Fix cycle when debug-printing opaque types
Fixes#61577
When printing an opaque type in non-verbose mode, we use the
`tcx.predicates_of` query to retrieve the opaque type's bounds for
pretty-printing. However, the pervasiveness of logging within librustc
means that we may already be executing `tcx.predicates_of` for the
opaque type we're trying to print, leading to a cycle error.
This commit adds a new 'no queries' thread-local flag to the pretty
printer. This flag is enabled during the computation of `predicates_of`
for opaque types, and causes us to print the opaque type in 'verbose'
mode (which does not require computing any additinal queries). This
should only affect debug logging for highly nested log messages, not any
user-visible output.
Add a proc-macro to derive HashStable in librustc dependencies
A second proc-macro is added to derive HashStable for crates librustc depends on.
This proc-macro HashStable_Generic (to bikeshed) allows to decouple code and some librustc's boilerplate.
Not everything is migrated, because `Span` and `TokenKind` require to be placed inside librustc.
Types using them stay there too.
Split out of #66279
r? @Zoxc
Aggregation of drive-by cosmetic changes for trait-upcasting PR
Cherry-picked from #60900.
As requested by @Centril (and @nikomatsakis, I believe).
r? @Centril
Fixes#61577
When printing an opaque type in non-verbose mode, we use the
`tcx.predicates_of` query to retrieve the opaque type's bounds for
pretty-printing. However, the pervasiveness of logging within librustc
means that we may already be executing `tcx.predicates_of` for the
opaque type we're trying to print, leading to a cycle error.
This commit adds a new 'no queries' thread-local flag to the pretty
printer. This flag is enabled during the computation of `predicates_of`
for opaque types, and causes us to print the opaque type in 'verbose'
mode (which does not require computing any additinal queries). This
should only affect debug logging for highly nested log messages, not any
user-visible output.
Derive TypeFoldable using a proc-macro
A new proc macro is added in librustc_macros.
It is used to derive TypeFoldable inside librustc and librustc_traits.
For now, the macro uses the `'tcx` lifetime implicitly, and does not allow for a more robust selection of the adequate lifetime.
The Clone-based TypeFoldable implementations are not migrated.
Closes#65674
Deleted unused labels from compiler and fixed or allowed
unused labels in tests. This patch removes some gratuitous
unused labels and turns off the warning for unused labels
that are a necessary part of tests. This will permit
setting the `unused_labels` lint to `warn`.
Avoid hashing the key twice in `get_query()`.
For a single-threaded parallel compiler, this reduces instruction counts
across several benchmarks, by up to 2.8%.
The commit also adds documentation about `Sharded`'s use of `FxHasher`.
r? @Zoxc
Split ConstValue into two enums
Hello,
Issue #59210 appeared abandoned, so I gave it a go.
Some further cleanup and refactoring may be mandated.
I did not test beyond `x.py check`, since my home computer dies compiling librustc.
Fixes#59210
Move self-profile infrastructure to data structures
The single dependency on queries (QueryName) can be fairly easily
abstracted via a trait and this further decouples Session from librustc
(the primary goal).
This is intended as a precursor to moving Session out of librustc, but since that involves lots of smaller steps that move around code I'm splitting it up into separate PRs.
Move Session fields to CrateStore
`allocator_kind` and `injected_panic_runtime` are both query-like, this moves them out of Session and into CrateStore, avoiding the `Once` they previously had by clearing separating initialization and de-initialization.
The single dependency on queries (QueryName) can be fairly easily
abstracted via a trait and this further decouples Session from librustc
(the primary goal).
This was essentially a "query" previously (with no key, just always run
once when resolving the crate dependencies), and remains so, just now in
a way that isn't on Session. This removes the need for the `Once` as
well.
rustc_target: inline abi::FloatTy into abi::Primitive.
This effectively undoes a small part of @oli-obk's #50967, now that the rest of the compiler doesn't use the `FloatTy` definition from `rustc_target`, post-#65884.
Have tidy ensure that we document all `unsafe` blocks in libcore
cc @rust-lang/libs
I documented a few and added ignore flags on the other files. We can incrementally document the files, but won't regress any files this way.
syntax: ABI-oblivious grammar
This PR has the following effects:
1. `extern $lit` is now legal where `$lit:literal` and `$lit` is substituted for a string literal.
2. `extern "abi_that_does_not_exist"` is now *syntactically* legal whereas before, the set of ABI strings was hard-coded into the grammar of the language. With this PR, the set of ABIs are instead validated and translated during lowering. That seems more appropriate.
3. `ast::FloatTy` is now distinct from `rustc_target::abi::FloatTy`. The former is used substantially more and the translation between them is only necessary in a single place.
