Hello. As the implementor of a native EPUB2/EPUB3.0.1/EPUB3.1 Wysiwyg editor, I have briefly v188.8.131.52 and its export to EPUB. If the EPUB packaging export itself is quite good and shows very good stuff (for instance the possibility to split the document at page breaks or section headers), the inner export to XHTML+CSS is surprisingly far behind the regular "Export to XHTML" in LibreOffice. You'll find a report of my light tests at . I hope that helps. Ping me any time if you need more info or if I can help.
Care to test master instead?
Daily builds are here:
I've implemented a couple of improvements since the libreoffice-6-0 branch-off:
(The git history of LibreOffice core.git and libepubgen.git has more details, a number of bugfixes as well.)
Also, while your feedback is nice, it's not really actionable. The ideal bugreport has one ODT file as a reproducer and points out one specific expected behavior in the EPUB result, describing what is the actual result. Thanks!
(In reply to Miklos Vajna from comment #1)
> Care to test master instead?
Sure, will do.
> Also, while your feedback is nice, it's not really actionable. The ideal
> bugreport has one ODT file as a reproducer and points out one specific
> expected behavior in the EPUB result, describing what is the actual result.
Yeah, but I wanted to reach out with the whole thing first. Will file individual bugs. Thanks.
@Daniel : is this specific to MacOS, or can I set all of these Epub export issues to ALL OSes ?
(In reply to Alex Thurgood from comment #3)
> @Daniel : is this specific to MacOS, or can I set all of these Epub export
> issues to ALL OSes ?
Probably not specific to OS X but I have no other platform here to test...
So, I'm unsure what to do with this report as there are no sample documents by which to compare (ODT, XHTML, HTML+CSS).
I see that you have opened a set of 4 bug reports with test ODT files included, so I'm inclined to close this particular report as INSUFFICIENTDATA. If you see a valid reason why we should keep this one open, please set back to UNCONFIRMED but you would need to provide a more specific test case (and different to the ones you have provided elsewhere).