Not all float/double tags in macro arguments are uniquely stringified

The bug

Not all float/double macro arguments are uniquely (in a way that can be distinguished at least from other values) stringified because the maximum number of fraction digits is limited to 15. This could result in multiple float/double tags being converted to the same string representation.

How to reproduce

$say $(a)

Float tags

Run

function a {a: 1.4E-45f}

Expected behavior

Unique representation like 0.000000000000000000000000000000000000000000001401298464324817

Actual behavior

0

Double tags

Run

function a {a: 4.9E-324d}

Expected behavior

Unique representation like 0.0000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000049

Actual behavior

0