U+105A U+103A should lose the subjoined component

[mhosken]

In the font NotoSansMyanmar-Regular.ttf and probably all Noto Myanmar fonts, the sequence U+105A U+103A does not remove the subjoined component from U+105A as expected. The rule is something like:

sub u105A' u103A by u1004;

That is to replace U+105A with U+1004 when followed by U+103A, a killer.

[dscorbett]

The subjoined component should be lost in <U+105A, U+1039 MYANMAR SIGN VIRAMA>, not in <U+105A, U+103A MYANMAR SIGN ASAT> as suggested here.

[ohbendy]

@dscorbett What's your source for this? I have also been informed by a scholar of Mon that the Mon letter Nga ၚ 105A loses its tail when combined with asat ◌် uni103A. It also loses the tail with virama, as you mention.

Conversely, in combination with vowels -u and -uu, I'm told it should keep its tail and trigger the post-base form of those vowels. This is currently not occurring.

Curiously, Noto Myanmar Serif seems to remove the tail when an above-base vowel is present, this is also incorrect.

[dscorbett]

I guess I was wrong.