Le 30/09/2021 à 09:53, Luke A. Guest a écrit :
>
> On 30/09/2021 07:23, J-P. Rosen wrote:
>
>>> Freely available iso asis spec would help here.
>> Actually, it is. Apart from ISO verbiage, all the interesting parts of
>> the ASIS standard is put as comments in the corresponding ASIS packages.
>>
>> Moreover, AdaCore kept this good habit for all the newly introduced
>> features that support up to Ada 2012, which would make retrofitting
>> them into an updated ASIS standard quite easy.
>>
>
> Are they GPL'd and where are they?

You can find them in the specifications of the various packages, with sentinels (as indicated in my previous message). Another excerpt:

-- Suggestions related to changing this specification to accept new Ada
-- features as defined in incoming revision of the Ada Standard (ISO 8652)
-- are marked by following comment sentinels:
--
-- --|A2005 start
-- ... the suggestion goes here ...
-- --|A2005 end
--
-- and the discussion items are marked by the comment sentinels of teh form:
--
-- --|D2005 start
-- ... the discussion item goes here ...
-- --|D2005 end

(and the same goes for 2012).

--
J-P. Rosen
Adalog
2 rue du Docteur Lombard, 92441 Issy-les-Moulineaux CEDEX
Tel: +33 1 45 29 21 52
https://www.adalog.fr

> But it's not available from AdaCore's community page. For most users,
> downloading and building from an FSF site is way too complicated. Call
> it asis-gcc or not, what is needed is a simple way to install ASIS support.

We have decided in any case to stop creating and distributing GNAT Community binaries, since this was causing too much confusion and misunderstanding wrt the license, so doing in the end more harm than good to the community, which we care very much about.

So in the future, GNAT will be available directly and only from the FSF versions, and Alire will make that easy.

Alire (https://alire.ada.dev/) already provides GCC 10.3 today, see e.g. "alr toolchain --select"

> (Making a tree generator separate from the compiler is for me another
> error, although I can live with it. One of the main benefits of ASIS is
> that the ASIS program has the same view of the code as the compiler -
> but that's a separate issue).

Right, and has never been the case for cross compilers where you already needed a native GNAT to build your ASIS application, and a cross GNAT to generate trees.

Arno

On 30/09/2021 09:13, J-P. Rosen wrote:
> Le 30/09/2021 à 09:53, Luke A. Guest a écrit :
>> On 30/09/2021 07:23, J-P. Rosen wrote:
>>> Le 30/09/2021 à 01:29, Luke A. Guest a écrit :
>>>>
>>>> On 29/09/2021 20:04, Emmanuel Briot wrote:
>>>>
>>>>> I suggested in an early message that perhaps the community could
>>>>> build an ASIS API on top of libadalang, if there is a need for that.
>>>>> I also suggested that libadalang documentation should be improved,
>>>>> I definitely agree with that one !
>>>>>
>>>>
>>>> Freely available iso asis spec would help here.
>>> Actually, it is. Apart from ISO verbiage, all the interesting parts
>>> of the ASIS standard is put as comments in the corresponding ASIS
>>> packages.
>>>
>>> Moreover, AdaCore kept this good habit for all the newly introduced
>>> features that support up to Ada 2012, which would make retrofitting
>>> them into an updated ASIS standard quite easy.
>>>
>>
>> Are they gpl'd?
> Yes. Here is a copy of the copyright notice of every ASIS module:

And that's an issue, why not release them PD or BSD? I've seen the asis
specs before and I'm certain they are not GPL'd, just like the packages
in the Ada RM.

On Thursday, September 30, 2021 at 9:52:39 AM UTC+2, J-P. Rosen wrote:
> But it's not available from AdaCore's community page. For most users,
> downloading and building from an FSF site is way too complicated. Call
> it asis-gcc or not, what is needed is a simple way to install ASIS support.

There are plenty of GNAT FSF 10 builds available:
- Linux distribs (Ubuntu/Debian, Arch, Fedora, and probably others that I don't know about)
- msys2 for Windows
- Simon Wright's builds for macOS
- Alire for Linux, Windows and macOS

Availability of GNAT FSF 10 is not an issue.

