kindlegen fails to convert epub

[codrutp]

Error(core):E1003: Unknown error in class String.

Error(parsing):E3013: More number of characters are hidden using display:none than allowed limit. Limit: 10000 in file: C:\Users\Codrut\AppData\Local\Temp\mbp_7E2_A_19_F_ 6_5_3A_684_2028_1\OEBPS\toc01.xhtml line: 13

KindlePreviewer 3.27.0, Windows 1809 64-bit

epub opens fine with other readers

[Doitsu]

Originally Posted by codrutp

Error(core):E1003: Unknown error in class String.

Error(parsing):E3013: More number of characters are hidden using display:none than allowed limit. Limit: 10000 in file: C:\Users\Codrut\AppData\Local\Temp\mbp_7E2_A_19_F_ 6_5_3A_684_2028_1\OEBPS\toc01.xhtml line: 13

These error messages are often displayed, if text is hidden via display: none CSS selectors, which are sometimes used to hide page targets. Check your stylesheets for display: none entries and comment them out, until you've found the one that's causing the problems.

It also couldn't hurt to check your book with EpubCheck.

[codrutp]

It seems that EpubCheck is raising a lot of error. So, now, the question is, which software/python script/etc. to use in order to fix my epubs such that they are beautifully converted to Kindle format using kindlegen (calibre converted them with many many alignment problems).

[Doitsu]

Originally Posted by codrutp

It seems that EpubCheck is raising a lot of error. So, now, the question is, which software/python script/etc. to use in order to fix my epubs

Calibre Editor can fix some problems. (Open your book in Calibre Editor then click Run check and Try to correct all fixable errors automatically.)

[codrutp]

Is there any way to maybe import the html files in Word, arrange it nicely, then use Kindle Creator to convert from doc to mobi? I am also willing to learn html/css to fix these epubs, but I don't know where to start. It seems you need to be a guru in order to convert files from one format to another. My platform is definitely Kindle. I need to learn how to convert other (DRM free) books to Kindle.

[Doitsu]

Originally Posted by codrutp

Is there any way to maybe import the html files in Word, arrange it nicely, then use Kindle Creator to convert from doc to mobi?

BTW, if you've used Calibre to convert your books to MOBI files, Calibre might have generated MOBI7 files. If that was the case, try converting your books to AZW3 files. This might take care of some of the formatting issues.

Originally Posted by codrutp

I am also willing to learn html/css to fix these epubs, but I don't know where to start.

It helps to be familiar with HTML/CSS, but you could also convert your epub to a docx file with Calibre, edit the docx file with MS Word and then convert it back to an epub/azw3 file with Calibre.

[codrutp]

Hi Doitsu,

Many thanks for your support. I have one last question
Kindlegen fails with this error:

Error(parsing):E3013: More number of characters are hidden using display:none than allowed limit. Limit: 10000 in file: C:\Users\Codrut\9781491958858\OEBPS\ch01.xhtml line: 9

Now, line 9 in that file is this:


<style type="text/css" id="column-width"> { max-width: &lt;%= column_width %&gt;% !important; margin: 0 auto !important; }</style>

any idea what could it be?

Thanks

[Doitsu]

Originally Posted by codrutp

<style type="text/css" id="column-width"> { max-width: &lt;%= column_width %&gt;% !important; margin: 0 auto !important; }</style>

This line contains angle brackets written as entities. However, even if they were correctly formatted they wouldn't make sense.
Simply delete the whole line.

[codrutp]

Hi Doitsu,

I was able to workaround the issue by unpacking the epub (it's just a zip file) and using Visual Code I replaced all occurrences in all the files of "display:none" to "display:hidden". Zipping back the files and directories into epub opened the book exactly as before.
Now kindlegen succeeded!

So, if anyone hits the same problem can try applying the above. I don't know what's the difference between "display:none" and "display:hidden". But it seems to work.
It looks right and the pictures are not elongated as when using Calibre.
I now have only one remaining problem. Some characters appear in my kindle book. I think kindlegen used my active code page and not UTF8.

I will try KindlePreviewer which is using a newer version of kindlegen.

Thanks,
Codrut