From 741bf9577e6484c086e637f3186b3166759979e2 Mon Sep 17 00:00:00 2001
From: Kovid Goyal <kovid@kovidgoyal.net>
Date: Tue, 24 Dec 2013 16:20:32 +0530
Subject: [PATCH] ToC Editor: When auto-geenrating the table of contents from
 headings or XPath, if an element is at the top of the file, link only to the
 file instead of to the element. Fixes #1261099 [[Enhancement]ToC Editor
 senses when link is to top of
 file](https://bugs.launchpad.net/calibre/+bug/1261099)

---
 .../ebooks/conversion/plugins/epub_output.py  | 20 ++------------
 src/calibre/ebooks/oeb/polish/toc.py          | 27 ++++++++++++++++++-
 2 files changed, 28 insertions(+), 19 deletions(-)

diff --git a/src/calibre/ebooks/conversion/plugins/epub_output.py b/src/calibre/ebooks/conversion/plugins/epub_output.py
index 90af27d4d2..bb59ee71dc 100644
--- a/src/calibre/ebooks/conversion/plugins/epub_output.py
+++ b/src/calibre/ebooks/conversion/plugins/epub_output.py
@@ -484,31 +484,15 @@ class EPUBOutput(OutputFormatPlugin):
         Perform toc link transforms to alleviate slow loading.
         '''
         from calibre.ebooks.oeb.base import urldefrag, XPath
+        from calibre.ebooks.oeb.polish.toc import item_at_top
 
         def frag_is_at_top(root, frag):
-            body = XPath('//h:body')(root)
-            if body:
-                body = body[0]
-            else:
-                return False
-            tree = body.getroottree()
             elem = XPath('//*[@id="%s" or @name="%s"]'%(frag, frag))(root)
             if elem:
                 elem = elem[0]
             else:
                 return False
-            path = tree.getpath(elem)
-            for el in body.iterdescendants():
-                epath = tree.getpath(el)
-                if epath == path:
-                    break
-                if el.text and el.text.strip():
-                    return False
-                if not path.startswith(epath):
-                    # Only check tail of non-parent elements
-                    if el.tail and el.tail.strip():
-                        return False
-            return True
+            return item_at_top(elem)
 
         def simplify_toc_entry(toc):
             if toc.href:
diff --git a/src/calibre/ebooks/oeb/polish/toc.py b/src/calibre/ebooks/oeb/polish/toc.py
index 7f119b2556..cb63a82afc 100644
--- a/src/calibre/ebooks/oeb/polish/toc.py
+++ b/src/calibre/ebooks/oeb/polish/toc.py
@@ -211,6 +211,28 @@ def elem_to_toc_text(elem):
         text = _('(Untitled)')
     return text
 
+def item_at_top(elem):
+    try:
+        body = XPath('//h:body')(elem.getroottree().getroot())[0]
+    except (TypeError, IndexError, KeyError, AttributeError):
+        return False
+    tree = body.getroottree()
+    path = tree.getpath(elem)
+    for el in body.iterdescendants(etree.Element):
+        epath = tree.getpath(el)
+        if epath == path:
+            break
+        try:
+            if el.tag.endswith('}img') or (el.text and el.text.strip()):
+                return False
+        except:
+            return False
+        if not path.startswith(epath):
+            # Only check tail of non-parent elements
+            if el.tail and el.tail.strip():
+                return False
+    return True
+
 def from_xpaths(container, xpaths):
     tocroot = TOC()
     xpaths = [XPath(xp) for xp in xpaths]
@@ -249,7 +271,10 @@ def from_xpaths(container, xpaths):
                 plvl -= 1
                 parent = level_prev[plvl]
             lvl = plvl + 1
-            dirtied, elem_id = ensure_id(item)
+            if item_at_top(item):
+                dirtied, elem_id = False, None
+            else:
+                dirtied, elem_id = ensure_id(item)
             text = elem_to_toc_text(item)
             item_dirtied = dirtied or item_dirtied
             toc = parent.add(text, name, elem_id)