Rollup of 7 pull requests
Successful merges:
- #106200 (Suggest `impl Fn*` and `impl Future` in `-> _` return suggestions)
- #106274 (Add JSON output to -Zdump-mono-stats)
- #106292 (Add codegen test for `Box::new(uninit)` of big arrays)
- #106327 (Add tidy check for dbg)
- #106361 (Note maximum integer literal for `IntLiteralTooLarge`)
- #106396 (Allow passing a specific date to `bump-stage0`)
- #106436 (Enable doctests for rustc_query_impl)
Failed merges:
r? `@ghost`
`@rustbot` modify labels: rollup
Suggest `impl Fn*` and `impl Future` in `-> _` return suggestions
Follow-up to #106172, only the last commit is relevant. Can rebase once that PR is landed for easier review.
Suggests `impl Future` and `impl Fn{,Mut,Once}` in `-> _` return suggestions.
r? `@estebank`
layout_of: `T: Thin` implies `sizeof(&T) == sizeof(usize)`
Use the `<T as Pointee>::Metadata` associated type to calculate the layout of a pointee's metadata, instead of hard-coding rules about certain types.
Maybe this approach is overkill -- we could instead hard-code this approach as a fallback, with the matching on `Slice`/`Dynamic`/etc. happening first
Fixes this issue here https://github.com/rust-lang/rust/pull/104338#issuecomment-1312595844 .. But is also useful with transmutes, for example, given the UI test I added below.
Refactoring report_method_error
While working on #105732, I found it's hard to follow this long function,
so I tried to make it shorter.
It's not easy for code reviewing, since so many lines of code changes,
but only the positions are changed.
Generally, I extract two sub-methods from `report_method_error`:
397b66e77b/compiler/rustc_hir_typeck/src/method/suggest.rs (L117)
to `note_candidates_on_method_error`
And this long block:
397b66e77b/compiler/rustc_hir_typeck/src/method/suggest.rs (L263)
to `report_no_match_method_error`.
r? `@compiler-errors`
Rollup of 8 pull requests
Successful merges:
- #104552 (warn newer available version of the x tool)
- #105681 (some fixes/improvements to mir::visit module)
- #106005 (Test the borrowck behavior of if-let guards)
- #106356 (clean: Remove `ctor_kind` from `VariantStruct`.)
- #106365 (Grammar : Missing "is" in format specifier diagnostic)
- #106388 (rustdoc: remove legacy box-sizing CSS)
- #106392 (`has_overflow` only if value is *not* within limit)
- #106402 (Fix dupe word typos)
Failed merges:
r? `@ghost`
`@rustbot` modify labels: rollup
Only specify `--target` by default for `-Zgcc-ld=lld` on wasm
On macOS, it's not yet clear which cases of clang/OS/target/SDK version impact how to find ld/lld/rust-lld. The `--target` argument is not needed on our current targets with a vanilla config to do so, but may be in some cases ?
That is, things look to be different and more subtle than suggested in https://github.com/rust-lang/rust/issues/97402#issuecomment-1147404520.
Specifying this argument unconditionally currently breaks `-Zgcc-ld=lld` on the 10.7+ targets on x64 macOS. Vanilla configs on x64 and aarch64 don't seem to need it to be able to find `rust-lld`.
This fixes#101653 on macOS (I've tried on x64 and aarch64: vanilla installs with the most recent command line tools available for each of the OS versions I have access to, `run-make/issue-71519` passes).
I didn't expect the previous PRs to fail because of the existing tests, but CI does not actually run those tests yet, which explains the regressions. I was hoping to fix those in this PR but it's more involved (building lld is required for the tests to run, llvm/lld is not built on the test builders but on the dist builders, the dist builders don't run tests). This PR is just to unblock current users on macOS who have reported issues, but a regression could happen in the future by mistake until then.
r? `@petrochenkov`
Only include metadata for non-dynamic libraries in rustc-dev
The actual object code should be linked from librustc_driver.so, which is still included in rustc-dev. This saves on download time and disk usage.
