Re: Could Git be smarter about object reuse?

From: Simon Richter <simon.richter@hogyros.de>
Date: Sep 22, 2025, 10:41 UTC
Message-ID: <f478fc6f-77ab-4d4e-a8d9-2d44622ba8dd@hogyros.de>
In-Reply-To: <pmKix6R7b3WVLrcK6ig1Lh7RhrB5G4Hm5yam_fEoC839aatB-OjJEmSJJ-weErGEnt4Mvgf5slxgu6Pm1xlGZ4mr_i4MIAAEMYy8DjJnWgk=@calamity.inc>

Hi,

On 9/22/25 7:06 PM, Sainan wrote:

> It's especially annoying because I know the server already has basically all the blobs needed and hence would only need to receive the commit and tree objects.

Git already does this. The receiver sends a list of commits it has, and the sender omits all objects (of any kind) that are reachable from any of these.

For this to work, the sender needs to be able to follow the commits from these references, so this does not work properly if the sender is operating from a shallow clone, or is missing branches, because the receiver only sends a list of branch tips (and, if the receiver is shallow, missing commits), not a full list of objects present, because that would be a lot.

    Simon