Lossless cast fuzz failure

[alexreinking]

Description

We tripped a fuzz_lossless_cast failure in #9162

It reproduces on main

Reproducing case

Seed: 15862464924509321052
lossless_cast failure
x = 0
buf_u8 = 37
buf_i8 = -80
out1 = -81
out2 = 175
Original: (int64)widening_sub(int32(int16(-16_i8 + (32_i8/int8((uint8)buf_u8(x))))), int32(uint16(223_u8*uint8(((-16_i8 + (32_i8/int8((uint8)buf_u8(x))))/-33_i8)*-33_i8))))
Lossless cast: (int16(-16_i8 + (32_i8/int8((uint8)buf_u8(x)))) - int16(223_u8*uint8(((-16_i8 + (32_i8/int8((uint8)buf_u8(x))))/-33_i8)*-33_i8)))

How did you get Halide?

Built from source

Halide version

main

Halide commit (if known)

3498eb4

Target

any

Operating system

all

Additional context

No response