Implement `tainted_by_errors` in MIR borrowck, use it to skip CTFE
Putting this up for initial review. The issue that I found is when we're evaluating a const, we're doing borrowck, but doing nothing with the fact that borrowck fails.
This implements a `tainted_by_errors` field for MIR borrowck like we have in infcx, so we can use that information to return an `Err` during const eval if our const fails to borrowck.
This PR needs some cleaning up. I should probably just use `Result` in more places, instead of `.expect`ing in the places I am, but I just wanted it to compile so I could see if it worked!
Fixes#93646
r? `@oli-obk`
feel free to reassign
Use const generics in SipHasher128's short_write
This was proposed by `@michaelwoerister` [here](https://github.com/rust-lang/rust/pull/93615#discussion_r799485554).
A few comments:
1) I tried to pass `&[u8; LEN]` instead of `[u8; LEN]`. Locally, it resulted in small icount regressions (about 0.5 %). When passing by value, there were no regressions (and no improvements).
2) I wonder if we should use `to_ne_bytes()` in `SipHasher128` to keep it generic and only use `to_le_bytes()` in `StableHasher`. However, currently `SipHasher128` is only used in `StableHasher` and the `short_write` method was private, so I couldn't use it directly from `StableHasher`. Using `to_le()` in the `StableHasher` was breaking this abstraction boundary before slightly.
```rust
debug_assert!(LEN <= 8);
```
This could be done at compile time, but actually I think that now we can remove this assert altogether.
r? `@the8472`
This is required to avoid creating large numbers of universes from each
Chalk query, while still having enough universe information for lifetime
errors.
Fix incorrect register conflict detection in asm!
This would previously incorrectly reject two subregisters that were
distinct but part of the same larger register, for example `al` and
`ah`.
Split `pauth` target feature
Per discussion on https://github.com/rust-lang/rust/issues/86941 we'd like to split `pauth` into `paca` and `pacg` in order to better support possible future environments that only have the keys available for address or generic authentication. At the moment LLVM has the one `pauth` target_feature while Linux presents separate `paca` and `pacg` flags for feature detection.
Because the use of [target_feature](https://rust-lang.github.io/rfcs/2045-target-feature.html) will "allow the compiler to generate code under the assumption that this code will only be reached in hosts that support the feature", it does not make sense to simply translate `paca` into the LLVM feature `pauth`, as it will generate code as if `pacg` is available.
To accommodate this we error if only one of the two features is present. If LLVM splits them in the future we can remove this restriction without making a breaking change.
r? ```@Amanieu```
Make `span_extend_to_prev_str()` more robust
Fixes#91560. The logic in `span_extend_to_prev_str()` is currently quite brittle and fails if there is extra whitespace or something else in between, and it also should return an `Option` but doesn't currently.
This option introduced in #15820 allows a custom crate to be imported in
the place of std, but with the name std. I don't think there is any
value to this. At most it is confusing users of a driver that uses this option. There are no users of
this option on github. If anyone still needs it, they can emulate it
injecting #![no_core] in addition to their own prelude.
Make all `hir::Map` methods consistently by-value
`hir::Map` only consists of a single reference (as part of the contained `TyCtxt`) anyways, so copying is literally zero overhead compared to passing a reference
Erase regions before calculating layout for packed field capture
Self-explanatory. We just erase region inferencing because we don't need that for layout computation... Q: layouts are always equal modulo regions, right?
Fixes#92240
add kernel target for RustyHermit
Currently, we are thinking to use *-unknown-none targets instead to define for every platform our own one (see hermitcore/rusty-hermit#197). However, the current target aarch64-unknown-none-softfloat doesn't support dynamic relocation. Our RustyHermit project uses this feature and consequently we define a new target aarch64-unknown-hermitkernel to support it.
> A tier 3 target must have a designated developer or developers (the "target maintainers") on record to be CCed when issues arise regarding the target. (The mechanism to track and CC such developers may evolve over time.)
I would be willing to be a target maintainer, though I would appreciate if others volunteered to help with that as well.
