`f16` and `f128` step 3: compiler support & feature gate
Continuation of https://github.com/rust-lang/rust/pull/121841, another portion of https://github.com/rust-lang/rust/pull/114607
This PR exposes the new types to the world and adds a feature gate. Marking this as a draft because I need some feedback on where I did the feature gate check. It also does not yet catch type via suffixed literals (so the feature gate test will fail, probably some others too because I haven't belssed).
If there is a better place to check all types after resolution, I can do that. If not, I figure maybe I can add a second gate location in AST when it checks numeric suffixes.
Unfortunately I still don't think there is much testing to be done for correctness (codegen tests or parsed value checks) until we have basic library support. I think that will be the next step.
Tracking issue: https://github.com/rust-lang/rust/issues/116909
r? `@compiler-errors`
cc `@Nilstrieb`
`@rustbot` label +F-f16_and_f128
Detect calls to .clone() on T: !Clone types on borrowck errors
When encountering a lifetime error on a type that *holds* a type that doesn't implement `Clone`, explore the item's body for potential calls to `.clone()` that are only cloning the reference `&T` instead of `T` because `T: !Clone`. If we find this, suggest `T: Clone`.
```
error[E0502]: cannot borrow `*list` as mutable because it is also borrowed as immutable
--> $DIR/clone-on-ref.rs:7:5
|
LL | for v in list.iter() {
| ---- immutable borrow occurs here
LL | cloned_items.push(v.clone())
| ------- this call doesn't do anything, the result is still `&T` because `T` doesn't implement `Clone`
LL | }
LL | list.push(T::default());
| ^^^^^^^^^^^^^^^^^^^^^^^ mutable borrow occurs here
LL |
LL | drop(cloned_items);
| ------------ immutable borrow later used here
|
help: consider further restricting this bound
|
LL | fn foo<T: Default + Clone>(list: &mut Vec<T>) {
| +++++++
```
```
error[E0505]: cannot move out of `x` because it is borrowed
--> $DIR/clone-on-ref.rs:23:10
|
LL | fn qux(x: A) {
| - binding `x` declared here
LL | let a = &x;
| -- borrow of `x` occurs here
LL | let b = a.clone();
| ------- this call doesn't do anything, the result is still `&A` because `A` doesn't implement `Clone`
LL | drop(x);
| ^ move out of `x` occurs here
LL |
LL | println!("{b:?}");
| ----- borrow later used here
|
help: consider annotating `A` with `#[derive(Clone)]`
|
LL + #[derive(Clone)]
LL | struct A;
|
```
Fix#48677.
coverage: Initial support for branch coverage instrumentation
(This is a review-ready version of the changes that were drafted in #118305.)
This PR adds support for branch coverage instrumentation, gated behind the unstable flag value `-Zcoverage-options=branch`. (Coverage instrumentation must also be enabled with `-Cinstrument-coverage`.)
During THIR-to-MIR lowering (MIR building), if branch coverage is enabled, we collect additional information about branch conditions and their corresponding then/else blocks. We inject special marker statements into those blocks, so that the `InstrumentCoverage` MIR pass can reliably identify them even after the initially-built MIR has been simplified and renumbered.
The rest of the changes are mostly just plumbing needed to gather up the information that was collected during MIR building, and include it in the coverage metadata that we embed in the final binary.
Note that `llvm-cov show` doesn't print branch coverage information in its source views by default; that needs to be explicitly enabled with `--show-branches=count` or similar.
---
The current implementation doesn't have any support for instrumenting `if let` or let-chains. I think it's still useful without that, and adding it would be non-trivial, so I'm happy to leave that for future work.
match lowering: don't collect test alternatives ahead of time
I'm very happy with this one. Before this, when sorting candidates into the possible test branches, we manually computed `usize` indices to determine in which branch each candidate goes. To make this work we had a first pass that collected the possible alternatives we'd have to deal with, and a second pass that actually sorts the candidates.
In this PR, I replace `usize` indices with a dedicated enum. This makes `sort_candidates` easier to follow, and we don't need the first pass anymore.
r? ``@matthewjasper``
pattern analysis: Store field indices in `DeconstructedPat` to avoid virtual wildcards
For a pattern like `Struct { field3: true, .. }`, in pattern analysis we represent it as `Struct { field1: _, field2: _, field3: true, field4: _ }`. This PR makes it so we store `Struct { field3: true, .. }` instead. This means we never have to create fake `_` patterns during lowering.
r? ``@compiler-errors``
Clarity improvements to `DropTree`
These changes are based on some points of confusion I had when initially trying to understand this code.
The only “functional” change is an additional assertion in `<ExitScopes as DropTreeBuilder>::link_entry_point`, checking that the dummy terminator is `TerminatorKind::UnwindResume` as expected.
match lowering: define a convenient struct
Small refactor PR: `bindings` and `ascriptions` always come together so I made a struct for them. I'll have one or two fields to add to it in a later PR as well.
Add asm goto support to `asm!`
Tracking issue: #119364
This PR implements asm-goto support, using the syntax described in "future possibilities" section of [RFC2873](https://rust-lang.github.io/rfcs/2873-inline-asm.html#asm-goto).
Currently I have only implemented the `label` part, not the `fallthrough` part (i.e. fallthrough is implicit). This doesn't reduce the expressive though, since you can use label-break to get arbitrary control flow or simply set a value and rely on jump threading optimisation to get the desired control flow. I can add that later if deemed necessary.
r? ``@Amanieu``
cc ``@ojeda``
This method would previously take a target scope, and then verify that it
was equal to the scope on top of the if-then scope stack.
In practice, this means that callers have to go out of their way to pass around
redundant scope information that's already on the if-then stack.
So it's easier to just retrieve the correct scope directly from the if-then
stack, and simplify the other code that was passing it around.
This allows us to use real field names instead of tuple element numbers.
Renaming `previous_drops` to `existing_drops_map` clarifies that "previous" was
unrelated to drop order.
This makes it easier to see that the call to `in_scope` returns both the then
block and the else block. The rather confusing `unpack!` step is confined to
its own separate line.
(This patch reindents several lines, so using "ignore whitespace" is
recommended in order to focus on the actual changes.)
Extract an arguments struct for `Builder::then_else_break`
Most of this method's arguments are usually or always forwarded as-is to recursive invocations.
Wrapping them in a dedicated struct allows us to document each struct field, and lets us use struct-update syntax to indicate which arguments are being modified when making a recursive call.
---
While trying to understand the lowering of `if` expressions, I found it difficult to keep track of the half-dozen arguments passed through to every call to `then_else_break`. I tried switching over to an arguments struct, and I found that it really helps to make sense of what each argument does, and how each call is modifying the arguments.
I have some further ideas for how to streamline these recursive calls, but I've kept those out of this PR so that it's a pure refactoring with no behavioural changes.
Most of this method's arguments are usually or always forwarded as-is to
recursive invocations.
Wrapping them in a dedicated struct allows us to document each struct field,
and lets us use struct-update syntax to indicate which arguments are being
modified when making a recursive call.
Add new `pattern_complexity` attribute to add possibility to limit and check recursion in pattern matching
Needed for https://github.com/rust-lang/rust-analyzer/issues/9528.
