[cfp-interest 3861] Re: Rounding terminology

[Damian McGuckin]

The following appears in the Introduction to the Annex F Special Cases.

The use of the term current rounding direction goes back to C99 when there 
was only the dynamic rounding direction mode, and no constant rounding 
mode. At that time, the meaning of current rounding direction was 
reasonably clear, being the value of the (dynamic) rounding direction mode 
at the pertinent point in the execution of the program. Such usages of 
"current rounding direction" have remained. But now, with C23 having 
constant rounding direction modes at play, the meaning of current rounding 
direction is now not always clear. In particular, the previous meaning 
that it just refers to the dynamic rounding direction mode is not 
generally suitable. This proposal eliminates the use of the term current 
rounding direction from Annex F. Where this term needs to change elsewhere 
in the standard will be addressed in a separate paper at a later date.

Thanks - Damian