Korean not working on iPhone

[monkeytypegeorge]

[neezacoto]

I'll look into this, it appears the inputs work fine, but it's how they are displayed to the user on mobile devices

[neezacoto]

Was able to test it on an android device. It seems to only be an IPhone problem.

[Miodec]

Are you still having issues on iPhone?

[flookey]

Well I don't know korean, but this looks normal to me.

[neezacoto]

Well I don't know korean, but this looks normal to me.

This was the problem, for some reason on apple devices some inputs aren't being picked up. I'm not sure how the inputs work on iPhone, but you can see as I type, some inputs aren't being picked up on the visualizer on the bottom.

Upload.from.GitHub.for.iOS.MOV

[rearrangement]

Well I don't know korean, but this looks normal to me.

This was the problem, for some reason on apple devices some inputs aren't being picked up. I'm not sure how the inputs work on iPhone, but you can see as I type, some inputs aren't being picked up on the visualizer on the bottom.

Upload.from.GitHub.for.iOS.MOV

i think this is happening because the letters are sometimes typed in with multiple characters.
you could try to fix by adding a certain exception to grade every input instead of every character?

[neezacoto]

@shxrklol Not sure if I understand. Currently the way Korean is checked is by each input. I don't remember exactly how I did this but, I think it went something like this: If you have "하다" it shows it on screen, but it's actually broken up into [['ㅎ','ㅏ'],['ㄷ','ㅏ']]. It's then treated as a non-ligature with extra conditions to reflect if a whole character is typed correctly.

You can confirm this: The same system that checks each input also displays each input below typing test.

If you're curious about implementation, you can filter for my pull requests. They should be more detailed about how I implement things. My suspicion is that iPhone at the time of checking, uses some other weird system that mixes both single and full character inputs.

[Kagami]

It seems to be related to multi-letter syllable completions which can be split into syllables in a different way.

E.g. for the word "이후", after you typed ㅇ, ㅣ, ㅎ, in any text field, system will suggest the syllable 잏, but after you typed ㅜ, it will become 이후 (because there can't be single syllable with those 4 letters).

But in monkeytype only 이 is displayed, when you type ㅎ it's not shown on the screen at all. However, if you type 이 and press "Enter" on a keyboard, "이" will be accepted and you will be able to type the next syllable "후".

Not sure if that helps, just trying to find the root cause of the issue.