This PR adds a new attribute only available when running rust testsuite called `pattern_complexity` which allows to set the maximum recursion for the pattern matching. It is quite useful to ensure the complexity doesn't grow, like in `tests/ui/pattern/usefulness/issue-118437-exponential-time-on-diagonal-match.rs`.
r? `@Nadrieril`
Rollup of 5 pull requests
Successful merges:
- #120761 (Add initial support for DataFlowSanitizer)
- #121622 (Preserve same vtable pointer when cloning raw waker, to fix Waker::will_wake)
- #121716 (match lowering: Lower bindings in a predictable order)
- #121731 (Now that inlining, mir validation and const eval all use reveal-all, we won't be constraining hidden types here anymore)
- #121841 (`f16` and `f128` step 2: intrinsics)
r? `@ghost`
`@rustbot` modify labels: rollup
match lowering: Lower bindings in a predictable order
After the recent refactorings, we can now lower bindings in a truly predictable order. The order in https://github.com/rust-lang/rust/pull/120214 was an improvement but not very clear. With this PR, we lower bindings from left to right, with the special case that `x @ pat` is traversed as `pat @ x` (i.e. `x` is lowered after any bindings in `pat`).
This description only applies in the absence of or-patterns. Or-patterns make everything complicated, because the binding place depends on the subpattern. Until I have a better idea I leave them to be handled in whatever weird order arises from today's code.
r? `@matthewjasper`
Before, the SwitchInt cases were computed in two passes: if the first
pass accepted e.g. 0..=5 and then 1, the second pass would not accept
0..=5 anymore because 1 would be listed in the SwitchInt options.
Now there's a single pass, so if we sort 0..=5 we must take care to not
sort a subsequent 1.
The ordinary lowering of `thir::ExprKind::Let` is unreachable
After desugaring, `let` expressions should only appear inside `if` expressions or `match` guards, possibly nested within a let-chain. In both cases they are specifically handled by the lowerings of those expressions, so this case is currently unreachable.
---
Context: https://rust-lang.zulipchat.com/#narrow/stream/182449-t-compiler.2Fhelp/topic/Lowering.20of.20.60thir.3A.3AExprKind.3A.3ALet.60.20is.20unreachable
My conclusions are partly based on the observation that stubbing out this match arm doesn't cause any test failures. So either this really is unreachable, or it can be reached in some obscure circumstances that our test suite doesn't cover.
If we end up needing this code (or something like it) for an implementation of https://github.com/rust-lang/rfcs/pull/3573, it should be easy enough to pull it back out of version control history.
I looked into having the `if`/`match` lowerings call back into `expr_into_dest`, but from what I can tell that won't work well, because there are extra scoping considerations that require some awareness of the enclosing if/match.
r? ```@Nadrieril```
After desugaring, `let` expressions should only appear inside `if` expressions
or `match` guards, possibly nested within a let-chain. In both cases they are
specifically handled by the lowerings of those expressions, so this case is
currently unreachable.
match lowering: pre-simplify or-patterns too
This is the final part of my work to simplify match pairs early: now we do it for or-patterns too. This makes it possible to collect fake borrows separately from the main match lowering algorithm. That'll enable more simplifications of or-pattern handling.
Note: I was tempted to have `Candidate` contain a `FlatPat`, but there are so many places that use `candidate.match_pairs` etc directly that I chose not to.
r? `@matthewjasper`
match lowering: Separate the `bool` case from other integers in `TestKind`
`TestKind::SwitchInt` had a special case for `bool` essentially everywhere it's used, so I made `TestKind::If` to handle the bool case on its own.
r? `@matthewjasper`
match lowering: Remove hacky branch in sort_candidate
Reusing `self.test()` there wasn't actually pulling a lot of weight. In particular the `TestKind::Len` cases were all already correctly handled.
r? `@matthewjasper`
Make the success arms of `if lhs || rhs` meet up in a separate block
Extracted from #118305, where this is necessary to avoid introducing a bug when injecting marker statements into the then/else arms.
---
In the previous code (#111752), the success block of `lhs` would jump directly to the success block of `rhs`. However, `rhs_success_block` could already contain statements that are specific to the RHS, and the direct goto causes them to be executed in the LHS success path as well.
This patch therefore creates a fresh block that the LHS and RHS success blocks can both jump to.
---
I think the reason we currently get away with this is that `rhs_success_block` usually doesn't contain anything other than StorageDead statements for locals used by the RHS, and those statements don't seem to cause problems in the LHS success path (which never makes those locals live).
But if we start adding meaningful statements for branch coverage (or MC/DC coverage), it's important to keep the LHS and RHS blocks separate.
Add stubs in IR and ABI for `f16` and `f128`
This is the very first step toward the changes in https://github.com/rust-lang/rust/pull/114607 and the [`f16` and `f128` RFC](https://rust-lang.github.io/rfcs/3453-f16-and-f128.html). It adds the types to `rustc_type_ir::FloatTy` and `rustc_abi::Primitive`, and just propagates those out as `unimplemented!` stubs where necessary.
These types do not parse yet so there is no feature gate, and it should be okay to use `unimplemented!`.
The next steps will probably be AST support with parsing and the feature gate.
r? `@compiler-errors`
cc `@Nilstrieb` suggested breaking the PR up in https://github.com/rust-lang/rust/pull/120645#issuecomment-1925900572
Skip unnecessary comparison with half-open range patterns
This is the last remaining detail in the implementation of half-open range patterns. Until now, a half-open range pattern like `10..` was converted to `10..T::MAX` before lowering to MIR, which generated an extra pointless comparison. With this PR we don't generate it.
In the previous code, the success block of `lhs` would jump directly to the
success block of `rhs`. However, `rhs_success_block` could already contain
statements that are specific to the RHS, and the direct goto causes them to be
executed in the LHS success path as well.
This patch therefore creates a fresh block that the LHS and RHS success blocks
can both jump to.
match lowering: Split off `test_candidates` into several functions and improve comments
The logic of `test_candidates` has three steps: pick a test, sort the candidates, and generate code for everything. So I split it off into three methods.
I also ended up reworking the comments that explain the algorithm. In particular I added detailed examples. I removed the digression about https://github.com/rust-lang/rust/issues/29740 because it's no longer relevant to how the code is structured today.
r? ``@matthewjasper``
PR #119097 made the decision to make all `IntoDiagnostic` impls generic,
because this allowed a bunch of nice cleanups. But four hand-written
impls were unintentionally overlooked. This commit makes them generic.
match lowering: test one or pattern at a time
This is a bit more opinionated than the previous PRs. On the face of it this is less efficient and more complex than before, but I personally found the loop that digs into `leaf_candidates` on each iteration very confusing. Instead this does "generate code for this or-pattern" then "generate further code for each branch if needed" in two steps.
Incidentally this way we don't _require_ or patterns to be sorted at the end. It's still an important optimization but I find it clearer to not rely on it for correctness.
r? `@matthewjasper`
Convert `delayed_bug`s to `bug`s.
I have a suspicion that quite a few delayed bug paths are impossible to reach, so I did an experiment.
I converted every `delayed_bug` to a `bug`, ran the full test suite, then converted back every `bug` that was hit. A surprising number were never hit.
This is too dangerous to merge. Increased coverage (fuzzing or a crater run) would likely hit more cases. But it might be useful for people to look at and think about which paths are genuinely unreachable.
r? `@ghost`
match lowering: simplify empty candidate selection
In match lowering, `match_simplified_candidates` is tasked with removing candidates that are fully matched and linking them up properly. The code that does that was needlessly complicated; this PR simplifies it.
The overall change isn't big but I split it up into tiny commits to convince myself that I was correctly preserving behavior. The test changes are all due to the first commit. Let me know if you'd prefer me to split up the PR to make reviewing easier.
r? `@matthewjasper`
match lowering: eagerly simplify match pairs
This removes one important complication from match lowering. Before this, match pair simplification (which includes collecting bindings and type ascriptions) was intertwined with the whole match lowering algorithm.