Fixes https://github.com/rust-lang/rust/issues/103538
do not add noalias in return position
`noalias` as a return attribute in LLVM indicates that the returned pointer does not alias anything else that is reachable from the caller, *including things reachable before this function call*. This is clearly not the case with a function like `fn id(Box<T>) -> Box<T>`, so we cannot use this attribute.
Fixes https://github.com/rust-lang/unsafe-code-guidelines/issues/385 (including an actual miscompilation that `@comex` managed to produce).
Document some of the AST nodes
Someone was confused about some of this on Zulip, added some docs
We probably should make sure every last field/variant in the AST/HIR is documented at some point
`@bors` rollup
Reduce HIR debug output
HIR debug output is currently very verbose, especially when used with the alternate (`#`) flag. This commit reduces the amount of noisy newlines by forcing a few small key types to stay on one line, which makes the output easier to read and scroll by.
```
$ rustc +after hello_world.rs -Zunpretty=hir-tree | wc -l
582
$ rustc +before hello_world.rs -Zunpretty=hir-tree | wc -l
932
```
On macOS, it's not yet clear which cases of clang/OS/target/SDK version impact
how to find ld/lld/rust-lld. The --target is not needed on our current targets with
a vanilla config, but may be in some cases. Specifying it all the time breaks the 10.7+
targets on x64 macOS.
We try to only specify it on macOS if the linker flavors are different,
for possible cases of cross-compilation with `-Zgcc-ld=lld` but the
expectation is that it should be passed manually when needed in these
situations.
This allows analyzing the output programatically; for example, finding
the item with the highest `total_estimate`.
I also took the liberty of adding `untracked` tests to `rustc_session` and documentation to the unstable book for `dump-mono-items`.
Reenable limited top-down MIR inlining
Reverts most of #105119 and uses an alternative strategy to prevent exponential blowup. Specifically, we allow doing top-down inlining up to depth at most five, and for at most one call site per nested body.
r? `@cjgillot`
Abolish `QueryVTable` in favour of more assoc items on `QueryConfig`
This may introduce additional mono _but_ may help const fold things better and especially may help not constructing a `QueryVTable` anymore which is cheap but not free.
This may introduce additional mono _but_ may help const fold things
better and especially may help not constructing a `QueryVTable` anymore
which is cheap but not free.
HIR debug output is currently very verbose, especially when used with
the alternate (`#`) flag. This commit reduces the amount of noisy
newlines by forcing a few small key types to stay on one line, which
makes the output easier to read and scroll by.
```
$ rustc +after hello_world.rs -Zunpretty=hir-tree | wc -l
582
$ rustc +before hello_world.rs -Zunpretty=hir-tree | wc -l
932
```
Encode spans relative to the enclosing item -- enable on nightly
Follow-up to #84373 with the flag `-Zincremental-relative-spans` set by default.
This PR seeks to remove one of the main shortcomings of incremental: the handling of spans.
Changing the contents of a function may require redoing part of the compilation process for another function in another file because of span information is changed.
Within one file: all the spans in HIR change, so typechecking had to be re-done.
Between files: spans of associated types/consts/functions change, so type-based resolution needs to be re-done (hygiene information is stored in the span).
The flag `-Zincremental-relative-spans` encodes local spans relative to the span of an item, stored inside the `source_span` query.
Trap: stashed diagnostics are referenced by the "raw" span, so stealing them requires to remove the span's parent.
In order to avoid too much traffic in the span interner, span encoding uses the `ctxt_or_tag` field to encode:
- the parent when the `SyntaxContext` is 0;
- the `SyntaxContext` when the parent is `None`.
Even with this, the PR creates a lot of traffic to the Span interner, when a Span has both a LocalDefId parent and a non-root SyntaxContext. They appear in lowering, when we add a parent to all spans, including those which come from macros, and during inlining when we mark inlined spans.
The last commit changes how queries of `LocalDefId` manage their cache. I can put this in a separate PR if required.
Possible future directions:
- validate that all spans are marked in HIR validation;
- mark macro-expanded spans relative to the def-site and not the use-site.