4. As a result of 2-3, libsyntax no longer depends on librustc_target, which should improve pipe-lining somewhat.
cc @rust-lang/lang -- the points 1-2 slightly change the definition of the language but in a way which seems consistent with our general principles (in particular wrt. the discussions of turning things into semantic errors). I expect this to be uncontroversial but it's worth letting y'all know. :)
r? @varkor
We also sever syntax's dependency on rustc_target as a result.
This should slightly improve pipe-lining.
Moreover, some cleanup is done in related code.
Rename `LocalInternedString` and more
This PR renames `LocalInternedString` as `SymbolStr`, removes an unnecessary `impl` from it, improves comments, and cleans up some `SymbolStr` uses.
r? @estebank
Move has_panic_handler to query
Moves us off of a global Once instead re-querying the lang item each time. The conditions on when we set it to true change a little (previously we'd make sure a few more lang items were `Some`) but I think they in practice don't matter, we won't compile later on if we don't have them.
For a single-threaded parallel compiler, this reduces instruction counts
across several benchmarks, by up to 2.8%.
The commit also adds documentation about `Sharded`'s use of `FxHasher`.
Type parameters are referenced in the error message after the previous
few commits (using span label). But when the main error message already
references the very same type parameter it becomes clumsy. Do not show
the additional label in this case as per code review comment by
@estebank.
Also this contains a small style fix.
Fixes#47319.
Shows the type parameter definition(s) on type mismatch errors so the
context is clearer. Pretty much changes the following:
```
LL | bar1(t);
| ^
| |
| expected enum `std::option::Option`, found type parameter `T`
```
into:
```
LL | fn foo1<T>(t: T) {
| - this type parameter
LL | bar1(t);
| ^
| |
| expected enum `std::option::Option`, found type parameter `T`
```
Part of #47319.
This just adds type parameter name to type mismatch error message, so
e.g. the following:
```
expected enum `std::option::Option`, found type parameter
```
becomes
```
expected enum `std::option::Option`, found type parameter `T`
```
Revert PR 64324: dylibs export generics again (for now)
As discussed on PR #65781, this is a targeted attempt to undo the main semantic change from PR #64324, by putting `dylib` back in the set of crate types that export generic symbols.
The main reason to do this is that PR #64324 had unanticipated side-effects that caused bugs like #64872, and in the opinion of @alexcrichton and myself, the impact of #64872 is worse than #64319.
In other words, it is better for us, in the short term, to reopen#64319 as currently unfixed for now than to introduce new bugs like #64872.
Fix#64872Reopen#64319
This changes the mechanism of `-Z dual-proc-macro` to record the host
proc macro hash in the transistive dependency information and use it
during dependency resolution instead of resolving only by name.
Add new EFIAPI ABI
Fixes#54527
Adds a new ABI, "efiapi", which reflects the calling convention as specified by [the current spec UEFI spec](https://uefi.org/sites/default/files/resources/UEFI%20Spec%202_7_A%20Sept%206.pdf#G6.999903). When compiling for x86_64, we should select the `win64` ABI, while on all other architectures (Itanium, x86, ARM and ARM64 and RISC-V), we should select the `C` ABI.
Currently, this is done by just turning it into the C ABI everywhere except on x86_64, where it's turned into the win64 ABI. Should we prevent this ABI from being used on unsupported architectures, and if so, how would this be done?
self-profiling: Record something more useful for crate metadata generation event.
Before this commit, we had an event that would only track the compression step
for proc-macros and Rust dylibs. After the commit we measure the time for
acutally generating the crate metadata bytes.
r? @wesleywiser
Before this commit, we had an event that would only track the compression step
for proc-macros and Rust dylibs. After the commit we measure the time for
acutally generating the crate metadata bytes.
Stabilize `const_constructor`
# Stabilization proposal
I propose that we stabilize `#![feature(const_constructor)]`.
Tracking issue: https://github.com/rust-lang/rust/issues/61456
Version target: 1.40 (2019-11-05 => beta, 2019-12-19 => stable).
## What is stabilized
### User guide
Tuple struct and tuple variant constructors are now considered to be constant functions. As such a call expression where the callee has a tuple struct or variant constructor "function item" type can be called:
```rust
const fn make_options() {
// These already work because they are special cased:
Some(0);
(Option::Some)(1);
// These also work now:
let f = Option::Some;
f(2);
{Option::Some}(3);
<Option<_>>::Some(5);
}
```
### Motivation
Consistency with other `const fn`. Consistency between syntactic path forms.
This should also ensure that constructors implement `const Fn` traits and can be coerced to `const fn` function pointers, if they are introduced.
