PDF Output: Fix Table of Contents being added tot he end of the PDF even
without the Add Table of Contents option being enabled. Fixes#1194836 [When convert to PDF, it always create TOC at the end](https://bugs.launchpad.net/calibre/+bug/1194836)
AZW3 Input: Add support for the page-progression-direction that is used
to indicate page turns should happen from right to left. The attribute
is passed into EPUB when converting. Fixes#1194766 [Incorrect conversion japanese MOBI](https://bugs.launchpad.net/calibre/+bug/1194766)
Conversion: Add an option to embed all fonts that are referenced in the
input document but are not already embedded. This will search your
system for the referenced font, and if found, the font will be embedded.
Only works if the output format supports font embedding (for example:
EPUB or AZW3).
When searching allow use of uppercase location names, such as AUTHOR
instead of author, automatically lowercasing them. Fixes#1192785 [search errors are replicated automatically](https://bugs.launchpad.net/calibre/+bug/1192785)
DOCX metadata: When reading covers from DOCX files use the first image
as specified in the actual markup instead of just the first image in the
container.
If a builtin plugin with the same name as a third party plugin exists,
then the builtin plagin was displayed in the get new plugins dialog as
installed (happened with the new DOCX Input plugin).
DOCX Input: If a large image that looks like a cover is present at the
start of the document, remove it and use it as the cover of the output
ebook. This can be turned off under the DOC Input section of the
conversion dialog.
Get Books: Fix error when using internal browser on some systems. I
cannot replicate this error so my fix is speculative, based on the idea
that not keeping explicit references to the python object is causing
them to be garbage collected. Fixes#1191199 [Python exception downloading books from Barnes & Noble](https://bugs.launchpad.net/calibre/+bug/1191199)
DOCX Input: Support for Table of Contents created using the Word Table
of Contents tool. calibre now first looks for such a Table of Contents
and only if one is not found does it generate a ToC from headings.
Unconditionally rendering the separator is wrong, for example, for the
TOC field it causes an extra space at the start of the first entry. In
any case, many common fields are webHidden, so they dont display anyway.
