Skip to content
bidi_data

bidi_data

Mojo module 🡭

bidi_data

Unicode data for bidi.mojo, generated – do not hand-edit.

Bidi classes from Unicode 15.0.0, as fixed-width records in one comptime string: six hex digits of range start, six of range end, then one letter naming the class. Sorted by start, so a lookup is a binary search over byte offsets with nothing allocated – a Mojo comptime List cannot be indexed at runtime, and a table rebuilt per call would cost more than the algorithm reading it.

Only runs whose class is not L are stored. L is the default for everything else, including unassigned codepoints, which is what UAX #9 assigns them outside the ranges its own defaults name.

Regenerate with the script in this commit’s message when moving to a newer Unicode version, and re-run tests/test_bidi.mojo.

Functions