On 30/09/2021 09:16, J-P. Rosen wrote:
> Le 30/09/2021 à 09:53, Luke A. Guest a écrit :
>>
>> On 30/09/2021 07:23, J-P. Rosen wrote:
>>
>>>> Freely available iso asis spec would help here.
>>> Actually, it is. Apart from ISO verbiage, all the interesting parts
>>> of the ASIS standard is put as comments in the corresponding ASIS
>>> packages.
>>>
>>> Moreover, AdaCore kept this good habit for all the newly introduced
>>> features that support up to Ada 2012, which would make retrofitting
>>> them into an updated ASIS standard quite easy.
>>>
>>
>> Are they GPL'd and where are they?
>
> You can find them in the specifications of the various packages, with
> sentinels (as indicated in my previous message). Another excerpt:
>
> --  Suggestions related to changing this specification to accept new Ada
> --  features as defined in incoming revision of the Ada Standard (ISO 8652)
> --  are marked by following comment sentinels:
> --
> --  --|A2005 start
> --   ... the suggestion goes here ...
> --  --|A2005 end
> --
> --  and the discussion items are marked by the comment sentinels of teh
> form:
> --
> --  --|D2005 start
> --   ... the discussion item goes here ...
> --  --|D2005 end
>
> (and the same goes for 2012).
>

I wanted to know where they are. I once found the entire directory of
asis specs from the iso doc, I think I have them somewhere still.

Where are the updated ones for post 95? There should be archive or
directory with them with no restrictive licensing comments.

Le 30/09/2021 à 10:26, Luke A. Guest a écrit :
>> Yes. Here is a copy of the copyright notice of every ASIS module:
>
> And that's an issue, why not release them PD or BSD? I've seen the asis
> specs before and I'm certain they are not GPL'd, just like the packages
> in the Ada RM.

If you are talking about the official ASIS specs ("like the packages in
the Ada RM"), they are part of an ISO standard, and as such under ISO
copyright. However, in the case of APIs, ISO allows their use by any
implementation (otherwise, they would be useless). This has of course
nothing common with the GPL or any other open license.

--
J-P. Rosen
Adalog
2 rue du Docteur Lombard, 92441 Issy-les-Moulineaux CEDEX
Tel: +33 1 45 29 21 52
https://www.adalog.fr

Le 30/09/2021 à 10:28, Luke A. Guest a écrit :
> I wanted to know where they are. I once found the entire directory of
> asis specs from the iso doc, I think I have them somewhere still.
>
> Where are the updated ones for post 95? There should be archive or
> directory with them with no restrictive licensing comments.

Just download ASIS for gnat CE 2019.

--
J-P. Rosen
Adalog
2 rue du Docteur Lombard, 92441 Issy-les-Moulineaux CEDEX
Tel: +33 1 45 29 21 52
https://www.adalog.fr

On 30/09/2021 11:54, J-P. Rosen wrote:
> Le 30/09/2021 à 10:26, Luke A. Guest a écrit :
>>> Yes. Here is a copy of the copyright notice of every ASIS module:
>>
>> And that's an issue, why not release them PD or BSD? I've seen the
>> asis specs before and I'm certain they are not GPL'd, just like the
>> packages in the Ada RM.
>
> If you are talking about the official ASIS specs ("like the packages in
> the Ada RM"), they are part of an ISO standard, and as such under ISO
> copyright. However, in the case of APIs, ISO allows their use by any
> implementation (otherwise, they would be useless).

Exactly, same as the ARM packages.

> This has of course
> nothing common with the GPL or any other open license.
>

But the issue is, if the specs for the extended ASIS have only been
released under GPL, they are useless to any non-gpl language
implementations as their use infects that implementation causing further
issues.

This GPL issue is the reason why I've looked at, in the past, creating
my own compiler, and now just wanting to develop my own language that I
can use anywhere.

On 30/09/2021 11:56, J-P. Rosen wrote:
> Le 30/09/2021 à 10:28, Luke A. Guest a écrit :
>> I wanted to know where they are. I once found the entire directory of
>> asis specs from the iso doc, I think I have them somewhere still.
>>
>> Where are the updated ones for post 95? There should be archive or
>> directory with them with no restrictive licensing comments.
>
> Just download ASIS for gnat CE 2019.
>

No. See my other message.

