Remove `Session::one_time_diagnostic`
This is untracked mutable state, which modified the behaviour of queries.
It was used for 2 things: some full-blown errors, but mostly for lint declaration notes ("the lint level is defined here" notes).
It is replaced by the diagnostic deduplication infra which already exists in the diagnostic emitter.
A new diagnostic level `OnceNote` is introduced specifically for lint notes, to deduplicate subdiagnostics.
As a drive-by, diagnostic emission takes a `&mut` to allow dropping the `SubDiagnostic`s.
resolve: Do not build expensive suggestions if they are not actually used
And remove a bunch of (conditionally) unused parameters from path resolution functions.
This helps with performance issues in https://github.com/rust-lang/rust/pull/94857, and should be helpful in general even without that.
Rollup of 5 pull requests
Successful merges:
- #94391 (Fix ice when error reporting recursion errors)
- #94655 (Clarify which kinds of MIR are allowed during which phases.)
- #95179 (Try to evaluate in try unify and postpone resolution of constants that contain inference variables)
- #95270 (debuginfo: Fix debuginfo for Box<T> where T is unsized.)
- #95276 (add diagnostic items for clippy's `trim_split_whitespace`)
Failed merges:
r? `@ghost`
`@rustbot` modify labels: rollup
debuginfo: Fix debuginfo for Box<T> where T is unsized.
Before this fix, the debuginfo for the fields was generated from the struct defintion of Box<T>, but (at least at the moment) the compiler pretends that Box<T> is just a (fat) pointer, so the fields need to be `pointer` and `vtable` instead of `__0: Unique<T>` and `__1: Allocator`.
This is meant as a temporary mitigation until we can make sure that simply treating Box as a regular struct in debuginfo does not cause too much breakage in the ecosystem.
r? ````@wesleywiser````
Also remove a redundant parameter from `fn resolve_path(_with_ribs)`, `crate_lint: CrateLint` is a more detailed version of `record_used: bool` with `CrateLint::No` meaning `false` and anything else meaning `true`.
Separate const prop lints from optimizations
r? `@oli-obk`
Separates lints and optimizations during const prop by moving the lints into their own file and checking them during post borrowck cleanup.
Thanks to `@oli-obk` for mentoring me.
Change Thir to lazily create constants
To allow `AbstractConst`s to work with the previous thir changes we made and those we want to make, i.e. to avoid problems due to `ValTree` and `ConstValue` conversions, we instead switch to a thir representation for constants that allows us to lazily create constants.
r? `@oli-obk`
Properly track `ImplObligations`
Instead of probing for all possible `impl`s that could have caused an
`ImplObligation`, keep track of its `DefId` and obligation spans for
accurate error reporting.
Follow to #89580. Addresses #89418.
Before this fix, the debuginfo for the fields was generated from the
struct defintion of Box<T>, but (at least at the moment) the compiler
pretends that Box<T> is just a (fat) pointer, so the fields need to be
`pointer` and `vtable` instead of `__0: Unique<T>` and `__1: Allocator`.
This is meant as a temporary mitigation until we can make sure that
simply treating Box as a regular struct in debuginfo does not cause too
much breakage in the ecosystem.
Instead of probing for all possible impls that could have caused an
`ImplObligation`, keep track of its `DefId` and obligation spans for
accurate error reporting.
Follow up to #89580. Addresses #89418.
Remove some unnecessary clones.
Tweak output for auto trait impl obligations.
Rollup of 5 pull requests
Successful merges:
- #94249 (Better errors when a Copy impl on a Struct is not self-consistent)
- #95069 (Fix auto traits in rustdoc)
- #95221 (interpret/memory: simplify check_and_deref_ptr)
- #95225 (remove `[async output]` from `impl Future` pretty-printing)
- #95238 (Stop emitting E0026 for struct enums with underscores)
Failed merges:
r? `@ghost`
`@rustbot` modify labels: rollup
interpret/memory: simplify check_and_deref_ptr
*Finally* I saw a way to make this code simpler. The odd preprocessing in `let ptr_or_addr =` has bothered me since forever, but it actually became unnecessary in the last provenance refactoring. :)
This also leads to slightly more explicit error messages as a nice side-effect. 🎉
r? `@oli-obk`
Better errors when a Copy impl on a Struct is not self-consistent
As discovered in a Zulip thread with `@nnethercote` and `@Mark-Simulacrum,` it's not immediately obvious why a field on an ADT doesn't implement `Copy`. This PR attempts to give slightly more detailed information by spinning up a fulfillment context to try to dig down and discover transitive fulfillment errors that cause `is_copy_modulo_regions` to fail on a ADT field.
