FYI: GNAT and GCC

FYI: GNAT and GCC

[Florian Weimer]

Richard Kenner has just started to populate the gcc/ada directory in
the CVS repository.

Re: FYI: GNAT and GCC

[Lutz Donnerhacke]

* Florian Weimer wrote:
>Richard Kenner has just started to populate the gcc/ada directory in
>the CVS repository.

Excellent!

Re: FYI: GNAT and GCC

[Robert Dewar]

Florian Weimer <Florian.Weimer@RUS.Uni-Stuttgart.DE> wrote in message news:<tgu1xikvmn.fsf@mercury.rus.uni-stuttgart.de>...
> Richard Kenner has just started to populate the gcc/ada directory in
> the CVS repository.


The GNAT sources are now at gcc. We are currently checking
to see that everything is there and is exactly as it should
be. For those who want to access these sources, I would
suggest waiting till tomorrow, to make sure everything is
settled down, we know for sure of one or two glitches that
need fixing.

Robert Dewar
(with his hat on as FSF maintainer for GNAT)

Re: FYI: GNAT and GCC

[Robert Dewar]

dewar@gnat.com (Robert Dewar) wrote in message news:<5ee5b646.0110021022.2843da52@posting.google.com>...
> The GNAT sources are now at gcc. We are currently checking
> to see that everything is there and is exactly as it should
> be. For those who want to access these sources, I would
> suggest waiting till tomorrow, to make sure everything is
> settled down, we know for sure of one or two glitches that
> need fixing.

I should note that this version is very much an experimental one,
suitable only for experimental use. It
is known to have deficiencies with respect to our main
test suite, so it is by no means as reliable as the
2.8.x official GNAT release yet.

> Robert Dewar
> (with his hat on as FSF maintainer for GNAT)

Re: FYI: GNAT and GCC

[mike]

In article <tgu1xikvmn.fsf@mercury.rus.uni-stuttgart.de>, Florian says...
>
>Richard Kenner has just started to populate the gcc/ada directory in
>the CVS repository.

This a historic moment in the life of Ada. 

3-13p or 3-14a1

[Frank]

Hi!

Is that GNAT 3-13p or GNAT 3-14a1?

Frank

Re: 3-13p or 3-14a1

[Larry Kilgallen]

In article <Vcpu7.6944$tu6.174479@news1.oke.nextra.no>, "Frank" <franjoe@frisurf.no> writes:

> Is that GNAT 3-13p or GNAT 3-14a1?

Is what ?

You started this topic.

Re: 3-13p or 3-14a1

[Ted Dennison]

In article <xIpqZQkdlYIu@eisner.encompasserve.org>, Larry Kilgallen says...
>
>In article <Vcpu7.6944$tu6.174479@news1.oke.nextra.no>, "Frank" <franjoe@frisurf.no> writes:
>
>> Is that GNAT 3-13p or GNAT 3-14a1?
>
>Is what ?
>
>You started this topic.

On my threaded reader, this is a reply to the original message about Gnat moving
into GCC CVS. Thus one could assume that he's talking about the version of Gnat
that is being put into there.

I don't know where to look in the sources for that info, but for anyone who
does, its in gcc's CVS. Unfortuntately, it looks like the gcc CVS server is
getting "slashdotted" by anxious Ada users, so it may be best to wait a while to
look it up. :-)

---
T.E.D.    homepage   - http://www.telepath.com/dennison/Ted/TED.html
          home email - mailto:dennison@telepath.com
No trees were killed in the sending of this message. 
However a large number of electrons were terribly inconvenienced.

Re: 3-13p or 3-14a1

[Robert Dewar]

Ted Dennison<dennison@telepath.com> wrote in message news:<%hqu7.13395$ev2.22118@www.newsranger.com>...
> In article <xIpqZQkdlYIu@eisner.encompasserve.org>, Larry Kilgallen says...
> >
> >In article <Vcpu7.6944$tu6.174479@news1.oke.nextra.no>, "Frank" <franjoe@frisurf.no> writes:
> >
> >> Is that GNAT 3-13p or GNAT 3-14a1?
> >
> >Is what ?
> >
> >You started this topic.
> 
> On my threaded reader, this is a reply to the original message about Gnat moving
> into GCC CVS. Thus one could assume that he's talking about the version of Gnat
> that is being put into there.
> 
> I don't know where to look in the sources for that info, but for anyone who
> does, its in gcc's CVS. Unfortuntately, it looks like the gcc CVS server is
> getting "slashdotted" by anxious Ada users, so it may be best to wait a while to
> look it up. :-)


This is neither 3.13p nor 3.14a1, it is the current
experimental development sources, which are known to
have problems with GCC 3.x, so this is not a version
that is appropriate for general use. It represents
work in progress on moving towards a stable GCC 3
compatible release of GNAT.

