ruff
b09f331a - [ty] Reuse overload prefix checks across calls (#30640)

Commit
15 hours ago
[ty] Reuse overload prefix checks across calls (#30640) ## Summary When an argument of type `Any` matches several overloads, ty checks whether the first overload, then the first two, and so on, accept every materialization of the arguments. Each step builds a union of parameter types. Calls to the same overloaded function can repeat the entire search with identical types. We now reuse the number of overloads to keep when the inferable type variables and the argument and parameter types at the distinguishing positions agree. We still compare the remaining return types at each call site and evaluate the search directly during cycle recovery. ## Performance After a full check of scikit-learn, the cache holds 4.01 MiB of keys, results and Salsa metadata. With one worker, total live Salsa storage grows 0.97% with the frozen inputs used by the CLI and falls 2.67% with mutable inputs. | Project | [Base instructions](https://github.com/astral-sh/ruff/commit/a866aff503fd777156c1ae4eae318a3795fa09ec) | [PR instructions](https://github.com/astral-sh/ruff/commit/ce4906c7203f761ff6b2b2748b3cbf1af653388e) | Change | Paired range | | --- | ---: | ---: | ---: | ---: | | [Black](https://github.com/psf/black/tree/55996d1fb8342725c8a19b6023cd9439c3c1d6f7) | 3,172,849,958.5 | 3,171,504,084 | -0.04% | -0.06% to -0.03% | | [discord.py](https://github.com/Rapptz/discord.py/tree/65232c38702be5844cf2ce865a4777eb1928b5d0) | 8,244,551,270 | 8,225,108,318.5 | -0.24% | -0.26% to -0.21% | | [Home Assistant](https://github.com/home-assistant/core/tree/c19ae285ede74cd64c3045a3a4aa43290c224f1b) | 159,255,076,114 | 159,160,797,237 | -0.06% | -0.07% to -0.05% | | [isort](https://github.com/pycqa/isort/tree/a514c71ed86c2216d0040c7d60e302ed76a4f7ab) | 1,860,603,867.5 | 1,858,126,810 | -0.13% | -0.21% to -0.06% | | [Jinja](https://github.com/pallets/jinja/tree/5ef70112a1ff19c05324ff889dd30405b1002044) (inconclusive) | 1,946,524,045.5 | 1,943,888,989.5 | -0.14% | -0.14% to -0.13% | | [pandas](https://github.com/pandas-dev/pandas/tree/a0225b65720018f4654390bb16d254daa6ee5e54) | 28,831,258,280.5 | 28,495,153,924 | -1.17% | -1.26% to -1.08% | | [pandas-stubs](https://github.com/pandas-dev/pandas-stubs/tree/8dd6d91b1c05b26e30f7ddbe3dc3d4042a92bdd6) (inconclusive) | 2,971,313,111 | 2,972,442,653 | +0.04% | -0.26% to +0.33% | | [Prefect](https://github.com/PrefectHQ/prefect/tree/a9452ac21852755311913cb27887b083b5007c03) | 4,957,712,369.5 | 4,952,626,259.5 | -0.10% | -0.11% to -0.09% | | [PyTorch](https://github.com/pytorch/pytorch/tree/ab92945cea42e47ae398aead69af0872d340ea0f) | 96,639,048,118.5 | 96,414,268,431.5 | -0.23% | -0.26% to -0.21% | | [HTTPX](https://github.com/encode/httpx/tree/b5addb64f0161ff6bfe94c124ef76f6a1fba5254) (inconclusive) | 1,683,810,371.5 | 1,682,014,111 | -0.11% | -0.13% to -0.08% | | [NumPy](https://github.com/numpy/numpy/tree/f9f1772ce010001208475f751d19253469788410) (inconclusive) | 8,497,337,652.5 | 8,480,204,308.5 | -0.20% | -0.32% to -0.09% | | [FastAPI](https://github.com/fastapi/fastapi/tree/7a0a3f0b286c6c865cf4398ea0e957a87b8bffa4) (inconclusive) | 3,008,218,552 | 3,004,472,580.5 | -0.12% | -0.22% to -0.03% | | [SQLAlchemy](https://github.com/sqlalchemy/sqlalchemy/tree/10f4d1e2d706a7722ca78e79ce05b00333a7aa5b) | 21,521,362,036.5 | 21,427,512,841 | -0.44% | -0.48% to -0.39% | | [Rich](https://github.com/Textualize/rich/tree/9d8f9a372cc5916fd4781fec207ced7ddac2f08f) (inconclusive) | 2,723,800,499.5 | 2,717,501,353.5 | -0.23% | -0.37% to -0.09% | | [Django](https://github.com/django/django/tree/b501b8493f31cf8cba31211d59b3b78642b8cb02) | 13,912,015,847 | 13,883,110,045 | -0.21% | -0.28% to -0.13% | | [SymPy](https://github.com/sympy/sympy/tree/85d3998a4517ffbc8a9709b8efa769bb417c174a) | 46,177,370,870 | 45,967,496,305.5 | -0.45% | -0.45% to -0.45% | | [OpenAI Python SDK](https://github.com/openai/openai-python/tree/f6e8480f9a5b2a4d02b17a1da54a1e57964bc312) (inconclusive) | 16,023,139,660 | 16,002,107,945.5 | -0.13% | -0.24% to -0.03% | | [pytest](https://github.com/pytest-dev/pytest/tree/8b005ed7f972987033ec34dc1e68b0635fb8142b) | 8,814,471,520.5 | 8,798,972,122.5 | -0.18% | -0.27% to -0.08% | | [aiohttp](https://github.com/aio-libs/aiohttp/tree/fe099491d275ef2c21bf683e024e34d51ac05d43) (inconclusive) | 4,175,149,063.5 | 4,172,710,283.5 | -0.06% | -0.15% to +0.04% | | [Ansible](https://github.com/ansible/ansible/tree/b4b2b1e82a535d27f902a71fc8396c3af138fca2) | 14,442,791,334.5 | 14,401,352,824 | -0.29% | -0.32% to -0.25% | | [botocore](https://github.com/boto/botocore/tree/e250abbe956da791c16086575e6ddb1f2da81a78) | 3,026,052,345 | 3,023,319,766 | -0.09% | -0.10% to -0.08% | | [Celery](https://github.com/celery/celery/tree/9c2f238334331db43f15956f131235fd3bb70fd1) | 4,823,425,865.5 | 4,812,542,246 | -0.23% | -0.29% to -0.16% | | [Dask](https://github.com/dask/dask/tree/249a51a1db4261791b93b4ce04e67341729939c3) | 34,967,356,287.5 | 33,935,527,972 | -2.95% | -2.98% to -2.93% | | [NetworkX](https://github.com/networkx/networkx/tree/6855cd586982cc2e4710b8eb87251ccf56674bf1) | 15,760,776,050.5 | 15,624,434,259.5 | -0.87% | -0.87% to -0.86% | | [Pydantic](https://github.com/pydantic/pydantic/tree/560f0f96a37d6913afdfddb2de22030ef08e68c3) | 13,064,494,398 | 13,039,730,470.5 | -0.19% | -0.19% to -0.19% | | [scikit-learn](https://github.com/scikit-learn/scikit-learn/tree/ee56c8b5d809dc7b970a140ed501269b63c69cf8) | 66,902,103,018 | 62,012,792,620 | -7.31% | -7.39% to -7.22% | | [Scrapy](https://github.com/scrapy/scrapy/tree/e63a574c767d7a9438c21db158f605f1c44c6e6a) | 8,974,801,335.5 | 8,962,767,833.5 | -0.13% | -0.17% to -0.10% | | [Transformers](https://github.com/huggingface/transformers/tree/3a134bc37f7c22c801b618e289b0255425f22647) | 102,173,739,178.5 | 101,044,134,037.5 | -1.11% | -1.21% to -1.01% |
Author
Parents
Loading