| Age | Commit message (Collapse) | Author | Lines |
|
internal: Remove generated md book files
|
|
|
|
Add missing name-ref parents to syntactic highlighting
|
|
minor: Show build scripts errors in server status
|
|
|
|
Make change annotations per text-edit
|
|
fix: Fix diagnostics being cleared right after being received
|
|
|
|
|
|
minor: Sync from downstream
|
|
|
|
internal: Run proc-macro server tests as separate CI job
|
|
|
|
|
|
Touch tt
|
|
fix: Normalize projections in evaluated const display and layout calculation
|
|
feat: Warn the user when a rename will change the meaning of the program
|
|
Fix `path` macro hygiene
|
|
|
|
|
|
|
|
Log build script error output in `load_cargo::load_workspace_at`
|
|
fix(hir): `VariantDef` is `impl HasSource`
|
|
fix: do not apply editorconfig to git commit msg
|
|
Fix syntax fixup producing invalid punctuation
|
|
Bump chalk for built-in supports of async closures
|
|
Rank ADT constructors as constructors for completion scoring
|
|
std: move stdio to `sys`
As per #117276, this moves the platform definitions of `Stdout` and friends into `sys`. This PR also unifies the UNIX and Hermit implementations and moves the `__rust_print_err` function needed by libunwind on SGX into the dedicated module for such helper functions.
|
|
r=fmease,GuillaumeGomez
doctests: fix merging on stable
Fixes #137898
The generated multi-test harness relies on nightly-only APIs, so the only way to run it on stable is to enable them.
To prevent the executing test case from getting at any of the stuff that the harness uses, they're built as two separate crates. The test bundle isn't built with RUSTC_BOOTSTRAP, while the runner harness is.
|
|
This prevents the included test case from getting at nightly-only
features when run on stable. The harness builds with
RUSTC_BOOTSTRAP, but the bundle doesn't.
|
|
Fixes #137898
The generated multi-test harness relies on nightly-only APIs,
so the only way to run it on stable is to enable them. Since
tests that use crate attrs don't be merged, there's no way to use
nightly-only features on it anyway.
|
|
As per #117276, this moves the platform definitions of `Stdout` and friends into `sys`. This PR also unifies the UNIX and Hermit implementations and moves the `__rust_print_err` function needed by libunwind on SGX into the dedicated module for such helper functions.
|
|
Always inline `query_get_at`.
r? `@saethlin`
|
|
Rollup of 7 pull requests
Successful merges:
- #122790 (Apply dllimport in ThinLTO)
- #137650 (Move `fs` into `sys`)
- #138228 (Use `disjoint_bitor` inside `borrowing_sub`)
- #138233 (Windows: Don't link std (and run-make) against advapi32, except on win7)
- #138253 (Continue to check attr if meet empty repr for adt)
- #138263 (Fix `repr128-dwarf` test)
- #138276 (Lazy load NtOpenFile for UWP)
r? `@ghost`
`@rustbot` modify labels: rollup
|
|
Lazy load NtOpenFile for UWP
Lazily load `NtOpenFile` to allow libraries targeting UWP to build and link.
Fixes #138257 .
r? `@ChrisDenton`
|
|
Fix `repr128-dwarf` test
The test now correctly ignores enums from `std`.
Fixes #138254
Unblocks #138200
|
|
Continue to check attr if meet empty repr for adt
Fixes #138241
Returning while checking ReprEmpty results in missing the check for the next repr
|
|
Windows: Don't link std (and run-make) against advapi32, except on win7
Std no longer depends on any functionality provided by advapi32, so we can remove it from the list of external libraries we link against. Except, the win7 targets do still rely on advapi32-provided functionality. This PR therefore moves linking against it to only occur on win7 targets, so that no new uses of it slip in without being noticed.
|
|
Use `disjoint_bitor` inside `borrowing_sub`
This makes the definition of `borrowing_sub` consistent with that of `carrying_add`.
|
|
Move `fs` into `sys`
Move platform definitions of `fs` into `std::sys`, as part of https://github.com/rust-lang/rust/issues/117276.
cc `@joboet`
|
|
Apply dllimport in ThinLTO
This partially reverts https://github.com/rust-lang/rust/pull/103353 by properly applying `dllimport` if `-Z dylib-lto` is passed. That PR should probably fully be reverted as it looks quite sketchy. We don't know locally if the entire crate graph would be statically linked.
This should hopefully be sufficient to make ThinLTO work for rustc on Windows.
r? ``@wesleywiser``
---
Edit: This PR is changed to just generally revert https://github.com/rust-lang/rust/pull/103353.
|
|
Split the `Edges` iterator.
Some nice performance wins here, mostly on the `wg-grammar` benchmark.
r? `@lcnr`
|
|
|
|
|
|
Rollup of 12 pull requests
Successful merges:
- #136127 (Allow `*const W<dyn A> -> *const dyn A` ptr cast)
- #136968 (Turn order dependent trait objects future incompat warning into a hard error)
- #137319 (Stabilize `const_vec_string_slice`)
- #137885 (tidy: add triagebot checks)
- #138040 (compiler: Use `size_of` from the prelude instead of imported)
- #138084 (Use workspace lints for crates in `compiler/`)
- #138158 (Move more layouting logic to `rustc_abi`)
- #138160 (depend more on attr_data_structures and move find_attr! there)
- #138192 (crashes: couple more tests)
- #138216 (bootstrap: Fix stack printing when a step cycle is detected)
- #138232 (Reduce verbosity of GCC build log)
- #138242 (Revert "Don't test new error messages with the stage 0 compiler")
r? `@ghost`
`@rustbot` modify labels: rollup
|
|
fix: Prevent wrong invocations of `needs_parens_in` with non-ancestral "parent"s
|
|
|
|
Revert "Don't test new error messages with the stage 0 compiler"
|
|
Reduce verbosity of GCC build log
It was a bit too spammy before.
Fixes: https://github.com/rust-lang/rust/issues/138213
r? ``@GuillaumeGomez``
|
|
bootstrap: Fix stack printing when a step cycle is detected
When bootstrap detects a step dependency cycle (which represents a bootstrap bug), it is supposed to print out the contents of the step stack as part of its panic message.
However, while investigating #138205 it was found that bootstrap was actually printing out several copies of `Any { .. }`, because that is the Debug implementation for `dyn Any`. This is sadly not very helpful.
This PR fixes that problem by introducing a `trait AnyDebug: Any + Debug` that delegates to the underlying type's Debug implementation, while still allowing downcasting via Any.
---
The fixed behaviour can be verified manually (and is tested automatically) via a new dummy command, `./x run cyclic-step`:
```
$ x run cyclic-step
Building bootstrap
Finished `dev` profile [unoptimized] target(s) in 0.02s
thread 'main' panicked at src/bootstrap/src/core/builder/mod.rs:1521:17:
Cycle in build detected when adding CyclicStep { n: 0 }
CyclicStep { n: 0 }
CyclicStep { n: 1 }
CyclicStep { n: 2 }
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
Build completed unsuccessfully in 0:00:00
```
|