Re: [PATCH 2/5] object.c: lookup_unknown_object() accept 'r' as parameter

[Taylor Blau <me@ttaylorr.com>]

On Wed, Feb 12, 2020 at 01:11:02PM -0800, Junio C Hamano wrote:
> Taylor Blau <me@ttaylorr.com> writes:
>
> > ... same question about why assigning:
> >
> >   struct repository *r = the_repository;
> >
> > and passing 'r' everywhere is preferable to simply passing
> > 'the_repository' in directly.
> > ...
> >>  static void mark_object_for_connectivity(const struct object_id *oid)
> >>  {
> >> -	struct object *obj = lookup_unknown_object(oid);
> >> +	struct repository *r = the_repository;
> >> +	struct object *obj = lookup_unknown_object(r, oid);
> >>  	obj->flags |= HAS_OBJ;
> >>  }
>
> I do not claim that it applies to this particular function, and the
> function is too small for it to matter, but when a function is large
> enough and it always works on one single repository, it would make
> it easier to later update the function further to set up 'r'
> upfront, making it point at the_repository for now, and to use 'r'
> throughout the function.  That way, when the time comes to update
> the function to work on an arbitrary repository, the only change
> required will be to turn the local variable 'r' to an incoming
> parameter the caller can supply.

Right, but my suggestion was that this advice doesn't apply to this
particular instance since I don't expect that we'd ever passing
something other than 'the_repository'.

Specifically, I was worried that we'd get bitten by re-assigning 'r' in
the middle of the function and then end up in some odd broken state.

Maybe I'm worrying too much.


Thanks,
Taylor