Re: [xml-dev] Why is < illegal in an attribute value but theequivalent h

[Pete Cordell <pete++xmldev@codalogic.com>]

On 17/03/2022 11:25, Roger L Costello wrote:
> So this is perfectly well-formed XML:
>
> <Test foo="&#x3C;x>blah&#x3C;/x>"/>
>
> And the numeric character references will be replaced during the parsing process to yield this:
>
> <Test foo="<x>blah</x>"/>
I'd say that parsing <Test foo="&#x3C;x>blah&#x3C;/x>"/> yields an 
attribute named "foo" with a value of "<x>blah</x>".

It's not creating an alternate piece of XML that is then parsed again.

The sequence is more like:

- Low level XML tokeniser reads attribute name "foo"

- Low level parse checks it's followed by "=" and quotes

- Low level parser reads until it finds end quote, getting 
"&#x3C;x>blah&#x3C;/x>"

- Internal logic converts "&#x3C;x>blah&#x3C;/x>" to "<x>blah</x>"

- Internal logic creates a data record for an attribute of name "foo" 
with value "<x>blah</x>" and associates it with the element "Test".

Regards,

Pete.
--
---------------------------------------------------------------------
Pete Cordell
Codalogic Ltd
Read & write XML in C++, http://www.xml2cpp.com
---------------------------------------------------------------------