Companies Only Offering Ada-95 Compilers

303 views
Skip to first unread message

Jeffrey Carter

unread,
May 16, 2012, 9:25:40 PM5/16/12
to
From what I can tell, the following companies only offer Ada-95 compilers:

DDCI
Greenhills
RR Software

The following offer compilers for current Ada:

AdaCore
Irvine

For the following, I can't tell:

Atego
IBM/Rational

Have I missed anyone? Corrections are welcome. I'm trying to determine the
extent of support for current Ada by compiler vendors.

--
Jeff Carter
"I blow my nose on you."
Monty Python & the Holy Grail
03

--- Posted via news://freenews.netfront.net/ - Complaints to ne...@netfront.net ---

Randy Brukardt

unread,
May 18, 2012, 12:17:38 AM5/18/12
to
"Jeffrey Carter" <spam.jrc...@spam.not.acm.org> wrote in message
news:jp1k2n$ouo$1...@adenine.netfront.net...
> From what I can tell, the following companies only offer Ada-95 compilers:
>
> DDCI
> Greenhills
> RR Software

Our (RRS) current beta compiler supports a handful of Ada 2005, and the
complete Ada 2005 syntax. Not quite just Ada 95.

> The following offer compilers for current Ada:
>
> AdaCore
> Irvine

Not sure if Irvine supports any Ada 2012, which is "current Ada" in my mind.

> For the following, I can't tell:
>
> Atego
> IBM/Rational

The Rational compiler supports at least most of Ada 2005 (there was an
announcement to this effect a couple years ago). No idea about Ada 2012. My
understanding was that Atego was Ada 95-only, but that may be old
information.

> Have I missed anyone? Corrections are welcome. I'm trying to determine the
> extent of support for current Ada by compiler vendors.

Randy.


Jeffrey Carter

unread,
May 18, 2012, 2:27:42 AM5/18/12
to
On 05/17/2012 09:17 PM, Randy Brukardt wrote:
>
> Our (RRS) current beta compiler supports a handful of Ada 2005, and the
> complete Ada 2005 syntax. Not quite just Ada 95.

Thanks for the update. To my mind, that's an Ada-95 compiler with a non-standard
mode.

> Not sure if Irvine supports any Ada 2012, which is "current Ada" in my mind.

IIUC, we don't yet have a final, ISO-approved, published version of next Ada
yet, so to my mind it's not current Ada yet.

> The Rational compiler supports at least most of Ada 2005 (there was an
> announcement to this effect a couple years ago). No idea about Ada 2012.

That's good to know.

So to recap, of 7 compilers, 3 implement the complete current standard (1 of
them also implements the draft standard for the next version). 5 years after
publication of the standard, that's not very encouraging.

--
Jeff Carter
Just as Khan was hindered by two-dimensional thinking in a
three-dimensional situation, so many developers are hindered
by sequential thinking in concurrent situations.
118

Lucretia

unread,
May 18, 2012, 7:11:12 AM5/18/12
to
On Friday, May 18, 2012 7:27:42 AM UTC+1, Jeffrey Carter wrote:

> So to recap, of 7 compilers, 3 implement the complete current standard (1 of
> them also implements the draft standard for the next version). 5 years after
> publication of the standard, that's not very encouraging.

And only 1 is available to the general public, RR's is affordable though.

Luke.

Martin

unread,
May 18, 2012, 7:57:42 AM5/18/12
to
On Friday, May 18, 2012 7:27:42 AM UTC+1, Jeffrey Carter wrote:
If their customers aren't calling for it, it's a hard agrument to make for spending time (i.e. money) changing a product and potentially introducing new bugs that might break the existing compiler...

I'm always amazed that more compiler vendors don't offer a $0.00 or $49.99 or $99.99 'home' version though...

And I guess, I thought maybe 1 or 2 might have taken the jump to 'open source' their compilers...just to try and hook more people into using the language.

Even if they just released their Win32 version and not the more commercial PPC604/VxWorks targeting versions.

-- Martin

Nasser M. Abbasi

unread,
May 18, 2012, 8:05:34 AM5/18/12
to
On 5/18/2012 6:11 AM, Lucretia wrote:

>
> And only 1 is available to the general public, RR's is affordable though.
>
> Luke.

well, we (i.e the general public :) only really need ONE good, open
source Ada compiler any way. right?

Look at Java for example, oracle's Java is all there is really. In
open source, it is openJDK. That is the Java that everyone uses. That
is what I use on Linux and windows.

I do not see a problem with having just one Ada compiler. As
long as it is open, and free to use, and is good.

What is needed for Ada, is more Ada libraries, not more Ada
compilers.

--Nasser



Robert A Duff

unread,
May 18, 2012, 9:40:38 AM5/18/12
to
Jeffrey Carter <spam.jrc...@spam.not.acm.org> writes:

> On 05/17/2012 09:17 PM, Randy Brukardt wrote:
>>
>> Our (RRS) current beta compiler supports a handful of Ada 2005, and the
>> complete Ada 2005 syntax. Not quite just Ada 95.
>
> Thanks for the update. To my mind, that's an Ada-95 compiler with a
> non-standard mode.

Well, as far as ISO standards are concerned, there is only
one Ada, and that's what we call Ada 2005, so it doesn't make
sense to talk about "nonstandard modes" with respect to Ada 95.

>> Not sure if Irvine supports any Ada 2012, which is "current Ada" in my mind.
>
> IIUC, we don't yet have a final, ISO-approved, published version of next
> Ada yet, so to my mind it's not current Ada yet.

Right, Randy's mind is a few months ahead of ISO's mind.

>> The Rational compiler supports at least most of Ada 2005 (there was an
>> announcement to this effect a couple years ago). No idea about Ada 2012.
>
> That's good to know.
>
> So to recap, of 7 compilers, 3 implement the complete current standard
> (1 of them also implements the draft standard for the next version). 5
> years after publication of the standard, that's not very encouraging.

Why do you say so? It's much better than the situation in the C world.

- Bob

Shark8

unread,
May 18, 2012, 2:53:31 PM5/18/12
to
On Friday, May 18, 2012 8:40:38 AM UTC-5, Robert A Duff wrote:
> Jeffrey Carter <spam.jrc...@spam.not.acm.org> writes:
> > So to recap, of 7 compilers, 3 implement the complete current standard
> > (1 of them also implements the draft standard for the next version). 5
> > years after publication of the standard, that's not very encouraging.
>
> Why do you say so? It's much better than the situation in the C world.
>
> - Bob

Could you elaborate on that? (What is the situation in the C/C++ world?)

