I'm making a window to retrieve some data from the net and show it on a list. Many lines of text have Unicode characters, in the format \uxxxx (where x are hexadecimal values). I changed them to &#xxxx; to made the conversion.
But then, I found the converted characters weren't right. After several test, I finally found the range of the missing characters.
The 0x0120 (G) character is correct, but 0x0121 and 0x0122 aren't right. The two other characters (s and T) are in fact 0x161 and 0x162 (this means a difference of 64).
The most annoying thing is that the DecodeString() function (used in RSS widget) show the right values.

Here's a picture of the test:


Here's the code I use to populate the list:

I found the error. Previously, I changed string.char(val) to Test_Convert(val), with Test_Convert() as:


Then, I realise the problem is that in many cases "192 + val / 64" isn't a integer value, and, for example, string.char(196.51) isn't the same as string.char(196). Changing it to string.char(math.floor(192 + val / 64)) solve this issue.