XPath Find and Replace

Shouldn't text editors for programmers have find-and-replace with
XPath? In the case of invalid XML, like a fragment or HTML, it should
try to wrap the text content or convert it to valid XML temporarily in
memory using a free tool[1].

This comes up for me all the time when editing HTML. It'd be so
convenient. Most of the time I want to do pattern matching it's on
markup, and regular expression that can do this have lots of problems:

1) Multiline matching isn't enabled in the tool I'm using.
2) I'm not using perl, so I can't use dynamic patterns.
3) The pattern is fragile and breaks on one of the hundreds of
assumptions I didn't take into account.

[1] Or something. I know some hack is possible, else Firefox
extensions like Firebug couldn't give me expressions to nodes on
arbitrary web pages.

- mwt