As long as the language remains the same, this will probably not be an issue.
But... when the user language changes, all grammars should be unloaded and the a fresh load of all grammars should be done. Now when changing from 'enx' to 'nld' for example, loaded grammars from the previous language remain loaded, including duplicate commands.
I would suggest this should be done in all cases of in loader.py on_change_callback, line 403.
I do not see immediately how all loaded grammars can be unloaded first, before the "trigger_load" function is called.
As long as the language remains the same, this will probably not be an issue.
But... when the user language changes, all grammars should be unloaded and the a fresh load of all grammars should be done. Now when changing from 'enx' to 'nld' for example, loaded grammars from the previous language remain loaded, including duplicate commands.
I would suggest this should be done in all cases of in loader.py on_change_callback, line 403.
I do not see immediately how all loaded grammars can be unloaded first, before the "trigger_load" function is called.