> Targets must use naming consistent with any existing targets; for instance, a target for the same CPU or OS as an existing Rust target should use the same name for that CPU or OS. Targets should normally use the same names and naming conventions as used elsewhere in the broader ecosystem beyond Rust (such as in other toolchains), unless they have a very good reason to diverge. Changing the name of a target can be highly disruptive, especially once the target reaches a higher tier, so getting the name right is important even for a tier 3 target.
Uses the same naming as the LLVM target, and the same convention as many other kernel targets (e.g. `x86_64_unknown_none_linuxkernel`). In contrast to the bare-metal target for the aarch64 architecture, the unikernel requires dynamic relocation.
> Target names should not introduce undue confusion or ambiguity unless absolutely necessary to maintain ecosystem compatibility. For example, if the name of the target makes people extremely likely to form incorrect beliefs about what it targets, the name should be changed or augmented to disambiguate it.
I don't believe there is any ambiguity here. It use the same convention on x86_64 architecture.
> Tier 3 targets may have unusual requirements to build or use, but must not create legal issues or impose onerous legal terms for the Rust project or for Rust developers or users.
I don't see any legal issues here.
> The target must not introduce license incompatibilities.
Anything added to the Rust repository must be under the standard Rust license (MIT OR Apache-2.0).
The target must not cause the Rust tools or libraries built for any other host (even when supporting cross-compilation to the target) to depend on any new dependency less permissive than the Rust licensing policy. This applies whether the dependency is a Rust crate that would require adding new license exceptions (as specified by the tidy tool in the rust-lang/rust repository), or whether the dependency is a native library or binary. In other words, the introduction of the target must not cause a user installing or running a version of Rust or the Rust tools to be subject to any new license requirements.
If the target supports building host tools (such as rustc or cargo), those host tools must not depend on proprietary (non-FOSS) libraries, other than ordinary runtime libraries supplied by the platform and commonly used by other binaries built for the target. For instance, rustc built for the target may depend on a common proprietary C runtime library or console output library, but must not depend on a proprietary code generation library or code optimization library. Rust's license permits such combinations, but the Rust project has no interest in maintaining such combinations within the scope of Rust itself, even at tier 3.
Targets should not require proprietary (non-FOSS) components to link a functional binary or library.
"onerous" here is an intentionally subjective term. At a minimum, "onerous" legal/licensing terms include but are not limited to: non-disclosure requirements, non-compete requirements, contributor license agreements (CLAs) or equivalent, "non-commercial"/"research-only"/etc terms, requirements conditional on the employer or employment of any particular Rust developers, revocable terms, any requirements that create liability for the Rust project or its developers or users, or any requirements that adversely affect the livelihood or prospects of the Rust project or its developers or users.
I see no issues with any of the above.
> Neither this policy nor any decisions made regarding targets shall create any binding agreement or estoppel by any party. If any member of an approving Rust team serves as one of the maintainers of a target, or has any legal or employment requirement (explicit or implicit) that might affect their decisions regarding a target, they must recuse themselves from any approval decisions regarding the target's tier status, though they may otherwise participate in discussions.
This requirement does not prevent part or all of this policy from being cited in an explicit contract or work agreement (e.g. to implement or maintain support for a target). This requirement exists to ensure that a developer or team responsible for reviewing and approving a target does not face any legal threats or obligations that would prevent them from freely exercising their judgment in such approval, even if such judgment involves subjective matters or goes beyond the letter of these requirements.
Only relevant to those making approval decisions.
> Tier 3 targets should attempt to implement as much of the standard libraries as possible and appropriate (core for most targets, alloc for targets that can support dynamic memory allocation, std for targets with an operating system or equivalent layer of system-provided functionality), but may leave some code unimplemented (either unavailable or stubbed out as appropriate), whether because the target makes it impossible to implement or challenging to implement. The authors of pull requests are not obligated to avoid calling any portions of the standard library on the basis of a tier 3 target not implementing those portions.
`core` and `alloc` can be used. For `std` exists already the target `aarch64_unknown_hermit`, which enables FPU support.
> The target must provide documentation for the Rust community explaining how to build for the target, using cross-compilation if possible. If the target supports running tests (even if they do not pass), the documentation must explain how to run tests for the target, using emulation if possible or dedicated hardware if necessary.
Use `--target=aarch64_unknown_hermitkernel` option to cross compile. The target does currently not support running tests.
> Tier 3 targets must not impose burden on the authors of pull requests, or other developers in the community, to maintain the target. In particular, do not post comments (automated or manual) on a PR that derail or suggest a block on the PR based on a tier 3 target. Do not send automated messages or notifications (via any medium, including via `@)` to a PR author or others involved with a PR regarding a tier 3 target, unless they have opted into such messages.
Backlinks such as those generated by the issue/PR tracker when linking to an issue or PR are not considered a violation of this policy, within reason. However, such messages (even on a separate repository) must not generate notifications to anyone involved with a PR who has not requested such notifications.
I don't foresee this being a problem.
> Patches adding or updating tier 3 targets must not break any existing tier 2 or tier 1 target, and must not knowingly break another tier 3 target without approval of either the compiler team or the maintainers of the other tier 3 target.
In particular, this may come up when working on closely related targets, such as variations of the same architecture with different features. Avoid introducing unconditional uses of features that another variation of the target may not have; use conditional compilation or runtime detection, as appropriate, to let each target run code supported by that target.
No other targets should be affected by the pull request.
Ensure that queries only return Copy types.
This should pervent the perf footgun of returning a result with an expensive `Clone` impl (like a `Vec` of a hash map).
I went for the stupid solution of allocating on an arena everything that was not `Copy`. Some query results could be made Copy easily, but I did not really investigate.
debuginfo: Fix DW_AT_containing_type vtable debuginfo regression
This PR brings back the `DW_AT_containing_type` attribute for vtables after it has accidentally been removed in #89597.
It also implements a more accurate description of vtables. Instead of describing them as an array of void pointers, the compiler will now emit a struct type description with a field for each entry of the vtable.
r? ``@wesleywiser``
This PR should fix issue https://github.com/rust-lang/rust/issues/93164.
~~The PR is blocked on https://github.com/rust-lang/rust/pull/93154 because both of them modify the `codegen/debug-vtable.rs` test case.~~
Better suggestions when user tries to collect into an unsized `[_]`
1. Extend the predicate on `rustc_on_unimplemented` to support substitutions like note, label, etc (i.e. treat it as a `OnUnimplementedFormatString`) so we can have slightly more general `rustc_on_unimplemented` special-cases.
2. Add a `rustc_on_unimplemented` if we fail on `FromIterator<A> for [A]` which happens when we don't explicitly collect into a `vec<A>`, but then pass the return from a `.collect` call into something that takes a slice.
Fixes#91423
Refactor query system to maintain a global job id counter
This replaces the per-shard counters with a single global counter, simplifying
the JobId struct down to just a u64 and removing the need to pipe a DepKind
generic through a bunch of code. The performance implications on non-parallel
compilers are likely minimal (this switches to `Cell<u64>` as the backing
storage over a `u64`, but the latter was already inside a `RefCell` so it's not
really a significance divergence). On parallel compilers, the cost of a single
global u64 counter may be more significant: it adds a serialization point in
theory. On the other hand, we can imagine changing the counter to have a
thread-local component if it becomes worrisome or some similar structure.
The new design is sufficiently simpler that it warrants the potential for slight
changes down the line if/when we get parallel compilation to be more of a
default.
A u64 counter, instead of u32 (the old per-shard width), is chosen to avoid
possibly overflowing it and causing problems; it is effectively impossible that
we would overflow a u64 counter in this context.
Delete -Zquery-stats infrastructure
These statistics are computable from the self-profile data and/or ad-hoc collectable as needed, and in the meantime contribute to rustc bootstrap times -- locally, this PR shaves ~2.5% from rustc_query_impl builds in instruction counts.
If this does lose some functionality we want to keep, I think we should migrate it to self-profile (or a similar interface) rather than this ad-hoc reporting.
Store rlink data in opaque binary format on disk
This removes one of the only uses of JSON decoding (to Rust structs) from the compiler, and fixes the FIXME comment. It's not clear to me what the reason for using JSON here originally was, and from what I can tell nothing outside of rustc expects to read the emitted information, so it seems like a reasonable step to move it to the metadata-encoding format (rustc_serialize::opaque).