I'm avoiding this by storing in each `MatchPair` the sub-`MatchPair`s that correspond to its subfields. This makes it possible to simplify everything (except or-patterns) in `Candidate::new()`.
This should open up further simplifications. It will also give us proper control over the order of bindings.
r? `@matthewjasper`
I have a suspicion that quite a few delayed bug paths are impossible to
reach, so I did an experiment.
I converted every `delayed_bug` to a `bug`, ran the full test suite,
then converted back every `bug` that was hit. A surprising number were
never hit.
The next commit will convert some more back, based on human judgment.
Currently many diagnostic modifier methods are available on both
`Diagnostic` and `DiagnosticBuilder`. This commit removes most of them
from `Diagnostic`. To minimize the diff size, it keeps them within
`diagnostic.rs` but changes the surrounding `impl Diagnostic` block to
`impl DiagnosticBuilder`. (I intend to move things around later, to give
a more sensible code layout.)
`Diagnostic` keeps a few methods that it still needs, like `sub`,
`arg`, and `replace_args`.
The `forward!` macro, which defined two additional methods per call
(e.g. `note` and `with_note`), is replaced by the `with_fn!` macro,
which defines one additional method per call (e.g. `with_note`). It's
now also only used when necessary -- not all modifier methods currently
need a `with_*` form. (New ones can be easily added as necessary.)
All this also requires changing `trait AddToDiagnostic` so its methods
take `DiagnosticBuilder` instead of `Diagnostic`, which leads to many
mechanical changes. `SubdiagnosticMessageOp` gains a type parameter `G`.
There are three subdiagnostics -- `DelayedAtWithoutNewline`,
`DelayedAtWithNewline`, and `InvalidFlushedDelayedDiagnosticLevel` --
that are created within the diagnostics machinery and appended to
external diagnostics. These are handled at the `Diagnostic` level, which
means it's now hard to construct them via `derive(Diagnostic)`, so
instead we construct them by hand. This has no effect on what they look
like when printed.
There are lots of new `allow` markers for `untranslatable_diagnostics`
and `diagnostics_outside_of_impl`. This is because
`#[rustc_lint_diagnostics]` annotations were present on the `Diagnostic`
modifier methods, but missing from the `DiagnosticBuilder` modifier
methods. They're now present.
The moment we get a candidate without guard, the return block becomes a
fresh block linked to nothing. So we can keep assigning a fresh block
every iteration to reuse the `next_prebinding` logic.
errors: only eagerly translate subdiagnostics
Subdiagnostics don't need to be lazily translated, they can always be eagerly translated. Eager translation is slightly more complex as we need to have a `DiagCtxt` available to perform the translation, which involves slightly more threading of that context.
This slight increase in complexity should enable later simplifications - like passing `DiagCtxt` into `AddToDiagnostic` and moving Fluent messages into the diagnostic structs rather than having them in separate files (working on that was what led to this change).
r? ```@nnethercote```
Subdiagnostics don't need to be lazily translated, they can always be
eagerly translated. Eager translation is slightly more complex as we need
to have a `DiagCtxt` available to perform the translation, which involves
slightly more threading of that context.
This slight increase in complexity should enable later simplifications -
like passing `DiagCtxt` into `AddToDiagnostic` and moving Fluent messages
into the diagnostic structs rather than having them in separate files
(working on that was what led to this change).
Signed-off-by: David Wood <david@davidtw.co>
This mostly works well, and eliminates a couple of delayed bugs.
One annoying thing is that we should really also add an
`ErrorGuaranteed` to `proc_macro::bridge::LitKind::Err`. But that's
difficult because `proc_macro` doesn't have access to `ErrorGuaranteed`,
so we have to fake it.
Continue compilation after check_mod_type_wf errors
The ICEs fixed here were probably reachable through const eval gymnastics before, but now they are easily reachable without that, too.
The new errors are often bugfixes, where useful errors were missing, because they were reported after the early abort. In other cases sometimes they are just duplication of already emitted errors, which won't be user-visible due to deduplication.
fixes https://github.com/rust-lang/rust/issues/120860
For some cases where it's clear that an error has already occurred,
e.g.:
- there's a comment stating exactly that, or
- things like HIR lowering, where we are lowering an error kind
The commit also tweaks some comments around delayed bug sites.
match lowering: simplify block creation
Match lowering was doing complicated things with block creation. As far as I can tell it was trying to avoid creating unneeded blocks, but of the three places that start out with `otherwise = &mut None`, two of them called `otherwise.unwrap_or_else(|| self.cfg.start_new_block())` anyway. As far as I can tell the only place where this PR makes a difference is in `lower_match_tree`, which did indeed sometimes avoid creating the unreachable final block + FakeRead. Unless this is important I propose we do the naive thing instead.
I have not checked all the graph isomorphisms by hand, but at a glance the test diff looks sensible.
r? `@matthewjasper`
Dejargonize `subst`
In favor of #110793, replace almost every occurence of `subst` and `substitution` from rustc codes, but they still remains in subtrees under `src/tools/` like clippy and test codes (I'd like to replace them after this)
These crates all needed specialization for `newtype_index!`, which will no
longer be necessary when the current nightly eventually becomes the next
bootstrap compiler.
Fix more `ty::Error` ICEs in MIR passes
Fixes#120791 - Add a check for `ty::Error` in the `ByMove` coroutine pass
Fixes#120816 - Add a check for `ty::Error` in the MIR validator
Also a drive-by fix for a FIXME I had asked oli to add
r? oli-obk
Invert diagnostic lints.
That is, change `diagnostic_outside_of_impl` and `untranslatable_diagnostic` from `allow` to `deny`, because more than half of the compiler has been converted to use translated diagnostics.
This commit removes more `deny` attributes than it adds `allow` attributes, which proves that this change is warranted.
r? ````@davidtwco````
Make `min_exhaustive_patterns` match `exhaustive_patterns` better
Split off from https://github.com/rust-lang/rust/pull/120742.
There remained two edge cases where `min_exhaustive_patterns` wasn't behaving like `exhaustive_patterns`. This fixes them, and tests the feature in a bunch more cases. I essentially went through all uses of `exhaustive_patterns` to see which ones would be interesting to compare between the two features.
r? `@compiler-errors`
Rollup of 9 pull requests
Successful merges:
- #119592 (resolve: Unload speculatively resolved crates before freezing cstore)
- #120103 (Make it so that async-fn-in-trait is compatible with a concrete future in implementation)
- #120206 (hir: Make sure all `HirId`s have corresponding HIR `Node`s)
- #120214 (match lowering: consistently lower bindings deepest-first)
- #120688 (GVN: also turn moves into copies with projections)
- #120702 (docs: also check the inline stmt during redundant link check)
- #120727 (exhaustiveness: Prefer "`0..MAX` not covered" to "`_` not covered")
- #120734 (Add `SubdiagnosticMessageOp` as a trait alias.)
- #120739 (improve pretty printing for associated items in trait objects)
r? `@ghost`
`@rustbot` modify labels: rollup
match lowering: consistently lower bindings deepest-first
Currently when lowering match expressions to MIR, we do a funny little dance with the order of bindings. I attempt to explain it in the third commit: we handle refutable (i.e. needing a test) patterns differently than irrefutable ones. This leads to inconsistencies, as reported in https://github.com/rust-lang/rust/issues/120210. The reason we need a dance at all is for situations like:
```rust
fn foo1(x: NonCopyStruct) {
let y @ NonCopyStruct { copy_field: z } = x;
// the above should turn into
let z = x.copy_field;
let y = x;
}
```
Here the `y ```````@```````` binding will move out of `x`, so we need to copy the field first.