Le 30/09/2021 à 14:27, Luke A. Guest a écrit :
>> If you are talking about the official ASIS specs ("like the packages
>> in the Ada RM"), they are part of an ISO standard, and as such under
>> ISO copyright. However, in the case of APIs, ISO allows their use by
>> any implementation (otherwise, they would be useless).
>
> Exactly, same as the ARM packages.
>
>> This has of course nothing common with the GPL or any other open license.
>>
>
> But the issue is, if the specs for the extended ASIS have only been
> released under GPL, they are useless to any non-gpl language
> implementations as their use infects that implementation causing further
> issues.
>
Right, currently AdaCore is the owner of these specifications. A
standardization effort would need a transfer of copyright, I hope that
AdaCore wouldn't object.

BTW, talking of copyright: LibAdalang has no header comment telling the
copyright status, therefore it is by default propriatory AdaCore!

--
J-P. Rosen
Adalog
2 rue du Docteur Lombard, 92441 Issy-les-Moulineaux CEDEX
Tel: +33 1 45 29 21 52
https://www.adalog.fr

Le 30/09/2021 à 14:27, Luke A. Guest a écrit :
> On 30/09/2021 11:56, J-P. Rosen wrote:
>> Le 30/09/2021 à 10:28, Luke A. Guest a écrit :
>>> I wanted to know where they are. I once found the entire directory of
>>> asis specs from the iso doc, I think I have them somewhere still.
>>>
>>> Where are the updated ones for post 95? There should be archive or
>>> directory with them with no restrictive licensing comments.
>>
>> Just download ASIS for gnat CE 2019.
>>
>
> No. See my other message.
>
If you want a separate, available document, I don't think there is. If
it is just out of curiosity, use ASIS for Gnat.

There is certainly work to do to get an updated standard!

--
J-P. Rosen
Adalog
2 rue du Docteur Lombard, 92441 Issy-les-Moulineaux CEDEX
Tel: +33 1 45 29 21 52
https://www.adalog.fr

"J-P. Rosen" <rosen@adalog.fr> wrote in message
news:sj3ktk$emb$1@dont-email.me...
....
> I'm afraid this is a red herring. I think rather that AdaCore has a hard
> time convincing people of moving from the well defined, carefully designed
> ASIS to the terrible mess of LibAdalang.

The ASIS design and definition is a mess (at least from the perspective of
explaining what is expected). We tried to clean it up in the previous ASIS
standardization update, but that was a lot of work and we probably didn't
match implementations very well.

The entire model of ASIS doesn't make much sense for static analysis
purposes, it's way too focused on syntax rather than semantics. And it
doesn't work well for syntax analysis because it requires a compilable
program. So it really has a very narrow use case (if any).

Your tool mainly proves that one can use anything with heroic enough
efforts. But the effort that your tools goes through to determine basic
semantics like whether a type is tagged demonstrates it's hardly a practical
way to build a tool. As far as I know, you're the only one that ever managed
to do anything beyond proof-of-concepts with ASIS. I can certainly see why
AdaCore might not want to support something solely for one usage.

I can easily believe that Libadalang is even more poorly defined than ASIS
(most vendor-generated things are, regardless of the vendor involved). I
would guess that the only way to build a tool like yours is to do your own
analysis (certainly, that is how I'd approach it). A true Ada Semantic
Interface would be a good thing, but ASIS isn't it.

Randy.

"J-P. Rosen" <rosen@adalog.fr> wrote in message
news:sj3rk5$na1$1@dont-email.me...
....
> -- This specification is adapted from the Ada Semantic
> Interface --
> -- Specification Standard (ISO/IEC 15291) for use with GNAT. In
> accordance --
> -- with the copyright of that document, you can freely copy and modify
> this --
> -- specification, provided that if you redistribute a modified version,
> any --
> -- changes that you have made are clearly
> --

Umm, someone is confusing the original ASIS drafts with the ISO Standard
(which has an ISO copyright with no exceptions). I would definitely not
reference the ISO Standard in anything you are freely giving away -- there
are copyright trolls out there that could easily decide to get your material
banned from the Internet.

For Ada, we are very carefully keeping the Ada Reference Manual as a
separate document from the ISO Standard, so that the Ada RM has the
permissive copyright while the ISO Standard for Ada definitely does not.
These are not the same thing!

That care was not taken for the ASIS Standard; I know of no public version
that was maintained. As such, my opinion is that ISO owns the copyright, and
any extensive use (like using all of the specs) would require a license from
ISO. This is by far the best reason for abandoning ASIS - I don't believe
that you can implement it without getting a license from ISO (since the bulk
of the ASIS Standard is Ada specifications, you are using too much to fall
under fair use). This is one reason that I would never consider implementing
ASIS in Janus/Ada.

