Add job duration changes to post-merge analysis report
This should help us observe large regressions in job duration changes.
I would also like to add quick links to GH jobs/workflow to the post-merge workflow, but for that I first need to store some CI metadata to the bootstrap metrics, to make it easier to lookup the corresponding GH workflows (otherwise we'd need to look them up by commit SHA, which would be much more complicated). The last commit adds this metadata. Once this PR is merged, and the metadata will be available in the metrics stored on S3, I'll send a follow-up PR that uses the metadata to add links to job names in the post-merge workflow report.
r? `@marcoieni`
Remove unneeded LLVM CI test assertions
The `download_ci_llvm` bootstrap test was checking implementation details of the LLVM CI download check, which isn't very useful. It was essentially testing "if function_that_checks_if_llvm_ci_is_available returns true, we enable CI LLVM", but the usage of the function was an implementation detail. After https://github.com/rust-lang/rust/pull/138704, the inner implementation has changed, so the test now breaks if LLVM is updated.
I don't think that it's very useful to test implementation details like this, without taking the outside git state into account. Ideally, we should mock the git state for the test, otherwise the test will randomly break when executed in environments which the test does not control (e.g. on CI when a LLVM change happens).
I only kept the part of the test that checks that LLVM CI isn't used when we specify `download-ci-llvm = false`, as that should hold under all conditions, CI/local, and all git states.
I also kept the `if-unchanged` assertion, but only on CI, and as a temporary measure. After https://github.com/rust-lang/rust/pull/138591, we should have a proper way of mocking the git state to make the test robust, and make it test what we actually want.
Fixes [this](https://github.com/rust-lang/rust/pull/138784#issuecomment-2751460456).
r? `@ghost`
Improve suggest construct with literal syntax instead of calling
Closing #138931
When constructing a structure through a format similar to calling a constructor, we can use verbose suggestions to hint at using literal syntax for clearer advice. The case of multiple fields is also considered here, provided that the field has the same number of arguments as CallExpr.
r? compiler
Report compiletest pass mode if forced
This is very non-obvious if it fails in PR CI, because the starting invocation is miles away from the final test suite outcome.
Clean up a few things in rustc_hir_analysis::check::region
Each commit is independent. They are all small clean-ups in rustc_hir_analysis::check::region.
Remove `kw::Empty` uses from `rustc_middle`.
There are several places in `rustc_middle` that check for an empty lifetime name. These checks appear to be totally unnecessary, because empty lifetime names aren't produced here. (Empty lifetime names *are* possible in `hir::Lifetime`. Perhaps there was some confusion between it and the `rustc_middle` types?)
This commit removes the `kw::Empty` checks.
r? `@lcnr`
expand: Leave traces when expanding `cfg` attributes
This is the same as https://github.com/rust-lang/rust/pull/138515, but for `cfg(true)` instead of `cfg_attr`.
The difference is that `cfg(true)`s already left "traces" after themselves - the `cfg` attributes themselves, with `expanded_inert_attrs` set to true, with full tokens, available to proc macros.
This is not a reasonably expected behavior, but it could not be removed without a replacement, because a [major rustdoc feature](https://github.com/rust-lang/rfcs/pull/3631) and a number of clippy lints rely on it. This PR implements a replacement.
This needs a crater run, because it changes observable behavior (in an intended way) - proc macros can no longer see expanded `cfg(true)` attributes.
(Some minor unnecessary special casing for `sym::cfg_attr` is also removed in this PR.)
r? `@nnethercote`
rustdoc: Rearrange `Item`/`ItemInner`.
The `Item` struct is 48 bytes and contains a `Box<ItemInner>`;
`ItemInner` is 104 bytes. This is an odd arrangement. Normally you'd
have one of the following.
- A single large struct, which avoids the allocation for the `Box`, but
can result in lots of wasted space in unused parts of a container like
`Vec<Item>`, `HashSet<Item>`, etc.
- Or, something like `struct Item(Box<ItemInner>)`, which requires the
`Box` allocation but gives a very small Item size, which is good for
containers like `Vec<Item>`.
`Item`/`ItemInner` currently gets the worst of both worlds: it always
requires a `Box`, but `Item` is also pretty big and so wastes space in
containers. It would make sense to push it in one direction or the
other. #138916 showed that the first option is a regression for rustdoc,
so this commit does the second option, which improves speed and reduces
memory usage.
r? `@GuillaumeGomez`
Rollup of 10 pull requests
Successful merges:
- #130883 (Add environment variable query)
- #138624 (Add mipsel maintainer)
- #138672 (Avoiding calling queries when collecting active queries)
- #138935 (Update wg-prio triagebot config)
- #138946 (Un-bury chapters from the chapter list in rustc book)
- #138964 (Implement lint against using Interner and InferCtxtLike in random compiler crates)
- #138977 (Don't deaggregate InvocationParent just to reaggregate it again)
- #138980 (Collect items referenced from var_debug_info)
- #138985 (Use the correct binder scope for elided lifetimes in assoc consts)
- #138987 (Always emit `native-static-libs` note, even if it is empty)
r? `@ghost`
`@rustbot` modify labels: rollup
Use the correct binder scope for elided lifetimes in assoc consts
Beyond diagnostics this has no real effect, and it's also just about a future incompat lint. But it causes ICEs in some refactorings that I'm doing, so trying to get it out of the way
Collect items referenced from var_debug_info
The collection is limited to full debuginfo builds to match behavior of FunctionCx::compute_per_local_var_debug_info.
Fixes#138942.
