Re: Git 2.26 fetches many times more objects than it should, wasting gigabytes

From: Jeff King <peff@peff.net>
Date: Apr 22, 2020, 19:18 UTC
Message-ID: <20200422191841.GA558336@coredump.intra.peff.net>
In-Reply-To: <20200422165358.GB140314@google.com>

On Wed, Apr 22, 2020 at 09:53:58AM -0700, Jonathan Nieder wrote:

> When process_acks sees an ACK, it passes it on to the negotiator.
> It wants to record that it received an ack to reset in_vain, but
> it forgets to!  The function is initialized and read but never
> written to.

I wondered if it might be something like this, too (and this might well be an independent bug), but...

Show 15 quoted lines

> So I'd expect the following to help:
> 
> diff --git i/fetch-pack.c w/fetch-pack.c
> index 1734a573b01..a1d743e1f61 100644
> --- i/fetch-pack.c
> +++ w/fetch-pack.c
> @@ -1287,6 +1287,8 @@ static int process_acks(struct fetch_negotiator *negotiator,
>  			struct object_id oid;
>  			if (!get_oid_hex(arg, &oid)) {
>  				struct commit *commit;
> +
> +				received_ack = 1;
>  				oidset_insert(common, &oid);
>  				commit = lookup_commit(the_repository, &oid);
>  				if (negotiator)

It doesn't. We never get any ACK from the server at all, because we give up on sending haves before hitting any common commit.

-Peff