Any discussion of more Perl-like regex capability inXPath 2.1?

[David Sewell]

Michael, thanks for the detailed feedback. My most important question
was answered, namely whether the working group has talked about
expanding regex/replacement functionality, and your reasons for not
adding the particular Perl extensions I mentioned are well taken,

David S.

On Fri, 21 May 2010, Michael Kay wrote:

>
> There is one other extensions to regexes in 2.1, support for non-capturing
> groups has been added.
>
> I don't think there is willingness in the WG to trawl what's in Perl and add
> every feature that seems to make sense in an XPath context. Rather, features
> have been added if there is a strong use case, if the specification is
> well-defined in an XML/XPath context, and if the feature (not necessarily
> with identical syntax) appears to be supported in most of the regular
> expression libraries that XPath/XQuery implementors are likely to be using.
>
> Quite often with requests like this the difficulty is in finding a precise
> specification of what the Perl construct actually does. My first attempt to
> lookup the spec for \l and \u gives:
>
> "lowercase|uppercase next char (think vi)"
>
> Given a joke of a specification like that (and this isn't untypical), we
> have in the past ended up running tests on implementations to see what the
> feature actually does, so that we can specify behaviour that will be
> consistent with Perl. Reverse engineering a spec by running tests against
> implementations is very labour-intensive, and the WGs are short of resource,
> so we're naturally disinclined to do it.
>
> As far as I can see this isn't actually an addition to the syntax of regular
> expressions you are proposing, it's an addition to the syntax of the
> replacement string in the replace() function. I think our view in the WG is
> that operations more complex than can be done using replace() require a
> feature like XSLT's xsl:analyze-string instruction, and to this end we have
> added the function fn:analyze-string() which will become available in XQuery
> 1.1: see
>
> http://www.w3.org/TR/2009/WD-xpath-functions-11-20091215/#func-analyze-strin
> g
>
> I think this achieves what you are looking for in a much more general and
> powerful way.
>
> Regards,
>
> Michael Kay
> http://www.saxonica.com/
> http://twitter.com/michaelhkay
>
>
>
> > -----Original Message-----
> > From: http://x-query.com/mailman/listinfo/talk
> > [mailto:http://x-query.com/mailman/listinfo/talk] On Behalf Of David Sewell
> > Sent: 20 May 2010 18:27
> > To: XQuery Talk
> > Subject:  Any discussion of more Perl-like regex
> > capability inXPath 2.1?
> >
> > So far as I can see in the working draft of "XPath and XQuery
> > Functions and Operators 1.1", the only real extension to
> > regular expressions is the addition of a "q" flag in matches
> > to block metacharacter
> > interpretation:
> >
> > http://www.w3.org/TR/xpath-functions-11/#flags
> >
> > I'm just curious whether there has been discussion of adding
> > any more Perl-like expressivity to XPath regular expressions.
> > For example, being able to use the escape sequences \u, \l,
> > etc. in replacement patterns would allow concision that is
> > currently not legal:
> >
> >  let $s := "from George Washington, 17 May 1789"
> >  return replace($s, '^(.+),.*', '\u$1')   (: ILLEGAL :)
> >
> >  ==> "From George Washington"
> >
> > instead of the verbosity required currently to do the same thing:
> >
> >   let $s := "from George Washington, 17 May 1789"
> >   return concat(
> >       upper-case(substring($s,1,1)),
> >       replace(substring($s,2), '^(.+),.*','$1')
> >       )
> >
> > I understand this would impose more burden on implementers,
> > but has there been demand from the user community (including
> > XSLT folk)?
> > (Count this email as "demand", of course.)
> >
> > DS
> >
> > --
> > David Sewell, Editorial and Technical Manager ROTUNDA, The
> > University of Virginia Press PO Box 801079, Charlottesville,
> > VA 22904-4318 USA
> > Courier: 310 Old Ivy Way, Suite 302, Charlottesville VA 22903
> > Email: http://x-query.com/mailman/listinfo/talk   Tel: +1 434 924 9973
> > Web: http://rotunda.upress.virginia.edu/
> > _______________________________________________
> > http://x-query.com/mailman/listinfo/talk
> > http://x-query.com/mailman/listinfo/talk
>
>

-- 
David Sewell, Editorial and Technical Manager
ROTUNDA, The University of Virginia Press
PO Box 801079, Charlottesville, VA 22904-4318 USA
Courier: 310 Old Ivy Way, Suite 302, Charlottesville VA 22903
Email: http://x-query.com/mailman/listinfo/talk   Tel: +1 434 924 9973
Web: http://rotunda.upress.virginia.edu/