Emit `check-cfg` lints during attribute parsing rather than evaluation #149215

JonathanBrouwer · 2025-11-22T14:39:14Z

The goal of this PR is to make the eval_config_entry not have any side effects, by moving the check-cfg lints to the attribute parsing. This also helps ensure we do emit the lint in situations where the attribute happens to be parsed, but never evaluated.

cc @jdonszelmann @Urgau for a vibe check if you feel like it

Urgau · 2025-11-22T15:09:21Z

Does it mean we would get warnings for non-active cfg attributes?

#[cfg(target_os = "lol")] // warn and not active 
#[cfg(unknown)] // currently doesn't warn, since above is not active, but will it with this PR?
fn foo() {}

Regarding the non-accessible TyCtxt, would it be possible to continue to buffer the lint in ParseSess::buffered_lints, using sess.parse_sess.buffer_lint? Or is the problem that you don't emit the lint as a builtin lint and as such no longer have access to TyCtxt?

JonathanBrouwer · 2025-11-25T19:02:49Z

Attributes that are configured out (like in your example) are currently not parsed, and therefore this would not cause this warning, preserving the old behavior. I didn't think of this edgecase so I'll make sure to add a test for this.

Indeed currently attribute lints are special and not emitted as a builtin lint. I'm gonna play around with this a bit and see if anything is stopping me from changing that

…r=Urgau Run `eval_config_entry` on all branches so we always emit lints Fixes rust-lang#149090 Ideally I'd have liked to fix this issue using rust-lang#149215, and this is still the long term plan, but this is slightly more annoying to implement than I'd have liked to, and this is also a nice and easy solution to the problem. r? `@tgross35`

Rollup merge of #149380 - JonathanBrouwer:cfg_select_lints, r=Urgau Run `eval_config_entry` on all branches so we always emit lints Fixes #149090 Ideally I'd have liked to fix this issue using #149215, and this is still the long term plan, but this is slightly more annoying to implement than I'd have liked to, and this is also a nice and easy solution to the problem. r? `@tgross35`

Run `eval_config_entry` on all branches so we always emit lints Fixes rust-lang/rust#149090 Ideally I'd have liked to fix this issue using rust-lang/rust#149215, and this is still the long term plan, but this is slightly more annoying to implement than I'd have liked to, and this is also a nice and easy solution to the problem. r? `@tgross35`

bors · 2025-11-28T09:40:12Z

☔ The latest upstream changes (presumably #149410) made this pull request unmergeable. Please resolve the merge conflicts.

Run `eval_config_entry` on all branches so we always emit lints Fixes rust-lang/rust#149090 Ideally I'd have liked to fix this issue using rust-lang/rust#149215, and this is still the long term plan, but this is slightly more annoying to implement than I'd have liked to, and this is also a nice and easy solution to the problem. r? `@tgross35`

…elmann Move even more early buffered lints to dyn lint diagnostics Follow-up to #145881 and #145747. I originally wanted to migrate most if not the entire rest of the early buffered lints. However, when trying to migrate the buffered lints used by check-cfg I encountered a roadblock. Namely, it depends on `TyCtxt` (well, `Option<TyCtxt>`) which makes it quite hard to migrate (see also #147634 (comment), #147634 (comment) & #149215). So for now, I won't migrate it (maybe #149215 will find a solution), nor will I migrate the rest since it's quite tedious to migrate these. I'll leave them for future me.

bors · 2025-12-05T02:45:37Z

☔ The latest upstream changes (presumably #149646) made this pull request unmergeable. Please resolve the merge conflicts.

JonathanBrouwer · 2025-12-05T14:16:48Z

Rebased on #149662

After that PR, this change becomes trivial, we don't need to move the lints and we have access to TyCtxt

rustbot · 2025-12-06T09:39:10Z

Some changes occurred in compiler/rustc_attr_parsing

cc @jdonszelmann

Some changes occurred in compiler/rustc_hir/src/attrs

cc @jdonszelmann

JonathanBrouwer · 2025-12-06T09:39:38Z

The PR is ready now :)

tests/ui/macros/cfg.stderr

jdonszelmann · 2025-12-08T14:56:43Z

tests/ui/macros/cfg_select.stderr

+   |     ^ help: found config with similar value: `target_feature = "a"`
+   |
+   = help: expected names are: `FALSE` and `test` and 31 more
+   = help: to expect this configuration use `--check-cfg=cfg(a)`


same here etc

jdonszelmann

some small comments. Let me know, if I'm wrong about them I'll approve after :)

View changes since this review

jdonszelmann · 2025-12-08T15:00:09Z

compiler/rustc_expand/src/expand.rs

        pos: usize,
    ) -> EvalConfigResult {
-        let res = self.cfg().cfg_true(&attr, node.node_id(), ShouldEmit::ErrorsAndLints);
+        let res = self.cfg().cfg_true(&attr, ShouldEmit::ErrorsAndLints);


do we still need the ShouldEmit. You removed the nodeid and that's usually relevant for emitting lints. So maybe we don't need the should_emit anymore either?

