GNAT 3.15 - GCC 3.x (Was Re: Format of .adp file).

GNAT 3.15 - GCC 3.x (Was Re: Format of .adp file).

[sk]

Hi,

> Preben Randhol <randhol+news@pvv.org> wrote:
> : 
> :> GNAT 3.15 has a totally new project file syntax; it looks like Ada (no
> :> surprise there :). You might want to wait until that becomes public.
> 
> Isn't the parsing code already in the current GCC tree?
> 
> --  georg

I suspect that 3.15 is not iminent and I am confused about
the GNAT/GCC relationship ...

Can someone enlighten me as to how future GNAT releases
and GCC release will relate to one another 

Better still, is there a web page or other resource which 
discusses the integration of GNAT into GCC etc ?

-- 
-------------------------------------
-- Merge vertically for real address
-------------------------------------
s n p @ t . o
 k i e k c c m
-------------------------------------

Re: GNAT 3.15 - GCC 3.x (Was Re: Format of .adp file).

[Simon Wright]

sk <noname@myob.com> writes:

> Better still, is there a web page or other resource which discusses
> the integration of GNAT into GCC etc ?

Well, http://gcc.gnu.org/ is the starting point ..

Re: GNAT 3.15 - GCC 3.x (Was Re: Format of .adp file).

[sk]

Hi,

Simon Wright <simon@pushface.org>
> Well, http://gcc.gnu.org/ is the starting point ..

My last excursion to http://gcc.gnu.org/ yielded
nothing particular to GNAT (or GNU Ada) ... and 
my visit 2 minutes ago, similar results. 

Do I need to sift through the "gcc" mailing list 
archives ? 

-- 
-------------------------------------
-- Merge vertically for real address
-------------------------------------
s n p @ t . o
 k i e k c c m
-------------------------------------

Re: GNAT 3.15 - GCC 3.x (Was Re: Format of .adp file).

[chris.danx]

sk wrote:

> Hi,
>
>
> >Preben Randhol  wrote:
> >:
> >:> GNAT 3.15 has a totally new project file syntax; it looks like Ada (no
> >:> surprise there :). You might want to wait until that becomes public.
> >
> >Isn't the parsing code already in the current GCC tree?
> >
> >--  georg
>
>
> I suspect that 3.15 is not iminent and I am confused about
> the GNAT/GCC relationship ...


Isn't a 3.15 (of some sort) already used by companies?  *I* think so. 
If you mean a public release, I doubt we will see one but you never 
know.  Act seem to be concentrating on the gcc 3.xx integration and 
probably would rather be doing that and keeping their customers happy 
than prepping' a public release of 3.15 which would soon become outdated 
by the 3.xx effort (if they prepared a 3.15p then they'd likely have to 
divert work from 3.xx and it would hold things up on that front...). 
Still it's a weird world and you never can be sure of anything.

> Can someone enlighten me as to how future GNAT releases
> and GCC release will relate to one another

Aren't they going to be the same?  I thought that once integration was 
complete, gcc and gnat would live together, and act wouldn't be inclined 
to make special public releases, just work on the gcc tree (and their 
customers needs).  Would it make any sense to do otherwise?


> Better still, is there a web page or other resource which
> discusses the integration of GNAT into GCC etc ?


Unfortunately not really!  I've been trying to find out how things are 
going, but there is no progress summary *publicly available* and ACT 
seem to work away at it and then drop things into gcc 3.xx in 'big 
merges', so we only see any fixes to incompatibilities with the existing 
incorporated gnat sources and gcc quickly and big differences less 
frequently.



Chris
-- 
for personal replies change spamoff to chris

Re: GNAT 3.15 - GCC 3.x (Was Re: Format of .adp file).

[sk]

Hi,


"chris.danx" <spamoff.danx@ntlworld.com>
> Isn't a 3.15 (of some sort) already used by companies?  

Sorry, I meant a "p" release ...

> ... Act seem to be concentrating on the gcc 3.xx integration and 
> probably would rather be doing that and keeping their customers happy 
> than prepping' a public release of 3.15 which would soon become outdated 
> by the 3.xx effort ...
> ...
> Aren't they going to be the same? ...
> ...
> Unfortunately not really! ...

