Jeff King <peff@xxxxxxxx> writes: > On Thu, Sep 17, 2020 at 10:28:53AM -0700, Junio C Hamano wrote: > >> > - if (diff_cache(revs, &ent->item->oid, ent->name, !!(option & DIFF_INDEX_CACHED))) >> > + >> > + if (option & DIFF_INDEX_MERGE_BASE) { >> > + diff_get_merge_base(revs, &oid); >> > + name = xstrdup(oid_to_hex(&oid)); >> >> Leak? > > Using oid_to_hex_r() avoids an extra copy, and the leak goes away too: > > char merge_base_hex[GIT_MAX_HEXSZ + 1]; > ... > name = oid_to_hex_r(merge_base_hex, &oid); > > -Peff Yes, I was debating myself if I should mention it or trust/assume that the contributor can easily figure it out.