[Cfp-interest 1893] Re: fabs and copysign cleanup
Mike Cowlishaw
mfc at speleotrove.com
Wed Jan 13 07:46:03 PST 2021
This looks good .. one minor suggestion; in:
The fabs functions compute the absolute value of [a floating-point number]
x.
instead of deleting all of 'a floating-point number', perhaps just change
'number' to 'value'? It's worth emphasising that these should/must be used
if the argument is double, etc.
[side comment .. I have, over the years, been surprised several times when I
forgot to use fabs() and used abs() -- as in other languages -- instead and
got apparently weird results ... :-)]
Mike
> -----Original Message-----
> From: Cfp-interest [mailto:cfp-interest-bounces at oakapple.net]
> On Behalf Of Jim Thomas
> Sent: 09 January 2021 18:12
> To: CFP
> Subject: [Cfp-interest 1883] fabs and copysign cleanup
>
> In addressing an action item to reword the signbit
> specification to account for it's being a bit-level
> operation, I also looked at fabs and copysign which are also
> bit-level operations, at least for IEC 60559 implementations.
> Here are some suggestions for these specification:
>
> https://wiki.edg.com/pub/CFP/WebHome/C23_proposal_-_fabs_and_c
opysign_cleanup.pdf
>
> - Jim Thomas
> _______________________________________________
> Cfp-interest mailing list
> Cfp-interest at oakapple.net
> http://mailman.oakapple.net/mailman/listinfo/cfp-interest
>
More information about the Cfp-interest
mailing list