The error message still kinda sucks, but should only show up in the case that an existing error message was totally missing... so I think it's a good compromise for now?
diagnostics: do not suggest `fn foo({ <body> }`
Instead of suggesting that the body always replace the last character on the line, presuming it must be a semicolon, the parser should instead check what the last character is, and append the body if it is anything else.
Fixes#83104
Rename `~const Drop` to `~const Destruct`
r? `@oli-obk`
Completely switching to `~const Destructible` would be rather complicated, so it seems best to add it for now and wait for it to be backported to beta in the next release.
The rationale is to prevent complications such as #92149 and #94803 by introducing an entirely new trait. And `~const Destructible` reads a bit better than `~const Drop`. Name Bikesheddable.
Modernize `alloc-no-oom-handling` test
- The edition should be 2021 to avoid warnings.
- The `external_crate` feature was removed in commit 45bf1ed1a1 ("rustc: Allow changing the default allocator").
Note that commit d620ae1070 ("Auto merge of #84266") removed the old test, but the new one introduced passed the `--cfg` like in the old one.
Signed-off-by: Miguel Ojeda <ojeda@kernel.org>
---
This is intended to align this test to the new `no_rc` and `no_sync` ones being added in https://github.com/rust-lang/rust/pull/89891, but it makes sense on its own too.
Fold aarch64 feature +fp into +neon
Arm's FEAT_FP and Feat_AdvSIMD describe the same thing on AArch64:
The Neon unit, which handles both floating point and SIMD instructions.
Moreover, a configuration for AArch64 must include both or neither.
Arm says "entirely proprietary" toolchains may omit floating point:
https://developer.arm.com/documentation/102374/0101/Data-processing---floating-point
In the Programmer's Guide for Armv8-A, Arm says AArch64 can have
both FP and Neon or neither in custom implementations:
https://developer.arm.com/documentation/den0024/a/AArch64-Floating-point-and-NEON
In "Bare metal boot code for Armv8-A", enabling Neon and FP
is just disabling the same trap flag:
https://developer.arm.com/documentation/dai0527/a
In an unlikely future where "Neon and FP" become unrelated,
we can add "[+-]fp" as its own feature flag.
Until then, we can simplify programming with Rust on AArch64 by
folding both into "[+-]neon", which is valid as it supersets both.
"[+-]neon" is retained for niche uses such as firmware, kernels,
"I just hate floats", and so on.
I am... pretty sure no one is relying on this.
An argument could be made that, as we are not an "entirely proprietary" toolchain, we should not support AArch64 without floats at all. I think that's a bit excessive. However, I want to recognize the intent: programming for AArch64 should be simplified where possible. For x86-64, programmers regularly set up illegal feature configurations because it's hard to understand them, see https://github.com/rust-lang/rust/issues/89586. And per the above notes, plus the discussion in https://github.com/rust-lang/rust/issues/86941, there should be no real use cases for leaving these features split: the two should in fact always go together.
- Fixesrust-lang/rust#95002.
- Fixesrust-lang/rust#95064.
- Fixesrust-lang/rust#95122.
Instead of suggesting that the body always replace the last character on the
line, presuming it must be a semicolon, the parser should instead check what
the last character is, and append the body if it is anything else.
Fixes#83104
Arm's FEAT_FP and Feat_AdvSIMD describe the same thing on AArch64:
The Neon unit, which handles both floating point and SIMD instructions.
Moreover, a configuration for AArch64 must include both or neither.
