Re: [PATCH v3 4/6] xdiff/xdl_cleanup_records: make limits more clear

From: Junio C Hamano <gitster@pobox.com>
Date: Mar 30, 2026, 19:59 UTC
Message-ID: <xmqqtstxdr6v.fsf@gitster.g>
In-Reply-To: <CAH=ZcbAKwtq9jiv=XWi_P0ZD1hz7XEpEtMPONB9n=_EcOPPSRg@mail.gmail.com>

Ezekiel Newren <ezekielnewren@gmail.com> writes:

Show 14 quoted lines

> On Fri, Mar 27, 2026 at 5:01 PM Junio C Hamano <gitster@pobox.com> wrote:
>> Updated code, when nm is not zero, does something different.  if
>> need_min is true, mlim1 is set to -1 and presumably nm is a count or
>> length that is bounded on its lower end with 0, so it is larger than
>> mlim1 (== -1), and we always take INVESTIGATE and never KEEP.
>>
>> So the rewritten code is broken when need_min is true?
>>
>> I suspect the remainder of the patch is broken exactly the same way,
>> so the remedy would be similar?
>
> Your assessment is correct, PTRDIFF_MAX should be used instead of
> SIZE_MAX. I realized my mistake a few hours after I pushed. This will
> be fixed in the next version.

Yeah, using PTRDIFF_MAX is fine. When I reported the breakage I was hinting that everything may want to become unsigned, but since the original does use signed quantities and variables, it is far safer to stick to signed arithmetic---until a full audit says it is safe to switch to size_t of course.

Thanks.