Type inference of inferred repeat expr counts is not as strong as it could be

[BoxyUwU]

See the FCP writeup for inference of inferred repeat expr counts: #139635 (comment)

Once marker traits are stable we should experiment with a RepeatExprWf lang item trait in core with marker impls for [T: Copy; N], [T; 0] and [T; 1] which will allow for proper ability to defer/stall the repeat expr wf checks.