Re: [xml-dev] newline/form feed valid as attribute value?

[Mike Sokolov <sokolov@ifactory.com>]

Maybe CDATA could be processed in the "first pass"? I also didn't see 
entity expansion: maybe that is supposed to happen up front?

At this point in this ever-recurring discussion I feel obliged to post 
(yet another) link to this internet classic:

http://stackoverflow.com/questions/1732348/regex-match-open-tags-except-xhtml-self-contained-tags

if you haven't read it, scroll down to the first answer

-Mike

On 07/02/2012 05:59 PM, Amelia A Lewis wrote:
> On Mon, 2 Jul 2012 23:27:23 +0200, Dan Shelton wrote:
>    
>>> ([^><]+)            # xml text
>>>
>>> isn't right either as>  is allowed in xml text.
>>>        
>> Do you have an example xml fragment for that? I can't imagine that
>> <atag>></atag>  can be valid. libxml2 at least chokes on that.
>>      
> <doc>
>    <expression>if x>  0</expression>
> </doc>
>
> "well-formed", not valid/invalid.  In XML, the distinction is
> significant.
>
> <atag>></atag>  is also well-formed, so if libxml2 won't accept it,
> someone should open a defect report.
>
> On the other hand, this:
>
> <atag>]]></atag>  is *not* well-formed (but, umm, good luck writing a
> regex that makes it ill-formed, while allowing<![CDATA[<atag>example
> of ill-formedness</btag>]]>, which is perfectly well-formed.
>
> Argue that they're corner cases, perhaps.
>
> Amy!
>