Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Don't make tools responsible for checking unknown and renamed lints #80524

Merged
merged 4 commits into from
Jan 17, 2021

Conversation

jyn514
Copy link
Member

@jyn514 jyn514 commented Dec 30, 2020

Previously, clippy (and any other tool emitting lints) had to have their
own separate UNKNOWN_LINTS pass, because the compiler assumed any tool
lint could be valid. Now, as long as any lint starting with the tool
prefix exists, the compiler will warn when an unknown lint is present.

This may interact with the unstable tool_lint feature, which I don't entirely understand, but it will take the burden off those external tools to add their own lint pass, which seems like a step in the right direction to me.

  • Don't mark ineffective_unstable_trait_impl as an internal lint
  • Use clippy's more advanced lint suggestions
  • Deprecate the UNKNOWN_CLIPPY_LINTS pass (and make it a no-op)
  • Say 'unknown lint clippy::x' instead of 'unknown lint x'

This is tested by existing clippy tests. When #80527 merges, it will also be tested in rustdoc tests. AFAIK there is no way to test this with rustc directly.

@jyn514 jyn514 added A-lint Area: Lints (warnings about flaws in source code) such as unused_mut. T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. T-dev-tools Relevant to the dev-tools subteam, which will review and decide on the PR/issue. labels Dec 30, 2020
@rust-highfive
Copy link
Collaborator

r? @matthewjasper

(rust-highfive has picked a reviewer for you, use r? to override)

@rust-highfive rust-highfive added the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Dec 30, 2020
@rust-log-analyzer
Copy link
Collaborator

The job x86_64-gnu-llvm-9 failed! Check out the build log: (web) (plain)

Click to see the possible cause of the failure (guessed by this bot)
   Compiling miniz_oxide v0.4.0
   Compiling object v0.22.0
   Compiling hashbrown v0.9.0
   Compiling addr2line v0.14.0
error: unknown lint: `rustc::existing_doc_keyword`
    |
    |
216 | #![deny(rustc::existing_doc_keyword)]
    |
    |
    = note: `-D unknown-lints` implied by `-D warnings`
error: aborting due to previous error

error: could not compile `std`

@flip1995
Copy link
Member

Something that is also important to mention:

tool_lints are semi-stable. So you can use them without adding a feature, but only explicitly allowed tools (defined in is_known_lint_tool can implement them. This also means, that this is a semi-breaking change. The reference says:

If a lint attribute, such as allow, references a nonexistent tool lint, the compiler will not warn about the nonexistent lint until you use the tool.

So it doesn't mention, that the handling of missing lints has to be implemented by the tool. With this and the fact, that we know all the tools that implement tool lints, this change should be fine.

@flip1995
Copy link
Member

So the CI error is because internal lints are only enabled in the Mode::Rustc, not when compiling std.

This makes the rustc::existing_doc_keywords lint useless, since it is only enabled in std. But enabling internal lints in std results in many warnings / allows of the rustc::default_hash_types lint.

@flip1995
Copy link
Member

flip1995 commented Dec 31, 2020

Oh well, this uncovered a whole new set of problems, that were silently ignored before: rustc::internal lints are also tool lints, but rustc doesn't implement a unknown_rustc_lints lint. So writing deny(rustc::internal) where internal lints aren't enabled (std or even rustc_macros for some reason), was just silently ignored. This PR fixes that. But now when a tool like Clippy is compiled with ./x.py, which also uses internal rustc lints it fails, because they are not enabled by tools. They can be enabled by adding Mode::ToolsRustc here:

rust/src/bootstrap/builder.rs

Lines 1268 to 1271 in 8b002d5

if mode == Mode::Rustc {
rustflags.arg("-Zunstable-options");
rustflags.arg("-Wrustc::internal");
}

But this results in build failures for other tools, that don't adhere to the rustc internal lints, like rustdoc.


Well, this is something for the new year. Have a great new years eve everyone 🎉

@jyn514
Copy link
Member Author

jyn514 commented Jan 1, 2021

But enabling internal lints in std results in many warnings / allows of the rustc::default_hash_types lint.

