[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

RE: regular expression question

• To: "'Paul Hermans'" <paul.hermans@a...>,<xml-dev@l...>
• Subject: RE: regular expression question
• From: "Michael Kay" <mike@s...>
• Date: Fri, 26 Aug 2005 10:56:20 +0100
• In-reply-to: <E7A0B91DC0CEE742BBD436ABF11D676C5ABBDF@d...>
• Thread-index: AcWqHs0HSRAin46QQJS90gGDOhGTeQABGi+g

There's a tortured history here. The original Schema Rec had an ambiguous
syntax for such regular expressions. To solve this problem, an erratum was
issued banning them. This erratum was incorporated in the PER of 18 March
2004. However, there were objections from vendors and users that this was an
incompatible change. Therefore, in the Second Edition of 28 Oct 2004, these
expressions were re-allowed, with accompanying prose that (supposedly)
explains how to disambiguate them. [However, I've recently raised comments
saying that the prose does nothing of the kind...].

Saxon-SA 8.5 is implementing the rules in the PER, the next release will
relax the restrictions on use of "-" in a character range to implement the
2nd Ed as best I can, given that the rules are still unclear.  

> But to my surprise the same regular expression is accepted by a
> dedicated regular expression engine (RegExBuddy), who clearly 
> indicates
> that it is the character itself we are after.

There are many different dialects of regular expressions, and the fact that
a particular tool accepts a particular expression is very little evidence of
anything.


Michael Kay
http://www.saxonica.com/

• References:
  • regular expression question
    • From: "Paul Hermans" <paul.hermans@a...>

• Prev by Date: Re: indexing and querying XML (not XQuery)
• Next by Date: Re: Python and JSON vs XML???
• Previous by thread: Re: regular expression question
• Index(es):
  • Date
  • Thread