Jeffrey Carter

unread,
May 18, 2012, 3:13:00 PM5/18/12
to
On 05/18/2012 06:40 AM, Robert A Duff wrote:
> Jeffrey Carter<spam.jrc...@spam.not.acm.org> writes:
>>
>> So to recap, of 7 compilers, 3 implement the complete current standard
>> (1 of them also implements the draft standard for the next version). 5
>> years after publication of the standard, that's not very encouraging.
>
> Why do you say so? It's much better than the situation in the C world.

I compare it to the situation in 2000, 5 yrs after publication of the Ada-95
standard, when everyone had implemented Ada 95 for several yrs.

--
Jeff Carter
"We use a large, vibrating egg."
Annie Hall
44

Mike Silva

unread,
May 18, 2012, 3:43:48 PM5/18/12
to
On Friday, May 18, 2012 12:13:00 PM UTC-7, Jeffrey Carter wrote:
> On 05/18/2012 06:40 AM, Robert A Duff wrote:
> > Jeffrey Carter<spam.jrc...@spam.not.acm.org> writes:
> >>
> >> So to recap, of 7 compilers, 3 implement the complete current standard
> >> (1 of them also implements the draft standard for the next version). 5
> >> years after publication of the standard, that's not very encouraging.
> >
> > Why do you say so? It's much better than the situation in the C world.
>
> I compare it to the situation in 2000, 5 yrs after publication of the Ada-95
> standard, when everyone had implemented Ada 95 for several yrs.

Yeah, it gives off a vague sense that Ada is dying.

Speaking of the C world, I'm hoping that with the AdaCore merger Ada Magic will come to support Ada 2005 (or better). That would be very nice.

Nasser M. Abbasi

unread,
May 18, 2012, 4:02:46 PM5/18/12
to
On 5/18/2012 2:43 PM, Mike Silva wrote:

>
> Yeah, it gives off a vague sense that Ada is dying.
>

It is not that Ada is dying (or Fortran or few others for that
matter), it is that all the developers are/have been moving to
the exciting new world of mobile and rich internet software
development, using HTML5 and Javascript.

Windows 8 all new apps development is/will be in HTML5 and
Javascript.

This is the new world order of software engineering. After 50 years
of computer science and programming languages research, the world has
decided it will be Javascript and HTML5.

If someone can make an Ada to Javascript compiler, may be there
is still a chance?

--Nasser


Dmitry A. Kazakov

unread,
May 18, 2012, 4:35:11 PM5/18/12
to
On Fri, 18 May 2012 15:02:46 -0500, Nasser M. Abbasi wrote:

> Windows 8 all new apps development is/will be in HTML5 and
> Javascript.

It will be what Microsoft decide it to be in.

BTW, chances are high that Window 8 will not make it. Customers tend to
refuse MS OSes when advantages are unclear, e.g. Me, Vista. XP had it very
hard in the beginning.

> This is the new world order of software engineering. After 50 years
> of computer science and programming languages research, the world has
> decided it will be Javascript and HTML5.

No. Due to opening a new market the technology (rather lack of) is
returning to the point where it was 20-30 years ago. Such retreats always
happen when there is no pressure to improve quality or safety.

In 10 years software development will recover things it managed to forget
now. Not for the first time. Remember how multi-tasking was forgotten when
PCs came? Microsoft "discovered" that decades later.

> If someone can make an Ada to Javascript compiler, may be there
> is still a chance?

Nope. Machine in the hands of a savage is scrap metal.

--
Regards,
Dmitry A. Kazakov
http://www.dmitry-kazakov.de

Nasser M. Abbasi

unread,
May 18, 2012, 4:52:00 PM5/18/12
to
On 5/18/2012 3:35 PM, Dmitry A. Kazakov wrote:

>> This is the new world order of software engineering. After 50 years
>> of computer science and programming languages research, the world has
>> decided it will be Javascript and HTML5.
>
> No. Due to opening a new market the technology (rather lack of) is
> returning to the point where it was 20-30 years ago. Such retreats always
> happen when there is no pressure to improve quality or safety.
>

That is funny you said that. I was just now thinking the same
thing as I was looking at a web site that tells which browsers
supports which feature of HTML5 and Javascript.

The reason is, I remember doing the same thing many many
years ago, when I was just learning the new Java and looking
to see which operating system supports which features of Java
and such.

So, the picture has changed from

app app
+----------+-----------+......
| Java VM | Java VM |
+----------+-----------+.....
| OS | OS |
+----------+-----------+.....

To now, almost 20 years later to become


HTML5/JS app HTML5/JS app
+--------------+--------------+.....
| browser 1 | browser 2 |
+--------------+--------------+.....
| OS | OS |
+--------------+--------------+.....

Well, it is progress I guess. The Java VM was replaced by
the browser, and Java is replaced by HTML/Javascript.

I can't wait to see what the picture will be 20 years from now :)

--Nasser

Nasser M. Abbasi

unread,
May 18, 2012, 5:07:25 PM5/18/12
to
On 5/18/2012 3:52 PM, Nasser M. Abbasi wrote:

>
> So, the picture has changed from
>
> app app
> +----------+-----------+......
> | Java VM | Java VM |
> +----------+-----------+.....
> | OS | OS |
> +----------+-----------+.....
>
> To now, almost 20 years later to become
>
>
> HTML5/JS app HTML5/JS app
> +--------------+--------------+.....
> | browser 1 | browser 2 |
> +--------------+--------------+.....
> | OS | OS |
> +--------------+--------------+.....
>

I'd like to make small improvement the above software architecture.

I think it should be like this:

20 years ago, the vision was

Java app
+
|
JVM
/|\
/ | \
/ | \
/ | \
OS1 OS2 OS3 .....

Now it evolved to

HTML5/JS app
+
/|\
/ | \
/ | \
/ | \
browser1 br2 browser3 .... (mobile platforms)
/|\
/ | \
/ | \
/ | \
OS1 OS2 OS3 .....

