Quote:
Originally Posted by davidfor
@SirBeardAlot: I tried the three scrambled versions that you sent. Only one of them caused any problems on my test device. This was the whose title started with "Finna dolda ting".
Firstly, one thing that was mentioned, but I didn't explicitly ask, is what format are you using? I didn't have any problems with the books as epubs. It was when I changed to using the KoboTouchExtended driver that the problem arose. This transformed the books to kepub. And as I suspected early, triggered a problem with the ToC. But, not one I had seen before.
The three books are ePub 3. The have the ToC as both an NCX and NAV document. For ePub3, the device is using the NAV document to get the ToC.
When I looked at the NAV document, called "ncc.xhtml" in each book, there where three sections: - The basic ToC for chapters. This matched the NCX.
- A page list.
- The landmarks.
They were in that order and the landmarks were the largest section. As a test, I deleted the landmarks section, and the book was able to be processed by the device. I tried another version with the landmarks, but, without the page list and it didn't work.
Looking at the NAV document, the calibre editor showed it as 153.4KB. When I removed the landmarks, it dropped to 20.4KB. My assumption is that it was the sheer size of the text in the landmarks section that is causing the problem. There were only about 70 entries, so it had to be that. Or there was an error in them.
Beyond that, I can't say much. I know very little about how the NAV document works and even less about the landmarks. I have seen them before but not really looked at them. One thing that does look strange is that each landmark has "epub:type" of either "annotation" or "sidebar". My vague understanding is that they should point to sections in the book, such as the frontmatter, the backmatter and the first page of "real" content. As such, I would expect less entries and for them to be smaller. I don't know if this makes them invalid or not. |
Thank you for looking into the matter! Regarding the format, I am using the KoboTouchExtended driver, and therefor converting the epubs into kepub.
I am looking in the epub you mentioned, and i can see that there's something odd going on in the ncc.xhtml. The landmarks looks way too long, and also the coding of the text looks incorrect. As if the file has been converted from UTF-8 to something else, which has turned letters like Ä into ä.
I am not sure what the landmarks are used for but I've had invalid landmarks in other books converted from daisy 2.02 into epub3 (found when checking the file in Calibre), and I have therefor removed them completely without noticing any problems whatsoever with the books when reading them.
I removed the landmarks from the book and it seems to be working just fine adding the book now. Will try with the other two as well. The only odd thing I am noticing is that during the processing, where it usually hung at 6%, it just stops at 6%. The processing goes very very fast, much faster than I am used to. I would expect it to go up to 100%, even if it would be fast.
But then again, this might be normal, since everything seems to be working (adding more books at the same time made the process longer, but it ended at 65%.).
Your help is very much appreciated! I couldn't have solved this on my own.