From: Junio C Hamano <gitster@pobox.com>
Date: Tue, 06 Oct 2026 19:57:41 GMT
Subject: Re: [PATCH v2 2/2] packfile: fix corruption due to stale delta base cache entries
Message-ID: <xmqqse2id2kq.fsf@gitster.g>
In-Reply-To: <20261006-pks-packfile-stale-delta-base-cache-v2-2-69669a2fc6ce@pks.im>

Patrick Steinhardt <ps@pks.im> writes:

> Note that the added test reliably reproduces the above bug on my machine
> that uses NixOS at c59305bab206 (cosmic-applets: add missing runtime
> dependency (#566040), 2026-10-01) with glibc 2.44-25. But as we rely on
> specific allocation behaviour of glibc it is very likely that the test
> will not work on other platforms.

In other words, the test will not detect the bug, when the fix is
reverted, unless the glibc allocator is used?

Adding an unreliable reproducer for a bug that is already fixed may
be of dubious value.  However, even if the test is unreliable (since
other allocators might hide the bug when the fix is reverted), it
may be OK as long as it catches the bug on widely used
configurations and does not trigger false positives.  On the other
hand, the earlier suggestion to write custom low-level code to
simulate a colliding allocation address somehow smells like a
maintenance burden to me.

Thanks.