• From: Liam R E Quin <liam@w...>
• To: David Carlisle <davidc@n...>
• Date: Mon, 02 Jul 2012 18:11:50 -0400

On Mon, 2012-07-02 at 22:46 +0100, David Carlisle wrote:
> On 02/07/2012 22:24, Dan Shelton wrote:
> > The +? after the pattern makes sure the previous expression is not
> > greedy. regex(5) should explain. So it won't swallow the whole
> > document, it'll travel along the characters until it finds the first
> > -->
> 
> There is no ? after the + in the code you quoted.

Which is a good reminder of why using regular expressions to parse XML
tends to be a bad idea.

A more traditional Unix shell-like approach might be to use a
stand-alone XML parser like xp or xmllint that can produce line-oriented
output; typically this is based on the old SGML "ESIS" idea, and has a
single character at the start of each line to indicate an event type,
such as "open element", "attribute", "text" or whatever.

Liam

-- 
Liam Quin - XML Activity Lead, W3C, http://www.w3.org/People/Quin/
Pictures from old books: http://fromoldbooks.org/

• Follow-Ups:
  • Re: newline/form feed valid as attribute value?
    • From: John Cowan <cowan@m...>

• References:
  • newline/form feed valid as attribute value?
    • From: Dan Shelton <dan.f.shelton@g...>
  • Re: newline/form feed valid as attribute value?
    • From: Michael Kay <mike@s...>
  • Re: newline/form feed valid as attribute value?
    • From: Dan Shelton <dan.f.shelton@g...>
  • Re: newline/form feed valid as attribute value?
    • From: David Carlisle <davidc@n...>
  • Re: newline/form feed valid as attribute value?
    • From: Mike Sokolov <sokolov@i...>
  • Re: newline/form feed valid as attribute value?
    • From: Dan Shelton <dan.f.shelton@g...>
  • Re: newline/form feed valid as attribute value?
    • From: David Carlisle <davidc@n...>