-
Notifications
You must be signed in to change notification settings - Fork 12.7k
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
Temporarily prohibit proc macro attributes placed after derives #54277
Conversation
... and also proc macro attributes used together with test/bench.
@bors: r+ p=1 |
📌 Commit 229df02 has been approved by |
Temporarily prohibit proc macro attributes placed after derives ... and also proc macro attributes used together with `#[test]`/`#[bench]`. Addresses item 6 from #50911 (comment). The end goal is straightforward predictable left-to-right expansion order for attributes. Right now derives are expanded last regardless of their relative ordering with macro attributes and right now it's simpler to temporarily prohibit macro attributes placed after derives than changing the expansion order. I'm not sure whether the new beta is already released or not, but if it's released, then this patch needs to be backported, so the solution needs to be minimal. How to fix broken code (derives): - Move macro attributes above derives. This won't change expansion order, they are expanded before derives anyway. Using attribute macros on same items with `#[test]` and `#[bench]` is prohibited for similar expansion order reasons, but this one is going to be reverted much sooner than restrictions on derives. How to fix broken code (test/bench): - Enable `#![feature(plugin)]` (don't ask why). r? @ghost
☀️ Test successful - status-appveyor, status-travis |
The restrictions were introduced in rust-lang#54277 and no longer necessary now because legacy plugins are now expanded in usual left-to-right order
fn deref_mut(&mut self) -> &mut T { | ||
&mut self.ptr | ||
} | ||
} |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This was intentionally not implemented, but I'm considering ways in which we can make ASTs fully immutable (and interned, I suppose).
`P<T>` has implemented `DerefMut` since rust-lang#54277. While this was lamented at the time [1], rustc now relies on it extensively via the many implementors of MutVisitor. Updates the docs to reflect that `P<T>` is fundamentally mutable, and a few other cleanups to make them nicer to browse. [1]: rust-lang#54277 (comment) [2]: https://doc.rust-lang.org/1.77.1/nightly-rustc/rustc_ast/mut_visit/trait.MutVisitor.html#implementors
`P<T>` has implemented `DerefMut` since rust-lang#54277. While this was lamented at the time [1], rustc now relies on it extensively via the many implementors of MutVisitor [2]. Updates the docs to reflect that `P<T>` is fundamentally mutable, and a few other cleanups to make them nicer to browse. [1]: rust-lang#54277 (comment) [2]: https://doc.rust-lang.org/1.77.1/nightly-rustc/rustc_ast/mut_visit/trait.MutVisitor.html#implementors
`P<T>` has implemented `DerefMut` since rust-lang#54277. While this was lamented at the time [1], rustc now relies on it extensively via the many implementors of MutVisitor [2]. Updates the docs to reflect that `P<T>` is fundamentally mutable, and a few other cleanups to make them nicer to browse. [1]: rust-lang#54277 (comment) [2]: https://doc.rust-lang.org/1.77.1/nightly-rustc/rustc_ast/mut_visit/trait.MutVisitor.html#implementors
rustc_ast: Update `P<T>` docs to reflect mutable status. `P<T>` has implemented `DerefMut` since rust-lang#54277. While this was lamented at the time [1], rustc now relies on it extensively via the many implementors of MutVisitor [2]. Updates the docs to reflect that `P<T>` is fundamentally mutable, and a few other cleanups to make them nicer to browse. [1]: rust-lang#54277 (comment) [2]: https://doc.rust-lang.org/1.77.1/nightly-rustc/rustc_ast/mut_visit/trait.MutVisitor.html#implementors r? `@Nilstrieb`
Rollup merge of rust-lang#123393 - aDotInTheVoid:ast-p-docs, r=Nilstrieb rustc_ast: Update `P<T>` docs to reflect mutable status. `P<T>` has implemented `DerefMut` since rust-lang#54277. While this was lamented at the time [1], rustc now relies on it extensively via the many implementors of MutVisitor [2]. Updates the docs to reflect that `P<T>` is fundamentally mutable, and a few other cleanups to make them nicer to browse. [1]: rust-lang#54277 (comment) [2]: https://doc.rust-lang.org/1.77.1/nightly-rustc/rustc_ast/mut_visit/trait.MutVisitor.html#implementors r? `@Nilstrieb`
... and also proc macro attributes used together with
#[test]
/#[bench]
.Addresses item 6 from #50911 (comment).
The end goal is straightforward predictable left-to-right expansion order for attributes.
Right now derives are expanded last regardless of their relative ordering with macro attributes and right now it's simpler to temporarily prohibit macro attributes placed after derives than changing the expansion order.
I'm not sure whether the new beta is already released or not, but if it's released, then this patch needs to be backported, so the solution needs to be minimal.
How to fix broken code (derives):
Using attribute macros on same items with
#[test]
and#[bench]
is prohibited for similar expansion order reasons, but this one is going to be reverted much sooner than restrictions on derives (UPDATE: reverted in #54336).How to fix broken code (test/bench):
#![feature(plugin)]
(don't ask why).r? @ghost