The Perl Toolchain Summit needs more sponsors. If your company depends on Perl, please support this very important event.

NAME

Mail::SpamAssassin::Plugin::Shortcircuit - short-circuit evaluation for certain rules

SYNOPSIS

  loadplugin     Mail::SpamAssassin::Plugin::Shortcircuit

  report Content analysis details:   (_SCORE_ points, _REQD_ required, s/c _SCTYPE_)

  add_header all Status "_YESNO_, score=_SCORE_ required=_REQD_ tests=_TESTS_ shortcircuit=_SCTYPE_ autolearn=_AUTOLEARN_ version=_VERSION_"

DESCRIPTION

This plugin implements simple, test-based shortcircuiting. Shortcircuiting a test will force all other pending rules to be skipped, if that test is hit. In addition, a symbolic rule, SHORTCIRCUIT, will fire.

Recommended usage is to use priority to set rules with strong S/O values (ie. 1.0) to be run first, and make instant spam or ham classification based on that.

CONFIGURATION SETTINGS

The following configuration settings are used to control shortcircuiting:

shortcircuit SYMBOLIC_TEST_NAME {ham|spam|on|off}

Shortcircuiting a test will force all other pending rules to be skipped, if that test is hit.

Recommended usage is to use priority to set rules with strong S/O values (ie. 1.0) to be run first, and make instant spam or ham classification based on that.

To override a test that uses shortcircuiting, you can set the classification type to off.

Note that DNS and other network lookups are launched when SA reaches priority -100. If you want to shortcircuit scanning before any network queries are sent, you need to set lower than -100 priority to any such rule, like -200 as in the examples below.

Shortcircuited test will be automatically set to priority -200, but only if the original priority is unchanged at default 0.

on

Shortcircuits the rest of the tests, but does not make a strict classification of spam or ham. Rather, it uses the default score for the rule being shortcircuited. This would allow you, for example, to define a rule such as

  body TEST /test/
  describe TEST test rule that scores barely over spam threshold
  score TEST 5.5
  priority TEST -200
  shortcircuit TEST on

The result of a message hitting the above rule would be a final score of 5.5, as opposed to 100 (default) if it were classified as spam.

off

Disables shortcircuiting on said rule.

spam

Shortcircuit the rule using a set of defaults; override the default score of this rule with the score from shortcircuit_spam_score, set the noautolearn tflag, and set priority to -200. In other words, equivalent to:

  shortcircuit TEST on
  priority TEST -200
  score TEST 100
  tflags TEST noautolearn
ham

Shortcircuit the rule using a set of defaults; override the default score of this rule with the score from shortcircuit_ham_score, set the noautolearn and nice tflags, and set priority to -200. In other words, equivalent to:

  shortcircuit TEST on
  priority TEST -200
  score TEST -100
  tflags TEST noautolearn nice
shortcircuit_spam_score n.nn (default: 100)

When shortcircuit is used on a rule, and the shortcircuit classification type is set to spam, this value should be applied in place of the default score for that rule.

shortcircuit_ham_score n.nn (default: -100)

When shortcircuit is used on a rule, and the shortcircuit classification type is set to ham, this value should be applied in place of the default score for that rule.

TAGS

The following tags are added to the set available for use in reports, headers etc.:

  _SC_              shortcircuit status (classification and rule name)
  _SCRULE_          rulename that caused the shortcircuit 
  _SCTYPE_          shortcircuit classification ("spam", "ham", "default", "none")

SEE ALSO

https://issues.apache.org/SpamAssassin/show_bug.cgi?id=3109