[Cfp-interest 2507] Re: 7.12.* versus F.*

[Jim Thomas]

> On Aug 4, 2022, at 7:38 AM, Fred J. Tydeman <tydeman at tybor.com> wrote:
> 
> Am I correct in saying that the main body of the standard, in particular,
> 7.12.*, lists all possible (general) error conditions, while F.* lists
> specific error conditions?  That is, F.* is a subset of 7.12.*.

If the implementation supports reporting errors by floating-point exceptions (required by Annex F), any required error in 7.12 must have a corresponding exception and no exception can occur unless it is allowed in 7.12. See footnote 452 in N3047. This footnote applies to spurious exceptions, but 7.12 allows implementation-defined errors if they fit the definition of overflow or underflow, so it is ok for Annex F to require an exception that isn't an explicit error in 7.12.

> 
> If that is the case, then 7.12.11.3 The nextafter functions
> is missing the underflow range error case.
> 
> ---
> 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