Fix re-export of doc hidden macro not showing up
It's part of the follow-up of https://github.com/rust-lang/rust/pull/109697.
Re-exports of doc hidden macros should be visible. It was the only kind of re-export of doc hidden item that didn't show up.
r? `@notriddle`
Perform MIR type ops locally in new solver
The new solver already does caching, and it's generally more correct to be using the infcx of the MIR typeck (which has the defining anchor set correctly and has already initialized all the opaques from HIR typeck).
This is based on #111918 so look at the final 3 commits.
This actually causes some tests to go from passing to failing, and failing to passing. Here's the full diff: https://www.diffchecker.com/hB4bh1A9/
Putting this up for exposure mostly.
r? `@lcnr`
Add #[inline] to array TryFrom impls
I was looking into https://github.com/rust-lang/rust/issues/111959 and I realized we don't have these. They seem like an uncontroversial addition.
IMO this PR does not fix that issue. I think the bad codegen is being caused by some underlying deeper problem but this change might cause the MIR inliner to paper over it in this specific case.
r? `@thomcc`
Include test suite metadata in the build metrics
This PR enhances the build metadata to include structured information about the test suites being executed, allowing external tools consuming the metadata to understand what was being tested.
The included metadata is:
* Target triple
* Host triple
* Stage number
* For compiletest tests:
* Suite name
* Mode
* Comparing mode
* For crate tests:
* List of crate names
This is implemented by replacing the `test` JSON node with a new `test_suite` node, which contains the metadata and the list of tests. This change also improves the handling of multiple test suites executed in the same step (for example in compiletest tests with a compare mode), as the multiple test suite executions will now be tracked in separate `test_suite` nodes.
This included a breaking change in the build metrics metadata format. To better handle this, in the second commit this PR introduces the `metadata_version` top-level field. The old version is considered to be `0`, while the new one `1`. Bootstrap will also gracefully handle existing metadata of a different version.
r? `@jyn514`
fix for `Self` not respecting tuple Ctor privacy
This PR fixes#111220 by checking the privacy of tuple constructors using `Self`, so the following code now errors
```rust
mod my {
pub struct Foo(&'static str);
}
impl AsRef<str> for my::Foo {
fn as_ref(&self) -> &str {
let Self(s) = self; // previously compiled, now errors correctly
s
}
}
```
Rollup of 8 pull requests
Successful merges:
- #111714 (Stop confusing specification levels when computing expectations.)
- #111927 (Migrate `item_static` to Askama)
- #111954 (improve error message for calling a method on a raw pointer with an unknown pointee)
- #111973 (Update current implementation comments for `select_nth_unstable`)
- #111976 (Generate docs for bootstrap itself)
- #111977 (Make errors from `x doc` less verbose)
- #111987 (do not prefer substs relate during coherence)
- #111991 (Change ty and const error's pretty printing to be in braces)
r? `@ghost`
`@rustbot` modify labels: rollup
Change ty and const error's pretty printing to be in braces
`[const error]` and `[type error]` are slightly confusing since they look like either a slice with an error type for the element ty or a slice with a const argument as the type ???. This PR changes them to display as `{const error}` and `{type error}` similar to `{integer}`.
This does not update the `Debug` impls for them which is done in #111988.
I updated some error logic to avoid printing the substs of trait refs when unable to resolve an assoc item for them, this avoids emitting errors with `{type error}` in them. The substs are not relevant for these errors since we don't take into account the substs when resolving the assoc item.
r? ``@compiler-errors``
Generate docs for bootstrap itself
This verifies the intra-doc links are correct, and hopefully makes things easier for new contributors.
Note that this will conflict with https://github.com/rust-lang/rust/pull/111955, i'm pretty sure i typo-ed some of the intra-doc links lol
Update current implementation comments for `select_nth_unstable`
This more accurately reflects the actual implementation, as it hasn't been a simple quickselect since #106997. While it does say that the current implementation always runs in O(n), I don't think it should require an FCP as it doesn't guarantee linearity in general and only points out that the current implementation is in fact linear.
r? `@Amanieu`
improve error message for calling a method on a raw pointer with an unknown pointee
The old error message had very confusing wording.