FWIW, adding -Wrustc::internal -Arustc::default-hash-types gives no warnings. So in theory we could special case std like that. I think existing_doc_keywords is niche enough it shouldn't be a big deal to remove it though (cc @GuillaumeGomez just in case - I kind of wish you'd pinged me on #79541, since I suggested it in #79464 (comment)).

But now when a tool like Clippy is compiled with ./x.py, which also uses internal rustc lints it fails, because they are not enabled by tools.

I don't quite understand this - I think to test it I would need x.py build --stage 2, which takes a while. Can you give an example of a lint that fails?

But this results in build failures for other tools, that don't adhere to the rustc internal lints, like rustdoc.

I'm happy to fix the lints for rustdoc, I would prefer to have them enabled anyway.

@GuillaumeGomez
Copy link
Member

existing_doc_keywords is aimed to rustc only, not rustdoc or anything else (which means you can still add doc on keywords that don't exist outside of rustc). The goal is to prevent rustc to have doc on keywords which don't exist while allowing other crates to (since it's aimed for proc-macros). So please don't remove it?

@jyn514
Copy link
Member Author

jyn514 commented Jan 1, 2021

@GuillaumeGomez it's aimed at libstd, which is the whole point - your PR did nothing because rustc lints aren't applied to libstd currently. I'm ok with denying that lint specifically for std though.

@GuillaumeGomez
Copy link
Member

libstd and libcore to be exact. If it's denied for both of them then it's all good for me.

@flip1995
Copy link
Member

flip1995 commented Jan 2, 2021

I don't quite understand this - I think to test it I would need x.py build --stage 2, which takes a while. Can you give an example of a lint that fails?

So Clippy has #[deny(rustc::internal)] in the lib.rs. But since for Mode::ToolsRustc internal lints aren't enabled currently, the compilation will fail (./x.py test src/tools/clippy). Your changes in #80573 should fix this though.

The question then is if other tools want those internal lints.

Copy link
Contributor

@matthewjasper matthewjasper left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

r=me with nit addressed (and CI passing)

compiler/rustc_lint/src/context.rs Outdated Show resolved Hide resolved
@jyn514
Copy link
Member Author

jyn514 commented Jan 15, 2021

I don't quite understand this - I think to test it I would need x.py build --stage 2, which takes a while. Can you give an example of a lint that fails?

So Clippy has #[deny(rustc::internal)] in the lib.rs. But since for Mode::ToolsRustc internal lints aren't enabled currently, the compilation will fail (./x.py test src/tools/clippy). Your changes in #80573 should fix this though.

The question then is if other tools want those internal lints.

So it turns out the issue is that rustc::ineffective_unstable_trait_impl is on unconditionally, because it's not actually an internal lint, it's a lint for staged_api. So even when internal lints are disabled, my new logic sees rustc:: and thinks all rustc:: lints are known.

FYI @m-ou-se I'm going to change this to just be ineffective_unstable_trait_impl, does that sound good? It will still be emitted only for staged_api.

There's a second question here which is why all rustc:: lints are not known when -W rustc::internal isn't passed, but I'm planning to leave that mystery to another day if possible.

@jyn514
Copy link
Member Author

jyn514 commented Jan 15, 2021

#76538 (comment):

Added it to the builtin lints as a rustc:: tool lint, just like the internal lints. That way it doesn't show up in -W help and an #[allow()] attribute for it doesn't trigger a unknown lint error (which happens when bootstrapping rustc+std, using the current beta).

Normally the fix for this is to add cfg(not(bootstrap)) to the allow. I don't have a fix for it showing up in -W help, but since people besides the standard library can use staged_api on nightly, I'm ok with adding it there.

@jyn514 jyn514 force-pushed the unknown-tool-lints branch 2 times, most recently from 4d7dcbe to 5db673c Compare January 15, 2021 20:28
@m-ou-se
Copy link
Member

m-ou-se commented Jan 15, 2021

@jyn514 Yeah, that lint is a bit of a weird one. It's mostly/only meant for std but still part of an unstable feature usable outside of it. Rustc doesn't really have a concept of "lint that's part of an unstable/private feature". Putting it under rustc:: was a hack. It seems fine to me to change that.