Before, we had one JVM on top of N number of OS's. Now we have
M browsers on top of the same N number of OS's. A progress, in
a way.

--Nasser

Dmitry A. Kazakov

unread,
May 18, 2012, 5:15:48 PM5/18/12
to
App (distributed)
+------+------+
| OS (distributed)
+------+------+

Layers above OS were only necessary because of retarded OSes like UNIX and
Windows. They sealed the toilet's pit. Alas, being not much more hygienic
(I would say most of them were actually much worse) they required another
layers above them, and so on.

Robert A Duff

unread,
May 18, 2012, 6:52:32 PM5/18/12
to
Shark8 <onewing...@gmail.com> writes:

> Could you elaborate on that? (What is the situation in the C/C++ world?)

I don't really know. My impression is that most C compilers do not
fully support the latest C standard, or even the second-to-latest one.

- Bob

Nasser M. Abbasi

unread,
May 18, 2012, 7:03:16 PM5/18/12
to
from http://en.wikipedia.org/wiki/ANSI_C

compilers supproting ansi c are

GCC
Microsoft Visual C++ (C90. A few features of C99)
LabWindows/CVI
ARM RealView
LCC
OpenWatcom (C89/90 and some C99)

per (http://en.wikipedia.org/wiki/C11_%28C_standard_revision%29)

c11 is the current C standard. From the above it says:

"GCC version 4.6 and Pelles C version 7.00 (RC1) have added
initial support for some features from the C11 draft.[5][6]"

--Nasser


Shark8

unread,
May 18, 2012, 9:36:11 PM5/18/12
to mai...@dmitry-kazakov.de
On Friday, May 18, 2012 3:35:11 PM UTC-5, Dmitry A. Kazakov wrote:
> On Fri, 18 May 2012 15:02:46 -0500, Nasser M. Abbasi wrote:
>
> > This is the new world order of software engineering. After 50 years
> > of computer science and programming languages research, the world has
> > decided it will be Javascript and HTML5.
>
> No. Due to opening a new market the technology (rather lack of) is
> returning to the point where it was 20-30 years ago. Such retreats always
> happen when there is no pressure to improve quality or safety.

Does this mean we'll be having PostScript -driven GUIs instead of this Frankenstein HTML5/CSS/Javascript monster? Because the way I see it PS has the advantage of having been designed to address the HTML5/CSS issues (i.e. PS was intended for layout, whereas HTML was not)?

Either way, I hope HTML5 dies.

> In 10 years software development will recover things it managed to forget
> now. Not for the first time. Remember how multi-tasking was forgotten when
> PCs came? Microsoft "discovered" that decades later.

Indeed; I find it slightly amusing that Ada's had tasking for 30 years and the C-type folks are bending over backwards trying to find not only "a way" to do it, but also "the right way."

> > If someone can make an Ada to Javascript compiler, may be there
> > is still a chance?
>
> Nope. Machine in the hands of a savage is scrap metal.

Ah, come on; we have Ada compilers targeting the JVM, surely we can have some that target ECMA-Script.

Nasser M. Abbasi

unread,
May 18, 2012, 10:02:39 PM5/18/12
to
On 5/18/2012 8:36 PM, Shark8 wrote:

> Either way, I hope HTML5 dies.

oh no! not after I just ordered 3 books on HTML5 and a book
on Javascript from amazon.

> Ah, come on; we have Ada compilers targeting the JVM, surely
> we can have some that target ECMA-Script.

I hope someone actually can do this. similar to google's
GWT :

https://developers.google.com/web-toolkit/overview

"The GWT SDK contains the Java API libraries, compiler,
and development server. It lets you write client-side
applications in Java and deploy them as JavaScript."

hang on, I just got an idea: how about

ada-to-java google GWT
Ada---------------> Java -------------> Javascript---> Mobile!
(we have this) .class


Do not know if the above will work, as many details not clear,
may be a brave person can try a "hello world" on the above.

--Nasser


tmo...@acm.org

unread,
May 19, 2012, 1:07:46 AM5/19/12
to
> Indeed; I find it slightly amusing that Ada's had tasking for 30 years and =
> the C-type folks are bending over backwards trying to find not only "a way"=
> to do it, but also "the right way."
I recently went to a local chapter ACM meeting with a speaker on
"Threads and Shared Variables in C++11". Lots of examples of potential
problems with shared variables and the main take-away seemed to be to
be careful to use semaphores carefully. Sad.

Dmitry A. Kazakov

unread,
May 19, 2012, 2:28:27 AM5/19/12
to
Last time I checked they didn't support all Ada.

But the point is that Ada does not fit into that picture.

Marco

unread,
May 19, 2012, 9:44:02 AM5/19/12
to n...@12000.org
On Friday, May 18, 2012 1:02:46 PM UTC-7, Nasser M. Abbasi wrote:
> On 5/18/2012 2:43 PM, Mike Silva wrote:
>
> >
> > Yeah, it gives off a vague sense that Ada is dying.
> >
>
> It is not that Ada is dying (or Fortran or few others for that
> matter), it is that all the developers are/have been moving to
> the exciting new world of mobile and rich internet software
> development, using HTML5 and Javascript.

More "Angry Birds" just what the world needs

there are other areas besides Mobile where most comp.lang.ada folks usually exist:

Industrial - manufacturing, robots, etc
Aerospace
Medical

Yannick Duchêne (Hibou57)

unread,
May 19, 2012, 11:16:48 AM5/19/12
to
Le Fri, 18 May 2012 23:15:48 +0200, Dmitry A. Kazakov
<mai...@dmitry-kazakov.de> a écrit:
Lacking things as fundamental as a minimal common and native standard UI
API. The “that layer over that layer” is often most required for UIs. By
the way, I still believe the web standards may be a chance in that matter.

