Compute generator saved locals on MIR
Generators are currently type-checked by introducing a `witness` type variable, which is unified with a `GeneratorWitness(captured types)` whose purpose is to ensure that the auto traits correctly migrate from the captured types to the `witness` type. This requires computing the captured types on HIR during type-checking, only to re-do it on MIR later.
This PR proposes to drop the HIR-based computation, and only keep the MIR one. This is done in 3 steps.
1. During type-checking, the `witness` type variable is never unified. This allows to stall all the obligations that depend on it until the end of type-checking. Then, the stalled obligations are marked as successful, and saved into the typeck results for later verification.
2. At type-checking writeback, `witness` is replaced by `GeneratorWitnessMIR(def_id, substs)`. From this point on, all trait selection involving `GeneratorWitnessMIR` will fetch the MIR-computed locals, similar to what opaque types do. There is no lifetime to be preserved here: we consider all the lifetimes appearing in this witness type to be higher-ranked.
3. After borrowck, the stashed obligations are verified against the actually computed types, in the `check_generator_obligations` query. If any obligation was wrongly marked as fulfilled in step 1, it should be reported here.
There are still many issues:
- ~I am not too happy having to filter out some locals from the checked bounds, I think this is MIR building that introduces raw pointers polluting the analysis;~ solved by a check specific to static variables.
- the diagnostics for captured types don't show where they are used/dropped;
- I do not attempt to support chalk.
cc `@eholk` `@jyn514` for the drop-tracking work
r? `@oli-obk` as you warned me of potential unsoundness
Sync rustc_codegen_cranelift
For cg_clif itself there have been a couple of bug fixes since the last sync, a Cranelift update and implemented all remaining simd platform intrinsics used by `std::simd`. (`std::arch` still misses a lot though) Most of the diff is from reworking of the cg_clif build system though.
r? `@ghost`
`@rustbot` label +A-codegen +A-cranelift +T-compiler
Move format_args!() into AST (and expand it during AST lowering)
Implements https://github.com/rust-lang/compiler-team/issues/541
This moves FormatArgs from rustc_builtin_macros to rustc_ast_lowering. For now, the end result is the same. But this allows for future changes to do smarter things with format_args!(). It also allows Clippy to directly access the ast::FormatArgs, making things a lot easier.
This change turns the format args types into lang items. The builtin macro used to refer to them by their path. After this change, the path is no longer relevant, making it easier to make changes in `core`.
This updates clippy to use the new language items, but this doesn't yet make clippy use the ast::FormatArgs structure that's now available. That should be done after this is merged.
use `LocalDefId` instead of `HirId` in trait resolution to simplify the obligation clause resolution
This commit introduces a refactoring suggested by `@lcnr` to simplify the obligation clause resolution.
This is just the first PR that introduces a type of refactoring, but others PRs will follow this to introduce name changing to change from the variable name from `body_id` to something else.
Fixes https://github.com/rust-lang/rust/issues/104827
Signed-off-by: Vincenzo Palazzo <vincenzopalazzodev@gmail.com>
`@rustbot` r? `@lcnr`
Update cargo
11 commits in 985d561f0bb9b76ec043a2b12511790ec7a2b954..3c5af6bed9a1a243a693e8e22ee2486bd5b82a6c 2023-01-20 14:39:28 +0000 to 2023-01-24 15:48:15 +0000
- Add a note about verifying your email address on crates.io (rust-lang/cargo#11620)
- Improve CI caching by skipping mtime checks for paths in $CARGO_HOME (rust-lang/cargo#11613)
- test: Update for clap 4.1.3 (rust-lang/cargo#11619)
- Fix unused attribute on Windows. (rust-lang/cargo#11614)
- [Doc]: Added links to the `Target` section of the glossary for occurences of `target triple` (rust-lang/cargo#11603)
- feat: stabilize auto fix note (rust-lang/cargo#11558)
- Clarify the difference between CARGO_CRATE_NAME and CARGO_PKG_NAME (rust-lang/cargo#11576)
- Temporarily pin libgit2-sys. (rust-lang/cargo#11609)
- Disable network SSH tests on windows. (rust-lang/cargo#11610)
- fix(toml): Add `default-features` to `TomlWorkspaceDependency` (rust-lang/cargo#11409)
- doc(contrib): remove rls in release process (rust-lang/cargo#11601)
