1. I didn't provide the solution, I commented on it
2. I don't have low standards for code - I suggested a different solution space (HTML::Strip or HTML::Parser) if this wasn't a one-off. Not all one-offs need to be robust.
3. I gave an example of three places where the solution was "broken" - script and style tags, tags nested in comments (illegal but common) and doctype declarations.
4. I make a suggestions for fixing those three specific issues.

Your point is taken, though - I don't say why I still don't think that the ammended solution is robust. Parsing HTML with a series of regexes is slow and difficult. style tags don't necessarily have endtags, for example: They could simply have a link to a .js file. Then, much later in the HTML document, if there was a closing script tag for another block, it would swallow and delete the enclosed valid content.

For performance, HTML::Stripper is an XS module, so it would be much, much faster than the multi-pass regex approach.