4.4 Message Matching

The functions in this section match the body of an email message against a regular expression. If the message is a multipart one (MIME), then its parts are descended recursively and each part that is recognized as a textual one is matched against the regexp.

The message is identified by its descriptor, a numeric value returned from a previous call to mailbox_get_message, message_from_stream, current_message, or similar function.

On success, these functions return a string identifying the part of the message that matched. For example, ‘1’ means the body of a plaintext email message. ‘1.2’ means second part (part numbers are 1-based) of a multipart message.

If the match isn’t found, empty string is returned.

Function: string pcre_match_message (number rx, number msg; string charset)

Checks whether the body of the message msg (or any of its parts, if it is a MIME message) matches regular expression referred to by descriptor rx.

If optional charset is supplied, each message part is re-coded to that charset before matching.

Function: string pcre_string_match_message (string rs, number msg; string charset)

Compiles rs and checks whether the body of the message msg (or any of its parts, if it is a MIME message) matches it.

If optional charset is supplied, each message part is re-coded to that charset before matching.

These functions use the Content-Type header to decide whether to apply the regular expression to the given message part, or just to skip it. An internal structure called text mime type list contains globbing patterns of the eligible content type values. The actual value of the Content-Type header is matched against each list element in turn, until either a match is found (in which case the part will be processed) or the end of list is reached (in which case the part will be skipped).

By default, the text mime type list is initialized with a single value: ‘text/*’. To maintain the list, the following functions are provided:

Function: void pcre_text_mime_init ()

Clears the list.

Function: void pcre_text_mime_add (...)

Arguments are strings in form ‘part/subpart’, where both part and subpart can be shell-style globbing patterns. The supplied strings are appended to the list.

Function: void pcre_text_mime_del (...)

Deletes from the list those patterns that match literally its arguments.