Expand the default rule set by ntBre · Pull Request #23385 · astral-sh/ruff

ntBre · 2026-02-17T23:40:39Z

Summary

This PR adds the new default rule set in preview. This ended up being pretty non-invasive because the DEFAULT_SELECTORS are only used in one place where preview isn't set to the default value of false.

I've currently listed each rule with a separate RuleSelector, which I generated with the script below. I thought about trying to be more clever by finding the smallest set of prefix selectors that yield the same rule set, but I figured this would be the easiest way to add and remove rules in the future anyway.

Script

import json
import subprocess
import tomllib
from pathlib import Path
from string import ascii_uppercase

RULES = {
    rule["code"]: rule["linter"]
    for rule in json.loads(
        subprocess.run(
            ["ruff", "rule", "--all", "--output-format=json"],
            check=True,
            text=True,
            capture_output=True,
        ).stdout
    )
}

for code, linter in RULES.items():
    if linter == "Ruff-specific rules":
        RULES[code] = "Ruff"


def kebab_to_pascal(s: str) -> str:
    return "".join(part.title() for part in s.split("-"))


rules = tomllib.loads(Path("proposal.toml").read_text())["lint"]["select"]

for rule in rules:
    linter = kebab_to_pascal(RULES[rule])
    suffix = rule.lstrip(ascii_uppercase)

    prefix = "_"
    match linter:
        case "Flake8Comprehensions":
            suffix = suffix.removeprefix("4")
        case "Pycodestyle":
            prefix = ""
            suffix = rule
        case "Flake8Gettext":
            linter = "Flake8GetText"
        case "Pep8Naming":
            linter = "PEP8Naming"
        case "Pylint":
            prefix = ""
            suffix = rule.removeprefix("PL")
        case "Flake8Debugger":
            suffix = suffix.removeprefix("10")

    print(
        " " * 4,
        f"RuleSelector::rule(RuleCodePrefix::{linter}("
        f"codes::{linter}::{prefix}{suffix})),"
        f" // {rule}",
        sep="",
    )

Test Plan

A new CLI test showing the preview default rules. I filtered down the snapshot to just the linter.rules.enabled section, which isn't strictly necessary but was a lot shorter. I also tested manually in VS Code to make sure I didn't miss wiring up the preview defaults in the server:

I also tested in the playground.

astral-sh-bot · 2026-02-17T23:48:12Z

`ruff-ecosystem` results

Linter (stable)

✅ ecosystem check detected no linter changes.

Linter (preview)

ℹ️ ecosystem check detected linter changes. (+21230 -35 violations, +0 -0 fixes in 16 projects; 40 projects unchanged)

aiven/aiven-client (+41 -0 violations, +0 -0 fixes)

ruff check --no-cache --exit-zero --no-fix --output-format concise --preview

+ aiven/client/__main__.py:1:1: I001 [*] Import block is un-sorted or un-formatted
+ aiven/client/argx.py:111:9: SIM102 Use a single `if` statement instead of nested `if` statements
+ aiven/client/argx.py:112:13: SIM102 Use a single `if` statement instead of nested `if` statements
+ aiven/client/argx.py:5:1: I001 [*] Import block is un-sorted or un-formatted
+ aiven/client/cli.py:2854:20: C417 Unnecessary `map()` usage (rewrite using a list comprehension)
+ aiven/client/cli.py:2903:22: C417 Unnecessary `map()` usage (rewrite using a list comprehension)
+ aiven/client/cli.py:324:35: FURB167 [*] Use of regular expression alias `re.I`
+ aiven/client/cli.py:4162:9: SIM102 Use a single `if` statement instead of nested `if` statements
+ aiven/client/cli.py:5209:13: SIM102 Use a single `if` statement instead of nested `if` statements
+ aiven/client/cli.py:559:27: TRY201 Use `raise` without specifying exception name
... 31 additional changes omitted for project

