Re: [PATCH v4 3/3] refs/reftable: reload locked stack when preparing transaction

From: Jeff King <peff@peff.net>
Date: Oct 1, 2024, 22:54 UTC
Message-ID: <20241001225425.GB2317071@coredump.intra.peff.net>
In-Reply-To: <Zvt6LxWm8gtJGw9S@pks.im>

On Tue, Oct 01, 2024 at 06:27:33AM +0200, Patrick Steinhardt wrote:

Show 19 quoted lines

> If this is causing problems for folks I'd say we can do the below change
> for now. It's of course only a stop-gap solution until I find the time
> to debug this, which should be later this week or early next week.
> 
> Patrick
> 
> diff --git a/t/t0610-reftable-basics.sh b/t/t0610-reftable-basics.sh
> index 2d951c8ceb..ad7bb39b79 100755
> --- a/t/t0610-reftable-basics.sh
> +++ b/t/t0610-reftable-basics.sh
> @@ -450,7 +450,7 @@ test_expect_success 'ref transaction: retry acquiring tables.list lock' '
>  	)
>  '
>  
> -test_expect_success 'ref transaction: many concurrent writers' '
> +test_expect_success !WINDOWS 'ref transaction: many concurrent writers' '
>  	test_when_finished "rm -rf repo" &&
>  	git init repo &&
>  	(

IMHO we can live with a flaky test for a little while. It's not like it's the only one. ;) And hopefully your digging turns up a real solution.

It also sounds from subsequent discussion that Josh's issue was on Linux, so it wouldn't help there.

-Peff