[9fans] APE isinf()

[Jeff Sickel]

Here's a little issue I found w/ the new Python 2.7.3 port:

acme# cd /n/sources/plan9/
acme# grep isinf */include/ape/math.h
arm/include/ape/math.h:#define isinf(x) isInf(x)
mips/include/ape/math.h:#define isinf(x) isInf(x)
power/include/ape/math.h:#define isinf(x) isInf(x)
sparc/include/ape/math.h:#define isinf(x) isInf(x)

isinf() is missing from 386 & amd64 APE headers.

Of course, having only one argument to isInf(x) will give
you the following if you actually try to use it:

/usr/jas/src/cmd/cpython/Plan9/../Python/bltinmodule.c:2173[stdin:16039] not enough function arguments: isInf

So it's probably safe to just have

#define isinf(x) isInf(x,0)

in /*/include/ape/math.h

[erik quanstrom]

> So it's probably safe to just have
>
> #define isinf(x) isInf(x,0)
>
> in /*/include/ape/math.h

yes, let's do that.

i'd like to move all the function definitions in /$objtype/include/math.h
to /sys/include/ape/math.h and move the system dependent goo to
/$objtype/include/u.h. this would require doing the same thing to
/sys/include/ape/u.h.

- erik

[Federico G. Benavento]

On Feb 11, 2013, at 6:46 PM, Jeff Sickel <j...@corpus-callosum.com> wrote:

> So it's probably safe to just have
>
> #define isinf(x) isInf(x,0)

yes, I've used that in the past.

---
Federico G. Benavento
bena...@gmail.com