I believe that the inconsistency came about when we fixed https://github.com/rust-lang/rust/issues/69971, and didn't notice that the fix didn't extend to refutable patterns. My guess then is that ordering bindings by "deepest-first, otherwise source order" is a sound choice. This PR implements that (at least I hope, match lowering is hard to follow 🥲).
Fixes https://github.com/rust-lang/rust/issues/120210
r? ```````@oli-obk``````` since you merged the original fix to https://github.com/rust-lang/rust/issues/69971
cc ```````@matthewjasper```````
update indirect structural match lints to match RFC and to show up for dependencies
This is a large step towards implementing https://github.com/rust-lang/rfcs/pull/3535.
We currently have five lints related to "the structural match situation":
- nontrivial_structural_match
- indirect_structural_match
- pointer_structural_match
- const_patterns_without_partial_eq
- illegal_floating_point_literal_pattern
This PR concerns the first 3 of them. (The 4th already is set up to show for dependencies, and the 5th is removed by https://github.com/rust-lang/rust/pull/116284.) nontrivial_structural_match is being removed as per the RFC; the other two are enabled to show up in dependencies.
Fixes https://github.com/rust-lang/rust/issues/73448 by removing the affected analysis.
pattern_analysis: use a plain `Vec` in `DeconstructedPat`
The use of an arena-allocated slice in `DeconstructedPat` dates to when we needed the arena anyway for lifetime reasons. Now that we don't, I'm thinking that if `thir::Pat` can use plain old `Vec`s, maybe so can I.
r? ```@ghost```
That is, change `diagnostic_outside_of_impl` and
`untranslatable_diagnostic` from `allow` to `deny`, because more than
half of the compiler has be converted to use translated diagnostics.
This commit removes more `deny` attributes than it adds `allow`
attributes, which proves that this change is warranted.
make matching on NaN a hard error, and remove the rest of illegal_floating_point_literal_pattern
These arms would never be hit anyway, so the pattern makes little sense. We have had a future-compat lint against float matches in general for a *long* time, so I hope we can get away with immediately making this a hard error.
This is part of implementing https://github.com/rust-lang/rfcs/pull/3535.
Closes https://github.com/rust-lang/rust/issues/41620 by removing the lint.
https://github.com/rust-lang/reference/pull/1456 updates the reference to match.
Because it's almost always static.
This makes `impl IntoDiagnosticArg for DiagnosticArgValue` trivial,
which is nice.
There are a few diagnostics constructed in
`compiler/rustc_mir_build/src/check_unsafety.rs` and
`compiler/rustc_mir_transform/src/errors.rs` that now need symbols
converted to `String` with `to_string` instead of `&str` with `as_str`,
but that' no big deal, and worth it for the simplifications elsewhere.
Borrow check inline const patterns
Add type annotations to MIR so that borrowck can pass constraints from inline constants in patterns to the containing function.
Also enables some inline constant pattern tests that were fixed by the THIR unsafeck stabilization.
cc #76001
Error codes are integers, but `String` is used everywhere to represent
them. Gross!
This commit introduces `ErrCode`, an integral newtype for error codes,
replacing `String`. It also introduces a constant for every error code,
e.g. `E0123`, and removes the `error_code!` macro. The constants are
imported wherever used with `use rustc_errors::codes::*`.
With the old code, we have three different ways to specify an error code
at a use point:
```
error_code!(E0123) // macro call
struct_span_code_err!(dcx, span, E0123, "msg"); // bare ident arg to macro call
\#[diag(name, code = "E0123")] // string
struct Diag;
```
With the new code, they all use the `E0123` constant.
```
E0123 // constant
struct_span_code_err!(dcx, span, E0123, "msg"); // constant
\#[diag(name, code = E0123)] // constant
struct Diag;
```
The commit also changes the structure of the error code definitions:
- `rustc_error_codes` now just defines a higher-order macro listing the
used error codes and nothing else.
- Because that's now the only thing in the `rustc_error_codes` crate, I
moved it into the `lib.rs` file and removed the `error_codes.rs` file.
- `rustc_errors` uses that macro to define everything, e.g. the error
code constants and the `DIAGNOSTIC_TABLES`. This is in its new
`codes.rs` file.
Remove unused/unnecessary features
~~The bulk of the actual code changes here is replacing try blocks with equivalent closures. I'm not entirely sure that's a good idea since it may have perf impact, happy to revert if that's the case/the change is unwanted.~~
I also removed a lot of `recursion_limit = "256"` since everything seems to build fine without that and most don't have any comment justifying it.
remove StructuralEq trait
The documentation given for the trait is outdated: *all* function pointers implement `PartialEq` and `Eq` these days. So the `StructuralEq` trait doesn't really seem to have any reason to exist any more.
One side-effect of this PR is that we allow matching on some consts that do not implement `Eq`. However, we already allowed matching on floats and consts containing floats, so this is not new, it is just allowed in more cases now. IMO it makes no sense at all to allow float matching but also sometimes require an `Eq` instance. If we want to require `Eq` we should adjust https://github.com/rust-lang/rust/pull/115893 to check for `Eq`, and rule out float matching for good.
Fixes https://github.com/rust-lang/rust/issues/115881
Rollup of 10 pull requests
Successful merges:
- #117910 (Refactor uses of `objc_msgSend` to no longer have clashing definitions)
- #118639 (Undeprecate lint `unstable_features` and make use of it in the compiler)
- #119801 (Fix deallocation with wrong allocator in (A)Rc::from_box_in)
- #120058 (bootstrap: improvements for compiler builds)
- #120059 (Make generic const type mismatches not hide trait impls from the trait solver)
- #120097 (Report unreachable subpatterns consistently)
- #120137 (Validate AggregateKind types in MIR)
- #120164 (`maybe_lint_impl_trait`: separate `is_downgradable` from `is_object_safe`)
- #120181 (Allow any `const` expression blocks in `thread_local!`)
- #120218 (rustfmt: Check that a token can begin a nonterminal kind before parsing it as a macro arg)
r? `@ghost`
`@rustbot` modify labels: rollup
Report unreachable subpatterns consistently
We weren't reporting unreachable subpatterns in function arguments and `let` expressions. This wasn't very important, but never patterns make it more relevant: a user might write `let (Ok(x) | Err(!)) = ...` in a case where `let Ok(x) = ...` is accepted, so we should report the `Err(!)` as redundant.
r? ```@compiler-errors```
Pack u128 in the compiler to mitigate new alignment
This is based on #116672, adding a new `#[repr(packed(8))]` wrapper on `u128` to avoid changing any of the compiler's size assertions. This is needed in two places:
* `SwitchTargets`, otherwise its `SmallVec<[u128; 1]>` gets padded up to 32 bytes.
* `LitKind::Int`, so that entire `enum` can stay 24 bytes.
* This change definitely has far-reaching effects though, since it's public.
never_patterns: typecheck never patterns
This checks that a `!` pattern is only used on an uninhabited type (modulo match ergonomics, i.e. `!` is allowed on `&Void`).
r? `@compiler-errors`
Exhaustiveness: simplify empty pattern logic
The logic that handles empty patterns had gotten quite convoluted. This PR simplifies it a lot. I tried to make the logic as easy as possible to follow; this only does logically equivalent changes.
The first commit is a drive-by comment clarification that was requested after another PR a while back.
r? `@compiler-errors`
Simplify `closure_env_ty` and `closure_env_param`
Random cleanup that I found when working on async closures. This makes it easier to separate the latter into a new tykind.
Use `zip_eq` to enforce that things being zipped have equal sizes
Some `zip`s are best enforced to be equal, since size mismatches suggest deeper bugs in the compiler.
