Re: Performance regression in "update" hooks

From: Adrian Ratiu <adrian.ratiu@collabora.com>
Date: Mar 2, 2026, 18:02 UTC
Message-ID: <87wlzu5cug.fsf@collabora.com>
In-Reply-To: <20260302175052.GA28275@coredump.intra.peff.net>

On Mon, 02 Mar 2026, Jeff King <peff@peff.net> wrote:

Show 14 quoted lines

> On Mon, Mar 02, 2026 at 03:27:22PM +0100, Patrick Steinhardt wrote:
>
>> > > I don't have the benchmark setup and it might be easier for you to
>> > > confirm?
>> 
>> All you need is a normal development infra and hyperfine. The
>> benchmarking scripts in the repo I linked should then "just work" with
>> the above invocation.
>
> Thanks, these were very cool and easy to use.
>
> Looking at the patch, my guess was that the problem is that we are now
> setting up and tearing down the sideband muxer for each hook invocation.
> This is expensive for the "update" hook, since it fires once per ref.

I independently root caused it and came up with (mostly) the same fix, so this is a very good confirmation, thanks!

Please wait for my patch because it needs fixing it 3 places, for 3 hooks which spin up/down no-op async threads. :)