After one and a half years of development, the release of the spam filtering platform SpamAssassin 4.0.2 has been published. The decision to block in SpamAssassin is made comprehensively. Initially, the message is checked using various methods, such as contextual analysis, black and white DNSBL lists, trainable Bayesian classifiers, signatures of known spam, and sender authentication via SPF and DKIM. As a result of the checks carried out, a weight coefficient is accumulated, which must not exceed a certain threshold; otherwise, the message is blocked or marked as spam. The package can be used on both client and server systems. It supports automatic updating of filtering rules. The SpamAssassin code is written in Perl and is distributed under the Apache 2.0 license.
In the new release:
- Compatibility with the July release of Perl 5.42 and new versions of some Perl modules is ensured.
- A new Mail::SpamAssassin::Plugin::Redirectors plugin has been added to identify URLs that are being redirected through shortened link services.
- A handler has been added to the HeaderEval plugin for identifying incorrect headers.
- Support for replication in the Redis database has been added. The connection parameters for Redis are divided into read_only and read_write.
- Support for using the Perl module IP::Geolocation::MMDB to determine location has been added to the built-in Mail::SpamAssassin::GeoDB module.
- MIME header detection has been improved.
- A 'replace_rules' operation has been added to the Mail::SpamAssassin::Plugin::HashBL module.
- CNAME record expansion in DNS has been ensured and added to the actual domain to the 'uri_detail_list'.
- The handling of internationalized domain names (IDN) has been improved.
- The TxRep plugin has added settings txrep_min_score and txrep_max_score to limit the weight coefficients issued by the plugin.
Source: opennet.ru
