Re: GCC 3.0 Release Criteria

["Trevor Jenkins" <Trevor dot Jenkins at suneidesis dot com>]

On Thursday, 27 April, 2000 15:59:54, Joe Buck <jbuck@possibly.synopsys.com>
wrote:

> [ Chill front end ]
>
> Trevor Jenkins
>> I'll play devil's advocate here; or should that be hell's-frozen-over
>> advocate? I believe there to be a very good reason for keeping the Chill
>> front-end maintained as part of gcc 3 and later. It provides a
>> reference/tutorial case of integrating a new language into the gcc
>> architecture.
>
> If we're going to maintain it, we need a maintainer;

I just knew you were going to say that. And why I didn't volunteer :-)

> the original author
> has gone on to other work, and almost none of the developers knows Chill.

It's a shame that there isn't an entrance examination for gcc developers,
which could include getting to know Chill and maintaining it until such time
as another apprentice came along. Chill could be a journeyman piece.

I have no objection to freezing the Chill language subset itself. I realy
would not like to see Chill disappear from the collection.

You didn't pick me up the most glaring omission in my original posting.
Chill is more than a front-end. The library needs to be maintained.

There is an equally glaring omission in the release criteria themselves
concerning Chill:

> The following languages will be supported by the release, but their
> behavior will not be a primary consideration in determining whether or not
> to ship a particular release candidate:
>
> * Chill
> * Objective-C
>
> In particular, no application testing, code quality, or compile-time
> performance testing will be required for these languages. However, the
> regression testing criteria documented below will apply to these languages.
>
> Regression Tests
>
> The GCC testsuite contains extensive C and C++ regression tests, as well as
> some Fortran, and Objective-C tests. GCC 3.0 will not fail any of these
> tests which the previous release GCC passed on any of the supported
> platforms. In particular, the current regression testsuite will be run
> using GCC 2.95.2 on each of the supported platforms; those results can then
> be compared with the output from a release candidate.

If "the regression testing criteria ... will apply to these languages"
(Chill and ObjC) is there a Chill regression suite? Such a suite would be
useful to anyone attempting to maintain Chill. If there is no Chill suite
then these clauses should be changed to reflect how Chill will be deemed
suitable for inclusion in gcc 3.0.

Regards, Trevor

British Sign Language is not inarticulate handwaving; it's a living
language. So recognise it now.

--

<>< Re: deemed!