Matthias Krüger [Sun, 29 Jan 2023 05:14:18 +0000 (06:14 +0100)]
Rollup merge of #107397 - Teapot4195:issue-107392-fix, r=albertlarsan68
Gracefully exit if --keep-stage flag is used on a clean source tree
Instead of quitting with an obscure No such file or directory error, give the user a clearer and easier to understand error (as well as suggesting a possible cause for the error).
This is the first time I have written rust since 2020, and the first PR I will ever make to rust, so please do point out any mistakes I have made :smile: .
Matthias Krüger [Sun, 29 Jan 2023 05:14:17 +0000 (06:14 +0100)]
Rollup merge of #107097 - tmiasko:ssa, r=cjgillot
Fix def-use dominance check
A definition does not dominate a use in the same statement. For example
in MIR generated for compound assignment x += a (when overflow checks
are disabled).
Matthias Krüger [Sun, 29 Jan 2023 05:14:16 +0000 (06:14 +0100)]
Rollup merge of #106618 - jmillikin:os-net-rustdoc-wasm32, r=JohnTitor
Disable `linux_ext` in wasm32 and fortanix rustdoc builds.
The `std::os::unix` module is stubbed out when building docs for these target platforms. The introduction of Linux-specific extension traits caused `std::os::net` to depend on sub-modules of `std::os::unix`, which broke rustdoc for the `wasm32-unknown-unknown` target.
Adding an additional `#[cfg]` guard solves that rustdoc failure by not declaring `linux_ext` on targets with a stubbed `std::os::unix`.
bors [Sat, 28 Jan 2023 22:11:11 +0000 (22:11 +0000)]
Auto merge of #103659 - clubby789:improve-partialord-derive, r=nagisa
Special-case deriving `PartialOrd` for enums with dataless variants
I was able to get slightly better codegen by flipping the derived `PartialOrd` logic for two-variant enums. I also tried to document the implementation of the derive macro to make the special-case logic a little clearer.
```rs
#[derive(PartialEq, PartialOrd)]
pub enum A<T> {
A,
B(T)
}
```
```diff
impl<T: ::core::cmp::PartialOrd> ::core::cmp::PartialOrd for A<T> {
#[inline]
fn partial_cmp(
&self,
other: &A<T>,
) -> ::core::option::Option<::core::cmp::Ordering> {
let __self_tag = ::core::intrinsics::discriminant_value(self);
let __arg1_tag = ::core::intrinsics::discriminant_value(other);
- match ::core::cmp::PartialOrd::partial_cmp(&__self_tag, &__arg1_tag) {
- ::core::option::Option::Some(::core::cmp::Ordering::Equal) => {
- match (self, other) {
- (A::B(__self_0), A::B(__arg1_0)) => {
- ::core::cmp::PartialOrd::partial_cmp(__self_0, __arg1_0)
- }
- _ => ::core::option::Option::Some(::core::cmp::Ordering::Equal),
- }
+ match (self, other) {
+ (A::B(__self_0), A::B(__arg1_0)) => {
+ ::core::cmp::PartialOrd::partial_cmp(__self_0, __arg1_0)
}
- cmp => cmp,
+ _ => ::core::cmp::PartialOrd::partial_cmp(&__self_tag, &__arg1_tag),
}
}
}
```
Godbolt: [Current](https://godbolt.org/z/GYjEzG1T8), [New](https://godbolt.org/z/GoK78qx15)
I'm not sure how common a case comparing two enums like this (such as `Option`) is, and if it's worth the slowdown of adding a special case to the derive. If it causes overall regressions it might be worth just manually implementing this for `Option`.
This has been unstably const for over a year now. And the code change simplifying the constness of the `new` constructors has been in stable Rust since 1.64 (a bit over a full release cycle). I'm not aware of any blockers to this stabilization.
