Ok, it's hard to explain what happens, but identifier's hygienic contexts need to be "adjusted" to modules/scopes before they are resolved in them.
To be resolved in all kinds on preludes the identifier needs to be adjusted to the root expansion (aka "no expansion").
Previously this was done for the `macro m() { ::my_crate::foo }` case, but forgotten for all other cases.
Also move macro stability checking closer to other checks performed on obtained resolutions.
Tighten the stability spans as well, it is an error to *refer* to and unstable entity in any way, not only "call" it.
This way we are processing all of them in a single point, rather than separately for each syntax extension kind.
Also, the standard expected/found wording is used.
It either returns the indeterminacy error, or valid (but perhaps dummy) `SyntaxExtension`.
With this change enum `Determinacy` is no longer used in libsyntax and can be moved to resolve.
The regressions in diagnosics are fixed in the next commits.
Create real working and registered (even if dummy) `SyntaxExtension`s for them.
This improves error recovery and allows to avoid all special cases for proc macro stubs (except for the error on use, of course).
The introduced dummy `SyntaxExtension`s can be used for any other inappropriately resolved macros as well.
We have to deal with dummy spans anyway
Remove def-site span from expander interfaces.
It's not used by the expansion infra, only by specific expanders, which can keep it themselves if they want it.
It's internal to resolve and always results in `Res::Err` outside of resolve.
Instead put `DefKind::Fn`s themselves into the macro namespace, it's ok.
Proc macro stubs are items placed into macro namespase for functions that define proc macros.
https://github.com/rust-lang/rust/pull/52383
The rustdoc test is changed because the old test didn't actually reproduce the ICE it was supposed to reproduce.
use visitor for #[structural_match] check
This changes the code so that we recur down the structure of a type of a const (rather than just inspecting at a shallow one or two levels) when we are looking to see if it has an ADT that did not derive `PartialEq` and `Eq`.
Fix#61188Fix#62307
Cc #62336
Prepare for LLVM 9 update
Main changes:
* In preparation for opaque pointer types, the `byval` attribute now takes a type. As such, the argument type needs to be threaded through to the function/callsite attribute application logic.
* On ARM the `+fp-only-sp` and `+d16` features have become `-fp64` and `-d32`. I've switched the target definitions to use the new names, but also added bidirectional emulation so either can be used on any LLVM version for backwards compatibility.
* The datalayout can now specify function pointer alignment. In particular on ARM `Fi8` is specified, which means that function pointer alignment is independent of function alignment. I've added this to our datalayouts to match LLVM (which is something we check) and strip the fnptr alignment for older LLVM versions.
* The fmul/fadd reductions now always respect the accumulator (including for unordered reductions), so we should pass the identity instead of undef.
Open issues:
* https://reviews.llvm.org/D62106 causes linker errors with ld.bdf due to https://sourceware.org/bugzilla/show_bug.cgi?id=24784. To avoid this I've enabled `RelaxELFRelocations`, which results in a GOTPCRELX relocation for `__tls_get_addr` and avoids the issue. However, this is likely not acceptable because relax relocations are not supported by older linker versions. We may need an LLVM option to keep using PLT for `__tls_get_addr` despite `RtLibUseGOT`.
The corresponding llvm-project PR is https://github.com/rust-lang/llvm-project/pull/19.
r? @ghost
Rollup of 9 pull requests
Successful merges:
- #62417 (Fix ICEs when `Self` is used in type aliases)
- #62450 (Raise the default recursion limit to 128)
- #62470 (Prevent shrinking of "crate select" element on Firefox)
- #62515 (cli: make help output for -l and -L consistent)
- #62520 (Regression test for issue 42574.)
- #62526 (normalize use of backticks in compiler messages for libsyntax/feature_gate.rs)
- #62527 (clarify that debug_assert does not completely omits the code)
- #62535 (ci: Configure $CI_JOB_NAME correctly)
- #62541 (Add spastorino for rustc-guide toolstate)
Failed merges:
r? @ghost
Regression test for issue 42574.
Cc #42574.
I'm not going to say this *closes* that issue yet, for two reasons:
1. I am still confused about some aspects of the behavior we are observing that bug
2. The "fix" to the diagnostic relies on full NLL (`#![feature(nll)]`); migration mode still has a subpar diagnostic.
Raise the default recursion limit to 128
The previous limit of 64 is being (just) barely hit by genuine code out there, which is causing issues like https://github.com/rust-lang/rust/issues/62059 to rear their end.
Ideally, we wouldn’t have such arbitrary limits at all, but while we do, it makes a lot of sense to just raise this limit whenever genuine use-cases end up hitting it.
r? @pnkfelix
Fixes https://github.com/rust-lang/rust/issues/62059
Fix ICEs when `Self` is used in type aliases
I think it is right just to disallow this at resolution stage rather than let typeck produce a cyclic error. This is in line with previous behaviour. There was probably no need at all for the change that introduced this bug in #57428, so I've simply reversed it.
Fixes#62263, #62364, #62305.
r? @eddyb
Normalize projections appearing in `impl Trait`
Fixes#60414
This does not try to do the same for `existential type`s (which have the same bug), since that always seems to lead to cycle errors.
Exit arm scopes
Due to a bug in the HIR CFG construction, borrows for arm scopes were incorrectly leaking into other arms.
This PR also includes some drive-by improvements to `-Zunpretty=hir,identified` that would have been helpful while investigating this.
Closes#62107
- uses a never-stable core::array::LengthAtMost32 to bound the impls
- includes a custom error message to avoid mentioning LengthAtMost32 too often
- doesn't use macros for the slice implementations to avoid #62433
rustdoc: set cfg(doctest) when collecting doctests
Note: This PR builds on top of https://github.com/rust-lang/rust/pull/61199; only the last commit is specific to this PR.
As discussed in https://github.com/rust-lang/rust/pull/61199, we want the ability to isolate items to only when rustdoc is collecting doctests, but we can't use `cfg(test)` because of libcore's `#![cfg(not(test))]`. This PR proposes a new cfg flag, `cfg(doctest)`, specific to this situation, rather than reusing an existing flag. I've isolated it behind a feature gate so that we can contain the effects to nightly only. (A stable workaround that can be used in lieu of `#[cfg(doctest)]` is `#[cfg(rustdoc)] #[doc(hidden)]`, at least once https://github.com/rust-lang/rust/pull/61351 lands.)
Tracking issue: https://github.com/rust-lang/rust/issues/62210
`non_ascii_idents` lint (part of RFC 2457)
RFC 2457 [declares](121bbeff50/text/2457-non-ascii-idents.md): "A `non_ascii_idents` lint is added to the compiler. This lint is allow by default."
(Part of #55467.)
r? @Manishearth
name the trait in ambiguous-associated-items fully qualified suggestion
We have the trait at this point, so we can name it in the error message, rather than using "Trait" as a (potentially confusing) placeholder.
Thanks to Yuki "@JohnTitor" Okushi for pointing out where to look (in the same file) for a closely related issue for ambiguous associated types (as opposed to items; that was #59225, except that one won't be
quite as easy to resolve, because we actually don't have the trait `DefId` at that point).
r? @petrochenkov
Now that procedural macros no longer link transitively to libsyntax,
this shouldn't be needed any more! This commit is an experiment in
removing all dynamic libraries from rustc except for librustc_driver
itself. Let's see how far we can get with that!