--
“Syntactic sugar causes cancer of the semi-colons.” [1]
“Structured Programming supports the law of the excluded muddle.” [1]
[1]: Epigrams on Programming — Alan J. — P. Yale University

Yannick Duchêne (Hibou57)

unread,
May 19, 2012, 11:23:38 AM5/19/12
to
Le Sat, 19 May 2012 03:36:11 +0200, Shark8 <onewing...@gmail.com> a
écrit:

> On Friday, May 18, 2012 3:35:11 PM UTC-5, Dmitry A. Kazakov wrote:
>> On Fri, 18 May 2012 15:02:46 -0500, Nasser M. Abbasi wrote:
>>
>> > This is the new world order of software engineering. After 50 years
>> > of computer science and programming languages research, the world has
>> > decided it will be Javascript and HTML5.
>>
>> No. Due to opening a new market the technology (rather lack of) is
>> returning to the point where it was 20-30 years ago. Such retreats
>> always
>> happen when there is no pressure to improve quality or safety.
>
> Does this mean we'll be having PostScript -driven GUIs instead of this
> Frankenstein HTML5/CSS/Javascript monster? Because the way I see it PS
> has the advantage of having been designed to address the HTML5/CSS
> issues (i.e. PS was intended for layout, whereas HTML was not)?

Obviously it was not, and it was on purpose. That's the role of CSS, not
HTML and HTML5, which describes structures and basic semantics traits. The
idea behind these web standards is the separation of presentation
(Cascading *Style* Sheet) and semantic (HyperText *Markup* Language).

> Either way, I hope HTML5 dies.

Sorry, it surely won't ;-) Too much a Holy Grail for so much people
all‑over the world.

Yannick Duchêne (Hibou57)

unread,
May 19, 2012, 11:32:20 AM5/19/12
to
Le Sat, 19 May 2012 04:02:39 +0200, Nasser M. Abbasi <n...@12000.org> a
écrit:
> "The GWT SDK contains the Java API libraries, compiler,
> and development server. It lets you write client-side
> applications in Java and deploy them as JavaScript."
>
> hang on, I just got an idea: how about
>
> ada-to-java google GWT
> Ada---------------> Java -------------> Javascript---> Mobile!
> (we have this) .class

Please, stop adding layers above layers! :-P (*). By the way, the target
is not only mobile, it's anything.

(*) and Google's Java to JS is a mess, and I know no good one, others are
not better, typically another nth script language whose purpose is to be
translatable to JS and optionally PHP or others, always with the same
issues as with all scripting languages: no typing, not proper modularity,
no “smart linking” (ending in nnnKb files for simple things), the “every
thing is a class” trap, scale very badly (OK for tiny stuff, but when
serious things begins, you have very hard times), and etc.

Nomen Nescio

unread,
May 20, 2012, 4:12:44 AM5/20/12
to
> well, we (i.e the general public :) only really need ONE good, open
> source Ada compiler any way. right?

No

> Look at Java for example, oracle's Java is all there is really. In
> open source, it is openJDK. That is the Java that everyone uses. That
> is what I use on Linux and windows.

No again. There are many JVMs and many compilers. Depends what you need and
where you need it. google a search on JVM and you will be surprised how many
companies and versions there are.

> I do not see a problem with having just one Ada compiler. As
> long as it is open, and free to use, and is good.

GNAT is certainly very good. But it is a pain in the ass in many ways. It is
GPL only and even if you use gcc Ada eventually that is going to cause
problems as GPL moves from GPL2 to 3, and who knows what's next. LGPL
libraries today, tomorrow everythings GPL, you trusted them, you put all
your eggs in one basket, you got screwed.

Green Hills (don't know about HQ but the international office I spoke with)
are a bunch of pricks. When they found out I wasn't calling from Hughes
Aircraft or Boeing they wouldn't even quote me a price.

Back to GNAT, it is often not (easily) available on many POSIX platforms and
architectures. We have discussed Librecore's abandonment of Solaris
recently. There is still a lot of SPARC UNIX being used and no place to get
a trusted build and no instructions on building yourself. I'm not even sure
there is an Intel Solaris GNAT available anywhere. Is it really that hard
for people who "have the technology" to make this available? Yes, we are
willing to pay for something as good as GNAT, just not $20,000. Not for one
person who isn't going to make any money on it. gcc Ada would be fine (until
Stallman drops his GPL bomb on the libraries) but it's only available on
Linux and Windows. There are other OS in the world. At least if there would
be instructions somewhere how to build it, it would be used a lot more. I
guess nobody in a position to do anything about that cares.

> What is needed for Ada, is more Ada libraries, not more Ada
> compilers.

I agree we need more non-GPL libraries but I think we need more compilers
and more competition. "Libre" is strangling the business and guess what, the
US government paid for GNAT to be developed in the first place. I'd like to
hear how that becomes a whole privately held company but corruption and
graft in the public sector is nothing new.

By the way, no offense intended to Dr. Dewar. Nice job if you can get it. I
blame the guys who funded it and then allowed it to become owned by one
company.

Many people on comp.lang.ada have expressed an interest in buying a
development system. But so far none of the vendors except RR are willing to
talk to that kind of potential customer. I tried RR a few years ago but the
sample executables already didn't run on my 64 bit Linux. It's hard to tell
from the website that anything much has happened since 1995. If I didn't
read this newsgroup once in awhile I would have no idea the product is still
being updated. I would like to hear more updates and what platforms the
current products work on. I don't think anybody would object to that here
because Mr. Brukhardt is a valued member and contributes plenty to the
discussion, he is obviously not selling his products on the newsgroup. I
wish he would a little though so we can understand exactly what is
offered.

Most people who buy any kind of compiler also need a good GUI debugger
to go with it. If that isn't available many people probably won't be
interested. I don't know what AONIX pricing is or if they have any suitable
versions.

I don't like to bother people if I am not sure I am going to spend the money
so personally I like websites with full info on supported environments and
simple to understand all-included products rather than "Contact US and a
salesman will call" or "Request a quote". And yes I do buy software and have
purchased a non-Ada development system in the past. I realize stuff does
cost people money to develop and I don't expect to get anything handed to
me. On the other hand I don't expect to have to beg anyone to sell me
something, indeed I don't ever do that.

Yannick Duchêne (Hibou57)

unread,
May 20, 2012, 5:15:45 AM5/20/12
to
Le Sun, 20 May 2012 10:12:44 +0200, Nomen Nescio <nob...@dizum.com> a
écrit:
> Green Hills (don't know about HQ but the international office I spoke
> with)
> are a bunch of pricks. When they found out I wasn't calling from Hughes
> Aircraft or Boeing they wouldn't even quote me a price.

Ouch. If it didn't change, at AdaCore they are way kinder than that (try
to call them at your country's office, if you ever feel a need).

