Lower let-else in MIR
This MR will switch to lower let-else statements in MIR building instead.
To lower let-else in MIR, we build a mini-switch two branches. One branch leads to the matching case, and the other leads to the `else` block. This arrangement will allow temporary lifetime analysis running as-is so that the temporaries are properly extended according to the same rule applied to regular `let` statements.
cc https://github.com/rust-lang/rust/issues/87335Fix#98672
Rollup of 10 pull requests
Successful merges:
- #98789 (rustdoc-json-types: Clean up derives.)
- #98848 (Build the Clippy book as part of x.py doc)
- #99020 (check non_exhaustive attr and private fields for transparent types)
- #99132 (Add some autolabels for A-bootstrap and T-infra)
- #99148 (Clarify that [iu]size bounds were only defined for the target arch)
- #99152 (Use CSS variables to handle theming (part 2))
- #99168 (Add regression test for #74713)
- #99176 (⬆️ rust-analyzer)
- #99183 (Mention rust-analyzer maintainers when `proc_macro` bridge is changed)
- #99185 (llvm-wrapper: adapt for LLVM API change)
Failed merges:
r? `@ghost`
`@rustbot` modify labels: rollup
Build the Clippy book as part of x.py doc
r? ``@ehuss`` since you said you would be interested in helping moving this forward.
cc ``@jyn514`` as part of the bootstrap team.
rustdoc-json-types: Clean up derives.
Closes#96189
Everything is `Clone, Debug, PartialEq, Eq, Hash, Serialize, Deserialize` except `Crate` and `Item` which arn't `Hash`, as they have `HashMap`'s. See linked issue for reasoning.
``@rustbot`` modify labels: +T-rustdoc +A-rustdoc-json
Remove most box syntax from librustdoc
This is the second attempt after the librustdoc specific changes have been reverted from #87781 in #89134, due to a minor, but exant regression caused by the changes. ~~There have been some changes to librustdoc in the past and maybe thanks to them there is no regression any more. If there is still a regression, one can investigate further and maybe find ways to fix the regressions. Thus, i request a perf run.~~ Edit: turns out there is still a regression, but it's caused only by a subset of the changes. So I've changed this PR to only contains the changes that don't cause any performance regressions, keeping the regression causing changes for a later PR.
Configure nightly branch name in `stage0.json`
The beta version number detection code relies on git to know how many merge commits were made since we branched off, and in doing so hardcodes `master` as the default branch name. This works for rust-lang/rust, but is problematic for forks that use a different default branch name (in Ferrocene we use `main` instead).
This PR changes the code to instead load the default branch name from `src/stage0.json`. `bump-stage0` has also been updated to remove the need to update it every time a new field is added to `stage0.json`.
Put back UI test regex
I just realized I overwrote these two commits in https://github.com/rust-lang/rust/pull/99055 when force pushing to fix the stdout output...
r? `@Dylan-DPC`
Fix last `let_chains` blocker
In order to forbid things like `let x = (let y = 1);` or `if let a = 1 && { let x = let y = 1; } {}`, the parser **HAS** to know the context of `let`.
This context thing is not a surprise in the parser because you can see **a lot** of ad hoc fixes mixing parsing logic with validation logic creating code that looks more like spaghetti with tomato sauce.
To make things even greater, a new ad hoc fix was added to only allow `let`s in a valid `let_chains` context by checking the previously processed token. This was the only solution I could think of and believe me, I thought about it for a long time 👍
In the long term, it should be preferable to segregate different responsibilities or create a more robust and cleaner parser framework.
cc https://github.com/rust-lang/rust/pull/94927
cc https://github.com/rust-lang/rust/issues/53667
Rollup of 8 pull requests
Successful merges:
- #97210 (Support `-A`, `-W`, `-D` and `-F` when running `./x.py clippy`)
- #99055 (Fix rustdoc help options)
- #99075 (Fix duplicated type annotation suggestion)
- #99124 (Fix sized check ICE in asm check)
- #99142 (fix(doctest): treat fatal parse errors as incomplete attributes)
- #99145 (Don't rerun the build script for the compiler each time on non-windows platforms)
- #99146 (Do not error during method probe on `Sized` predicates for types that aren't the method receiver)
- #99161 (compiletest: trim edition before passing as flag)
Failed merges:
r? `@ghost`
`@rustbot` modify labels: rollup
Do not error during method probe on `Sized` predicates for types that aren't the method receiver
Fixes#61525
This is safe even though we're skipping an error because we end up confirming the method, which means we're still checking the `Sized` predicate in the end. It just means that we don't emit an erroneous message as below:
```
error: the `query` method cannot be invoked on a trait object
--> src/lib.rs:14:11
|
14 | 1.query::<dyn ToString>("")
| ^^^^^
|
= note: another candidate was found in the following trait, perhaps add a `use` for it:
`use crate::Example;`
```
Also fixes erroneously suggesting the same trait over again, as seen in the `issue-35976.rs` UI test.
