[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]
Re: Trying to get sendmail-8.15.1 from pkgsrc working on NetBSD-5, libresolv issues
hello. Following up on my own question yet again, I think I found the
trouble and I don't think it's NetBSD-5 specific.
has the following change:
RCS file: /cvsroot/pkgsrc/mail/sendmail/files/site.config.m4-sasl2,v
retrieving revision 1.2
retrieving revision 1.3
diff -u -r1.2 -r1.3
--- site.config.m4-sasl2 30 Aug 2004 20:16:28 -0000 1.2
+++ site.config.m4-sasl2 20 Jun 2014 16:26:55 -0000 1.3
@@ -1,5 +1,8 @@
-# $NetBSD: site.config.m4-sasl2,v 1.2 2004/08/30 20:16:28 adrianp Exp $
+# $NetBSD: site.config.m4-sasl2,v 1.3 2014/06/20 16:26:55 jnemeth Exp $
# enable SASL 2
+# for some strange reason SASL, and NAMED_RESN and MILTER are incompatible
The cvs log says this:
date: 2014-06-20 16:26:55 +0000; author: jnemeth; state: Exp; lines: +4 -1; commitid: z2V9mO8kpncGohFx;
Update to sendmail 8.14.9nb2: this is a pkgsrc bugfix update.
This should be the last update during the freeze.
PR/48566 - Emmanuel Dreyfus -- typo in patch-aw leading to build failure
PR/48913 - Matthias Scheler -- libmilter fails on unprivileged builds
By switching NAMED_RESN to 1, sendmail builds and uses the proper
version of the resolver routines such that they work with the libpthread library
linked in. Preliminary testing suggests this is a building problem, rather
than a run time problem, as all the milters we have configured seem to be
working fine with the resultant binaries.
So, my question is, is this change in place solely to accomodate the
fact that we do unprivileged builds on the bulk builders? If so, might
it be useful to put a note in the Makefile or a README file in the package
directory explaining this issue and the work around that's available? The
work around seems to be to perform privileged builds, which work fine.
If you want to use sendmail with the latest version of the openssl
libraries, which was what I was trying to do, you absolutely need to use
libpthread, so you need to build with NAMED_RESN set to 1. Let me be
clear, I'm not objecting to the change, but I think it should be more
clearly documented so that folks building this package can avoid the tail
chasing I've engaged in for the past few days. Any suggestion on the best
way to provide that documentation?
Main Index |
Thread Index |