We have `span_delayed_bug` and often pass it a `DUMMY_SP`. This commit
adds `delayed_bug`, which matches pairs like `err`/`span_err` and
`warn`/`span_warn`.
Because it takes an error code after the span. This avoids the confusing
overlap with the `DiagCtxt::struct_span_err` method, which doesn't take
an error code.
Exhaustiveness: Statically enforce revealing of opaques
In https://github.com/rust-lang/rust/pull/116821 it was decided that exhaustiveness should operate on the hidden type of an opaque type when relevant. This PR makes sure we consistently reveal opaques within exhaustiveness. This makes it possible to remove `reveal_opaque_ty` from the `TypeCx` trait which was an unfortunate implementation detail.
r? `@compiler-errors`
Check yield terminator's resume type in borrowck
In borrowck, we didn't check that the lifetimes of the `TerminatorKind::Yield`'s `resume_place` were actually compatible with the coroutine's signature. That means that the lifetimes were totally going unchecked. Whoops!
This PR implements this checking.
Fixes#119564
r? types
Fix scoping for let chains in match guards
If let guards were previously represented as a different type of guard in HIR and THIR. This meant that let chains in match guards were not handled correctly because they were treated exactly like normal guards.
- Remove `hir::Guard` and `thir::Guard`.
- Make the scoping different between normal guards and if let guards also check for let chains.
closes#118593
Match guards with an if let guard or an if let chain guard should have a
temporary scope of the whole arm. This is to allow ref bindings to
temporaries to borrow check.
custom mir: make it clear what the return block is
Custom MIR recently got support for specifying the "unwind action", so now there's two things coming after the actual call part of `Call` terminators. That's not very self-explaining so I propose we change the syntax to imitate keyword arguments:
```
Call(popped = Vec::pop(v), ReturnTo(drop), UnwindContinue())
```
Also fix some outdated docs and add some docs to `Call` and `Drop`.
`Diagnostic` has 40 methods that return `&mut Self` and could be
considered setters. Four of them have a `set_` prefix. This doesn't seem
necessary for a type that implements the builder pattern. This commit
removes the `set_` prefixes on those four methods.
Clean up some lifetimes in `rustc_pattern_analysis`
This PR removes some redundant lifetimes. I figured out that we were shortening the lifetime of an arena-allocated `&'p DeconstructedPat<'p>` to `'a DeconstructedPat<'p>`, which forced us to carry both lifetimes when we could otherwise carry just one.
This PR also removes and elides some unnecessary lifetimes.
I also cherry-picked 0292eb9bb9b897f5c0926c6a8530877f67e7cc9b, and then simplified more lifetimes in `MatchVisitor`, which should make #119233 a very simple PR!
r? Nadrieril
Remove `DiagCtxt` API duplication
`DiagCtxt` defines the internal API for creating and emitting diagnostics: methods like `struct_err`, `struct_span_warn`, `note`, `create_fatal`, `emit_bug`. There are over 50 methods.
Some of these methods are then duplicated across several other types: `Session`, `ParseSess`, `Parser`, `ExtCtxt`, and `MirBorrowckCtxt`. `Session` duplicates the most, though half the ones it does are unused. Each duplicated method just calls forward to the corresponding method in `DiagCtxt`. So this duplication exists to (in the best case) shorten chains like `ecx.tcx.sess.parse_sess.dcx.emit_err()` to `ecx.emit_err()`.
This API duplication is ugly and has been bugging me for a while. And it's inconsistent: there's no real logic about which methods are duplicated, and the use of `#[rustc_lint_diagnostic]` and `#[track_caller]` attributes vary across the duplicates.
This PR removes the duplicated API methods and makes all diagnostic creation and emission go through `DiagCtxt`. It also adds `dcx` getter methods to several types to shorten chains. This approach scales *much* better than API duplication; indeed, the PR adds `dcx()` to numerous types that didn't have API duplication: `TyCtxt`, `LoweringCtxt`, `ConstCx`, `FnCtxt`, `TypeErrCtxt`, `InferCtxt`, `CrateLoader`, `CheckAttrVisitor`, and `Resolver`. These result in a lot of changes from `foo.tcx.sess.emit_err()` to `foo.dcx().emit_err()`. (You could do this with more types, but it gets into diminishing returns territory for types that don't emit many diagnostics.)
After all these changes, some call sites are more verbose, some are less verbose, and many are the same. The total number of lines is reduced, mostly because of the removed API duplication. And consistency is increased, because calls to `emit_err` and friends are always preceded with `.dcx()` or `.dcx`.
r? `@compiler-errors`
Give temporaries in if let guards correct scopes
Temporaries in if-let guards have scopes that escape the match arm, this causes problems because the drops might be for temporaries that are not storage live. This PR changes the scope of temporaries in if-let guards to be limited to the arm:
```rust
_ if let Some(s) = std::convert::identity(&Some(String::new())) => {}
// Temporary for Some(String::new()) is dropped here ^
```
We also now deduplicate temporaries between copies of the guard created for or-patterns:
```rust
// Only create a single Some(String::new()) temporary variable
_ | _ if let Some(s) = std::convert::identity(&Some(String::new())) => {}
```
This changes MIR building to pass around `ExprId`s rather than `Expr`s so that we have a way to index different expressions.
cc #51114Closes#116079
`IntoDiagnostic` defaults to `ErrorGuaranteed`, because errors are the
most common diagnostic level. It makes sense to do likewise for the
closely-related (and much more widely used) `DiagnosticBuilder` type,
letting us write `DiagnosticBuilder<'a, ErrorGuaranteed>` as just
`DiagnosticBuilder<'a>`. This cuts over 200 lines of code due to many
multi-line things becoming single line things.
Exhaustiveness: reveal opaque types properly
Previously, exhaustiveness had no clear policy around opaque types. In this PR I propose the following policy: within the body of an item that defines the hidden type of some opaque type, exhaustiveness checking on a value of that opaque type is performed using the concrete hidden type inferred in this body.
I'm not sure how consistent this is with other operations allowed on opaque types; I believe this will require FCP.
From what I can tell, this doesn't change anything for non-empty types.
The observable changes are:
- when the real type is uninhabited, matches within the defining scopes can now rely on that for exhaustiveness, e.g.:
```rust
#[derive(Copy, Clone)]
enum Void {}
fn return_never_rpit(x: Void) -> impl Copy {
if false {
match return_never_rpit(x) {}
}
x
}
```
- this properly fixes ICEs like https://github.com/rust-lang/rust/issues/117100 that occurred because a same match could have some patterns where the type is revealed and some where it is not.
Bonus subtle point: if `x` is opaque, a match like `match x { ("", "") => {} ... }` will constrain its type ([playground](https://play.rust-lang.org/?version=nightly&mode=debug&edition=2021&gist=901d715330eac40339b4016ac566d6c3)). This is not the case for `match x {}`: this will not constain the type, and will only compile if something else constrains the type to be empty.
Fixes https://github.com/rust-lang/rust/issues/117100
r? `@oli-obk`
Edited for precision of the wording
[Included](https://github.com/rust-lang/rust/pull/116821#issuecomment-1813171764) in the FCP on this PR is this rule:
> Within the body of an item that defines the hidden type of some opaque type, exhaustiveness checking on a value of that opaque type is performed using the concrete hidden type inferred in this body.
- Make temporaries in if-let guards be the same variable in MIR when
the guard is duplicated due to or-patterns.
- Change the "destruction scope" for match arms to be the arm scope rather
than the arm body scope.