r? `@ghost`
11 commits in 985d561f0bb9b76ec043a2b12511790ec7a2b954..3c5af6bed9a1a243a693e8e22ee2486bd5b82a6c
2023-01-20 14:39:28 +0000 to 2023-01-24 15:48:15 +0000
- Add a note about verifying your email address on crates.io (rust-lang/cargo#11620)
- Improve CI caching by skipping mtime checks for paths in $CARGO_HOME (rust-lang/cargo#11613)
- test: Update for clap 4.1.3 (rust-lang/cargo#11619)
- Fix unused attribute on Windows. (rust-lang/cargo#11614)
- [Doc]: Added links to the `Target` section of the glossary for occurences of `target triple` (rust-lang/cargo#11603)
- feat: stabilize auto fix note (rust-lang/cargo#11558)
- Clarify the difference between CARGO_CRATE_NAME and CARGO_PKG_NAME (rust-lang/cargo#11576)
- Temporarily pin libgit2-sys. (rust-lang/cargo#11609)
- Disable network SSH tests on windows. (rust-lang/cargo#11610)
- fix(toml): Add `default-features` to `TomlWorkspaceDependency` (rust-lang/cargo#11409)
- doc(contrib): remove rls in release process (rust-lang/cargo#11601)
use LocalDefId instead of HirId in trait resolution to simplify
the obligation clause resolution
Signed-off-by: Vincenzo Palazzo <vincenzopalazzodev@gmail.com>
remove error code from `E0789`, add UI test/docs
`E0789` shouldn't have an error code, it's explicitly internal-only and is tiny in scope. (I wonder if we can tighten the standard for this in the RFC?) I also added a UI test and error docs (done like `E0208`, they are "no longer emitted").
r? `@GuillaumeGomez` (shouldn't need a compiler review, it's pretty minor)
Enable sanitizers for s390x-linux
Include sanitizers supported by LLVM on s390x (asan, lsan, msan, tsan) in the target definition, as well as in the compiletest supported list.
Build sanitizer runtime for the target. Enable sanitizers in the CI.
Implement some more predicates in the new solver
Implement a few more goals. The subtype goal specifically is important, since it's required for this code to compile:
```
fn main() {
let mut x = vec![];
x.push(1i32);
}
```
(I think we emit a subtype goal here because of coercion).
Drive-by: Also implements `--compare-mode=next-solver` -- I've been using this locally a lot to find out what works and what doesn't. I'm also happy to split this out into another PR.
r? `@lcnr`
Remap paths in UI tests by default
If you think this needs further discussions / something RFC-like, please let me know the best forum for that.
This PR runs UI tests with a remapped "src base" directory by default.
Why? Because some UI tests currently depend on the length of the absolute path to the `src/test/ui` directory. Remapping makes the tests independent of the absolute path.
The path to the source file (which is absolute on CI) is part of the type name of closures. `rustc` diagnostic output depends on the length of type names (long type names are truncated). So a long absolute path leads to long closure type names, which leads to truncation and changed diagnostics.
(I initially tried just disabling type name truncation, but that made some error messages stupid long (thousands of characters, IIRC)).
Additional changes:
* All boolean `compiletest` directives now support explicit `no-` versions to disable them.
* Adapt existing tests when necessary:
* Disable remapping for individual tests that fail with it enabled (when there's no obvious alternative fix).
* For tests that already check something remapping related switch to the new option unless we gain something significant by keeping the manual remap.
Passed Windows CI in https://github.com/rust-lang/rust/actions/runs/3933100590
Use UnordMap and UnordSet for id collections (DefIdMap, LocalDefIdMap, etc)
This PR changes the `rustc_data_structures::define_id_collections!` macro to use `UnordMap` and `UnordSet` instead of `FxHashMap` and `FxHashSet`. This should account for a large portion of hash-maps being used in places where they can cause trouble.
The changes required are moderate but non-zero:
- In some places the collections are extracted into sorted vecs.
- There are a few instances where for-loops have been changed to extends.
~~Let's see what the performance impact is. With a bit more refactoring, we might be able to get rid of some of the additional sorting -- but the change set is already big enough. Unless there's a performance impact, I'd like to do further changes in subsequent PRs.~~
Performance does not seem to be negatively affected ([perf-run here](https://github.com/rust-lang/rust/pull/106977#issuecomment-1396776699)).