Re: 3-13p or 3-14a1

[Frank]

Ok.

Frank

Re: FYI: GNAT and GCC

[Ian S. Nelson]

Florian Weimer wrote:

> Richard Kenner has just started to populate the gcc/ada directory in
> the CVS repository.
> 


So GNAT or rather Ada will be in GCC 3.1? Ada will be part of *the* 
compiler?!?

This is some of the best news I've heard in quite a while.

Ian Nelson

Re: FYI: GNAT and GCC

[Robert Dewar]

"Ian S. Nelson" <nelsonis@earthlink.net> wrote in message news:<3BBA9724.2020009@earthlink.net>...
> Florian Weimer wrote:
> So GNAT or rather Ada will be in GCC 3.1? Ada will be part of *the* 
> compiler?!?
> 
> This is some of the best news I've heard in quite a while.


People unfamiliar with gcc may be reading far too much into
this announcement. For example, this definitely does not 
mean that all installations of gcc will include Ada, no
more than they include g77 (or chill for that matter)
right now. It is just one of the front ends that is
available, it is not automatically included.

Robert Dewar

Re: FYI: GNAT and GCC

[Preben Randhol]

On 3 Oct 2001 09:12:20 -0700, Robert Dewar wrote:

> People unfamiliar with gcc may be reading far too much into
> this announcement. For example, this definitely does not 
> mean that all installations of gcc will include Ada, no
> more than they include g77 (or chill for that matter)
> right now. It is just one of the front ends that is
> available, it is not automatically included.

No, but it will probably be easier to install as it is just to add the
gcc-gnat DEB, RPM, SLP etc. package to the installation. Also I think
that you will get a few more people who wonders what is gcc-gnat (or
gcc-ada I don't know what it will be called) as they see it listed
together with the gcc packages and check it out. At least I read a bit
about object C after seeing that there was a gcc-objc package. And quite
possibly you will have more linux distributions adding this package to
the standard distribution.

Keep up the excellent work!

Preben Randhol

Re: FYI: GNAT and GCC

[Robert Dewar]

Preben Randhol <randhol+abuse@pvv.org> wrote in message news:<slrn9rn5tj.uu.randhol+abuse@kiuk0156.chembio.ntnu.no>...
> No, but it will probably be easier to install as it is just to add the
> gcc-gnat DEB, RPM, SLP etc. package to the installation. 

Yes, although of course unlike the other gcc front ends,
to install from sources requires that you go and fetch
at least a gnat1 starter executable from somewhere. That's
not significantly less work than installing a complete
GNAT in the first place.

I just don't want people to get the wrong impression here.
I think a fair analogy is the extent to which chill got
picked up as part of typical gcc installs.

Robert Dewar

Re: FYI: GNAT and GCC

[Florian Weimer]

dewar@gnat.com (Robert Dewar) writes:

> I just don't want people to get the wrong impression here.
> I think a fair analogy is the extent to which chill got
> picked up as part of typical gcc installs.

With the same prospect for the future?  I hope not. ;-)

Re: FYI: GNAT and GCC

[Ian S. Nelson]

Robert Dewar wrote:
> "Ian S. Nelson" <nelsonis@earthlink.net> wrote in message news:<3BBA9724.2020009@earthlink.net>...
> 
>>Florian Weimer wrote:
>>So GNAT or rather Ada will be in GCC 3.1? Ada will be part of *the* 
>>compiler?!?
>>
>>This is some of the best news I've heard in quite a while.
>>
> 
> 
> People unfamiliar with gcc may be reading far too much into
> this announcement. For example, this definitely does not 
> mean that all installations of gcc will include Ada, no
> more than they include g77 (or chill for that matter)
> right now. It is just one of the front ends that is
> available, it is not automatically included.
> 
> Robert Dewar


It doesn't mean it will be there but it means it will be there when I 
build GCC..  And most likely Mandrake, Redhat, *BSD, etc.. will start 
adding support.

