page getting dropped in conversion from epub

[RAH]

I have been trying to use Calibre to convert an epub file which I created (using Sigil) to a mobi file (I want to make it available for Kindle users). The results are useable with one big problem - ONE particular page keeps getting dropped.

I tried converting the epub to PDF, just to see if the same problem would occur and indeed it did - this one particular page gets left out of the conversion output. I have examined the epub code using Sigil, and the offending page looks the same, code-wise, to 2 similar pages that come before it. And I see nothing wrong with the page that follows the one that gets dropped.

I should mention that the epub file itself works perfectly in the Calibre viewer, in Adobe Digital Editions, with the Firefox plugin, and also on the Nook - i.e. the page that is getting dropped on these conversions works fine in the epub file on all viewers I have tried it on.

Does anyone have any suggestions on how I can fix this problem? I'm not sure if there is something wrong with the epub file (my first inclination) or if there is something wrong with Calibre.

[RAH]

I haven't received any answers to this, so I guess no one has any ideas. Is there a way for me to contact the developer? The more I look at this, the more I think that this is a bug in Calibre, and not a fault in the epub file.

[Starson17]

Originally Posted by RAH

I haven't received any answers to this, so I guess no one has any ideas.

You didn't give enough information. Post a copy of the book or enough of the book to show the problem.

Is there a way for me to contact the developer? The more I look at this, the more I think that this is a bug in Calibre, and not a fault in the epub file.

Bugs are reported here, but he'll need something to work with beyond "a page disappears." A thorough report here would probably have received his comments already, as he seems to read all posts.

[RAH]

OK, thanks for the advice. I'll see if I can create a smaller version of the file which illustrates the problem. If not, I'll try posting the full one. I didn't want to do this until I had contact with the developer first, but I see that this is not the way it is done.

[chaley]

Originally Posted by RAH

OK, thanks for the advice. I'll see if I can create a smaller version of the file which illustrates the problem. If not, I'll try posting the full one. I didn't want to do this until I had contact with the developer first, but I see that this is not the way it is done.

If it is copyrighted material, use the bug report link that Starson17 gave you. Mobileread rules strongly discourage putting such attachments here.

[RAH]

The material is not copyrighted, but I went the official route and created a ticket on the link I was given. Thanks for everyone's help!

[Starson17]

Originally Posted by chaley

If it is copyrighted material, use the bug report link that Starson17 gave you. Mobileread rules strongly discourage putting such attachments here.

He wrote that it was a book he'd created him/herself, so copyright shouldn't be a problem. OTOH, I suspect anyone looking at the problem would prefer the book be stripped down to just the problem area.

[RAH]

I tried to strip it down but wasn't able to in any meaningful way. It has very little text but many small jpg images in it. I could not see any easy way to remove all the images using Sigil (the page in question does not use any of the images and is pretty close to the front of the epub). Just removing one takes about 10 seconds and there are hundreds. So I submitted the entire epub (only 1.2M) plus the mobi, with an explanation of which page it is that is getting dropped.

[RAH]

I just want to add to this thread the reasons that have been found why my conversion of this epub file to other formats was not working correctly.

I opened a Calibre bug ticket about this problem (ticket #7469), using essentially the same description about the problem I reported here, and attaching the epub and mobi files.

Turns out that there was a problem in the content.opf file contained in the epub file. From what I understand, this file is the Table of Contents file. Although the TOC appeared correct when the epub file was viewed in a reader, and when it was viewed in the Sigil TOC tool, there was an incorrect line in it in the <guide> area of the content file, appearing as this:

<guide>
<reference type="cover" title="Cover" href="Text/Choose_note3.xhtml"/>
</guide>

This line was apparently generated by Sigil at some point during my editing of the epub file. It is incorrectly pointing to the page Choose_note3.xhtml as the "Cover".