>> Actually, it is. Apart from ISO verbiage, all the interesting parts of
>> the ASIS standard is put as comments in the corresponding ASIS packages.
>>
>> Moreover, AdaCore kept this good habit for all the newly introduced
>> features that support up to Ada 2012, which would make retrofitting them
>> into an updated ASIS standard quite easy.

It's only easy if you think that giving AdaCore's work to ISO under the
exclusive copyright that they (ISO) will insist on is something that is
legally and ethically appropriate.

You need to come to grips with the reality that ASIS is dead. It's legally
dangerous to implement it, it isn't a good match for either syntax or
semantic analysis (doing neither very well), and it is poor match for modern
compilers (hardly anyone builds trees much like the ASIS ones, unless you
are trying to implement ASIS).

Randy.

"J-P. Rosen" <rosen@adalog.fr> wrote in message
news:sj4519$rg0$1@dont-email.me...
....
> If you are talking about the official ASIS specs ("like the packages in
> the Ada RM"), they are part of an ISO standard, and as such under ISO
> copyright. However, in the case of APIs, ISO allows their use by any
> implementation (otherwise, they would be useless).

This is news to me. As far as I can tell, any API that is the bulk ofan ISO
Standard *is* useless, because you need an ISO license to use it. Smaller
APIs might be usable as "fair use", but they can't be 90% of the Standard's
text, as in ASIS.

On what basis do you think that "ISO allows their use by any
implementation"; I see no such thing in any ISO Standard??

Randy.

"Luke A. Guest" <laguest@archeia.com> wrote in message
news:sj4ahc$e05$1@gioia.aioe.org...
> On 30/09/2021 11:54, J-P. Rosen wrote:
....
>> If you are talking about the official ASIS specs ("like the packages in
>> the Ada RM"), they are part of an ISO standard, and as such under ISO
>> copyright. However, in the case of APIs, ISO allows their use by any
>> implementation (otherwise, they would be useless).
>
> Exactly, same as the ARM packages.

No, you can use RM packages because we've carefully kept the original
copyrights and document intact; the ISO versions are *derived* from that
document. When you are implementing Ada, you are using the packages from the
RM, which have an appropriate exception.

That was not done for ASIS, and there is no appropriate exception to use the
libraries. And surely not to copy the extensive comments along with the spec
(even if you argue that fair use allows you to implement the spec -- and the
Oracle case appears to have ruled that there is no such right -- you don't
need to copy the comments that make up the bulk of the ASIS Standard).

Randy.

Le 01/10/2021 à 02:18, Randy Brukardt a écrit :
> "J-P. Rosen" <rosen@adalog.fr> wrote in message
> The ASIS design and definition is a mess (at least from the perspective of
> explaining what is expected). We tried to clean it up in the previous ASIS
> standardization update, but that was a lot of work and we probably didn't
> match implementations very well.
That was mainly an attempt to introduce more static and tagged typing,
and it failed due to the complexity involved (and that AdaCore said they
would never implement it). LibAdalang made the same error, and got the
same unnecessary complexity.

> The entire model of ASIS doesn't make much sense for static analysis
> purposes, it's way too focused on syntax rather than semantics.
It is a exact image of the program, from which you can derive all
information you need. Some higher level queries are needed, but they can
be provided as secondary queries or added to the standard.

> And it
> doesn't work well for syntax analysis because it requires a compilable
> program. So it really has a very narrow use case (if any).
On the contrary. There is no semantic you can analyze in a
non-compilable program. And since it analyzes the output of a validated
compiler, you can trust it better than any custom analyzer without known
pedigree.

> Your tool mainly proves that one can use anything with heroic enough
> efforts. But the effort that your tools goes through to determine basic
> semantics like whether a type is tagged demonstrates it's hardly a practical
> way to build a tool.
I'm afraid you are confused here. It is very easy to check whether a
type is tagged. You may confusing this with checking whether a type is
limited or not: yes, an extra query would be useful for this case. No
big deal.

