Brandon Williams <bmw...@google.com> writes:
> @@ -132,15 +131,14 @@ int cmd_diff_tree(int argc, const char **argv, const
> char *prefix)
> * reverse the order of the objects if the second one
> * is marked UNINTERESTING.
> */
> - nr_sha1 = opt->pending.nr;
> - switch (nr_sha1) {
> + switch (opt->pending.nr) {
Unrelated to the topic of this patch, but I notice that the comment
whose tail shows in the pre-context of this hunk is incorrect.
We expect "a..b" to expand to "^a b" but it is perfectly possible
for revision range parser to produce "b ^a", which means the same
thing. We we get the latter, i.e. the second one is marked
UNINTERESTING, we recover the original "a..b" by swapping the
order.
is probably what we wanted to say. These days, I suspect that we
may be able to do a lot better by using rev_cmdline_info thing,
which did not exist when this code was written.
