exp() versus the POSIX standard

The POSIX standard says this about the exp(3) function:

    If x is -Inf, +0 shall be returned.

At least on my Linux box, our version does no such thing:

regression=# select exp('-inf'::float8);
ERROR:  value out of range: underflow

Does anyone disagree that that's a bug?  Should we back-patch
a fix, or just change it in HEAD?  Given the lack of user
complaints, I lean a bit towards the latter, but am not sure.

            regards, tom lane