From 333cb0652ef3f8ff4b2045db730650b78df92709 Mon Sep 17 00:00:00 2001
From: ldolse <ldolse@yahoo.com>
Date: Mon, 22 Nov 2010 18:35:44 +0800
Subject: [PATCH] fix for mobi files which put style tags outside of paragraph
 tags, which upsets lxml and BeautifulSoup

---
 src/calibre/ebooks/mobi/reader.py | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/src/calibre/ebooks/mobi/reader.py b/src/calibre/ebooks/mobi/reader.py
index c55b94cac5..d2bc25f8eb 100644
--- a/src/calibre/ebooks/mobi/reader.py
+++ b/src/calibre/ebooks/mobi/reader.py
@@ -476,6 +476,9 @@ class MobiReader(object):
         self.processed_html = self.processed_html.replace('> <', '>\n<')
         self.processed_html = self.processed_html.replace('<mbp: ', '<mbp:')
         self.processed_html = re.sub(r'<?xml[^>]*>', '', self.processed_html)
+        # Rearrange various style tags & paragraph tags so that lxml and BeautifulSoup don't get upset
+        self.processed_html = re.sub(r'(?i)(?P<styletags>(<(i|b|u|em|small|tt)>\s*){1,})(?P<para><p[^>]*>)', '\g<para>'+'\g<styletags>', self.processed_html)
+        self.processed_html = re.sub(r'(?i)(?P<para></p[^>]*>)\s*(?P<styletags>(</(i|b|u|em|small|tt)>\s*){1,})', '\g<styletags>'+'\g<para>', self.processed_html)
 
     def remove_random_bytes(self, html):
         return re.sub('\x14|\x15|\x19|\x1c|\x1d|\xef|\x12|\x13|\xec|\x08',