Rework treatment of `$crate` in procedural macros
Important clarification: `$crate` below means "processed `$crate`" or "output `$crate`". In the input of a decl macro `$crate` is just two separate tokens, but in the *output of a decl macro* `$crate` is a single keyword identifier (https://github.com/rust-lang/rust/issues/55640#issuecomment-435692791).
First of all, this PR removes the `eliminate_crate_var` hack.
`$crate::foo` is no longer replaced with `::foo` or `::crate_name::foo` in the input of derive proc macros, it's passed to the macro instead with its precise span and hygiene data, and can be treated as any other path segment keyword (like `crate` or `self`) after that. (Note: `eliminate_crate_var` was never used for non-derive proc macros.)
This creates an annoying problem - derive macros still may stringify their input before processing and expect `$crate` survive that stringification and refer to the same crate (the Rust 1.15-1.29 way of doing things).
Moreover, the input of proc macro attributes and derives (but not fn-like proc macros) also effectively survives stringification before being passed to the macro (also for legacy implementation reasons).
So we kind of resurrect the `eliminate_crate_var` hack in reduced form, but apply it only to AST pretty-printing.
If an AST fragment is pretty-printed, the resulting *text* will have `$crate` replaced with `crate` or `::crate_name`. This should be enough to keep all the legacy cases working.
Closes https://github.com/rust-lang/rust/issues/55640
Closes https://github.com/rust-lang/rust/issues/56622
r? @ghost
trigger unsized coercions keyed on Sized bounds
This PR causes unsized coercions to not be disabled by `$0: Unsize<dyn
Object>` coercion obligations when we have an `$0: Sized` obligation somewhere.
This should be mostly backwards-compatible, because in these cases not doing the unsize coercion should have caused the `$0: Sized` obligation to fail.
Note that `X: Unsize<dyn Object>` obligations can't fail *as obligations* if `X: Sized` holds, so this still maintains some version of monotonicity (I think that an unsized coercion can't be converted to no coercion by unifying type variables).
Fixes#49593 (unblocking never_type).
r? @eddyb
cc @nikomatsakis
Rollup of 15 pull requests
Successful merges:
- #56363 (Defactored Bytes::read)
- #56663 (Remove lifetime from Resolver)
- #56689 (add a lint group for lints emitted by rustdoc)
- #56772 (fix issue 54153 by not testing issue-18804 on Windows nor OS X.)
- #56820 (format-related tweaks)
- #56881 (Implement Eq, PartialEq and Hash for atomic::Ordering)
- #56907 (Fix grammar in compiler error for array iterators)
- #56908 (rustc: Don't ICE on usage of two new target features)
- #56910 (Do not point at delim spans for complete correct blocks)
- #56913 (Enable stack probes for UEFI images)
- #56918 (Profiler: simplify total_duration, improve readability)
- #56931 (Update release notes for Rust 1.31.1)
- #56947 (Add targets thumbv7neon-linux-androideabi and thumbv7neon-unknown-linux-gnueabihf)
- #56948 (Update LLVM submodule)
- #56959 (Fix mobile menu rendering collision with tooltip.)
Failed merges:
- #56914 (Ignore ui/target-feature-gate on sparc, sparc64, powerpc, powerpc64 and powerpc64le)
r? @ghost
Add targets thumbv7neon-linux-androideabi and thumbv7neon-unknown-linux-gnueabihf
These two targets enable both thumb-mode and NEON for ARMv7 CPUs.
This another attempt at #49902, which cannot be reopened. Between that PR and this one, some subrepos with C code whose build systems were failing went away.
Enable stack probes for UEFI images
When building UEFI images, we don't link to any CRT libraries so we need to provide a stack probe. Without `__rust_probestack`, the linker looks for `__chkstk` and fails to link if there is a function with large local variables.
r? @alexcrichton
Fix grammar in compiler error for array iterators
This fixes a small grammatical mistake in the message the compiler gives when attempting to iterate directly over an array `arr` without calling `arr.iter()` or borrowing `&arr`.
add a lint group for lints emitted by rustdoc
As rustdoc adds more lints that it specifically manages, it would be nice to be able to lump them all together. This gives us a new group just for that.
I deliberately didn't include `missing_docs` because this is kind of a stepping stone for moving our lints into tool lints (i.e. `#![warn(rustdoc::private_doc_tests)]`), since all of these are specifically emitted by rustdoc. If we want to move `missing_docs` out of the compiler, that's also an option, but it would create a surprising change of behavior.
I also took the chance to rewrite the lint descriptions of these lints to better match the style of the other lints. `>_>`
Defactored Bytes::read
Removed unneeded refactoring of read_one_byte, which removed the unneeded dynamic dispatch (`dyn Read`) used by that function.
This function is only used in one place in the entire Rust codebase; there doesn't seem to be a reason for it to exist (and there especially doesn't seem to be a reason for it to use dynamic dispatch)
Short-circuit Rc/Arc equality checking on equal pointers where T: Eq
based on #42965
Is the use of the private trait ok this way? Is there anything else needed for this to get pulled?
Search other library paths when loking for link objects
Support the case when link objects are not located in Rust sysroot
but in other locations which could be specify through library paths.
fix trait objects with a Self-containing projection values
Fixes#56288.
This follows ALT2 in the issue.
beta-nominating since this is a regression.
r? @nikomatsakis
Fix various aspects around `let` bindings inside const functions
* forbid `let` bindings in const contexts that use short circuiting operators
* harden analysis code against derefs of mutable references
Initially this PR was about stabilizing `let` bindings, but too many flaws were exposed that need some more testing on nightly
