Bug 39795

When an OpenDocument Text (file) is exported to XHTML, the exported code does not contain lang attributes that identify the document's default language or the language changes inside the document.

Steps to reproduce the issue:
1. Create a new Writer document and insert some text in English.
2. Add a paragraph in French (e.g. copy something from fr.wikipedia.org).
3. Go to File > Export > and choose XHTML.
4. Inspect the exported XHTML file in a source code editor and search for 'lang="'.

What the XHTML *should* have is:
1. lang="en" (possibly lang="en-US" or lang="en-GB", depending on the language specified for the Writer document) on the HTML element;
2. lang="..." on elements where the language changes compared to the immediate context (i.e. nearest ancestor).

Notes:
* xml:lang is also in use, but is not supported by screen readers or software for dyslexics; screen readers are used by blind users to convert content to synthetic speech and/or Braille, and correct language identification is essential for both synthetic speech and Braille.
* Using Dublin Core metadata (e.g. <meta name="DCTERMS.language" content="en-US"...) specifies the expected audience language, but not the text processing language.

Background:
* <http://www.w3.org/International/tutorials/language-decl/#Slide0140>: "Declaring the text-processing language" (in W3C tutorial);
* WCAG 2.0 technique H57: Using language attributes on the html element: <http://www.w3.org/TR/2010/NOTE-WCAG20-TECHS-20101014/H57>
* WCAG 2.0 technique H58: Using language attributes to identify changes in the human language: <http://www.w3.org/TR/2010/NOTE-WCAG20-TECHS-20101014/H58.html>