[oXygen-user] oXygen to WordPress

Ben McGinnes
Mon Jan 11 19:47:19 CST 2016

On 11/01/2016 8:34 am, Eliot Kimber wrote:
> Please report any failure to conform to the EPUB3 spec as an issue
> on GitHub.

It looks to be mostly fine.  I got about half a dozen errors, but
three of those were produced by extra things I'd added to either the
oXygen webhelp css directory or the DITA-OT XHTML CSS directory,
probably the latter.  That can be fixed by simply removing the added
files (which I did after I appended theapprox. 250 lines of CSS
previously patched to the D4P-HTML5 transformation to commonltr.css
and commonrtl.css (and it does indeed make small capitals behave
properly with bold and underline or strike through and all the rest.

The other three errors were .html vs. .xhtml extension complaints.  In
those cases it was due to having three parts with no corresponding
.dita file, just navtitles and for the epub the XHTML is produced .
Those topic files default to having a .html extension, but can no
doubt be avoided entirely by making them actual .dita files to
transform with all the rest.

> I haven't had the bandwidth to test the output against epubcheck
> in the last 6 months or so (the last major update was for Harper-Collins,
> who funded the EPUB3 work and seem to be happy with the output being
> produced).

As long as they remain pedantic about every file it ought to be fine
from what I can see.

> Steven Calderwood (now at RSI but previously at Human Kinetics) did
> a lot of work to sort out epubcheck details but epubcheck is a
> moving target.

And it still does.  I'll add my explicit dividers for each part of my
little test book and confirm that.  Which means the only thing left
over which I tend to do manually is the metadata.  In some cases this
is because the automatic method doesn't seem quite right, but mainly
it's because I figure it's worth being fairly pedantic about and will
eventually have a large portion of it being much the same across
projects anyway.  The one bit of it that doesn't work at all by
default in oXygen and is the same in the Docbook template is the meta
element for dcterms:modified being set to the timestamp for the local
timezone, but fudged to have the Z on the end appended.

After that the only build error in oXygen, but not an error in a
subsequent epubcheck is not including a title on the d4pcover page.
Obviously that's expected.

Since writing the above I made the relevant adjustments so there would
be no topic generated HTML files and got rid of the extra CSS files
(leaving only my modified ones in org.dita.xhtml/resources/) and it
builds with no errors, passes the epubcheck with no errors or warnings
and only has the same build warning mentioned as expects.  The only
changes I make post buid are to pasts in a previously prepared block
of metadata and change the filename.  At which point it's ready for
release (unless it's a draft in which case it's ready for review).


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 630 bytes
Desc: OpenPGP digital signature
URL: <http://www.oxygenxml.com/pipermail/oxygen-user/attachments/20160112/91bcf068/attachment.sig>

More information about the oXygen-user mailing list