Arm says "entirely proprietary" toolchains may omit floating point:
https://developer.arm.com/documentation/102374/0101/Data-processing---floating-point
In the Programmer's Guide for Armv8-A, Arm says AArch64 can have
both FP and Neon or neither in custom implementations:
https://developer.arm.com/documentation/den0024/a/AArch64-Floating-point-and-NEON
In "Bare metal boot code for Armv8-A", enabling Neon and FP
is just disabling the same trap flag:
https://developer.arm.com/documentation/dai0527/a
In an unlikely future where "Neon and FP" become unrelated,
we can add "[+-]fp" as its own feature flag.
Until then, we can simplify programming with Rust on AArch64 by
folding both into "[+-]neon", which is valid as it supersets both.
"[+-]neon" is retained for niche uses such as firmware, kernels,
"I just hate floats", and so on.
- The edition should be 2021 to avoid warnings.
- The `external_crate` feature was removed in commit 45bf1ed1a1
("rustc: Allow changing the default allocator").
Note that commit d620ae1070 ("Auto merge of #84266") removed
the old test, but the new one introduced passed the `--cfg` like
in the old one.
Signed-off-by: Miguel Ojeda <ojeda@kernel.org>
Suggest constraining param for unary ops when missing trait impl
This PR adds a suggestion of constraining param for unary ops `-` and `!` when the corresponding trait implementation
is missing.
Fixs #94543.
BTW, this is my first time to touch rustc, please correct me if I did anything wrong.
Fixes 5270
Previously, rustfmt only checked the `merge_derives` configuration value
to determine if it should merge_derives. This lead to derives being
merged even when annotated with the `rustfmt::skip` attribute.
Now, rustfmt also checks if derives are explicitly being skipped in the
current context via the `rustfmt::skip` attribute.
Return err instead of ICE
Having `escaping_bound_vars` results in ICE when trying to create `ty::Binder::dummy`, to avoid it we return err like the line above. I think this requires a more sophisticated fix, I would love to investigate if mentorship is available 🤓Fixes#95023 and #85350
Don't run UB in test suite
This splits `ui/unsafe/union.rs` to make it so only the non-UB parts are run. It also means we can do more testing of the location of error messages (which are a bit different with the THIR unsafety checker). `union-modification.rs` has no UB (according to Miri), and `union.rs` has errors (but would have UB if not for those errors).
Closes#95075.
r? `@bjorn3`
Fixes 5273
Previously, rustfmt searched for the start of a struct body after the
opening `{`. In most cases this works just fine, but const values can
also be defined between `{ }`, which lead to issues when rewriting the
struct body.
Now, rustfmt will search for the `{` after the generic argument list to
guarantee that the `{` it finds is the start of the struct body.
suggest removing type ascription in bad parsing position
Not sure how to test this with the non-nightly suggestion. Didn't add a new UI test because it already manifests in an existing UI test.
Fixes#95014
Fix debuginfo tests with GDB 11.2
GDB 11.2 added support for DW_ATE_UTF, which caused some test
failures. This fixes these tests by changing the format that is used,
and adds a new test to verify that characters are emitted as something
that GDB can print in a char-like way.
Fixes#94458
solaris build environment should include libsendfile/liblgrp
As of version 0.2.120 of the libc crate, the solaris target now requires
some additional libraries to be present in the sysroot. Note that the
solaris target doesn't really build against files from Solaris, but
rather against some files from DilOS (a platform similar to both Solaris
and illumos). Pull in the extra libraries and their compilation links
from that apt repository.
This aims to assist with rust-lang/rust#94052.