Matthias Krüger [Sat, 28 Jan 2023 10:11:05 +0000 (11:11 +0100)]
Rollup merge of #104012 - chenyukang:yukang/fix-103882-deli-indentation, r=petrochenkov
Improve unexpected close and mismatch delimiter hint in TokenTreesReader
Fixes #103882
Fixes #68987
Fixes #69259
The inner indentation mismatching will be covered by outer block, the new added function `report_error_prone_delim_block` will find out the error prone candidates for reporting.
bors [Sat, 28 Jan 2023 10:00:56 +0000 (10:00 +0000)]
Auto merge of #106916 - lukas-code:overlapping-substs, r=estebank
Remove overlapping parts of multipart suggestions
This PR adds a debug assertion that the parts of a single substitution cannot overlap, fixes a overlapping substitution from the testsuite, and fixes https://github.com/rust-lang/rust/issues/106870.
Note that a single suggestion can still have multiple overlapping substitutions / possible edits, we just don't suggest overlapping replacements in a single edit anymore.
I've also included a fix for an unrelated bug where rustfix for `explicit_outlives_requirements` would produce multiple trailing commas for a where clause.
bors [Sat, 28 Jan 2023 06:46:42 +0000 (06:46 +0000)]
Auto merge of #107400 - matthiaskrgr:rollup-l6bycds, r=matthiaskrgr
Rollup of 8 pull requests
Successful merges:
- #107022 (Implement `SpecOptionPartialEq` for `cmp::Ordering`)
- #107100 (Use proper `InferCtxt` when probing for associated types in astconv)
- #107103 (Use new solver in `evaluate_obligation` query (when new solver is enabled))
- #107190 (Recover from more const arguments that are not wrapped in curly braces)
- #107306 (Correct suggestions for closure arguments that need a borrow)
- #107339 (internally change regions to be covariant)
- #107344 (Minor tweaks in the new solver)
- #107373 (Don't merge vtables when full debuginfo is enabled.)
Matthias Krüger [Sat, 28 Jan 2023 04:20:19 +0000 (05:20 +0100)]
Rollup merge of #107373 - michaelwoerister:dont-merge-vtables-when-debuginfo, r=WaffleLapkin
Don't merge vtables when full debuginfo is enabled.
This PR makes the compiler not emit the `unnamed_addr` attribute for vtables when full debuginfo is enabled, so that they don't get merged even if they have the same contents. This allows debuggers to more reliably map from a dyn pointer to the self-type of a trait object by looking at the vtable's debuginfo.
The PR only changes the behavior of the LLVM backend as other backends don't emit vtable debuginfo (as far as I can tell).
The performance impact of this change should be small as [measured](https://github.com/rust-lang/rust/pull/103514#issuecomment-1290833854) in a previous PR.
Matthias Krüger [Sat, 28 Jan 2023 04:20:19 +0000 (05:20 +0100)]
Rollup merge of #107344 - compiler-errors:new-solver-tweaks, r=lcnr
Minor tweaks in the new solver
1. `InferCtxt::probe` is not needed in `compute_subtype_goal` and `compute_well_formed_goal`.
2. Add a handful of comments.
3. Add a micro-optimization in `consider_assumption` where we check the def-ids of the assumption and goal match before instantiating any binders.
Matthias Krüger [Sat, 28 Jan 2023 04:20:18 +0000 (05:20 +0100)]
Rollup merge of #107339 - aliemjay:covariant, r=lcnr
internally change regions to be covariant
Surprisingly, we consider the reference type `&'a T` to be contravaraint in its lifetime parameter. This is confusing and conflicts with the documentation we have in the reference, rustnomicon, and rustc-dev-guide. This also arguably not the correct use of terminology since we can use `&'static u8` in a place where `&' a u8` is expected, this implies that `&'static u8 <: &' a u8` and consequently `'static <: ' a`, hence covariance.
Because of this, when relating two types, we used to switch the argument positions in a confusing way:
`Subtype(&'a u8 <: &'b u8) => Subtype('b <: 'a) => Outlives('a: 'b) => RegionSubRegion('b <= 'a)`
The reason for the current behavior is probably that we wanted `Subtype('b <: 'a)` and `RegionSubRegion('b <= 'a)` to be equivalent, but I don' t think this is a good reason since these relations are sufficiently different in that the first is a relation in the subtyping lattice and is intrinsic to the type-systems, while the the second relation is an implementation detail of regionck.
