rustc_resolve: overhaul `#![feature(uniform_paths)]` error reporting.
Fixes#53408 by only considering external crates to conflict within their (type/module) namespace, *not* with the value or macro namespaces, and also by adding a special-cased error for redundant `use crate_name;` imports (without actually allowing them).
Also, all canaries for a given import are grouped into one diagnostic per namespace, in order to make block-scoped ambiguities clearer.
See changed/added tests for more details.
r? @petrochenkov cc @aturon @joshtriplett
Visit all attributes for feature collection
Previously feature attributes were just collected on item-like "things" as well as exported macros and crate attributes. This ignored some places feature attributes could be specified, such as on enum variants.
Fixes https://github.com/rust-lang/rust/issues/53391.
This shim is generated elsewhere in the compiler so this commit adds support to
ensure it goes through similar paths as the rest of the compiler to set llvm
function attributes like target features.
cc #53372
Rollup of 10 pull requests
Successful merges:
- #52946 (Documented impl From on line 367 of libserialize/json.rs)
- #53234 (Remove Travis shutdown debug scripts, and remove CI-specific DNS settings)
- #53313 (Two small improvements)
- #53360 (Addressed #51602)
- #53364 (Warn if the user tries to use GATs)
- #53373 (Tweak unclosed delimiter parser error)
- #53377 (std: Use target_pointer_width for BACKTRACE_ELF_SIZE)
- #53395 (Use #[non_exhaustive] on internal enums)
- #53399 (Tidy: ignore non-Markdown files when linting for the Unstable Book)
- #53412 (syntax_ext: remove leftover span_err_if_not_stage0 macro.)
Originally added in #52887 this commit disables passing `--strip-debug` to LLD
when optimized. This bring back the original bug of emitting broken debuginfo
but currently it *also* strips the `name` section which makes it very difficult
to inspect the final binary. A real fix is happening at
https://reviews.llvm.org/D50729 and we can reevaluate once we've updated LLD to
have that commit.
Remove Travis shutdown debug scripts, and remove CI-specific DNS settings
Since the cause of the host shutdown (travis-ci/travis-ci#4924) is found, we could revert the shutdown debug attempts to shorten the logs.
Additionally, we're pretty sure a custom DNS (added in will not help travis-ci/travis-ci#9696, so reverting that part of #51420 to reduce CI-specific settings.
syntax_ext: remove leftover span_err_if_not_stage0 macro.
I believe this is the right fix for #53380, although I'm not sure what happened.
My guess is this copy of the macro was accidentally missed when others were removed?
cc @matthewjasper @varkor (please do not put this in a rollup, in case it fails)
Tidy: ignore non-Markdown files when linting for the Unstable Book
Previously, any file would be assumed a `.md`, which is annoying when files like `.DS_STORE` slip in and then cause tidy errors.
Use #[non_exhaustive] on internal enums
This replaces `__Nonexhaustive` variants (and variants thereof) with `#[non_exhaustive]`. These were mostly unstable previously, with the exception of the `cloudabi` enums and `Level` in proc_macro: these were `#[doc(hidden)]`, so clearly intended not to be used directly. It should be safe to replace all of these.
std: Use target_pointer_width for BACKTRACE_ELF_SIZE
The former code used `target.contains("64")` to detect Elf64 targets,
but this is inaccurate in a few cases:
- `s390x-unknown-linux-gnu` is 64-bit
- `sparcv9-sun-solaris` is 64-bit
- `x86_64-unknown-linux-gnux32` is 32-bit
Instead the `std` build script can use `CARGO_CFG_TARGET_POINTER_WIDTH`
to reliably detect 64-bit targets for libbacktrace.
Also update to backtrace-sys 0.1.24 for alexcrichton/backtrace-rs#122.
Addressed #51602Fixed#51602
r? @estebank
here I have addressed the case where `in` was not expected right after `if` block. Speaking of `type ascription` I am not sure if this the best approach which I have implemented. Plus I think one more test case can be added to test `type-ascription` case, though I don't have any at this point of time. I will ping you again if all existing testcases pass.
Two small improvements
In `librustc_apfloat/ieee.rs`, use the iterator.[r]find methods to simplify the code. In `libserialize/json.rs`, make use of the fact that `Vec.last` on an empty `Vec` returns `None` to simplify the code to a single match.
Documented impl From on line 367 of libserialize/json.rs
This is for the impl From mentioned in #51430 assigned to @skade .
Hopefully I didn't miss anything/get anything wrong. I looked over another PR for another part of this same issue to see what the proper formatting was, etc.
Thanks!
TokenStream::extend
Two new insta-stable impls in libproc_macro:
```rust
impl Extend<TokenTree> for TokenStream
impl Extend<TokenStream> for TokenStream
```
`proc_macro::TokenStream` already implements `FromIterator<TokenTree>` and `FromIterator<TokenStream>` so I elected to support the same input types for `Extend`.
**This commit reduces compile time of Serde derives by 60% (takes less than half as long to compile)** as measured by building our test suite:
```console
$ git clone https://github.com/serde-rs/serde
$ cd serde/test_suite
$ cargo check --tests --features proc-macro2/nightly
$ rm -f ../target/debug/deps/libtest_*.rmeta
$ time cargo check --tests --features proc-macro2/nightly
Before: 20.8 seconds
After: 8.6 seconds
```
r? @alexcrichton
Don't accept non-string literals for the format string in writeln
This is to improve diagnostics.
`println` and `eprintln` were already fixed by #52394.
Fixes#30143
Start adding an `aarch64-pc-windows-msvc` target
This commit adds the necessary definitions for target specs and such as well as
the necessary support in libstd to compile basic `aarch64-pc-windows-msvc`
binaries. The target is not currently built on CI, but it can be built locally
with:
./configure --target=aarch64-pc-windows-msvc --set rust.lld
./x.py build src/libstd --target aarch64-pc-windows-msvc
Currently this fails to build `libtest` due to a linker bug (seemingly in LLD?)
which hasn't been investigate yet. Otherwise though with libstd you can build a
hello world program (linked with LLD). I've not tried to execute it yet, but it
at least links!
Full support for this target is still a long road ahead, but this is hopefully a
good stepping stone to get started.
Points of note about this target are:
* Currently defaults to `panic=abort` as support is still landing in LLVM for
SEH on AArch64.
* Currently defaults to LLD as a linker as I was able to get farther with it
than I was with `link.exe`
For move errors, suggest match ergonomics instead of `ref`
Partially fixes issue #52423. Also makes errors and suggestions more consistent between move-from-place and move-from-value errors.
Limitations:
- Only the first pattern in a match arm can have a "consider removing this borrow operator" suggestion.
- Suggestions don't always compile as-is (see the TODOs in the test for details).
Sorry for the really long test. I wanted to make sure I handled every case I could think of, and it turned out there were a lot of them.
Questions:
- Is there any particular applicability I should set on those suggestions?
- Are the notes about the `Copy` trait excessive?
`HybridIdxSetBuf` is a sparse-when-small but dense-when-large index set
that is very efficient for sets that (a) have few elements, (b) have
large `universe_size` values, and (c) are cleared frequently. Which
makes it perfect for the `gen_set` and `kill_set` sets used by the new
borrow checker.
This patch reduces the execution time of the five slowest NLL benchmarks
by 55%, 21%, 16%, 10% and 9%. It also reduces the max-rss of three
benchmarks by 53%, 33%, and 9%.
A few cleanups and minor improvements for the lexer
- improve readability by adjusting the formatting of some function signatures and adding some newlines
- reorder some functions for easier reading
- remove redundant `'static` in `const`s
- remove some explicit `return`s
- read directly to a `String` in `gather_comments_and_literals`
- change `unwrap_or!` (macro) to `unwrap_or` (function)
- move an `assert!`ion from `try_next_token` (called in a loop) to `try_real_token` after all calls to `try_next_token`
- `#[inline]` some one-liner functions
- assign directly from an `if-else` expression
- refactor a `match` to `map_or`
- add a `token::is_irrelevant` function to detect tokens that are not "`real`"
This commit adds the necessary definitions for target specs and such as well as
the necessary support in libstd to compile basic `aarch64-pc-windows-msvc`
binaries. The target is not currently built on CI, but it can be built locally
with:
./configure --target=aarch64-pc-windows-msvc --set rust.lld
./x.py build src/libstd --target aarch64-pc-windows-msvc
Currently this fails to build `libtest` due to a linker bug (seemingly in LLD?)
which hasn't been investigate yet. Otherwise though with libstd you can build a
hello world program (linked with LLD). I've not tried to execute it yet, but it
at least links!
Full support for this target is still a long road ahead, but this is hopefully a
good stepping stone to get started.
Points of note about this target are:
* Currently defaults to `panic=abort` as support is still landing in LLVM for
SEH on AArch64.
* Currently defaults to LLD as a linker as I was able to get farther with it
than I was with `link.exe`
