The previous behaviour ignored slice lengths above a certain length
because it could not do otherwise. We now have VarLenSlice however, that
can represent the ignored lengths to make the algorithm more consistent.
This does not change the correctness of the algorithm, but makes it
easier to reason about.
As a nice side-effect, exhaustiveness errors have improved: they now
capture all missing lengths instead of only the shortest.
Note that where we previously ran `max_slice_len` with input having not
only matrix.heads() but also v.head(). Now we run it on matrix.heads()
only, but also take into account the currently processed constructor.
This preserves behavior since `pat_constructors` returns only one
constructor in the case that is of interest for us.
This improves error messages by indicating when slices above a certain
lengths have not been matched. Previously, we would only report examples
of such lengths, but of course never all of them.
rustc_target: rename {Fn,Arg}Type to {Fn,Arg}Abi.
I was trying to tweak the API of `FnType` (now `FnAbi`) and the name kept bothering me.
`FnAbi` is to a function signature a bit like a layout is to a type, so the name still isn't perfect yet, but at least it doesn't have the misleading `Type` in it anymore.
If this can't land I think I can continue my original refactor without it, so I'm not strongly attached to it.
r? @nagisa cc @oli-obk
Clarify pattern-matching usefulness algorithm
This PR clarifies a bit the usefulness algorithm by emphasizing that each row of the matrix can be seen as a sort of stack from which we pop constructors. It also moves code around to increase separation of concerns.
This is part of my splitting of https://github.com/rust-lang/rust/pull/65160 into smaller PRs.
expand: Feature gate out-of-line modules in proc macro input
Extracted from https://github.com/rust-lang/rust/pull/64273.
We are currently gating attributes applied directly to `mod` items because there are unresolved questions about out-of-line modules and their behavior is very likely to change.
However, you can sneak an out-of-line module into an attribute macro input using modules nested into other items like
```rust
#[my_attr]
fn m() {
#[path = "zzz.rs"]
mod n; // what tokens does the `my_attr` macro see?
}
```
This PR prevents that and emits a feature gate error for this case as well.
r? @Centril
It would be great to land this before beta.
Remove LintBuffer from Session
This moves the `LintBuffer` from `Session` into the `Resolver`, where it is used until lowering is done and then consumed by early lint passes. This also happily removes the failure mode of buffering lints too late where it would have previously lead to ICEs; it is statically no longer possible to do so.
I suspect that with a bit more work a similar move could be done for the lint buffer inside `ParseSess`, but this PR doesn't touch it (in part to keep itself small).
The last commit is the "interesting" commit -- the ones before it don't work (though they compile) as they sort of prepare the various crates for the lint buffer to be passed in rather than accessed through Session.
Reduce amount of errors given unclosed delimiter
When in a file with a non-terminated item, catch the error and consume
the block instead of trying to recover it on a more granular way in order to
reduce the amount of unrelated errors that would be fixed after adding
the missing closing brace. Also point out the possible location of the
missing closing brace.
Fix#63690.
Move help popup generation code
The first commit is just a small cleanup.
The idea behind this PR is to reduce a bit more the generated HTML files by moving the duplicated code into one place instead.
r? @kinnison
Allow foreign exceptions to unwind through Rust code and Rust panics to unwind through FFI
This PR fixes interactions between Rust panics and foreign (mainly C++) exceptions.
C++ exceptions (and other FFI exceptions) can now safely unwind through Rust code:
- The FFI function causing the unwind must be marked with `#[unwind(allowed)]`. If this is not the case then LLVM may optimize landing pads away with the assumption that they are unreachable.
- Drop code will be executed as the exception unwinds through the stack, as with a Rust panic.
- `catch_unwind` will *not* catch the exception, instead the exception will silently continue unwinding past it.
Rust panics can now safely unwind through C++ code:
- C++ destructors will be called as the stack unwinds.
- The Rust panic can only be caught with `catch (...)`, after which it can be either rethrown or discarded.
- C++ cannot name the type of the Rust exception object used for unwinding, which means that it can't be caught explicitly or have its contents inspected.
Tests have been added to ensure all of the above works correctly.
Some notes about non-C++ exceptions:
- `pthread_cancel` and `pthread_exit` use unwinding on glibc. This has the same behavior as a C++ exception: destructors are run but it cannot be caught by `catch_unwind`.
- `longjmp` on Windows is implemented using unwinding. Destructors are run on MSVC, but not on MinGW. In both cases the unwind cannot be caught by `catch_unwind`.
- As with C++ exceptions, you need to mark the relevant FFI functions with `#[unwind(allowed)]`, otherwise LLVM will optimize out the destructors since they seem unreachable.
I haven't updated any of the documentation, so officially unwinding through FFI is still UB. However this is a step towards making it well-defined.
Fixes#65441
cc @gnzlbg
r? @alexcrichton
Validate error patterns and error annotation in ui tests when present
Previously, when compilation succeeded, neither error patterns nor error
annotation would be validated. Additionally, when compilation failed,
only error patterns would be validated if both error patterns and error
annotation were present.
Now both error patterns and error annotation are validated when present,
regardless of compilation status. Furthermore, for test that should run,
the error patterns are matched against executable output, which is what
some of tests already expect to happen, and when #65506 is merged even
more ui tests will.
Fixes#56277
Remove a loop which runs exactly once
Though the code seems to work properly, it is worth removing the loop entirely in order to not confuse the reader.
r? @estebank
The concrete type that will be too big is target dependent. Avoid
matching it in error annotation to make test work correctly across
different targets.
Previously, when compilation succeeded, neither error patterns nor error
annotation would be validated. Additionally, when compilation failed,
only error patterns would be validated if both error patterns and error
annotation were present.
Now both error patterns and error annotation are validated when present,
regardless of compilation status. Furthermore, for test that should run,
the error patterns are matched against executable output, which is what
some of tests already expect to happen, and when #65506 is merged even
more ui tests will.
When both error patterns and error annotations are present in an ui
test, only error patterns are validated against the output.
Replace the error pattern with an error annotation to avoid silently
ignoring the other error annotation.
