rustdoc: collect trait impls as an early pass
Fixes https://github.com/rust-lang/rust/issues/52545, fixes https://github.com/rust-lang/rust/issues/41480, fixes https://github.com/rust-lang/rust/issues/36922
Right now, rustdoc pulls all its impl information by scanning a crate's HIR for any items it finds. However, it doesn't recurse into anything other than modules, preventing it from seeing trait impls that may be inside things like functions or consts. Thanks to https://github.com/rust-lang/rust/pull/53002, now these items actually *exist* for rustdoc to see, but they still weren't getting collected for display.
But there was a secret. Whenever we pull in an item from another crate, we don't have any of its impls in the local HIR, so instead we ask the compiler for *everything* and filter out after the fact. This process is only triggered if there's a cross-crate re-export in the crate being documented, which can sometimes leave this info out of the docs. This PR instead moves this collection into an early pass, which occurs immediately after crate cleaning, so that that collection occurs regardless. In addition, by including the HIR's own `trait_impls` in addition to the existing `all_trait_implementations` calls, we can collect all these tricky trait impls without having to scan for them!
A few cleanups and minor improvements to rustc/traits
It's a little bigger than usual, so bear with me ^^:
- introduce `TyCtxt::all_impls` and use it to avoid inefficiently allocating push loops
- modify `ArgKind::from_expected_ty` to take an `Option<Span>` argument to make it more versatile
- replace `ArgKind::Arg("_".to_owned(), "_".to_owned())` with `ArgKind::empty`
- move early `return`s earlier where possible
- if all branches of a `match` end with the same expression, move it after it
- change a hacky `match` expression to an `if else` chain
- move the `push` out from a push loop closure to reduce the number of allocations
- correct the vector size for `pretty_predicates` (under `specialize`)
- take advantage of the fact that `if else` is an expression
- prefer `cloned()` to `map(|&x| x)` and `map(|x| *x)`
- prefer `vec![x; y.len()]` to `y.map(|_| x).collect()`
- use `unwrap_or_else` instead of `match` where applicable
- use `if let` instead of `match` when only one branch matters
- prefer `to_owned` to `to_string` for string literals
- remove explicit `return`s
- remove superfluous braces
- whitespace fixes
- several other minor improvements
Switch linker for aarch64-pc-windows-msvc from LLD to MSVC
The MSVC linker does not seem to have the same problems with Rust symbols that LLD currently has on Windows (see https://github.com/rust-lang/rust/issues/54190#issuecomment-421288625). This PR makes MSVC the default linker for `aarch64-pc-windows-msvc`.
r? @alexcrichton
Enable fatal warnings for the wasm32 linker
Historically LLD has emitted warnings for various reasons but all the bugs have
since been fixed (yay!) and by enabling fatal warnings we should be able to head
off bugs like #53390 sooner.
Switch wasm math symbols to their original names
The names `Math_*` were given to help undefined symbol messages indicate how to
implement them, but these are all implemented in compiler-rt now so there's no
need to rename them! This change should make it so wasm binaries by default, no
matter the math symbols used, will not have unresolved symbols.
Warn about metadata loader errors
Output when writing corrupting to libcore.rlib
```
warning: no metadata found: failed to read rlib metadata in '/Users/bjorn/Documents/rust_fork/build/x86_64-apple-darwin/stage1-std/x86_64-apple-darwin/release/deps/libcore-857d662d379c5d0c.rlib': File too small to be an archive
error[E0463]: can't find crate for `core`
error: aborting due to previous error
```
Fixes#53381
Duration div mul extras
Successor of #52556.
