Robert Treat <xzilla(at)users(dot)sourceforge(dot)net> writes:
> The current wording implies that you replace arg1 with arg2 in arg3, but
> replace actually replaces occurences of arg2 with arg3 in arg1. Attached
> patch makes the function comment more closely resemble that, and is more
> in-line with the docs wording as well.
Applied, thanks.
regards, tom lane