[Cfp-interest 3232] Re: Various Updates to Annex F+G

[Damian McGuckin]

On Tue, 20 Aug 2024, Jim Thomas wrote:

> I believe the parts ?For Consistency Between F+G? need to be reviewed 
> together.

So, is this small breakdown of F.10.1 and G.4.1 OK, or do you want me to 
itemize and compare paragraph by paragraph so it is more obvious what I 
was trying to achieve.

Once you have that consistency, it makes it easier to read.

And both F.10.1 and G.4.1 have virtually no mathematics in them so they 
make easier reading.

> But you might consider separating out other parts. That way CFP could 
> have agenda slots for focused discussion of the separate parts.

I will. But I was trying to get these first two chunk right first before I 
attacked the others.

> Keeping it all together (as is) should suffice for getting an initial 
> round of review and comments from CFP. Then we could have an agenda slot 
> to start through comments.

I was probably going to keep the special cases together for now and let 
CFP decide on how to break those down into small chunks. They are changes 
that follow that syntax of

      <function-name> ( <arguments> )
      [ STORES <value-expression> INTO <pointer-name> ]
      [ RETURNS <result> [ FOR <domain> ] [ ( <comment> ) ] ]
      [ [optionally] RAISES <exception> [ WHEN <scenario> ] ]

By the way, that syntax is not mine. Except for the WHEN clause (which is 
introduced to handle that case in 'nextafter') and the fact that the 
domain (FOR clause) appears before the exception (RAISES clause), the 
above syntax is a just distillation of the syntax that currently occurs 
most commonly throughout Annex F and Annex G. So, it is not something new. 
That syntax was already there. It was just not obeyed all the time. All I 
am doing is forcing every special case to follow the same rule. That way, 
the reader does not have to worry about whether they are interpreting 
subtle variants the right way.

Thanks for looking at it - Damian