-
Notifications
You must be signed in to change notification settings - Fork 1.6k
RFC: Make the as
keyword consider Into
Trait implementations
#2308
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
Closed
Kerollmops
wants to merge
1
commit into
rust-lang:master
from
Kerollmops:as-keyword-consider-into-trait
Closed
Changes from all commits
Commits
File filter
Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
There are no files selected for viewing
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,130 @@ | ||
- Feature Name: `as_keyword_consider_into_trait` | ||
- Start Date: 2018-01-21 | ||
- RFC PR: | ||
- Rust Issue: | ||
|
||
# Summary | ||
[summary]: #summary | ||
|
||
Permit to use the common `as` keyword with any type that implement the `Into` Trait, | ||
allowing explicit conversions whose primitives already benefit, more visible than simple function calls. | ||
|
||
# Motivation | ||
[motivation]: #motivation | ||
|
||
Many operators are allowed to be implemented permitting a much better programming flow: | ||
- the `+`, `-`, `*`, `\` operators are the result of the `Add`, `Sub`, `Mul`, `Div` Traits implementations. | ||
- the `+=`, `-=`, `*=`, `\=` operators are the result of the `AddAssign`, `SubAssign`, `MulAssign`, `DivAssign` Traits implementations. | ||
- the `container[index]` operator is the result of the `Index` and `IndexMut` Traits implementations. | ||
- and many others can be found on the [`std::ops`](https://doc.rust-lang.org/std/ops/) documentation page. | ||
|
||
The `as` operator is actually reserved to primitives only but these given primitives already implement `Into` Traits for all possible conversions: | ||
- the `u32` implement `Into<u64>` in the form of an [`impl From<u32> for u64`](https://doc.rust-lang.org/std/primitive.u64.html#impl-From%3Cu32%3E). | ||
- the `u8` implement `Into<u16>` in the form of an [`impl From<u8> for u16`](https://doc.rust-lang.org/std/primitive.u16.html#impl-From%3Cu8%3E). | ||
|
||
All of these conversions can be found in the [`libcore/num/mod.rs` file](https://github.com/rust-lang/rust/blob/d9d5c667d819ce400fc7adb09dcd6482b0aa519e/src/libcore/num/mod.rs#L3343-L3400). | ||
|
||
Some special primitives like `usize` and `isize` doesn't implement many `From` Traits because these are [the representation of the address register](https://users.rust-lang.org/t/cant-convert-usize-to-u64/6243) and it can break compilation on some architectures. | ||
|
||
Adding this feature to the compiler will probably unify the [actual conversion error detections](https://users.rust-lang.org/t/cant-convert-usize-to-u64/6243/8) of the `as` keyword and the `into` method. | ||
|
||
# Guide-level explanation | ||
[guide-level-explanation]: #guide-level-explanation | ||
|
||
The actual design is not far from what we already know of the `as` keyword. | ||
It doesn't add new grammar to the language, it add more freedom to the actual syntax: | ||
|
||
```rust | ||
struct Foo(i32); | ||
|
||
struct Bar(i32); | ||
|
||
impl Into<Bar> for Foo { | ||
fn into(self) -> Bar { | ||
Bar(self.0) | ||
} | ||
} | ||
|
||
// this old syntax | ||
let x = Foo(42); | ||
let y: Bar = x.into(); | ||
|
||
// is equivalent to the new one | ||
let x = Foo(42); | ||
let y = x as Bar; | ||
``` | ||
|
||
The actual syntax **is not deprecated in any way**, it has to mimic the `Add::add` or the `Deref::deref` methods and permit the user to directly call these methods or use the operators as he likes. | ||
|
||
# Reference-level explanation | ||
[reference-level-explanation]: #reference-level-explanation | ||
|
||
The `as` keyword is use to reflect the actual `Into::into` method call, if the `Into` Trait is implemented for the type to be converted, the compiler needs to emit an error informing the user that the conversion is not possible. | ||
|
||
The actual `as` keyword is reserved for primitives to convert into other primitives but it shouldn't be a breaking change to change the behavior and to work with any type implementing `Into` because the primitives already implement these Traits. | ||
|
||
The internal desing could be a simple syntax sugar to something like that: | ||
|
||
```rust | ||
// this actual syntax | ||
let x = Foo(42); | ||
let y = x as Bar; | ||
|
||
// is a syntax sugar for | ||
let x = Foo(42); | ||
let y = Into::<Bar>::into(x); | ||
``` | ||
|
||
# Drawbacks | ||
[drawbacks]: #drawbacks | ||
|
||
The `Into` Trait is not in the `std::ops` page and there is a reason why, | ||
it's probably not considered overridable to be used on an operator/keyword like the `as` one. | ||
|
||
The `as` keyword differ from the `Into::into` named method, it can be confusing. | ||
|
||
# Rationale and alternatives | ||
[alternatives]: #alternatives | ||
|
||
With this design we doesn't introduce a new syntax, we just enlarge its possibilities. | ||
|
||
The `as` keyword can be confusing with the `Into::into` named method, | ||
we can add a new `into` keyword and/or deprecate the `as` one. | ||
|
||
# Unresolved questions | ||
[unresolved]: #unresolved-questions | ||
|
||
What about `Copy` on the actual types ? Can it be a problem ? | ||
|
||
Is the actual restriction on primitives due to the fact that they implement `Copy` and the conversion is low performance impact ? | ||
|
||
How can we handle generic type conversions ? | ||
Do we need to disallow them and only accpet non-generic ones ? | ||
|
||
```rust | ||
struct Foo<T>(T); | ||
|
||
struct Bar<T>(T); | ||
|
||
impl Into<Bar<i64>> for Foo<i32> { | ||
fn into(self) -> Bar<i64> { | ||
Bar(self.0 as i64) | ||
} | ||
} | ||
|
||
// this old syntax | ||
let x = Foo(42_i32); | ||
let y: Bar<i64> = x.into(); | ||
|
||
// is equivalent to the new one | ||
let x = Foo(42); | ||
let y = x as Bar<i64>; | ||
|
||
// we can let the type inference guess the generic type | ||
let x = Foo(42_i32); | ||
let y: Bar<_> = x.into(); | ||
|
||
// equivalent to | ||
let x = Foo(42); | ||
let y = x as Bar<_>; | ||
``` |
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Uh oh!
There was an error while loading. Please reload this page.
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 won't work since
as
is not equivalent to a call to.into()
- for example,as
can convert fromf32
tou32
, while.into()
can't (this is a lossy conversion,so)..into()
shouldn't work hereUh oh!
There was an error while loading. Please reload this page.
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.
So what is the rule of
as
conversion andinto
conversion ?I think this is related to this thread about lossy conversions between
as
andinto
(I linked it in the Rendered).