Simon Wright

unread,
May 20, 2012, 6:56:35 AM5/20/12
to
Nomen Nescio <nob...@dizum.com> writes:

> GNAT is certainly very good. But it is a pain in the ass in many ways. It is
> GPL only and even if you use gcc Ada eventually that is going to cause
> problems as GPL moves from GPL2 to 3, and who knows what's next. LGPL
> libraries today, tomorrow everythings GPL, you trusted them, you put all
> your eggs in one basket, you got screwed.

Not sure it's quite as black as that?

FSF GCC has been GPLv3 for a couple of years now, and its libraries
(excepting classpath) are released with the GCC Runtime Library
Exception: for example, C++ headers say

// Under Section 7 of GPL version 3, you are granted additional
// permissions described in the GCC Runtime Library Exception, version
// 3.1, as published by the Free Software Foundation.

FSF seem to have gone to a lot of trouble to make this change[1].

> Back to GNAT, it is often not (easily) available on many POSIX platforms and
> architectures. We have discussed Librecore's abandonment of Solaris
> recently. There is still a lot of SPARC UNIX being used and no place to get
> a trusted build and no instructions on building yourself. I'm not even sure
> there is an Intel Solaris GNAT available anywhere.

There's a GNAT GPL 2007 at [2]. It would take a while (for someone who
had Solaris x86 set up for it, and the inclination) to go from here to
an up-to-date compiler, but it should be possible.

> gcc Ada would be fine (until
> Stallman drops his GPL bomb on the libraries) but it's only available on
> Linux and Windows. There are other OS in the world. At least if there would
> be instructions somewhere how to build it, it would be used a lot
> more.

It's certainly available on Mac OS X; for instructions, see [3].

[1] http://www.gnu.org/licenses/gcc-exception-3.1-faq.html
[2] http://sourceforge.net/projects/gnuada/files/GNAT_GPL%20Solaris%2010/
[3] http://forward-in-code.blogspot.co.uk/2011/11/building-gcc-again.html

Simon Clubley

unread,
May 20, 2012, 7:30:06 AM5/20/12
to
On 2012-05-20, Nomen Nescio <nob...@dizum.com> wrote:
>
> GNAT is certainly very good. But it is a pain in the ass in many ways. It is
> GPL only and even if you use gcc Ada eventually that is going to cause
> problems as GPL moves from GPL2 to 3, and who knows what's next. LGPL
> libraries today, tomorrow everythings GPL, you trusted them, you put all
> your eggs in one basket, you got screwed.
>

I am assuming above that by "gcc Ada", you mean the FSF branch; I use the
FSF branch in order to avoid GPL issues and because when I need a RTOS, the
RTOS I use (RTEMS) also uses it.

The GPL 2 to GPL 3 comment has caught my attention. Given that the FSF Ada
runtime libraries come with the GMGPL exception, what issues are raised
by the GPL 2 to GPL 3 conversion ? I am assuming I have missed something,
but I don't know what.

On a more general note, one of the problems is that if you want to do hard
real time/bare metal/low level programming using a Wirth type language,
which I do, then Ada appears to be the only viable choice. I would really
like to find another choice as a backup option, but what I have looked at
so far (Modula-2, a Oberon variant, etc) appear to all be one-off ports
or lacking in other ways.

There are more options available if you want to do non-realtime application
level programming under a mainstream operating system, but for system level
work, Ada appears to be the only real viable option. Has anyone here used
another Wirth type language in a real time system level environment, and if
so, what kind of experience did you have ?

>
> Back to GNAT, it is often not (easily) available on many POSIX platforms and
> architectures. We have discussed Librecore's abandonment of Solaris
> recently. There is still a lot of SPARC UNIX being used and no place to get
> a trusted build and no instructions on building yourself. I'm not even sure
> there is an Intel Solaris GNAT available anywhere. Is it really that hard
> for people who "have the technology" to make this available? Yes, we are
> willing to pay for something as good as GNAT, just not $20,000. Not for one
> person who isn't going to make any money on it. gcc Ada would be fine (until
> Stallman drops his GPL bomb on the libraries) but it's only available on
> Linux and Windows.

Can you point to this "GPL bomb" issue, please ? It's something I need to
become aware of and a quick search didn't reveal anything. BTW, wouldn't it
affect C++ code just as much as Ada code ? If that's the case, then we may
end up with a XFree86/X.Org type situation once again.

BTW, as well as the native support, FSF Ada does work as a cross compiler
for a couple of targets. I have used it in the past to run Ada code on AVR
MCUs and more recently, on ARM based MCUs using RTEMS.

Simon.

--
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Microsoft: Bringing you 1980s technology to a 21st century world

Britt

unread,
May 20, 2012, 10:24:58 PM5/20/12
to
The "GPL Bomb" in this context is a "FUD". I just downloaded the GNAT
FSF sources contained in the latest GCC 4.7.0 from gnu.org and looked
in the header of the system.ads and other runtime files. They all
contain this:

-- As a special exception under Section 7 of GPL version 3, you are
granted --
-- additional permissions described in the GCC Runtime Library
Exception, --
-- version 3.1, as published by the Free Software
Foundation. --

which is the same as the file headers in the current GNAT Pro sources
(but different than the GNAT GPL sources).

The GCC Runtime Library Exception is here: http://www.gnu.org/licenses/gcc-exception.html

It doesn't seem any different in effect than the older GMGPL runtime
exception statement except that the definitions and wording in the
full text of the new runtime library exception make for a clearer and
more complete statement of intent. The key statements are:

"When you use GCC to compile a program, GCC may combine portions of
certain GCC header files and runtime libraries with the compiled
program. The purpose of this Exception is to allow compilation of non-
GPL (including proprietary) programs to use, in this way, the header
files and runtime libraries covered by this Exception."

and

"1. Grant of Additional Permission.
You have permission to propagate a work of Target Code formed by
combining the Runtime Library with Independent Modules, even if such
propagation would otherwise violate the terms of GPLv3, provided that
all Target Code was generated by Eligible Compilation Processes. You
may then convey such a combination under terms of your choice,
consistent with the licensing of the Independent Modules"

Simon Clubley

unread,
May 21, 2012, 8:12:08 AM5/21/12
to
On 2012-05-20, Britt <britt.s...@gmail.com> wrote:
> The "GPL Bomb" in this context is a "FUD". I just downloaded the GNAT
> FSF sources contained in the latest GCC 4.7.0 from gnu.org and looked
> in the header of the system.ads and other runtime files. They all
> contain this:
>
> -- As a special exception under Section 7 of GPL version 3, you are
> granted --
> -- additional permissions described in the GCC Runtime Library
> Exception, --
> -- version 3.1, as published by the Free Software
> Foundation. --
>
> which is the same as the file headers in the current GNAT Pro sources
> (but different than the GNAT GPL sources).
>

Which is basically the same as the last time I looked.

