Entire book blank in EPUB

[sgirsberger]

After making some small text corrections to an existing reflowable EPUB3 today, I ran it through epubcheck 3.0b5 again to ensure there were no errors, and there weren’t. I have tested earlier versions of this EPUB on the same iPad Pro and have had no problems with the file.

This time, however, every page of the book was blank. Cover displayed fine. No error messages with line references were presented for troubleshooting. There was no nav toc to access, either, but the pages could be turned as if they had content.

I deleted the EPUB, rebooted the iPad Pro, reloaded the EPUB, and the result was the same.

I then loaded the same EPUB onto my iPhone. It displays and functions perfectly in iBooks (4.12)— same iOS (10.3.3).

I loaded it onto two other iPads, one with an older iOS and older version of iBooks, and one the same iOS 10.3.3. It displays and functions perfectly.

I transferred it via AirDrop and via e-mail. Didn’t make any difference.

I have to conclude that this is an issue with the iPad Pro, perhaps?

Also need to point out that the MOBI file that was converted from this EPUB displays and functions perfectly on a Kindle Fire HD.

I’m attaching it here in case anyone has any clues. It was originally a print book done in InDesign CS-something in 2004. Yes, I know there is some CSS that is unnecessary, but it’s a book of lists, and doesn’t take long to load anyway. I’m really most concerned about why it’s working everywhere BUT the iPad Pro when it did before.

Your guidance is greatly appreciated!

[jhowell]

One problem I noticed is that the font obfuscation in that EPUB is broken.

Did you perhaps change the book identifier in the OPF?

[sgirsberger]

Thanks for the quick response!

I did change the book identifier to the eISBN for this book, but I don't understand what you mean by the font obfuscation being broken. I didn't change anything to do with the obfuscation when exporting, unless removing the arbitrarily assigned book identifier resulted in that? I usually keep just the eISBN as the book identifier.

Would appreciate further explanation. When I opened this in Sigil 0.9.6, I check to see if any font obfuscation was present, and none was checked. If I should check an option, should it be Adobe's method or IDPF's method? I know that often, Adobe's is applied when exporting with fonts embedded from InDesign.

[JSWolf]

The version of epubcheck you are using is rather old. You should be using epubcheck 4.0.2.

[jhowell]

Originally Posted by sgirsberger

I did change the book identifier to the eISBN for this book, but I don't understand what you mean by the font obfuscation being broken. I didn't change anything to do with the obfuscation when exporting, unless removing the arbitrarily assigned book identifier resulted in that? I usually keep just the eISBN as the book identifier.

Would appreciate further explanation. When I opened this in Sigil 0.9.6, I check to see if any font obfuscation was present, and none was checked. If I should check an option, should it be Adobe's method or IDPF's method? I know that often, Adobe's is applied when exporting with fonts embedded from InDesign.

Font obfuscation encrypts embedded font files as a weak from of copy protection. The obfuscation algorithm uses the identifier of the book in the OPF file as the encryption key. The presence of an excryption.xml file along with the garbled contents of the fonts in your book shows that font obfuscation was applied to the book at some point. Changing the book identifier after the fact has rendered the embedded fonts unusable.

You can read more detail in the blog post Font obfuscation in EPUBs and why you shouldn't change the dc:identifier after exporting from InDesign.

I am not familiar with how Sigil handles font obfuscation. Someone else will have to comment on that.

[sgirsberger]

Originally Posted by jhowell

Font obfuscation encrypts embedded font files as a weak from of copy protection. The obfuscation algorithm uses the identifier of the book in the OPF file as the encryption key. The presence of an excryption.xml file along with the garbled contents of the fonts in your book shows that font obfuscation was applied to the book at some point. Changing the book identifier after the fact has rendered the embedded fonts unusable.

You can read more detail in the blog post Font obfuscation in EPUBs and why you shouldn't change the dc:identifier after exporting from InDesign.

I am not familiar with how Sigil handles font obfuscation. Someone else will have to comment on that.

I read your blog post and it was very helpful. However, restoring the identifier did not solve the problem, nor did re-exporting and replacing the info in the content.opf and encryption.xml files.

There was a lot of little things I had to do in HTML and CSS, once I exported from InDesign, that I would have to redo in the main XHTML and CSS files if I started from scratch again. Naturally, I prefer to avoid that, especially when it's on ONLY the iPad Pro that the book does not function. I would think that the obfuscation and encryption key issues would have rendered the book as blank on all devices if that was what was causing this?

My best option may be to re-export WITHOUT fonts embedded and edit the CSS files accordingly, then replace and edit the stylesheet references. I know how to do this; I'm just frustrated that there's not a clear solution to the problem.

Thanks for your help. No more tinkering with the identifier.

[sgirsberger]

Originally Posted by JSWolf

The version of epubcheck you are using is rather old. You should be using epubcheck 4.0.2.

Thanks, JSWolf. Downloaded and installed it. Works just fine.

[sgirsberger]

I removed the @font references and the encryption.xml and everything has worked out just fine. : ) Thank you both for your help.