From 604530e569efb3a7899f626fd6fb630630486757 Mon Sep 17 00:00:00 2001
From: Kovid Goyal <kovid@kovidgoyal.net>
Date: Sun, 8 Sep 2013 12:40:52 +0530
Subject: [PATCH] DOCX Input: Line breaks at the end of a paragraph

DOCX Input: Fix a single line break at the end of a paragraph not being
rendered as a blank line.
---
 src/calibre/ebooks/docx/cleanup.py | 14 ++++++++++++++
 1 file changed, 14 insertions(+)

diff --git a/src/calibre/ebooks/docx/cleanup.py b/src/calibre/ebooks/docx/cleanup.py
index 33f869faca..60db64b621 100644
--- a/src/calibre/ebooks/docx/cleanup.py
+++ b/src/calibre/ebooks/docx/cleanup.py
@@ -8,8 +8,12 @@ __copyright__ = '2013, Kovid Goyal <kovid at kovidgoyal.net>'
 
 import os
 
+from lxml.html.builder import P
+
 from calibre.ebooks.docx.names import XPath
 
+NBSP = '\xa0'
+
 def mergeable(previous, current):
     if previous.tail or current.tail:
         return False
@@ -161,6 +165,16 @@ def cleanup_markup(log, root, styles, dest_dir, detect_cover):
     for span in root.xpath('//span[not(@class) and not(@id)]'):
         lift(span)
 
+    # If a paragraph ends with a <br>, that <br> is not rendered in HTML, but
+    # it is in Word, so move it out
+    for br in root.xpath('//p/node()[position()=last()]/self::br'):
+        if not br.tail:
+            p = br.getparent()
+            p.remove(br)
+            gp = p.getparent()
+            blank = P(NBSP)
+            gp.insert(gp.index(p)+1, blank)
+
     if detect_cover:
         # Check if the first image in the document is possibly a cover
         img = root.xpath('//img[@src][1]')