math/slatec - worth fixing deleted language features?

[Steve Kargl]

On Sat, Jul 20, 2013 at 06:37:45PM +0100, Anton Shterenlikht wrote:
> Anybody uses math/slatec?

I don't use it currently.  When I was actively contributing to
gfortran (to ensure that gfortran worked on FreeBSD), I fed
routinely fed slatec into gfortran looking for bugs.

> This is a collection of >1400 f77 routines.
> The problem is that there is no upstream anymore,
> and the code has deleted features, e.g.:

(snip)

> Fixing each individual deleted feature is not that hard.

There is nothing to fix.  You will be hard pressed to find
a Fortran, which works on FreeBSD, that does not support
the deleted features used by slatec.

> In addition, there is a risk of introducing new bugs.
> The library, as it is, is supposed to be thoroughly tested.

This is why compilers still support the deleted features.  Code
written 50 years will still compile and work.

> 
> So I'll asking for opinions on what to do
> with math/slatec; (1) nothing - leave as it is,

yes

> (2) try updating deleted (and obsolete) language constructs,

No.  Do nothing.  If the warnings bother you, add FFLAGS+=-w
or FFLAGS+=-std=legacy to the Makefile.

> (3) spawn another port, with
> no deleted or obsolete features, but with little testing?

Do nothing.

-- 
Steve