cvs commit: ports/lang/mozart Makefile

[Volker Stolz]

Am 31. Jan 2005 um 11:03 CET schrieb Alexey Dokuchaev:
> >     lang/mozart          Makefile 
> >   Log:
> >   - Fix build on FreeBSD 5
> >     gcc 3.x makes new assumptions on private c++ data
> >     gcc 2.95 compiles mozart with no problems
> 
> For this very case, I don't want to test it and fix myself since this port
> want emacs and stuff, and I do not want to pollute my working system with
> this, but since submitter states that the problem with newer GCC is because
> it "makes new assumptions on private C++ data", I'd rather see a patch
> committed, or some reasoning why it cannot be.

Especially in the programming languages-department (and I'll speak for
haskell@ here, too), things are often highly compiler dependent. Mostly
applications break with older gccs because of c99, but here we're
probably (I didn't look too closely) talking about fixes which cannot
be come up with by simply looking at the source. In such cases it would be
appropriate for the maintainer to file a bug-report upstream, but we
shouldn't expect them to be able to handle all possible problems. And
since gcc295 is still readily available as a package (insert usual rant
about modem-users here) without any limitations I'm aware of, I don't
think we should push this.

It looks like we're talking about the two bugs mentioned in
http://www.mozart-oz.org/cgi-bin/oz-bugs/incoming?expression=gcc-3;user=guest
which have not been fixed by the developers yet.

Conclusion: Feel free to yell "Send patches", but please bear in mind that
a working port is better than a suspended PR. Life's too short.

Cheers,
  Volker
-- 
http://www-i2.informatik.rwth-aachen.de/stolz/ *** PGP *** S/MIME