René's URL Explorer Experiment

{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"__signature__ in the inspect module","articleBody":"# Documentation\r\n\r\nThe `__signature__` attribute is treated by inspect.signature (and inspect.Signature.from_callable) as an override which is returned instead of computing the signature through the ordinary process. `__text_signature__` serves a similar purpose but containing a string version of the signature.\r\nThat's for the implementation.\r\n\r\nNone of the two are mentioned in the signature or Signature sections of the inspect module's doc page. However, the [inspect.unwrap](https://docs.python.org/3/library/inspect.html#inspect.unwrap) function's doc entry mentions, in an example, that inspect.signature does recognize the `__signature__` attribute, without saying what it does with it.\r\nThis is the only mention of `__signature__` or `__text_signature__` in the doc, excluding changelogs, whatsnew, and the Argument clinic section which is overtly implementation-dependent.\r\nHowever, it is also mentioned in the [PEP 362](https://peps.python.org/pep-0362/#implementation) which introduced Signature, albeit in the Implementation section.\r\n\r\nIt should be clarified whether what `__signature__` does is documented or not. It's a very useful feature to be able to cache or lie about a function's signature, and I think it should be supported.\r\nIf not, then inspect.unwrap should stop mentioning it and use another example.\r\n\r\n\u003c!-- gh-linked-prs --\u003e\r\n### Linked PRs\r\n* gh-106311\r\n* gh-111145\n* gh-111146\n\u003c!-- /gh-linked-prs --\u003e\r\n","author":{"url":"https://github.com/Gouvernathor","@type":"Person","name":"Gouvernathor"},"datePublished":"2023-07-01T15:20:35.000Z","interactionStatistic":{"@type":"InteractionCounter","interactionType":"https://schema.org/CommentAction","userInteractionCount":0},"url":"https://github.com/106310/cpython/issues/106310"}