Adds an I512 intermediate type so I256 fixed-point mul_div operations get the same phantom-overflow handling i128 already has (falling back to a wider intermediate when x * y overflows), instead of the current behavior where large I256 × I256 products can overflow with no fallback.
Adds an
I512intermediate type soI256fixed-pointmul_divoperations get the same phantom-overflow handlingi128already has (falling back to a wider intermediate whenx * yoverflows), instead of the current behavior where largeI256 × I256products can overflow with no fallback.