GNAT GPL 2005 Edition is now available
Jamie Ayre
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/
Jacob Sparre Andersen
> 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.
David Trudgett
> 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
Ludovic Brenta
> 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.
woj...@power.com.pl
Do you intend to remove the Modified clause from all the software you
maintain? What about the sources in gcc.gnu.org?
:pserver:ano...@libre.adacore.com:/anoncvs is not working anymore. Is
it on purpose?
Regards,
Wojtek Narczynski
Ludovic Brenta
> Jamie Ayre <ay...@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.
woj...@power.com.pl
> 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
Martin Dowie
> 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
Brian May
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 <b...@snoopy.apana.org.au>
Jacob Sparre Andersen
> 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."
Ludovic Brenta
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.
Ludovic Brenta
Brian May a écrit :
> >>>>> "Ludovic" == Ludovic Brenta <lud...@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.
Brian May
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 <b...@snoopy.apana.org.au>
Friess Michael
> 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
Samuel Tardieu
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 -- s...@rfc1149.net -- http://www.rfc1149.net/sam
Stephane Riviere
> 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
Simon Clubley
>
> 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
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.
Jeff Creem
> Jamie Ayre <ay...@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
Samuel Tardieu
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.
Ludovic Brenta
Samuel Tardieu a écrit :
> >>>>> "Ludovic" == Ludovic Brenta <lud...@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.
Samuel Tardieu
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.
Dr. Adrian Wrigley
...
> 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
Thomas Quinot
> :pserver:ano...@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. ** qui...@adacore.com ** Senior Software Engineer
AdaCore -- Paris, France -- New York, USA
Samuel Tardieu
>>>>> "Ludovic" == Ludovic Brenta <lud...@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.
Ludovic Brenta
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.
Larry Kilgallen
> 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.
Ludovic Brenta
Larry Kilgallen a écrit :
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.
Andreas Schwarz
> 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.
Martin Dowie
[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
Niklas Holsti
> ...
> As of now, there are several possible choices for the next default Ada
> compiler in Debian:
>
> 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
. @ .
Larry Kilgallen
>
> Larry Kilgallen a =E9crit :
>
>> In article <432919be$0$10539$4d4e...@read.news.fr.uu.net>, Jamie Ayre <a=
> y...@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.
Larry Kilgallen
There is no reason why Randy would have to do anything GPL to offer
a compiler sufficiently inexpensive to be used on GPL projects.
Jeffrey Carter
>
> 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
Georg Bauhaus
> 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 $$$, and so on.
There are some offerings of toolsets in the range $$-$$$,
but most of them can only be used for "console applications",
unless you build a host of tools, libraries and OS bindings yourself,
or get them elsewhere, with or without sources.
Digging the archives will probably reveal comments made by/about
ACT, indicating that supporting less than 3 or 5 "seats" per
contract isn't the preferred basis for a support contract. (insert
standard disclaimer) But I'm only speculating here, things
might even change, since obviously free riding isn't possible any
longer by using the GNAT GPL edition, unless someone pays you for
producing software that is GPLed.
In fact, someone is paying ACT for producing GPLed software.
Part of the GNAT library is used by the competition, that's good
for everyone.
Perhaps we can dream of a GNAT Std Edition, on a par with
what you get from Borland, Microsoft, or Intel, with limited support,
but with GMGPL?
Georg Bauhaus
> What AdaCore has done is force a fork in the GNAT sources. The GCC/FSF
> version currently allows the creation of non-GPL programs. If they
> change that for one language, they will certainly have to change it for
> all. Luckily, people can and do have their own copies of the sources
> with the exceptions, so there should always be versions that allow the
> creation of non-GPL programs.
Some companies have invested in GCC, not all of them producing
GPLed software. I find it hard to imagine all of them switching
compilers and libraries because at the same time they decide to
change GCC licensing.
Niklas Holsti
> GNAT Pro users have indicated a cost of $$$$ or more, not just
> $$$.
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.
Adrien Plisson
> 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
Simon Clubley
> 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.
Dr. Adrian Wrigley
> 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
chris
> 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.
Simon Clubley
>
> 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.
Georg Bauhaus
> 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.
Martin Krischik
> 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://kris...@users.sourceforge.net
Ada programming at: http://ada.krischik.com
Martin Krischik
>> 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 Krischik
Where is the problem? The GNAT/GPL does not cost any money.
Ludovic Brenta
> 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.
Ludovic Brenta
> 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.
Martin Dowie
> 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
Samuel Tardieu
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
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
Tapio Kelloniemi
>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
Adrien Plisson
> 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
Georg Bauhaus
> 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?
Georg Bauhaus
> 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.
Tapio Kelloniemi
>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
Tapio Kelloniemi
>Georg Bauhaus <bau...@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
Andreas Schwarz
> 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.
Björn Persson
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.
--
Björn Persson PGP key A88682FD
omb jor ers @sv ge.
r o.b n.p son eri nu
Brian May
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 <b...@snoopy.apana.org.au>
Britt Snodgrass
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
Björn Persson
> [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.
Brian May
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 <b...@snoopy.apana.org.au>
Samuel Tardieu
Björn> You *can* take source code with a GPL-compatible license, such
Björn> as the X11 license, link it with a GPL library, such as those
Björn> of the GNAT GPL 2005 Edition, and distribute the resulting
Björn> binary, *but* the binary you distribute will be
Björn> GPL-licensed. This is because the X11 license is GPL-compatible
Björn> 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
Björn> I'm very puzzled about the requirement for
Björn> registration. There's not a word of explanation on the page
Björn> that I can see. Why would I have to log in to download free
Björn> software? It's not like I need yet another website account with
Björn> 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.
Jeff Creem
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.
Jeff Creem
> 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.
Dr. Adrian Wrigley
...
> 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
Jeff Creem
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.
...
Ludovic Brenta
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.
David Trudgett
> 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
David Trudgett
> 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"
Larry Kilgallen
> 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".
Tapio Kelloniemi
>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
Jeff Creem
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.
Ludovic Brenta
> 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.
Samuel Tardieu
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.
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
Ludovic Brenta
> 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.
Stephane Riviere
> 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
Niklas Holsti
> ...
> 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
. @ .
Ludovic Brenta
> 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.
Marc A. Criley
> 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
Jeffrey Carter
>
> 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
Georg Bauhaus
> 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."
Jeffrey Carter
>
> 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).
Martin Krischik
> , 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://kris...@users.sourceforge.net
Ada programming at: http://ada.krischik.com
Tapio Kelloniemi
>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
Marc A. Criley
> 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.
Björn Persson
> 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.
--
Björn Persson PGP key A88682FD
omb jor ers @sv ge.
r o.b n.p son eri nu
Björn Persson
> I don't think that current paying customers of
> 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.
Björn Persson
> 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.
David Trudgett
>> 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
Simon Wright
> 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.
Martin Dowie
[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
Jacob Sparre Andersen
> 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 Björn's e-mail address.
--
»Altså hør nu, jeg ved snart ikke... Jeg var ikke så vild
med sidste gang på Jorden... Korset, sømmene, tornene og
alt det der...« -- Sønnen (ifgl. Mézières & Christin)
Jacob Sparre Andersen
> 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.
Georg Bauhaus
> 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.
Marc A. Criley
> 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.
Jacob Sparre Andersen
> 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."