Re: [PATCH v1 6/7] mv: from in-cone to out-of-cone

[Derrick Stolee <derrickstolee@github.com>]

On 8/3/2022 7:50 AM, Shaoxuan Yuan wrote:
> On 7/20/2022 2:14 AM, Derrick Stolee wrote:
>>> -        if ((mode & SPARSE) &&
>>> -            (path_in_sparse_checkout(dst, &the_index))) {
>>> -            int dst_pos;
>>> +        if (ignore_sparse &&
>>> +            core_apply_sparse_checkout &&
>>> +            core_sparse_checkout_cone) {
>>>
>>> -            dst_pos = cache_name_pos(dst, strlen(dst));
>>> -            active_cache[dst_pos]->ce_flags &= ~CE_SKIP_WORKTREE;
>>> +            /* from out-of-cone to in-cone */
>>> +            if ((mode & SPARSE) &&
>>> +                path_in_sparse_checkout(dst, &the_index)) {
>>> +                int dst_pos = cache_name_pos(dst, strlen(dst));
>>> +                struct cache_entry *dst_ce = active_cache[dst_pos];
>>>
>>> -            if (checkout_entry(active_cache[dst_pos], &state, NULL, NULL))
>>> -                die(_("cannot checkout %s"), active_cache[dst_pos]->name);
>>> +                dst_ce->ce_flags &= ~CE_SKIP_WORKTREE;
>>> +
>>> +                if (checkout_entry(dst_ce, &state, NULL, NULL))
>>> +                    die(_("cannot checkout %s"), dst_ce->name);
>>> +                continue;
>>> +            }
>>
>> Here, it helps to ignore whitespace changes. This out to in was already
>> handled by the existing implementation.
> 
> Yes, I think it would be better to let `diff` ignore the existing
> implementation. Are you suggesting the `-w` (--ignore-all-space) option
> of `diff`? I tried this option and it does not work. But another reason
> is that there *are* some changes that are different from the original
> out-to-in implementation, so even though it looks a bit messy, I think
> it makes sense.

I'm just making a note that I looked at this not in its patch form,
but as a commit diff where I could use the '-w' option to get a nice
view. It's not possible to do that in the patch.
 
>>> +            /* from in-cone to out-of-cone */
>>> +            if ((dst_mode & SKIP_WORKTREE_DIR) &&
>>
>> This is disjoint from the other case (because of !path_in_sparse_checkout()),
>> so maybe we could short-circuit with an "else if" here? You could put your
>> comments about the in-to-out or out-to-in inside the if blocks.
> 
> I tried an else-if but it does clutter the code a bit. I think I'll
> leave it as-is. Or do you mind show me a diff of your approach? To be
> honest, this disjoint here looks logically cleaner to me.

Here's the diff I had in mind:

--- >8 ---

diff --git a/builtin/mv.c b/builtin/mv.c
index df1f69f1a7..111aafb69a 100644
--- a/builtin/mv.c
+++ b/builtin/mv.c
@@ -455,10 +455,9 @@ int cmd_mv(int argc, const char **argv, const char *prefix)
 		if (ignore_sparse &&
 		    core_apply_sparse_checkout &&
 		    core_sparse_checkout_cone) {
-
-			/* from out-of-cone to in-cone */
 			if ((mode & SPARSE) &&
 			    path_in_sparse_checkout(dst, &the_index)) {
+				/* from out-of-cone to in-cone */
 				int dst_pos = cache_name_pos(dst, strlen(dst));
 				struct cache_entry *dst_ce = active_cache[dst_pos];
 
@@ -466,13 +465,10 @@ int cmd_mv(int argc, const char **argv, const char *prefix)
 
 				if (checkout_entry(dst_ce, &state, NULL, NULL))
 					die(_("cannot checkout %s"), dst_ce->name);
-				continue;
-			}
-
-			/* from in-cone to out-of-cone */
-			if ((dst_mode & SKIP_WORKTREE_DIR) &&
-			    !(mode & SPARSE) &&
-			    !path_in_sparse_checkout(dst, &the_index)) {
+			} else if ((dst_mode & SKIP_WORKTREE_DIR) &&
+				   !(mode & SPARSE) &&
+				   !path_in_sparse_checkout(dst, &the_index)) {
+				/* from in-cone to out-of-cone */
 				int dst_pos = cache_name_pos(dst, strlen(dst));
 				struct cache_entry *dst_ce = active_cache[dst_pos];
 				char *src_dir = dirname(xstrdup(src));

--- >8 ---

I agree with you that the whitespace breaking the two cases is nice,
but relying on that "continue;" to keep these cases disjoint is easy
to miss and I'd rather let the code be clearer about the cases.

Thanks,
-Stolee