Fix TOC processing cutting off Markdown content

[jdhoek]

The footnotes Python-Markdown extension appends a HTML div containing
the footnotes to the end of the generated HTML. This commit prevents the
toc class from accidentally moving that div to the TOC HTML output
instead of the main HTML output.

[jdhoek]

I'm not too well-versed in Python development, so this pull-request may be lacking in that respect. This code works, but I would like to add a test for this use case (using the footnotes Python Markdown extension) as well, how do I run the test suite? (Or, where can I find out how to do so?)

[d0ugal]

This change doesn't seem to be quite right, a bunch of tests are failing because there is now some extra HTML. https://travis-ci.org/tomchristie/mkdocs/jobs/44571618

I can't see the reason, but lots of the failures are due to an extra \n</p> being added to the end of the HTML.

[d0ugal]

Have you tested this change without enabling the footnotes extension?

[jdhoek]

Hmm, that seems to break the TOC… I'll resubmit this pull request after refining the fix.

[d0ugal]

Thanks :) Once we get the current tests passing I'll help you add a new one if we need it.

[d0ugal]

Closing this for now, happy to consider an updated version if you get time.