MIR: Fix value moved diagnose messages
#45960. I believe this will take a different approach. Simply replacing all nouns to verbs (`desired_action`) messes up the message `use of moved value` (although fixes the message in original issue). Here is what happens:
<pre>
$ rustc -Zborrowck-mir src/test/ui/borrowck/borrowck-reinit.rs
error[E0382]: <b>used</b> of moved value: `x` (Mir)
--> src/test/ui/borrowck/borrowck-reinit.rs:18:16
|
17 | drop(x);
| - value moved here
18 | let _ = (1,x);
| ^ value used here after move
error: aborting due to 2 previous errors
</pre>
(Notice: *"**used** of moved value: `x`"* instead of *"**use**"*)
Which does not seem to be okay.
After experimenting a bit, it looks like [`report_use_of_moved_value()`](1dc0b573e7/src/librustc_mir/borrow_check.rs (L1319)) tries to handle both these messages by taking in only one form of`desired_action`.
These messages rise from: *"[{noun} of moved value](1dc0b573e7/src/librustc_mir/borrow_check.rs (L1338-L1342))"* and *"[value {verb} here after move](1dc0b573e7/src/librustc_mir/borrow_check.rs (L1343))"*.
This PR fixes *"value {verb} here after move"* type messages by passing a corresponding verb (`desired_action`) instead of the original noun.
Stabilize spin_loop_hint
Stabilize `spin_loop_hint` in release `1.23.0`.
I've also renamed feature `hint_core_should_pause` to `spin_loop_hint`.
cc #41196
Implement From<RecvError> for TryRecvError and RecvTimeoutError
According to the documentation, it looks to me that `TryRecvError` and `RecvTimeoutError` are strict extensions of `RecvError`. As such, it makes sense to allow conversion from the latter type to the two former types without constraining future developments.
This permits to write `input.recv()?` and `input.recv_timeout(timeout)?` in the same function for example.
Make accesses to fields of packed structs unsafe
To handle packed structs with destructors (which you'll think are a rare
case, but the `#[repr(packed)] struct Packed<T>(T);` pattern is
ever-popular, which requires handling packed structs with destructors to
avoid monomorphization-time errors), drops of subfields of packed
structs should drop a local move of the field instead of the original
one.
That's it, I think I'll use a strategy suggested by @Zoxc, where this mir
```
drop(packed_struct.field)
```
is replaced by
```
tmp0 = packed_struct.field;
drop tmp0
```
cc #27060 - this should deal with that issue after codegen of drop glue
is updated.
The new errors need to be changed to future-compatibility warnings, but
I'll rather do a crater run first with them as errors to assess the
impact.
cc @eddyb
Things which still need to be done for this:
- [ ] - handle `repr(packed)` structs in `derive` the same way I did in `Span`, and use derive there again
- [ ] - implement the "fix packed drops" pass and call it in both the MIR shim and validated MIR pipelines
- [ ] - do a crater run
- [ ] - convert the errors to compatibility warnings
mention nightly in -Z external-macro-backtrace note
Fix#46167 by mentioning that you need nightly in the message that tells you to pass `-Z external-macro-backtrace`.
Rationale:
1. The reason for having this message is to increase discoverability of the functionality. If the message is only shown on nightly it's less disoverable.
2. The same approach is taken if you call a const fn in const context without its feature gate (previously, if you called it without `#![feature(const_fn)]`).
cc @kennytm
Add a MIR-borrowck-only output mode
Removes the `-Z borrowck-mir` flag in favour of a `-Z borrowck=mode` flag where mode can be `mir`, `ast`, or `compare`.
* The `ast` mode represents the current default, passing `-Z borrowck=ast` is equivalent to not passing it at all.
* The `compare` mode outputs both the output of the MIR borrow checker and the AST borrow checker, each error with `(Ast)` and `(Mir)` appended. This mode has the same behaviour as `-Z borrowck-mir` had before this commit.
* The `mir` mode only outputs the results of the MIR borrow checker, while suppressing the errors of the ast borrow checker
The PR also updates the tests to use the new flags.
closes #46097
rustc_trans: don't apply noalias on returned references.
In #45225 frozen returned `&T` were accidentally maked `noalias`, unlike `&mut T`.
Return value `noalias` is only sound for functions that return dynamic allocations, e.g. `Box`, and using it on anything else can lead to miscompilation, as LLVM assumes certain usage patterns.
Fixes#46239.
There's a trailing whitespace problem in one of the tests. @nrc said
he'll go fix it quickly, but until then I'll like to land this PR.