PlasmaPy/PlasmaPy (+57 -0 violations, +0 -0 fixes)

ruff check --no-cache --exit-zero --no-fix --output-format concise --preview

+ .github/scripts/authors_in_cff.py:60:9: LOG015 `info()` call on root logger
+ .github/scripts/authors_in_cff.py:92:5: LOG015 `info()` call on root logger
+ docs/conf.py:59:5: LOG015 `info()` call on root logger
+ docs/conf.py:63:5: LOG015 `warning()` call on root logger
+ src/plasmapy/__init__.py:13:11: RUF022 [*] `__all__` is not sorted
+ src/plasmapy/analysis/swept_langmuir/__init__.py:6:11: RUF022 [*] `__all__` is not sorted
+ src/plasmapy/analysis/swept_langmuir/floating_potential.py:3:11: RUF022 [*] `__all__` is not sorted
+ src/plasmapy/analysis/swept_langmuir/ion_saturation_current.py:5:11: RUF022 [*] `__all__` is not sorted
+ src/plasmapy/diagnostics/charged_particle_radiography/detector_stacks.py:6:11: RUF022 [*] `__all__` is not sorted
+ src/plasmapy/diagnostics/langmuir.py:3:11: RUF022 [*] `__all__` is not sorted
... 37 additional changes omitted for rule RUF022
... 47 additional changes omitted for project

docker/docker-py (+209 -0 violations, +0 -0 fixes)

ruff check --no-cache --exit-zero --no-fix --output-format concise --preview

+ docker/api/build.py:128:9: SIM102 Use a single `if` statement instead of nested `if` statements
+ docker/api/build.py:133:13: SIM118 Use `key in dict` instead of `key in dict.keys()`
+ docker/api/client.py:259:17: TRY004 Prefer `TypeError` exception for invalid type
+ docker/api/client.py:296:21: PERF403 Use `dict.update` instead of a for-loop
+ docker/api/client.py:512:23: TRY201 Use `raise` without specifying exception name
+ docker/api/container.py:867:13: SIM114 [*] Combine `if` branches using logical `or` operator
+ docker/api/container.py:884:13: SIM114 [*] Combine `if` branches using logical `or` operator
+ docker/api/image.py:401:19: RUF059 Unpacked variable `repo_name` is never used
+ docker/api/image.py:476:19: RUF059 Unpacked variable `repo_name` is never used
+ docker/api/plugin.py:132:19: RUF059 Unpacked variable `repo_name` is never used
... 199 additional changes omitted for project

facebookresearch/chameleon (+303 -0 violations, +0 -0 fixes)

ruff check --no-cache --exit-zero --no-fix --output-format concise --preview

+ chameleon/inference/alignment.py:13:38: FA102 Missing `from __future__ import annotations`, but uses PEP 585 collection
+ chameleon/inference/alignment.py:13:43: FA102 Missing `from __future__ import annotations`, but uses PEP 585 collection
+ chameleon/inference/alignment.py:17:41: FA102 Missing `from __future__ import annotations`, but uses PEP 585 collection
+ chameleon/inference/alignment.py:17:46: FA102 Missing `from __future__ import annotations`, but uses PEP 585 collection
+ chameleon/inference/alignment.py:31:38: FA102 Missing `from __future__ import annotations`, but uses PEP 585 collection
+ chameleon/inference/alignment.py:31:43: FA102 Missing `from __future__ import annotations`, but uses PEP 585 collection
... 261 additional changes omitted for rule FA102
+ chameleon/inference/chameleon.py:433:12: FA100 Add `from __future__ import annotations` to simplify `typing.Union`
+ chameleon/inference/chameleon.py:434:12: FA100 Add `from __future__ import annotations` to simplify `typing.Union`
+ chameleon/inference/chameleon.py:435:17: FA100 Add `from __future__ import annotations` to simplify `typing.Union`
+ chameleon/inference/chameleon.py:436:22: FA100 Add `from __future__ import annotations` to simplify `typing.Union`
... 293 additional changes omitted for project

ing-bank/probatus (+55 -0 violations, +0 -0 fixes)

ruff check --no-cache --exit-zero --no-fix --output-format concise --preview

+ probatus/feature_elimination/__init__.py:1:1: I001 [*] Import block is un-sorted or un-formatted
+ probatus/feature_elimination/__init__.py:4:11: RUF022 [*] `__all__` is not sorted
+ probatus/feature_elimination/early_stopping_feature_elimination.py:1:1: I001 [*] Import block is un-sorted or un-formatted
+ probatus/feature_elimination/feature_elimination.py:1018:16: RET504 Unnecessary assignment to `ranking` before `return` statement
+ probatus/feature_elimination/feature_elimination.py:1:1: I001 [*] Import block is un-sorted or un-formatted
+ probatus/feature_elimination/feature_elimination.py:669:13: PLR1730 [*] Replace `if` statement with `current_step = max(current_step, 1)`
+ probatus/feature_elimination/feature_elimination.py:866:13: PLW0133 Missing `raise` statement on exception
+ probatus/feature_elimination/feature_elimination.py:976:30: PLW0211 First argument of a static method should not be named `self`
+ probatus/feature_elimination/feature_elimination.py:989:20: SIM210 Remove unnecessary `True if ... else False`
+ probatus/feature_elimination/feature_elimination.py:991:16: RET504 Unnecessary assignment to `support` before `return` statement
... 45 additional changes omitted for project

