GNAT and GCC 3.02

GNAT and GCC 3.02

[Mr. Caffiene]

I've been considering upgrading to GCC 3.02, however I'm not sure if GNAT will work or even compile with this GCC release.

Anyone point me to some documents which will give me the lowdown on this?

Mr. Caffiene chris@dont.spam.me

Re: GNAT and GCC 3.02

[Ted Dennison]

In article <20020102140849.2bb221a5.nospam@attbi.com>, Mr. Caffiene says...
>
>I've been considering upgrading to GCC 3.02, however I'm not sure if GNAT will work or even compile with this GCC release.
>
>Anyone point me to some documents which will give me the lowdown on this?

It depends on what you mean by "work with". The latest public release of Gnat
use gcc 2.8.1. The only way to even attempt to use 3.02 instead would be to try
to build them together from sources, which wouldn't likely work at all. However,
there's nothing preventing you from having *both* compilers on your system,
which is what most folks have been doing for a while now.

Gnat has been checked into the GCC baseline, but no stable version of gcc has
been released since then. You may or may not be able to grab a gcc source
snapshot from their CVS server and build yourself a post 3.0.3 gcc with Gnat
included. If you want to go that route, I'd monitor the mailing list
(http://gcc.gnu.org/ml/gcc/) to help out with any problems you might encounter.


---
T.E.D.    homepage   - http://www.telepath.com/dennison/Ted/TED.html

No trees were killed in the sending of this message. 
However a large number of electrons were terribly inconvenienced.

Re: GNAT and GCC 3.02

[Florian Weimer]

Ted Dennison<dennison@telepath.com> writes:

> Gnat has been checked into the GCC baseline, but no stable version of gcc has
> been released since then.

IIRC GCC 3.0.4 has been released in the meantime, but this is the
stable branch of GCC, and GNAT has been comitted to the development
branch (which will end up in GCC 3.1 eventually).

Re: GNAT and GCC 3.02

[Bobby D. Bryant]

On Wed, 02 Jan 2002 15:46:56 -0600, Florian Weimer wrote:

> Ted Dennison<dennison@telepath.com> writes:
> 
>> Gnat has been checked into the GCC baseline, but no stable version of
>> gcc has been released since then.
> 
> IIRC GCC 3.0.4 has been released in the meantime, but this is the
> stable branch of GCC, and GNAT has been comitted to the development
> branch (which will end up in GCC 3.1 eventually).

FWIW, I noticed that there are GCC 3.1 RPMs out at
ftp://rawhide.redhat.com/pub/redhat/linux/rawhide/i386/RedHat/RPMS/.

I don't know whether they include GNAT, because upgrading to those RPMs
will also require upgrading some libraries and probably a lot else, and
I've got better things to do right now than download everything just to
find out.

Bobby Bryant
Austin, Texas

Re: GNAT and GCC 3.02

[Preben Randhol]

On Thu, 03 Jan 2002 03:49:41 -0600, Bobby D. Bryant wrote:
> On Wed, 02 Jan 2002 15:46:56 -0600, Florian Weimer wrote:
> 
> FWIW, I noticed that there are GCC 3.1 RPMs out at
> ftp://rawhide.redhat.com/pub/redhat/linux/rawhide/i386/RedHat/RPMS/.
> 
> I don't know whether they include GNAT, because upgrading to those RPMs
> will also require upgrading some libraries and probably a lot else, and
> I've got better things to do right now than download everything just to
> find out.

As far as I can see there are only gcc-c++, gcc-g77 (FORTRAN 77),
gvv-java and gcc-objc pacakges. If gnat was included I would expect it
was called gcc-gnat or gcc-ada or something like that.

Preben
-- 
 ()   Join the worldwide campaign to protect fundamental human rights.
'||}
{||'                                           http://www.amnesty.org/

Re: GNAT and GCC 3.02

[Ted Dennison]

In article <slrna38gp3.6mo.randhol+abuse@kiuk0156.chembio.ntnu.no>, Preben
Randhol says...
>As far as I can see there are only gcc-c++, gcc-g77 (FORTRAN 77),
>gvv-java and gcc-objc pacakges. If gnat was included I would expect it
>was called gcc-gnat or gcc-ada or something like that.

Perhaps I'm a bit biased (nooooo), but if it includes Fortran, Java, and
ObjectiveC then it ought to include Ada too. Ada's just as much a part of the
3.1 baseline as those languages.

---
T.E.D.    homepage   - http://www.telepath.com/dennison/Ted/TED.html

No trees were killed in the sending of this message. 
However a large number of electrons were terribly inconvenienced.