Mostly intended as a FIXME fix, though potentially a stepping stone to dropping the support for Decodable to be used to decode JSON entirely (allowing for better/faster APIs on the Decoder trait).
cc #64191
Improve opaque type higher-ranked region error message under NLL
Currently, any higher-ranked region errors involving opaque types
fall back to a generic "higher-ranked subtype error" message when
run under NLL. This PR adds better error message handling for this
case, giving us the same kinds of error messages that we currently
get without NLL:
```
error: implementation of `MyTrait` is not general enough
--> $DIR/opaque-hrtb.rs:12:13
|
LL | fn foo() -> impl for<'a> MyTrait<&'a str> {
| ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ implementation of `MyTrait` is not general enough
|
= note: `impl MyTrait<&'2 str>` must implement `MyTrait<&'1 str>`, for any lifetime `'1`...
= note: ...but it actually implements `MyTrait<&'2 str>`, for some specific lifetime `'2`
error: aborting due to previous error
```
To accomplish this, several different refactoring needed to be made:
* We now have a dedicated `InstantiateOpaqueType` struct which
implements `TypeOp`. This is used to invoke `instantiate_opaque_types`
during MIR type checking.
* `TypeOp` is refactored to pass around a `MirBorrowckCtxt`, which is
needed to report opaque type region errors.
* We no longer assume that all `TypeOp`s correspond to canonicalized
queries. This allows us to properly handle opaque type instantiation
(which does not occur in a query) as a `TypeOp`.
A new `ErrorInfo` associated type is used to determine what
additional information is used during higher-ranked region error
handling.
* The body of `try_extract_error_from_fulfill_cx`
has been moved out to a new function `try_extract_error_from_region_constraints`.
This allows us to re-use the same error reporting code between
canonicalized queries (which can extract region constraints directly
from a fresh `InferCtxt`) and opaque type handling (which needs to take
region constraints from the pre-existing `InferCtxt` that we use
throughout MIR borrow checking).
Drop tracking: track borrows of projections
Previous efforts to ignore partially consumed values meant we were also not considering borrows of a projection. This led to cases where we'd miss borrowed types which MIR expected to be there, leading to ICEs.
This PR also includes the `-Zdrop-tracking` flag from #93313. If that PR lands first, I'll rebase to drop the commit from this one.
Fixes#93648
rustc_query_impl: reduce visibility of some modules/fn's
Locally this reduces number of exported functions from 15221 -> 14952 and size a little.
Perf run please?
Remove defaultness from ImplItem.
This information is not really used anywhere, except HIR pretty-printing. This makes ImplItem and TraitItem more similar.
Rollup of 7 pull requests
Successful merges:
- #91950 (Point at type when a `static` `#[global_allocator]` doesn't `impl` `GlobalAlloc`)
- #92715 (Do not suggest char literal for zero-length strings)
- #92917 (Don't constrain projection predicates with inference vars in GAT substs)
- #93206 (Use `NtCreateFile` instead of `NtOpenFile` to open a file)
- #93732 (add fut/back compat tests for implied trait bounds)
- #93764 (⬆️ rust-analyzer)
- #93767 (deduplicate `lcnr` in mailmap)
Failed merges:
r? `@ghost`
`@rustbot` modify labels: rollup
This replaces the per-shard counters with a single global counter, simplifying
the JobId struct down to just a u64 and removing the need to pipe a DepKind
generic through a bunch of code. The performance implications on non-parallel
compilers are likely minimal (this switches to `Cell<u64>` as the backing
storage over a `u64`, but the latter was already inside a `RefCell` so it's not
really a significance divergence). On parallel compilers, the cost of a single
global u64 counter may be more significant: it adds a serialization point in
theory. On the other hand, we can imagine changing the counter to have a
thread-local component if it becomes worrisome or some similar structure.
The new design is sufficiently simpler that it warrants the potential for slight
changes down the line if/when we get parallel compilation to be more of a
default.
A u64 counter, instead of u32 (the old per-shard width), is chosen to avoid
possibly overflowing it and causing problems; it is effectively impossible that
we would overflow a u64 counter in this context.
