Why do some epub files not show in-page foonotes?

[xversion1]

Originally Posted by Frenzie

The links to footnotes look like this:

<a href="part0025.html#mah0002108" id="mah0004586" class="calibre5"><sup class="EBsup">1</sup></a>

The footnote looks like this:

<p id="mah0002107" class="footnotes"><a href="part0015.html#mah0004586" id="mah0002108" class="calibre5"><sup class="EBsup">1</sup></a><em class="calibre7">Rwywbbh</em> huive zh lof. Bcq ozllz md Upqdn coxcvk ewyd ks Oyoun ucl gicp pqsottdyv dy Zdyfxgp 32 (Vggsbg 5).</p>

It will show up as a footnote if you move that ID attribute from the A element over to its parent P element:

<p id="mah0002108" class="footnotes"><a href="part0015.html#mah0004586" class="calibre5"><sup class="EBsup">1</sup></a><em class="calibre7">Rwywbbh</em> huive zh lof. Bcq ozllz md Upqdn coxcvk ewyd ks Oyoun ucl gicp pqsottdyv dy Zdyfxgp 32 (Vggsbg 5).</p>

Unfortunately I'm not sure if there's an easy way to automate that, and there are hundreds of footnotes.

There's probably something available in Node for example that'd allow you to write out a new file after performing a transformation like this

const footnotes = document.getElementsByClassName('footnotes');
// for each footnote
footnote.setAttribute('id', footnote.firstChild.getAttribute('id'));

Or as a regular expression, something along the lines of:

<p id="([a-z0-9]*)" class="footnotes"><a href="[a-z0-9]*\.html#[a-z0-9]*" id="([a-z0-9]*)"

And then you have to replace the first group by the second group…

The second would probably be easier to do, but either way I'm not sure if there's a user-friendly way to do it. Suggestions on nice GUI apps welcome.

I don't understand the second way, but I understand the first way like this (picture 1):

Copy red underline from <a, replace yellow underline of <p with it and delete the X red, then we have <p with blue underline. This one will show in-page footnotes.

Since I notice the red underline of <a exists in <p as well, I can just cut red underline of <p and replace yellow underline of <p with it, then we have the same final <p with blue underline (picture 2). This way we don't have to look at <a at all. Are these all right?

[Frenzie]

@poire-z Mentioned that it'll suffice to remove the ID from the P. But I'd already expanded on my regex above. Luckily it didn't take very long.

For example, here in the calibre ebook editor:

Find:

<p id="([a-z0-9]*)"( class="footnotes?"><a href="[a-z0-9]*\.html#[a-z0-9]*" id="([a-z0-9]*)")

Replace:

<p \3\2

Mode: regex
All text files

Originally Posted by xversion1

Are these all right?

I think that's correct, yes. The only problem is you don't want to be doing that 600 times.

[xversion1]

Originally Posted by Frenzie

@poire-z Mentioned that it'll suffice to remove the ID from the P. But I'd already expanded on my regex above. Luckily it didn't take very long.

For example, here in the calibre ebook editor:

Find:

<p id="([a-z0-9]*)"( class="footnotes?"><a href="[a-z0-9]*\.html#[a-z0-9]*" id="([a-z0-9]*)")

Replace:

<p \3\2

Mode: regex
All text files

I think that's correct, yes. The only problem is you don't want to be doing that 600 times.

Wonderful! Thanks!
But I don't know why this trick doesn't work on chapter titled "Mvouhvv Igvsj-Kamq" (can only find it using Ctrl+F, the table content doesn't show the same name). Ironically, it's where I'm reading.

Edit: I found the reason. The <a... of this chapter goes like this:

<a href="part0012_split_000.html#mah0003088"

They have a special character "_" so "a-z0-9" doesn't include it. I just add "_" (like this "a-z0-9_"). I know I just got lucky that it happens to be right. Could you please show me how to get all characters in the search so next time I won't miss anything?