I was under similar impressions until someone mentioned a 3.15 release
and I remembered that Ada, as in GCC 3.x, and GNAT are currently
divergent. 

Is there going to be a continued distinction between the ACT GNAT 
releases (private or public) and the Ada gcc front-end ? 

I would assume otherwise, but as noted ...

> ... it's a weird world and you never can be sure of anything.

:-)

...

Has anyone built a gcc list digest to exclusively track the 
Ada/GNAT issues/discussions and be willing to share ?

-- 
-------------------------------------
-- Merge vertically for real address
-------------------------------------
s n p @ t . o
 k i e k c c m
-------------------------------------

Re: GNAT 3.15 - GCC 3.x (Was Re: Format of .adp file).

[Stephen Leake]

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

> Isn't a 3.15 (of some sort) already used by companies?  

Yes, that's the currently supported release.

> *I* think so. If you mean a public release, I doubt we will see one
> but you never know. 

I believe Robert Dewar said here that the gcc releases would be the
public releases of GNAT, but I'm not sure, and that may change.

> Act seem to be concentrating on the gcc 3.xx integration and
> probably would rather be doing that and keeping their customers
> happy than prepping' a public release of 3.15 which would soon
> become outdated by the 3.xx effort (if they prepared a 3.15p then
> they'd likely have to divert work from 3.xx and it would hold things
> up on that front...). 

As a paying customer of ACT, that's what I want them to do :).

> > Can someone enlighten me as to how future GNAT releases
> > and GCC release will relate to one another
> 
> Aren't they going to be the same?  

Not precisely. I'm not speaking for ACT, but what I remember of
previous discussions, and can extrapolate from current behavior, is
that ACT will continue to make "internal, customer" releases of GNAT.
They will also continue to dump changes into the gcc tree. There will
not necessarily be synchronized releases; there is no need. But in
general, the gcc tree will track the GNAT releases. I would not be
surprised to see lead/lag times of a year, though (in either
direction). 

> > Better still, is there a web page or other resource which
> > discusses the integration of GNAT into GCC etc ?

apparently the gcc developer mailing list is where this is mostly
discussed. I have not tried to wade thru it.

-- 
-- Stephe

Re: GNAT 3.15 - GCC 3.x (Was Re: Format of .adp file).

[chris.danx]

Stephen Leake wrote:

 > >*I* think so. If you mean a public release, I doubt we will see one
 > >but you never know.
 >
 >
 > I believe Robert Dewar said here that the gcc releases would be the
 > public releases of GNAT, but I'm not sure, and that may change.

That's what I thought.

 > >Act seem to be concentrating on the gcc 3.xx integration and
 > >probably would rather be doing that and keeping their customers
 > >happy than prepping' a public release of 3.15 which would soon
 > >become outdated by the 3.xx effort (if they prepared a 3.15p then
 > >they'd likely have to divert work from 3.xx and it would hold things
 > >up on that front...).
 >
 >
 > As a paying customer of ACT, that's what I want them to do :).

As a Gnat user that's what I want too :)


 > >Aren't they going to be the same?
 >
 >
 > Not precisely. I'm not speaking for ACT, but what I remember of
 > previous discussions, and can extrapolate from current behavior, is
 > that ACT will continue to make "internal, customer" releases of GNAT.
 > They will also continue to dump changes into the gcc tree. There will
 > not necessarily be synchronized releases; there is no need. But in
 > general, the gcc tree will track the GNAT releases. I would not be
 > surprised to see lead/lag times of a year, though (in either
 > direction).

I know of people who have a problem with this and I agree with them.
I'd prefer it if there where external maintainers as well as ACT who
could improve Gnat for all.  All the other gcc frontends are improved by
the communities they serve, why not Gnat?


 > >>Better still, is there a web page or other resource which
 > >>discusses the integration of GNAT into GCC etc ?
 >
 >
 > apparently the gcc developer mailing list is where this is mostly
 > discussed. I have not tried to wade thru it.

I gave up with that list.  There's just too much traffic and not enough
Ada/Gnat topics to justify deleting 100-200 messages a day on the off
chance that something on Gnat would appear.  Best to search the
archives, but even then it's tough.


