This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH]: fix the --with-mpfr-dir=PATH configure flag

On Sun, 19 Nov 2006, Mark Mitchell wrote:

> Gabriel Dos Reis wrote:
> > If it is a regression and goes into 4.2, then it makes sense to
> > propagate ot up to 4.0.x.
> This is OK for 4.2, but let's not backport it to 4.1.  People using 4.1
> have presumably found a way to live with it the way that it is.
> Thanks,
> Mark Mitchell


I would respectfully disagree.  Older versions of these libs used to work
out of the box with GCC, so the fact that the latest release of mpfr fails
in this situation is a true regression.

What if "live with it" means they use an older gmp/mpfr that is buggy?
Those bugs inadvertedly may be blamed on GCC in the public's mind because
that's the tool they invoked.  Or worse, whatif "live with it" means they
opt not to configure and install GNU Fortran, and instead they bought a
proprietary compiler?

As a consequence of not including gmp/mpfr in the repo, we need to ensure
that the various versions of these packages, and certainly the latest
ones, work with GCC.  As these tools evolve over time (like adding
libtool, which changes where the archive gets created) we should try our
best on active branches to ensure they continue to work IMHO.

Would you please reconsider for 4.1?

Kaveh R. Ghazi

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]