Update cargo
9 commits in 65c82664263feddc5fe2d424be0993c28d46377a..109bfbd055325ef87a6e7f63d67da7e838f8300b
2022-03-09 02:32:56 +0000 to 2022-03-17 21:43:09 +0000
- Refactor RegistryData::load to handle management of the index cache (rust-lang/cargo#10482)
- Separate VCS command paths with "--" (rust-lang/cargo#10483)
- Fix panic when artifact target is used for `[target.'cfg(<target>)'.dependencies` (rust-lang/cargo#10433)
- Bump git2@0.14.2 and libgit2-sys@0.13.2 (rust-lang/cargo#10479)
- vendor: Don't allow multiple values for --sync (rust-lang/cargo#10448)
- Use types to make clere (credential process || token) (rust-lang/cargo#10471)
- Warning on conflicting keys (rust-lang/cargo#10316)
- Registry functions return Poll to enable parallel fetching of index data (rust-lang/cargo#10064)
- Refine the contributor guide (rust-lang/cargo#10468)
Extend --check-cfg tests to all predicate inside all/any
Now that https://github.com/rust-lang/rust/pull/94295 is merged it's time to add more tests to check that all predicate inside `all` and `any` are always checked.
r? `@petrochenkov`
Re-enable parallel debuginfo tests
Debuginfo tests are serialized due to some older version of LLDB.
However, that comment was last touched in 2014, so presumably these
older versions are long since obsolete.
Partially fixes bug #72719.
Compare installed browser-ui-test version to the one used in CI
I happened a few times to run into (local) rustdoc GUI tests errors because I forgot to update my browser-ui-test version. I know at least two others who encountered the same problem so I think emitting a warning to let us know about this version mismatch would make it easier to figure out.
So now, I'm not too sure that this PR is the right approach because it requires to parse a Dockerfile, which feels pretty bad. I had the idea to instead store the browser-ui-test version into a docker ARG like:
```docker
ARG BROWSER_UI_TEST_VERSION=0.8.0
```
And then use it as such in the command to make the parsing more reliable.
Or we could store this version into a file and import this file into the Dockerfile and read it from the builder.
Any preference or maybe another solution?
r? ``@Mark-Simulacrum``
Always evaluate all cfg predicate in all() and any()
This pull-request adjust the handling of the `all()` and `any()` to always evaluate every cfg predicate because not doing so result in accepting incorrect `cfg`:
```rust
#[cfg(any(unix, foo::bar))] // Should error on foo::bar, but does not on unix platform (but does on non unix platform)
fn foo1() {}
#[cfg(all(foo, foo::bar))] // Should error on foo::bar, but does not
fn foo2() {}
#[cfg(all(foo::bar, foo))] // Correctly error on foo::bar
fn foo3() {}
#[cfg(any(foo::bar, foo))] // Correctly error on foo::bar
fn foo4() {}
```
This pull-request take the side to directly turn it into a hard error instead of having a future incompatibility lint because the combination to get this incorrect behavior is unusual and highly probable that some code have this without noticing.
A [search](https://cs.github.com/?scopeName=All+repos&scope=&q=lang%3Arust+%2Fany%5C%28%5Ba-zA-Z%5D%2C+%5Ba-zA-Z%5D%2B%3A%3A%5Ba-zA-Z%5D%2B%2F) on Github reveal no such instance nevertheless a Crater run should probably be done before merging this.
This was discover in https://github.com/rust-lang/rust/pull/94175 when trying to lint on the second predicate. Also note that this seems to have being introduce with Rust 1.27.0: https://rust.godbolt.org/z/KnfqKv15f.
r? `@petrochenkov`
Debuginfo tests are serialized due to some older version of LLDB.
However, that comment was last touched in 2014, so presumably these
older versions are long since obsolete.
Partially fixes bug #72719.
Rollup of 4 pull requests
Successful merges:
- #95013 (Update browser-ui-test version to 0.8.2)
- #95039 (Make negative coherence work when there's impl negative on super predicates)
- #95047 (Refactor: remove an unnecessary pattern for ignoring all parts)
- #95048 (update Miri)
Failed merges:
r? `@ghost`
`@rustbot` modify labels: rollup
Update browser-ui-test version to 0.8.2
It brings mostly debugging improvements: it doesn't stop at the first failing command but rather at the first "fatal error".
r? `@notriddle`
Implement -Z oom=panic
This PR removes the `#[rustc_allocator_nounwind]` attribute on `alloc_error_handler` which allows it to unwind with a panic instead of always aborting. This is then used to implement `-Z oom=panic` as per RFC 2116 (tracking issue #43596).
