[Python-Dev] HUGE_VAL and INFINITY

[Tim Peters]

[Jack Jansen]
> I found a couple of references to INFINITY being a float. ...

Yes, INFINITY must expand to a constant expression of type float, although
your compiler isn't doing that (see below).  The header files you're using
are still braindead for the reasons I explained last time regardless.

>> #define Py_HUGE_VAL ((double)INFINITY)

> Is the intention of this define that it would first convert the
> constant "1e50" to an IEEE float "Infinity", and that this float
> would then be promoted to a double "Infinity"?

No.  As the comments before this code said, the explicit cast to double was
for the benefit of some other broken compiler.

The literal "1e50" has type double in C, so if they're really #define'ing
INFINITY as 1e50 then they're violating that INFINITY must expand to an
expression of type float.  They could have made it a float literal by
appending "f" or "F", but then it wouldn't be a legal float literal.
They're screwed either way -- they're doing this part incorrectly no matter
how you cut it.  They can look at any other compiler for a correct way to do
it <wink>.

> If it is indeed stated somewhere in the C standard that this is the
> course of action to take then the compiler is wrong,

The compiler is wrong, but for other reasons.

> because what it actually seems to be doing is parsing the "1e50" as a
> double because of the cast (speculating here, but this is consistent
> with the results).

1e50 is a double with or without the cast.

> The patch is a simple
> #ifdef __APPLE__
> #undef INFINITY
> #endif

Bleech.  I'm going to remove all this crap.  If some Crays still have broken
HUGE_VAL definitions, tough -- someone on a Cray can fix it.  Putting this
junk in the core just ensures it will always stay broken.