David Rowley <dgrowleyml(at)gmail(dot)com> writes:
> On Thu, 30 Oct 2025 at 12:55, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
>> Seems like it's important that the value be signed, so maybe ssize_t?
>> Or ptrdiff_t?
> The reason I left it an unsigned type was that the check is doing: if
> (chsize + ALLOC_CHUNKHDRSZ != blk_used), so we'd still catch this with
> an unsigned type, even if it wrapped due to going negative due to a
> bogus freeptr. Changing to a signed type would leave us with a few
> tests comparing signed to unsigned types.
Yeah, that's fair. I'm content with what you have.
regards, tom lane