I suspect this happened because one of the dependencies in the
Cargo.lock I updated had changed the format of the XML they emit, but
that has to be investigated.
Fix the derive implementation for repr(packed) structs to move the
fields out instead of calling functions on references to each subfield.
That's it, `#[derive(PartialEq)]` on a packed struct now does:
```Rust
fn eq(&self, other: &Self) {
let field_0 = self.0;
let other_field_0 = other.0;
&field_0 == &other_field_0
}
```
Instead of
```Rust
fn eq(&self, other: &Self) {
let ref field_0 = self.0;
let ref other_field_0 = other.0;
&*field_0 == &*other_field_0
}
```
Taking (unaligned) references to each subfield is undefined, unsound and
is an error with MIR effectck, so it had to be prevented. This causes
a borrowck error when a `repr(packed)` struct has a non-Copy field (and
therefore is a [breaking-change]), but I don't see a sound way to avoid
that error.
To handle packed structs with destructors (which you'll think are a rare
case, but the `#[repr(packed)] struct Packed<T>(T);` pattern is
ever-popular, which requires handling packed structs with destructors to
avoid monomorphization-time errors), drops of subfields of packed
structs should drop a local move of the field instead of the original
one.
cc #27060 - this should deal with that issue after codegen of drop glue
is updated.
The new errors need to be changed to future-compatibility warnings, but
I'll rather do a crater run first with them as errors to assess the
impact.
MIR: adopt borrowck test
Fix trailing whitespace
span_bug! on unexpected action
Make RegionVid use newtype_index!
Closes#45843
Check rvalue aggregates during check_stmt in tycheck, add initial, (not passing) test
Fix failing test
Remove attributes and test comments accidentally left behind, add in span_mirbugs
Normalize LvalueTy for ops and format code to satisfy tidy check
only normalize operand types when in an ADT constructor
avoid early return
handle the active field index in unions
normalize types in ADT constructor
Fixes#45940
Fix borrowck compiler errors for upvars contain "spurious" dereferences
Fixes#46003
added associated function Box::leak
Box::leak - improve documentation
Box::leak - fixed bug in documentation
Box::leak - relaxed constraints wrt. lifetimes
Box::leak - updated documentation
Box::leak - made an oops, fixed now =)
Box::leak: update unstable issue number (46179).
Add test for #44953
Add missing Debug impls to std_unicode
Also adds #![deny(missing_debug_implementations)] so they don't get
missed again.
Amend RELEASES for 1.22.1
and fix the date for 1.22.0
Rename param in `[T]::swap_with_slice` from `src` to `other`.
The idea of ‘source’ and ‘destination’ aren’t very applicable for this
operation since both slices can both be considered sources and
destinations.
Clarify stdin behavior of `Command::output`.
Fixes#44929.
Add hints for the case of confusing enum with its variants
Add failing testcases
Add module population and case of enum in place of expression
Use for_each_child_stable in find_module
Use multiline text for crate conflict diagnostics
Make float::from_bits transmute (and update the documentation to reflect this).
The current implementation/documentation was made to avoid sNaN because of
potential safety issues implied by old/bad LLVM documentation. These issues
aren't real, so we can just make the implementation transmute (as permitted
by the existing documentation of this method).
Also the documentation didn't actually match the behaviour: it said we may
change sNaNs, but in fact we canonicalized *all* NaNs.
Also an example in the documentation was wrong: it said we *always* change
sNaNs, when the documentation was explicitly written to indicate it was
implementation-defined.
This makes to_bits and from_bits perfectly roundtrip cross-platform, except
for one caveat: although the 2008 edition of IEEE-754 specifies how to
interpet the signaling bit, earlier editions didn't. This lead to some platforms
picking the opposite interpretation, so all signaling NaNs on x86/ARM are quiet
on MIPS, and vice-versa.
NaN-boxing is a fairly important optimization, while we don't even guarantee
that float operations properly preserve signalingness. As such, this seems like
the more natural strategy to take (as opposed to trying to mangle the signaling
bit on a per-platform basis).
This implementation is also, of course, faster.
Simplify an Iterator::fold to Iterator::any
This method of once-diagnostics doesn't allow nesting
UI tests extract the regular output from the 'rendered' field in json
Merge cfail and ui tests into ui tests
Add a MIR pass to lower 128-bit operators to lang item calls
Runs only with `-Z lower_128bit_ops` since it's not hooked into targets yet.
Include tuple projections in MIR tests
Add type checking for the lang item
As part of doing so, add more lang items instead of passing u128 to the i128 ones where it doesn't matter in twos-complement.
