On 6/5/2026 1:14 AM, Richard Biener wrote:
Ah, finally found the bug I was thinking of. It was just 32bit hosts, not 32bit HOST_WIDE_INT that we weren't handling correctly in a couple places. So let's drop this concern.HOST_WIDE_INT is always 64bit nowadays, but if you need a 64bit entity use [u]int64_t. HOST_WIDE_INT should be reserved to the case where our internal data structures would expose those.
jeff
