Unity screws up some fonts? (Testing Tamil)

So, I copied that string and pasted it into a text file. I saved that text file as UTF-16. If I do a hex dump of that file I get:

graham$ od -x tamil.txt
0000000 fffe a80b c70b b00b


Ignore the byte-order-mark, and look at the last 3 columns. These tell me that the sequence of glyphs starts with U+0x0BA8, which is the ந character in Tamil. Now, in your forum post, this glyph appears in the middle of the string. So, there must be some rules about Tamil language that changes the order in which glyphs are displayed. I don't know what these rules are, but guess they are encoded in the font's GSUB or GPOS tables. GSUB controls the way sequences of glyphs are displayed, so, the f-followed-by-i sequence in some fonts can be swapped out for a single fi ligature (U+FB01). GPOS controls the spacing of glyphs, so can move sequences together/apart, so is another way to display fi using two glyphs. Some fonts display accented characters by rendering the character, and then the accent having moved the accent backwards to appear over the character. Anyways, this is just "knowledge sharing". Unity does not process the GPOS or GSUB tables, which is why the order of your string in Unity is the order of the glyphs in the string. You may need to implement some string munging code, either offline, or in your script that takes strings, and orders the characters correctly knowing that Unity will display them in the order they are in the string. (Ashkan did this for Farsi, which has complex rules for the casing of characters. In Farsi there are different glyphs used when a letter appears at the start, end, or middle of a word. The GSUB table usually handles this.)