Regular Expression Extensions in AOP
====================================

[Ant][10/11/2002]

The current implementation of the AOP allows regular expressions that match a
number of occurences of a set of characters. For example,

  number = [ "read_all", [["0"-"9"]] ];

will match any string that contains only numeric characters.

Regular expressions of this format are very limited. Some bugs exist in the
built-in agent templates that allow incorrect inputs to pass through the parser
(e.g. number and ident). One major limitation is that if you use 'read_prefix'
together with 'read_all' then you allow whitespace to occur between the two
regular expressions.

Eden (1.43 onwards) includes functions to handle Perl Compatible Regular
Expressions. The scripts in this package add extra operations to make use of
this. These operations are: "literal_re", "prefix_re" and "suffix_re". All
require a second argument that is the PCRE as string. For example:

  number = [ "literal_re", "[0-9]+" ];

The "prefix_re" and "suffix_re" have a third argument which is the name of the
agent to apply to the remaining string (the child agent). The following example
matches any number of lowercase letters at the start of the string, and passes
the remaining string to the "anything" agent:

  lowercase = [ "prefix_re", "[a-z]*", "anything" ];

These new operations all have a similar behaviour to their non-RE siblings. For
more information on "literal", "prefix", and "suffix", see documentation.