More restrictive 2 phase borrows - take 2
Signal lint diagnostic `mutable_borrow_reservation_conflict` when borrow-check finds a 2-phase borrow's reservation overlapping with a shared borrow.
(pnkfelix updated description)
cc #56254 , #59159
blocks PR #59114
r? @pnkfelix
cc @RalfJung @nikomatsakis
Remove invalid assertion back:🔗:from add_upstream_rust_crates().
This removes a misplaced assertion. The function containing the assertion is actually only ever called for upstream crates that are not considered for LTO, so we don't care whether upstream code has been merged in by LTO or not.
Fixes#59137
r? @alexcrichton
Fix invalid bounds string generation in rustdoc
Fixes#58737.
Very weird and I'm not sure this is the best fix around. However, trying to fix it beforehand seems overly complicated compared to the gain (in `clean`, it wouldn't change anything since we **have to** return something so that wouldn't work, and in `hir`, I'm afraid I'd break something else for very little gain).
Also, I wasn't able to make a small code to reproduce the issue. The only way to test is to document `crossbeam` directly and check the `Scope` struct...
r? @QuietMisdreavus
Unsized rvalues: implement boxed closure impls. (2nd try)
This is a rebase of S-blocked-closed PR #55431 to current master. LLVM has moved forward since then, so maybe we can check whether the new LLVM 8.0 version unblocked this work.
This commit changes the behavior of Formatter::debug_struct,
debug_tuple, debug_list, debug_set, and debug_map to render trailing
commas in {:#?} mode, which is the dominant style in modern Rust code.
Before:
Language {
name: "Rust",
trailing_commas: false
}
After:
Language {
name: "Rust",
trailing_commas: true,
}
std: Avoid usage of `Once` in `Instant`
This commit removes usage of `Once` from the internal implementation of
time utilities on OSX and Windows. It turns out that we accidentally hit
a deadlock today (#59020) via events that look like:
* A thread invokes `park_timeout`
* Internally, only on OSX, `park_timeout` calls `Instant::elapsed`
* Inside of `Instant::elapsed` on OSX we enter a `Once` to initialize
global timer data
* Inside of `Once`, it attempts to `park`
This means on the same stack frame, when there's contention, we're
calling `park` from inside `park_timeout`, causing a deadlock!
The solution implemented in this commit was to remove usage of `Once`
and instead just do a small dance with atomics. There's no real need we
need to guarantee that the global information is only learned once, only
that it's only *stored* once. This implementation may have multiple
threads invoke `mach_timebase_info`, but only one will store the global
information which will amortize the cost for all other threads.
A similar fix has been applied to windows to be uniform across our
implementations, but looking at the code on Windows no deadlock was
possible. This is purely just a consistency update for Windows and in
theory a slightly leaner implementation.
Closes#59020
This commit removes usage of `Once` from the internal implementation of
time utilities on OSX and Windows. It turns out that we accidentally hit
a deadlock today (#59020) via events that look like:
* A thread invokes `park_timeout`
* Internally, only on OSX, `park_timeout` calls `Instant::elapsed`
* Inside of `Instant::elapsed` on OSX we enter a `Once` to initialize
global timer data
* Inside of `Once`, it attempts to `park`
This means on the same stack frame, when there's contention, we're
calling `park` from inside `park_timeout`, causing a deadlock!
The solution implemented in this commit was to remove usage of `Once`
and instead just do a small dance with atomics. There's no real need we
need to guarantee that the global information is only learned once, only
that it's only *stored* once. This implementation may have multiple
threads invoke `mach_timebase_info`, but only one will store the global
information which will amortize the cost for all other threads.
A similar fix has been applied to windows to be uniform across our
implementations, but looking at the code on Windows no deadlock was
possible. This is purely just a consistency update for Windows and in
theory a slightly leaner implementation.
Closes#59020
Never return uninhabited values at all
Functions with uninhabited return values are already marked `noreturn`,
but we were still generating return instructions for this. When running
with `-C passes=lint`, LLVM prints:
Unusual: Return statement in function with noreturn attribute
The LLVM manual makes a stronger statement about `noreturn` though:
> This produces undefined behavior at runtime if the function ever does
dynamically return.
