Implement `#ruff:file-ignore` file-level suppressions by amyreese · Pull Request #23599 · astral-sh/ruff

amyreese · 2026-02-27T02:21:13Z

Follow-up to #23404

Add support for #ruff:file-ignore[code] style file-level suppressions
as own-line comments at global module scope. The range covered by these
suppressions is always the entirety of the file:

# ruff:file-ignore[ARG001]

def foo(
	arg1,
	arg2,
):
	pass

def bar(
	arg1,
	arg2,
):
	pass

This currently requires having preview mode enabled. Without preview mode,
the comments are parsed and processed, but not materialized into active
suppressions at runtime.

astral-sh-bot · 2026-02-27T02:33:58Z

`ruff-ecosystem` results

Linter (stable)

✅ ecosystem check detected no linter changes.

Linter (preview)

✅ ecosystem check detected no linter changes.

ntBre

Looks reasonable to me! I'll just be curious to get Micha's thoughts on this and #23404 next week.

ntBre · 2026-02-27T14:20:22Z

+# ruff: ignore-all[F401, ARG001]
+```
+
+The full-level suppression comment specification is as follows:


Suggested change

The full-level suppression comment specification is as follows:

The full file-level suppression comment specification is as follows:

I think

dylwil3

I think this is a great idea! Especially because users have reported accidentally doing ruff: noqa: code when they meant to just do noqa: code. So I think eventually we should deprecate the ruff: noqa: code option in favor of this one.

One question I have is whether we want to take this opportunity to be more stringent about the requirement that the file-level suppression be "towards the top". Maybe before any code? (possibly excluding module-level docstrings)?

It always seemed strange to me that someone could stick # ruff: noqa: code randomly in the middle of the file and it would affect the whole file.

ntBre · 2026-02-27T14:56:58Z

One question I have is whether we want to take this opportunity to be more stringent about the requirement that the file-level suppression be "towards the top". Maybe before any code? (possibly excluding module-level docstrings)?

I thought about this as well, but I guess it could be a separate lint rule or something as an alternative.

MichaReiser

We'll need to update is_pragma_comment to recognize the new pragma comments that this PR and the ruff: ignore PR add.

amyreese · 2026-03-06T22:50:16Z

We'll need to update is_pragma_comment to recognize the new pragma comments that this PR and the ruff: ignore PR add.

That function already recognizes all comments starting with ruff: as pragma comments. :)

…nused suppressions

amyreese · 2026-04-21T16:04:05Z

Rebased, ensured suppressions are sorted by (action, textrange, rule) for consistent behavior/precedence, and added integration tests.

MichaReiser · 2026-04-22T12:57:51Z

Coming back to this and not remembering what this PR was about, I misinterpreted the PR title as meaning this PR introduces ruff: ignore-all as an equivalent to ruff: ignore[all] (assuming we supported categories within ignore comments). This triggered me to think a bit more about the naming and overall design. I'll lay out my thoughts because I haven't reached a conclusion on which I prefer, and I'm interested in your opinion.

The term all is overloaded because of rule selection, and I made the mistake of assuming that ruff: ignore-all is the same as ruff: ignore[all]. That's why I think we should avoid all. We could consider using ruff: ignore-file[code] or ruff: file-ignore[code].
Semantically, ruff: ignore-all feels closer to a ruff: disable comment because it suppresses a block of code and not just the next statement. That's why I think I'd prefer ruff: disable-all (or ruff: disable-file) over ruff: ignore-all.
When using ESLint, I always struggled to remember all the suppression styles. Which is why I think fewer suppression styles are generally preferred. One way to accomplish this is to special-case ruff: disable[...] at the start of a file. However, this would require adding an exception to unmatched-suppression-comment to not warn about ruff: disable comments at the start of a file, and open the door for accidental file-level suppressions. This is the part I have the least clarity on what I prefer. Do I prefer fewer suppression styles or avoiding accidental suppressions at all costs? Note mostly for myself: ESLint has a single eslint: disable comment, and so does ESLint-inspired Oxc; Rust has #![allow] for module-level suppression and #[allow] for block/next-definition suppression; Pyright uses pyright: E501=ignore.

amyreese · 2026-04-22T20:41:32Z

For reference, Biome uses ignore, ignore-all, ignore-start, and ignore-end — the last two being analogs of disable and enable. We also don't allow # ruff:ignore[all], (or ignoring by prefix for that matter) so I think the chance of confusion is low.

