L2/08-114

Source: Mark Davis
Subject: Recommended Regex Changes
Date: 2008-02-08

Please add this to the doc registry and agenda:

We've done some work recently on accepted practices in regex implementations, and have become aware that we should update the notation in UTS #18.

1. Recommend doubled characters for set operations

 
Operator
 		 Name
 
 ||
 		 union (normally suppressed)
 && 
 		 intersection
 --  
 		 set difference (removal)
 ~~   
 		 symmetric difference

2. Record the recommended precedence (union before others).

3. Note the use of the single characters, perhaps with different precedence and restrictions on syntax, and the fact that that a single hyphen can be problematic because of its use in ranges.

4. Add clarifications and wordsmithing where necessary.

--
Mark