Inconsistency of generated code when using different methods to declare nullable structures

[dktapps]

Describe the bug

The following diff:

diff --git a/schema/PluginListYml.json b/schema/PluginListYml.json
index 4500662..7d5fdea 100644
--- a/schema/PluginListYml.json
+++ b/schema/PluginListYml.json
@@ -14,17 +14,13 @@
                        "description": "The listed plugins will either be allowed or disallowed based on this setting"
                },
                "plugins": {
-                       "anyOf": [
-                               {
-                                       "type": "array",
-                                       "items": {
-                                               "type": "string"
-                                       }
-                               },
-                               {
-                                       "type": "null"
-                               }
+                       "type": [
+                               "array",
+                               "null"
                        ],
+                       "items": {
+                               "type": "string"
+                       },
                        "default": [],
                        "description": "List of plugins to allow or disallow"
                }

produces the following change:
pmmp/DataModels@46eab45

You can see that there's much less generated code for some reason, but what bothers me more is that the ?array native types disappeared from both getters and setters.

Expected behavior
The code should be the same in both cases, since they are semantically equivalent to the best of my understanding.

However, more importantly, the native types should not be missing after the change: the calculated types are exactly the same.

Schema
https://github.com/pmmp/DataModels/blob/46eab454f5e3c2498f3a84e8065ed9702f27a58f/schema/PluginListYml.json

https://github.com/pmmp/DataModels/blob/46eab454f5e3c2498f3a84e8065ed9702f27a58f/generate-schemas.php

Version:
0.12.7.

[wol-soft]

see #50 (comment)