[cfp-interest 3819] Re: Annex F Special Cases

[Damian McGuckin]

On Tue, 17 Mar 2026, Jim Thomas wrote:

> Here?s a link to some comments:

The change for scalbn is needed.

 	scalbn(x, 0) returns a quiet NaN when x is a signaling NaN.

The replacement domain explicitly excludes a signaling NaN.

See my explanation in the update.

In the totalorder functions, I do not keep these sentence separate

 	These functions are full specified by ISO/IEC 60559

because this sentence says one thing about "these functions" and the 
next sentence says two things about "these functions".

I chose one sentence to say the three things about "these functions".

We could have three separate consecutive sentences that say those 
things but then they all start the same way which is poor English
as I was taught and Professor Strunk from Yale mentions in his book on 
English style (I think). Your call.

I will update the file shortly but I have yet to fix all those poow and 
rootn cases where I did not put in the red text. My fingers, head and eyes 
hurt.

Thanks - Damian