Separator (dinkus) treatment on Kobo devices

[TrisSherliker]

The original text uses a separator, which I think is called a dinkus, to separate scenes or logical breaks within chapters. The symbol used is ■ black square (U+25A0). For example, this can be seen in Monarch 16.1.

I think this is being ignored in conversion, leading to some jarring changes especially in later chapters. I haven't investigated this in the code to work out the cause.

[domenic]

Can you give an example? There's actually code specifically to preserve these.

[domenic]

Here's a screenshot of 16.1:

Spoilers

Let me know if you find any cases where it's actually missing, and I'll reopen.

[TrisSherliker]

I can see what you mean. On my Kobo Libre Color it does not display properly in-book, though oddly does appear on the search results page. I've tried various fonts and it doesn't seem to be a specific font issue, so I assume it's a unicode support problem, which would be annoying. I see there is a smattering of low-traction Reddit posts complaining about unicode support. If so, it isn't your problem. But I suppose it might be worth baking in a solution anyway. Perhaps replacing with a different glyph for Kobo support purposes, or providing a compatibility option.

…

On Wed, 10 Dec 2025, 01:34 Domenic Denicola, ***@***.***> wrote: *domenic* left a comment (domenic/worm-scraper#78) <#78 (comment)> Here's a screenshot of 16.1: Spoilers image.png (view on web) <https://github.com/user-attachments/assets/4f3ed832-c966-44a0-adec-eccd3ab4a88b> Let me know if you find any cases where it's actually missing, and I'll reopen. — Reply to this email directly, view it on GitHub <#78 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABKONRZQ2LIH3AEWGOISOYD4A52C3AVCNFSM6AAAAACOQTAFMCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTMMZUHE4DCNZTGY> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[domenic]

Ah, I see. See #39. I was hoping that in the last few years Kobo would get its act together.

I guess a non-default option to replace all of the separators with <hr /> would be reasonable. Plumbing it through from the main index into

worm-scraper/lib/convert-worker.js

Lines 531 to 556 in a5632b6

	function cleanSceneBreaks(xml) {
	// Normalize scene breaks. <hr> would be more semantically appropriate, but loses the author's intent. This is
	// especially the case in Ward, which uses a variety of different scene breaks.
	xml = xml.replace(/<p(?:[^>]*)>■<\/p>/vg, `<p style="text-align: center;">■</p>`);
	xml = xml.replace(
	/<p style="text-align: center;"><strong>⊙<\/strong><\/p>/vg,
	`<p style="text-align: center;">⊙</p>`
	);
	xml = xml.replace(
	/<p style="text-align: center;"><em><strong>⊙<\/strong><\/em><\/p>/vg,
	`<p style="text-align: center;">⊙</p>`
	);
	xml = xml.replace(
	/<p style="text-align: center;"><strong>⊙⊙<\/strong><\/p>/vg,
	`<p style="text-align: center;">⊙</p>`
	);
	xml = xml.replace(
	/<p style="text-align: center;"><strong>⊙ *⊙ *⊙ *⊙ *⊙<\/strong><\/p>/vg,
	`<p style="text-align: center;">⊙ ⊙ ⊙ ⊙ ⊙</p>`
	);
	return xml;
	}

would be a bit of a chore, but not impossible.

It'd be better if we could find some sort of workaround though, which preserves the different separators and doesn't bifurcate the matrix of outputs much further than --chapter-titles already does. Some googling shows that it's not that Kobo has zero Unicode support, so, maybe this is fixable by e.g. bundling some free fonts or something like that. Unfortunately without a Kobo to test, I can't really iterate myself.

[TrisSherliker]

Well, at least we've identified the problem between us. If you were to replace the glyph with mininal disruption, I'd suggest a string of three asterisks separated by spaces, (perhaps centered): `* * *` - So it's just a text for text replacement. At the very least that's guaranteed to display. More generally, I'd be happy to test on Kobo if helpful. For example if you wanted to send me a minimal lorem ipsum example I can report back results.

…

On Wed, 10 Dec 2025, 08:33 Domenic Denicola, ***@***.***> wrote: *domenic* left a comment (domenic/worm-scraper#78) <#78 (comment)> Ah, I see. See #39 <#39>. I was hoping that in the last few years Kobo would get its act together. I guess a non-default option to replace all of the separators with <hr /> would be reasonable. Plumbing it through from the main index into https://github.com/domenic/worm-scraper/blob/a5632b69646f4eb926b50680fa9983c3afeef952/lib/convert-worker.js#L531-L556 would be a bit of a chore, but not impossible. It'd be better if we could find some sort of workaround though, which preserves the different separators and doesn't bifurcate the matrix of outputs much further than --chapter-titles already does. Some googling shows that it's not that Kobo has zero Unicode support, so, maybe this is fixable by e.g. bundling some free fonts or something like that. Unfortunately without a Kobo to test, I can't really iterate myself. — Reply to this email directly, view it on GitHub <#78 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABKONRZKECXB6K7S2QJAYQ34A7LFPAVCNFSM6AAAAACOQTAFMCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTMMZVHE2TENRUHA> . You are receiving this because you authored the thread.Message ID: ***@***.***>