I did have trouble while working on the epub file getting the cover to display correctly on my Nook, so probably some of my manipulations caused this bogus line to be placed in the TOC file. Looking on the Sigil forum, apparently the Symantics menu is what is used to manipulate this <guide> data, but while trying to correct the cover problem I only knew about the Symantics menu in relation to the cover.jpg file, which I did assign a Symantics of "Cover Image", which made it work correctly as a cover.

Although the <guide> error did not affect the display of the cover or the display of the Choose_note3 page in any viewer, it did cause Calibre to drop that Choose_note3 page from the conversion to any other format (or at least the ones I tried, mobi and pdf).

Fortunately, kovidgoyal was savvy enough to look inside the epub archive and spot the offending <guide> lines in the content.opf file. He added this info to the report thread and that was the resolution - a problem with the file, not Calibre, which is not surprising but very hard to find (at least for me).

I have fixed the problem by reading some on the Sigil forum, renaming the .epub file to .zip, extracting the content.opf file from it, editing it and removing those bad lines and inserting instead several innocuous <guide> lines like this:

<guide>
<reference type="cover" title="Cover" href="Text/Cover.xhtml"/>
<reference type="title-page" title="Title Page" href="Text/Title.xhtml" />
</guide>

After replacing this file back into the zip file and renaming it back to epub, it now converts correctly into the mobi file I was trying to create.

Having read some more in the Sigil doc, I probably could have used the Symantics menu on the Choose_note3 page to fix the <guide> error (there are a bunch of choices when you right-mouse a .xhtml file in Sigil and look in the Symantics menu), instead of doing it the brute-force way I did by manipulating the archive.

Anyway, all's swell that ends swell. I think that this <guide> area is not mentioned much with regard to these files, and since it can in fact mess you up somewhat, I thought it might help some people later to know that they should pay attention to it, and the format of the content.opf file itself. Apparently the Sigil folks have plans to allow direct editing of this file in the future, which would be a big help.

[DoctorOhh]

Originally Posted by RAH

Having read some more in the Sigil doc, I probably could have used the Symantics menu on the Choose_note3 page to fix the <guide> error (there are a bunch of choices when you right-mouse a .xhtml file in Sigil and look in the Symantics menu), instead of doing it the brute-force way I did by manipulating the archive.

Interesting, I'm glad Kovid solved this mystery.

The day you posted the bug I checked it out in Sigil and noticed that Choose_note3 had cover set under Symantics. I unchecked cover saved the book and then tried converting it to Mobi and still had the same issue you were reporting. Maybe I did something wrong, but it looks like getting into the OPF might have been the only way to fix it.

[RAH]

Originally Posted by dwanthny

The day you posted the bug I checked it out in Sigil and noticed that Choose_note3 had cover set under Symantics. I unchecked cover saved the book and then tried converting it to Mobi and still had the same issue you were reporting. Maybe I did something wrong, but it looks like getting into the OPF might have been the only way to fix it.

Thanks for letting me know this. I'm thinking that unchecking the cover in the Symantics for that page still left some kind of bad <guide> entry for it there. Seems like it must have, to cause Calibre to still drop it. Interesting. It sure will be useful if/when Sigil gives you the ability to easily view the code for the TOC file!

[jackie_w]

Originally Posted by RAH

It sure will be useful if/when Sigil gives you the ability to easily view the code for the TOC file!

Maybe I've misunderstood the problem, but couldn't you use Calibre's Tweak-epub feature to view the TOC file (toc.ncx) or any other internal epub file you want to examine/edit?

[RAH]

Up until this moment, I was not aware of Calibre's Tweak-epub feature. I just tried it and it looks very handy. I could have used that had I known about it. I think this comes from pretty much thinking of Sigil as the epub editing tool and Calibre as a utility for conversion, organizing, and ebook viewing. But yes, the tweak tool looks nice and a much faster way to go than the manual way I did it. Thanks for the suggestion!

[jackie_w]

You're welcome. Glad to help