Perf and binary size tests show negligible impact.
GDB 11.2 added support for DW_ATE_UTF, which caused some test
failures. This fixes these tests by changing the format that is used,
and adds a new test to verify that characters are emitted as something
that GDB can print in a char-like way.
Fixes#94458
Rollup of 10 pull requests
Successful merges:
- #91133 (Improve `unsafe` diagnostic)
- #93222 (Make ErrorReported impossible to construct outside `rustc_errors`)
- #93745 (Stabilize ADX target feature)
- #94309 ([generator_interior] Be more precise with scopes of borrowed places)
- #94698 (Remove redundant code from copy-suggestions)
- #94731 (Suggest adding `{ .. }` around a const function call with arguments)
- #94960 (Fix many spelling mistakes)
- #94982 (Add deprecated_safe feature gate and attribute, cc #94978)
- #94997 (debuginfo: Fix ICE when generating name for type that produces a layout error.)
- #95000 (Fixed wrong type name in comment)
Failed merges:
r? `@ghost`
`@rustbot` modify labels: rollup
[generator_interior] Be more precise with scopes of borrowed places
Previously the generator interior type checking analysis would use the nearest temporary scope as the scope of a borrowed value. This ends up being overly broad for cases such as:
```rust
fn status(_client_status: &Client) -> i16 {
200
}
fn main() {
let client = Client;
let g = move || match status(&client) {
_status => yield,
};
assert_send(g);
}
```
In this case, the borrow `&client` could be considered in scope for the entirety of the `match` expression, meaning it would be viewed as live across the `yield`, therefore making the generator not `Send`.
In most cases, we want to use the enclosing expression as the scope for a borrowed value which will be less than or equal to the nearest temporary scope. This PR changes the analysis to use the enclosing expression as the scope for most borrows, with the exception of borrowed RValues which are true temporary values that should have the temporary scope. There's one further exception where borrows of a copy such as happens in autoref cases also should be ignored despite being RValues.
Joint work with `@nikomatsakis`
Fixes#57017
r? `@tmandry`
Make ErrorReported impossible to construct outside `rustc_errors`
There are a few places were we have to construct it, though, and a few
places that are more invasive to change. To do this, we create a
constructor with a long obvious name.
cc #69426 `@varkor` `@eddyb` `@estebank`
I actually didn't see that I was assigned to this issue until now...
erase late-bound regions in dyn projection types for debuginfo
simply skipping the binder leaves late-bound regions that will cause debug assertions to fail when checking the layout of the projection ty, so let's erase the regions instead.
sorry for taking so long to put this up, had trouble getting rustc set up on a new computer.
fixes#94998
There are a few places were we have to construct it, though, and a few
places that are more invasive to change. To do this, we create a
constructor with a long obvious name.
There are a few places were we have to construct it, though, and a few
places that are more invasive to change. To do this, we create a
constructor with a long obvious name.
Update books
## reference
8 commits in 9d289c05fce7254b99c6a0d354d84abb7fd7a032..0a2fe6651fbccc6416c5110fdf5b93fb3cb29247
2022-02-23 08:58:20 -0800 to 2022-03-15 09:32:25 -0700
- Documentation PR for cfg_panic (rust-lang/reference#1157)
- Document aarch64 `target_feature` options (rust-lang/reference#1102)
- Try to clarify destructor not being run scenario. (rust-lang/reference#1107)
- Add undocumented Punctuation token Tilde `~` (rust-lang/reference#1149)
- update UB list for safe target_feature (rust-lang/reference#1050)
- Update const_eval.md for feature stabilization (rust-lang/reference#1166)
- Remove `.intel_syntax`/`.att_syntax` support entirely.
