Recover gracefully from `struct` parse errors
Currently the parser tries to recover from finding a keyword where a field name was expected, but this causes extra knock down parse errors that are completely irrelevant. Instead, bail out early in the parsing of the field and consume the remaining tokens in the block. This can reduce output significantly.
_Improvements based on the narrative in https://fasterthanli.me/articles/i-am-a-java-csharp-c-or-cplusplus-dev-time-to-do-some-rust_
merge `as_local_hir_id` with `local_def_id_to_hir_id`
`as_local_hir_id` was defined as just calling `local_def_id_to_hir_id` and I think that having two different ways to call the same method is somewhat confusing.
Don't really care about which of these 2 methods we want to keep.
Does this require an MCP, considering that these methods are fairly frequently used?
Set CMAKE_SYSTEM_NAME when cross-compiling
Configure CMAKE_SYSTEM_NAME when cross-compiling in `configure_cmake`,
to tell CMake about target system. Previously this was done only for
LLVM step and now applies more generally to steps using cmake.
Helps with #74576.
Remove most specialization use in serialization
Switching from specialization to min_specialization in the compiler made the unsoundness of how we used these traits pretty clear. This changes how the `Encodable` and `Decodable` traits work to be more friendly for types need a `TyCtxt` to deserialize.
The alternative design of having both `Encodable` and `TyEncodable` traits was considered, but doesn't really work because the following impls would conflict:
```
impl<E: Ecodable> TyEncodable for Encodable
impl<E: TyEcodable> TyEncodable for [E]
```
## How-to guide
- `Rustc(De|En)codable` is now spelled `Ty(De|En)coable` in `rustc_middle`, `Metadata(En|De)codable` in `rustc_metadata` where needed, and `(De|En)codable` everywhere else.
- Manual implementations of `(De|En)codable` shouldn't be much different.
- If you're adding a new interned type that needs to be en/decodable then the simplest thing way to handle this is:
- Have the type be a wrapper around a reference to the interned data (i.e. do what `ty::Predicate` does, and not what all of the other interned types do)
- Derive `Ty(En|De)codable` on the inner type
- Implement `Encodable<impl TyEncoder>` by forwarding to the inner type.
- Implement `Decodable<impl TyDecoder>` by decoding the inner type and then creating the wrapper around that (using the `tcx` from the decoder as needed).
cc @rust-lang/compiler for opinions on this change
r? @oli-obk
The match branch simplification is applied when target blocks contain
statements that are either equal or perform a const bool assignment with
different values to the same place.
Previously, when constructing new statements, only statements from a
single block had been examined. This lead to a misoptimization when
statements are equal because the assign the *same* const bool value to
the same place.
Fix the issue by examining statements from both blocks when deciding on
replacement.
This commit adjusts polymorphization's handling of predicates so that
after ensuring that `T` is used in `I: Foo<T>` if `I` is used, it now
ensures that `I` is used if `T` is used in `I: Foo<T>`. This is
necessary to mark generic parameters that only exist in impl parameters
as used - thereby avoiding symbol clashes when using the new mangling
scheme.
Signed-off-by: David Wood <david@davidtw.co>
Do not emit E0228 when it is implied by E0106
Emit E0288 (lifetime bound for trait object cannot be deduced) only on bare trait objects. When the trait object is in the form of `&dyn Trait`, E0106 (missing lifetime specifier) will have been emitted, making the former redundant.
- Move the type parameter from `encode` and `decode` methods to
the trait.
- Remove `UseSpecialized(En|De)codable` traits.
- Remove blanket impls for references.
- Add `RefDecodable` trait to allow deserializing to arena-allocated
references safely.
- Remove ability to (de)serialize HIR.
- Create proc-macros `(Ty)?(En|De)codable` to help implement these new
traits.
LLVM IR coverage encoding aligns closer to Clang's
I found some areas for improvement while attempting to debug the
SegFault issue when running rust programs compiled using MSVC, with
coverage instrumentation.
