unicode-org / icu4x / 8235832001

73%

Build Type

push

github

Committed by web-flow

Commit Message

Prepare for future normalization data for Unicode 16 (#4538)

ICU4X's data representation doesn't have a place for marking the first
character of an expansion as a starter that combines backwards.

To accommodate the Unicode 16 characters discussed in section 5 of
https://www.unicode.org/L2/L2024/24009-utc178-properties-recs.pdf I
expect the least disruptive change to be that characters that can occur
as the first character of a composition whose second character may occur
as the first character of an expansion decomposition be represented the
way non-BMP singleton decompositions are currently represented. This
should exclude them from NFC passthrough eligibility without having to
find a way to mark the second character.

Without this patch, these characters would be incorrectly be claimed as
having a singleton decomposition (to self) in
`CanonicalDecomposition::decompose()`.

There are no tests, because we don't actually have data like this, yet.
The purpose is to enable a copy of ICU4X binary that has this code to be
able to consume dynamically-loaded data from the future.

This change is low-risk in case we don't end up using the data trick
anticipated by this patch.

Coverage Stats

0 of 3 new or added lines in 1 file covered. (0.0%)

2 existing lines in 1 file now uncovered.

49623 of 65297 relevant lines covered (76.0%)

520515.09 hits per line