Hyphenation is not working properly

[brave171]

I've been seeing a lot of hyphens where they don't belong: They're not placed as according to the (German) hyphenation dictionary.
For instance, the word "Krankenzimmer" has a proper entry in the hyphenation dictionary as "Kran-ken-zim-mer", but is rendered as "Krankenz-immer" in text. This is also not an issue caused by the file; there are no soft hyphens.

Is there a way to fix this?

The used version is 2025.04 on a PW3.

[jonnyl]

Is your document language set to German? (Top menu, 2nd tab, Typography rules: German).

I tested your example and it worked fine (see attachment).

[brave171]

Yes, the document language is German. I can see that the hyphenation dictionary properly contains "Kran-ken-zim-mer", but somehow it is rendered falsely.

Attachments: This is how it renders for me. Also, "jüdis-chzartes" on this page is not hyphenated correctly too. As you can see, the hyphenation dictionary is not the issue, but something else.

[pazos]

While you think you answered the question "Is your document language set to German?" you missed the steps to verify the german typography rules are applied on the current document.

So please verify. If you still can reproduce on a German document where typography rules are set to german please say so and attach a sample document that showcases the issue.

[brave171]

I did verify it's set to German, I just didn't bother sending a screenshot for this since it's a trivial task. But if it's needed, I attached another screenshot.

I couldn't find a hyphenation error in a short test text I made, but as said, it's not all words that are hyphenated incorrectly. I would share the document in the screenshot, but I'm afraid MobileRead forbids this, as the author died too recently.

[pazos]

Originally Posted by brave171

I did verify it's set to German, I just didn't bother sending a screenshot for this since it's a trivial task. But if it's needed, I attached another screenshot.

I couldn't find a hyphenation error in a short test text I made, but as said, it's not all words that are hyphenated incorrectly. I would share the document in the screenshot, but I'm afraid MobileRead forbids this, as the author died too recently.

No need for screenshots, we trust your words

Unfortunately without a test case there's nothing we can do (and scramble isn't an option to debug hyphenation).

The test case doesn't need to be the full document. Just the minimum chunk of html + css required to trigger the bug.

I'm not sure. If you cannot reproduce in any other document on your collection you might want to try to reproduce disabling "embedded style" and see if that, somehow, affects the issue.

In any case feel free to open a ticket on github. Poire-z is there and not here and he's probably the only one it could help if this issue ends deemed as a bug

[brave171]

Alright, I found the issue. Even though the book metadata had the language set as German (thus setting the hyphenation rules to German), the html files actually contained the tag xml:lang="en". And because "Respect embedded language tags" was enabled by default, English hyphenation rules were applied. The properly hyphenated words just happened to have their hyphenation correctly guessed.

So, for anyone who has the same problem and finds this post: The easiest solution is just disabling [Typography Rules > Respect embedded language tags]. If you prefer having a properly formatted book or if the book contains multiple languages, you'll need to manually change the language tags in the files.

[oje]

@brave171: Thanks man! I was having the same problem and was already 'improving' Koreader's Dutch hyphenation patterns ...

[Moonbase59]

Don’t have a running KOReader here ATM, and not the book. But shouldn’t the HTML tags correctly indicate the document language, and thus make KOReader use the correct set of typo rules?

Looks like a problem in the ebook to me, possible?

Did a quick unpack and check of one of my ebooks, where it correctly shows "de-DE":

<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.1//EN" "http://www.w3.org/TR/xhtml11/DTD/xhtml11.dtd">
<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="de-DE">
<head>
  <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
  <meta http-equiv="Content-Style-Type" content="text/css" />
  <meta name="generator" content="pandoc" />
  <title>ch001.xhtml</title>
  <link rel="stylesheet" type="text/css" href="../styles/stylesheet1.css" />
</head>

[JSWolf]

Originally Posted by Moonbase59

Don’t have a running KOReader here ATM, and not the book. But shouldn’t the HTML tags correctly indicate the document language, and thus make KOReader use the correct set of typo rules?

Looks like a problem in the ebook to me, possible?

Did a quick unpack and check of one of my ebooks, where it correctly shows "de-DE":

<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.1//EN" "http://www.w3.org/TR/xhtml11/DTD/xhtml11.dtd">
<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="de-DE">
<head>
  <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
  <meta http-equiv="Content-Style-Type" content="text/css" />
  <meta name="generator" content="pandoc" />
  <title>ch001.xhtml</title>
  <link rel="stylesheet" type="text/css" href="../styles/stylesheet1.css" />
</head>

It could also be code in the CSS to turn off hyphenation.