Re: [PATCH] t1092: use GIT_PROGRESS_DELAY for consistent results

From: Junio C Hamano <gitster@pobox.com>
Date: May 25, 2021, 02:41 UTC
Message-ID: <xmqqa6ojilhb.fsf@gitster.g>
In-Reply-To: <e35b5217-362f-fc4f-7396-08218a9abf63@gmail.com>

Derrick Stolee <stolee@gmail.com> writes:

Show 12 quoted lines

> The test is comparing the same exact Git command just with
> different configurations. Any change to what Git writes to
> stderr should be consistent across these, unless there is
> an explicit reason why it would behave differently across
> these options (for example, saying "You are in a sparse
> checkout" in 'git status').
>
> There are no expectations that stderr is stable across
> versions of Git. These tests don't add friction to developers
> making new features or changing the error messages that appear
> over stderr. It's just that these tests should catch any
> unintended inconsistency across these modes.

If it just happens that an auto-gc gets triggered, and millions of other similar reasons in the future, will break that expectation, without running two different vintages of Git.

I agree with Jonathan that it fundamentally is flakey to expect two invocations of Git will behave exactly the same. Even repacking a repository starting from exactly the same state into a single pack may not produce byte-for-byte identical result due to thread scheduling.