Log message:
amavisd-new: Requires p5-CryptX at runtime.

Reported by jfqd on #smartos.  Bump PKGREVISION.

Log message:
update package to version 2.12.2

It's the successor to amavisd-new both in sources and in versioning,
but strictly speaking it's now "amavis". I kept the package name to
make it easier to update existing installations.

Upstream release notes:
                                                              October 13, 2021
   amavis-2.12.2 release notes

   BUG FIXES

   - Allow $timestamp_fmt_mysql to be used with the DBD::MariaDB driver.
     Reported by Marcel Evenson.
     Issue: https://gitlab.com/amavis/amavis/issues/79
     MR:    https://gitlab.com/amavis/amavis/merge_requests/78

   - Resolve utf8mb4 problems when using DBD-MariaDB.
     Reported by Marcel Evenson.
     Issue: https://gitlab.com/amavis/amavis/issues/67
     MR:    https://gitlab.com/amavis/amavis/merge_requests/80

   - Set correct originating flag when using milter/AM.PDP without policy banks.
     Reported by Henrik K.
     Issue: https://gitlab.com/amavis/amavis/issues/61
     MR:    https://gitlab.com/amavis/amavis/merge_requests/81

   - Resolve crash on reload with insufficient permissions.
     Amavis now tests whether it is able read and evaluate its
     configuration files with dropped privileges.
     In case it cannot, amavis fails to start and refuses to reload.
     Reported by Michael Orlitzky.
     Suggestions by Ralph Seichter and Michael Orlitzky.
     Issue: https://gitlab.com/amavis/amavis/issues/10
     MR:    https://gitlab.com/amavis/amavis/merge_requests/74
     MR:    https://gitlab.com/amavis/amavis/merge_requests/75

   - Resolve SSL client connection hangups with broken pipe
     Reported by @kolbma.
     Issue: https://gitlab.com/amavis/amavis/issues/73
     MR:    https://gitlab.com/amavis/amavis/merge_requests/71

   ---------------------------------------------------------------------------
                                                             November 13, 2020
   amavis-2.12.1 release notes

   BUG FIXES

   - Generate DKIM record comment line including the 's=' (selector) tag
     instead of an 'i=' (identifier) tag when using "amavisd showkeys".
     Reported by Juan Orti Alcaine.
     Issue: https://gitlab.com/amavis/amavis/issues/63
     MR:    https://gitlab.com/amavis/amavis/merge_requests/41

   - Make sure generated Authentication-Results follow RFC specification.
     MR: https://gitlab.com/amavis/amavis/merge_requests/57

   - Prevent re-encoding of notification templates.
     Fixed by Henrik Krohns and Michael Weiser.
     Issue: https://gitlab.com/amavis/amavis/issues/62
     MR:    https://gitlab.com/amavis/amavis/merge_requests/40

   - Compare inode numbers as strings.
     Reported and implemented by Mark Martinec.
     Issue: https://gitlab.com/amavis/amavis/issues/48
     MR:    https://gitlab.com/amavis/amavis/merge_requests/21

   - Resolve MySQL invalid utf8mb4 clause.
     Reported by Henrik Krohns.
     Issue: https://gitlab.com/amavis/amavis/issues/33
     MR:    https://gitlab.com/amavis/amavis/merge_requests/14

   ---------------------------------------------------------------------------
                                                                 July 25, 2019
   amavis-2.12.0 release notes

   This software is renamed from amavisd-new to amavis.

   NEW FEATURES

   - Introduce Rspamd client extension
     With this extension, Amavis can use Rspamd either running on the same
     server or remotely. Connections are made using HTTP/HTTPS depending on
     configuration, the latter requiring a HTTPS-capable proxy (like NGINX or
     Apache) for Rspamd, which does not natively support HTTPS. Basic
     authentication with name/password pairs is also available.
     Implemented by Ralph Seichter.
     MR: https://gitlab.com/amavis/amavis/merge_requests/11

   BUG FIXES

   - Treat "not an OLE file" as a successful result
     Amavis supports calling the ripOLE program to extract embedded objects
     from Microsoft OLE documents. However, not all Microsoft documents
     contain said objects, and the underlying file format changes when they
     do. Since Amavis can't tell the difference, it passes everything to
     ripOLE unconditionally.  Amavis now treats the "not an OLE file" error
     code of ripOLE as a successful result, proceeds normally and scans the
     file as a whole.
     Fixed by Michael Orlitzky.
     MR: https://gitlab.com/amavis/amavis/merge_requests/9

   - Fix unix socket path extraction that has prevented a socket based
     policy bank to be loaded;
     Fixed by Boris Gulay.
     MR: https://gitlab.com/amavis/amavis/merge_requests/2
     MR: https://gitlab.com/amavis/amavis/merge_requests/19

   - Fix DKIM signing for outbound messages.
     Fixed by Ralph Seichter.
     MR: https://gitlab.com/amavis/amavis/merge_requests/1

   - Fix unescaped left brace regex warning in run_av() subroutine.
     Fixed by Ralph Seichter.
     MR:    https://gitlab.com/amavis/amavis/merge_requests/10
     Issue: https://gitlab.com/amavis/amavis/issues/31

   - Mention default value for $myprogram_name in minimal amavisd.conf.
     MR:    https://gitlab.com/amavis/amavis/merge_requests/12
     Issue: https://gitlab.com/amavis/amavis/issues/36

Log message:
*: recursive bump for perl 5.36

Log message:
mail: Replace RMD160 checksums with BLAKE2s checksums

All checksums have been double-checked against existing RMD160 and
SHA512 hashes

The following distfiles were unfetchable (possibly fetched
conditionally?):

./mail/qmail/distinfo netqmail-1.05-TAI-leapsecs.patch

Log message:
mail: Remove SHA1 hashes for distfiles

Log message:
*: recursive bump for perl 5.34

Log message:
*: bump PKGREVISION for perl-5.32.

Log message:
mail/amavisd-new: Fix build after removal of replace-interpreter

- Add NO_BUILD=yes
- Move SUBST_STAGE to pre-configure target

Log message:
erlang, amavisd, policyd: remove redundant replace-interpreter

Since do-configure-pre-hook already depends on replace-interpreter, there
is no point in making any other stage depend on that as well.  At best,
it has no effect.  At worst it creates a hard-to-find difference between
builds that run "bmake install" directly and builds that split the build
into "bmake configure && bmake build && bmake \ 
install", as bulk builds
do.

Log message:
Bump PKGREVISIONs for perl 5.30.0