B<minc> incorporates mail from a maildir to a mh folder hierarchy. It
takes mail from a maildir folder (not a maildir folder hierarchy),
-optionally checks for spam with a user-defined spam-checking function,
-and optionally filters mail into separate mh folders.
-
-The filtering is quite sophisticated, as it is done using real Perl
-matching (m//) commands.
+running each message through regular expression based filter and hook
+functions to determine in which folder to store it or whether it's
+spam. Post-processing hooks may be applied to each message.
+
+As it processes each message, B<minc> prints a line for each message
+similar to B<inc(1)> and B<scan(1)>. This line includes the folder
+and message number in which the message was stored, the last 'From'
+header, and the last 'Subject' header. These fields are truncated to
+fit in the user's terminal (see COLUMNS in B<ENVIRONMENT> below) in
+the following proportions: folder (0.1), message number (0.0625), from
+header (0.175). Any of these may be overridden with $SCAN_P_FOLDER,
+$SCAN_P_MESSAGE, or $SCAN_P_FROM. The subject always fills out the
+rest of the line.
=cut
=item B<-d>
Dump (using Data::Dumper) the FILTERS list and exit. This is useful
-for testing the syntax of .mincfilter.
+for testing the syntax of .minc.
=item B<-h>
=over 4
+=item COLUMNS
+
+How many columns the user's terminal can hold, used to print scan
+lines for each processed message. Defaults to 80.
+
=item HOME
-Where configuration files (.mincfilter) are found. Also,
+Where the configuration file (.minc) is found. Also,
$HOME/Maildir is used for the maildir if MAILDIR is not set.
=item MAILDIR
=over 4
-=item $HOME/.mincfilter
-
-This file is Perl code (included via the 'require' directive) which is
-expected to define the FILTERS list.
+=item $HOME/.minc
-=item $HOME/.mincspam
-
-If this file exists, B<minc> will include it with the expectation that
-it will define a B<spam_check> function. This function takes a
-message filename as an argument and returns 1 if the message is spam,
-else 0. If this file does not exist, B<minc> will define a simple
-function that always returns 0.
-
-One of B<minc>'s global variables is available to the user-defined
-B<spam_check> function: $run. This boolean should be honored;
-B<spam_check> should only take real action (i.e. removing or creating
-files, running external programs, etc.) if $run is non-zero.
+This file is Perl code (included via the 'require' directive) which
+may define the FILTERS list, @start_hooks, @filter_hooks,
+@post_store_hooks, and @stop_hooks.
-This file may also declare two other functions: B<spam_start_hook> and
-B<spam_stop_hook>. The former is passed no arguments and is expected
-to return a list. This list is a "baton" that will also be passed to
-B<spam_check> and B<spam_stop_hook>. It can hold anything
-B<spam_check> will need to do its job, whether network sockets, pipes,
-or whatever.
+=item `mhpath +`/.folders
-XXX: need more details about the spam-handling process; for now read
-the code.
+B<minc> adds all folders it filters into to this file, which is used
+by lukem's B<new(1)> (XXX need a link).
=item `mhpath +`/logs/minc.log
=head1 THE FILTERS STRUCTURE
-The user's .mincfilter file must define the @FILTERS structure. This
-structure is an array. Each element of @FILTERS is a filter. A
-filter is itself an array. The first element of a filter is a string,
-the name of the header this filter acts upon. The header name is not
-case-sensitive. Each subsequent element of a filter is a pair (i.e. a
-two-element array): first, a regular expression B<minc> uses to
-determine whether this filter matches or not, and second, an
-expression which B<minc> evaluates to get the folder name.
+Each element of @FILTERS is a filter. A filter is itself an array.
+The first element of a filter is a string, the name of the header this
+filter acts upon. The header name is not case-sensitive. Each
+subsequent element of a filter is a pair (i.e. a two-element array):
+first, a regular expression B<minc> uses to determine whether this
+filter matches or not, and second, an expression which B<minc>
+evaluates to get the folder name.
B<minc> decides where to store a message by iterating over the
@FILTERS array. It tests each regexp of each filter against all
including) the To. Which headers to use is determined by a regular
expression borrowed from procmail.
+=head1 HOOKS
+
+Filter hooks take a reference to a hash of batons, a reference to a
+hash of headers, and the message filename as arguments. It returns
+undef to decline filtering of this message (thus falling back to
+subsequent filter hooks, and finally @FILTERS), or the name of the
+folder to store this message into.
+
+One of B<minc>'s global variables is available to the user-defined
+hooks: $run. This boolean should be honored; hooks should only take
+real action (i.e. removing or creating files, running external
+programs, etc.) if $run is non-zero.
+
+The baton hash is created simply from the start hooks; if the hook
+returns at least one defined value, this value is used as the key and
+all other return values are put into a list reference as the value.
+This hash is then passed by reference to all filter, post-store, and
+stop hooks.
+
+Post store hooks take a reference to a hash of batons, a reference to
+a hash of headers, the folder this message was stored in, and its new
+message number.
+
+XXX: need more details about the hook process; for now read the code.
+
=head1 EXAMPLES
@FILTERS = (
list such as dev@httpd.apache.org, this filter will create the folder
name l/apache/httpd/dev.
-For an example B<spam_check> function, see
-L<http://pretzelnet.org/cvs/dotfiles/.mincspam>
+XXX Need hook examples.
=head1 AUTHORS
projects / minc / commitdiff