Remove a HACK by instead inferring opaque types during expected/formal type checking
I was wondering why I couldn't come up with a test that hits the code path of the argument check checking the types we inferred from the return type... Turns out we reject those attempts early during fudging.
I have absolutely no information for you as to what kind of type inference changes this may incur, but I think we should just land this out of two reasons:
* had I found the other place to use opaque type inference on before I added the hack, we'd be using that today and this PR would never have happened
* if it is possible to hit this path, it requires some god awful recursive RPIT logic that I doubt anyone would have written without actively trying to write obscure code
r? ``@ghost``
Add the missing inttoptr when we ptrtoint in ptr atomics
Ralf noticed this here: https://github.com/rust-lang/rust/pull/122220#discussion_r1535172094
Our previous codegen forgot to add the cast back to integer type. The code compiles anyway, because of course all locals are in-memory to start with, so previous codegen would do the integer atomic, store the integer to a local, then load a pointer from that local. Which is definitely _not_ what we wanted: That's an integer-to-pointer transmute, so all pointers returned by these `AtomicPtr` methods didn't have provenance. Yikes.
Here's the IR for `AtomicPtr::fetch_byte_add` on 1.76: https://godbolt.org/z/8qTEjeraY
```llvm
define noundef ptr `@atomicptr_fetch_byte_add(ptr` noundef nonnull align 8 %a, i64 noundef %v) unnamed_addr #0 !dbg !7 {
start:
%0 = alloca ptr, align 8, !dbg !12
%val = inttoptr i64 %v to ptr, !dbg !12
call void `@llvm.lifetime.start.p0(i64` 8, ptr %0), !dbg !28
%1 = ptrtoint ptr %val to i64, !dbg !28
%2 = atomicrmw add ptr %a, i64 %1 monotonic, align 8, !dbg !28
store i64 %2, ptr %0, align 8, !dbg !28
%self = load ptr, ptr %0, align 8, !dbg !28
call void `@llvm.lifetime.end.p0(i64` 8, ptr %0), !dbg !28
ret ptr %self, !dbg !33
}
```
r? `@RalfJung`
cc `@nikic`
Rust passes a *const &OperandBundleDef to these APIs, usually from a
Vec<&OperandBundleDef> or so. Previously we were dereferencing that
pointer and passing it to the ArrayRef constructor with some length (N).
This meant that if the length was 0, we were dereferencing a pointer to
nowhere, and if the length was >1 then loading the *second* element
somewhere in LLVM would've been reading past the end.
Since Rust can't hold OperandBundleDef by-value we're forced to indirect
through a vector that copies out the OperandBundleDefs from the
by-reference list on the Rust side in order to match the LLVM expected
API.
Discard overflow obligations in `impl_may_apply`
Hacky fix for #123493. Throws away obligations that are overflowing in `impl_may_apply` when we recompute if an impl applies, since those will lead to fatal overflow if processed during fulfillment.
Something about #114811 (I think it's the predicate reordering) caused us to evaluate predicates differently in error reporting leading to fatal overflow, though I believe the underlying overflow is possible to hit since this code was rewritten to use fulfillment.
Fixes#123493
Does not necessarily change much, but we never overwrite it, so I see no reason
for it to be in the `Successors` trait. (+we already have a similar `is_cyclic`)
Add add/sub methods that only panic with debug assertions to rustc
This mitigates the perf impact of enabling overflow checks on rustc. The change to use overflow checks will be done in a later PR.
For rust-lang/compiler-team#724, based on data gathered in #119440.
Rollup of 7 pull requests
Successful merges:
- #123530 (Enable building tier2 target riscv32im-unknown-none-elf)
- #123642 (do not allow using local llvm while using rustc from ci)
- #123716 (Update documentation of Path::to_path_buf and Path::ancestors)
- #123876 (Update backtrace submodule)
- #123888 (Replace a `DefiningOpaqueTypes::No` with `Yes` by asserting that one side of the comparison is a type variable.)
- #123890 (removed (mostly) unused code)
- #123891 (Miri subtree update)
r? `@ghost`
`@rustbot` modify labels: rollup
Replace a `DefiningOpaqueTypes::No` with `Yes` by asserting that one side of the comparison is a type variable.