latchbio/latch (+6 -0 violations, +0 -0 fixes)

ruff check --no-cache --exit-zero --no-fix --output-format concise --preview

+ src/latch_cli/centromere/ctx.py:158:20: BLE001 Do not catch blind exception: `Exception`
+ src/latch_cli/main.py:1438:16: BLE001 Do not catch blind exception: `Exception`
+ src/latch_cli/services/launch/type_converter.py:340:16: BLE001 Do not catch blind exception: `Exception`
+ src/latch_cli/services/local_dev_old.py:344:16: BLE001 Do not catch blind exception: `Exception`
+ src/latch_cli/services/local_dev_old.py:430:24: BLE001 Do not catch blind exception: `Exception`
+ src/latch_cli/services/login.py:67:12: BLE001 Do not catch blind exception: `Exception`

prefecthq/prefect (+16246 -0 violations, +0 -0 fixes)

ruff check --no-cache --exit-zero --no-fix --output-format concise --preview

+ benches/__main__.py:18:11: PLW1510 `subprocess.run` without explicit `check` argument
+ client/client_flow.py:18:35: PLW1508 Invalid type for environment variable default; expected `str` or `None`
+ examples/ai_database_cleanup_with_approval.py:255:24: BLE001 Do not catch blind exception: `Exception`
+ examples/per_worker_task_concurrency.py:120:12: RET504 Unnecessary assignment to `result` before `return` statement
+ examples/run_api_sourced_etl.py:105:12: RET504 Unnecessary assignment to `df` before `return` statement
+ examples/run_api_sourced_etl.py:132:9: PERF401 Use a list comprehension to create a transformed list
+ integration-tests/test_automation_assessments.py:132:19: TRY002 Create your own exception
+ integration-tests/test_automation_assessments.py:180:19: TRY002 Create your own exception
+ integration-tests/test_automation_assessments.py:243:19: TRY002 Create your own exception
+ integration-tests/test_automation_assessments.py:30:17: UP006 [*] Use `dict` instead of `Dict` for type annotation
+ integration-tests/test_automation_assessments.py:315:19: TRY002 Create your own exception
+ integration-tests/test_automation_assessments.py:31:21: UP006 [*] Use `dict` instead of `Dict` for type annotation
+ integration-tests/test_automation_assessments.py:6:1: UP035 [*] Import from `collections.abc` instead: `AsyncGenerator`
+ integration-tests/test_automation_assessments.py:6:1: UP035 `typing.Dict` is deprecated, use `dict` instead
+ integration-tests/test_automation_assessments.py:83:19: PLR1704 Redefining argument with the local name `event`
+ integration-tests/test_automation_assessments.py:87:11: TRY002 Create your own exception
+ integration-tests/test_client_context_lifespan.py:114:19: TRY201 Use `raise` without specifying exception name
+ integration-tests/test_client_context_lifespan.py:4:1: UP035 [*] Import from `collections.abc` instead: `Callable`
+ integration-tests/test_concurrency_leases.py:41:16: RET504 Unnecessary assignment to `limit` before `return` statement
+ integration-tests/test_concurrency_leases.py:69:5: SIM117 [*] Use a single `with` statement with multiple contexts instead of nested `with` statements
+ integration-tests/test_load_flows_concurrently.py:37:16: BLE001 Do not catch blind exception: `Exception`
+ integration-tests/test_load_flows_concurrently.py:38:60: RUF010 [*] Use explicit conversion flag
+ integration-tests/test_runner_resilience.py:83:12: BLE001 Do not catch blind exception: `Exception`
+ integration-tests/test_schedule_statefulness.py:19:1: UP035 [*] Import from `collections.abc` instead: `Callable`
+ integration-tests/test_schedule_statefulness.py:49:16: BLE001 Do not catch blind exception: `Exception`
+ integration-tests/test_task_worker.py:6:1: UP035 [*] Import from `collections.abc` instead: `AsyncGenerator`
+ integration-tests/test_worker.py:16:39: UP006 [*] Use `list` instead of `List` for type annotation
+ integration-tests/test_worker.py:25:13: PERF401 Use `list.extend` with an async comprehension to create a transformed list
+ integration-tests/test_worker.py:28:32: UP006 [*] Use `list` instead of `List` for type annotation
+ integration-tests/test_worker.py:34:45: RUF100 [*] Unused `noqa` directive (unused: `F821`)
+ integration-tests/test_worker.py:35:13: UP006 [*] Use `list` instead of `List` for type annotation
+ integration-tests/test_worker.py:5:1: UP035 `typing.List` is deprecated, use `list` instead
... 793 additional changes omitted for rule UP035
+ scripts/all_links_should_be_ok.py:38:12: BLE001 Do not catch blind exception: `Exception`
+ scripts/all_links_should_be_ok.py:48:12: BLE001 Do not catch blind exception: `Exception`
... 237 additional changes omitted for rule BLE001
+ scripts/backfill_release_notes.py:134:12: PIE810 Call `startswith` once with a `tuple`
+ scripts/backfill_release_notes.py:167:12: PIE810 Call `startswith` once with a `tuple`
+ scripts/backfill_release_notes.py:199:13: PLR1714 Consider merging multiple comparisons. Use a `set` if the elements are hashable.
+ scripts/backfill_release_notes.py:224:13: SIM102 Use a single `if` statement instead of nested `if` statements
... 16208 additional changes omitted for project

