPL. This page lists many other free software licenses, including many 
that are incompatible with the GPL:

http://www.fsf.org/licensing/licenses/index_html

None of those licenses (with reservation for possible rare exceptions) 
could be used on a program linked to the runtime library of the GNAT GPL 
2005 Edition. Not even those that are listed as GPL-compatible, because 
that's a one-way compatibility.

The GNAT GPL 2005 Edition is not for free software developers. It's for 
GPL-only developers. You can't even compile LGPL or GMGPL software (for 
distribution) with it.

-- 
Bj�rn Persson                              PGP key A88682FD
                    omb jor ers @sv ge.
                    r o.b n.p son eri nu

Re: GNAT GPL 2005 Edition is now available

[Samuel Tardieu]

>>>>> "Bj�rn" == Bj�rn Persson <spam-away@nowhere.nil> writes:

Bj�rn> I'm very puzzled about the requirement for
Bj�rn> registration. There's not a word of explanation on the page
Bj�rn> that I can see. Why would I have to log in to download free
Bj�rn> software? It's not like I need yet another website account with
Bj�rn> yet another password to remember.

There exists a service for that: BugMeNot.

See http://www.bugmenot.com/ as well as the excellent BugMeNot Firefox
plugin.

  Sam
-- 
Samuel Tardieu -- sam@rfc1149.net -- http://www.rfc1149.net/sam