Also added some more test cases besides the single edition test.
r? `@compiler-errors`
Migrate `item_static` to Askama
This pull request addresses the type signature of the item_static function in our codebase. Previously, this function accepted a mutable reference to a Buffer for writing output. The current changes modify this to instead accept a mutable reference to any type that implements the Write trait.
Referes https://github.com/rust-lang/rust/issues/108868
Adds support for LLVM [SafeStack] which provides backward edge control
flow protection by separating the stack into two parts: data which is
only accessed in provable safe ways is allocated on the normal stack
(the "safe stack") and all other data is placed in a separate allocation
(the "unsafe stack").
SafeStack support is enabled by passing `-Zsanitizer=safestack`.
[SafeStack]: https://clang.llvm.org/docs/SafeStack.html
Previously, we would normalize *all* of
- the absolute path to the repository checkout
- the /rustc/$sha for stage1 (if `remap-debuginfo` was enabled)
- the /rustc/$sha for download-rustc
- the sysroot for download-rustc
Now, we consistently only normalize /rustc/FAKE_PREFIX. Not only is this
much simpler, but it also avoids ongoing maintenance for download-rustc
and makes it much less likely that tests break by accident.
- Change `tests/ui/track-diagnostics/track6.rs` to use a relative path
instead of an absolute one. I am not actually sure why `track_caller`
works here, but it does seem to work 🤷
- Pass `-Zsimulate-remapped-rust-src-base=/rustc/FAKE_PREFIX` to all
suites, not just UI. In particular, mir-opt tests emit /rustc/ paths
in their output.
Previously, we used the following info to determine whether to load the crate:
1. The METADATA_HEADER, which includes a METADATA_VERSION constant
2. The embedded rustc version
3. Various metadata in the `CrateRoot`, including the SVH
This worked ok most of the time. Unfortunately, when building locally the rustc version is always
the same because `omit-git-hash` is on by default. That meant that we depended only on 1 and 3, and
we are not very good about bumping METADATA_VERSION (it's currently at 7) so in practice we were
only depending on 3. `CrateRoot` is a very large struct and changes somewhat regularly, so this led
to a steady stream of crashes from trying to load it.
Change the logic to add an intermediate step between 2 and 3: introduce a new `CrateHeader` struct
that contains only the minimum info needed to decide whether the crate should be loaded or not. That
avoids having to load all of `CrateRoot`, which in practice means we should crash much less often.
Note that this works because the SVH should be different between any two dependencies, even if the
compiler has changed, because we use `-Zbinary-dep-depinfo` in bootstrap. See
https://github.com/rust-lang/rust/pull/111329#issuecomment-1538303474 for more details about how the
original crash happened.
Improve startup time of bootstrap
~~If the user has a `build/host` symlink set up, we can determine the target triple by reading it rather than invoking rustc. This significantly reduces startup time of bootstrap once any kind of build has been done~~
New approach explained below
```
➜ hyperfine -p 'git checkout -q master' -N './x.py -h' -r 50
Benchmark 1: ./x.py -h
Time (mean ± σ): 140.7 ms ± 2.6 ms [User: 99.9 ms, System: 39.3 ms]
Range (min … max): 136.8 ms … 149.6 ms 50 runs
➜ rust git:(master) hyperfine -p 'git checkout -q speedup-bootstrap-py' -N './x.py -h' -r 50
Benchmark 1: ./x.py -h
Time (mean ± σ): 95.2 ms ± 1.5 ms [User: 67.7 ms, System: 26.7 ms]
Range (min … max): 92.9 ms … 99.6 ms 50 runs
```
Also a small microoptimisation in using string splitting rather than regex when reading toml, which saves a few more milliseconds (2-5 testing locally), but less important.
Profiling shows the remaining runtime is around half setting up the Python runtime, and the vast majority of the remaining time is spent in subprocess building and running bootstrap itself, so probably can't be improved much further.
