On Mon, 15 Oct 2018 01:55:24 +1100 Harshula via Unicode <[email protected]> wrote:
> 3) However, what you have observed is an issue with *explicit* > conjunct creation. After the segmentation is completed, the > layout/shaping engine needs to first check if there is a > corresponding lookup for the explicit conjunct, if not, then it needs > to remove the ZWJ and redo the segmentation and lookup(s). Perhaps > that is not happening in Harfbuzz. This indeed seems to be the problem with HarfBuzz and with Windows 7 Uniscribe. Curiously, they almost adopt this behaviour when touching letters are not available. (The ZWJ seems not to be completely removed - in HarfBuzz at least it can result in the al-lakuna not interacting properly with the base character.) But where is this usually useful behaviour specified? 1. There may be nothing but time and money to stop fallbacks being built into the font. For example, what prohibits the rendering of a conjunct falling back to touching letters or a missing glyph symbol? 2. One could argue that the current behaviour falls back to a <consonant, al-lakuna, consonant> display; Pali in Thai script does use sequences of <left matra, consonant, vowel-killer, consonant>. The problem is that al-lakuna also acts as a vowel modifier. 3. What stops one arguing that a conjunct is an abstract character and that to render it with a sequence using a visible al-lakuna would violate its identity? Richard.