Thus there will never be an opaque type involved in a way that constrains its hidden type, as the other side of the comparison is always a generator witness type
r? ``@compiler-errors``
Linker flavors next steps: linker features
This is my understanding of the first step towards `@petrochenkov's` vision for the future of linker flavors, described in https://github.com/rust-lang/rust/pull/119906#issuecomment-1895693162 and the discussion that followed.
To summarize: having `Cc` and `Lld` embedded in linker flavors creates tension about naming, and a combinatorial explosion of flavors for each new linker feature we'd want to use. Linker features are an extension mechanism that is complementary to principal flavors, with benefits described in #119906.
The most immediate use of this flag would be to turn self-contained linking on and off via features instead of flavors. For example, `-Clinker-features=+/-lld` would toggle using lld instead of selecting a precise flavor, and would be "generic" and work cross-platform (whereas linker flavors are currently more tied to targets). Under this scheme, MCP510 is expected to be `-Clink-self-contained=+linker -Zlinker-features=+lld -Zunstable-options` (though for the time being, the original flags using lld-cc flavors still work).
I purposefully didn't add or document CLI support for `+/-cc`, as it would be a noop right now. I only expect that we'd initially want to stabilize `+/-lld` to begin with.
r? `@petrochenkov`
You had requested that minimal churn would be done to the 230 target specs and this does none yet: the linker features are inferred from the flavor since they're currently isomorphic. We of course expect this to change sooner rather than later.
In the future, we can allow targets to define linker features independently from their flavor, and remove the cc and lld components from the flavors to use the features instead, this actually doesn't need to block stabilization, as we discussed.
(Best reviewed per commit)
Detect borrow checker errors where `.clone()` would be an appropriate user action
When a value is moved twice, suggest cloning the earlier move:
```
error[E0509]: cannot move out of type `U2`, which implements the `Drop` trait
--> $DIR/union-move.rs:49:18
|
LL | move_out(x.f1_nocopy);
| ^^^^^^^^^^^
| |
| cannot move out of here
| move occurs because `x.f1_nocopy` has type `ManuallyDrop<RefCell<i32>>`, which does not implement the `Copy` trait
|
help: consider cloning the value if the performance cost is acceptable
|
LL | move_out(x.f1_nocopy.clone());
| ++++++++
```
When a value is borrowed by an `fn` call, consider if cloning the result of the call would be reasonable, and suggest cloning that, instead of the argument:
```
error[E0505]: cannot move out of `a` because it is borrowed
--> $DIR/variance-issue-20533.rs:53:14
|
LL | let a = AffineU32(1);
| - binding `a` declared here
LL | let x = bat(&a);
| -- borrow of `a` occurs here
LL | drop(a);
| ^ move out of `a` occurs here
LL | drop(x);
| - borrow later used here
|
help: consider cloning the value if the performance cost is acceptable
|
LL | let x = bat(&a).clone();
| ++++++++
```
otherwise, suggest cloning the argument:
```
error[E0505]: cannot move out of `a` because it is borrowed
--> $DIR/variance-issue-20533.rs:59:14
|
LL | let a = ClonableAffineU32(1);
| - binding `a` declared here
LL | let x = foo(&a);
| -- borrow of `a` occurs here
LL | drop(a);
| ^ move out of `a` occurs here
LL | drop(x);
| - borrow later used here
|
help: consider cloning the value if the performance cost is acceptable
|
LL - let x = foo(&a);
LL + let x = foo(a.clone());
|
```
This suggestion doesn't attempt to square out the types between what's cloned and what the `fn` expects, to allow the user to make a determination on whether to change the `fn` call or `fn` definition themselves.