This PR adds the following `impl`s:
- `impl Mul<Duration> for u32` (to allow `10*SECOND` in addition to `SECOND*10`)
- `impl Mul<f64> for Duration` (to allow `2.5*SECOND` vs `2*SECOND + 500*MILLISECOND`)
- `impl Mul<Duration> for f64`
- `impl MulAssign<f64> for Duration`
- `impl Div<f64> for Duration`
- `impl DivAssign<f64> for Duration`
- `impl Div<Duration> for Duration` (`Output = f64`, can be useful e.g. for `duration/MINUTE`)
`f64` is chosen over `f32` to minimize rounding errors. (52 bits fraction precision vs `Duration`'s ~94 bit)
constraints:
- clean/inline.rs needs this map to fill in traits when inlining
- fold.rs needs this map to allow passes to fold trait items
- html/render.rs needs this map to seed the Cache.traits map of all
known traits
The first two are the real problem, since `DocFolder` only operates on
`clean::Crate` but `clean/inline.rs` only sees the `DocContext`. The
introduction of early passes means that these two now exist at the same
time, so they need to share ownership of the map. Even better, the use
of `Crate` in a rustc thread pool means that it needs to be Sync, so it
can't use `Lrc<Lock>` to manually activate thread-safety.
`parking_lot` is reused from elsewhere in the tree to allow use of its
`ReentrantMutex`, as the relevant parts of rustdoc are still
single-threaded and this allows for easier use in that context.
Remove usages of span_suggestion without Applicability
Use `Applicability::Unspecified` for all of them instead.
Shall deprecations for the non-`_with_applicability` functions be added?
Shall clippy be addressed somehow?
r? @estebank
This is a safe wrapper around ptr::copy, for regions within a single
slice. Previously, safe in-place copying was only available as a side
effect of Vec::drain.
Update some `*-sys` dependencies of Cargo/RLS
This is intended to help solve #54206 on nightly where the RLS on MinGW is
having build issues with accidentally building a `curl` library which links to
pthread symbols on Windows (where it should use native mutex locking instead).
The build system for these `*-sys` crates have all been rewritten to be based on
`cc` to bypass native build systems and platform detection to make sure we
configure them correctly.
This is intended to help solve #54206 on nightly where the RLS on MinGW is
having build issues with accidentally building a `curl` library which links to
pthread symbols on Windows (where it should use native mutex locking instead).
The build system for these `*-sys` crates have all been rewritten to be based on
`cc` to bypass native build systems and platform detection to make sure we
configure them correctly.
Split `Liveness::users` into three.
This reduces memory usage on some benchmarks because no space is wasted
for padding. For a `check-clean` build of `keccak` it reduces `max-rss`
by 20%.
r? @nikomatsakis, but I want to do a perf run. Locally, I had these results:
- instructions: slight regression
- max-rss: big win on "Clean" builds
- faults: big win on "Clean" and "Nll" builds
- wall-time: small win on "Clean" and "Nll" builds
So I want to see how a different machine compares.
Currently, `BitSet` doesn't actually know its own domain size; it just
knows how many words it contains. To improve things, this commit makes
the following changes.
- It changes `BitSet` and `SparseBitSet` to store their own domain size,
and do more precise bounds and same-size checks with it. It also
changes the signature of `BitSet::to_string()` (and puts it within
`impl ToString`) now that the domain size need not be passed in from
outside.
- It uses `derive(RustcDecodable, RustcEncodable)` for `BitSet`. This
required adding code to handle `PhantomData` in `libserialize`.
- As a result, it removes the domain size from `HybridBitSet`, making a
lot of that code nicer.
- Both set_up_to() and clear_above() were overly general, working with
arbitrary sizes when they are only needed for the domain size. The
commit removes the former, degeneralizes the latter, and removes the
(overly general) tests.
- Changes `GrowableBitSet::grow()` to `ensure()`, fixing a bug where a
(1-based) domain size was confused with a (0-based) element index.
- Changes `BitMatrix` to store its row count, and do more precise bounds
checks with it.
- Changes `ty_params` in `select.rs` from a `BitSet` to a
`GrowableBitSet` because it repeatedly failed the new, more precise
bounds checks. (Changing the type was simpler than computing an
accurate domain size.)
- Various other minor improvements.
