Epubcheck: New Error under upgrade to Validator 4.0.2

[Phaedrus Wolfe]

Manuscript templates that have worked for years have suddenly stopped working. Smashwords, Stealth, and others who have recently upgraded their epubcheck validation are all rejecting my epub files created in Calibre.

The error occurs immediately before each generation of a "Section Break". No problem with "Page Breaks".

Two general types of errors occur, the first is a "Span" error. Here is the Error Warning Notice:

Error while parsing file 'element "span" not allowed here; expected the element end-tag or element "address", "blockquote", "del", "div", "dl", "h1", "h2", "h3", "h4", "h5", "h6", "hr", "ins", "noscript", "ns:svg", "ol", "p", "pre", "script", "table" or "ul" (with xmlns:ns="http://www.w3.org/2000/svg")

Here is the Code in the epub creating the error:

<span class="calibre4"><span class="calibre5"><br class="calibre7"/>
</span></span>

Here is the Code in the HTML file generated by MS Word that causes Calibre to build this error in the epubs.

<u><span style='font-size:10.0pt;font-family:"Arial","sans-serif";color:blue'><br style=''>
</span></u>

The second type of error is an 'element "b" not allowed here', and the Warning Error Notice is:

Error while parsing file 'element "b" not allowed here; expected the element end-tag or element "address", "blockquote", "del", "div", "dl", "h1", "h2", "h3", "h4", "h5", "h6", "hr", "ins", "noscript", "ns:svg", "ol", "p", "pre", "script", "table" or "ul" (with xmlns:ns="http://www.w3.org/2000/svg")'.

Here is the Code in the epub file creating the error (Notice its identical to the code in the HTML file generated by MS Word):

<b><span style='font-size:12.0pt;font-family:"Arial","sans-serif"'><br style=''>
</span></b>

Is there something I can do in MS Word to prevent this error, or is the something Calibre can do to fix it? Deleting the offending code from the epub or html files prevents this error. But the nature of it as it changes from manuscript to manuscript makes writing a batch file an impossible solution. Does anyone have any ideas please?

[PeterT]

What version of calibre?

[JSWolf]

It's most likely the <br/> in the span that's the problem. You need to use prober styles in Word. Don't just highlight some text and click something in the toolbar. Word can make a real mess if you don't use it correctly. That mess then becomes a messy eBook that is not so easy to clean up. You should have clean code in your eBook.

One other thing, there's a plugin for Calibre's eBook editor to allow you to validate using epubcheck 4.0.2. Install that and it will help you not get rejected when online shops validate your eBook.

[Phaedrus Wolfe]

Originally Posted by PeterT

What version of calibre?

Hello Sir,

Thanks. I've tried versions:
calibre-64bit-3.3.0.msi
calibre-64bit-3.6.0.msi

Its definitely connected with the "Section Break" occurring immediately after. Its also definitely connected to the new updates in epubcheck validation now checking something it never checked before.

Any assistance you could give would be appreciated.

Thanks
Phaedrus

[Phaedrus Wolfe]

Originally Posted by JSWolf

It's most likely the <br/> in the span that's the problem. You need to use prober styles in Word. Don't just highlight some text and click something in the toolbar. Word can make a real mess if you don't use it correctly. That mess then becomes a messy eBook that is not so easy to clean up. You should have clean code in your eBook.

One other thing, there's a plugin for Calibre's eBook editor to allow you to validate using epubcheck 4.0.2. Install that and it will help you not get rejected when online shops validate your eBook.

We are quite experienced at using styles. Thanks for the notice regarding the new plugin. However, when used, it does not show any epubcheck errors in Calibre, yet the IDPF Epub validator continues to report the error.

Any more suggestions?

Phaedrus

[DNSB]

Originally Posted by Phaedrus Wolfe

We are quite experienced at using styles. Thanks for the notice regarding the new plugin. However, when used, it does not show any epubcheck errors in Calibre, yet the IDPF Epub validator continues to report the error.

Any more suggestions?

Phaedrus

Interesting that you don't see the error running epubcheck 4.0.2 locally while the remote validation shows the error. I don't use the Calibre editor but the epubcheck plugin for Sigil uses the same Java jar file as the online checker and I would suspect the Calibre version does the same. It's possible the dc:type metadata in the epub is hooped so it is not running the correct profile.