> As far as I know, you're the only one that ever managed
> to do anything beyond proof-of-concepts with ASIS.
For years, AdaCore tools (gnatelim, gnatstub) used ASIS, not counting
Gnatcheck that has not yet been able to migrate to LibAadalang.
The interface generator of AWS is also based on ASIS. Out of the top of
my mind, I think certain document generators as well as some real-time
properties analyzers also use ASIS.

> I can certainly see why
> AdaCore might not want to support something solely for one usage.
> I can easily believe that Libadalang is even more poorly defined than ASIS
> (most vendor-generated things are, regardless of the vendor involved). I
> would guess that the only way to build a tool like yours is to do your own
> analysis (certainly, that is how I'd approach it). A true Ada Semantic
> Interface would be a good thing, but ASIS isn't it.
>
But you didn't use it. True, a first approach or a casual reading of the
interface is not very friendly. But the more you use it, the more you
realize that it is very consistently defined, and allows you to do
whatever you need.

--
J-P. Rosen
Adalog
2 rue du Docteur Lombard, 92441 Issy-les-Moulineaux CEDEX
Tel: +33 1 45 29 21 52
https://www.adalog.fr

Le 01/10/2021 à 02:30, Randy Brukardt a écrit :
> Umm, someone is confusing the original ASIS drafts with the ISO Standard
> (which has an ISO copyright with no exceptions). I would definitely not
> reference the ISO Standard in anything you are freely giving away -- there
> are copyright trolls out there that could easily decide to get your material
> banned from the Internet.
>
Strangely enough, my copy of ISO 15291 has no copyright statement at
all; might be a "last draft" version.

However, the headers of every ASIS-for-Gnat package state:
"This specification is adapted from the Ada Semantic
Interface Specification Standard (ISO/IEC 15291) for use with GNAT. In
accordance with the copyright of that document, you can freely copy and
modify this specification, provided that if you redistribute a modified
version, any changes that you have made are clearly indicated."

(and since that statement dates back to Robert Dewar's times, I'm pretty
certain it is reliable).

My memory is that all "interesting" part of the standard was deliberatly
put as comments in the specification, precisely to circumvent the ISO
copyright, and allow the use of ASIS without paying an outrageous price
to ISO.

--
J-P. Rosen
Adalog
2 rue du Docteur Lombard, 92441 Issy-les-Moulineaux CEDEX
Tel: +33 1 45 29 21 52
https://www.adalog.fr

Le 30/09/2021 à 10:21, Arnaud Charlet a écrit :
> We have decided in any case to stop creating and distributing GNAT Community binaries, since this was causing too much confusion and misunderstanding wrt the license, so doing in the end more harm than good to the community, which we care very much about.
>
> So in the future, GNAT will be available directly and only from the FSF versions, and Alire will make that easy.
>
> Alire (https://alire.ada.dev/) already provides GCC 10.3 today, see e.g. "alr toolchain --select"
>
And what will happen to other versions of Gnat that were useful for
promoting Ada, like JGnat and Lego-mindstorm? (I know you freezed these
some years ago, but it was very useful to be able to mention them).

And what will happen for fixes to asis-gcc? Will they be propagated to
GCC 10.3? Even after you move to GCC 11.x?

--
J-P. Rosen
Adalog
2 rue du Docteur Lombard, 92441 Issy-les-Moulineaux CEDEX
Tel: +33 1 45 29 21 52
https://www.adalog.fr

Le 30/09/2021 à 10:21, Arnaud Charlet a écrit :
> Alire (https://alire.ada.dev/) already provides GCC 10.3 today, see e.g. "alr toolchain --select"
And does it provide a matching version of ASIS?

--
J-P. Rosen
Adalog
2 rue du Docteur Lombard, 92441 Issy-les-Moulineaux CEDEX
Tel: +33 1 45 29 21 52
https://www.adalog.fr

"J-P. Rosen" <rosen@adalog.fr> wrote in message
news:sj6l3g$f9g$1@dont-email.me...
> Le 01/10/2021 � 02:30, Randy Brukardt a �crit :
>> Umm, someone is confusing the original ASIS drafts with the ISO Standard
>> (which has an ISO copyright with no exceptions). I would definitely not
>> reference the ISO Standard in anything you are freely giving away --
>> there
>> are copyright trolls out there that could easily decide to get your
>> material
>> banned from the Internet.
>>
> Strangely enough, my copy of ISO 15291 has no copyright statement at all;
> might be a "last draft" version.
>
> However, the headers of every ASIS-for-Gnat package state:
> "This specification is adapted from the Ada Semantic
> Interface Specification Standard (ISO/IEC 15291) for use with GNAT. In
> accordance with the copyright of that document, you can freely copy and
> modify this specification, provided that if you redistribute a modified
> version, any changes that you have made are clearly indicated."
>
> (and since that statement dates back to Robert Dewar's times, I'm pretty
> certain it is reliable).

