[Cfp-interest 2679] GB 287

[Fred J. Tydeman]

Thoughts on GB 287

IEEE-754-2019 has:

  5.4.2 Conversion operations for floating-point formats and decimal
  character sequences

  Implementations shall provide the following formatOf conversion
  operations from all supported floating-point formats to all supported
  floating-point formats, as well as conversions to and from decimal
  character sequences.


So, to convert from non-arithmetic binary128 to arithmetic decimal128
in C requires one of:
1)  d128_from_b128()
2)  d128 <-(rnd) hex string <-(exact) b128     [w/ 29 hex digits]
3)  d128 <-(rnd) decimal string <-(exact) b128 [w/ 16496 digits]
4)  d128 <-(exact) decimal string <-(rnd) b128 [w/ 34 digits]

Option 1 would require many similar functions.
         Or, one function that takes the largest binary format
         and the kind of decimal FP format and 
         returns a union of all the decimal formats.
Option 2 is what is in C23.
Option 3 requires many more digits than C23 requires.
Option 4 has double rounding for subnormal decimal numbers.
         Unless the conversion to string is told the decimal FP format.


---
Fred J. Tydeman        Tydeman Consulting
tydeman at tybor.com      Testing, numerics, programming
+1 (702) 608-6093      Vice-chair of INCITS/C (ANSI "C")
Sample C17+FPCE tests: http://www.tybor.com
Savers sleep well, investors eat well, spenders work forever.