... Truncated remaining completed project reports due to GitHub comment length restrictions

Changes by rule (180 rules affected)

code	total	+ violation	- violation
UP045	5834	5834	0
FA102	2916	2916	0
UP006	2219	2219	0
C408	1387	1387	0
UP037	1130	1130	0
UP007	902	902	0
UP035	798	798	0
RUF100	671	671	0
FA100	484	484	0
RET504	457	457	0
B008	446	446	0
BLE001	345	345	0
SIM117	300	300	0
RUF012	185	185	0
SIM102	178	178	0
RUF059	173	173	0
TRY002	122	122	0
I001	120	120	0
TRY300	116	116	0
PERF401	109	109	0
PLR0402	86	86	0
RUF022	84	84	0
C419	78	78	0
PYI036	73	73	0
DTZ001	68	68	0
SIM118	66	66	0
TRY004	66	66	0
TRY201	65	65	0
RUF015	63	63	0
SIM115	60	60	0
FLY002	59	59	0
PYI063	56	56	0
B018	52	52	0
B006	47	47	0
RUF010	45	45	0
PIE790	44	44	0
PYI034	42	42	0
S110	40	40	0
PLR1714	35	35	0
RUF013	33	33	0
B010	32	32	0
SIM114	32	32	0
PLR1704	30	30	0
G201	30	30	0
SIM103	30	30	0
B009	30	30	0
DTZ005	29	29	0
C401	27	27	0
RET501	27	27	0
C402	27	27	0
PERF402	26	26	0
PIE804	26	26	0
PYI032	26	26	0
PYI041	23	23	0
LOG015	21	21	0
UP032	21	21	0
PLC0414	21	21	0
PLW1510	19	19	0
B015	19	19	0
UP012	19	19	0
UP030	19	19	0
PLW1508	18	18	0
PLR1711	18	18	0
PT031	18	18	0
C403	18	18	0
PT014	17	17	0
C405	17	17	0
PIE807	17	17	0
PYI055	17	17	0
C413	16	16	0
S102	16	16	0
PYI066	15	15	0
PLW0602	14	14	0
PYI019	14	14	0
F403	14	0	14
PYI016	13	13	0
UP004	13	13	0
E402	13	0	13
PIE808	11	11	0
PIE810	11	11	0
PERF102	11	11	0
C417	10	10	0
PLC0208	10	10	0
B017	10	10	0
UP009	10	10	0
SIM201	9	9	0
C414	9	9	0
PYI059	9	9	0
SIM101	8	8	0
UP034	8	8	0
D419	8	8	0
TRY401	8	8	0
TRY203	8	8	0
PLR0133	8	8	0
E712	8	0	8
B023	7	7	0
PLR1722	7	7	0
PLC0206	7	7	0
UP024	7	7	0
C400	7	7	0
B026	7	7	0
FURB177	7	7	0
G202	7	7	0
LOG014	7	7	0
PLW0127	6	6	0
PLR2044	6	6	0
PLW0120	6	6	0
RUF018	6	6	0
RUF023	6	6	0
FURB167	5	5	0
PERF403	5	5	0
EXE001	5	5	0
PIE800	5	5	0
UP036	5	5	0
PLR1733	5	5	0
TC004	5	5	0
PLR0124	5	5	0
SIM905	5	5	0
SIM210	4	4	0
DTZ007	4	4	0
FURB136	4	4	0
DTZ011	4	4	0
PTH124	4	4	0
PLE0704	3	3	0
LOG009	3	3	0
TC005	3	3	0
UP031	3	3	0
FURB188	3	3	0
B020	3	3	0
PYI030	3	3	0
B033	3	3	0
N999	3	3	0
PIE796	2	2	0
FURB129	2	2	0
EXE002	2	2	0
DTZ004	2	2	0
SIM113	2	2	0
PLE1205	2	2	0
PLR1730	2	2	0
PLW0128	2	2	0
PIE794	2	2	0
YTT103	2	2	0
UP010	2	2	0
SIM211	2	2	0
PLC0205	2	2	0
UP028	2	2	0
PLC0105	2	2	0
SIM223	2	2	0
PYI061	2	2	0
PYI025	2	2	0
UP018	2	2	0
FURB105	2	2	0
FURB157	1	1	0
PLW1509	1	1	0
PLW0133	1	1	0
PLW0211	1	1	0
UP014	1	1	0
RUF007	1	1	0
S112	1	1	0
B012	1	1	0
PLW0642	1	1	0
B013	1	1	0
YTT301	1	1	0
DTZ006	1	1	0
B039	1	1	0
UP011	1	1	0
DTZ901	1	1	0
FURB163	1	1	0
PLW0129	1	1	0
B030	1	1	0
RUF051	1	1	0
LOG001	1	1	0
DTZ002	1	1	0
RUF019	1	1	0
PLR1736	1	1	0
DTZ003	1	1	0
PYI044	1	1	0
FURB122	1	1	0
PYI006	1	1	0
PYI045	1	1	0