Re: GNAT and GCC 3.02

[Preben Randhol]

On Thu, 03 Jan 2002 13:06:34 GMT, Ted Dennison wrote:
> Perhaps I'm a bit biased (nooooo), but if it includes Fortran, Java, and
> ObjectiveC then it ought to include Ada too. Ada's just as much a part of the
> 3.1 baseline as those languages.

Except for gcc-java (as I don't use RedHat anymore I don't know when
they first made gcc-java) the other packages has been available for
years. I mean that they have been updated for years. So they are already
integrated into the distros and only need to be compiled for the new
version. 

But I would rather have a tested gcc-gnat out in a while than something
ASAP that might be bugging and thus give gcc-gnat a bad rep.

Preben
-- 
 ()   Join the worldwide campaign to protect fundamental human rights.
'||}
{||'                                           http://www.amnesty.org/

Re: GNAT and GCC 3.02

[Ted Dennison]

In article <slrna38rib.btr.randhol+abuse@kiuk0156.chembio.ntnu.no>, Preben
Randhol says...
>But I would rather have a tested gcc-gnat out in a while than something
>ASAP that might be bugging and thus give gcc-gnat a bad rep.

True. But if its a 3.1 distro, the other stuff is just as liable to be buggy.

---
T.E.D.    homepage   - http://www.telepath.com/dennison/Ted/TED.html

No trees were killed in the sending of this message. 
However a large number of electrons were terribly inconvenienced.

Re: GNAT and GCC 3.02

[Preben Randhol]

On Thu, 03 Jan 2002 15:27:18 GMT, Ted Dennison wrote:
> In article <slrna38rib.btr.randhol+abuse@kiuk0156.chembio.ntnu.no>, Preben
> Randhol says...
>>But I would rather have a tested gcc-gnat out in a while than something
>>ASAP that might be bugging and thus give gcc-gnat a bad rep.
> 
> True. But if its a 3.1 distro, the other stuff is just as liable to be buggy.

Yes


-- 
 ()   Join the worldwide campaign to protect fundamental human rights.
'||}
{||'                                           http://www.amnesty.org/

Re: GNAT and GCC 3.02

[Florian Weimer]

"Bobby D. Bryant" <bdbryant@mail.utexas.edu> writes:

> FWIW, I noticed that there are GCC 3.1 RPMs out at
> ftp://rawhide.redhat.com/pub/redhat/linux/rawhide/i386/RedHat/RPMS/.

It is certainly a bad idea to use RPMs at this stage.  If these RPMs
were built on a bad day, you can get into arbitrary problems.

If you want to use GCC 3.1 now, better checkout the CVS and compile it
directly from source.  This way, you have got the source code at hand
if anything unexpected happens.

Re: GNAT and GCC 3.02

[Florian Weimer]

Ted Dennison<dennison@telepath.com> writes:

> Perhaps I'm a bit biased (nooooo), but if it includes Fortran, Java, and
> ObjectiveC then it ought to include Ada too. Ada's just as much a part of the
> 3.1 baseline as those languages.

Currently, building the Ada-specific part of the GNAT toolchain
requires reading the build instructions and obtaining a working GNAT
binary.  That's why binary distributions do not automatically include
the GCC Ada front end.

Re: off-topic (wandering away from GNAT and GCC 3.02)

[Wes Groleau]

Florian Weimer wrote:
> It is certainly a bad idea to use RPMs at this stage.  If these RPMs
> were built on a bad day, you can get into arbitrary problems.

Like the advice to never buy a car assembled on Monday morning
or Friday afternoon?  Someone I know took his car to the shop
several times and no one could explain the rattling sound.
Finally someone removed the door panel and found a paper sack
containing bolts inside the door.  Moral of the story: don't
by a car that had a lunch break during its assembly.

-- 
Wes Groleau
http://freepages.rootsweb.com/~wgroleau

Re: off-topic (wandering away from GNAT and GCC 3.02)

[Florian Weimer]

Wes Groleau <wesgroleau@spamcop.net> writes:

> Florian Weimer wrote:
>> It is certainly a bad idea to use RPMs at this stage.  If these RPMs
>> were built on a bad day, you can get into arbitrary problems.
>
> Like the advice to never buy a car assembled on Monday morning
> or Friday afternoon?

No, the problem with the GCC CVS is less esoteric, I think.  It's
software in (sometiems heavy) development.  Rather static RPM
snapshots are a bit like lottery.  (I don't know how much pre-release
testing is done on these RPMs, though.)