My concern here is that the OP may found discussions of a possible license
change on, say, a mailing list somewhere which has not yet been implemented.

A search with google has not turned up anything however; I have only found
the discussions about allowing gcc plugins around the time of the 4.4 series.

NatarovVI

unread,
May 21, 2012, 11:36:42 AM5/21/12
to
> It is not that Ada is dying (or Fortran or few others for that matter),
> it is that all the developers are/have been moving to the exciting new
> world of mobile and rich internet software development, using HTML5 and
> Javascript.

looks more like developers OR good OR "moves to exciting... html5 and
javascript" ))

> Windows 8 all new apps development is/will be in HTML5 and Javascript.
> This is the new world order of software engineering. After 50 years of
> computer science and programming languages research, the world has
> decided it will be Javascript and HTML5.

oh my... marketing say "we lead you to perfect new world"...
again...))

NatarovVI

unread,
May 21, 2012, 11:46:49 AM5/21/12
to
> Well, it is progress I guess. The Java VM was replaced by the browser,
> and Java is replaced by HTML/Javascript.

i think javascript is your bad karma))

> I can't wait to see what the picture will be 20 years from now :)
> --Nasser

essentially? same is. because people sins remains the same.
lasiness, greed, egoism...

Stephen Leake

unread,
May 22, 2012, 8:01:20 AM5/22/12
to
Simon Clubley <clubley@remove_me.eisner.decus.org-Earth.UFP> writes:

> On 2012-05-20, Nomen Nescio <nob...@dizum.com> wrote:
>>
>> GNAT is certainly very good. But it is a pain in the ass in many ways. It is
>> GPL only and even if you use gcc Ada eventually that is going to cause
>> problems as GPL moves from GPL2 to 3, and who knows what's next. LGPL
>> libraries today, tomorrow everythings GPL, you trusted them, you put all
>> your eggs in one basket, you got screwed.
>>
>
> I am assuming above that by "gcc Ada", you mean the FSF branch; I use the
> FSF branch in order to avoid GPL issues and because when I need a RTOS, the
> RTOS I use (RTEMS) also uses it.
>
> The GPL 2 to GPL 3 comment has caught my attention. Given that the FSF Ada
> runtime libraries come with the GMGPL exception, what issues are raised
> by the GPL 2 to GPL 3 conversion ?

None; he's spreading FUD.

> I am assuming I have missed something, but I don't know what.

The language has changed in detail, because GPL 3 provides a general
mechanism for specifying things like GMGPL. But the meaning is the same;
the GNAT runtime and generics are open source, but usuable in a
proprietary system.

AdaCore provides the same exception as FSF GNAT on their GNATPro
sources; that's their core business, so you can believe they got it
right!

--
-- Stephe

Nomen Nescio

unread,
May 22, 2012, 12:43:22 PM5/22/12
to
Simon Clubley <clubley@remove_me.eisner.decus.org-Earth.UFP> wrote:

> I am assuming above that by "gcc Ada", you mean the FSF branch; I use the
> FSF branch in order to avoid GPL issues and because when I need a RTOS, the
> RTOS I use (RTEMS) also uses it.

Yes, that's what I mean.

> The GPL 2 to GPL 3 comment has caught my attention. Given that the FSF Ada
> runtime libraries come with the GMGPL exception, what issues are raised
> by the GPL 2 to GPL 3 conversion ? I am assuming I have missed something,
> but I don't know what.

It could be FUD but I have read FSF is eventually planning to do away with
the LGPL for libraries and has started moving to GPL3 for everything. I
personally wouldn't put anything past Stallman. I think it's only a matter
of time. He knows what he can get away with and what he can't. He's just
biding his time. It gives the FSF guys indigestion knowing there is any
library LGPLd.

> On a more general note, one of the problems is that if you want to do hard
> real time/bare metal/low level programming using a Wirth type language,
> which I do, then Ada appears to be the only viable choice. I would really
> like to find another choice as a backup option, but what I have looked at
> so far (Modula-2, a Oberon variant, etc) appear to all be one-off ports
> or lacking in other ways.

I don't know what to say on this. Maybe Modula-3 is worth a look?

> Can you point to this "GPL bomb" issue, please ? It's something I need to
> become aware of and a quick search didn't reveal anything. BTW, wouldn't
> it affect C++ code just as much as Ada code ? If that's the case, then we
> may end up with a XFree86/X.Org type situation once again.

Following various various newsgroups and posts for the last few years this
is my feeling. Nobody comes out and says it. Yes, it will affect all the gcc
stuff, not just Ada. I think projects like FreeBSD are aware of it and
concerned and that is also motivating the move to clang (llvm). The BSD
projects will start moving off gcc and then gcc will eventually reduce
support for all non-Linux platforms (most people using gcc don't realize
anything but Linux exists anyway) and then nobody will be left to object to
everything being plain GPL 3 (or 4 or 5..)

Thanks to you and the other Simon for the info. I'd like to get at least gcc
Ada with lgpl libraries to run on Solaris.

Yannick Duchêne (Hibou57)

unread,
May 22, 2012, 7:25:12 PM5/22/12
to
Le Tue, 22 May 2012 14:01:20 +0200, Stephen Leake
<stephe...@stephe-leake.org> a écrit:
>> The GPL 2 to GPL 3 comment has caught my attention. Given that the FSF
>> Ada
>> runtime libraries come with the GMGPL exception, what issues are raised
>> by the GPL 2 to GPL 3 conversion ?
>
> None; he's spreading FUD.
>
>> I am assuming I have missed something, but I don't know what.
>
> The language has changed in detail, because GPL 3 provides a general
> mechanism for specifying things like GMGPL. But the meaning is the same;

Do I understand correctly if I understand there will be no need for the
GMGPL specific text anymore, and now just simply the GPL v3 exception will
do the trick?

