A long-delayed reply:

From: Richard Smith
Sent: Saturday, February 22, 2014 3:13 PM

> I'm a little concerned about providing a digit separators feature test macro. It seems to encourage people to use digit separators in #ifdef'd code, which doesn't work, so I'm not sure how much value the macro adds.

I agree that there's a problem here. Of course I would say, without
qualification, that the problem is caused entirely by the syntax selected
for digit separators.

We could choose to shine a light on the problem, or to help conceal it. But
neither of those choices would make the slightest difference about the fact
that the problem exists. After all, testing for the existence of new
standard features is not the only reason for using conditional compilation.

I predict that the decision we make will be based entirely on political
considerations. The technical arguments against providing a feature-test
macro for this new feature are really against the feature itself, not
against the macro itself.

Clark