rurban / asterisk-matcher Goto Github PK

2003 I added a dynamic regular expression matcher to asterisk,
which has a weird synatx, but otherwise the matcher looked fine.

* a regex must start with "_"
* regex patterns are case-insensitive except characters inside []
* "." matches zero or more characters (as in * in glob)
* character ranges as in [0-9a-zA-Z]
* X,Z,N match 0-9,1-9,2-9 resp.    
* "?" matches any character
* negation as in [^0] ("any char but 0")
               or [^a-z]
* explicit quantifiers as in X{2,4} ("from 2 to 4 digits"),
                           or X{2,}  ("at least 2 digits"),
                           or X{2}   ("exactly 2 digits"),
* regex-style quantifiers like ?, + and * are supported
   by "{}" grouping.
     ? <=> {0,1}
     + <=> {1,}
     * <=> {0,}
* grouping as in N(1X){1,2}  ("one or two sequences of 1X")
* capturing
   with () grouped matches are stored in subsequent numbered global
   variables, starting with $1, $2 and so on.
* alternation as in (01|0|99) ("01 or 0 or 99")
   This is not yet supported, the subsequent alternatives are not
   checked yet.

Copyright 2003 Reini Urban, MIT Licensed.

In the end it was removed from CVS before a release without me
noticing because the variable capturing was not thread-safe. Would
have been trivial to fix.