From ac8ccceef86423fbacb500fb6f5da842cf785573 Mon Sep 17 00:00:00 2001
From: James Ralston <>
Date: Sun, 21 Feb 2010 10:06:40 -0800
Subject: [PATCH] remove br from top of page in chm conversion

---
 src/calibre/ebooks/chm/input.py | 19 ++++++++++++++++++-
 1 file changed, 18 insertions(+), 1 deletion(-)

diff --git a/src/calibre/ebooks/chm/input.py b/src/calibre/ebooks/chm/input.py
index a2976c944a..784848929d 100644
--- a/src/calibre/ebooks/chm/input.py
+++ b/src/calibre/ebooks/chm/input.py
@@ -11,7 +11,7 @@ from mimetypes import guess_type as guess_mimetype
 from htmlentitydefs import name2codepoint
 from pprint import PrettyPrinter
 
-from BeautifulSoup import BeautifulSoup
+from BeautifulSoup import BeautifulSoup, NavigableString
 from lxml import html, etree
 from pychm.chm import CHMFile
 from pychm.chmlib import (
@@ -35,6 +35,17 @@ def match_string(s1, s2_already_lowered):
             return True
     return False
 
+def check_all_prev_empty(tag):
+    if tag is None:
+        return True
+    if tag.__class__ == NavigableString and not check_empty(tag):
+        return False
+    return check_all_prev_empty(tag.previousSibling)
+
+def check_empty(s, rex = re.compile(r'\S')):
+    return rex.search(s) is None
+
+
 def option_parser():
     parser = OptionParser(usage=_('%prog [options] mybook.chm'))
     parser.add_option('--output-dir', '-d', default='.', help=_('Output directory. Defaults to current directory'), dest='output')
@@ -160,6 +171,12 @@ class CHMReader(CHMFile):
                 t[-1].extract()
         # for some very odd reason each page's content appears to be in a table
         # too. and this table has sub-tables for random asides... grr.
+        
+        # remove br at top of page if present after nav bars removed
+        br = html('br')
+        if br:
+            if check_all_prev_empty(br[0].previousSibling):
+                br[0].extract()
 
         # some images seem to be broken in some chm's :/
         for img in html('img'):