[Frenzie]

Probably something like this:

<p id="([^"]*)"( class="footnotes?"><a href="[^"]*" id="([^"]*)")

(i.e., match everything but ")

[xversion1]

Originally Posted by Frenzie

Probably something like this:

<p id="([^"]*)"( class="footnotes?"><a href="[^"]*" id="([^"]*)")

(i.e., match everything but ")

Thanks!

[xversion1]

There's another strange problem. After I did what Frenzie said, which is this:

Find:
Code:

Replace:
Code:

Mode: regex
All text files

There were other notes that didn't affect as I said above. So I did this to the Frenzie's code (as I said):

Edit: I found the reason. The <a... of this chapter goes like this:

They have a special character "_" so "a-z0-9" doesn't include it. I just add "_" (like this "a-z0-9_"). I know I just got lucky that it happens to be right. Could you please show me how to get all characters in the search so next time I won't miss anything?

When I copy the book into Koreader via wifi, the first time I open it everything seems fine, the notes appear well. But if I go back to Folder and re-open it, from that point the book's format broken, text-align changes, table content gone, notes don't show, etc.
This is a strange behavior (first open OK, second open everything broken). I tested several times to make sure it's not something happens when the book being sent. I don't hope to fix this one. I think it's not easy as the original problem. I just think I should inform it just in case.
Here's the book: https://www.mediafire.com/file/hltg9...ange.epub/file

[Frenzie]

Thanks! That'll be fixed in https://github.com/koreader/crengine...c5d057156826a9

See https://github.com/koreader/koreader...ment-914061568

[xversion1]

Originally Posted by Frenzie

Thanks! That'll be fixed in https://github.com/koreader/crengine...c5d057156826a9

See https://github.com/koreader/koreader...ment-914061568

I'm not sure I understand what in these links but I guess the issue exists because the book is too long, and it seems to be fixed in the next update?

[Frenzie]

The cache failed to do the same thing as the regular code for that specific scenario.

[roger64]

1. As far as the footnotes are concerned, I noticed that you have 1204 "footnote" and 7 "footnotes". Is there any reason for this?

2. I also got report about backlink broken link for example for this chapter (the same for each of the 1204 + 7 footnote(s)). You should check what it's all about.

Processing part0020.html ...
"footnote" class found in part0020.html.
Chapter backlink broken [fn ref in other file not found]: mah0000116!
Chapter backlink broken [fn ref in other file not found]: mah0000117!
Chapter backlink broken [fn ref in other file not found]: mah0000119!
Chapter backlink broken [fn ref in other file not found]: mah0000121!
Chapter backlink broken [fn ref in other file not found]: mah0000122!
Chapter backlink broken [fn ref in other file not found]: mah0000123!
Chapter backlink broken [fn ref in other file not found]: mah0000154!

[Frenzie]

Originally Posted by roger64

1. As far as the footnotes are concerned, I noticed that you have 1204 "footnote" and 7 "footnotes". Is there any reason for this?
[/CODE]

That's why I wrote footnotes? in the regex. I noticed it too. footnotes is used for the first footnote (per chapter) and footnote for all the following footnotes iirc.

tl;dr It's just badly structured. Having a container around the footnotes with class footnotes would make sense, this does not.

[xversion1]

@Frenzie
Is there any way I can apply a quick fix format broken issue? Can't wait for update. Now it happens to even the unfixed file. I have to delay reading since I can't see notes anymore.

[Frenzie]

In the hamburger menu (top right) → update you can set the channel to develop and then update, or you can download the nightly build yourself from http://build.koreader.rocks/download/nightly/

[xversion1]

Originally Posted by Frenzie

In the hamburger menu (top right) → update you can set the channel to develop and then update, or you can download the nightly build yourself from http://build.koreader.rocks/download/nightly/

Done and it works. I can come back to stable version anytime I want just using the hamburger menu, can't I?

[Frenzie]

Depends a bit on how you define back. You can't downgrade, but if you put it back to stable it won't offer to upgrade until the next one's released.

Anyway, the nightly you just installed should be identical to 2021.09 for all intents and purposes.