I discovered that LLVM's coverage writer was generating incomplete
function name variable names (that's not a typo: the name of the
variable that holds a function name).
The existing implementation used one-up numbers to distinguish
variables, and correcting the names did not fix the MSVC coverage bug,
but the fix in this PR makes the names and resulting LLVM IR easier to
follow and more consistent with Clang's implementation.
I also changed the way the `-Zinstrument-coverage` option is supported in
symbol_export.rs. The original implementation was incorrect, and the
corrected version matches the handling for `-Zprofile-generate`, as it
turns out.
(An argument could be made that maybe `-Zinstrument-coverage` should
automatically enable `-Cprofile-generate`. In fact, if
`-Cprofile-generate` is analagous to Clang's `-fprofile-generate`, as
some documentation implies, Clang always requires this flag for its
implementation of source-based code coverage. This would require a
little more validation, and if implemented, would probably require
updating some of the user-facing messages related to
`-Cprofile-generate` to not be so specific to the PGO use case.)
None of these changes fixed the MSVC coverage problems, but they should
still be welcome improvements.
Lastly, I added some additional FIXME comments in instrument_coverage.rs
describing issues I found with the generated LLVM IR that would be
resolved if the coverage instrumentation is injected with a `Statement`
instead of as a new `BasicBlock`. I describe seven advantages of this
change, but it requires some discussion before making a change like
this.
r? @tmandry
I found some areas for improvement while attempting to debug the
SegFault issue when running rust programs compiled using MSVC, with
coverage instrumentation.
I discovered that LLVM's coverage writer was generating incomplete
function name variable names (that's not a typo: the name of the
variable that holds a function name).
The existing implementation used one-up numbers to distinguish
variables, and correcting the names did not fix the MSVC coverage bug,
but the fix in this PR makes the names and resulting LLVM IR easier to
follow and more consistent with Clang's implementation.
I also changed the way the `-Zinstrument-coverage` option is supported
in symbol_export.rs. The original implementation was incorrect, and the
corrected version matches the handling for `-Zprofile-generate`, as it
turns out.
(An argument could be made that maybe `-Zinstrument-coverage` should
automatically enable `-Cprofile-generate`. In fact, if
`-Cprofile-generate` is analagous to Clang's `-fprofile-generate`, as
some documentation implies, Clang always requires this flag for its
implementation of source-based code coverage. This would require a
little more validation, and if implemented, would probably require
updating some of the user-facing messages related to
`-Cprofile-generate` to not be so specific to the PGO use case.)
None of these changes fixed the MSVC coverage problems, but they should
still be welcome improvements.
Lastly, I added some additional FIXME comments in instrument_coverage.rs
describing issues I found with the generated LLVM IR that would be
resolved if the coverage instrumentation is injected with a `Statement`
instead of as a new `BasicBlock`. I describe seven advantages of this
change, but it requires some discussion before making a change like
this.
Clean up some mir transform passes
I noticed a few places where there were intermediates being created
in MIR optimization passes, so I removed them.
I also changed some `Some(..)` into just `..` and wrap `Some(..)` at the function end, doing early returns for `None`.
I was generally looking for some easy optimizations in theses passes, and hopefully these should improve the runtime of these passes by a tinnnyyyyy bit.
r? @oli-obk
Run standard library unit tests without optimizations in `nopt` CI jobs
This was discussed in #73288 as a way to catch similar issues in the future. This builds an unoptimized standard library with the bootstrap compiler and runs the unit tests. This takes about 2 minutes on my laptop.
I confirmed that this method works locally, although there may be a better way of implementing it. It would be better to use the stage 2 compiler instead of the bootstrap one.
