René's URL Explorer Experiment

{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"Populate the initial per-interpreter interned_strings dict with runtime-global singleton strings","articleBody":"# Feature or enhancement\r\n\r\nRe-use the runtime-global singleton strings inside the interned_strings dict to reduce duplication of\r\nsingleton strings (improving performance where such strings are used).\r\n\r\n# Pitch\r\n\r\nbpo-46430 (https://github.com/python/cpython/pull/30683) caused an interesting side effect; the code\r\n`x = 'a'; x[0] is x` no longer returned True. This in turn\r\nis because there are two different cached versions of 'a':\r\n\r\n- One that was cached when code in frozen modules was compiled\r\n  (and is stored in the interned_dict)\r\n- One that is stored as a runtime-global object that is used\r\n  during function calls (and is stored in _Py_SINGLETON(strings))\r\n\r\nHowever, some characters do not have this behaviour (for example,\r\n'g', 'u', and 'z'). I suspect it because these characters are not\r\nused in co_consts of frozen modules.\r\n\r\nThe interned_dict is per interpreter, and is initialized by\r\n`init_interned_dict(PyInterpreterState *)`. Currently, it is\r\ninitialized to an empty dict, which allows code in frozen modules\r\nto use their (different and per interpreter) singleton strings\r\ninstead of the runtime-global one.\r\n\r\nUsing the synthetic test case:\r\n\r\n```python\r\ndef test():\r\n    total = 0\r\n    for ch in 'abc':\r\n        if ch in {'a', 'c'}:\r\n            total += 1\r\n    return total\r\n\r\ntimeit.timeit(`test()`, globals=globals())\r\n```\r\n\r\nI get a ~5.43% improvement when the interned_strings dict reuses the runtime-global singleton strings.\r\n# Previous discussion\r\n\r\n\u003c!--\r\n  New features to Python should first be discussed elsewhere before creating issues on GitHub,\r\n  for example in the \"ideas\" category (https://discuss.python.org/c/ideas/6) of discuss.python.org,\r\n  or the python-ideas mailing list (https://mail.python.org/mailman3/lists/python-ideas.python.org/).\r\n  Use this space to post links to the places where you have already discussed this feature proposal:\r\n--\u003e\r\n\r\n\r\n\u003c!--\r\nYou can freely edit this text. Remove any lines you believe are unnecessary.\r\n--\u003e\r\n\n\n\u003c!-- gh-linked-prs --\u003e\n### Linked PRs\n* gh-103572\n\u003c!-- /gh-linked-prs --\u003e\n","author":{"url":"https://github.com/Christopher-Chianelli","@type":"Person","name":"Christopher-Chianelli"},"datePublished":"2023-04-16T00:56:28.000Z","interactionStatistic":{"@type":"InteractionCounter","interactionType":"https://schema.org/CommentAction","userInteractionCount":2},"url":"https://github.com/103571/cpython/issues/103571"}