On Fri, 30 Oct 2009, Sverre Rabbelier wrote:
> Heya,
>
> On Thu, Oct 29, 2009 at 22:56, Daniel Barkalow <barkalow@xxxxxxxxxxxx> wrote:
> >> +++ b/builtin-clone.c
> >> @@ -526,8 +526,9 @@ int cmd_clone(int argc, const char **argv, const char *prefix)
> >>
> >> refs = transport_get_remote_refs(transport);
> >> if (refs) {
> >> - mapped_refs = wanted_peer_refs(refs, refspec);
> >> - transport_fetch_refs(transport, mapped_refs);
> >> + struct ref *ref_cpy = wanted_peer_refs(refs, refspec);
> >> + mapped_refs = ref_cpy;
> >> + transport_fetch_refs(transport, ref_cpy);
> >
> > Just drop this hunk; the change doesn't actually do anything. Otherwise,
> > the patch matches what I have.
>
> Am I missing something? I thought we need this hunk since mapped_refs
> is const, and transport_fetch_refs takes a non-const argument?
>
> builtin-clone.c: In function ‘cmd_clone’:
> builtin-clone.c:531: warning: passing argument 2 of
> ‘transport_fetch_refs’ discards qualifiers from pointer target type
Ah, actually, mapped_refs should just be made non-const; unlike "refs",
it's set from wanted_peer_refs(), which returns a non-const value.
-Daniel
*This .sig left intentionally blank*