- Add tests.
match lowering: Remove the `make_target_blocks` hack
This hack was introduced 4 years ago in [`a1d0266` (#60730)](a1d0266878) to improve LLVM optimization time, specifically noticed in the `encoding` benchmark. Measurements today indicate it is no longer needed.
r? `@matthewjasper`
Make exhaustiveness usable outside of rustc
With this PR, `rustc_pattern_analysis` compiles on stable (with the `stable` feature)! `rust-analyzer` will be able to use it to provide match-related diagnostics and refactors.
Two questions:
- Should I name the feature `nightly` instead of `rustc` for consistency with other crates? `rustc` makes more sense imo.
- `typed-arena` is an optional dependency but tidy made me add it to the allow-list anyway. Can I avoid that somehow?
r? `@compiler-errors`
And make all hand-written `IntoDiagnostic` impls generic, by using
`DiagnosticBuilder::new(dcx, level, ...)` instead of e.g.
`dcx.struct_err(...)`.
This means the `create_*` functions are the source of the error level.
This change will let us remove `struct_diagnostic`.
Note: `#[rustc_lint_diagnostics]` is added to `DiagnosticBuilder::new`,
it's necessary to pass diagnostics tests now that it's used in
`into_diagnostic` functions.
Annotate some bugs
Gives a semi-helpful message to some `bug!()`/`unreachable!()`/`panic!()`. This also works around some other bugs/panics/etc that weren't needed, and also makes some of them into `span_bug!`s so they also have a useful span.
Note to reviewer: best to disable whitespace when comparing for some cases where indentation changed.
cc #118955
rustc_mir_build: Enforce `rustc::potential_query_instability` lint
Stop allowing `rustc::potential_query_instability` on all of `rustc_mir_build` and instead allow it on a case-by-case basis if it is safe to do so. In this crate there was only one instance of the lint, and it was safe to allow.
Part of https://github.com/rust-lang/rust/issues/84447 which is E-help-wanted.
Stop allowing `rustc::potential_query_instability` on all of
`rustc_mir_build` and instead allow it on a case-by-case basis if it is
safe to do so. In this crate there was no instance of the lint
remaining.
Renamings:
- find -> opt_hir_node
- get -> hir_node
- find_by_def_id -> opt_hir_node_by_def_id
- get_by_def_id -> hir_node_by_def_id
Fix rebase changes using removed methods
Use `tcx.hir_node_by_def_id()` whenever possible in compiler
Fix clippy errors
Fix compiler
Apply suggestions from code review
Co-authored-by: Vadim Petrochenkov <vadim.petrochenkov@gmail.com>
Add FIXME for `tcx.hir()` returned type about its removal
Simplify with with `tcx.hir_node_by_def_id`
Extract exhaustiveness into its own crate
It now makes sense to extract exhaustiveness into its own crate! This was much-requested by rust-analyzer (they currently maintain by hand a copy of the algorithm), and I hope this can serve other projects e.g. clippy.
This is the churny PR: it exclusively moves code around. It's not yet useable outside of rustc but I wanted the churny parts to be out of the way.
r? `@compiler-errors`
remove redundant imports
detects redundant imports that can be eliminated.
for #117772 :
In order to facilitate review and modification, split the checking code and removing redundant imports code into two PR.
r? `@petrochenkov`
Don't print host effect param in pretty `path_generic_args`
Make `own_args_no_defaults` pass back the `GenericParamDef`, so that we can pass both the args *and* param definitions into `path_generic_args`. That allows us to use the `GenericParamDef` to filter out effect params.
This allows us to filter out the host param regardless of whether it's `sym::host` or `true`/`false`.
This also renames a couple of `const_effect_param` -> `host_effect_param`, and restores `~const` pretty printing to `TraitPredPrintModifiersAndPath`.
cc #118785
r? `@fee1-dead` cc `@oli-obk`
detects redundant imports that can be eliminated.
for #117772 :
In order to facilitate review and modification, split the checking code and
removing redundant imports code into two PR.
Don't warn an empty pattern unreachable if we're not sure the data is valid
Exhaustiveness checking used to be naive about the possibility of a place containing invalid data. This could cause it to emit an "unreachable pattern" lint on an arm that was in fact reachable, as in https://github.com/rust-lang/rust/issues/117119.
This PR fixes that. We now track whether a place that is matched on may hold invalid data. This also forced me to be extra precise about how exhaustiveness manages empty types.
Note that this now errs in the opposite direction: the following arm is truly unreachable (because the binding causes a read of the value) but not linted as such. I'd rather not recommend writing a `match ... {}` that has the implicit side-effect of loading the value. [Never patterns](https://github.com/rust-lang/rust/issues/118155) will solve this cleanly.
```rust
match union.value {
_x => unreachable!(),
}
```
I recommend reviewing commit by commit. I went all-in on the test suite because this went through a lot of iterations and I kept everything. The bit I'm least confident in is `is_known_valid_scrutinee` in `check_match.rs`.
Fixes https://github.com/rust-lang/rust/issues/117119.
- `ConstructorSet` knows about both empty and nonempty constructors;
- If an empty constructor is present in the column, we output it in
`split().present`.
When MIR is built for an if-not expression, the `!` part of the condition
doesn't correspond to any MIR statement, so coverage instrumentation normally
can't see it.
We can fix that by deliberately injecting a dummy statement whose sole purpose
is to associate that span with its enclosing block.
There are cases where coverage instrumentation wants to show a span for some
syntax element, but there is no MIR node that naturally carries that span, so
the instrumentor can't see it.
MIR building can now use this new kind of coverage statement to deliberately
include those spans in MIR, attached to a dummy statement that has no other
effect.
Remove the `precise_pointer_size_matching` feature gate
`usize` and `isize` are special for pattern matching because their range might depend on the platform. To make code portable across platforms, the following is never considered exhaustive:
```rust
let x: usize = ...;
match x {
0..=18446744073709551615 => {}
}
```
Because of how rust handles constants, this also unfortunately counts `0..=usize::MAX` as non-exhaustive. The [`precise_pointer_size_matching`](https://github.com/rust-lang/rust/issues/56354) feature gate was introduced both for this convenience and for the possibility that the lang team could decide to allow the above.
Since then, [half-open range patterns](https://github.com/rust-lang/rust/issues/67264) have been implemented, and since #116692 they correctly support `usize`/`isize`:
```rust
match 0usize { // exhaustive!
0..5 => {}
5.. => {}
}
```
I believe this subsumes all the use cases of the feature gate. Moreover no attempt has been made to stabilize it in the 5 years of its existence. I therefore propose we retire this feature gate.
Closes https://github.com/rust-lang/rust/issues/56354
Exhaustiveness: allocate memory better
Exhaustiveness is a recursive algorithm that allocates a bunch of slices at every step. Let's see if I can improve performance by improving allocations.
Already just using `Vec::with_capacity` is showing impressive improvements on my local measurements.
r? `@ghost`
Fix `PartialEq` args when `#[const_trait]` is enabled
This is based off of your PR that enforces effects on all methods, so just see the last commits.
r? fee1-dead
Add `never_patterns` feature gate
This PR adds the feature gate and most basic parsing for the experimental `never_patterns` feature. See the tracking issue (https://github.com/rust-lang/rust/issues/118155) for details on the experiment.
`@scottmcm` has agreed to be my lang-team liaison for this experiment.
effects: Run `enforce_context_effects` for all method calls
So that we also perform checks when overloaded `PartialEq`s are called.
r? `@compiler-errors`
Rewrite exhaustiveness in one pass
This is at least my 4th attempt at this in as many years x) Previous attempts were all too complicated or too slow. But we're finally here!
