Currently, pickle can only save some very simple data types into bytes. And the object itself contains reference to some builtin data or variable, only reference will be saved.
I am wondering if it is possible recursively persist all the data into pickle representation. Even if some data might be builtin function of variable, it will represent it in very raw (memory level). If it contains function it will persist all its attribute, and code object. Method-wrapper and all its attribute will be persist entirely. If it contains two attribute with same memory location `id(x)` is same, then they can be persist as one, and reference each other. For example, if user pickle recursively dump a numpy ndarray into a pickle file, when user pickle load this file from a system which doesn't install numpy, its method should all works properly like transpose, reshape ... When applying recursively dump, the output pickle file or bytes might be very large. All related memory data will be dumped. As the example of numpy ndarray, when doing recursively dump, large part of numpy library's content will be persist except those really lose or none coupled part. This is actually a cool feature, because pickle is to keep python object, completeness is a great advantage of pickle than json and other data persist library. Sent with ProtonMail Secure Email. _______________________________________________ Python-ideas mailing list -- python-ideas@python.org To unsubscribe send an email to python-ideas-le...@python.org https://mail.python.org/mailman3/lists/python-ideas.python.org/ Message archived at https://mail.python.org/archives/list/python-ideas@python.org/message/JGYMOCZVONB25FXXUQ67H6PRV6WCTQYE/ Code of Conduct: http://python.org/psf/codeofconduct/
