I found this issue when generating an epub document using Sphinx and the Spinx extension opengraph.
The opengraph extension can generate a description based on the beginning text on the page. If this text happens to have something resembling an HTML tag (enclosed with "<" and ">"), an error can be found when validating the xhtml used when generating an EPUB document.
PR #120 was created to address this issue by cleaning up the HTML that gets inserted into the description.
python3 -m venv .env
. .env/bin/activate
pip install -r dev-requirements.txt
make html epub
Running a check on the xhtml produces this error:
$ htmlcheck page _build/epub/index.xhtml
10:39:50 - Launching validation for 1 paths
10:39:51 - .../sphinx-epub-issue/_build/epub/index.xhtml
10:39:51 - From line 11 column 1 to line 11 column 60
10:39:51 - attribute values may not contain '<'
10:39:51 - issue" />\n<meta property="og:description" content="This is a test of <tags>
10:39:51 - From line 1 column 16 to line 3 column 85
10:39:51 - Consider adding a “lang” attribute to the “html” start tag to declare the language of this document.
10:39:51 - TYPE html>\n\n<html xmlns="http://www.w3.org/1999/xhtml" xmlns:epub="http://www.idpf.org/2007/ops">\n <heAnd, after importing the generated EPUB into the MacOS Books app, this error is displayed:
“This page contains the following errors:
error on line 11 at column 60: Unescaped '<' not allowed in attributes values
Below is a rendering of the page up to the first error.”