This seems like it is an oversight in our CI -- is it possible to write a test that would fail without this code change 🤔

Nice catch

The problem may not related to ci, because I tested it on cli, and it also worked. @alamb

❯ select array_resize([1,2],5);
+------------------------------------------------------+
| array_resize(make_array(Int64(1),Int64(2)),Int64(5)) |
+------------------------------------------------------+
| [1, 2, , , ]                                         |
+------------------------------------------------------+
1 row in set. Query took 0.009 seconds.

The problem may not related to ci, because I tested it on cli, and it also worked. @alamb

Right, sorry, what I was saying is that given there is a bug you found in the code (by manual inspection) but

1. it wasn't found by tests
2. you didn't have to update any tests when you changed the code

I conclude there is a gap in our test coverage. Thus I was suggesting we find a way to write a test that would fail prior to this code change, but will pass with the change

The additional tests is in #8868 @alamb

Thanks again @Weijun-H