Clippy lints / Suspicious
Clippy lint · Suspicious · checked with Clippy 0.1.99
clippy::type_id_on_box: calling `.type_id()` on `Box<dyn Any>`
What it meansLooks for calls to .type_id() on a Box<dyn _>. It is in the suspicious group and warns by default.
What you see
warning: calling `.type_id()` on `Box<dyn Any>` --> main.rs:6:12 | 6 | assert_eq!(any_box.type_id(), TypeId::of::<i32>()); // ⚠️ this fails! | -------^^^^^^^^^^ | | | help: consider dereferencing first: `(*any_box)` | = note: this returns the type id of the literal type `Box<_>` instead of the type id of the boxed value, which is most likely not what you want = note: if this is intentional, use `TypeId::of::<Box<dyn Any>>()` instead, which makes it more clear = help: for further information visit https://rust-lang.github.io/rust-clippy/stable/index.html#type_id_on_box = note: `#[warn(clippy::type_id_on_box)]` on by default
Real output of Clippy 0.1.99 (Rust 1.99) on the example below. Names and line numbers in yours will differ.
Does it matter?
This almost certainly does not do what the user expects and can lead to subtle bugs. Calling .type_id() on a Box<dyn Trait> returns a fixed TypeId of the Box itself, rather than returning the TypeId of the underlying type behind the trait object. For Box<dyn Any> specifically (and trait objects that have Any as its supertrait), this lint will provide a suggestion, which is to dereference the receiver explicitly to go from Box<dyn Any> to dyn Any. This makes sure that .type_id() resolves to a dynamic call on the trait object and not on the box. If the fixed TypeId of the Box is the intended behavior, it’s better to be explicit about it and write TypeId::of::<Box<dyn Trait>>(): this makes it clear that a fixed TypeId is returned and not the TypeId of the implementor.
Code that is most likely wrong or useless. Warns by default.
How to fix it
Clippy does not fix this one automatically (its suggestion may change behaviour, so it only shows it). Change the code by hand:
Flagged
use std::any::{Any, TypeId};
let any_box: Box<dyn Any> = Box::new(42_i32);
assert_eq!(any_box.type_id(), TypeId::of::<i32>()); // ⚠️ this fails!
Write this instead
use std::any::{Any, TypeId};
let any_box: Box<dyn Any> = Box::new(42_i32);
assert_eq!((*any_box).type_id(), TypeId::of::<i32>());
// ^ dereference first, to call `type_id` on `dyn Any`
Turn it off, or on
Allow it where it does not fit your code, as narrowly as you can:
On one item
#[allow(clippy::type_id_on_box)]
fn my_function() { /* ... */ }Whole crate (Cargo.toml)
[lints.clippy]
type_id_on_box = "allow"One run
cargo clippy -- -A clippy::type-id-on-boxThe [lints] table needs Rust 1.74 or newer. In a workspace, put it under [workspace.lints.clippy] and add [lints] workspace = true to each member.
Clippy's own documentation
What it does
Looks for calls to .type_id() on a Box<dyn _>.
Why is this bad?
This almost certainly does not do what the user expects and can lead to subtle bugs.
Calling .type_id() on a Box<dyn Trait> returns a fixed TypeId of the Box itself,
rather than returning the TypeId of the underlying type behind the trait object.
For Box<dyn Any> specifically (and trait objects that have Any as its supertrait),
this lint will provide a suggestion, which is to dereference the receiver explicitly
to go from Box<dyn Any> to dyn Any.
This makes sure that .type_id() resolves to a dynamic call on the trait object
and not on the box.
If the fixed TypeId of the Box is the intended behavior, it’s better to be explicit about it
and write TypeId::of::<Box<dyn Trait>>():
this makes it clear that a fixed TypeId is returned and not the TypeId of the implementor.
Example
use std::any::{Any, TypeId};
let any_box: Box<dyn Any> = Box::new(42_i32);
assert_eq!(any_box.type_id(), TypeId::of::<i32>()); // ⚠️ this fails!
Use instead:
use std::any::{Any, TypeId};
let any_box: Box<dyn Any> = Box::new(42_i32);
assert_eq!((*any_box).type_id(), TypeId::of::<i32>());
// ^ dereference first, to call `type_id` on `dyn Any`
From the Clippy lint list (0.1.99), part of the Rust project, MIT or Apache-2.0. Official entry.
Related lints
- suspicious_unary_op_formattingby not having a space between `&&` and `!` it looks like `&&!` is a single operator
- swap_ptr_to_refcall to `core::mem::swap` with a parameter derived from a raw pointer
- test_attr_in_doctestunit tests in doctest are not executed
- unconditional_recursionfunction cannot return without recursing
- unnecessary_clippy_cfgno need to put clippy lints behind a `clippy` cfg
- unnecessary_option_map_or_elseunused "map closure" when calling `Option::map_or_else` value
All 847 Clippy lints · Rust compiler errors · Cargo.toml checker
Up next
Related lintclippy::suspicious_unary_op_formattingby not having a space between `&&` and `!` it looks like `&&!` is a single operatorRead it next →Did this answer your question?