I guess you're getting the nodeid out of self now. Maybe should emit as well? Or does that change based on where we call this

The ShouldEmit depends on where we call this indeed. Some cfg attributes are evaluated multiple times and we want to be careful to only emit errors once, that's why the ShouldEmit is an argument here

The node.node_id() was actually incorrect, because that turned out to be always the dummy node id

I see, cool!

jdonszelmann · 2025-12-08T15:03:33Z

@rustbot author

rustbot · 2025-12-08T15:03:38Z

Reminder, once the PR becomes ready for a review, use @rustbot ready.

JonathanBrouwer · 2025-12-08T15:44:31Z

@rustbot ready

jdonszelmann · 2025-12-08T22:47:40Z

Looks good then I think. @bors r+ rollup

bors · 2025-12-08T22:47:42Z

📌 Commit 01576a6 has been approved by jdonszelmann

It is now in the queue for this repository.

…szelmann Emit `check-cfg` lints during attribute parsing rather than evaluation The goal of this PR is to make the `eval_config_entry` not have any side effects, by moving the check-cfg lints to the attribute parsing. This also helps ensure we do emit the lint in situations where the attribute happens to be parsed, but never evaluated. cc `@jdonszelmann` `@Urgau` for a vibe check if you feel like it

Rollup of 12 pull requests Successful merges: - #147572 (std: Use more `unix.rs` code on WASI targets) - #148491 ( Correctly provide suggestions when encountering `async fn` with a `dyn Trait` return type) - #149215 (Emit `check-cfg` lints during attribute parsing rather than evaluation) - #149652 (Add release notes for 1.92.0) - #149720 (rustdoc book: mention inner doc attribute) - #149730 (lint: emit proper diagnostic for unsafe binders in improper_ctypes instead of ICE) - #149754 (Retire `opt_str2` from compiletest cli parsing) - #149755 (bootstrap: Use a `CompiletestMode` enum instead of bare strings) - #149763 (Add inline attribute to generated delegation function if needed) - #149772 (test: Add a test for 146133) - #149779 (Fix typo "an" → "and") - #149782 (Remove `[no-mentions]` handler in the triagebot config) r? `@ghost` `@rustbot` modify labels: rollup

Rollup of 11 pull requests Successful merges: - #147585 (Suppress the error for private fields with non_exhaustive attribute) - #149215 (Emit `check-cfg` lints during attribute parsing rather than evaluation) - #149652 (Add release notes for 1.92.0) - #149720 (rustdoc book: mention inner doc attribute) - #149730 (lint: emit proper diagnostic for unsafe binders in improper_ctypes instead of ICE) - #149754 (Retire `opt_str2` from compiletest cli parsing) - #149755 (bootstrap: Use a `CompiletestMode` enum instead of bare strings) - #149763 (Add inline attribute to generated delegation function if needed) - #149772 (test: Add a test for 146133) - #149779 (Fix typo "an" → "and") - #149782 (Remove `[no-mentions]` handler in the triagebot config) Failed merges: - #148491 ( Correctly provide suggestions when encountering `async fn` with a `dyn Trait` return type) r? `@ghost` `@rustbot` modify labels: rollup

Rollup merge of #149215 - JonathanBrouwer:cfg_lints2, r=jdonszelmann Emit `check-cfg` lints during attribute parsing rather than evaluation The goal of this PR is to make the `eval_config_entry` not have any side effects, by moving the check-cfg lints to the attribute parsing. This also helps ensure we do emit the lint in situations where the attribute happens to be parsed, but never evaluated. cc ``@jdonszelmann`` ``@Urgau`` for a vibe check if you feel like it

This comment has been minimized.

Sign in to view

JonathanBrouwer force-pushed the cfg_lints2 branch from abb82ad to 4ee0f66 Compare November 22, 2025 15:21

This comment has been minimized.

Sign in to view

JonathanBrouwer force-pushed the cfg_lints2 branch from 4ee0f66 to f5e9e93 Compare November 22, 2025 19:55

This comment has been minimized.