The previous version of the exhaustiveness algorithm computed reachability for each arm then exhaustiveness of the whole match. Since each of these steps does roughly the same things, this rewrites the algorithm to do them all in one go. I also think this makes things much simpler.
I also rewrote the documentation of the algorithm in depth. Hopefully it's up-to-date and easier to follow now. Plz comment if anything's unclear.
r? `@oli-obk` I think you're one of the rare other people to understand the exhaustiveness algorithm?
cc `@varkor` I know you're not active anymore, but if you feel like having a look you might enjoy this :D
Fixes https://github.com/rust-lang/rust/issues/79307
Currently we always do this:
```
use rustc_fluent_macro::fluent_messages;
...
fluent_messages! { "./example.ftl" }
```
But there is no need, we can just do this everywhere:
```
rustc_fluent_macro::fluent_messages! { "./example.ftl" }
```
which is shorter.
The `fluent_messages!` macro produces uses of
`crate::{D,Subd}iagnosticMessage`, which means that every crate using
the macro must have this import:
```
use rustc_errors::{DiagnosticMessage, SubdiagnosticMessage};
```
This commit changes the macro to instead use
`rustc_errors::{D,Subd}iagnosticMessage`, which avoids the need for the
imports.
By default, `newtype_index!` types get a default `Encodable`/`Decodable`
impl. You can opt out of this with `custom_encodable`. Opting out is the
opposite to how Rust normally works with autogenerated (derived) impls.
This commit inverts the behaviour, replacing `custom_encodable` with
`encodable` which opts into the default `Encodable`/`Decodable` impl.
Only 23 of the 59 `newtype_index!` occurrences need `encodable`.
Even better, there were eight crates with a dependency on
`rustc_serialize` just from unused default `Encodable`/`Decodable`
impls. This commit removes that dependency from those eight crates.
This disentangles the row-specific tracking of `parent_row` etc from the
logical operation of specialization. This means `wildcard_row` doesn't
need to provide dummy values for `parent_row` etc anymore.
interpret: simplify handling of shifts by no longer trying to handle signed and unsigned shift amounts in the same branch
While we're at it, also update comments in codegen and MIR building related to shifts, and fix the overflow error printed by Miri on negative shift amounts.
Build pre-coroutine-transform coroutine body on error
I was accidentally building the post-transform coroutine body, rather than the pre-transform coroutine body. There's no pinning expected here yet, and the return type isn't yet transformed into `CoroutineState`.
Fixes#117670
Custom MIR: Support cleanup blocks
Cleanup blocks are declared with `bb (cleanup) = { ... }`.
`Call` and `Drop` terminators take an additional argument describing the unwind action, which is one of the following:
* `UnwindContinue()`
* `UnwindUnreachable()`
* `UnwindTerminate(reason)`, where reason is `ReasonAbi` or `ReasonInCleanup`
* `UnwindCleanup(block)`
Also support unwind resume and unwind terminate terminators:
* `UnwindResume()`
* `UnwindTerminate(reason)`
Cleanup blocks are declared with `bb (cleanup) = { ... }`.
`Call` and `Drop` terminators take an additional argument describing the
unwind action, which is one of the following:
* `UnwindContinue()`
* `UnwindUnreachable()`
* `UnwindTerminate(reason)`, where reason is `ReasonAbi` or `ReasonInCleanup`
* `UnwindCleanup(block)`
Also support unwind resume and unwind terminate terminators:
* `UnwindResume()`
* `UnwindTerminate(reason)`
generator layout: ignore fake borrows
fixes#117059
We emit fake shallow borrows in case the scrutinee place uses a `Deref` and there is a match guard. This is necessary to prevent the match guard from mutating the scrutinee: fab1054e17/compiler/rustc_mir_build/src/build/matches/mod.rs (L1250-L1265)
These fake borrows end up impacting the generator witness computation in `mir_generator_witnesses`, which causes the issue in #117059. This PR now completely ignores fake borrows during this computation. This is sound as thse are always removed after analysis and the actual computation of the generator layout happens afterwards.
Only the second commit impacts behavior, and could be backported by itself.
r? types
patterns: reject raw pointers that are not just integers
Matching against `0 as *const i32` is fine, matching against `&42 as *const i32` is not.
This extends the existing check against function pointers and wide pointers: we now uniformly reject all these pointer types during valtree construction, and then later lint because of that. See [here](https://github.com/rust-lang/rust/pull/116930#issuecomment-1784654073) for some more explanation and context.
Also fixes https://github.com/rust-lang/rust/issues/116929.
Cc `@oli-obk` `@lcnr`
Thir unsafeck fixes
- Recognise thread local statics in THIR unsafeck
- Add suggestion for unsafe_op_in_unsafe_fn
- Fix unsafe checking of let expressions
Cleanup `rustc_mir_build/../check_match.rs`
The file had become pretty unwieldy, with a fair amount of duplication. As a bonus, I discovered that we weren't running some pattern checks in if-let chains.
I recommend looking commit-by-commit. The last commit is a whim, I think it makes more sense that way but I don't hold this opinion strongly.
Support enum variants in offset_of!
This MR implements support for navigating through enum variants in `offset_of!`, placing the enum variant name in the second argument to `offset_of!`. The RFC placed it in the first argument, but I think it interacts better with nested field access in the second, as you can then write things like
```rust
offset_of!(Type, field.Variant.field)
```
Alternatively, a syntactic distinction could be made between variants and fields (e.g. `field::Variant.field`) but I'm not convinced this would be helpful.
[RFC 3308 # Enum Support](https://rust-lang.github.io/rfcs/3308-offset_of.html#enum-support-offset_ofsomeenumstructvariant-field_on_variant)
Tracking Issue #106655.
Match usize/isize exhaustively with half-open ranges
The long-awaited finale to the saga of [exhaustiveness checking for integers](https://github.com/rust-lang/rust/pull/50912)!
```rust
match 0usize {
0.. => {} // exhaustive!
}
match 0usize {
0..usize::MAX => {} // helpful error message!
}
```
Features:
- Half-open ranges behave as expected for `usize`/`isize`;
- Trying to use `0..usize::MAX` will tell you that `usize::MAX..` is missing and explain why. No more unhelpful "`_` is missing";
- Everything else stays the same.
This should unblock https://github.com/rust-lang/rust/issues/37854.
Review-wise:
- I recommend looking commit-by-commit;
- This regresses perf because of the added complexity in `IntRange`; hopefully not too much;
- I measured each `#[inline]`, they all help a bit with the perf regression (tho I don't get why);
- I did not touch MIR building; I expect there's an easy PR there that would skip unnecessary comparisons when the range is half-open.
Rollup of 7 pull requests
Successful merges:
- #116862 (Detect when trait is implemented for type and suggest importing it)
- #117389 (Some diagnostics improvements of `gen` blocks)
- #117396 (Don't treat closures/coroutine types as part of the public API)
- #117398 (Correctly handle nested or-patterns in exhaustiveness)
- #117403 (Poison check_well_formed if method receivers are invalid to prevent typeck from running on it)
- #117411 (Improve some diagnostics around `?Trait` bounds)
- #117414 (Don't normalize to an un-revealed opaque when we hit the recursion limit)
r? `@ghost`
`@rustbot` modify labels: rollup
Correctly handle nested or-patterns in exhaustiveness
I had assumed nested or-patterns were flattened, and they mostly are but not always.
Fixes https://github.com/rust-lang/rust/issues/117378
- Sort dependencies and features sections.
- Add `tidy` markers to the sorted sections so they stay sorted.
- Remove empty `[lib`] sections.
- Remove "See more keys..." comments.
Excluded files:
- rustc_codegen_{cranelift,gcc}, because they're external.