Fix sized check ICE in asm check
Fixes (beta nominated, so doesn't close) #99122
1. Moves a check for unresolved inference variables to _before_ other checks that could possibly ICE. We're not changing behavior here, just doing the same thing earlier in the function.
2. Erases region variables in sized check (which are not resolved at this point) because rustc will also ICE when region vars are passed to a query which does not canonicalize them.
Fix duplicated type annotation suggestion
Before, there was more or less duplicated suggestions to add type hints.
Fix by clearing more generic suggestions when a more specific suggestion
is possible.
This fixes#93506 .
Fix rustdoc help options
Fixes#98976.
Since you're the one who found out about the problem and also provided the solution (thanks for both!):
r? ```@jyn514```
Support `-A`, `-W`, `-D` and `-F` when running `./x.py clippy`
Resolves#97059
This PR adds support for `-A`, `-W`, `-D` and `-F` when running `./x.py clippy`.
Create fresh lifetime parameters for bare fn trait too
The current code fails to account for the equivalence between `dyn FnMut(&mut u8)` and bare `FnMut(&mut u8)`, and treated them differently.
This PR introduces a special case for `Fn` traits, which are always fully resolved.
Fixes#98616Fixes#98726
This will require a beta-backport, as beta contains that bug.
r? `@petrochenkov`
This extracts the linux-isms into variables, so that the script can be
extended to do PGO on windows. These variables will be overriden in a
few spots, in windows-specific blocks.
When building LLVM/LLD as part of a build that asks LLVM to generate profiles, e.g. when
doing PGO, cmake or clang-cl don't automatically link clang's profiler runtime in,
causing undefined reference errors at link-time.
We do that manually, by adding clang's resource library folder to the library search path:
- for LLVM itself, by extending the linker args that `rustc_llvm`'s build script
uses, to avoid the linker errors when linking `rustc_driver`.
- for LLD, by extending cmake's linker flags during the LLD build step.
Mention similarly named associated type even if it's not clearly in supertrait
Due to query cycle avoidance, we sometimes restrict the candidates in `complain_about_assoc_type_not_found` too much so that we can't detect typo replacements from just supertraits.
This creates a more general note of the existence of a similarly named associated type from _all_ visible traits when possible.
Fixes#55673
explain doc comments in macros a bit
Open to suggestions on improving this... macro parsing is very foreign to me.
Should we have a structured suggestion to turn them into their regular non-doc comments?
Fixes#92846Fixes#97850
Currently, for the enums and comparison traits we always check the tag
for equality before doing anything else. This is a bit clumsy. This
commit changes things so that the tags are handled very much like a
zeroth field in the enum.
For `eq`/ne` this makes the code slightly cleaner.
For `partial_cmp` and `cmp` it's a more notable change: in the case
where the tags aren't equal, instead of having a tag equality check
followed by a tag comparison, it just does a single tag comparison.
The commit also improves how `Hash` works for enums: instead of having
duplicated code to hash the tag for every arm within the match, we do
it just once before the match.
All this required replacing the `EnumNonMatchingCollapsed` value with a
new `EnumTag` value.
For fieldless enums the new code is particularly improved. All the code
now produced is close to optimal, being very similar to what you'd write
by hand.
Remove unsupported options in configure.py
I've seen people using `optimize = false` and `full-bootstrap = true` in the past, without knowing
that they're not recommended. Remove `optimize` and a few other options that are always a bad idea,
and document that full-bootstrap is only for testing reproducible builds.
Use `tag` in names of things referring to tags, instead of the
mysterious `vi`.
Also change `arg_N` in output to `argN`, which has the same length as
`self` and so results in nicer vertical alignments.
By producing `&T` expressions for fields instead of `T`. This matches
what the existing comments (e.g. on `FieldInfo`) claim is happening, and
it's also what most of the trait-specific code needs.
The exception is `PartialEq`, which needs `T` expressions for lots of
special case error messaging to work. So we now convert the `&T` back to
a `T` for `PartialEq`.
E.g. improving code like this:
```
match &*self {
&Enum1::Single { x: ref __self_0 } => {
::core:#️⃣:Hash::hash(&*__self_0, state)
}
}
```
to this:
```
match self {
Enum1::Single { x: __self_0 } => {
::core:#️⃣:Hash::hash(&*__self_0, state)
}
}
```
by removing the `&*`, the `&`, and the `ref`.
I suspect the current generated code predates deref-coercion.
The commit also gets rid of `use_temporaries`, instead passing around
`always_copy`, which makes things a little clearer. And it fixes up some
comments.