Yannick Duchêne (Hibou57)

unread,
May 22, 2012, 7:36:36 PM5/22/12
to
Le Tue, 22 May 2012 18:43:22 +0200, Nomen Nescio <nob...@dizum.com> a
écrit:
> It gives the FSF guys indigestion knowing there is any
> library LGPLd.

At least, this one is not FUD (lol). I myself, have read some text from
the FSF which left me with a similar feeling. Anyway, there is still no
real‑life issue with the GPL and GNAT. If there are issues, these are
elsewhere, like with suspected (erroneous or not) intents, political
opinions and others. I don't believe the runtime libraries will ever
become pure GPL, because as much they would like, if they could, make
every thing pure GPL, *they are as much aware they can't* and aware doing
so will be counterproductive for the FSF lobbying, as making runtime
libraries pure GPL, would prevent many people from using it at all. They
know it, and although they don't like it, they have to do with it.
Conclusion: it's unlikely there will ever be an concrete issue with it, at
least not before a rather far future.

May be you are less spreading FUD (which is always too bad to say to
someone) than doing unprobable extrapolations, for some reasons you are
the only one to know.

Britt

unread,
May 22, 2012, 9:00:25 PM5/22/12
to
On May 22, 12:43 pm, Nomen Nescio <nob...@dizum.com> wrote:
> Simon Clubley <clubley@remove_me.eisner.decus.org-Earth.UFP> wrote:
> > I am assuming above that by "gcc Ada", you mean the FSF branch; I use the
> > FSF branch in order to avoid GPL issues and because when I need a RTOS, the
> > RTOS I use (RTEMS) also uses it.
>
> Yes, that's what I mean.
>
> > The GPL 2 to GPL 3 comment has caught my attention. Given that the FSF Ada
> > runtime libraries come with the GMGPL exception, what issues are raised
> > by the GPL 2 to GPL 3 conversion ? I am assuming I have missed something,
> > but I don't know what.
>
> It could be FUD but I have read FSF is eventually planning to do away with
> the LGPL for libraries and has started moving to GPL3 for everything. I
> personally wouldn't put anything past Stallman. I think it's only a matter
> of time. He knows what he can get away with and what he can't. He's just
> biding his time. It gives the FSF guys indigestion knowing there is any
> library LGPLd.
>
<snip>
>
> Thanks to you and the other Simon for the info. I'd like to get at least gcc
> Ada with lgpl libraries to run on Solaris.

It seems you are confused, as I once was*, about GNAT and the "Lesser
GPL" (LGPL). The LGPL has never been applied to GNAT. GNAT Pro and FSF
GNAT have always been licensed as either GPLv2 with the GMGPL special
exception or, currently, GPLv3 with the GCC Runtime Library Exception
v3.1. The relatively recent GNAT GPL editions are GPLv3 without the
exception. So no LGPL in the mix anywhere.

Here is a good reference on the topic:
http://people.debian.org/~lbrenta/debian-ada-policy.html#The-variants-of-GNAT

* See this very old thread where I was wrong and got re-calibrated:
https://groups.google.com/group/comp.lang.ada/browse_frm/thread/f6ad09be517b338c/e716c9bfdc1b0612?#e716c9bfdc1b0612

Fritz Wuehler

unread,
May 23, 2012, 8:16:19 AM5/23/12
to
Yannick Duchêne (Hibou57
)
<yannick...@yahoo.fr> wrote:

> opinions and others. I don't believe the runtime libraries will ever
> become pure GPL, because as much they would like, if they could, make
> every thing pure GPL, *they are as much aware they can't* and aware doing
> so will be counterproductive for the FSF lobbying, as making runtime
> libraries pure GPL, would prevent many people from using it at all.

I explained how this will work. Right now everyone realizes Linux is the
major focus of FSF because Linux kernel is GPL, Linux people love GPL,
etc. Autotools is supposed to work cross platform but it's mostly for Linux.

The BSD groups (Open, Free, Net, DragonFly) don't like GPL and they are
moving at different rates to clang/LLVM and other ways of reducing
dependence on gnu toolchain and other GPL code and have very little GPL
compared to Linux.

What I think is going to happen is as everybody who doesn't like GPL moves
off GPL-anything, the people who like GPL (Linux developers and users) will
not be upset at all, indeed they will be very happy, to see all their
libraries GPLd. So no, FSF can't go GPL libraries today, but things are
moving in the right direction (for everybody, really) so sooner or later
only Linux code will be GPL, including all the libraries, and BSD and other
projects will have virtually no GPL code. Then there will be no war and
nobody upset when FSF GPLs all the libraries.

> May be you are less spreading FUD (which is always too bad to say to
> someone) than doing unprobable extrapolations, for some reasons you are
> the only one to know.

I've been right about stuff like this before and I think the extrapolations
are reasonable conclusions based on actual trends. Having a major project
like FreeBSD openly state their goal to get off gcc and move to clang/LLVM
is something that will have far reaching effects. Right now it is possible
to compile ALL of FreeBSD with clang. This shows people there is no reason
to be serfs to Stallman or enslaved by GPL, or FSF, etc. That's REAL
freedom!


Shark8

unread,
May 23, 2012, 5:44:15 PM5/23/12
to
On Tuesday, May 22, 2012 11:43:22 AM UTC-5, Nomen Nescio wrote:
> Simon Clubley <clubley@remove_me.eisner.decus.org-Earth.UFP> wrote:
> > On a more general note, one of the problems is that if you want to do hard
> > real time/bare metal/low level programming using a Wirth type language,
> > which I do, then Ada appears to be the only viable choice. I would really
> > like to find another choice as a backup option, but what I have looked at
> > so far (Modula-2, a Oberon variant, etc) appear to all be one-off ports
> > or lacking in other ways.
>
> I don't know what to say on this. Maybe Modula-3 is worth a look?

You might also want to look into Oberon.