@flip1995
Copy link
Member

flip1995 commented Jan 15, 2021

lints are not known when -W rustc::internal isn't passed, but I'm planning to leave that mystery to another day if possible.

Oh it isn't because of the -W flag, it's the -Zunstable-options flag, that gates the registration of internal lints. If internal lints were registered unconditionally, it could have a negative impact on performance.

We chose to (ab)use the unstable-options option for internal lints, because we didn't think it was worth to add an extra option for it (or smth like this).


Clippy sets this flag in .cargo/config.

@rust-log-analyzer

This comment has been minimized.

@jyn514
Copy link
Member Author

jyn514 commented Jan 15, 2021

Ahh, that makes sense.

So to summary, this does the following things:

  1. Rename rustc::ineffective_unstable_trait_impl to ineffective_unstable_trait_impl.
  2. Move checking of unknown and renamed lints to the lint machinery, so that each new tool doesn't have to add its own pass checking it. Assume that if any tool:: lint is present, all lints for tool are known. In particular this is necessary to avoid warning on #[deny(rustc::internal)] when -Z unstable-options isn't passed, and in general we have to assume at some point that the lints are known or we can't warn about unknown lints.
  3. Fix 'unknown lint x' to say 'unknown lint tool::x' instead.

@flip1995 does all of that sound good? I think for 2 you'll want to deprecate/remove the clippy pass that currently checks for unknown lints - is that simple to do? Happy to do it myself if you let me know how.

@flip1995
Copy link
Member

Yes, that sounds good to me.

If everything in this PR works as expected, every error message in the test for unknown_clippy_lints should be duplicated. If that is the case, I will deprecate the Clippy lint for you (I can either post a patch file or you can give me push access to your fork, so I can push directly to this branch).

@rust-log-analyzer

This comment has been minimized.

@rust-log-analyzer

This comment has been minimized.

@rust-log-analyzer

This comment has been minimized.

It's not an internal lint:
- It's not in the rustc::internal lint group
- It's on unconditionally, because it actually lints `staged_api`, not
  the compiler

This fixes a bug where `#[deny(rustc::internal)]` would warn that
`rustc::internal` was an unknown lint.
compiler/rustc_lint/src/context.rs Outdated Show resolved Hide resolved
src/tools/clippy/clippy_lints/src/deprecated_lints.rs Outdated Show resolved Hide resolved
src/tools/clippy/clippy_lints/src/lib.rs Outdated Show resolved Hide resolved
Comment on lines +1 to +7
error: unknown lint: `clippy::All`
--> $DIR/unknown_clippy_lints.rs:5:10
|
LL | #![allow(clippy::All)]
| ^^^^^^^^^^^ help: did you mean: `clippy::all`
|
= note: `-D unknown-lints` implied by `-D warnings`
Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Hmm, I'm curious to know why this changed. This is a bug fix, right, it should have given a warning before?

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Hm? The order is just different. And it has the default weirdness with linting attributes, that it is duplicated for the first inner attribute.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Ah ok, I misread. No need to change anything.

And it has the default weirdness with linting attributes, that it is duplicated for the first inner attribute.

Does this have an issue open? It seems worth fixing.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I don't know of an issue for this.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'm not sure if it's just in a different order; at the bottom it shows that it emitted 8 errors previously and now it emits 9.

This copies the unknown_lints code clippy uses for its
unknown_clippy_lints lint to rustc. The unknown_clippy_lints code is
more advanced, because it doesn't suggest renamed or removed lints and
correctly suggest lower casing lints.
This is now handled by unknown_lints
@jyn514
Copy link
Member Author

jyn514 commented Jan 17, 2021

Alright, I think this is good to go!

@bors r=flip1995,matthewjasper

@bors
Copy link
Contributor

bors commented Jan 17, 2021

📌 Commit 13728b8 has been approved by flip1995,matthewjasper

@bors 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 Jan 17, 2021
@bors
Copy link
Contributor

bors commented Jan 17, 2021

⌛ Testing commit 13728b8 with merge 1f0fc02...

