[Cfp-interest 1237] Re: Action Item: Evaluation of Literals

[Jim Thomas]

> On Jan 23, 2019, at 12:35 PM, Ian McIntosh <ianmc at eol.ca> wrote:
> 
> Last meeting I was wrong.  C++ does not have a pragma to control 
> literal evaluation precision.  It just does it the right way.

What is “the right way”? 

- Jim

> 
> Some approaches we could take:
> 
> 1.  Do nothing.
> 2.  Treat this as a bug in the current standard and change it to 
> match C++ (likely getting complaints).
> 3.  Make it an implementation option, so the compiler vendor has to 
> document what they do.
> 4.  Add a compiler-defined macro so the program can detect what the 
> compiler will do, and possibly select alternate code like explicitly 
> converting precision.
> 5.  Add a user-defined macro so the program can tell the compiler 
> what to do (eg, new improved behavior or stay compatible with old).
> 6.  Same as #5 but with a pragma.
> 7.  Other?
> 
>  - Ian McIntosh
> 
> _______________________________________________
> Cfp-interest mailing list
> Cfp-interest at oakapple.net
> http://mailman.oakapple.net/mailman/listinfo/cfp-interest