I'm certain that is something that predates the ISO version of ASIS. There's
no such permission in the ISO document that I was sent as editor during our
last (aborted) revision attempt. Robert probably was using the pre-ISO
version as the source, all

> My memory is that all "interesting" part of the standard was deliberatly
> put as comments in the specification, precisely to circumvent the ISO
> copyright, and allow the use of ASIS without paying an outrageous price to
> ISO.

I don't see how using comments helps anything. The Oracle case makes it
pretty clear an API iteself can be covered by a copyright, and surely the
comments are covered by the copyright. And the ISO version has no copyright
statement other than the usual "All rights reserved".

Disclaimer: I am not a lawyer and cannot say anything for certain in these
matters.

Randy.

"J-P. Rosen" <rosen@adalog.fr> wrote in message
news:sj6k41$7i4$1@dont-email.me...
> Le 01/10/2021 � 02:18, Randy Brukardt a �crit :
>> "J-P. Rosen" <rosen@adalog.fr> wrote in message
>> The ASIS design and definition is a mess (at least from the perspective
>> of
>> explaining what is expected). We tried to clean it up in the previous
>> ASIS
>> standardization update, but that was a lot of work and we probably didn't
>> match implementations very well.
> That was mainly an attempt to introduce more static and tagged typing, and
> it failed due to the complexity involved (and that AdaCore said they would
> never implement it). LibAdalang made the same error, and got the same
> unnecessary complexity.

No, the existing part also had major problems. Many of the terms were never
defined, many possible effects were missing from the various lists of
results, and loads of other things as well. That all would have had to be
fixed even if the semantic interface was completely forgotten (indeed, it
was kept pretty separate for that reason).

>> The entire model of ASIS doesn't make much sense for static analysis
>> purposes, it's way too focused on syntax rather than semantics.
> It is a exact image of the program, from which you can derive all
> information you need. Some higher level queries are needed, but they can
> be provided as secondary queries or added to the standard.

That's exactly the problem. You start with the source code, which is way too
low a level for any useful analysis. At most, you want a simple connection
to the source in the semantic information, not trying to preserve every
punctuation mark and comment. (Janus/Ada discards all of that stuff as soon
as parsing succeeds.) If you need to refer to the original source, say for
error handling purposes, then do that, but don't waste vast amounts of space
and time trying to keep loads of irrelevant material.

>> And it
>> doesn't work well for syntax analysis because it requires a compilable
>> program. So it really has a very narrow use case (if any).
> On the contrary. There is no semantic you can analyze in a non-compilable
> program. And since it analyzes the output of a validated compiler, you can
> trust it better than any custom analyzer without known pedigree.

No sane compiler (validated or not) keeps all of the irrelevant syntactic
detail required by ASIS. It ends up getting reconstructed solely for the use
of ASIS, and how a rarely used interface is somehow more reliable escapes
me.

A true semantic interface on the lines of the one proposed for ASIS would
make good sense (design of types), but the vast majority of the existing
ASIS belongs in a rubbish bin. Good riddance.

...
> But you didn't use it. True, a first approach or a casual reading of the
> interface is not very friendly. But the more you use it, the more you
> realize that it is very consistently defined, and allows you to do
> whatever you need.

I don't use it because implementing it would require adding loads of useless
cruft to our Ada compiler. And even then, it doesn't make much sense based
on our compilation model and our generic unit model. Supporting it would be
like building a whole new Ada compiler. Ergo, it is a lie, it claims to be
an "interface to a compiler", but it requires many things that most
compilers would not waste time on. (I think it is a fairly close
representation of the internals of early Rational compilers, which is
probably why they were so slow and memory hogs. ;-) So what is it really?
Just a very complex way to do stuff that you can easily do with a parser. No
worth anyone's time, IMHO.

I've assumed most people used it because it was there and because some
people had spent a lot of time trying to define it as some sort of Standard.
Just because people put a lot of work into something doesn't mean that it is
a useful thing.

