Closed JayPanoz closed 5 years ago
So there are two different things here:
epub:type
values are now reported as USAGE
(since #1013, following a decision made in the EPUB 3 CG call of March 14, 2019). In other words, the message you got in in v4.2.0-beta is still reported (with the same wording), but at the USAGE
level.doc-endnote
role is not reported as it is allowed on div
elements (I guess the confusion comes from the role being explicitly listed as allowed on li
elements… but note that it is also implicitly allowed wherever "Any role
" is allowed, which is the case of div
elements).So I think there are nothing wrong with EPUBCheck here (no false-negative). I'm closing this as invalid
, but feel free to keep discussing or reopen if I missed something!
Eew thanks, sorry for the inconvenience.
God I'm happy this is my last eprdctn gig, I can't even tell whether the most basic a11y stuff I'm doing is right or not anymore. 😩
Eew thanks, sorry for the inconvenience
No problem! ☺️
The epub:type
/ARIA role mapping is a bit of a red herring (in that it’s not really a mapping), so it’s prone to confusion…
So sorry to be that guy. :-(
In an EPUB I’ve been working on for a few weeks, I temporarily used a placeholder like:
Then I reported issue #1010 so I wanted to come back to it and correct that because
doc-endnote
require<li>
and, out of curiosity, I tested version 4.2.0In epubcheck 4.2.0 beta I got this warning:
In epubcheck 4.2.0 rc, it went completely unnoticed – since handling of deprecated epub:types was changed somehow?
So as far as I can tell, epubcheck may be stopping validation on the deprecated epub:types and not check the
role
that follows.Logs: