Enable use-after-scope checks by default when using AddressSanitizer.
They allow to detect incorrect use of stack objects after their scope
have already ended. The detection is based on LLVM lifetime intrinsics.
To facilitate the use of this functionality, the lifetime intrinsics are
now emitted regardless of optimization level if enabled sanitizer makes
use of them.
* Generate links to the primitive type docs for re-exports.
* Don't ICE on cross crate primitive type re-exports.
* Make primitive type re-exports show up cross crate.
rustc_span: return an impl Iterator instead of a Vec from macro_backtrace.
Having `Span::macro_backtrace` produce an `impl Iterator<Item = ExpnData>` allows #67359 to use it instead of rolling its own similar functionality.
The move from `MacroBacktrace` to `ExpnData` (which the first two commits are prerequisites for) both eliminates unnecessary allocations, and is strictly more flexible (exposes more information).
r? @petrochenkov
```
error[E0412]: cannot find type `T` in this scope
--> file.rs:3:12
|
3 | impl Trait<T> for Struct {}
| - ^ not found in this scope
| |
| help: you might be missing a type parameter: `<T>`
```
Fix#64298.
perf: Avoid creating a SmallVec if nothing changes during a fold
Not sure if this helps but in theory it should be less work than what
the current micro optimization does for `ty::Predicate` lists.
(It would explain the overhead I am seeing from `perf`.)
add a test for #60976
The test fails on 1.36.0 but passes on master.
Huge thanks for @hellow554 actually digging out the minimized version of the
repro.
Fixes#60976.
Do not ICE on multipart suggestions touching multiple files
When encountering a multipart suggestion with spans belonging to
different contexts, skip that suggestion.
Fix#68449. Similar to #68256.
rustc: Allow cdylibs to link against dylibs
Previously, rustc mandated that cdylibs could only link against rlibs as dependencies (not dylibs).
This commit disables that restriction and tests that it works in a simple case.
I don't have a windows rustc dev environment, so I guessed at the MSVC test code, I'm hoping the CI can run that for me.
Additionally, we might want to consider emitting (through cargo or rustc) some metadata to help C users of a cdylib figure out where all the dylibs they need are. I don't think that should be needed to land this change, as it will still be usable by homogeneous build systems without it.
My new test was templated off the `tests/run-make-fulldeps/cdylib` test. It seemed more appropriate to have it as a separate test, since both foo.rs and bar.rs would need to be replicated to make that test cover both cases, but I can do that if it would be preferred.
If I'm doing anything out of order/process, please let me know; this is only my second change to rustc and the prior one was trivial.
r? alexcrichton
Fix invalid link to the C++ Exception Handling ABI documentation
The original link is longer valid(404). I am assuming it's meant to be pointed to the Itanium C++ Exception Handling ABI documentation.