Special case move errors caused by `FnOnce`:
```
error[E0382]: use of moved value: `blk`
--> $DIR/once-cant-call-twice-on-heap.rs:8:5
|
LL | fn foo<F:FnOnce()>(blk: F) {
| --- move occurs because `blk` has type `F`, which does not implement the `Copy` trait
LL | blk();
| ----- `blk` moved due to this call
LL | blk();
| ^^^ value used here after move
|
note: `FnOnce` closures can only be called once
--> $DIR/once-cant-call-twice-on-heap.rs:6:10
|
LL | fn foo<F:FnOnce()>(blk: F) {
| ^^^^^^^^ `F` is made to be an `FnOnce` closure here
LL | blk();
| ----- this value implements `FnOnce`, which causes it to be moved when called
```
Account for redundant `.clone()` calls in resulting suggestions:
```
error[E0507]: cannot move out of dereference of `S`
--> $DIR/needs-clone-through-deref.rs:15:18
|
LL | for _ in self.clone().into_iter() {}
| ^^^^^^^^^^^^ ----------- value moved due to this method call
| |
| move occurs because value has type `Vec<usize>`, which does not implement the `Copy` trait
|
note: `into_iter` takes ownership of the receiver `self`, which moves value
--> $SRC_DIR/core/src/iter/traits/collect.rs:LL:COL
help: you can `clone` the value and consume it, but this might not be your desired behavior
|
LL | for _ in <Vec<usize> as Clone>::clone(&self).into_iter() {}
| ++++++++++++++++++++++++++++++ ~
```
We use the presence of `&mut` values in a move error as a proxy for the user caring about side effects, so we don't emit a clone suggestion in that case:
```
error[E0505]: cannot move out of `s` because it is borrowed
--> $DIR/borrowck-overloaded-index-move-index.rs:53:7
|
LL | let mut s = "hello".to_string();
| ----- binding `s` declared here
LL | let rs = &mut s;
| ------ borrow of `s` occurs here
...
LL | f[s] = 10;
| ^ move out of `s` occurs here
...
LL | use_mut(rs);
| -- borrow later used here
```
We properly account for `foo += foo;` errors where we *don't* suggest `foo.clone() += foo;`, instead suggesting `foo += foo.clone();`.
---
Each commit can be reviewed in isolation. There are some "cleanup" commits, but kept them separate in order to show *why* specific changes were being made, and their effect on tests' output.
Fix#49693, CC #64167.
Thus there will never be an opaque type involved in a way that constrains its hidden type, as the other side of the comparison is always a generator witness type
Re-enable `has_thread_local` for i686-msvc
A few years back, `has_thread_local` was disabled as a workaround for a compiler issue. While the exact cause was never tracked down, it was suspected to be caused by the compiler inlining a thread local access across a dylib boundary. This should be fixed now so let's try again.
Add `/System/iOSSupport` to the library search path on Mac Catalyst
On macOS, `/System/iOSSupport` contains iOS frameworks like UIKit, which is the whole idea of Mac Catalyst.
To link to these, we need to explicitly tell the linker about the support library stubs provided in the macOS SDK under the same path.
Concretely, when building a binary for Mac Catalyst, Xcode passes the following flags to the linker:
```
-iframework /Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX14.2.sdk/System/iOSSupport/System/Library/Frameworks
-L/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX14.2.sdk/System/iOSSupport/usr/lib
```
This is not something that can be disabled (it's enabled as soon as you enable `SUPPORTS_MACCATALYST`), so I think it's pretty safe to say that we don't need an option to turn these off.
I've chosen to slightly deviate from what Xcode does and use `-F` instead of `-iframework`, since we don't need to change the header search path, and this way the flags nicely match on all the linkers. From what I could tell by reading Clang sources, there shouldn't be a difference when just running the linker.
CC `@BlackHoleFox,` `@shepmaster` (I accidentally let rustbot choose the reviewer).
typeck: fix `?` suggestion span
Noticed in <https://github.com/rust-lang/rust/pull/112043#issuecomment-2043565292>, if the
```
use the `?` operator to extract the `Result<(), std::fmt::Error>` value, propagating a `Result::Err` value to the caller
```
suggestion is applied to a macro that comes from a non-local crate (e.g. the stdlib), the suggestion span can become non-local, which will cause newer rustfix versions to fail.
This PR tries to remedy the problem by recursively probing ancestors of the expression span, trying to identify the most ancestor span that is (1) still local, and (2) still shares the same syntax context as the expression.
This is the same strategy used in https://github.com/rust-lang/rust/pull/112043.
The test unfortunately cannot `//@ run-rustfix` because there are two conflicting MaybeIncorrect suggestions that when collectively applied, cause the fixed source file to become non-compilable.
Also avoid running `//@ run-rustfix` for `tests/ui/typeck/issue-112007-leaked-writeln-macro-internals.rs` because that also contains conflicting suggestions.
cc `@ehuss` who noticed this. This question mark span fix + not running rustfix on the tests containing conflicting MaybeIncorrect suggestions should hopefully unblock rustfix from updating.
The suggestion to use `let else` with an uninitialized refutable `let`
statement was erroneous: `let else` cannot be used with deferred
initialization.