@bors
Copy link
Contributor

bors commented Jan 17, 2021

☀️ Test successful - checks-actions
Approved by: flip1995,matthewjasper
Pushing 1f0fc02 to master...

@bors bors added the merged-by-bors This PR was explicitly merged by bors. label Jan 17, 2021
@bors bors merged commit 1f0fc02 into rust-lang:master Jan 17, 2021
@rustbot rustbot added this to the 1.51.0 milestone Jan 17, 2021
@jyn514 jyn514 deleted the unknown-tool-lints branch January 17, 2021 20:49
teor2345 added a commit to teor2345/zebra that referenced this pull request Jan 18, 2021
The clippy unknown lints attribute was deprecated in
nightly in rust-lang/rust#80524. The old lint name now produces a
warning.

Since we're using `allow(unknown_lints)` to suppress warnings, we need to
adopt the canonical name, so we can continue to build without warnings.
teor2345 added a commit to teor2345/zebra that referenced this pull request Jan 18, 2021
The clippy unknown lints attribute was deprecated in
nightly in rust-lang/rust#80524. The old lint name now produces a
warning.

Since we're using `allow(unknown_lints)` to suppress warnings, we need to
adopt the canonical name, so we can continue to build without warnings.

This change was automatically generated using the following script:

```sh
sd clippy::unknown_clippy_lints unknown_lints \
   $(fd . --extension rs zebra*)`
```
teor2345 added a commit to teor2345/zebra that referenced this pull request Jan 18, 2021
The clippy unknown lints attribute was deprecated in
nightly in rust-lang/rust#80524. The old lint name now produces a
warning.

Since we're using `allow(unknown_lints)` to suppress warnings, we need to
add the canonical name, so we can continue to build without warnings on
nightly.

But we also need to keep the old name, so we can continue to build
without warnings on stable.

And therefore, we also need to disable the "removed lints" warning,
otherwise we'll get warnings about the old name on nightly.

We'll need to keep this transitional clippy config until rustc 1.51 is
stable.
teor2345 added a commit to teor2345/zebra that referenced this pull request Jan 18, 2021
The clippy unknown lints attribute was deprecated in
nightly in rust-lang/rust#80524. The old lint name now produces a
warning.

Since we're using `allow(unknown_lints)` to suppress warnings, we need to
add the canonical name, so we can continue to build without warnings on
nightly.

But we also need to keep the old name, so we can continue to build
without warnings on stable.

And therefore, we also need to disable the "removed lints" warning,
otherwise we'll get warnings about the old name on nightly.

We'll need to keep this transitional clippy config until rustc 1.51 is
stable.
teor2345 added a commit to teor2345/zebra that referenced this pull request Jan 19, 2021
The clippy unknown lints attribute was deprecated in
nightly in rust-lang/rust#80524. The old lint name now produces a
warning.

Since we're using `allow(unknown_lints)` to suppress warnings, we need to
add the canonical name, so we can continue to build without warnings on
nightly.

But we also need to keep the old name, so we can continue to build
without warnings on stable.

And therefore, we also need to disable the "removed lints" warning,
otherwise we'll get warnings about the old name on nightly.

We'll need to keep this transitional clippy config until rustc 1.51 is
stable.
dconnolly pushed a commit to ZcashFoundation/zebra that referenced this pull request Jan 19, 2021
The clippy unknown lints attribute was deprecated in
nightly in rust-lang/rust#80524. The old lint name now produces a
warning.

Since we're using `allow(unknown_lints)` to suppress warnings, we need to
add the canonical name, so we can continue to build without warnings on
nightly.

But we also need to keep the old name, so we can continue to build
without warnings on stable.

And therefore, we also need to disable the "removed lints" warning,
otherwise we'll get warnings about the old name on nightly.

We'll need to keep this transitional clippy config until rustc 1.51 is
stable.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
A-lint Area: Lints (warnings about flaws in source code) such as unused_mut. merged-by-bors This PR was explicitly merged by bors. S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. T-dev-tools Relevant to the dev-tools subteam, which will review and decide on the PR/issue.
Projects
None yet
Development

Successfully merging this pull request may close these issues.

10 participants