Currently, any higher-ranked region errors involving opaque types
fall back to a generic "higher-ranked subtype error" message when
run under NLL. This PR adds better error message handling for this
case, giving us the same kinds of error messages that we currently
get without NLL:
```
error: implementation of `MyTrait` is not general enough
--> $DIR/opaque-hrtb.rs:12:13
|
LL | fn foo() -> impl for<'a> MyTrait<&'a str> {
| ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ implementation of `MyTrait` is not general enough
|
= note: `impl MyTrait<&'2 str>` must implement `MyTrait<&'1 str>`, for any lifetime `'1`...
= note: ...but it actually implements `MyTrait<&'2 str>`, for some specific lifetime `'2`
error: aborting due to previous error
```
To accomplish this, several different refactoring needed to be made:
* We now have a dedicated `InstantiateOpaqueType` struct which
implements `TypeOp`. This is used to invoke `instantiate_opaque_types`
during MIR type checking.
* `TypeOp` is refactored to pass around a `MirBorrowckCtxt`, which is
needed to report opaque type region errors.
* We no longer assume that all `TypeOp`s correspond to canonicalized
queries. This allows us to properly handle opaque type instantiation
(which does not occur in a query) as a `TypeOp`.
A new `ErrorInfo` associated type is used to determine what
additional information is used during higher-ranked region error
handling.
* The body of `try_extract_error_from_fulfill_cx`
has been moved out to a new function `try_extract_error_from_region_constraints`.
This allows us to re-use the same error reporting code between
canonicalized queries (which can extract region constraints directly
from a fresh `InferCtxt`) and opaque type handling (which needs to take
region constraints from the pre-existing `InferCtxt` that we use
throughout MIR borrow checking).
Don't constrain projection predicates with inference vars in GAT substs
cc #91762
Not a fix, but a mitigation to prevent a backwards-compatible hazard where we normalize using a predicate only because it's the only one available, but shouldn't. This would constrain an inference variable which didn't really want. We already do this when selecting a projection candidate, which isn't always correct. But changing that is a problem for a different day.
Also found out that a suggestion for `await`ing a future was using the wrong substs.
r? ``@nikomatsakis``
Do not suggest char literal for zero-length strings
PR #92507 adds a hint to switch to single quotes when a char is expected and a single-character string literal is provided.
The check to ensure the string literal is one character long missed the 0-char case, and would incorrectly offer the hint.
This PR adds the missing check, and a test case to confirm the new behavior.
Currently, we are thinking to use *-unknown-none targets instead
to define for every platform our own one (see hermitcore/rusty-hermit#197).
However, the current target aarch64-unknown-none-softfloat doesn't support
dynamic relocation. Our kernel uses this feature and consequently
we define a new target aarch64-unknown-hermitkernel to support it.
Add in ValuePair::Term
This adds in an enum when matching on positions which can either be types or consts.
It will default to emitting old special cased error messages for types.
r? `@oli-obk`
cc `@matthiaskrgr`
Fixes#93578
update comment wrt const param defaults
after #93669 i looked through all other uses of `GenericParamKind::Const` again to detect if we missed the `default` there as well, but afaict we really only missed lifetime resolution '^^ at least i found an outdated comment :3
Add more *-unwind ABI variants
The following *-unwind ABIs are now supported:
- "C-unwind"
- "cdecl-unwind"
- "stdcall-unwind"
- "fastcall-unwind"
- "vectorcall-unwind"
- "thiscall-unwind"
- "aapcs-unwind"
- "win64-unwind"
- "sysv64-unwind"
- "system-unwind"
cc `@rust-lang/wg-ffi-unwind`
This commit fixes two issues.
One, sometimes break or continue have a block target instead of an
expression target. This seems to mainly happen with try blocks. Since
the drop tracking analysis only works on expressions, if we see a block
target for break or continue, we substitute the last expression of the
block as the target instead.
Two, break and continue were incorrectly being treated as the same, so
continue would also show up as an exit from the loop or block. This
patch corrects the way continue is handled by keeping a stack of loop
entry points and uses those to find the target of the continue.