We now emit an `abort` anywhere that would have tried to return an
uninhabited value.
Fixes#48227
cc #7463#48229
r? @eddyb
Functions with uninhabited return values are already marked `noreturn`,
but we were still generating return instructions for this. When running
with `-C passes=lint`, LLVM prints:
Unusual: Return statement in function with noreturn attribute
The LLVM manual makes a stronger statement about `noreturn` though:
> This produces undefined behavior at runtime if the function ever does
dynamically return.
We now emit an `abort` anywhere that would have tried to return an
uninhabited value.
Fix stack overflow when generating debuginfo for 'recursive' type
By using 'impl trait', it's possible to create a self-referential
type as follows:
fn foo() -> impl Copy { foo }
This is a function which returns itself.
Normally, the signature of this function would be impossible
to write - it would look like 'fn foo() -> fn() -> fn() ...'
e.g. a function which returns a function, which returns a function...
Using 'impl trait' allows us to avoid writing this infinitely long
type. While it's useless for practical purposes, it does compile and run
However, issues arise when we try to generate llvm debuginfo for such a
type. All 'impl trait' types (e.g. ty::Opaque) are resolved when we
generate debuginfo, which can lead to us recursing back to the original
'fn' type when we try to process its return type.
To resolve this, I've modified debuginfo generation to account for these
kinds of weird types. Unfortunately, there's no 'correct' debuginfo that
we can generate - 'impl trait' does not exist in debuginfo, and this
kind of recursive type is impossible to directly represent.
To ensure that we emit *something*, this commit emits dummy
debuginfo/type names whenever it encounters a self-reference. In
practice, this should never happen - it's just to ensure that we can
emit some kind of debuginfo, even if it's not particularly meaningful
Fixes#58463
resolve: collect trait aliases along with traits
It seems trait aliases weren't being collected as `TraitCandidates` in resolve, this should change that. (I can't compile the full compiler locally, so relying on CI...)
Fixes https://github.com/rust-lang/rust/issues/56485
r? @alexreg
async fn now lowers directly to an existential type declaration
rather than reusing the `impl Trait` return type lowering.
As part of this, it lowers all argument-position elided lifetimes
using the in-band-lifetimes machinery, creating fresh parameter
names for each of them, using each lifetime parameter as a generic
argument to the generated existential type.
This doesn't currently successfully allow multiple
argument-position elided lifetimes since `existential type`
doesn't yet support multiple lifetimes where neither outlive
the other. This requires a separate fix.
By using 'impl trait', it's possible to create a self-referential
type as follows:
fn foo() -> impl Copy { foo }
This is a function which returns itself.
Normally, the signature of this function would be impossible
to write - it would look like 'fn foo() -> fn() -> fn() ...'
e.g. a function which returns a function, which returns a function...
Using 'impl trait' allows us to avoid writing this infinitely long
type. While it's useless for practical purposes, it does compile and run
However, issues arise when we try to generate llvm debuginfo for such a
type. All 'impl trait' types (e.g. ty::Opaque) are resolved when we
generate debuginfo, which can lead to us recursing back to the original
'fn' type when we try to process its return type.
To resolve this, I've modified debuginfo generation to account for these
kinds of weird types. Unfortunately, there's no 'correct' debuginfo that
we can generate - 'impl trait' does not exist in debuginfo, and this
kind of recursive type is impossible to directly represent.
To ensure that we emit *something*, this commit emits dummy
debuginfo/type names whenever it encounters a self-reference. In
practice, this should never happen - it's just to ensure that we can
emit some kind of debuginfo, even if it's not particularly meaningful
Fixes#58463
Fix LLVM IR generated for C-variadic arguments
It is possible to create malformed LLVM IR given variadic arguments that
are aggregate types. This occurs due to improper tracking of the current
argument in the functions list of arguments.
Fixes: #58881
It is possible to create malformed LLVM IR given variadic arguments that
are aggregate types. This occurs due to improper tracking of the current
argument in the functions list of arguments.
Fallback to `static_impl_trait` for nice error message by peeking at the
return type and the lifetime type. Point at the return type instead of
the return expr/stmt in NLL mode.