Part of [MCP 533](https://github.com/rust-lang/compiler-team/issues/533).
r? `@ghost`
fix check macro expansion
If the only argument to `check!` is the module name I get this error:
```
error: expected expression, found `,`
--> src/tools/tidy/src/main.rs:63:42
|
57 | / macro_rules! check {
58 | | ($p:ident $(, $args:expr)* ) => {
59 | | drain_handles(&mut handles);
60 | |
... |
63 | | $p::check($($args),* , &mut flag);
| | ^ expected expression
... |
69 | | }
70 | | }
| |_________- in this expansion of `check!`
...
117 | check!(hey);
| ----------- in this macro invocation
```
This change makes it so commas are added only when there are `args`.
r? ```@albertlarsan68```
check for x version updates
This PR adds a check to tidy to assert that the installed version of `x` is equal to the version in `src/tools/x/Cargo.toml`. It checks the installed version of `x` by parsing the output of `cargo install --list` (as an option proposed in this [issue](https://github.com/rust-lang/rust/issues/106469)).
It does not warn if `x` has not yet been installed, on the assumption that the user isn't interested in using it.
Update cargo
3 commits in 50eb688c2bbea5de5a2e8496230a7428798089d1..985d561f0bb9b76ec043a2b12511790ec7a2b954
2023-01-19 10:09:05 +0000 to 2023-01-20 14:39:28 +0000
- Stabilize sparse-registry (rust-lang/cargo#11224)
- Wrapper type for data that should never be logged (rust-lang/cargo#11545)
- Add semver rule for lints (rust-lang/cargo#11596)
r? `@ghost`
Include sanitizers supported by LLVM on s390x (asan, lsan, msan, tsan)
in the target definition, as well as in the compiletest supported list.
Build sanitizer runtime for the target. Enable sanitizers in the CI.
3 commits in 50eb688c2bbea5de5a2e8496230a7428798089d1..985d561f0bb9b76ec043a2b12511790ec7a2b954
2023-01-19 10:09:05 +0000 to 2023-01-20 14:39:28 +0000
- Stabilize sparse-registry (rust-lang/cargo#11224)
- Wrapper type for data that should never be logged (rust-lang/cargo#11545)
- Add semver rule for lints (rust-lang/cargo#11596)
Clippy: Make sure to include in beta: Move `unchecked_duration_subtraction` to pedantic
This was merged one day too late in order to make it into the last sync. But since we talked about moving this lint to `pedantic` in the meeting, I'd like to get this in rather sooner than later.
https://github.com/rust-lang/rust-clippy/pull/10194
r? ``@Manishearth``
3 commits in a5d47a72595dd6fbe7d4e4f6ec20dc5fe724edd1..50eb688c2bbea5de5a2e8496230a7428798089d1
2023-01-16 18:51:50 +0000 to 2023-01-19 10:09:05 +0000
- Normalize git deps when doing `cargo vendor` for resolving deps inherited from a workspace (rust-lang/cargo#11414)
- Ignore `workspace.default-members` when running `cargo install` on root package of a non-virtual workspace (rust-lang/cargo#11067)
- Corrected documentation of how to cache binaries installed with `cargo install` in CI workflows (rust-lang/cargo#11592)
Remove extra removal from test path
I don't know how to describe it shortly so better show what it's doing instead. Currently, there is one extra "rust/" before the test folder when running tests:
```
failures:
---- [rustdoc] rust/tests/rustdoc/redirect.rs stdout ----
```
This is a bit annoying when copying the test path. This is due to the moving of the `tests` folder one level up, meaning we were trimming too much of the `root_path`.
Now it is again displaying the correct path:
```
failures:
---- [rustdoc] tests/rustdoc/redirect.rs stdout ----
```
document + UI test `E0208` and make its output more user-friendly
Cleans up `E0208`'s output a lot. It could actually be useful for someone learning about variance now. I also added a UI test for it in `tests/ui/error-codes/` and wrote some docs for it.
r? `@GuillaumeGomez` another error code, can't be bothered to find the issue :P. Obviously there's some compiler stuff, so you'll have to hand it off.
Part of https://github.com/rust-lang/rust/issues/61137.
