– HTML is cleaned and normalised to well‑formed XHTML (entity replacement, void‑tag handling) so the output passes EPUB validators and works with strict readers.
– manifest, spine, NCX and EPUB 3 navigation document rebuilt to meet current spec requirements; automatic TOC page added when needed.
– images are now always converted to PNG before being added to the EPUB.

These changes fix validation errors that caused some EPUBs to be rejected by stricter readers.

New icons for version 1.2.1