## Tests
* [ui/consts/const_constructor/const-construct-call.rs](0d75ab2293/src/test/ui/consts/const_constructor/const-construct-call.rs) - Tests various syntactic forms, use in both `const fn` and `const` items, and constructors in both the current and extern crates.
* [ui/consts/const_constructor/const_constructor_qpath.rs](1850dfcdab/src/test/ui/consts/const_constructor/const_constructor_qpath.rs) - Tests that type qualified paths to enum variants are also considered to be `const fn`.(#64247)
r? @oli-obk
Closes#61456
Closes #64247
rustc: add `Span`s to `inferred_outlives_of` predicates.
This would simplify #59789, and I suspect it has some potential in diagnostics (although we don't seem to use the predicate `Span`s much atm).
trait-based structural match implementation
Moves from using a `#[structural_match]` attribute to using a marker trait (or pair of such traits, really) instead.
Fix#63438.
(This however does not remove the hacks that I believe were put into place to support the previous approach of injecting the attribute based on the presence of both derives... I have left that for follow-on work.)
self-profiling: Update measureme to 0.4.0 and remove non-RAII methods from profiler.
This PR removes all non-RAII based profiling methods from `SelfProfilerRef` 🎉
It also delegates the `TimingGuard` implementation to `measureme`, now that that is available there.
r? @wesleywiser
Adds a new ABI for the EFIAPI calls. This ABI should reflect the latest
version of the UEFI specification at the time of commit (UEFI spec 2.8,
URL below). The specification says that for x86_64, we should follow the
win64 ABI, while on all other supported platforms (ia32, itanium, arm,
arm64 and risc-v), we should follow the C ABI.
To simplify the implementation, we will simply follow the C ABI on all
platforms except x86_64, even those technically unsupported by the UEFI
specification.
https://uefi.org/sites/default/files/resources/UEFI_Spec_2_8_final.pdf
(Or more precisely, a pair of such traits: one for `derive(PartialEq)` and one
for `derive(Eq)`.)
((The addition of the second marker trait, `StructuralEq`, is largely a hack to
work-around `fn (&T)` not implementing `PartialEq` and `Eq`; see also issue
rust-lang/rust#46989; otherwise I would just check if `Eq` is implemented.))
Note: this does not use trait fulfillment error-reporting machinery; it just
uses the trait system to determine if the ADT was tagged or not. (Nonetheless, I
have kept an `on_unimplemented` message on the new trait for structural_match
check, even though it is currently not used.)
Note also: this does *not* resolve the ICE from rust-lang/rust#65466, as noted
in a comment added in this commit. Further work is necessary to resolve that and
other problems with the structural match checking, especially to do so without
breaking stable code (adapted from test fn-ptr-is-structurally-matchable.rs):
```rust
fn r_sm_to(_: &SM) {}
fn main() {
const CFN6: Wrap<fn(&SM)> = Wrap(r_sm_to);
let input: Wrap<fn(&SM)> = Wrap(r_sm_to);
match Wrap(input) {
Wrap(CFN6) => {}
Wrap(_) => {}
};
}
```
where we would hit a problem with the strategy of unconditionally checking for
`PartialEq` because the type `for <'a> fn(&'a SM)` does not currently even
*implement* `PartialEq`.
----
added review feedback:
* use an or-pattern
* eschew `return` when tail position will do.
* don't need fresh_expansion; just add `structural_match` to appropriate `allow_internal_unstable` attributes.
also fixed example in doc comment so that it actually compiles.
Intern place projection
This should sit on top of https://github.com/rust-lang/rust/pull/65197. After that one merged, I'm gonna rebase on top of it.
The important commits are the last three and there's a bunch of code repetition that I'm going to remove but for that I need to refactor some things that probably need to be added before this PR.
Anyway this work helps as is because we can run perf tests :).
r? @oli-obk /cc @nikomatsakis
Remove `InternedString`
This PR removes `InternedString` by converting all occurrences to `Symbol`. There are a handful of places that need to use the symbol chars instead of the symbol index, e.g. for stable sorting; local conversions `LocalInternedString` is used in those places.
r? @eddyb
Lockless LintStore
This removes mutability from the lint store after registration. Each commit stands alone, for the most part, though they don't make sense out of sequence.
The intent here is to move LintStore to a more parallel-friendly architecture, although also just a cleaner one from an implementation perspective. Specifically, this has the following changes:
* We no longer implicitly register lints when registering lint passes
* For the most part this means that registration calls now likely want to call something like:
`lint_store.register_lints(&Pass::get_lints())` as well as `register_*_pass`.
* In theory this is a simplification as it's much easier for folks to just register lints and then have passes that implement whichever lint however they want, rather than necessarily tying passes to lints.
