[3D Models] Added workaround for KiCad 7

- Fails to export VRMLs for PCBs using paths relative to the footprint.

See #417

CHANGELOG.md

@@ -48,6 +48,10 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
   - Problems when using an output and no variant specified.
 - PCB/SCH parity test:
   - Workaround for bogus net codes generated by KiCad (#410)
+- 3D Models:
+  - Problems to download KiCad 7 models (#417)
+  - Added workaround for KiCad 7 failing to export VRMLs for PCBs using paths
+    relative to the footprint. (See #417)
 
 ### Changed:
 - Some R, L and C values that were rejected are accepted now. You just get a

kibot/out_base_3d.py

@@ -88,6 +88,8 @@ def do_expand_env(fname, used_extra, extra_debug, lib_nickname):
         else:
             # We still missing the 3D model
             # Try relative to the footprint lib
+            # This was introduced in 7.0.0, but it doesn't work for all things in 7.0.1.
+            # I.e. You can't export a VRML when using this feature
             aliases = KiConf.get_fp_lib_aliases()
             lib_alias = aliases.get(lib_nickname)
             if lib_alias is not None:
@@ -96,7 +98,9 @@ def do_expand_env(fname, used_extra, extra_debug, lib_nickname):
                     logger.debug("- Using path relative to `{}` for `{}` ({})".format(lib_nickname, fname, full_name_lib))
                     full_name = full_name_lib
                     # KiCad 5 and 6 will need help
-                    force_used_extra = not GS.ki7
+                    # force_used_extra = not GS.ki7
+                    # Even KICad 7.0.1 needs help
+                    force_used_extra = True
         if force_used_extra:
             used_extra[0] = True
         return full_name