Notably, there are currently four `libstd` unit tests that fail in debug mode on `i686-unkown-linux-gnu` (a tier one target):
```
failures:
f32::tests::test_float_bits_conv
f32::tests::test_total_cmp
f64::tests::test_float_bits_conv
f64::tests::test_total_cmp
```
These are the tests that prompted #73288 as well as the ones added in #72568, which is currently broken due to #73328.
Emit E0288 (lifetime bound for trait object cannot be deduced) only on
bare trait objects. When the trait object is in the form of
`&dyn Trait`, E0106 (missing lifetime specifier) will have been emitted,
making the former redundant.
This also explicitly checks that the types are `bool`. `try_eval_bool` also appears to just
succeed for `u8`, so this ensures that it actually is a bool before casting.
Disable zlib in LLVM on aarch64-apple-darwin
For some reason, building rustc on the Apple Silicon DTK fails with some undefined symbols from zlib, which I guess is similar to some issues that appeared on *-apple-ios and *-apple-tvos.
Change registered "program name" for -Cllvm-args usage messages
While debugging a codegen issue, I tried adding LLVM options with
the rustc -Cllvm-args option, and was confused by the error and usage
messaging.
The LLVM "program name" argument is set to "rustc", and command line
error messages make it look like invalid arguments are "rustc"
arguments, not LLVM.
I changed this argument so error messages and the "-help" usage feedback
is easier to understand and react to. (Clang does something similar.)
r? @wesleywiser
fix LocalInfo doc comment
The doc comment makes it sound like this is diagnostics-only, but that is not true -- even [unsafety checking uses this information](ded20c98be/src/librustc_mir/transform/check_unsafety.rs (L206)), so it is crucial for soundness, not just "nice to have".
Cc @oli-obk
self-profile: Cache more query key strings when doing self-profiling.
This PR adds optimized `SpecIntoSelfProfilingString` implementations for two common query key types (`LocalDefId` and `WithOptConstParam`). This makes raw self-profiling data on disk 8-9% smaller for my two test cases (`regex` and `cargo`).
The on-disk format is not affected, so no tooling updates need to happen.
I also tried adding an impl for `Ty<'tcx>` (which should reduce size quite a bit) but the compiler did not allow me to add a specialized impl parameterized with `'tcx`. I don't know if there is an actual problem with that or if the implementation of specialization just doesn't support it yet.
cc @wesleywiser @Mark-Simulacrum
First iteration of simplify match branches
This is a simple MIR pass that attempts to convert
```
bb0: {
StorageLive(_2);
_3 = discriminant(_1);
switchInt(move _3) -> [0isize: bb2, otherwise: bb1];
}
bb1: {
_2 = const false;
goto -> bb3;
}
bb2: {
_2 = const true;
goto -> bb3;
}
```
into
```
bb0: {
StorageLive(_2);
_3 = discriminant(_1);
_2 = _3 == 0;
goto -> bb3;
}
```
There are still missing components(like checking if the assignments are bools).
Was hoping that this could get some review though.
Handles #75141
r? @oli-obk
We already have builders which built standard library *test*s without
optimizations, but we previously did not have builders which built the standard
library itself without optimizations and then tested that.
This adds those builds for i686 and x86_64 linux.
Link sanitizers when creating dynamic libraries on macOS
Link sanitizer runtime when creating dynamic libraries on macOS
to resolve sanitizer runtime symbols and avoid failure at link time.
Closes#74571.
Configure CMAKE_SYSTEM_NAME when cross-compiling in `configure_cmake`,
to tell CMake about target system. Previously this was done only for
LLVM step and now applies more generally to steps using cmake.
Add `array` lang item and `[T; N]::map(f: FnMut(T) -> S)`
This introduces an `array` lang item so functions can be defined on top of `[T; N]`. This was previously not done because const-generics was not complete enough to allow for this. Now it is in a state that is usable enough to start adding functions.