* Lint passes still have a list of associated lints, but a followup PR could plausibly change that
* This list must be known for a given pass type, not instance, i.e., `fn get_lints()` is the signature instead of `fn get_lints(&self)` as before.
* We do not store pass objects, instead storing constructor functions. This means we always get new passes when running lints (this happens approximately once though for a given compiler session, so no behavior change is expected).
* Registration API is _much_ simpler: generally all functions are just taking `Fn() -> PassObject` rather than several different `bool`s.
Rename `ConstValue::Infer(InferConst::Canonical(..))` to `ConstValue::Bound(..)`
It already has the right form, so this is just a renaming. Fixes https://github.com/rust-lang/rust/issues/65655.
r? @eddyb
Change untagged_unions to not allow union fields with drop
This is a rebase of #56440, massaged to solve merge conflicts and make the test suite pass.
Change untagged_unions to not allow union fields with drop
Union fields may now never have a type with attached destructor. This for example allows unions to use arbitrary field types only by wrapping them in `ManuallyDrop` (or similar).
The stable rule remains, that union fields must be `Copy`. We use the new rule for the `untagged_union` feature.
Tracking issue: https://github.com/rust-lang/rust/issues/55149
It's a full conversion, except in `DefKey::compute_stable_hash()` where
a `Symbol` now is converted to an `InternedString` before being hashed.
This was necessary to avoid test failures.
rustc: arena-allocate the slice in `ty::GenericsPredicate`, not the whole struct.
While rebasing #59789 I noticed we can do this now. However, it doesn't help much without changing `inferred_outlives_of` to the same type, which I might try next.
Rollup of 8 pull requests
Successful merges:
- #65237 (Move debug_map assertions after check for err)
- #65316 (make File::try_clone produce non-inheritable handles on Windows)
- #65319 (InterpCx: make memory field public)
- #65461 (Don't recommend ONCE_INIT in std::sync::Once)
- #65465 (Move syntax::ext to a syntax_expand and refactor some attribute logic)
- #65475 (add example for type_name)
- #65478 (fmt::Write is about string slices, not byte slices)
- #65486 (doc: fix typo in OsStrExt and OsStringExt)
Failed merges:
r? @ghost
Optimize dropck
This does two things: caches the `trivial_dropck` check by making it a query, and shifts around the implementation of the primary dropck itself to avoid allocating many small vectors.
Return `false` from `needs_drop` for all zero-sized arrays.
Resolves#65348.
This changes the result of the `needs_drop` query from `true` to `false` for types such as `[Box<i32>; 0]`. I believe this change to be sound because a zero-sized array can never actually hold a value. This is an elegant way of resolving #65348 and #64945, but obviously it has much broader implications.
Optimize `try_expand_impl_trait_type`
A lot of time was being spent expanding some large `impl Future` types in fuchsia. This PR takes the number of types being visited in one expansion from >3 billion to about a thousand, and eliminates the compile time regression in https://github.com/rust-lang/rust/issues/65147 (in fact, compile times are better than they were before).
Thanks to @Mark-Simulacrum for helping identify the issue and to @matthewjasper for suggesting this change.
Fixes#65147.
r? @matthewjasper,@nikomatsakis
Remove last uses of gensyms
Underscore bindings now use unique `SyntaxContext`s to avoid collisions. This was the last use of gensyms in the compiler, so this PR also removes them.
closes#49300
cc #60869
r? @petrochenkov
Deduplicate is_{freeze,copy,sized}_raw
Fixes#65259
Deduplicates `is_{freeze,copy,sized}_raw` by delegating to a new method which takes in a `LangItem`.
Union fields may now never have a type with attached destructor.
This for example allows unions to use arbitrary field types only by
wrapping
them in ManuallyDrop.
The stable rule remains, that union fields must be Copy. We use the new
rule for the `untagged_union` feature.
See RFC 2514.
Note for ui tests:
We can't test move out through Box's deref-move since we can't
have a Box in a union anymore.
This ICEs in MIR currently, which I think is to be expected since none of the MIR plumbing is set up. I added a test which confirms that the shim is being used for reifying a track_caller function.
Remove query-related macros
The query system has a few macros that only have one or two call sites, and I find they hurt readability. This PR removes them.
r? @michaelwoerister
Remove loaded_from_cache map from DepGraph
It's now unused, even with -Zquery-dep-graph
From https://github.com/rust-lang/rust/pull/63756/files#r316039379 -- it'll simplify that PR to get this landed separately so we can just remove some of the code that it touches.
r? @Zoxc or @michaelwoerister
fix bug in folding for constants
These was a bug in the folding for constants that caused it to overlook bound regions. This branch includes some other little things that I did while trying to track the bug down.
r? @oli-obk
