Re: [PATCH v2 2/4] odb/source-packed: support flags when iterating an object prefix

[Patrick Steinhardt <ps@pks.im>]

On Wed, Jun 24, 2026 at 07:02:48PM +0200, Christian Couder wrote:
> On Wed, Jun 24, 2026 at 12:37 PM Patrick Steinhardt <ps@pks.im> wrote:
> > diff --git a/odb/source-packed.c b/odb/source-packed.c
> > index 3afc4bf01f..6f31f0ff94 100644
> > --- a/odb/source-packed.c
> > +++ b/odb/source-packed.c
> > @@ -171,6 +172,20 @@ static int for_each_prefixed_object_in_midx(
> >                         const struct object_id *current = NULL;
> >                         struct object_id oid;
> >
> > +                       if (opts->flags) {
> > +                               uint32_t pack_id = nth_midxed_pack_int_id(m, i);
> > +                               struct packed_git *pack;
> > +
> > +                               if (prepare_midx_pack(m, pack_id)) {
> > +                                       pack_errors = true;
> > +                                       continue;
> > +                               }
> > +
> > +                               pack = nth_midxed_pack(m, pack_id);
> > +                               if (should_exclude_pack(pack, opts->flags))
> > +                                       continue;
> > +                       }
> > +
> >                         current = nth_midxed_object_oid(&oid, m, i);
> >
> >                         if (!match_hash(len, opts->prefix->hash, current->hash))
> 
> It looks like this is:
> 
>                         if (!match_hash(len, opts->prefix->hash, current->hash))
>                                 break;
> 
> and I wonder if the `if (opts->flags) { ... }` block would be better
> after that prefix check rather than before it.
> 
> Putting it after the prefix check would make sure we don't continue
> when the prefix doesn't match.

Hm, that's a good point indeed. Will adapt, thanks!

Patrick