US 196-315 32.5.8.3p5, 14 [atomics.types.int] Use integral-type and floating-point-type, not T

[wg21bot]

These should say integral-type and floating-point-type respectively rather than T.

Proposed change:

Modify accordingly.

[timsong-cpp]

Also applies to 32.5.8.4p5, 15

[jensmaurer]

Accepted.