The test file you sent me has
Code:
<p>(non breaking space as an HTML entity)</p>
So if you want to search for it use
Code:
<p>(non breaking space as an entity)</p>
in the search box. With that it found 41 matches for me. When you use Beautiofy calibre will remove all these ugly entities, replacing them with their unicode characters, which is why searching for <p>\s+</p> will now work