Subject: Re: postfix broken by AF_LOCAL semantics change
To: Jaromir Dolecek <>
From: Perry E.Metzger <>
List: tech-kern
Date: 11/29/2003 09:24:23
Jaromir Dolecek <> writes:
> This sounds reasonable and I don't see any API incompatibility
> in this. Everyone agrees that connect() can block, right?

It doesn't work that way on virtually any Unix -- the exceptions are
one point release of Solaris (long dead), a couple of long dead point
releases of Linux, and Dynix (I think). Code therefore gets written
that depends on the behavior that got restored.

BTW, the dependency in Postfix was really subtle -- it took getting
a full queue before you could detect it at all. I wouldn't be
surprised if other programs breaking broke very very subtly.

Anyway, Matt has made his new behavior dependent on a sockopt (I
believe) so all is now well.

Perry E. Metzger