The function added is a monadic (I think?) map from `[T; N] -> [S; N]`. Until transmute can function on arrays, it also allocates an extra temporary array, but this can be removed at some point.
r? @lcnr
Add a bunch of const-generic revisions for `min_const_generics`
This adds a bunch of revisions to `const-generic` tests which is part of #75279, but doesn't cover everything.
r? @lcnr
Move platform support to the rustc book.
This moves the [Platform Support](https://forge.rust-lang.org/release/platform-support.html) page from the forge to the rustc book. There are several reasons for doing this:
* The forge is not really oriented towards end-users (it mostly contains infrastructure, governance and policy, internal team pages, etc.). This platform support page is useful to user to know which targets are supported.
* This page can now be updated in-sync with any PRs that add or remove a target, or change its status.
* This is now automatically checked on CI to verify the list does not get out of sync. Currently it only checks the presence/absence of an entry, but more sophisticated checks could be added in the future.
I'm not 100% certain this is the best location, but I think it fits. I'd like to see the rustc guide continue to grow, including things like linking information and more platform-specific details.
Update cargo
7 commits in 1653f354644834073d6d2541e27fae94588e685e..ab32ee88dade1b50c77347599e82ca2de3fb8a51
2020-08-04 23:14:37 +0000 to 2020-08-10 17:44:43 +0000
- Build manpage archive deterministically (rust-lang/cargo#8600)
- doc: Qualify GNU licenses in example license field (rust-lang/cargo#8604)
- Fix jobserver_exists test on single-cpu systems (rust-lang/cargo#8598)
- Fix small typo in reference/profiles.md (rust-lang/cargo#8605)
- Default cargo publish to the alt registry if it's the only allowed one (rust-lang/cargo#8571)
- cargo install with specific yanked version gives confusing "not found" error (rust-lang/cargo#8565)
- Fix typo (rust-lang/cargo#8589)
Add basic test
And also run fmt which is where the other changes are from
Fix mut issues
These only appear when running tests, so resolved by adding mut
Swap order of forget
Add pub and rm guard impl
Add explicit type to guard
Add safety note
Change guard type from T to S
It should never have been T, as it guards over [MaybeUninit<S>; N]
Also add feature to test
This creates the language item for arrays, and adds the map fn which is like map in options or
iterators. It currently allocates an extra array, unfortunately.
Added fixme for transmuting
Fix typo
Add drop guard
Fix suggestion to use lifetime in type and in assoc const
_Do not merge until #75363 has landed, as it has the test case for this._
* Account for associated types
* Associated `const`s can't have generics (fix#74264)
* Do not suggest duplicate lifetimes and suggest `for<'a>` more (fix#72404)
While debugging a codegen issue, I tried adding LLVM options with
the rustc -Cllvm-args option, and was confused by the error and usage
messaging.
The LLVM "program name" argument is set to "rustc", and command line
error messages make it look like invalid arguments are "rustc"
arguments, not LLVM.
I changed this argument so error messages and the "-help" usage feedback
is easier to understand and react to. (Clang does something similar.)
Miri: improve spans of required_const failures
In https://github.com/rust-lang/rust/pull/75339 I added a loop evaluating all consts required by a function body. Unfortunately, if one of their evaluations fails, then the span used for that was that of the first statement in the function body, which happened to work form some existing test but is not sensible in general.
This PR changes it to point to the whole function instead, which is at least not wrong.
r? @oli-obk
Handle projection predicates in the param env for auto-trait docs
Fixes#72213
Any predicates in the param env are guaranteed to hold, so we don't need
to do any additional processing of them if we come across them as
sub-obligations of a different predicate. This allows us to avoid adding
the same predicate to the computed ParamEnv multiple times (but with
different regions each time), which causes an ambiguity error during
fulfillment.
