There are two big categories of changes in here
- Removing lifetimes from common traits that can essentially never user a lifetime from an input (particularly `Drop` & `Debug`)
- Forwarding impls that are only possible because the lifetime doesn't matter (like `impl<R: Read + ?Sized> Read for &mut R`)
I omitted things that seemed like they could be more controversial, like the handful of iterators that have a `Item: 'static` despite the iterator having a lifetime or the `PartialEq` implementations where the flipped one cannot elide the lifetime.
Fix span for closure return type when annotated.
Fixes#58053.
This PR adjusts the span used to label closure return types so that
if the user specifies the return type, i.e. `|_| -> X {}` instead of
`|_| {}`, we correctly highlight all of it and not just the last
character.
r? @pnkfelix
Rollup of 16 pull requests
Successful merges:
- #57259 (Update reference of rlibc crate to compiler-builtins crate)
- #57740 (Use `to_ne_bytes` for converting IPv4Addr to octets)
- #57926 (Tiny expansion to docs for `core::convert`.)
- #58157 (Add Cargo.lock automatically adding message)
- #58203 (rustdoc: display sugared return types for async functions)
- #58243 (Add trait alias support in rustdoc)
- #58262 (Add #[must_use] message to Fn* traits)
- #58295 (std::sys::unix::stdio: explain why we do into_raw)
- #58297 (Cleanup JS a bit)
- #58317 (Some writing improvement, conciseness of intro)
- #58324 (miri: give non-generic functions a stable address)
- #58332 (operand-to-place copies should never be overlapping)
- #58345 (When there are multiple filenames, print what got interpreted as filenames)
- #58346 (rpath computation: explain why we pop())
- #58350 (Fix failing tidy (line endings on Windows))
- #58352 (miri value visitor: use `?` in macro)
Failed merges:
r? @ghost
When there are multiple filenames, print what got interpreted as filenames
I have written code that crafts command lines for rustc, and when I get "multiple input filenames provided" it can be quite hard to figure out where in this long list of arguments the mistake is hiding. Probably I passed an argument to a flag that does not expect an argument, but which flag would that be?
This changes the error message to print the first two filenames, to make it easier to debug what is going on.
Tiny expansion to docs for `core::convert`.
This is not really significant, accept or reject as you wish. I just want to make sure I understand how the PR process works and that I'm doing it right before doing a bigger one for #33417.
Make -Zdump-mir dump shims
Fixes https://github.com/rust-lang/rust/issues/53532 by (a) making the MIR shim generation use the MIR pass infrastructure, and (b) fixing said infrastructure to handle the fallout.
Cc @eddyb @oli-obk
Add a query type which is always marked as red if it runs
This is useful for queries which produce results which are very likely to change if their inputs do.
I also expect this to be useful for end to end queries because 1) we don't need `HashStable` impls and 2) we avoid the overhead of hashing the result of large results like the AST or the HIR map.
r? @michaelwoerister
Factor out error reporting from `smart_resolve_path_fragment` fn
This function was ridiculously monolithic before. We now have three rather-less-monolithic-and-horrifying functions.
r? @centril
Deduplicate mismatched delimiter errors
Delay unmatched delimiter errors until after the parser has run to deduplicate them when parsing and attempt recovering intelligently.
Second attempt at #54029, follow up to #53949. Fix#31528.
Avoid committing to autoderef in object method probing
This fixes the "leak" introduced in #57835 (see test for details, also apparently #54252 had no tests for the "leaks" that were fixed in it, so go ahead and add one).
Maybe beta-nominating because regression, but I'm against landing things on beta we don't have to.
r? @nikomatsakis
Make `intern_lazy_const` actually intern its argument.
Currently it just unconditionally allocates it in the arena.
For a "Clean Check" build of the the `packed-simd` benchmark, this
change reduces both the `max-rss` and `faults` counts by 59%; it
slightly (~3%) increases the instruction counts but the `wall-time` is
unchanged.
For the same builds of a few other benchmarks, `max-rss` and `faults`
drop by 1--5%, but instruction counts and `wall-time` changes are in the
noise.
Fixes#57432, fixes#57829.
