[Cfp-interest] AI about C vs IEEE 754 specifications of pow

[Jim Thomas]

>     Jim: pow: Add a note to F10.4.4 pow to say it is the same as IEEE-754.
> 


C11 annex F already has a general statement about this. With the changes in TS 18661-1, it appears in F.1 as

> [3] An implementation that defines __STDC_IEC_60559_BFP__ to 201ymmL shall conform to the specifications in this annex.356) Where a binding between the C language and IEC 60559 is indicated, the IEC 60559-specified behavior is adopted by reference, unless stated otherwise.
> 

Also, C11 F.10 #3 says

> Special cases for functions in <math.h> are covered directly or indirectly by IEC 60559. The functions that IEC 60559 specifies directly are identified in F.3. 

TS 18661-1 adds in F.10

> [4a] The functions bound to operations in IEC 60559 (F.3) are fully specified by IEC 60559, including rounding behaviors and floating-point exceptions.
> 


Is this sufficient? 

We observed that for pow it’s especially difficult to compare the C and IEC 60559 specifications. We could add a (redundant) note in F10.4.4:

NOTE	The pow functions, like the other functions bound to operations in IEC 60559 (F.3), are specified by IEC 60559. The IEC 60559 specification is adopted by reference. The specification here is intended to to match IEC 60559.


Jim Thomas

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.oakapple.net/pipermail/cfp-interest/attachments/20180206/2f9f62da/attachment-0001.html