[cfp-interest 3930] Re: Trailing Decimal Exponent on a Hexadecimal standard floating type

[Damian McGuckin]

With reference to the numerous emails on this topic, the following might
avoid wasting time discussing their details in about 36 hours time.

There will soon be a submission to IEEE 754 proposing that the Trailing 
Decimal Exponent on a Hexadecimal standard floating type literal be made 
optional when used as a literal within a program.  My original question
was raised before this submission seemed a certainty and made much of the 
discussion in CFP a bit moot.

C has long accepted such a format as input to strtod().

This has been supported in the Chapel language for 8 years. Works well.

I would assume the submission to be discussed and pass 754 by August at 
which time, I might suggest that C do the same. This would be a technical 
change for C.

IEEE 754 has no remit to standardize the output style.

C can sort of enforce an output style by retaining how "%a" is currently 
interpreted. That said, it is a bit loose when dealing with subnormals as 
Paul highlighted  (and on which my own comment was less than precise).

Thanks - Damian