Rollup of 9 pull requests
Successful merges:
- #74521 (older toolchains not valid anymore)
- #74960 (Fix regionck failure when converting Index to IndexMut)
- #75234 (Update asm! documentation in unstable book)
- #75368 (Move to doc links inside the prelude)
- #75371 (Move to doc links inside std/time.rs)
- #75394 (Add a function to `TyCtxt` for computing an `Allocation` for a `static` item's initializer)
- #75395 (Switch to intra-doc links in library/std/src/os/*/fs.rs)
- #75422 (Accept more safety comments)
- #75424 (fix wrong word in documentation)
Failed merges:
r? @ghost
Accept more safety comments
This accepts more `// SAFETY:` comments from `tidy`.
This is done after the current behaviour of requiring text one the same line (because spaces are stripped so the last space never pass if there is no text on the same line) bit me once more in #75066
This could potentially accept empty `// SAFETY:` comments but `tidy` is an internal tool used only here so my reasoning is reviews will catch those.
Fix regionck failure when converting Index to IndexMut
Fixes#74933
Consider an overloaded index expression `base[index]`. Without knowing whether it will be mutated, this will initially be desugared into `<U as Index<T>>::index(&base, index)` for some `U` and `T`. Let `V` be the `expr_ty_adjusted` of `index`.
If this expression ends up being used in any mutable context (or used in a function call with `&mut self` receiver before #72280), we will need to fix it up. The current code will rewrite it to `<U as IndexMut<V>>::index_mut(&mut base, index)`. In most cases this is fine as `V` will be equal to `T`, however this is not always true when `V/T` are references, as they may have different region.
This issue is quite subtle before #72280 as this code path is only used to fixup function receivers, but after #72280 we've made this a common path.
The solution is basically just rewrite it to `<U as IndexMut<T>>::index_mut(&mut base, index)`. `T` can retrieved in the fixup path using `node_substs`.
A few updates:
- Some minor wording and formatting changes.
- Remove the `cargo` column.
- Explain the columns up-front.
- Add no-wrap on the target-triple, which looks better to me.
- Minor mention on how to install support for a built-in target via rustup.
Reland #74069
Investigation in #74716 has concluded that this PR is indeed not a regression (and in fact the rollup itself is not either).
This reverts the revert in #74611.
r? @nnethercote cc @eddyb
This reverts commit 25670749b4.
This commit does not actually fix the problem. It merely removes the name of
the argument from the LLVM output. Even without the name, Rust codegen still
spills the (nameless) variable onto the stack which is the root cause. The root
cause is solved in the next commit.
move Deaggregate pass to post_borrowck_cleanup
Reopen of #71946
Only the second commit is from this PR, the other commit is a bugfix that's in the process of getting merged. I'll rebase once that's done
In #70073 MIR pass handling got reorganized, but with the goal of not changing behavior (except for disabling some optimizations on opt-level = 0). But there we realized that the Deaggregator pass, while conceptually more of a "cleanup" pass (and one that should be run before optimizations), was run in the middle of the optimization chain. Likely this is an accident of history, so I suggest we try and clean that up by making it a proper cleanup pass.
This does change mir-opt output, because deaggregation now runs before const-prop instead of after.
r? @wesleywiser @rust-lang/wg-mir-opt
cc @RalfJung
Add sample fix for E0749
Even though the description is clear but the solution may not be as straightforward.
Adding a suggested fix from documentation side.
r? @GuillaumeGomez
However, this suggestion should be shown in rustc itself for easy fix, the documentation should also reflect on the changes in rustc. Currently,
```
error[E0749]: negative impls cannot have any items
--> test.rs:6:5
|
6 | type Foo = i32; // error!
| ^^^^^^^^^^^^^^^
error: aborting due to previous error
For more information about this error, try `rustc --explain E0749`.
```
rustc should tell the user to remove it.
Tweak conditions for E0026 and E0769
When we have a tuple struct used with struct we don't want to suggest using the (valid) struct syntax with numeric field names. Instead we want to suggest the expected syntax.
