I'm trying to remove all content-less tags from document. After node
removal it's parent should be reconsidered as content-less.

from bs4 import BeautifulSoup as bs

def normalize(soup):
    no_content_patt = re.compile(r'^[^\w]+$', re.U | re.S)

    while True:
        bad_blocks = soup(True, text=no_content_patt)
        if bad_blocks:
            for item in bad_blocks:
                item.decompose()
        else:
            break

# ... see below

soup = bs(s)
normalize(soup)
print soup.body.decode_contents()

What confuses me a lot, is a reaction to "\n" within tag's contens:

s = """
<p>Some content</p>       <-- keeps   (ok)
<p>...</p>                <-- removes (ok)
<p>!!<span>...</span></p> <-- removes (ok)
"""

...

s = """
<p>Some content</p>   <-- keeps (ok)
<p>...</p>            <-- removes (ok)
<p>!!<span>...</span> <-- keeps <p>!!\n</p> untouched (why?)
</p>
"""

So question is clear -- how can "\n" break my regex or soup behavior?