Fix `x build library/std compiler/rustc`
Previously, this was broken because of improper caching:
1. `StepDescription::maybe_run` builds `Compile::Std`, which only built `std` and not `proc_macro`
1. `Std` calls `builder.ensure(StdLink)`
1. `Rustc` calls `ensure(Std)`, which builds all crates, including `proc_macro`
1. `Rustc` calls `ensure(StdLink)`. `ensure` would see that it had already been run and do nothing. <-- bug is here
1. Cargo gives an error that `proc_macro` doesn't exist.
This fixes the caching by adding `crates` to `StdLink`, so it will get rerun if the crates that are built change.
Fixes https://github.com/rust-lang/rust/issues/99129.
Before:
```
thread 'main' panicked at 'Unable to build RLS', dist.rs:42:9
```
After:
```
thread 'main' panicked at 'Unable to build submodule tool RLS (use `missing-tools = true` to ignore this failure)
note: not all tools are available on all nightlies
help: see https://forge.rust-lang.org/infra/toolstate.html for more information', dist.rs:43:9
```
Allow using `download-ci-llvm = true` outside the git checkout
`@bjorn3` noticed that this is already allowed today when download-llvm is disabled, but breaks with it enabled:
```
$ ./rust2/x.py build
fatal: not a git repository (or any of the parent directories): .git
thread 'main' panicked at 'command did not execute successfully: "git" "rev-list" "--author=bors@rust-lang.org" "-n1" "--first-parent" "HEAD" "--" "/home/jnelson/rust-lang/rust2/src/llvm-project" "/home/jnelson/rust-lang/rust2/src/bootstrap/download-ci-llvm-stamp" "/home/jnelson/rust-lang/rust2/src/version"
expected success, got: exit status: 128', src/bootstrap/native.rs:134:20
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
```
Support it too for consistency. It's unclear to me when anyone would need to use this, but `@bjorn3`
feels we should support it, and it's not much additional effort to get it working.
Previously, this was broken because of improper caching:
1. `StepDescription::maybe_run` builds `Compile::Std`, which only built `std` and not `proc_macro`
1. `Std` calls `builder.ensure(StdLink)`
1. `Rustc` calls `ensure(Std)`, which builds all crates, including `proc_macro`
1. `Rustc` calls `ensure(StdLink)`. `ensure` would see that it had already been run and do nothing. <-- bug is here
1. Cargo gives an error that `proc_macro` doesn't exist.
This fixes the caching by adding `crates` to `StdLink`, so it will get rerun if the crates that are
built change. This also does the same for `RustcLink`; it doesn't matter in practice currently
because nothing uses it except `impl Step for Rustc`, but it will avoid bugs if we start using it in
the future (e.g. to build individual crates for rustfmt).
Fix `download-ci-llvm` NixOS patching for binaries
LLVM tools should also be patched, since they are used in some tests, specially,
- src/test/run-make-fulldeps/cross-lang-lto (llvm-ar)
- src/test/run-make-fulldeps/cross-lang-lto-upstream-rlibs (llvm-ar)
- src/test/run-make-fulldeps/issue-64153 (llvm-objdump)
To be more future proof, we should patch all binaries in `bin`, which is done in this PR.
Group .test-arrow CSS rules and fix rgb/rgba property
Surprisingly, the web browsers were handling the `rgb`/`rgba` typo correctly. At least it now is as expected.
For the rest, it's simply grouping `.test-arrow` rules.
r? ``@Dylan-DPC``
promote placeholder bounds to 'static obligations
In NLL, when we are promoting a bound out from a closure, if we have a requirement that `T: 'a` where `'a` is in a higher universe, we were previously ignoring that, which is totally wrong. We should be promoting those constraints to `'static`, since universes are not expressible across closure boundaries.
Fixes#98693
~~(Marking as WIP because I'm still running tests, haven't add the new test, etc)~~
r? ``@jackh726``
LLVM tools should also be patched, since they are used in some tests,
specially,
- src/test/run-make-fulldeps/cross-lang-lto (llvm-ar)
- src/test/run-make-fulldeps/cross-lang-lto-upstream-rlibs (llvm-ar)
- src/test/run-make-fulldeps/issue-64153 (llvm-objdump)
To be more future proof, we should patch all binaries in `bin`.
@bjorn3 noticed that this is already allowed today when download-llvm is disabled, but breaks with it enabled:
```
$ ./rust2/x.py build
fatal: not a git repository (or any of the parent directories): .git
thread 'main' panicked at 'command did not execute successfully: "git" "rev-list" "--author=bors@rust-lang.org" "-n1" "--first-parent" "HEAD" "--" "/home/jnelson/rust-lang/rust2/src/llvm-project" "/home/jnelson/rust-lang/rust2/src/bootstrap/download-ci-llvm-stamp" "/home/jnelson/rust-lang/rust2/src/version"
expected success, got: exit status: 128', src/bootstrap/native.rs:134:20
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
```
Support it too for consistency. It's unclear to me when anyone would need to use this, but @bjorn3
feels we should support it, and it's not much additional effort to get it working.
