(libf2c) libu77 struct timezone configuration fix

[Dave Love]

>>>>> "MK" == Mumit Khan <khan@xraylith.wisc.edu> writes:

 MK> In other words, the existing code is non-conformant. 

I don't understand.  (I read what doc I could find at the time, but I
don't have access to the original BSD stuff.)

Are you saying that on _all_ systems the null pointer is OK?  (I can't
tell but it may well be true.)  If so, why not just use that?

Otherwise are you saying that on some systems that define struct
timezone, passing a non-null pointer loses?  If so, presumably there
should be an execute configure test for it.

 MK> Let me know if you want more clarification,

As before, what is this in aid of?

	* libU77/aclocal.m4: New file.
	* libU77/configure.in (LIBU77_HAVE_STRUCT_TIMEZONE): Add test.
	* libU77/acconfig.h (HAVE_STRUCT_TIMEZONE): Add macro.

How is the standard autoconf test deficient, and why can't the test
you need go inline in configure.in (as elsewhere), avoiding the extra
complication of aclocal/acconfig?

 MK> and/or any particular way you want me to handle this.

It's not up to me, but it's entirely unclear to me what's going on and
I suspect others trying to maintain it would have the same problem.

BTW, presumably dmg would appreciate such changes for the canonical
libf2c.