Formatter (stable)

✅ ecosystem check detected no format changes.

Formatter (preview)

✅ ecosystem check detected no format changes.

#23203 (reply in thread)

ntBre · 2026-02-18T16:11:20Z

README.md

+In [preview](https://docs.astral.sh/ruff/preview/), Ruff enables an expanded set of default rules
+that includes rules from the `B`, `UP`, and `RUF` categories, as well as many more. If you give the
+new defaults a try, feel free to leave feedback in the [GitHub
+discussion](https://github.com/astral-sh/ruff/discussions/23203), where you can also find the new
+rule set listed in full.
+


I'm not totally sold on including this, especially in the README (but this is how the top of our Rules page is generated). I'd be happy just to put something like this in the release notes instead, but I'm curious what others think.

I like including this, but should probably also say something along the lines of "and this will be the default set in a future minor release"

To me that's kind of implied by referencing preview, but I don't feel too strongly.

I like it here, and I agree with you that "in a future minor" is implicit. (In fact - if people don't like them in preview, then they won't be stabilized in a future minor!)

ntBre · 2026-02-18T18:24:00Z

I think there are a few too many ecosystem results to verify them all manually, but I wanted to at least address a few of the top rules:

UP045: Optional[T] -> T | None and has an autofix
FA102: checks for annotations that are missing from __future__ import annotations

This one is a bit surprising because the code shouldn't actually work as written if these are true positives. Some projects must have their Python versions misconfigured. I should probably look into this one more closely. It also had 0 hits in my previous ecosystem check, although I may have been running with --isolated.
UP006: typing.List -> list, seems preferable and has an autofix
C408, UP037, UP007, and UP035 all seem reasonable and have fixes

Those are a few high level thoughts. I'm working on a script to check a sample of the diagnostics directly just to rule out some obvious false positives. I think this is otherwise ready for review!

amyreese

I'm also curious: how many of the rules that are currently default will get disabled by default in preview mode? Is that worth calling out somewhere?

ntBre · 2026-02-18T19:59:39Z

I'm pretty happy with the ecosystem results. It looks like prefect only recently (PrefectHQ/prefect#19273) dropped Python 3.9 support, which explains why they have so many UP violations, despite their requires-python>=3.10.