MichaReiser · 2026-04-22T21:15:53Z

We also don't allow # ruff:ignore[all],

I could see us allowing this in the future (using categories in suppressions). It's something I find useful in clippy, specifically at the file level.

amyreese · 2026-04-23T21:01:00Z

Updated to use # ruff: file-ignore terminology.

dylwil3

Looks good, thank you!

dylwil3 · 2026-04-24T13:41:19Z

+                    }
+                } else {
+                    warn_user_once!(
+                        "#ruff:file-ignore comment found but not active, enable preview mode"


Suggested change

"#ruff:file-ignore comment found but not active, enable preview mode"

"#ruff:file-ignore comment found but has no effect, enable preview mode"

I'm not fond of "has no effect" because it seems to imply that it's "unused", ie, doesn't suppress anything, rather than it being ignored by Ruff because it's not running in preview mode... 🤔

Also this matches what we already used for #ruff:ignore

dylwil3 · 2026-04-24T13:45:11Z

-                    covered_source: "# ruff: disable[bar]\n    print('hello')\n\n",
-                    code: "bar",
+                    covered_source: "# ruff: disable[foo]\nprint('hello')\n\ndef foo():\n    # ruff: disable[bar]\n    print('hello')\n\n",
+                    code: "foo",


did these snapshots change just because of the ordering change? (I think so but just making sure I'm not missing something)

Yes, this is because the sorting changed the order of the suppressions in the snapshots, and so the diff is a bit weird.

MichaReiser · 2026-04-24T15:24:40Z

+    def bar(self, arg1, arg2):
+        print("hello")
+
+# ruff: file-ignore[ARG002]  should cover the class method above!


This is a bit tricky, and I can see an argument in both directions:

Marking the inline comment as unused has the advantage that it reduces the total number of suppressions.

Marking the file-level suppression as unused has the advantage that it keeps suppressions as local as possible (unless the file-level suppression is used in at least one other place).

I wanted to argue that ty implements the second behavior, but it doesn't (https://play.ty.dev/47a21efb-d9e7-46f4-964e-69ff82bb5467). So I think this is fine.

* main: (248 commits) [ty] bump conformance suite commit (astral-sh#24848) [ty] Pass unmapped type variables to `SpecializationBuilder::build_with` (astral-sh#24809) [ty] Avoid bookkeeping for unannotated functions (astral-sh#24842) [ty] Optimize signature checking based on number of arguments (astral-sh#24674) [ty] Avoid eagerly inferring legacy generic context (astral-sh#24841) [ty] Skip decorator inference for undecorated functions (astral-sh#24839) [ty] solve unions against generic protocols (astral-sh#24837) [ty] Lazily allocate parameter type builders (astral-sh#24838) [ty] Lazily compute call argument fallbacks (astral-sh#24836) [ty] Avoid collecting type context callables (astral-sh#24835) Bump ecosystem-analyzer to e7576e6 (astral-sh#24834) [ty] Support basic narrowing with aliased conditional expressions (astral-sh#24302) [ty] Fix project and workspace selection (astral-sh#24824) [ty] Fix missing memory usage attributes (astral-sh#24823) Fix setting selection for multi folder workspace (astral-sh#24819) Improve diagnostics for implicit calls to a possibly unbound unary operator. (astral-sh#24816) [ty] Update review pool (astral-sh#24818) Bump 0.15.12 (astral-sh#24815) [ty] Complete support for more detailed diagnostics on possibly unbound errors from implicit dunder calls against unions. (astral-sh#24676) Implement `#ruff:file-ignore` file-level suppressions (astral-sh#23599) ...

astral-sh-bot Bot assigned ntBre Feb 27, 2026

amyreese changed the base branch from main to amy/ruff-ignore-end-of-line February 27, 2026 02:21

amyreese changed the title ~~Implement #ruff:ignore-all range suppressions~~ Implement #ruff:ignore-all file-level suppressions Feb 27, 2026

amyreese added suppression Related to supression of violations e.g. noqa preview Related to preview mode features labels Feb 27, 2026

amyreese requested review from MichaReiser and ntBre February 27, 2026 02:32

ntBre reviewed Feb 27, 2026

View reviewed changes

dylwil3 reviewed Feb 27, 2026

View reviewed changes

Comment thread docs/linter.md Outdated

amyreese force-pushed the amy/ruff-ignore-end-of-line branch 2 times, most recently from c9df742 to aef5860 Compare March 4, 2026 23:04

amyreese force-pushed the amy/ruff-ignore-all branch from 7bce278 to dfb0d89 Compare March 4, 2026 23:10