- Fix `.intel_syntax` directive
## book
3 commits in 3f255ed40b8c82a0434088568fbed270dc31bf00..036e88a4f135365de85358febe5324976a56030a
2022-02-27 21:26:12 -0500 to 2022-03-04 21:53:33 -0500
- Fix some links and small wordings
- Snapshot of chapter 19 for nostarch
- Clarify fully-qualified syntax explanation
## rust-by-example
2 commits in 2a928483a20bb306a7399c0468234db90d89afb5..d504324f1e7dc7edb918ac39baae69f1f1513b8e
2022-02-28 11:36:59 -0300 to 2022-03-07 09:26:32 -0300
- Fixed extra indentation at line 43 in Phantom Testcase example. (rust-lang/rust-by-example#1515)
- Typo fixed in description of inline ASM cpuid function (rust-lang/rust-by-example#1514)
## rustc-dev-guide
3 commits in 32f2a5b4e7545318846185198542230170dd8a42..0e4b961a9c708647bca231430ce1b199993e0196
2022-03-01 10:45:24 -0600 to 2022-03-14 08:40:37 -0700
- update winget install instructions to ensure proper packages are installed (-e for --exact, and full package names to ensure arbitrary packages from
the msstore source aren't installed)
- Add missing rustdoc tests explanations
- Fix incorrectly escaped backtick
Rollup of 5 pull requests
Successful merges:
- #94868 (Format core and std macro rules, removing needless surrounding blocks)
- #94951 (Extend the irrefutable_let_patterns lint to let chains)
- #94955 (Refactor: Use `format_args_capture` in some parts of `rustc_parse`)
- #94957 (Improve the explanation about the behaviour of read_line)
- #94974 (Ensure that `let_else` does not interact with `let_chains`)
Failed merges:
r? `@ghost`
`@rustbot` modify labels: rollup
Extend the irrefutable_let_patterns lint to let chains
Implements the suggestion from https://github.com/rust-lang/rust/pull/94927#issuecomment-1067078300
We only look for complete suffixes or prefixes of irrefutable let patterns, so
that an irrefutable let pattern in a chain surrounded by refutable ones is
not linted, as it is an useful pattern that has no low-cost replacement (unlike suffixes or prefixes which can just be copied outside of the `if`: either into the `if`'s block, or the block surrounding the `if`).
If all patterns in a let chain are irrefutable, we lint as well.
Depends on #94958 ~~so I included it into the PR for now~~ *which has been merged since*.
r? `@estebank`
cc `@joshtriplett` `@c410-f3r`
Format core and std macro rules, removing needless surrounding blocks
Many of the asserting and printing macros in `core` and `std` are written with prehistoric-looking formatting, like this:
335ffbfa54/library/std/src/macros.rs (L96-L101)
In modern Rust style this would conventionally be written as follows instead, always using braces and a trailing semicolon on the macro arms:
af53809c87/library/std/src/macros.rs (L98-L105)
Getting rid of the unneeded braces inside the expansion reduces extraneous indentation in macro-expanded code. For example:
```rust
println!("repro {}", true);
```
```rust
// before:
{
::std::io::_print(
::core::fmt::Arguments::new_v1(
&["repro ", "\n"],
&[::core::fmt::ArgumentV1::new_display(&true)],
),
);
};
```
```rust
// after:
::std::io::_print(
::core::fmt::Arguments::new_v1(
&["repro ", "\n"],
&[::core::fmt::ArgumentV1::new_display(&true)],
),
);
```
Only look for complete suffixes or prefixes of irrefutable let patterns, so
that an irrefutable let pattern in a chain surrounded by refutable ones is
not linted, as it is an useful pattern.
As of version 0.2.120 of the libc crate, the solaris target now requires
some additional libraries to be present in the sysroot. Note that the
solaris target doesn't really build against files from Solaris, but
rather against some files from DilOS (a platform similar to both Solaris
and illumos). Pull in the extra libraries and their compilation links
from that apt repository.
Add GUI test to check URLs
The first commit merges both file doing kinda the same thing. Didn't make much sense to keep them separated.