Similarly, chameleon and qdrant-client do not declare versions, so we default them to an older Python version and thus report FA102. I guess we could consider defaulting to the latest Python version for this rule instead of the oldest since the semantics are reversed from most rules that use TargetVersion::linter_version.

C408 is one case where the results might be changing my mind. I think this came up in an earlier discussion, but I kind of like to use the dict(key=value) format when the keys are valid identifiers, and that's one of the common patterns in prefect (1288/1387 are from prefect). I think I might drop this rule for now.

RUF100 also seems somewhat troubling. The examples I checked for prefect were blanket noqas, so it's not even possible to configure around them with lint.external. I'd be curious to get other thoughts on this. I think when I added it to the set I was only picturing Ruff rules, where I think this would be a nice default to have, but it can potentially interact badly with other linters.

The prefect hits for B008 also look like false positives, but they're in functions decorated with decorators from typer and fastapi, so I think it's reasonable. And they can be fixed by configuring lint.flake8-bugbear.extend-immutable-calls.

It looks like PIE796 technically has 2 false positives of this form:

class Store(Enum):
    overwrite = object()
    write = object()
    skip = object()
    ^ PIE796: Enum contains duplicate value: `object()`

    def handle(self, getter: Callable[[], str], path: str) -> None:

but I guess it's probably better to use enum.auto() here anyway.

There's a strange false positive for B030 in setuptools:

        try:
            parsed = self.parsers.get(option_name, lambda x: x)(value)
        except (Exception,) * self.ignore_option_errors:
               ^ B030: `except` handlers should only be exception classes or tuples of exception classes
            return

I assume self.ignore_option_errors is either 1 or 0 and can bypass this exception handler. That seems unusual enough not to worry about.

A couple other rules looked more pedantic to me today, but I don't want to second guess too many of the rules. Hopefully we're about to get some feedback on them anyway! I'm leaning toward moving ahead with this set, with the possible exception of C408.

ntBre · 2026-02-18T20:02:25Z

I'm also curious: how many of the rules that are currently default will get disabled by default in preview mode? Is that worth calling out somewhere?

I think only a handful of the E rules were removed, but I can double check and try to work that in somewhere!

amyreese · 2026-02-18T20:05:12Z

RUF100 also seems somewhat troubling. The examples I checked for prefect were blanket noqas, so it's not even possible to configure around them with lint.external. I'd be curious to get other thoughts on this. I think when I added it to the set I was only picturing Ruff rules, where I think this would be a nice default to have, but it can potentially interact badly with other linters.

Rather than turn off RUF100 (I think that would be a huge miss) maybe we should narrow RUF100 to not fire on blanket noqa, and have a separate code for that? We could combine that with narrowing RUF100 to not overlap with RUF102?

amyreese · 2026-02-18T20:08:00Z

Maybe C408 could similarly be narrowed to only fire on dict() with no arguments? Even the documentation for it implies that it only cares about empty dict() vs {}, etc.

ntBre · 2026-02-18T20:11:19Z

I like those suggestions. It does make some sense to me to leave these rules in the default set and potentially update their behavior instead of removing them.

dscorbett · 2026-02-18T20:16:03Z

Maybe C408 could similarly be narrowed to only fire on dict() with no arguments? Even the documentation for it implies that it only cares about empty dict() vs {}, etc.

That could be effected by keeping the rule as is but defaulting allow-dict-calls-with-keyword-arguments to true.

ntBre · 2026-02-18T20:23:56Z

I'm also curious: how many of the rules that are currently default will get disabled by default in preview mode? Is that worth calling out somewhere?

I think only a handful of the E rules were removed, but I can double check and try to work that in somewhere!