- rustc_lexer, because it has external use.
- stable_mir, because it has external use.
Allow partially moved values in match
This PR attempts to unify the behaviour between `let _ = PLACE`, `let _: TY = PLACE;` and `match PLACE { _ => {} }`.
The logical conclusion is that the `match` version should not check for uninitialised places nor check that borrows are still live.
The `match PLACE {}` case is handled by keeping a `FakeRead` in the unreachable fallback case to verify that `PLACE` has a legal value.
Schematically, `match PLACE { arms }` in surface rust becomes in MIR:
```rust
PlaceMention(PLACE)
match PLACE {
// Decision tree for the explicit arms
arms,
// An extra fallback arm
_ => {
FakeRead(ForMatchedPlace, PLACE);
unreachable
}
}
```
`match *borrow { _ => {} }` continues to check that `*borrow` is live, but does not read the value.
`match *borrow {}` both checks that `*borrow` is live, and fake-reads the value.
Continuation of ~https://github.com/rust-lang/rust/pull/102256~ ~https://github.com/rust-lang/rust/pull/104844~
Fixes https://github.com/rust-lang/rust/issues/99180https://github.com/rust-lang/rust/issues/53114
Never consider raw pointer casts to be trival
HIR typeck tries to figure out which casts are trivial by doing them as
coercions and seeing whether this works. Since HIR typeck is oblivious
of lifetimes, this doesn't work for pointer casts that only change the
lifetime of the pointee, which are, as borrowck will tell you, not
trivial.
This change makes it so that raw pointer casts are never considered
trivial.
This also incidentally fixes the "trivial cast" lint false positive on
the same code. Unfortunately, "trivial cast" lints are now never emitted
on raw pointer casts, even if they truly are trivial. This could be
fixed by also doing the lint in borrowck for raw pointers specifically.
fixes#113257
Lint `non_exhaustive_omitted_patterns` by columns
This is a rework of the `non_exhaustive_omitted_patterns` lint to make it more consistent. The intent of the lint is to help consumers of `non_exhaustive` enums ensure they stay up-to-date with all upstream variants. This rewrite fixes two cases we didn't handle well before:
First, because of details of exhaustiveness checking, the following wouldn't lint `Enum::C` as missing:
```rust
match Some(x) {
Some(Enum::A) => {}
Some(Enum::B) => {}
_ => {}
}
```
Second, because of the fundamental workings of exhaustiveness checking, the following would treat the `true` and `false` cases separately and thus lint about missing variants:
```rust
match (true, x) {
(true, Enum::A) => {}
(true, Enum::B) => {}
(false, Enum::C) => {}
_ => {}
}
```
Moreover, it would correctly not lint in the case where the pair is flipped, because of asymmetry in how exhaustiveness checking proceeds.
A drawback is that it no longer makes sense to set the lint level per-arm. This will silently break the lint for current users of it (but it's behind a feature gate so that's ok).
The new approach is now independent of the exhaustiveness algorithm; it's a separate pass that looks at patterns column by column. This is another of the motivations for this: I'm glad to move it out of the algorithm, it was akward there.
This PR is almost identical to https://github.com/rust-lang/rust/pull/111651. cc `@eholk` who reviewed it at the time. Compared to then, I'm more confident this is the right approach.
This allows coverage information to be attached to the function as a whole when
appropriate, instead of being smuggled through coverage statements in the
function's basic blocks.
As an example, this patch moves the `function_source_hash` value out of
individual `CoverageKind::Counter` statements and into the per-function info.
When synthesizing unused functions for coverage purposes, the absence of this
info is taken to indicate that a function was not eligible for coverage and
should not be synthesized.
THIR unsafety checking was getting a cycle of
function unsafety checking
-> building THIR for the function
-> evaluating pattern inline constants in the function
-> building MIR for the inline constant
-> checking unsafety of functions (so that THIR can be stolen)
This is fixed by not stealing THIR when generating MIR but instead when
unsafety checking.
This leaves an issue with pattern inline constants not being unsafety
checked because they are evaluated away when generating THIR.
To fix that we now represent inline constants in THIR patterns and
visit them in THIR unsafety checking.
use `PatKind::Error` when an ADT const value has violation
Fixes#115599
Since the [to_pat](https://github.com/rust-lang/rust/pull/111913/files#diff-6d8d99538aca600d633270051580c7a9e40b35824ea2863d9dda2c85a733b5d9R126-R155) behavior has been changed in the #111913 update, the kind of `inlined_const_ast_pat` has transformed from `PatKind::Leaf { pattern: Pat { kind: Wild, ..} } ` to `PatKind::Constant`. This caused a scenario where there are no matched candidates, leading to a testing of the candidates. This process ultimately attempts to test the string const, triggering the `bug!` invocation finally.
r? ``@oli-obk``
Format all the let-chains in compiler crates
Since rust-lang/rustfmt#5910 has landed, soon we will have support for formatting let-chains (as soon as rustfmt syncs and beta gets bumped).
This PR applies the changes [from master rustfmt to rust-lang/rust eagerly](https://rust-lang.zulipchat.com/#narrow/stream/122651-general/topic/out.20formatting.20of.20prs/near/374997516), so that the next beta bump does not have to deal with a 200+ file diff and can remain concerned with other things like `cfg(bootstrap)` -- #113637 was a pain to land, for example, because of let-else.
I will also add this commit to the ignore list after it has landed.
The commands that were run -- I'm not great at bash-foo, but this applies rustfmt to every compiler crate, and then reverts the two crates that should probably be formatted out-of-tree.
```
~/rustfmt $ ls -1d ~/rust/compiler/* | xargs -I@ cargo run --bin rustfmt -- `@/src/lib.rs` --config-path ~/rust --edition=2021 # format all of the compiler crates
~/rust $ git checkout HEAD -- compiler/rustc_codegen_{gcc,cranelift} # revert changes to cg-gcc and cg-clif
```
cc `@rust-lang/rustfmt`
r? `@WaffleLapkin` or `@Nilstrieb` who said they may be able to review this purely mechanical PR :>
cc `@Mark-Simulacrum` and `@petrochenkov,` who had some thoughts on the order of operations with big formatting changes in https://github.com/rust-lang/rust/pull/95262#issue-1178993801. I think the situation has changed since then, given that let-chains support exists on master rustfmt now, and I'm fairly confident that this formatting PR should land even if *bootstrap* rustfmt doesn't yet format let-chains in order to lessen the burden of the next beta bump.
exhaustiveness: Rework constructor splitting
`SplitWildcard` was pretty opaque. I replaced it with a more legible abstraction: `ConstructorSet` represents the set of constructors for patterns of a given type. This clarifies responsibilities: `ConstructorSet` handles one clear task, and diagnostic-related shenanigans can be done separately.
I'm quite excited, I had has this in mind for years but could never quite introduce it. This opens up possibilities, including type-specific optimisations (like using a `FxHashSet` to collect enum variants, which had been [hackily attempted some years ago](https://github.com/rust-lang/rust/pull/76918)), my one-pass rewrite (https://github.com/rust-lang/rust/pull/116042), and future librarification.
Fix overflow checking in range patterns
When a range pattern contains an overflowing literal, if we're not careful we might not notice the overflow and use the wrapped value. This makes for confusing error messages because linting against overflowing literals is only done in a later pass. So when a range is invalid we check for overflows to provide a better error.
This check didn't use to handle negative types; this PR fixes that. First commit adds tests, second cleans up without changing behavior, third does the fix.
EDIT: while I was at it, I fixed a small annoyance about the span of the overflow lint on negated literals.
Fixes https://github.com/rust-lang/rust/issues/94239
