Re: The new 'isinf.c' function in port.

Marc G. Fournier <scrappy@hub.org>

From: The Hermit Hacker <scrappy@hub.org>
To: "Billy G. Allie" <Bill.Allie@mug.org>
Cc: hackers@postgreSQL.org
Date: 1998-02-06T12:59:56Z
Lists: pgsql-hackers
On Thu, 5 Feb 1998, Billy G. Allie wrote:

> Hi Marc,
> 
> I was looking at the new 'isinf.c' function in the src/backend/port directory.
> I have a question about the inclusion of cases for FP_SNAN and FP_QNAN.
> 
> 	Does the 'isinf' function mean "not finite" or "is infinity"?
> 
> If it is to mean "not finite", then (IMHO) then checking for FP_SNAN and 
> FP_QNAN is appropiate.  If it is to mean "is infinity", then the function 
> should only be checking (again, IHMO) for FP_NINF and FP_PINF.

	Good question...

	From FreeBSD's man page:

The isninf() function returns 1 if the number is ``infinity'', otherwise
0.

	In my defence, those functions were just pulled in from the
various ports themselves...I believe you are correct about removing
FP_SNAN/FP_QNAN...