Randy.

Le 02/10/2021 à 11:34, Randy Brukardt a écrit :

> No, the existing part also had major problems. Many of the terms were never
> defined, many possible effects were missing from the various lists of
> results, and loads of other things as well.
Huh? ASIS uses the terms of, and as defined in, the LRM. And I never hit
a "missing result". Curious to see what you are aluding too.
>
>>> The entire model of ASIS doesn't make much sense for static analysis
>>> purposes, it's way too focused on syntax rather than semantics.
Yes, it is a description of the syntactic. Where else can you start from?

>> It is a exact image of the program, from which you can derive all
>> information you need. Some higher level queries are needed, but they can
>> be provided as secondary queries or added to the standard.
>
> That's exactly the problem. You start with the source code, which is way too
> low a level for any useful analysis. At most, you want a simple connection
> to the source in the semantic information, not trying to preserve every
> punctuation mark and comment. (Janus/Ada discards all of that stuff as soon
> as parsing succeeds.) If you need to refer to the original source, say for
> error handling purposes, then do that, but don't waste vast amounts of space
> and time trying to keep loads of irrelevant material.
The liaison to the source is not fundamental to any serious analysis,
and anyway you are not required to provide it.

> No sane compiler (validated or not) keeps all of the irrelevant syntactic
> detail required by ASIS. It ends up getting reconstructed solely for the use
> of ASIS, and how a rarely used interface is somehow more reliable escapes
> me.
Except for the not-required text interface, I see very few of these
"irrelevant syntactic details". OK, you have to keep a boolean to
remember if "end" is followed by a name. Big deal...

> A true semantic interface on the lines of the one proposed for ASIS would
> make good sense (design of types), but the vast majority of the existing
> ASIS belongs in a rubbish bin. Good riddance.
Says someone who didn't use or implement ASIS. BTW, I understand that
ASIS would be difficult to implement in Janus Ada, especially when it
comes to generic expansion. But it's not a reason to deprive others from
it...

>> [...] (I think it is a fairly close
> representation of the internals of early Rational compilers
True, the design was based on ideas from Diana. But it was designed with
inputs from various compilers.

> I've assumed most people used it because it was there and because some
> people had spent a lot of time trying to define it as some sort of Standard.
> Just because people put a lot of work into something doesn't mean that it is
> a useful thing.
It allowed me to build a very sophisticated tool, valued at 1.24M$ (see
https://www.adacontrol.fr), and used by very serious customers. Seems
enough to qualify it "useful".

--
J-P. Rosen
Adalog
2 rue du Docteur Lombard, 92441 Issy-les-Moulineaux CEDEX
Tel: +33 1 45 29 21 52
https://www.adalog.fr

Le 02/10/2021 à 11:14, Randy Brukardt a écrit :
>> My memory is that all "interesting" part of the standard was deliberatly
>> put as comments in the specification, precisely to circumvent the ISO
>> copyright, and allow the use of ASIS without paying an outrageous price to
>> ISO.
> I don't see how using comments helps anything. The Oracle case makes it
> pretty clear an API iteself can be covered by a copyright, and surely the
> comments are covered by the copyright. And the ISO version has no copyright
> statement other than the usual "All rights reserved".

1) It seems to me that you are confusing the copyright owner with the
right to use the interface. Undoubtedly, ISO is the copyright owner. But
they may authorize unlimited use of the specification, otherwise NO
standard would make sense. Do you infringe copyright if you build an
electrical plug that conforms to you electrical standard?

2) Comments help, because they describe precisely what is expected by
every function, and what it provides. Actually, I never open the ASIS
standard, everything I need is detailed in the comments.

--
J-P. Rosen
Adalog
2 rue du Docteur Lombard, 92441 Issy-les-Moulineaux CEDEX
Tel: +33 1 45 29 21 52
https://www.adalog.fr

"J-P. Rosen" <rosen@adalog.fr> wrote in message
news:sjerti$5s1$1@dont-email.me...
> Le 02/10/2021 � 11:34, Randy Brukardt a �crit :
>
>> No, the existing part also had major problems. Many of the terms were
>> never
>> defined, many possible effects were missing from the various lists of
>> results, and loads of other things as well.
> Huh? ASIS uses the terms of, and as defined in, the LRM. And I never hit a
> "missing result". Curious to see what you are aluding too.

