mentalisttraceur

Umm... wow. ```python >>> import wrapt, importlib, os >>> os.environ['WRAPT_DISABLE_EXTENSIONS'] = 'x' >>> _ = importlib.reload(wrapt.wrappers) >>> wrapt.wrappers.CallableObjectProxy is wrapt._wrappers.CallableObjectProxy False >>> dict(self='works') {'self': 'works'} >>> wrapt._wrappers.CallableObjectProxy(dict)(self='works') {'self': 'works'} >>>...

...you do realize that most people use a very similar reasoning shape to dismiss the value of Correctly(tm) implementing decorators, the descriptor protocol, and so on, right? Do you not...

By the way, to me your first and second reply felt dismissively and condescendingly presumptious - as if I couldn't possibly be any righter in bringing this up than a...

In retrospect, your reaction makes sense. I probably should've been more sympathetic to it. After all, this is very relatable to me: > On first quick glance it seemed to...

Actually I want to take emphasis off `CallableObjectProxy.__call__` and refocus onto stuff like `FunctionWrapper.__call__`. `CallableObjectProxy.__call__` is so simple that I can trivially make sure my code polyfills the desired behavior...

I think if a property raises `AttributeError`, it's totally fine to treat it as if it does not exist. After all, that's how we say "this attribute does not exist"...

@GrahamDumpleton re: "does pyodide have its own specific platform designator for Python wheels" - per [PEP-425 platform tag rules](https://peps.python.org/pep-0425/#platform-tag), it ought to be `emscripten_1_0_wasm32`: ```python Welcome to the Pyodide terminal...

For the record, this is the main reason why I use Qutebrowser instead of Vieb as my primary browser. Actually the only reason that I can remember.