From cafe81f2e4528b596517108d501cd7a934864d15 Mon Sep 17 00:00:00 2001
From: Kovid Goyal <kovid@kovidgoyal.net>
Date: Sat, 4 Dec 2010 20:20:20 -0700
Subject: [PATCH] EPUB Output: Mangle filenames to ensure they are unique, to
 support broken EPUB Readers like Aldiko and Stanza. Currently disabled
 pending further testing.

---
 src/calibre/ebooks/epub/output.py             |   3 +
 .../ebooks/oeb/transforms/filenames.py        | 130 ++++++++++++++++++
 2 files changed, 133 insertions(+)
 create mode 100644 src/calibre/ebooks/oeb/transforms/filenames.py

diff --git a/src/calibre/ebooks/epub/output.py b/src/calibre/ebooks/epub/output.py
index 17f6eb9f46..c5d11edc2b 100644
--- a/src/calibre/ebooks/epub/output.py
+++ b/src/calibre/ebooks/epub/output.py
@@ -142,6 +142,9 @@ class EPUBOutput(OutputFormatPlugin):
     def convert(self, oeb, output_path, input_plugin, opts, log):
         self.log, self.opts, self.oeb = log, opts, oeb
 
+        #from calibre.ebooks.oeb.transforms.filenames import UniqueFilenames
+        #UniqueFilenames()(oeb, opts)
+
         self.workaround_ade_quirks()
         self.workaround_webkit_quirks()
         self.upshift_markup()
diff --git a/src/calibre/ebooks/oeb/transforms/filenames.py b/src/calibre/ebooks/oeb/transforms/filenames.py
new file mode 100644
index 0000000000..2b22474d30
--- /dev/null
+++ b/src/calibre/ebooks/oeb/transforms/filenames.py
@@ -0,0 +1,130 @@
+#!/usr/bin/env python
+# vim:fileencoding=UTF-8:ts=4:sw=4:sta:et:sts=4:ai
+
+__license__   = 'GPL v3'
+__copyright__ = '2010, Kovid Goyal <kovid@kovidgoyal.net>'
+__docformat__ = 'restructuredtext en'
+
+import posixpath
+from urlparse import urldefrag
+
+from lxml import etree
+import cssutils
+
+from calibre.ebooks.oeb.base import rewrite_links, urlnormalize
+
+class RenameFiles(object):
+
+    '''
+    Rename files and adjust all links pointing to them. Note that the spine
+    and manifest are not touched by this transform.
+    '''
+
+    def __init__(self, rename_map):
+        self.rename_map = rename_map
+
+    def __call__(self, oeb, opts):
+        self.log = oeb.logger
+        self.opts = opts
+        self.oeb = oeb
+
+        for item in oeb.manifest.items:
+            self.current_item = item
+            if etree.iselement(item.data):
+                rewrite_links(self.current_item.data, self.url_replacer)
+            elif hasattr(item.data, 'cssText'):
+                cssutils.replaceUrls(item.data, self.url_replacer)
+
+        if self.oeb.guide:
+            for ref in self.oeb.guide.values():
+                href = urlnormalize(ref.href)
+                href, frag = urldefrag(href)
+                replacement = self.rename_map.get(href, None)
+                if replacement is not None:
+                    nhref = replacement
+                    if frag:
+                        nhref += '#' + frag
+                    ref.href = nhref
+
+        if self.oeb.toc:
+            self.fix_toc_entry(self.oeb.toc)
+
+
+    def fix_toc_entry(self, toc):
+        if toc.href:
+            href = urlnormalize(toc.href)
+            href, frag = urldefrag(href)
+            replacement = self.rename_map.get(href, None)
+
+            if replacement is not None:
+                nhref = replacement
+                if frag:
+                    nhref = '#'.join((nhref, frag))
+                toc.href = nhref
+
+        for x in toc:
+            self.fix_toc_entry(x)
+
+    def url_replacer(self, orig_url):
+         url = urlnormalize(orig_url)
+         path, frag = urldefrag(url)
+         href = self.current_item.abshref(path)
+         replacement = self.rename_map.get(href, None)
+         if replacement is None:
+             return orig_url
+         replacement = self.current_item.relhref(replacement)
+         if frag:
+             replacement += '#' + frag
+         return replacement
+
+class UniqueFilenames(object):
+
+    'Ensure that every item in the manifest has a unique filename'
+
+    def __call__(self, oeb, opts):
+        self.log = oeb.logger
+        self.opts = opts
+        self.oeb = oeb
+
+        self.seen_filenames = set([])
+        self.rename_map = {}
+
+        for item in list(oeb.manifest.items):
+            fname = posixpath.basename(item.href)
+            if fname in self.seen_filenames:
+                suffix = self.unique_suffix(fname)
+                data = item.data
+                base, ext = posixpath.splitext(item.href)
+                nhref = base + suffix + ext
+                nhref = oeb.manifest.generate(href=nhref)[1]
+                nitem = oeb.manifest.add(item.id, nhref, item.media_type, data=data,
+                        fallback=item.fallback)
+                self.seen_filenames.add(posixpath.basename(nhref))
+                self.rename_map[item.href] = nhref
+                if item.spine_position is not None:
+                    oeb.spine.insert(item.spine_position, nitem, item.linear)
+                    oeb.spine.remove(item)
+                oeb.manifest.remove(item)
+            else:
+                self.seen_filenames.add(fname)
+
+        if self.rename_map:
+            self.log('Found non-unique filenames, renaming to support broken'
+                    ' EPUB readers like FBReader, Aldiko and Stanza...')
+            from pprint import pformat
+            self.log.debug(pformat(self.rename_map))
+
+            renamer = RenameFiles(self.rename_map)
+            renamer(oeb, opts)
+
+
+    def unique_suffix(self, fname):
+        base, ext = posixpath.splitext(fname)
+        c = 0
+        while True:
+            c += 1
+            suffix = '_u%d'%c
+            candidate = base + suffix + ext
+            if candidate not in self.seen_filenames:
+                return suffix
+