W3C home > Mailing lists > Public > public-epub3@w3.org > June 2022

RE: HTML incorrect specification?

From: Matt Garrish <matt.garrish@gmail.com>
Date: Tue, 21 Jun 2022 08:51:15 -0300
To: "'Jeremy Greenfield'" <jeremyjgreenfield@gmail.com>, <public-epub3@w3.org>
Message-ID: <011c01d88565$36c45fe0$a44d1fa0$@gmail.com>
It’s nearly impossible to guess what might be going on without a code sample that reproduces the error, but the first question I’d ask is if you’ve produced an EPUB 3 or EPUB 2 book? EPUB 2 doesn’t support the Legacy module elements and attributes <https://www.w3.org/TR/xhtml-modularization/abstract_modules.html#s_legacymodule> , so if that’s what you’re producing that would explain why the attribute isn’t allowed on list items.

 

I tried some variations on the value attribute in an EPUB 3 book and epubcheck correctly reported them all as valid/invalid as appropriate.

 

(And, FYI, there are some graphical user interfaces for epubcheck. There’s a list in the epubcheck repository <https://www.w3.org/publishing/epubcheck/docs/apps-and-tools/#epubcheck-apps> .)

 

Matt

 

From: Jeremy Greenfield <jeremyjgreenfield@gmail.com> 
Sent: June 19, 2022 12:03 AM
To: public-epub3@w3.org
Subject: HTML incorrect specification?

 

Hi all,
Just a minor potential bug, on behalf of a friend of mine who is an author who was asked to use your tool in preparation for online publishing.

He was having trouble using epubcheck, due to its current lack of front end and his lack of technical knowledge, so I helped him out by installing the tool myself and running his first book through it. It came up with three errors, which when I dug down, all indicated that one particular HTML file within the epub was using an attribute tag incorrectly, but as far as I can see, it's not.

Specifically, this chapter of the book contains an ordered list, with items 1 to 9. Inside item 6, there is another ordered list, with items 'a' through to 'd'. However that list is in turn broken up by a series of quotes, which means that the HTML parser has broken that second list into two; the first has one item, named 'a', and the second list (after the set of quotes) has three items, defined with the value attribute of '2', which parses correctly as 'b' through to 'd'.
The only restriction I can see is that this value attribute can only apply to ordered lists, not unordered lists, which makes sense.
However the epubcheck tool appears to think that the value attribute can't apply to any list object,
Without downloading the source code, and an environment editor, and then trying to find the exact line that creates this error, I can't confirm whether this is a bug or not. But it seems like it.

Meanwhile, I've explained to my friend what's happening and he's decided that it's easier for him to modify the list formatting in that page of the book rather than get involved in the technical aspects here.

Please let me know if someone can confirm that this was a bug, I'm curious and hopeful that this email is useful to you.

Thanks for your time and effort!
Jeremy.
Received on Tuesday, 21 June 2022 11:51:34 UTC

This archive was generated by hypermail 2.4.0 : Tuesday, 21 June 2022 11:51:35 UTC