URL-encoded characters are always normalized; cannot be used for escaping

[Suggon]

Originally reported in 2024 on Discord (link). I vaguely remember getting a dev reply saying they've made a ticket and will be working on it, but can't seem to find the reply anymore.

---

I'm trying to use https://x.com/search?q=(%‍23かんたべりーでいず)(from:guardiantalesjp) as the link to Official Raw in https://mangadex.org/title/4df219da-4b5f-48ea-9fac-23641ae8a584, since it's the official publication medium.

Since the hashtag character (#) needs to be escaped when using inside URL queries, I've tried percent-encoding it in multiple ways:

• %23 -> #
• %2523 -> %23
• %25%32%33 -> %23

No matter what kind of cursed percent-encoded combination I throw at it, the backend has no problem accepting it:


But annoyingly, the frontend always normalizes that combination into a single #:


Which breaks the Twitter/X hashtag search query:

Because the browser treats the unescaped # as the URI fragment marker, and everything after it is ignored.

The intended behavior should be:

 

[Valle_reading_manga]

As you have already noted, this sounds like a frontend bug.
After checking the known issues, it either was forgotten or is incorrectly missing from the list, so after double checking the bug forum for an existing report, I'd report it over there to get added to the currently used issue tracker... (maybe a mod can move the thread for you)

---

On another note, I don't think I've ever really seen twitter search links used for official links, maybe it's precisely due to this bug. Usually folks just link to the first chapter or something...