Go back and read many of the SI99s. I'm not going to waste time doing it
again.

>>>> The entire model of ASIS doesn't make much sense for static analysis
>>>> purposes, it's way too focused on syntax rather than semantics.
> Yes, it is a description of the syntactic. Where else can you start from?

The semantics, of course. You have a list of entities in a scope. The only
interesting thing from the source code is the line/position of the
declaration and/or use. The details of the source are irrelevant (such as
whether optional keywords are given).

>>> It is a exact image of the program, from which you can derive all
>>> information you need. Some higher level queries are needed, but they can
>>> be provided as secondary queries or added to the standard.
>>
>> That's exactly the problem. You start with the source code, which is way
>> too
>> low a level for any useful analysis. At most, you want a simple
>> connection
>> to the source in the semantic information, not trying to preserve every
>> punctuation mark and comment. (Janus/Ada discards all of that stuff as
>> soon
>> as parsing succeeds.) If you need to refer to the original source, say
>> for
>> error handling purposes, then do that, but don't waste vast amounts of
>> space
>> and time trying to keep loads of irrelevant material.
> The liaison to the source is not fundamental to any serious analysis, and
> anyway you are not required to provide it.

It's a basic part of the ASIS definition, and I fail to see how you could do
much without it. You have to keep nonsense such as whether someone specified
"in" or a matching id after "end". A waste of time and space. You also have
to keep the entire program in a tree form, which is also dubious (Janus/Ada
certainly does not do that).

>> No sane compiler (validated or not) keeps all of the irrelevant syntactic
>> detail required by ASIS. It ends up getting reconstructed solely for the
>> use
>> of ASIS, and how a rarely used interface is somehow more reliable escapes
>> me.
> Except for the not-required text interface, I see very few of these
> "irrelevant syntactic details". OK, you have to keep a boolean to remember
> if "end" is followed by a name. Big deal...

It is a big deal when multiplied by dozens of such things. And that
information has to be carried through from the front end -- we discard
everything possible from the front end, as it cuts the I/O load and
effectively doubled the speed of the compiler. Maybe doesn't matter as much
these days, but back then it was a difference between running the ACVC in 5
days or 10 days.

On the rest, we're going to agree to disagree. Don't expect any support from
me for doing anything with ASIS in the ARG.

Randy.

"J-P. Rosen" <rosen@adalog.fr> wrote in message
news:sjes64$79r$1@dont-email.me...
> Le 02/10/2021 � 11:14, Randy Brukardt a �crit :
>>> My memory is that all "interesting" part of the standard was deliberatly
>>> put as comments in the specification, precisely to circumvent the ISO
>>> copyright, and allow the use of ASIS without paying an outrageous price
>>> to
>>> ISO.
>> I don't see how using comments helps anything. The Oracle case makes it
>> pretty clear an API iteself can be covered by a copyright, and surely the
>> comments are covered by the copyright. And the ISO version has no
>> copyright
>> statement other than the usual "All rights reserved".
>
> 1) It seems to me that you are confusing the copyright owner with the
> right to use the interface. Undoubtedly, ISO is the copyright owner. But
> they may authorize unlimited use of the specification, otherwise NO
> standard would make sense. Do you infringe copyright if you build an
> electrical plug that conforms to you electrical standard?

That's clearly covered by "fair use". But API Standards are different: you
have to copy large parts of the Standard to implement them (and ASIS is an
extreme case -- you have to copy 90% of it to use it). That certainly is not
covered by "fair use".

It's my (semi-informed) opinion that API Standards are useless, because you
have to violate the ISO copyright to use them (or buy a license).

> 2) Comments help, because they describe precisely what is expected by
> every function, and what it provides. Actually, I never open the ASIS
> standard, everything I need is detailed in the comments.

Exactly. Someone copied 90% of the ASIS standard without permission, and
*that* is what you are using. And that is depriving ISO of possible revenue.

It's clear to me that anyone using ASIS specs is skating on thin ice.
Whether it ever would become a problem for ISO is certainly unknown, but I
wouldn't want to build a business on top of such a thing. It's definitely
not open source by any reasonable definition.

We've spent a huge amount of effort to ensure that the Ada language (and
it's language-defined packages) do not fall into the same trap. But it's way
too late to do that for ASIS.

Randy.