[Cfp-interest] llogb(1.)

[Jim Thomas]

llogb and ilogb return in integer format, so we don’t need to say anything about the sign of a zero result for them. 

For logb we do seem to be missing specification for the sign of a zero result. 754 says

logB(1) is +0.

Shouldn’t that be

logB(+/-1) is +0.

Whichever, we could add an equivalent statement in F.10.3.11.

Alternatively, we could add something in 7.12.6.11, though would have to account for the possibility of unsigned zeros. Maybe

If the argument is +/-1, the result is positive or unsigned zero.


- Jim Thomas

> On Aug 3, 2018, at 1:35 PM, Fred J. Tydeman <tydeman at tybor.com> wrote:
> 
> Do we need to add words along the lines of:
> 
> llogb(1.) is +0L
> ilogb(1.) is +0
> logb(1.) is +0.0 if signed zeros are supported, otherwise, unsigned zero.
> 
> This is a IEEE-754 requirement.
> 
> 
> ---
> Fred J. Tydeman        Tydeman Consulting
> tydeman at tybor.com      Testing, numerics, programming
> +1 (702) 608-6093      Vice-chair of PL22.11 (ANSI "C")
> Sample C99+FPCE tests: http://www.tybor.com
> Savers sleep well, investors eat well, spenders work forever.
> 
> _______________________________________________
> Cfp-interest mailing list
> Cfp-interest at oakapple.net
> http://mailman.oakapple.net/mailman/listinfo/cfp-interest

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.oakapple.net/pipermail/cfp-interest/attachments/20180821/04872137/attachment.html