Chris
-- 
for personal replies change spamoff to chris

Re: GNAT 3.15 - GCC 3.x (Was Re: Format of .adp file).

[Stephen Leake]

"chris.danx" <chris.danx@ntlworld.com> writes:

> Stephen Leake wrote:
> 
>  > Not precisely. I'm not speaking for ACT, but what I remember of
>  > previous discussions, and can extrapolate from current behavior, is
>  > that ACT will continue to make "internal, customer" releases of GNAT.
>  > They will also continue to dump changes into the gcc tree. There will
>  > not necessarily be synchronized releases; there is no need. But in
>  > general, the gcc tree will track the GNAT releases. I would not be
>  > surprised to see lead/lag times of a year, though (in either
>  > direction).
> 
> I know of people who have a problem with this and I agree with them.
> I'd prefer it if there where external maintainers as well as ACT who
> could improve Gnat for all.  All the other gcc frontends are improved by
> the communities they serve, why not Gnat?

Nothing is stoping other people from working on the gcc tree. If you
have something significant to contribute, it would help to coordinate
with ACT. I guess that could be complicated, especially if they decide
not to support what you want to add. But I don't see that as being any
more or less likely than RedHat deciding not to support some addition
to gcc C++.

-- 
-- Stephe

Re: GNAT 3.15 - GCC 3.x (Was Re: Format of .adp file).

[Adrian Knoth]

sk <noname@myob.com> wrote:

> Can someone enlighten me as to how future GNAT releases
> and GCC release will relate to one another 

A propos: I would like to get rid of the bootstrapping-problem of GNAT.
Compiling an Ada-compiler with an Ada-compiler is not really extending
the use on different platforms.

What about the common way of using a subset-language-stage0-compiler
written in shellscript or C?

-- 
mail: adi@thur.de  	http://adi.thur.de	PGP: v2-key via keyserver

Microsoft Motto: 'Wait for us, we're the leaders!'

Re: GNAT 3.15 - GCC 3.x

[Florian Weimer]

Adrian Knoth <adi@drcomp.erfurt.thur.de> writes:

> What about the common way of using a subset-language-stage0-compiler
> written in shellscript or C?

This is still a very significant amount of work.  Such work would be
spent better designing an efficient (register-based?) VM and GCC
backend for it.

Re: GNAT 3.15 - GCC 3.x (Was Re: Format of .adp file).

[Pascal Obry]

Adrian Knoth <adi@drcomp.erfurt.thur.de> writes:

> sk <noname@myob.com> wrote:
> 
> > Can someone enlighten me as to how future GNAT releases
> > and GCC release will relate to one another 
> 
> A propos: I would like to get rid of the bootstrapping-problem of GNAT.
> Compiling an Ada-compiler with an Ada-compiler is not really extending
> the use on different platforms.

As already said the bootstrapping problem is not really an issue. You can
cross-compiler GNAT to get the minimum set of executables needed. The real
problem is to port the runtime.

Pascal.

-- 

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

Porting the GNAT runtime (was: Re: GNAT 3.15 - GCC 3.x)

[SteveD]

"Pascal Obry" <p.obry@wanadoo.fr> wrote in message
news:u4ragf384.fsf@wanadoo.fr...
[snip]
>
> As already said the bootstrapping problem is not really an issue. You can
> cross-compiler GNAT to get the minimum set of executables needed. The real
> problem is to port the runtime.

Is any document available describing just what is required to port the
runtime?  Or do you just have to analyze the sources and figure it out.

SteveD

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

Re: Porting the GNAT runtime (was: Re: GNAT 3.15 - GCC 3.x)

[sk]

Hi,

>Is any document available describing just what is required to port the
>runtime?  Or do you just have to analyze the sources and figure it out.
>
>SteveD

I haven't had chance to read this yet but "Javier Miranda" has
documented
the GNAT source which should be very useful in a porting effort if no
one
else has an alternative answer for you.

http://www.iuma.ulpgc.es/users/jmiranda/gnat-rts/



-- 
-------------------------------------
-- Merge vertically for real address
-------------------------------------
s n p @ t . o
 k i e k c c m
-------------------------------------