[boost] Re: Decimal Library re-review begins today

Thanks for the explanation. It makes sense to document the status of FP exception support and wait until SG6 comes up with a reasonable solution. Matt Borland wrote:

An example is at the top of the <cmath> page it talks about none of the functions actually meet the IEEE standard of 0.5 ULP precision since this is an unreasonable expectation [1].

I'd suggest splitting that statement on CR into a mandatory-operation part (IEEE 754 Clause 5, e.g. conversions, arithmetic operations, fma, sqrt, …) and an optional-operation part (Clause 9). This would help clarify which parts are required for conformance and which are outside the standard's mandatory scope. Regards, Michel