One question that comes to mind is what is wrapped around those spans?

Could you try adding a <p> or <div> at the start and </p> or </div> at the end of the error lines to see what that does for the error message?

[theducks]

Originally Posted by DNSB

Interesting that you don't see the error running epubcheck 4.0.2 locally while the remote validation shows the error. I don't use the Calibre editor but the epubcheck plugin for Sigil uses the same Java jar file as the online checker and I would suspect the Calibre version does the same. It's possible the dc:type metadata in the epub is hooped so it is not running the correct profile.

One question that comes to mind is what is wrapped around those spans?

Could you try adding a <p> or <div> at the start and </p> or </div> at the end of the error lines to see what that does for the error message?

Spans can not be Naked. They must be inside a block (level)

[kovidgoyal]

Dont convert word html, convert docx files directly. https://manual.calibre-ebook.com/faq...e-is-not-valid

[Phaedrus Wolfe]

Originally Posted by kovidgoyal

Dont convert word html, convert docx files directly. https://manual.calibre-ebook.com/faq...e-is-not-valid

Thank you very much for this suggestion. We gave it a try, and you are quite correct that this error magically disappears when converting our doc files to docx and then to epub. However, for us this adds a number of other difficulties. If I force auto-generation of a TOC, then the reverse links from each chapter back to the TOC do not work. If I change the reverse links to the top of the document, then we need to recreate these each time new info is added to the top of the manuscript template.

Therefore we have always used the following string in a manually created TOC:

//*[((name()='h1' or name()='h2' or name()='h3' or name()='h4' or name()='h5' or name()='h6' or name()='h7') and re:test(., '\s*((chapter|book|section|volume|part)\s+)|((A|B|C|D|E|F|G|H|I|J|K|L|M|N|O|P|Q|R|S|T|U|V|W|X|Y|Z|0|1|2|3|4|5|6|7|8|9|Ą|Ć|Ę|Ł|Ń|Ó|Ś|Ź|Ż|ą|ć|ę|ł|ń|ó|ś|ź|ż)(\s+|$))', 'i')) or @class = 'chapter']

This works great when we convert from HTML to epub, but from DOCX it only detects the first three levels of headers.

Is there a way to get Calibre to detect up to 6 or 7 levels of contents when converting to docx? If we could solve this, we would stop using the html format even though it gives us greater control in managing the output through various custom bat files and powershell scripts we run.

Thanks for your time.

[kovidgoyal]

Simply create a Table of COntents in the docx file, and calibre will read that instead of trying to build it from headings. See the sameple docx file https://manual.calibre-ebook.com/con...word-documents

for an example of doing this.

[Phaedrus Wolfe]

Originally Posted by kovidgoyal

Simply create a Table of COntents in the docx file, and calibre will read that instead of trying to build it from headings. See the sameple docx file https://manual.calibre-ebook.com/con...word-documents

for an example of doing this.

Thank you sir,

You have me convinced to make the change. Fiddling with the settings I was able to achieve the goal. We had been creating our own TOC's but Calibre had not been reading them below the 3rd level until now on any files except html.

That part is now working. One last question if you don't mind my asking. Is there a way to make all heading level hyperlinks show up as black instead of blue, while allowing the internal links to outside web pages to remain blue?

Thanks for your assistance.
Phaedrus

[JSWolf]

Originally Posted by Phaedrus Wolfe

We are quite experienced at using styles. Thanks for the notice regarding the new plugin. However, when used, it does not show any epubcheck errors in Calibre, yet the IDPF Epub validator continues to report the error.

Any more suggestions?

Phaedrus

Given that you are quite experienced with styles in Word, how did you end up with that sloppy code?

[DNSB]

Originally Posted by JSWolf

Given that you are quite experienced with styles in Word, how did you end up with that sloppy code?

Hmmm... very likely answer would be:

Word => HTML (likely filtered) => Calibre => epub.

[kovidgoyal]

Use a style in word for heading hyperlinks that specifies the color as black.

[JSWolf]

Originally Posted by DNSB

Hmmm... very likely answer would be:

Word => HTML (likely filtered) => Calibre => epub.

Can't blame Calibre for this one. The sloppy code was in the HTML. Calibre just did a garbage in, garbage out.