This PR changes this behavior to use covariance, so..
`Subtype(&'a u8 <: &'b u8) => Subtype('a <: 'b) => Outlives('a: 'b) => RegionSubRegion('b <= 'a) `
Matthias Krüger [Sat, 28 Jan 2023 04:20:18 +0000 (05:20 +0100)]
Rollup merge of #107306 - compiler-errors:correct-sugg-for-closure-arg-needs-borrow, r=oli-obk
Correct suggestions for closure arguments that need a borrow
Fixes #107301 by dealing with binders correctly
Fixes another issue where we were suggesting adding just `&` when we expected `&mut _` in a closure arg
Matthias Krüger [Sat, 28 Jan 2023 04:20:17 +0000 (05:20 +0100)]
Rollup merge of #107190 - fmease:fix-81698, r=compiler-errors
Recover from more const arguments that are not wrapped in curly braces
Recover from some array, borrow, tuple & arithmetic expressions in const argument positions that lack curly braces and provide a suggestion to fix the issue continuing where #92884 left off. Examples of such expressions: `[]`, `[0]`, `[1, 2]`, `[0; 0xff]`, `&9`, `("", 0)` and `(1 + 2) * 3` (we previously did not recover from them).
I am not entirely happy with my current solution because the code that recovers from `[0]` (coinciding with a malformed slice type) and `[0; 0]` (coinciding with a malformed array type) is quite fragile as the aforementioned snippets are actually successfully parsed as types by `parse_ty` since it itself already recovers from them (returning `[⟨error⟩]` and `[⟨error⟩; 0]` respectively) meaning I have to manually look for `TyKind::Err`s and construct a separate diagnostic for the suggestion to attach to (thereby emitting two diagnostics in total).
(Which <https://alive2.llvm.org/ce/z/-rop5r> says LLVM *could* just do itself, but there's probably an issue already open for that problem from when this was originally looked at for `Option<NonZeroU8>` and friends.)
bors [Sat, 28 Jan 2023 01:05:29 +0000 (01:05 +0000)]
Auto merge of #101692 - cjgillot:generator-lazy-witness, r=oli-obk
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
Yuki Okushi [Fri, 27 Jan 2023 15:23:14 +0000 (00:23 +0900)]
Rollup merge of #107321 - lcnr:comment, r=compiler-errors
solver comments + remove `TyCtxt::evaluate_goal`
from the `RustcContributor::explore` session yesterday.
This also removes `TyCtxt::evaluate_goal` because to canonicalize you have to use an `InferCtxt` anyways at which point we should just always get people to use `evaluate_root_goal`.
bors [Fri, 27 Jan 2023 15:02:44 +0000 (15:02 +0000)]
Auto merge of #107055 - kylematsuda:eb-fn-sig, r=lcnr
Switch to `EarlyBinder` for `fn_sig` query
Part of the work to finish #105779 (also see https://github.com/rust-lang/types-team/issues/78).
Several queries `X` have a `bound_X` variant that wraps the output in [`EarlyBinder`](https://doc.rust-lang.org/nightly/nightly-rustc/rustc_middle/ty/subst/struct.EarlyBinder.html). This adds `EarlyBinder` to the return type of the `fn_sig` query and removes `bound_fn_sig`.
bors [Fri, 27 Jan 2023 09:01:05 +0000 (09:01 +0000)]
Auto merge of #107054 - petrochenkov:effvisdoc3, r=GuillaumeGomez
rustdoc: Collect "rustdoc-reachable" items during early doc link resolution
This pass only needs to know about visibilities, attributes and reexports, so it can be run early, similarly to `compute_effective_visibilities` in rustc.
Results of this pass can be used to prune the list of extern impls early thus improving performance of https://github.com/rust-lang/rust/pull/94857.