Hello, On Tue, 23 Feb 2021 17:52:20 +1100 Chris Angelico <rosuav@gmail.com> wrote:
On Tue, Feb 23, 2021 at 5:33 PM Paul Sokolovsky <pmiscml@gmail.com> wrote:
Because that's the right analogy of what happens with a function - in the function "prolog", there's a series of assignments: "formal_paramX = actual_argX". Conceptually, those *are* in the function context (or you would not be able to reference function's formal params).
But they aren't actual lines of code. Your analogy breaks down because you're trying to put a line into a backtrace for a mythical and conceptual action that happens as part of the function call.
But that's the whole point - the externally visible behavior should correspond to the conceptual model, even if a particular implementation does something different underlyingly. So again, it's CPython's implementation detail that, for bytecode functions only, it implements: def foo(params): # foo takes care of its params ... foo(args) as: param_check_and_call(foo, __param_meta__[foo], *args) And due to this implementation detail raises exceptions from param_check_and_call(). []
The way I see it, this is perfect as an "additional information" field.
I started there, and proceeded to where no "additional information" hacks are needed, and everything fits well into the existing framework. Come along with me. [] -- Best regards, Paul mailto:pmiscml@gmail.com