Re: [External] Re: Missing Promisor Objects in Partial Repo Design Doc

From: Han Young <hanyang.tony@bytedance.com>
Date: Oct 12, 2024, 03:30 UTC
Message-ID: <CAG1j3zFFYcwYg7b9_xGRGHAOHm+qTHY=WpngqtJCrmDznhD+HA@mail.gmail.com>
In-Reply-To: <20241012020512.217383-1-jonathantanmy@google.com>

On Sat, Oct 12, 2024 at 10:05 AM Jonathan Tan <jonathantanmy@google.com> wrote:

Show 5 quoted lines

> So I think we'll need to use do_not_die_on_missing_objects. It does have
> the weakness that if the object is not supposed to be missing, we don't
> inform the user, but perhaps this is OK here because we know that all
> objects we encounter on this object walk are promisor objects, so if
> it's missing, it's OK.

And I think users would prefer the git command to succeed if possible, rather than die on the first (noncritical) error. Maybe show a warning and swallow the error?

> In addition to do_not_die_on_missing_objects, we'll also need the actual
> code that stops iteration through objects that pass our "best effort"
> promisor object check. Probably the best place is in get_revision_1()
> after the NULL check

get_revision_1() only does commit limiting though. Some callers of rev-list also do tree walking on commits, in a (corrupted) partial repo, tree could also be missing. There isn't a central place we can stop tree walking, callers using this feature would have to implement "tree walking early termination" themself.