Kobo Bug thread

[ownedbycats]

I'll try cleaning out the bezels again and see if that helps.

[Aromatisse]

Product: Touch
Firmware: 4.24.15676
Plugins: none
File types: epub
Sideloaded or from Kobobooks: sideloaded
On external SD: yes
Steps to reproduce: since the last firmware update, often when I open a book it will display a blank page instead, and say that I am on page 1 of 1. Attempting to go to the next page will close the book and mark it as completed. If I re-open the book and it does load (whatever it does or not seems random) it will go back to the page it should have in the first place, and progress will go back to displaying normally.
Workaround: none that I can find, other than reopening books until things work normally.

[DNSB]

Originally Posted by Aromatisse

Product: Touch
Firmware: 4.24.15676
Plugins: none
File types: epub
Sideloaded or from Kobobooks: sideloaded
On external SD: yes
Steps to reproduce: since the last firmware update, often when I open a book it will display a blank page instead, and say that I am on page 1 of 1. Attempting to go to the next page will close the book and mark it as completed. If I re-open the book and it does load (whatever it does or not seems random) it will go back to the page it should have in the first place, and progress will go back to displaying normally.
Workaround: none that I can find, other than reopening books until things work normally.

Try checking the structure of the ebook. The only consistent trigger for that issue that I've found is epubs where the entire text is in one overly large file. If you are using calibre and have the Quality Check plugin installed, use it's Check epub structure => Check oversize html files. You can also try opening the book with calibre's editor or a Sigil user can try opening it in Sigil. If you use neither, try opening the book with a zip file utility (I like 7Zip). If you see a single or two html files and one of them is over 500K, it's a likely troublemaker. See attached image for an example (sadly this epub is a fine example of a total mess since all the files other than the container.xml are stored in the root.

A quick and dirty fix is to do a calibre epub to epub conversion.

[Aromatisse]

Originally Posted by DNSB

A quick and dirty fix is to do a calibre epub to epub conversion.

Thank you, I will try doing that. The problem appeared with multiple books of various sizes, some of which opened and loaded fine before the firmware update.

[Rev. Bob]

Originally Posted by DNSB

A quick and dirty fix is to do a calibre epub to epub conversion.

A slightly cleaner and slower fix, provided the book is arranged in chapters, is to open the ebook in the calibre editor and manually split it at a chapter heading.

If the book has a table of contents, open that and do some rough math. Using DNSB’s example, in which all the text is in one big 1.1 MB file, try to split it into four or five pieces; you want each file to be 250K or less. So, if the book has forty chapters, try splitting at the tops of chapters 11, 21, and 31 – so click on each of those entries, use the split tool (icon located under the preview pane) to make the divisions, and the ebook navigation should square itself away automatically.

Without a TOC, you can search the text for “chapter” (case insensitive) and eyeball the vertical scroll bar to find a good split point that way. (Or just split on every chapter, for that matter. Won’t hurt anything, but it’s a little more work.)

One thing to watch out for with both methods are ID attributes or page breaks near the chapter header. You want to make sure you split right after a page break (if you spot one) and/or before any ID attributes, if possible.

It’s a relatively low learning curve, and once you’ve done it a couple of times, it’s usually pretty easy.

[compurandom]

Originally Posted by Rev. Bob

A slightly cleaner and slower fix, provided the book is arranged in chapters, is to open the ebook in the calibre editor and manually split it at a chapter heading.

Frequently, the automatic conversion does it right and manual intervention isn't needed like this (although I have done it a few times).

Fixing it manually isn't too bad if the book is small and only has a handful of chapters, but for larger books, I look for other ways.

When the automatic conversion doesn't work (probably because the document isn't using proper headings for chapters or something), sometimes it's possible to spot a pattern of HTML that is used at the top of each chapter. If it's simple and unique, you can construct an xpath expression and tell calibre to split chapters there. If it's messier than that, use a global search and replace to insert something that can be easily matched with xpath and manually add/remove exceptions.

If the TOC isn't also badly formed, sometimes you can use the TOC as a guide on where to split it.

[Rev. Bob]

Originally Posted by compurandom

Fixing it manually isn't too bad if the book is small and only has a handful of chapters, but for larger books, I look for other ways.

Which is why I suggested this compromise:

Originally Posted by Rev. Bob

Using DNSB’s example, in which all the text is in one big 1.1 MB file, try to split it into four or five pieces; you want each file to be 250K or less. So, if the book has forty chapters, try splitting at the tops of chapters 11, 21, and 31 – so click on each of those entries, use the split tool (icon located under the preview pane) to make the divisions, and the ebook navigation should square itself away automatically.

That also has the advantage, once you find one chapter header, of allowing one to, as you describe:

Originally Posted by compurandom

spot a pattern of HTML that is used at the top of each chapter. If it's simple and unique, you can construct an xpath expression and tell calibre to split chapters there. If it's messier than that, use a global search and replace to insert something that can be easily matched with xpath and manually add/remove exceptions.

Only, I usually don’t mess with xpath. Regex usually does the job just fine.

[bigwoof]

Originally Posted by JSWolf

I have found a bug with KePub. The page numbers for the full book are not correct and when changing the page, the page numbers are not updating correctly. The total number of pages is not correct and thus, the page number that shows what page you are on does not update correctly. So if anyone is using the total number of pages to judge how long the eBook is, that won't be correct.

