Using sectionstart para contents for epub ncx/toc (including "notoc" instruction)

[mattretzer]

@lsquill, please review, @ericawarren fyi :) I looked at doing this a couple of ways in the htmltohtmlbook script:

• capturing the text and putting it in the section element as an attribute (encoding double quotes & ampersands), or
• using the text to populate the existing non-printing "header" element, essentially a meta-element Nellie added to conform to the HTMLbook spec.

I went with the latter b/c leveraging the header meant I could leave the existing O'Reilly ncx.xsl intact, and this seemed like a reasonable use of the header element.

An accompanying PR will be submitted shortly from the bookmaker_addons repo; moving the header removal from epubmaker_preprocessing to epubmaker_postprocessing.

[mattretzer]

Approved in Asana task: https://app.asana.com/0/435955636292439/423253954875564