Don't deaggregate InvocationParent just to reaggregate it again
Also makes it easier to add more things to it in the future (which I am doing in some local experiments, so not really a reason to do this just now, but I think this PR stands on its own).
Implement lint against using Interner and InferCtxtLike in random compiler crates
Often `Interner` defines similar methods to `TyCtxt` (but often simplified due to the simpler API surface of the type system layer for the new solver), which people will either unintentionally or intentionally import and use. Let's discourage that.
r? lcnr
Update wg-prio triagebot config
This completes the Zulip channel renaming after https://github.com/rust-lang/compiler-team/issues/848
Just nits: fixed a documentation link and the name of the Zulip channel for prioritization alerts.
r? ````@davidtwco````
Avoiding calling queries when collecting active queries
This PR changes active query collection to no longer call queries. Instead the fields needing queries have their computation delayed to when an cycle error is emitted or when printing the query backtrace in a panic.
This is done by splitting the fields in `QueryStackFrame` needing queries into a new `QueryStackFrameExtra` type. When collecting queries `QueryStackFrame` will contain a closure that can create `QueryStackFrameExtra`, which does make use of queries. Calling `lift` on a `QueryStackFrame` or `CycleError` will convert it to a variant containing `QueryStackFrameExtra` using those closures.
This also only calls queries needed to collect information on a cycle errors, instead of information on all active queries.
Calling queries when collecting active queries is a bit odd. Calling queries should not be done in the deadlock handler at all.
This avoids the out of memory scenario in https://github.com/rust-lang/rust/issues/124901.
Add environment variable query
Generally, `rustc` prefers command-line arguments, but in some cases, an environment variable really is the most sensible option. We should make sure that this works properly with the compiler's change-tracking mechanisms, such that changing the relevant environment variable causes a rebuild.
This PR is a first step forwards in doing that.
Part of the work needed to do https://github.com/rust-lang/rust/issues/118204, see https://github.com/rust-lang/rust/pull/129342 for some discussion.
r? ``@petrochenkov``
Instantiate binder before registering nested obligations for auto/built-in traits
Instead of turning a `Binder<Vec<Ty>>` into a bunch of higher-ranked predicates, instantiate the binder eagerly *once* and turn them into a bunch of non-higher-ranked predicates.
Right now this feels like a noop, but this `enter_forall_and_leak_universe` call would be the singular place where we could instantiate bound lifetime assumptions for coroutine witnesses... if we had them. Thus consolidating the binder instantiation here is useful if we want to fix the coroutine-auto-trait problem.
r? lcnr
Update cargo
10 commits in 307cbfda3119f06600e43cd38283f4a746fe1f8b..a6c604d1b8a2f2a8ff1f3ba6092f9fda42f4b7e9
2025-03-20 20:00:39 +0000 to 2025-03-26 18:11:00 +0000
- fix(package): update tracking issue for `--message-format` (rust-lang/cargo#15354)
- docs(contrib): Expand the description of team meetings (rust-lang/cargo#15349)
- feat(package): add unstable `--message-format` flag (rust-lang/cargo#15311)
- feat(complete): Added completion for `--profile` (rust-lang/cargo#15308)
- Uplift windows Cygwin DLL import libraries (rust-lang/cargo#15193)
- do not pass cdylib link args to test (rust-lang/cargo#15317)
- fix: revert the behavior checking lockfile's VCS status (rust-lang/cargo#15341)
- Temporarily ignore cargo_test_doctest_xcompile_ignores (rust-lang/cargo#15348)
- docs: fix typo in the "Shared cache" section (rust-lang/cargo#15346)
- Fix some issues with future-incompat report generation (rust-lang/cargo#15345)
r? ghost
Use `cfg_match!` in core
All of these uses of `cfg_if!` do not utilize that `cfg_if!` works with auto-`doc(cfg)`, so no functionality is lost from switching to use `cfg_match!` instead. We *do* lose what rustfmt special support exists for `cfg_if!`, though.
Tracking issue: #115585
Remove `prev_index_to_index` field from `CurrentDepGraph`
The dep graph currently has 2 ways to map a previous index into a current index. The `prev_index_to_index` map stores the current index equivalent of a previous index. For indices which are marked green, we also store the same information in the `DepNodeColorMap`. We actually only need to known the mapping for green nodes however, so this PR removes `prev_index_to_index` and instead makes use of the `DepNodeColorMap`.
To avoid racing when promoting a node from the previous session, the encoder lock is now used to ensure only one thread encodes the promoted node. This was previously done by the lock in `prev_index_to_index`.
This also changes `nodes_newly_allocated_in_current_session` used to detect duplicate dep nodes to contain both new and previous nodes, which is simpler and can better catch duplicates.
The dep node index encoding used in `DepNodeColorMap` is tweak to avoid subtraction / addition to optimize accessing the current equivalent of a previous index.
r? `@oli-obk`
Along with `TyCtx::env_var` helper. These can be used to track
environment variable accesses in the query system.
Since `TyCtx::env_var_os` uses `OsStr`, this commit also adds the
necessary trait implementations for that to work.
hygiene: Ensure uniqueness of `SyntaxContextData`s
`SyntaxContextData`s are basically interned with `SyntaxContext`s working as indices, so they are supposed to be unique.
However, currently duplicate `SyntaxContextData`s can be created during decoding from metadata or incremental cache.
This PR fixes that.
cc https://github.com/rust-lang/rust/pull/129827#discussion_r1759074553
This currently works because it's part of expansion, and that isn't yet
tracked by the query system. But we want to ensure it continues working,
even if that is changed.