The second commit add checks to ensure the URL has the expected value depending if the search is displayed or not.
r? `@notriddle`
fix typos
Rework of #94603 which got closed as I was trying to unmerge and repush. This is a subset of changes from the original pr as I sed'd whatever typos I remembered from the original PR
thanks to `@cuishuang` for the original PR
debuginfo: Fix bug in type name generation for dyn types with associated types but no other generic arguments.
For types like `&dyn Future<Output=bool>` the compiler currently emits invalid types names in debuginfo. This PR fixes this.
Before:
```txt
// DWARF
&dyn core::future::future::Future, Output=bool>
// CodeView
ref$<dyn$<core::future::future::Future,assoc$<Output,bool> > > >
```
After:
```txt
// DWARF
&dyn core::future::future::Future<Output=bool>
// CodeView
ref$<dyn$<core::future::future::Future<assoc$<Output,bool> > > >
```
These syntactically incorrect type names can cause downstream tools (e.g. debugger extensions) crash when trying to parse them.
r? `@wesleywiser`
debuginfo: Refactor debuginfo generation for types
This PR implements the refactoring of the `rustc_codegen_llvm::debuginfo::metadata` module as described in MCP https://github.com/rust-lang/compiler-team/issues/482.
In particular it
- changes names to use `di_node` instead of `metadata`
- uniformly names all functions that build new debuginfo nodes `build_xyz_di_node`
- renames `CrateDebugContext` to `CodegenUnitDebugContext` (which is more accurate)
- removes outdated parts from `compiler/rustc_codegen_llvm/src/debuginfo/doc.md`
- moves `TypeMap` and functions that work directly work with it to a new `type_map` module
- moves enum related builder functions to a new `enums` module
- splits enum debuginfo building for the native and cpp-like cases, since they are mostly separate
- uses `SmallVec` instead of `Vec` in many places
- removes the old infrastructure for dealing with recursion cycles (`create_and_register_recursive_type_forward_declaration()`, `RecursiveTypeDescription`, `set_members_of_composite_type()`, `MemberDescription`, `MemberDescriptionFactory`, `prepare_xyz_metadata()`, etc)
- adds `type_map::build_type_with_children()` as a replacement for dealing with recursion cycles
- adds many (doc-)comments explaining what's going on
- changes cpp-like naming for C-Style enums so they don't get a `enum$<...>` name (because the NatVis visualizer does not apply to them)
- fixes detection of what is a C-style enum because some enums where classified as C-style even though they have fields
- changes cpp-like naming for generator enums so that NatVis works for them
- changes the position of discriminant debuginfo node so it is consistently nested inside the top-level union instead of, sometimes, next to it
The following could be done in subsequent PRs:
- add caching for `closure_saved_names_of_captured_variables`
- add caching for `generator_layout_and_saved_local_names`
- fix inconsistent handling of what is considered a C-style enum wrt to debuginfo
- rename `metadata` module to `types`
- move common generator fields to front instead of appending them
This PR is based on https://github.com/rust-lang/rust/pull/93644 which is not merged yet.
Right now, the changes are all done in one big commit. They could be split into smaller commits but hopefully the list of changes above makes it tractable to review them as a single commit too.
For now: r? `@ghost` (let's see if this affects compile times)
More robust fallback for `use` suggestion
Our old way to suggest where to add `use`s would first look for pre-existing `use`s in the relevant crate/module, and if there are *no* uses, it would fallback on trying to use another item as the basis for the suggestion.
But this was fragile, as illustrated in issue #87613
This PR instead identifies span of the first token after any inner attributes, and uses *that* as the fallback for the `use` suggestion.
Fix#87613
More robust fallback for `use` suggestion
Our old way to suggest where to add `use`s would first look for pre-existing `use`s in the relevant crate/module, and if there are *no* uses, it would fallback on trying to use another item as the basis for the suggestion.
But this was fragile, as illustrated in issue #87613
This PR instead identifies span of the first token after any inner attributes, and uses *that* as the fallback for the `use` suggestion.
Fix#87613
check ~Projection~ all supertrait bounds when confirming dyn candidate
I'm pretty sure Projection is the only other PredicateKind that we care about enforcing here.
Fixes#80800
