bug#31718: 26.1; Strange behavior of `cond'

[Noam Postavsky]

Paul Eggert <eggert@cs.ucla.edu> writes:

> A small problem: gensym does not guarantee that the resulting symbol
> is unique so the generated symbol could in theory appear in the input

What?

    gensym is a compiled Lisp function in ‘subr.el’.

    (gensym &optional PREFIX)

    Return a new uninterned symbol.
             ^^^

How could a new symbol have already appeared in the input?

(I have no objection to the alternative you used, just wondering why
gensym is not also correct)