As described in #57374, NLL currently produces unhelpful higher-ranked
trait bound (HRTB) errors when '-Zno-leak-check' is enabled.
This PR tackles one half of this issue - making the error message point
at the proper span. The error message itself is still the very generic
"higher-ranked subtype error", but this can be improved in a follow-up
PR.
The root cause of the bad spans lies in how NLL attempts to compute the
'blamed' region, for which it will retrieve a span for.
Consider the following code, which (correctly) does not compile:
```rust
let my_val: u8 = 25;
let a: &u8 = &my_val;
let b = a;
let c = b;
let d: &'static u8 = c;
```
This will cause NLL to generate the following subtype constraints:
d :< c
c :< b
b <: a
Since normal Rust lifetimes are covariant, this results in the following
region constraints (I'm using 'd to denote the lifetime of 'd',
'c to denote the lifetime of 'c, etc.):
'c: 'd
'b: 'c
'a: 'b
From this, we can derive that 'a: 'd holds, which implies that 'a: 'static
must hold. However, this is not the case, since 'a refers to 'my_val',
which does not outlive the current function.
When NLL attempts to infer regions for this code, it will see that the
region 'a has grown 'too large' - it will be inferred to outlive
'static, despite the fact that is not declared as outliving 'static
We can find the region responsible, 'd, by starting at the *end* of
the 'constraint chain' we generated above. This works because for normal
(non-higher-ranked) lifetimes, we generally build up a 'chain' of
lifetime constraints *away* from the original variable/lifetime.
That is, our original lifetime 'a is required to outlive progressively
more regions. If it ends up living for too long, we can look at the
'end' of this chain to determine the 'most recent' usage that caused
the lifetime to grow too large.
However, this logic does not work correctly when higher-ranked trait
bounds (HRTBs) come into play. This is because HRTBs have
*contravariance* with respect to their bound regions. For example,
this code snippet compiles:
```rust
let a: for<'a> fn(&'a ()) = |_| {};
let b: fn(&'static ()) = a;
```
Here, we require that 'a' is a subtype of 'b'. Because of
contravariance, we end up with the region constraint 'static: 'a,
*not* 'a: 'static
This means that our 'constraint chains' grow in the opposite direction
of 'normal lifetime' constraint chains. As we introduce subtypes, our
lifetime ends up being outlived by other lifetimes, rather than
outliving other lifetimes. Therefore, starting at the end of the
'constraint chain' will cause us to 'blame' a lifetime close to the original
definition of a variable, instead of close to where the bad lifetime
constraint is introduced.
This PR improves how we select the region to blame for 'too large'
universal lifetimes, when bound lifetimes are involved. If the region
we're checking is a 'placeholder' region (e.g. the region 'a' in
for<'a>, or the implicit region in fn(&())), we start traversing the
constraint chain from the beginning, rather than the end.
There are two (maybe more) different ways we generate region constraints for NLL:
requirements generated from trait queries, and requirements generated
from MIR subtype constraints. While the former always use explicit
placeholder regions, the latter is more tricky. In order to implement
contravariance for HRTBs, TypeRelating replaces placeholder regions with
existential regions. This requires us to keep track of whether or not an
existential region was originally a placeholder region. When we look for
a region to blame, we check if our starting region is either a
placeholder region or is an existential region created from a
placeholder region. If so, we start iterating from the beginning of the
constraint chain, rather than the end.
Rollup of 10 pull requests
Successful merges:
- #63674 (syntax: Support modern attribute syntax in the `meta` matcher)
- #63931 (Stabilize macros in some more positions)
- #64887 (syntax: recover trailing `|` in or-patterns)
- #64895 (async/await: improve not-send errors)
- #64896 (Remove legacy grammar)
- #64907 (A small amount of tidying-up factored out from PR #64648)
- #64928 (Add tests for some issues)
- #64930 (Silence unreachable code lint from await desugaring)
- #64935 (Improve code clarity)
- #64937 (Deduplicate closure type errors)
Failed merges:
r? @ghost
Deduplicate closure type errors
Closure typing obligations flow in both direcitons to properly infer
types. Because of this, we will get 2 type errors whenever there's
an unfulfilled obligation. To avoid this, we deduplicate them in the
`InferCtxt`.
Silence unreachable code lint from await desugaring
Fixes#61798.
This PR silences the unreachable code lint when it originates from within an await desugaring.
A small amount of tidying-up factored out from PR #64648
As requested by @Mark-Simulacrum, I put this in a separate commit to make it easier to review. (As far as I can tell, no violations of the policy here, and they are simply in a separate PR because they're not directly related to the import of that PR.)
r? @Mark-Simulacrum
Remove legacy grammar
Revival of #50835 & #55545
On the #wg-grammar discord there was agreement that enough progress has been made to be able to remove the legacy grammar.
r? @Centril @qmx
cc @rust-lang/wg-grammar
async/await: improve not-send errors
cc #64130.
```
note: future does not implement `std::marker::Send` because this value is used across an await
--> $DIR/issue-64130-non-send-future-diags.rs:15:5
|
LL | let g = x.lock().unwrap();
| - has type `std::sync::MutexGuard<'_, u32>`
LL | baz().await;
| ^^^^^^^^^^^ await occurs here, with `g` maybe used later
LL | }
| - `g` is later dropped here
```
r? @nikomatsakis
syntax: recover trailing `|` in or-patterns
Fixes https://github.com/rust-lang/rust/issues/64879.
