Adding a fix in soup2text for a common pathological case: <br><br> used instead

of <p /> to indicate paragraph breaks.

This changes the failed diff for /iesg/telechat/detail/354/ to show only three
differences, where two are whitespace differences and one shows a difference
between '@ietf.org. The' and '@ietf.org . The' and is an artifact of the text
extraction.  Will look at fixing that next.
 - Legacy-Id: 300

Author: levkowetz

ietf/utils/soup2text.py

@@ -66,6 +66,8 @@ def __str__(self, encoding='latin-1',
         return str
 
 def soup2text(html):
+    # some preprocessing to handle common pathological cases
+    html = re.sub("<br */?>[ \t\r\n]*(<br */?>)+", "<p/>", html)
     soup = TextSoup(html)
     return str(soup)