Handle shifts properly
* The overflow-checking shift items need to take a full 128-bit type, since they need to be able to detect idiocy like `1i128 << (1u128 << 127)`
* The unchecked ones just take u32, like the `*_sh?` methods in core
* Because shift-by-anything is allowed, cast into a new local for every shift
incr.comp.: Make sure we don't lose unused green results from the query cache.
rustbuild: Update LLVM and enable ThinLTO
This commit updates LLVM to fix#45511 (https://reviews.llvm.org/D39981) and
also reenables ThinLTO for libtest now that we shouldn't hit #45768. This also
opportunistically enables ThinLTO for libstd which was previously blocked
(#45661) on test failures related to debuginfo with a presumed cause of #45511.
Closes#45511
std: Flag Windows TLS dtor symbol as #[used]
Turns out ThinLTO was internalizing this symbol and eliminating it. Worse yet if
you compiled with LTO turns out no TLS destructors would run on Windows! The
`#[used]` annotation should be a more bulletproof implementation (in the face of
LTO) of preserving this symbol all the way through in LLVM and ensuring it makes
it all the way to the linker which will take care of it.
Add enum InitializationRequiringAction
Fix tidy tests
introduce macros for type-foldable and lift, convert stuff to use them
A random commit from a branch I've shelved for the time being that made `TypeFoldable` stuff a bit less annoying to write.
r? @eddyb
InstCombine Len([_; N]) => const N in MIR
A small opportunity I noticed in passing.
Not super exciting on its own, but opens the door for a const propagation pass that could completely remove const bounds checks from arrays at MIR time, for example.
Implement `Rc`/`Arc` conversions for string-like types
Provides the following conversion implementations:
* `From<`{`CString`,`&CStr`}`>` for {`Arc`,`Rc`}`<CStr>`
* `From<`{`OsString`,`&OsStr`}`>` for {`Arc`,`Rc`}`<OsStr>`
* `From<`{`PathBuf`,`&Path`}`>` for {`Arc`,`Rc`}`<Path>`
Closes#45008
Provides the following conversion implementations:
* `From<`{`CString`,`&CStr`}`>` for {`Arc`,`Rc`}`<CStr>`
* `From<`{`OsString`,`&OsStr`}`>` for {`Arc`,`Rc`}`<OsStr>`
* `From<`{`PathBuf`,`&Path`}`>` for {`Arc`,`Rc`}`<Path>`
Be more obvious when suggesting dereference
Include `&` span when suggesting dereference on a span that is already a reference:
```
error: non-reference pattern used to match a reference (see issue #42640)
--> dont-suggest-dereference-on-arg.rs:16:19
|
16 | .filter(|&(ref a, _)| foo(a))
| ^^^^^^^^^^^ help: consider using: `&&(ref k, _)`
|
= help: add #![feature(match_default_bindings)] to the crate attributes to enable
```
Fix#45925.
rustc: Add support for some more x86 SIMD ops
This commit adds compiler support for two basic operations needed for binding
SIMD on x86 platforms:
* First, a `nontemporal_store` intrinsic was added for the `_mm_stream_ps`, seen
in rust-lang-nursery/stdsimd#114. This was relatively straightforward and is
quite similar to the volatile store intrinsic.
* Next, and much more intrusively, a new type to the backend was added. The
`x86_mmx` type is used in LLVM for a 64-bit vector register and is used in
various intrinsics like `_mm_abs_pi8` as seen in rust-lang-nursery/stdsimd#74.
This new type was added as a new layout option as well as having support added
to the trans backend. The type is enabled with the `#[repr(x86_mmx)]`
attribute which is intended to just be an implementation detail of SIMD in
Rust.
I'm not 100% certain about how the `x86_mmx` type was added, so any extra eyes
or thoughts on that would be greatly appreciated!
This commit adds compiler support for two basic operations needed for binding
SIMD on x86 platforms:
* First, a `nontemporal_store` intrinsic was added for the `_mm_stream_ps`, seen
in rust-lang-nursery/stdsimd#114. This was relatively straightforward and is
quite similar to the volatile store intrinsic.
* Next, and much more intrusively, a new type to the backend was added. The
`x86_mmx` type is used in LLVM for a 64-bit vector register and is used in
various intrinsics like `_mm_abs_pi8` as seen in rust-lang-nursery/stdsimd#74.
This new type was added as a new layout option as well as having support added
to the trans backend. The type is enabled with the `#[repr(x86_mmx)]`
attribute which is intended to just be an implementation detail of SIMD in
Rust.
I'm not 100% certain about how the `x86_mmx` type was added, so any extra eyes
or thoughts on that would be greatly appreciated!
