Bad encodings in Chakma?

simoncozens commented 1 year ago

We noticed the following in udhr_ccp.xml:

The syllable aa + aa-matra is encoded as aa + virama + aa; Noto Sans Chakma displays this combination as aa + aa-matra (𑄃𑄳𑄃), presumably so that this precise text can be correctly displayed. But there's nothing in the Chakma encoding proposal (L2/09-187R) about this aa + virama + aa combination. We're removing the virama+aa -> aa-matra substitution from the font (see https://github.com/notofonts/chakma/issues/4), so soon the UDHR text will look funny. Should it be U+11103 U+11145 instead of U+11103 U+11133 U+11103?
There are a number of doubled vowel signs: e.g. 𑄖𑄧𑄧𑄱 is U+11116 U+11127 U+11127 U+11131. Visually this works in certain shapers, but will cause broken clusters in some shapers/fonts. This should presumably be U+11116 U+11127 U+11131.

simoncozens commented 1 year ago

Another issue: 𑄖𑄧𑄖𑄧𑄱 should be 𑄖𑄧𑄖𑄮, or the Universal Shaping Engine gets upset.

eric-muller commented 1 year ago

I am trying to reach Bivuti Chakma, who provided the translation.

eric-muller / udhr