Stephen Leake

unread,
May 23, 2012, 9:12:27 PM5/23/12
to
"Yannick Duchêne (Hibou57)" <yannick...@yahoo.fr> writes:

> Le Tue, 22 May 2012 14:01:20 +0200, Stephen Leake
> <stephe...@stephe-leake.org> a écrit:
>>> The GPL 2 to GPL 3 comment has caught my attention. Given that the
>>> FSF Ada
>>> runtime libraries come with the GMGPL exception, what issues are raised
>>> by the GPL 2 to GPL 3 conversion ?
>>
>> None; he's spreading FUD.
>>
>>> I am assuming I have missed something, but I don't know what.
>>
>> The language has changed in detail, because GPL 3 provides a general
>> mechanism for specifying things like GMGPL. But the meaning is the same;
>
> Do I understand correctly if I understand there will be no need for
> the GMGPL specific text anymore, and now just simply the GPL v3
> exception will do the trick?

Yes.

--
-- Stephe

Stephen Leake

unread,
May 23, 2012, 9:14:43 PM5/23/12
to
Nomen Nescio <nob...@dizum.com> writes:

> Simon Clubley <clubley@remove_me.eisner.decus.org-Earth.UFP> wrote:
>
>> I am assuming above that by "gcc Ada", you mean the FSF branch; I use the
>> FSF branch in order to avoid GPL issues and because when I need a RTOS, the
>> RTOS I use (RTEMS) also uses it.
>
> Yes, that's what I mean.
>
>> The GPL 2 to GPL 3 comment has caught my attention. Given that the FSF Ada
>> runtime libraries come with the GMGPL exception, what issues are raised
>> by the GPL 2 to GPL 3 conversion ? I am assuming I have missed something,
>> but I don't know what.
>
> It could be FUD but I have read FSF is eventually planning to do away with
> the LGPL for libraries and has started moving to GPL3 for everything.

Actually, "GPL3 + library exceptions".

The point is that GPL 3 provides a mechanism to specify exceptions, so
the intent of LGPL (which is similar to the intent of GMGPL) can be
achieved by GPL 3 plus an appropriate exception.

--
-- Stephe

Yannick Duchêne (Hibou57)

unread,
May 24, 2012, 3:57:59 AM5/24/12
to
Le Wed, 23 May 2012 03:00:25 +0200, Britt <britt.s...@gmail.com> a
écrit:
Also of interest in that topic context, is the appendix D, at the very
bottom of the same page.

Simon Wright

unread,
May 24, 2012, 1:56:58 PM5/24/12
to
Simon Wright <si...@pushface.org> writes:

> There's a GNAT GPL 2007 at [2]. It would take a while (for someone who
> had Solaris x86 set up for it, and the inclination) to go from here to
> an up-to-date compiler, but it should be possible.

I've built[1] GCC 4.7.0 for Solaris 11 i386/amd64 and uploaded[2] the
result.

[1] http://forward-in-code.blogspot.co.uk/2012/05/building-gcc-with-ada-on-solaris-x86.html
[2] https://sourceforge.net/projects/gnuada/files/GNAT_GCC%20Solaris%2011%20i386%2C%20amd64/

Nomen Nescio

unread,
May 24, 2012, 5:58:03 PM5/24/12
to
Simon Wright <si...@pushface.org> wrote:

> Simon Wright <si...@pushface.org> writes:
>
> > There's a GNAT GPL 2007 at [2]. It would take a while (for someone who
> > had Solaris x86 set up for it, and the inclination) to go from here to
> > an up-to-date compiler, but it should be possible.

There is no longer any GNAT for Solaris on Libre. The only thing they have
is SPARK for SPARC.
Thank you very much. Unfortunately I'm using Solaris 10. I don't know if
that will work. But even more, I would like to understand how to build it
myself so I don't have to rely on anybody. I'll look at your packages,
and hopefully you explained how you did it on your blog. If so, I'll try it
myself.

Simon Clubley

unread,
May 25, 2012, 7:58:37 AM5/25/12
to
On 2012-05-23, Fritz Wuehler <fr...@spamexpire-201205.rodent.frell.theremailer.net> wrote:
> <yannick...@yahoo.fr> wrote:
>
>> opinions and others. I don't believe the runtime libraries will ever
>> become pure GPL, because as much they would like, if they could, make
>> every thing pure GPL, *they are as much aware they can't* and aware doing
>> so will be counterproductive for the FSF lobbying, as making runtime
>> libraries pure GPL, would prevent many people from using it at all.
>
> I explained how this will work. Right now everyone realizes Linux is the
> major focus of FSF because Linux kernel is GPL, Linux people love GPL,
> etc. Autotools is supposed to work cross platform but it's mostly for Linux.
>
> The BSD groups (Open, Free, Net, DragonFly) don't like GPL and they are
> moving at different rates to clang/LLVM and other ways of reducing
> dependence on gnu toolchain and other GPL code and have very little GPL
> compared to Linux.
>
> What I think is going to happen is as everybody who doesn't like GPL moves
> off GPL-anything, the people who like GPL (Linux developers and users) will
> not be upset at all, indeed they will be very happy, to see all their
> libraries GPLd. So no, FSF can't go GPL libraries today, but things are
> moving in the right direction (for everybody, really) so sooner or later
> only Linux code will be GPL, including all the libraries, and BSD and other
> projects will have virtually no GPL code. Then there will be no war and
> nobody upset when FSF GPLs all the libraries.
>

If this played out as you describe, then the day that the FSF applies the
pure GPL (without any exceptions, including the LGPL, allowed) to it's
libraries will be the day that commercial closed source development will
end on Linux. Day+1 will be the day that RedHat, with it's large commercial
user base (and support contracts to match) will fork the libraries in
question under the original license and commercial development on Linux
will resume, but with the FSF no longer been relevant.

This has happened before and can happen again. (XFree86 -> X.Org)

For this reason, I am not worried about the FSF making everything pure GPL
even though some may wish to. However, a more targetted move to the GPL,
such as when ACT removed the GMGPL exception on the public version of
GtkAda, is always possible, but I don't see any evidence of anyone moving
in this direction yet with GNAT itself.

However, it's always nice to have backup options (even though I prefer Ada)
in case this does happen, so thanks to everyone for their suggestions.

Nicholas Paul Collin Gloucester

unread,
Jun 21, 2012, 7:28:37 AM6/21/12
to
PowerAda from OC Systems
AdaMagic

Nicholas Paul Collin Gloucester

unread,
Jun 21, 2012, 8:05:20 AM6/21/12
to
Oh, I forgot to mention MAXAda but I never had access to it.
Reply all
Reply to author
Forward
0 new messages