RFR: 8356652: Input field ignores custom input source characters

[Andy Goryachev]

On Thu, 15 May 2025 17:26:39 GMT, Alexander Zuev <kizune at openjdk.org> wrote:

>> Noticed a bit of a problem.
>> 
>> To reproduce, install Himyarit Musnad keyboard https://keyman.com/keyboards/himyarit_musnad
>> 
>> typing 'h' causes this character to appear: 𐩱
>> typing 'n' followed by 'h' generates garbage input rather than expected 𐩱:
>> 
>> ![Screenshot 2025-05-15 at 08 51 18](https://github.com/user-attachments/assets/5bf31611-46f9-4417-9720-f485a236c80e)
>> 
>> (I am using the monkey tester, but it will work with any application that uses a TextArea)
>
>> typing 'h' causes this character to appear: 𐩱 typing 'n' followed by 'h' generates garbage input rather than expected 𐩱:
> 
> The same happens without the fix. For some reason events for the keys not mapped on the active keyman keyboard still pass to the component - at least partially. That affects the character composition.

You are right, @azuev-java !  Looks like this is a separate bug, where the same text involving surrogate pairs is rendered differently by the TextArea:

![Screenshot 2025-05-15 at 10 48 11](https://github.com/user-attachments/assets/a3663976-2066-49aa-8773-2f5440497824)

-------------

PR Comment: https://git.openjdk.org/jfx/pull/1805#issuecomment-2884616372