For example (this also shows that we are sensitive to the typo `||`):
```
error: a trailing `|` is not allowed in an or-pattern
--> $DIR/remove-leading-vert.rs:33:11
|
LL | A || => {}
| - ^^ help: remove the `||`
| |
| while parsing this or-pattern starting here
|
= note: alternatives in or-patterns are separated with `|`, not `||`
```
r? @estebank
syntax: Support modern attribute syntax in the `meta` matcher
Where "modern" means https://github.com/rust-lang/rust/pull/57367:
```
PATH
PATH `(` TOKEN_STREAM `)`
PATH `[` TOKEN_STREAM `]`
PATH `{` TOKEN_STREAM `}`
```
Unfortunately, `meta` wasn't future-proofed using the `FOLLOW` token set like other matchers (https://github.com/rust-lang/rust/issues/34011), so code like `$meta:meta {` or `$meta:meta [` may break, and we need a crater run to find out how often this happens in practice.
Closes https://github.com/rust-lang/rust/issues/49629 (by fully supporting `meta` rather than removing it.)
Rollup of 9 pull requests
Successful merges:
- #64377 (Add long error explanation for E0493)
- #64786 (Use https for curl when building for linux)
- #64828 (Graphviz debug output for generic dataflow analysis)
- #64838 (Add long error explanation for E0550)
- #64891 (Fix `vec![x; n]` with null raw fat pointer zeroing the pointer metadata)
- #64893 (Zero-initialize `vec![None; n]` for `Option<&T>`, `Option<&mut T>` and `Option<Box<T>>`)
- #64911 (Fixed a misleading documentation issue #64844)
- #64921 (Add test for issue-64662)
- #64923 (Add missing links for mem::needs_drop)
Failed merges:
- #64918 (Add long error explanation for E0551)
r? @ghost
Closure typing obligations flow in both direcitons to properly infer
types. Because of this, we will get 2 type errors whenever there's
an unfulfilled obligation. To avoid this, we deduplicate them in the
`InferCtxt`.
This commit improves obligation errors for async/await:
```
note: future does not implement `std::marker::Send` because this value is used across an
await
--> $DIR/issue-64130-non-send-future-diags.rs:15:5
|
LL | let g = x.lock().unwrap();
| - has type `std::sync::MutexGuard<'_, u32>`
LL | baz().await;
| ^^^^^^^^^^^ await occurs here, with `g` maybe used later
LL | }
| - `g` is later dropped here
```
Signed-off-by: David Wood <david@davidtw.co>
Fix `vec![x; n]` with null raw fat pointer zeroing the pointer metadata
https://github.com/rust-lang/rust/pull/49496 introduced specialization based on:
```rust
unsafe impl<T: ?Sized> IsZero for *mut T {
fn is_zero(&self) -> bool {
(*self).is_null()
}
}
```
… to call `RawVec::with_capacity_zeroed` for creating `Vec<*mut T>`, which is incorrect for fat pointers since `<*mut T>::is_null` only looks at the data component. That is, a fat pointer can be “null” without being made entirely of zero bits.
This commit fixes it by removing the `?Sized` bound on this impl (and the corresponding `*const T` one). This regresses `vec![x; n]` with `x` a null raw slice of length zero, but that seems exceptionally uncommon. (Vtable pointers are never null, so raw trait objects would not take the fast path anyway.)
An alternative to keep the `?Sized` bound (or even generalize to `impl<U: Copy> IsZero for U`) would be to cast to `&[u8]` of length `size_of::<U>()`, but the optimizer seems not to be able to propagate alignment information and sticks with comparing one byte at a time:
https://rust.godbolt.org/z/xQFkwL
----
Without the library change, the new test fails as follows:
```rust
---- vec::vec_macro_repeating_null_raw_fat_pointer stdout ----
[src/liballoc/tests/vec.rs:1301] ptr_metadata(raw_dyn) = 0x00005596ef95f9a8
[src/liballoc/tests/vec.rs:1306] ptr_metadata(vec[0]) = 0x0000000000000000
thread 'vec::vec_macro_repeating_null_raw_fat_pointer' panicked at 'assertion failed: vec[0] == null_raw_dyn', src/liballoc/tests/vec.rs:1307:5
```
Graphviz debug output for generic dataflow analysis
A follow up to #64566.
This outputs graphviz diagrams in the generic dataflow engine when `#[rustc_mir(borrowck_graphviz_postflow="suffix.dot")]` is set on an item. This code is based on [`dataflow/graphviz.rs`](https://github.com/rust-lang/rust/blob/master/src/librustc_mir/dataflow/graphviz.rs), but displays different information since there are no "gen"/"kill" sets and transfer functions cannot be coalesced in the generic analysis. As a result, we show the dataflow state at the start and end of each block, as well as the changes resulting from each statement. I also render the state bitset in full (`{_1,_2}`) instead of hex-encoded as the current renderer does (`06`).
Use https for curl when building for linux
I noticed that the dist-x86_64-linux builder uses http to fetch curl and doesn't do any signature verification. It should probably use https.
r? @alexcrichton @Mark-Simulacrum