The % read in the footer seems to be working and the page numbers for the chapter in the header are working correctly.

I'm having this same problem with fanfictions updated with calibre on my libra H2O and the latest firmware.

The epub on the device is definitely updated. when I view it using other ebook-viewer directly from the Kobo mounted drive, the new chapters etc are there.

When viewing on the Kobo itself, the intro chapter shows the correct page count and number of chapters in the text but the new chapters are not accessible in the reader. It basically only allows access to the chapters it saw when it *first* opened the book. any new chapter added after that point are unaccessible.

The only solution I have found is to mark the book as unread and then reboot the Kobo. that clears whatever cache it has and it reads the full book just fine next time.

The books are all sent to the Kobo using calibre and the extended driver to convert them automatically to kepub btw. The metadata is being updated properly and collections are being created etc.

Any other solution?

Thanks!

[bigwoof]

Originally Posted by bigwoof

The only solution I have found is to mark the book as unread and then reboot the Kobo. that clears whatever cache it has and it reads the full book just fine next time.

turns out this "solution" is not very reliable. It's failing miserably on another book. I just updated to the newest 4.25 firmware as well. no difference.

I've tried updating the metadata using the calibre driver option, compressing the database, and loading the file in koreader installed via the kmon/nicklemenu all-in-one package (which is super slow for some reason). no difference yet.

[davidfor]

Originally Posted by bigwoof

I'm having this same problem with fanfictions updated with calibre on my libra H2O and the latest firmware.

Um, no you are not. The problem that @JSWolf is talking about is with the page count while reading. Sometimes, it doesn't increase as you turn the page.

The epub on the device is definitely updated. when I view it using other ebook-viewer directly from the Kobo mounted drive, the new chapters etc are there.

When viewing on the Kobo itself, the intro chapter shows the correct page count and number of chapters in the text but the new chapters are not accessible in the reader. It basically only allows access to the chapters it saw when it *first* opened the book. any new chapter added after that point are unaccessible.

The only solution I have found is to mark the book as unread and then reboot the Kobo. that clears whatever cache it has and it reads the full book just fine next time.

The books are all sent to the Kobo using calibre and the extended driver to convert them automatically to kepub btw. The metadata is being updated properly and collections are being created etc.

Unfortunately, that is working as designed. What happens when you add a book to the device is that it imports details of it into an internal database. This includes the ToC. If you replace the book, the behaviour of the device it to remove it from the internal database and import the book as a new book. This means that you lose the reading position and annotations from the book.

Because of those last two points, calibre cheats. When you resend a book, it will update the device in such a way that it will not remove the book. That will retain your reading status and annotations, but, it does not update the ToC or metadata. It means that if you make minor changes (spelling, styles, etc.) the book will load with no problems. But, if you make structural changes, such as to the ToC or the internal file structure, these will not be picked up.

You can turn this behaviour off in the driver. Uncheck the option "Do not treat replacements as new books". After that, any changed book will be treated as new and reimported.

[bigwoof]

Originally Posted by davidfor

Um, no you are not. The problem that @JSWolf is talking about is with the page count while reading. Sometimes, it doesn't increase as you turn the page.

ahh apologies!

Originally Posted by davidfor

You can turn this behaviour off in the driver. Uncheck the option "Do not treat replacements as new books". After that, any changed book will be treated as new and reimported.

Thanks! I'll try this!

[Bookpossum]

I'm not sure if my problem is a Kobo bug or not, but hope someone can advise me.

I always download books from Kobo as I read on my iPad. I got some books in November and downloaded them without any trouble. I have just bought a couple of books on special in the last few days, and can do the download. However when I hit the URL Link to open the book in ADE, I keep getting a message:

Error getting licence. Licence server communication problem: E_ACT_NOT_READY

I have tried redoing the download of the URL Link a couple of times, but the problem continues.

Any thoughts or suggestions please?

[Rev. Bob]

Originally Posted by Bookpossum

However when I hit the URL Link to open the book in ADE, I keep getting a message:

Error getting licence. Licence server communication problem: E_ACT_NOT_READY

I have tried redoing the download of the URL Link a couple of times, but the problem continues.

Any thoughts or suggestions please?

Not a Kobo problem. The URL Link/ACSM file is not the book itself, but a “ticket” that ADE redeems through one of Adobe’s servers to validate the purchase and allow you access through the copy protection. As your error message says, ADE’s having trouble talking to the server to manage that process.

The easy but frustrating answer is to wait a couple of days, redownload a fresh ACSM file from the retailer, and try again. If you’re impatient and have access to a VPN, you might try virtually relocating to a different country and see if that connects you to a different server that is working better.

I see you’re in Australia. Any natural disasters or other known infrastructure disruptions going on?

[chrisridd]

Relying on the permanent availability of some commercial servers seems risky in the long run, and you should consider stripping the DRM from your books. I believe we're allowed to suggest googling for "Apprentice Alf" nowadays? (Other search engines are available.)

[DNSB]

Originally Posted by chrisridd

I believe we're allowed to suggest googling for "Apprentice Alf" nowadays? (Other search engines are available.)

AFAIK, posting a link to Apprentice Alf’s Blog or Apprentice Harper's Github site is permitted.

Step by step instructions and detailed discussions are still to be avoided.