On Wed, 7 Sept 2022 at 14:17, Thomas Köppe via SG10
<sg10_at_[hidden]> wrote:
>
> Aha, thanks -- yes, if the macros are normative, then it might indeed be best if we resolve this four-fold value bump collision with an LWG issue that captures the discussion.

They seem plenty normative to me, or that has least been the intent.
The standard requires the presence and values
of a set of macros, and they affect the meaning of portable programs
written against a particular standard.

Sure, the phrasing in [version.syn]/1, "The header <version> supplies
implementation-dependent information
about the C ++ standard library (e.g.,version number and release
date)." is a bit funny, because there shouldn't
be anything implementation-dependent about the macros and their
values. Luckily, "implementation-dependent"
is not a Term of Power. :)