On second thought, I think this will be too verbose to work in smoothly. For reference, these are the removed rules:

E401
E402
E701
E702
E703
E711
E712
E713
E714
E721
E731
E741
E742
E743
F403
F405
F406
F722

so it is mostly E rules (only E902 and E722 are preserved) plus a couple of F rules.

amyreese · 2026-02-18T20:28:32Z

On second thought, I think this will be too verbose to work in smoothly. For reference, these are the removed rules:

Could be worth documenting somewhere and reference it in the changelog at the same time. My main consideration would be the UX of users expecting those rules to be enabled, or unexpectedly getting RUF100 for any existing #noqa for those violations.

ntBre · 2026-02-18T20:39:15Z

I'd be open to suggestions for where to document it! I just think it feels out of place in the README/top of the Rules page, but that's the main place where we discuss the defaults. And it's now documented here in the PR, which is linked from the release notes, at least. I could also add it to the body of the discussion and/or the changelog itself if nothing else.

dylwil3

Amazing! Thanks for all the analysis and discussion that went into this

dylwil3 · 2026-02-19T19:24:56Z

README.md

+In [preview](https://docs.astral.sh/ruff/preview/), Ruff enables an expanded set of default rules
+that includes rules from the `B`, `UP`, and `RUF` categories, as well as many more. If you give the
+new defaults a try, feel free to leave feedback in the [GitHub
+discussion](https://github.com/astral-sh/ruff/discussions/23203), where you can also find the new
+rule set listed in full.
+


I like it here, and I agree with you that "in a future minor" is implicit. (In fact - if people don't like them in preview, then they won't be stabilized in a future minor!)

dylwil3 · 2026-02-19T19:28:05Z

I like those suggestions. It does make some sense to me to leave these rules in the default set and potentially update their behavior instead of removing them.

We also don't have to feel too bad about modifying this default set while it's still in preview.

ntBre added 2 commits February 17, 2026 18:26

add PREVIEW_DEFAULT_RULES and a test

78210c0

include proposed rules as individual selectors

bd206cd

ntBre added preview Related to preview mode features rule-selection Related to enabling or disabling rules labels Feb 17, 2026

ntBre added 5 commits February 17, 2026 18:51

clippy and update other tests

339b8b5

update rule docs

5608fba

remove rules from discussion

0d2bf9b

#23203 (reply in thread)

remove rules from discussion

24e34ed

#23203 (reply in thread)

update snapshots

5acf5d5

ntBre commented Feb 18, 2026

View reviewed changes

allow links to preview

189c7e5

ntBre marked this pull request as ready for review February 18, 2026 18:24

ntBre changed the title ~~[WIP] Add new default rules~~ Expand the default rule set Feb 18, 2026

amyreese approved these changes Feb 18, 2026

View reviewed changes

dylwil3 approved these changes Feb 19, 2026

View reviewed changes

ntBre merged commit 222574a into main Feb 19, 2026
42 checks passed

ntBre deleted the brent/default-rules branch February 19, 2026 19:45

BrewTestBot mentioned this pull request Feb 19, 2026

ruff 0.15.2 Homebrew/homebrew-core#268449

Merged

Conversation

ntBre commented Feb 17, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Test Plan

Uh oh!

astral-sh-bot bot commented Feb 17, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

ruff-ecosystem results

Linter (stable)

Linter (preview)

Formatter (stable)

Formatter (preview)

Uh oh!

ntBre Feb 18, 2026

Choose a reason for hiding this comment

Uh oh!

amyreese Feb 18, 2026

Choose a reason for hiding this comment

Uh oh!

ntBre Feb 18, 2026

Choose a reason for hiding this comment

Uh oh!

dylwil3 Feb 19, 2026

Choose a reason for hiding this comment

Uh oh!

ntBre commented Feb 18, 2026

Uh oh!

amyreese left a comment

Choose a reason for hiding this comment

Uh oh!

ntBre commented Feb 18, 2026

Uh oh!

ntBre commented Feb 18, 2026

Uh oh!

amyreese commented Feb 18, 2026

Uh oh!

amyreese commented Feb 18, 2026

Uh oh!

ntBre commented Feb 18, 2026

Uh oh!

dscorbett commented Feb 18, 2026

Uh oh!

ntBre commented Feb 18, 2026

Uh oh!

amyreese commented Feb 18, 2026

Uh oh!

ntBre commented Feb 18, 2026

Uh oh!

dylwil3 left a comment

Choose a reason for hiding this comment

Uh oh!

dylwil3 Feb 19, 2026

Choose a reason for hiding this comment

Uh oh!

dylwil3 commented Feb 19, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

Comments

ntBre commented Feb 17, 2026 •

edited

Loading

astral-sh-bot bot commented Feb 17, 2026 •

edited

Loading

`ruff-ecosystem` results