Quoted from @dominiccooney
The presence of an EmptyCodeTable could suppress encoding the length for a given array, which must be zero at that point.
If I understand correctly, the idea is to use EmptyCodeTable for 0-length arrays, instead of a UnitCodeTable followed by a 32-bit length.
Quoted from @dominiccooney
If I understand correctly, the idea is to use
EmptyCodeTablefor 0-length arrays, instead of aUnitCodeTablefollowed by a 32-bit length.