Given
```rust
fn main() {
match MyOption::MySome(42) {
MyOption::MySome { x: 42 } => (),
_ => (),
}
}
```
We now emit E0769 "tuple variant `MyOption::MySome` written as struct variant" instead of E0026 "variant `MyOption::MySome` does not have a field named `x`".
move stack size check to const_eval machine
This is consistent with how we enforce the step limit. In particular, we do not want this limit checked for Miri-the-tool.
polymorphize: constrain unevaluated const handling
This PR constrains the support added for handling unevaluated consts in polymorphization (introduced in #75260) by:
- Skipping associated constants as this causes cycle errors.
- Skipping promoted constants when they contain `Self` as this ensures `T` is used in constants of the form `<Self as Foo<T>>`.
Due to an oversight on my part, when landing #75260 and #75255, some tests started failing when polymorphization was enabled that I didn't notice until after landing - this PR fixes the regressions from #75260.
r? @lcnr
Miri: Renamed "undef" to "uninit"
Renamed remaining references to "undef" to "uninit" when referring to Miri.
Impacted directories are:
- `src/librustc_codegen_llvm/consts.rs`
- `src/librustc_middle/mir/interpret/`
- `src/librustc_middle/ty/print/pretty.rs`
- `src/librustc_mir/`
- `src/tools/clippy/clippy_lints/src/consts.rs`
Upon building Miri based on the new changes it was verified that no changes needed to be made with the Miri project.
Related issue #71193
Use existing `infcx` when emitting trait impl diagnostic
Fixes#75361Fixes#74918
Previously, we were creating a new `InferCtxt`, which caused an ICE when
used with type variables from the existing `InferCtxt`
instance: only polymorphize upvar substs
This PR restricts the substitution polymorphization added in #75255 to only apply to the tupled upvar substitution, rather than all substitutions, fixing a bunch of regressions when polymorphization is
enabled.
Due to an oversight on my part, when landing #75260 and #75255, some tests started failing when polymorphization was enabled that I didn't notice until after landing - this PR fixes the regressions from #75255. #75336 has been filed to make sure that we don't forget to try make this change again in future, as it does enable some optimisations.
r? @lcnr
Don't try to use wasm intrinsics on vectors
This commit fixes an issue with #74695 where the fptosi and fptoui
specializations on wasm were accidentally used on vector types by the
`simd_cast` intrinsic. This issue showed up as broken CI for the stdsimd
crate. Here this commit simply skips the specialization on vector kinds
flowing into `fpto{s,u}i`.
Avoid deleting temporary files on error
Previously if the compiler error'd, fatally, then temporary directories which
should be preserved by -Csave-temps would be deleted due to fatal compiler
errors being implemented as panics.
cc @infinity0
(Hopefully) fixes#75275, but I haven't tested
Clean up errors in typeck and resolve
* Tweak ordering of suggestions
* Do not suggest similarly named enclosing item
* Point at item definition in foreign crates
* Add missing primary label
CC #34255.
Remove restriction on type parameters preceding consts w/ feature const-generics
Removed the restriction on type parameters preceding const parameters when the feature const-generics is enabled.
Builds on #74676, which deals with unsorted generic parameters. This just lifts the check in lowering the AST to HIR that permits consts and types to be reordered with respect to each other. Lifetimes still must precede both
This change is not intended for min-const-generics, and is gated behind the `#![feature(const_generics)]`.
One thing is that it also permits type parameters without a default to come after consts, which I expected to not work, and was hoping to get more guidance on whether that should be permitted or how to prevent it otherwise.
I did not go through the RFC process for this pull request because there was prior work to get this feature added. In the previous PR that was cited, work was done to enable this change.
r? @lcnr
This commit constrains the support added for handling unevaluated consts
in polymorphization (introduced in #75260) by:
- Skipping associated constants as this causes cycle errors.
- Skipping promoted constants when they contain `Self` as this ensures
`T` is used in constants of the form `<Self as Foo<T>>`.
Signed-off-by: David Wood <david@davidtw.co>
