Collection of test cases for text segmentation/itemization issues

[lianghai]

(Action per Meeting 8.)

Preliminary notes:

By the way, the HarfBuzz test suite incorporates the Unicode text rendering tests; HarfBuzz checks glyph positions and glyph IDs [see source code] without doing a “visual” comparison on glyph paths.

See also #41

[NeilSureshPatel]

I have attached test case info that I have been discussing in w3c/afrlreg showing script itemization issues for both Adlam and N'ko, specifically in regards to how shared Arabic punctuation falls back to a system font in browsers. At this time, I am thinking the issue related to the fact that Unicode Script Extensions table does not include Adlam and N'ko as valid script options for the shared Arabic punctuation. I have submitted a ticket to Unicode with recommended changes to Script Extensions. In the example I posted on Feb 3rd it also appears as if N'ko is not being recognized properly as word spaces are being rendered in a separate font from the N'ko text.

https://github.com/w3c/afrlreq/issues/18#issue-796943511