Sign in to view

JonathanBrouwer mentioned this pull request Nov 27, 2025

Run eval_config_entry on all branches so we always emit lints #149380

Merged

fmease mentioned this pull request Nov 30, 2025

Move even more early buffered lints to dyn lint diagnostics #147634

Merged

JonathanBrouwer force-pushed the cfg_lints2 branch from f5e9e93 to 1b85568 Compare December 4, 2025 19:27

JonathanBrouwer mentioned this pull request Dec 4, 2025

Move attribute lints to rustc_lint #149662

Merged

JonathanBrouwer force-pushed the cfg_lints2 branch from 1b85568 to 557fcb0 Compare December 5, 2025 14:13

JonathanBrouwer added the S-blocked Status: Blocked on something else such as an RFC or other implementation work. label Dec 5, 2025

JonathanBrouwer force-pushed the cfg_lints2 branch from 557fcb0 to 5010285 Compare December 5, 2025 14:15

JonathanBrouwer changed the title ~~Move check-cfg lints to rustc_attr_parsing~~ Emit check-cfg during attribute parsing rather than evaluation Dec 5, 2025

JonathanBrouwer changed the title ~~Emit check-cfg during attribute parsing rather than evaluation~~ Emit check-cfg lints during attribute parsing rather than evaluation Dec 5, 2025

JonathanBrouwer force-pushed the cfg_lints2 branch from 5010285 to b93aaa7 Compare December 5, 2025 14:23

jdonszelmann self-assigned this Dec 5, 2025

Add regression test for nested cfgs

01576a6

JonathanBrouwer marked this pull request as ready for review December 6, 2025 09:39

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Dec 6, 2025

JonathanBrouwer removed the S-blocked Status: Blocked on something else such as an RFC or other implementation work. label Dec 6, 2025

JonathanBrouwer mentioned this pull request Dec 6, 2025

Port doc attributes to new attribute API #149645

Merged

jdonszelmann reviewed Dec 8, 2025

View reviewed changes

tests/ui/macros/cfg.stderr Show resolved Hide resolved

jdonszelmann reviewed Dec 8, 2025

View reviewed changes

rustbot added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Dec 8, 2025

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Dec 8, 2025

bors added S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Dec 8, 2025

Zalathar mentioned this pull request Dec 9, 2025

Rollup of 12 pull requests #149796

Closed

matthiaskrgr mentioned this pull request Dec 9, 2025

Rollup of 11 pull requests #149798

Merged

bors merged commit ece9977 into rust-lang:main Dec 9, 2025
11 checks passed

rustbot added this to the 1.94.0 milestone Dec 9, 2025

Uh oh!

Emit check-cfg lints during attribute parsing rather than evaluation #149215

Emit check-cfg lints during attribute parsing rather than evaluation #149215

Uh oh!

Conversation

JonathanBrouwer commented Nov 22, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

This comment has been minimized.

Urgau commented Nov 22, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

This comment has been minimized.

This comment has been minimized.

JonathanBrouwer commented Nov 25, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

bors commented Nov 28, 2025

Uh oh!

bors commented Dec 5, 2025

Uh oh!

JonathanBrouwer commented Dec 5, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rustbot commented Dec 6, 2025

Uh oh!

JonathanBrouwer commented Dec 6, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

jdonszelmann Dec 8, 2025

Choose a reason for hiding this comment

Uh oh!

jdonszelmann left a comment • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

jdonszelmann Dec 8, 2025

Choose a reason for hiding this comment

Uh oh!

jdonszelmann Dec 8, 2025

Choose a reason for hiding this comment

Uh oh!

JonathanBrouwer Dec 8, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

jdonszelmann Dec 8, 2025

Choose a reason for hiding this comment

Uh oh!

jdonszelmann commented Dec 8, 2025

Uh oh!

rustbot commented Dec 8, 2025

Uh oh!

JonathanBrouwer commented Dec 8, 2025

Uh oh!

jdonszelmann commented Dec 8, 2025

Uh oh!

bors commented Dec 8, 2025

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

Emit `check-cfg` lints during attribute parsing rather than evaluation #149215

Emit `check-cfg` lints during attribute parsing rather than evaluation #149215

JonathanBrouwer commented Nov 22, 2025 •

edited

Loading

Urgau commented Nov 22, 2025 •

edited

Loading

JonathanBrouwer commented Nov 25, 2025 •

edited

Loading

JonathanBrouwer commented Dec 5, 2025 •

edited

Loading

JonathanBrouwer commented Dec 6, 2025 •

edited

Loading

jdonszelmann left a comment •

edited by rustbot

Loading

JonathanBrouwer Dec 8, 2025 •

edited

Loading