Skip to content

properties: Update arrays atomically and remove LongArrayBitVector

vlorentz requested to merge properties-atomic into master

It is unsound to update an array from multiple threads using non-atomic operations, even though two threads can't update the same cell.

Additionally, this changes the format of is_skipped.bits to be an actual bitvec instead of LongArrayBitVector (which would be is_skipped.bin).

Merge request reports