Arabic speech recognition returns incorrect string length

[tatiang]

Platform
Thunkable X

Issue/Request Origin
This issue is described in more detail in the forums post here and the one right below it.

Steps to Reproduce
When using speech recognition in ARABIC with partial result, the partial result value has a length one greater than what is displayed on the device.

For example, if I say "shukraan lakum" (Arabic for thank you), the device correctly decodes the speech into the Arabic script شكرا لكم but if I check the length of that decoded value, it is 9. If I check the length of the same script, شكرا لكم, in an app variable, the length is 8.

Furthermore, if I compare the two strings -- decoded from speech vs. app variable value -- character by character in a loop, it shows that the two characters in the same position in both strings are unequal even though visibly, they look identical.

Note that this does not seem to be a problem with French, for example. I changed the app variable to "Bonjour" and spoke the word "Bonjour" and the result was that the length of both strings was 7, as expected, and the character by character comparison was equal for every letter in both strings.

Actual Behavior
See above.

[tatiang]

This might be relevant:

https://stackoverflow.com/questions/45858510/swift-string-characters-count-returns-wrong-number-for-an-arabic-string

[tatiang]

This Swift code I ran is also indicative of an incorrect count for Extended Grapheme Clusters:

[mark-friedman]

Interesting problem, @tatiang.

[jane-d-alt]

Hey @tatiang, just checking if this behavior is still present in the platform for you?

[tatiang]

You know, I don't think I kept the demo project I shared a link to in the forum post mentioned above. Sorry! I don't speak Arabic nor do I develop apps that use it. I was just trying to help @davidsobhysabetb on the forums. You might check with him to see if the problem still exists.