This also updates a bunch of other git commands that were similarly depending on the current directory.
I've seen people using `optimize = false` and `full-bootstrap = true` in the past, without knowing
that they're not recommended. Remove `optimize` and a few other options that are always a bad idea,
and document that full-bootstrap is only for testing reproducible builds.
Previously, drop-tracking would incorrectly assume the struct would be dropped immediately, which
was not true: when the field had a type with a manual `Drop` impl, the drop becomes observable and
has to be dropped after the await instead.
For reasons I don't understand, this also fixes another error crater popped up related to type parameters.
#98476
proc_macro: Fix expand_expr expansion of bool literals
Previously, the expand_expr method would expand bool literals as a
`Literal` token containing a `LitKind::Bool`, rather than as an `Ident`.
This is not a valid token, and the `LitKind::Bool` case needs to be
handled seperately.
Tests were added to more deeply compare the streams in the expand-expr
test suite to catch mistakes like this in the future.
improve print styles
this change removes some interactive elements in ``@media` print` form.
more specifically, it removes the sidebar, source links, the expand/collapse toggle buttons, and the `#copy-path` button.
it also adjusts some spacing and removes the `.top-doc` description completely if it's currently collapsed.
Partially stabilize const_slice_from_raw_parts
This doesn't stabilize methods working on mutable pointers.
This pull request continues from #94946.
Pinging `@rust-lang/wg-const-eval` this because I use `rustc_allow_const_fn_unstable`. I believe this is justifiable as it's already possible to use `slice::from_raw_parts` in stable by abusing `transmute`. The stable alternative to this would be to provide a stable const implementation of `std::ptr::from_raw_parts` (as it can already be implemented in stable).
```rust
use std::mem;
#[repr(C)]
struct Slice<T> {
data: *const T,
len: usize,
}
fn main() {
let data: *const i32 = [1, 2, 3, 4].as_ptr();
let len = 4;
println!("{:?}", unsafe {
mem::transmute::<Slice<i32>, &[i32]>(Slice { data, len })
});
}
```
`@rustbot` modify labels: +T-libs-api
don't allow ZST in ScalarInt
There are several indications that we should not ZST as a ScalarInt:
- We had two ways to have ZST valtrees, either an empty `Branch` or a `Leaf` with a ZST in it.
`ValTree::zst()` used the former, but the latter could possibly arise as well.
- Likewise, the interpreter had `Immediate::Uninit` and `Immediate::Scalar(Scalar::ZST)`.
- LLVM codegen already had to special-case ZST ScalarInt.
So I propose we stop using ScalarInt to represent ZST (which are clearly not integers). Instead, we can add new ZST variants to those types that did not have other variants which could be used for this purpose.
Based on https://github.com/rust-lang/rust/pull/98831. Only the commits starting from "don't allow ZST in ScalarInt" are new.
r? `@oli-obk`
Fix several issues during cross compiling
- When cross compiling LLVM on an arm64 macOS machine to x86_64, CMake will produce universal binaries by default, causing link errors. Explicitly set `CMAKE_OSX_ARCHITECTURES` to the one single target architecture so that the executables and libraries will be single architecture.
- When cross compiling rustc with `llvm.clang = true`, `CLANG_TABLEGEN` has to be set to the host `clang-tblgen` executable to build clang.
Implement support for DWARF version 5.
DWARF version 5 brings a number of improvements over version 4. Quoting from
the announcement [1]:
> Version 5 incorporates improvements in many areas: better data compression,
> separation of debugging data from executable files, improved description of
> macros and source files, faster searching for symbols, improved debugging
> optimized code, as well as numerous improvements in functionality and
> performance.
On platforms where DWARF version 5 is supported (Linux, primarily), this commit
adds support for it behind a new `-Z dwarf-version=5` flag.
[1]: https://dwarfstd.org/Public_Review.php
r? ``@michaelwoerister``
Before, there was more or less duplicated suggestions to add type hints.
Fix by clearing more generic suggestions when a more specific suggestion
is possible.
This fixes#93506 .
Track implicit `Sized` obligations in type params
When we evaluate `ty::GenericPredicates` we introduce the implicit
`Sized` predicate of type params, but we do so with only the `Predicate`
its `Span` as context, we don't have an `Obligation` or
`ObligationCauseCode` we could influence. To try and carry this
information through, we add a new field to `ty::GenericPredicates` that
tracks both which predicates come from a type param and whether that
param has any bounds already (to use in suggestions).
We also suggest adding a `?Sized` bound if appropriate on E0599.
Address part of #98539.