Either way, I think it's good news.  No more patching GCC and hoping I 
can still compile ada and c++.

Ian

Re: FYI: GNAT and GCC

[Dale Pontius]

In article <3BBBCF3B.9050403@earthlink.net>,
        "Ian S. Nelson" <nelsonis@earthlink.net> writes:
> Robert Dewar wrote:
>> "Ian S. Nelson" <nelsonis@earthlink.net> wrote in message news:<3BBA9724.2020009@earthlink.net>...
>>
>>>Florian Weimer wrote:
>>>So GNAT or rather Ada will be in GCC 3.1? Ada will be part of *the*
>>>compiler?!?
>>>
>>>This is some of the best news I've heard in quite a while.
>>>
>>
>>
>> People unfamiliar with gcc may be reading far too much into
>> this announcement. For example, this definitely does not
>> mean that all installations of gcc will include Ada, no
>> more than they include g77 (or chill for that matter)
>> right now. It is just one of the front ends that is
>> available, it is not automatically included.
>>
>> Robert Dewar
>
>
> It doesn't mean it will be there but it means it will be there when I
> build GCC..  And most likely Mandrake, Redhat, *BSD, etc.. will start
> adding support.
>
Just for a minor chunk of other good news, at least rpms for gnat,
gtk-ada, and a few other things are on the RedHat 7.1 PowerTools
CD. I don't have the system with it installed handy, so I can't
comment on the versions. Plus I know it's on PowerTools and not on
either of the base install CDs, but at least it's in the package.

Dale Pontius
(NOT speaking for IBM)

Re: FYI: GNAT and GCC

[Lutz Donnerhacke]

* Ian S. Nelson wrote:
>Either way, I think it's good news.  No more patching GCC and hoping I
>can still compile ada and c++.

With a very small and obvious Makefile patch, I got a working:
GNU Ada version 3.1 20011008 (experimental) (i686-pc-linux-gnu)
        compiled by GNU C version 3.1 20011008 (experimental).
and start looking on assembler outputs.

I'm very pleased compared with GNAT 3.13p: Much better inlining, more
efficient stack handling, some new error messages� and definitly shorter
code (especially on bound checking). I'm still checking if the code is
correct. But even -fssa seems to work fine. The vast improvement (to me) are
the enumeration converters 'Val, 'Image. The necessary tables look very much
like my old C64 BASIC ROM.
\f
1  Following Errors are new to me:

 - pragma Inline in package specification do not longer work for code in
   package bodies, but can not reproduce this in a very short example.

 - Instantiation of a generic with a derivated type now cleanly rejects
   nonderivated integral types.

Re: FYI: GNAT and GCC

[Robert Dewar]

lutz@iks-jena.de (Lutz Donnerhacke) wrote in message news:<slrn9s3gv0.i5.lutz@taranis.iks-jena.de>...
>  
> 1  Following Errors are new to me:
> 
>  - pragma Inline in package specification do not longer work for code in
>    package bodies, but can not reproduce this in a very short example.
> 
>  - Instantiation of a generic with a derivated type now cleanly rejects
>    nonderivated integral types.

Good to hear that you got a clean build. Remember to report
any problems to report@gnat.com (you can also use the standard gcc
process), but errors reported to CLA go into
the ether :-)

Re: FYI: GNAT and GCC

[Lutz Donnerhacke]

* Robert Dewar wrote:
>Good to hear that you got a clean build.

Today I managed a full bootstrap and build even the libs without any problem.
Linking a small example to the real executable works. Example does the
requested job. I'm missing gnatmake ;-/.

>Remember to report any problems to report@gnat.com (you can also use the
>standard gcc process), but errors reported to CLA go into the ether :-)

I do not report errors without a detailed reproducable example.
I'd like to inform cla about the success and good quality. (Small examples
tend to grow in Size and Time when switching on optimization, because all
working data is required to go though CPU registers even if this is not
necessary.)