FFe Update to opendkim 2.3.1

Binary package hint: opendkim

This is mostly a bug fix update with some minor, low risk feature changes. Upstream generally just supports the most recent minor version so we would be a lot better off to release with that one.

Here's the excerpt from the upstream RELEASE_NOTES (which serves the function of changelog for opendkim):

2.3.1 2011/03/28
        Feature request #SF3196060: Log retrieved reputation values.
                Patch from Andreas Schulze.
        Feature request #SF3199824: Don't log "no signature data" unless
                LogWhy is enabled or the action for unsigned mail is something
                other than "accept".
        Fix bug #SF3196069: Log more detail when dkimf_config_setlib() fails.
                Problem noted by Andreas Schulze.
        Fix bug #SF3197890: Update ARF report format for DKIM verification
                failures to use what's in draft-ietf-dkim-marf-reporting.
        Fix bug #SF3204725: Add "ReplaceHeaders" data set to constrain what
                header fields are affected by "ReplaceRules". Problem
                noted by Chris Callegari.
        Fix bug #SF3211114: Extract "BaseDirectory" from configuration
                file properly.
        Fix bug #SF3234178: Correct handling of a KeyTable entry for which the
                key data is empty. Based on a patch from Yoshiaki Yanagihara.
        Patch #SF3205699: Make all header field name and domain name data
                sets case insensitive. Submitted by Richard Rognlie.
        Add a case-insensitivity flag to odkim.db_open().
        Minor correction to DNS function return codes for unbound, which
                doesn't have an expired/no-reply distinction like libar does.
        LIBOPENDKIM: Fix up timeout handling with a DNS callback is in use.
        LIBOPENDKIM: Update for draft-kucherawy-dkim-atps-03.
        STATS: Fix bug #SF3195449: Omit data from reporters that are not
                currently marked "enabled" in the reporters table.
        STATS: Evaluate "i=" against the "d=" domain, not the "From" domain.
                Problem noted by Mark Martinec.
        STATS: A few more reports.
        MILTERTEST: Don't use non-standard printf() conversions.
        MILTERTEST: Get the SMFIP_HDR_LEADSPC logic right.
        TOOLS: Feature request #SF3200848: Add a backgrounding option
                (on by default) to opendkim-spam.