[Plugin] ePub3-itizer - epub3 output plugin for Sigil

[2manymovies]

Looks like I'm late to the party, lol. So just remove that whole line, THEN convert from epub 2 to epub 3 and the OEBPS error will be gone. Thanks!

[KevinH]

Yes, just delete that meta line.

[DiapDealer]

I've been seeing more and more commercial EPUBs that contain empty dc metadata:

<dc:contributor/>
<dc:description/>
<dc:coverage/>
<dc:source/>

(a Sigil edit/save changes these to
<dc:contributor />
<dc:description />
<dc:coverage />
<dc:source />)

EpubCheck for EPUB2 doesn't seem to care about these empty metadata items, but it considers them errors in an EPUB3.

Col: 37: ERROR(RSC-005): Error while parsing file 'character content of element "dc:contributor" invalid; must be a string with length at least 1 (actual length was 0)'.

Do you think the plugin should remove these empty items upon conversion, or do you think it's better to let the user clean them up themselves before or after using the plugin?

[KevinH]

Yes, I can modify the plugin to check for and delete empty metadata elements in the opf easy enough.

Thanks for the head's up.

[KevinH]

There is a now a new version posted in this thread.

Changes Include:

- restrict fix that forced focus to Mac systems only to prevent problems on other platforms

- remove display:none css from parts of the nav since they are redudant given the hidden attribute (Thank you Doitsu)

- remove empty dc: metadata since for epub3 epubcheck barfs on it

- Add LGPL V2 or LGPL V3 license info to plugin.py

Have fun!

[Julie34]

Hello !
Great job !
But using Sigil 0.9.8 for W10 with epub-itiser 0.3.8 and epubcheck 0.4.02 on epub2 done with Amanuensis green Snake 24 (passing epubcheck before conversion) I get the message with the epub3 converted on all xhtml files :
line 6 (that is : <link type="text/css" href="../Styles/style.css" charset="UTF-8" rel="stylesheet"/> ) offset 290 Col: 84: ERROR(RSC-005): Erreur lors de l'analyse du fichier 'lattribut {0} nest pas autorisé ici ; lattribut {0} est attendu'.
The same withe any epub 2 of this type.
Does I anything wrong or is there a problem ?
Thanks

[Doitsu]

This error is caused by the optional charset="UTF-8" attribute, which isn't allowed in epub3 books (charset="UTF-8" isn't flagged in epub2 books).

[KevinH]

According to the Mozilla Develloper Network docs on xhtml, the "charset" is an obsolete attribute in a link tag and should NOT be used. So use search and replace to remove the charset="UTF-8" from each link tag. Does that then allow it to pass epub3 epubcheck?

If not, please copy and paste the top 10 or so lines from the xhtml file (up to the body tag) so that we can see if the nesting of tags is correct.

[Doitsu]

Originally Posted by KevinH

So use search and replace to remove the charset="UTF-8" from each link tag. Does that then allow it to pass epub3 epubcheck?

@KevinH: Removing charset="UTF-8" will allow the epub3 to pass epubcheck. IMHO, charset="UTF-8" should also be flagged by epubcheck in epub2 books.

[KevinH]

Thanks for letting me know. I will add checking for and removing the charset attribute from link tags to the next version of ePub3-itizer since it should be an easy fix.

Originally Posted by Doitsu

@KevinH: Removing charset="UTF-8" will allow the epub3 to pass epubcheck. IMHO, charset="UTF-8" should also be flagged by epubcheck in epub2 books.

[KevinH]

Just committed this change to ePub3-itizer master. It will appear in the next release.

[Julie34]

Thanks Doitsui and KevinH for your answers !
And KevinH for the change in the next release !
I tried to supress charset="UTF-8" : it does work ! And the epub is fine.
Now I have to think how to pass this plugin on several epubs...

[KevinH]

The current design of the plugin itself means it is not particularly scriptable as it is a Sigil output plugin and relies on Sigil doing its thing first to order the epub, mass rename file extensions, run the epubcheck plugin, before doing the final conversion to an epub3.

[Julie34]

I know (sigh.)

But thanks, really for this work ! Really great !

[Julie34]

A last (small) remarque. When removing charset="utf-8" manually (until the new release of the plugin), if you use "replace all" in xhtml files it also remove the part of <meta charset="utf-8"/> in the nav.xhtml file (line 6). With a new error message from epubcheck. Then, or you suppress the rest (<meta> alone) else you manage to keep it entier (wich I did).