[RFC/WIP PATCH 2/3] merge-recursive: fix handling of submodules in modify/delete conflicts

[Elijah Newren <newren@gmail.com>]

Similar to commit c641ca670729 ("merge-recursive: handle addition of
submodule on our side of history", 2017-11-14) a submodule can be
confused for a D/F conflict for modify/delete and rename/delete
conflicts.  (To the code, it appears there is a directory in the way of
us writing our expected path to the working tree, because our path is a
submodule; i.e. the submodule is itself the directory and any directory
is assumed to be a D/F conflict that is in the way.)  So, when we are
dealing with a submodule, avoid checking the working copy for a
directory being in the way.

Signed-off-by: Elijah Newren <newren@gmail.com>
---

It might be better to first check that the submodule existed on the HEAD
side of the merge, because there could be a directory in the way of the
submodule.  But that's starting to get ugly quick, and the real fix to
make this cleaner is the rewrite of merge-recursive that does an index-only
merge first, then updates the working copy later...

 merge-recursive.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/merge-recursive.c b/merge-recursive.c
index 1446e92bea..4832234073 100644
--- a/merge-recursive.c
+++ b/merge-recursive.c
@@ -1466,7 +1466,7 @@ static int handle_change_delete(struct merge_options *o,
 	const char *update_path = path;
 	int ret = 0;
 
-	if (dir_in_way(path, !o->call_depth, 0) ||
+	if (dir_in_way(path, !o->call_depth && !S_ISGITLINK(changed_mode), 0) ||
 	    (!o->call_depth && would_lose_untracked(path))) {
 		update_path = alt_path = unique_path(o, path, change_branch);
 	}
-- 
2.18.0.550.g44d6daf40a.dirty