Milter-greylist

John Thiltges <jthiltges2@...> wrote:

> Attached is a patch to replace the logging of ACL line number with ACL
> id string (if an id is specified). It uses fstring_expand(..., "%a") to
> build the string.

What tabout the performance impact of this change? Now we go to
fstring_expand() regardless of the existance of an ACL id string, and
that function at least performs two mallocs. Perhaps we should test for
it before entering fstring_expand()?

-- 
Emmanuel Dreyfus
http://hcpnet.free.fr/pubz
manu@...