Update E0138 to new format
Part of #35233Fix#35510
r? @jonathandturner
Question: How can I only underline the function name ? I have observed the debug output and the struct of item, but I can't find the `Span` for function name. Should I modify the struct I get to save function name's position or there is another way to get it ? (I can only find `Span`s for function attributes, inputs, outputs, blocks)
Implemented a smarter TokenStream concatenation system
The new algorithm performs 'aggressive compacting' during concatenation as follows:
- If the nodes' combined total total length is less than 32, we copy both of
them into a new vector and build a new leaf node.
- If one node is an internal node and the other is a 'small' leaf (length<32),
we recur down the internal node on the appropriate side.
- Otherwise, we construct a new internal node that points to them as left and
right.
This should produce notably better behavior than the current concatenation implementation.
std: Optimize panic::catch_unwind slightly
The previous implementation of this function was overly conservative with
liberal usage of `Option` and `.unwrap()` which in theory never triggers. This
commit essentially removes the `Option`s in favor of unsafe implementations,
improving the code generation of the fast path for LLVM to see through what's
happening more clearly.
cc #34727
macros: Make metavariables hygienic
This PR makes metavariables hygienic. For example, consider:
```rust
macro_rules! foo {
($x:tt) => { // Suppose that this token tree argument is always a metavariable.
macro_rules! bar { ($x:expr, $y:expr) => { ($x, $y) } }
}
}
fn main() {
foo!($z); // This currently compiles.
foo!($y); // This is an error today but compiles after this PR.
}
```
Today, the `macro_rules! bar { ... }` definition is only valid when the metavariable passed to `foo` is not `$y` (since it unhygienically conflicts with the `$y` in the definition of `bar`) or `$x` (c.f. #35450).
After this PR, the definition of `bar` is always valid (and `bar!(a, b)` always expands to `(a, b)` as expected).
This can break code that was allowed in #34925 (landed two weeks ago). For example,
```rust
macro_rules! outer {
($t:tt) => {
macro_rules! inner { ($i:item) => { $t } }
}
}
outer!($i); // This `$i` should not interact with the `$i` in the definition of `inner!`.
inner!(fn main() {}); // After this PR, this is an error ("unknown macro variable `i`").
```
Due to the severe limitations on nested `macro_rules!` before #34925, this is not a breaking change for stable/beta.
Fixes#35450.
r? @nrc
Add --test-threads option to test binaries
This change allows parallelism of test runs to be specified by a
command line flag names --test-threads in addition to the existing
environment variable RUST_TEST_THREADS. Fixes#25636.
[MIR] Add explicit SetDiscriminant StatementKind for deaggregating enums
cc #35186
To deaggregate enums, we need to be able to explicitly set the discriminant. This PR implements a new StatementKind that does that.
I think some of the places that have `panics!` now could maybe do something smarter.
Implement `impl Trait` in return type position by anonymization.
This is the first step towards implementing `impl Trait` (cc #34511).
`impl Trait` types are only allowed in function and inherent method return types, and capture all named lifetime and type parameters, being invariant over them.
No lifetimes that are not explicitly named lifetime parameters are allowed to escape from the function body.
The exposed traits are only those listed explicitly, i.e. `Foo` and `Clone` in `impl Foo + Clone`, with the exception of "auto traits" (like `Send` or `Sync`) which "leak" the actual contents.
The implementation strategy is anonymization, i.e.:
```rust
fn foo<T>(xs: Vec<T>) -> impl Iterator<Item=impl FnOnce() -> T> {
xs.into_iter().map(|x| || x)
}
// is represented as:
type A</*invariant over*/ T> where A<T>: Iterator<Item=B<T>>;
type B</*invariant over*/ T> where B<T>: FnOnce() -> T;
fn foo<T>(xs: Vec<T>) -> A<T> {
xs.into_iter().map(|x| || x): $0 where $0: Iterator<Item=$1>, $1: FnOnce() -> T
}
```
`$0` and `$1` are resolved (to `iter::Map<vec::Iter<T>, closure>` and the closure, respectively) and assigned to `A` and `B`, after checking the body of `foo`. `A` and `B` are *never* resolved for user-facing type equality (typeck), but always for the low-level representation and specialization (trans).
The "auto traits" exception is implemented by collecting bounds like `impl Trait: Send` that have failed for the obscure `impl Trait` type (i.e. `A` or `B` above), pretending they succeeded within the function and trying them again after type-checking the whole crate, by replacing `impl Trait` with the real type.
While passing around values which have explicit lifetime parameters (of the function with `-> impl Trait`) in their type *should* work, regionck appears to assign inference variables in *way* too many cases, and never properly resolving them to either explicit lifetime parameters, or `'static`.
We might not be able to handle lifetime parameters in `impl Trait` without changes to lifetime inference, but type parameters can have arbitrary lifetimes in them from the caller, so most type-generic usecases (or not generic at all) should not run into this problem.
cc @rust-lang/lang
The previous implementation of this function was overly conservative with
liberal usage of `Option` and `.unwrap()` which in theory never triggers. This
commit essentially removes the `Option`s in favor of unsafe implementations,
improving the code generation of the fast path for LLVM to see through what's
happening more clearly.
cc #34727
E0072 update error format
Part of #35233Fixes#35506
r? @jonathandturner
The bonus for this issue currently seems to be impossible to do reliably, as the compiler seems to lack span information for item names alone, like `Foo` in `struct Foo { ... }`. It would be possible to hack something together by computing span offsets, but that seems like a solution that would be begging for trouble.
A proper solution to this would, of course, be to add span information to the right place (seems to be `rustc::hir::Item::name` but I may be wrong).