MichaReiser reviewed Mar 5, 2026

View reviewed changes

Comment thread crates/ruff_linter/src/suppression.rs

Comment thread crates/ruff_linter/src/suppression.rs

Comment thread crates/ruff_linter/src/suppression.rs Outdated

amyreese force-pushed the amy/ruff-ignore-end-of-line branch from aef5860 to d03c94e Compare March 12, 2026 21:23

amyreese force-pushed the amy/ruff-ignore-all branch from dfb0d89 to 64ca042 Compare March 12, 2026 21:24

amyreese force-pushed the amy/ruff-ignore-end-of-line branch from d03c94e to fdfae04 Compare March 17, 2026 20:18

amyreese force-pushed the amy/ruff-ignore-end-of-line branch from 20cf527 to 3cb3a2f Compare April 8, 2026 22:46

amyreese force-pushed the amy/ruff-ignore-end-of-line branch from 7089290 to 029f481 Compare April 20, 2026 14:59

Base automatically changed from amy/ruff-ignore-end-of-line to main April 20, 2026 18:38

Implement and document ignore-all

cce519c

amyreese force-pushed the amy/ruff-ignore-all branch from 64ca042 to f87a79a Compare April 20, 2026 19:17

amyreese added 3 commits April 20, 2026 15:26

Add preview mode warning

fa2e30a

Sort final suppressions by action to ensure consistent ordering for u…

3717f76

…nused suppressions

Add integration test case

02e59f5

amyreese force-pushed the amy/ruff-ignore-all branch from 8298523 to 87d0adf Compare April 21, 2026 15:03

Demonstrate lower ignore-all taking precedence over inline ignore

28bd03c

Rename ruff:ignore-all to ruff:file-ignore

0ba1467

amyreese force-pushed the amy/ruff-ignore-all branch from f1d6b14 to 0ba1467 Compare April 23, 2026 20:57

amyreese changed the title ~~Implement #ruff:ignore-all file-level suppressions~~ Implement #ruff:file-ignore file-level suppressions Apr 23, 2026

dylwil3 approved these changes Apr 24, 2026

View reviewed changes

MichaReiser approved these changes Apr 24, 2026

View reviewed changes

apply suggested changes

a635ce8

amyreese enabled auto-merge (squash) April 24, 2026 15:58

amyreese merged commit ed669ea into main Apr 24, 2026
43 checks passed

amyreese deleted the amy/ruff-ignore-all branch April 24, 2026 16:02

BrewTestBot mentioned this pull request Apr 24, 2026

ruff 0.15.12 Homebrew/homebrew-core#279335

Merged

This was referenced Apr 27, 2026

Always include panic payload in panic diagnostic message #24873

Merged

ASCII identifier fast path #24876

Draft

	The full-level suppression comment specification is as follows:
	The full file-level suppression comment specification is as follows:

	"#ruff:file-ignore comment found but not active, enable preview mode"
	"#ruff:file-ignore comment found but has no effect, enable preview mode"

Conversation

amyreese commented Feb 27, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

astral-sh-bot Bot commented Feb 27, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

ruff-ecosystem results

Linter (stable)

Linter (preview)

Uh oh!

ntBre left a comment

Choose a reason for hiding this comment

Uh oh!

ntBre Feb 27, 2026

Choose a reason for hiding this comment

Uh oh!

dylwil3 left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

ntBre commented Feb 27, 2026

Uh oh!

MichaReiser left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

amyreese commented Mar 6, 2026

Uh oh!

amyreese commented Apr 21, 2026

Uh oh!

MichaReiser commented Apr 22, 2026

Uh oh!

amyreese commented Apr 22, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

MichaReiser commented Apr 22, 2026

Uh oh!

amyreese commented Apr 23, 2026

Uh oh!

dylwil3 left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

dylwil3 Apr 24, 2026

Choose a reason for hiding this comment

Uh oh!

amyreese Apr 24, 2026

Choose a reason for hiding this comment

Uh oh!

amyreese Apr 24, 2026

Choose a reason for hiding this comment

Uh oh!

dylwil3 Apr 24, 2026

Choose a reason for hiding this comment

Uh oh!

amyreese Apr 24, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

MichaReiser Apr 24, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

amyreese commented Feb 27, 2026 •

edited

Loading

astral-sh-bot Bot commented Feb 27, 2026 •

edited

Loading

`ruff-ecosystem` results

amyreese commented Apr 22, 2026 •

edited

Loading