LLVM's memory dependence analysis doesn't properly account for calls
that could unwind and thus effectively act as a branching point. This
can lead to stores that are only visible when the call unwinds being
removed, possibly leading to calls to drop() functions with b0rked
memory contents.
As there is no fix for this in LLVM yet and we want to keep
compatibility to current LLVM versions anyways, we have to workaround
this bug by omitting the noalias attribute on &mut function arguments.
Benchmarks suggest that the performance loss by this change is very
small.
Thanks to @RalfJung for pushing me towards not removing too many
noalias annotations and @alexcrichton for helping out with the test for
this bug.
Fixes#29485
As demonstrated in the `resolve_socket_addr` change, this is less awkward than re-creating a new address from the other parts.
If this is to be accepted, pleas open a tracking issue (I can’t set the appropriate tags) and I’ll update the PR with the tracking issue number.
r? @Manishearth
I just noticed they can't be rolled up (often modifying the same line(s) in imports). So once I reach the critical amount for them to be merged I'll create a PR that merges all of them.
We no longer have a separate powerpc64 and powerpc64le target_arch, and instead use target_endian to select between the two. These patches fix a couple of remaining issues.
This PR adds to `NameBinding` so it can more fully represent bindings from imports as well from items, refactors away `Target`, generalizes `ImportResolution` to a simpler type `NameResolution`, and uses a single `NameResolution`-valued map in place the existing maps `children` and `import_resolutions` (of `NameBinding`s and `ImportResolution`s, respectively), simplifying duplicate checking and name resolution.
It also unifies the `resolve_name_in_module` in `lib.rs` with its namesake in `resolve_imports.rs`, clarifying and improving the core logic (fixes#31403 and fixes#31404) while maintaining clear future-comparability with shadowable globs (i.e., never reporting that a resolution is a `Success` or is `Failing` unless this would also be knowable with shadowable globs).
Since it fixes#31403, this is technically a [breaking-change], but it is exceedingly unlikely to cause breakage in practice. The following is an example of code that would break:
```rust
mod foo {
pub mod bar {} // This defines bar in the type namespace
pub use alpha::bar; // This defines bar in the value namespace
// This should define baz in both namespaces, but it only defines baz in the type namespace.
pub use self::bar as baz;
pub fn baz() {} // This should collide with baz, but now it does not.
}
pub fn f() {}
mod alpha {
pub use self::f as bar; // Changing this to `pub fn bar() {}` causes the collision right now.
pub use super::*;
}
```
r? @nrc
This allows printing pointers to unsized types with the {:p} formatting
directive. The following impls are extended to unsized types:
- impl<'a, T: ?Sized> Pointer for &'a T
- impl<'a, T: ?Sized> Pointer for &'a mut T
- impl<T: ?Sized> Pointer for *const T
- impl<T: ?Sized> Pointer for *mut T
- impl<T: ?Sized> fmt::Pointer for Box<T>
- impl<T: ?Sized> fmt::Pointer for Rc<T>
- impl<T: ?Sized> fmt::Pointer for Arc<T>
These commits are an implementation of https://github.com/rust-lang/rfcs/pull/1359 which is tracked via https://github.com/rust-lang/rust/issues/31398. The `before_exec` implementation fit easily with the current process spawning framework we have, but unfortunately the `exec` implementation required a bit of a larger refactoring. The stdio handles were all largely managed as implementation details of `std::process` and the `exec` function lived in `std::sys`, so the two didn't have access to one another.
I took this as a sign that a deeper refactoring was necessary, and I personally feel that the end result is cleaner for both Windows and Unix. The commits should be separated nicely for reviewing (or all at once if you're feeling ambitious), but the changes made here were:
* The process spawning on Unix was refactored in to a pre-exec and post-exec function. The post-exec function isn't allowed to do any allocations of any form, and management of transmitting errors back to the parent is managed by the pre-exec function (as it's the one that actually forks).
* Some management of the exit status was pushed into platform-specific modules. On Unix we must cache the return value of `wait` as the pid is consumed after we wait on it, but on Windows we can just keep querying the system because the handle stays valid.
* The `Stdio::None` variant was renamed to `Stdio::Null` to better reflect what it's doing.
* The global lock on `CreateProcess` is now correctly positioned to avoid unintended inheritance of pipe handles that other threads are sending to their child processes. After a more careful reading of the article referenced the race is not in `CreateProcess` itself, but rather the property that handles are unintentionally shared.
* All stdio management now happens in platform-specific modules. This provides a cleaner implementation/interpretation for `FromFraw{Fd,Handle}` for each platform as well as a cleaner transition from a configuration to what-to-do once we actually need to do the spawn.
With these refactorings in place, implementing `before_exec` and `exec` ended up both being pretty trivial! (each in their own commit)
LLVM's memory dependence analysis doesn't properly account for calls
that could unwind and thus effectively act as a branching point. This
can lead to stores that are only visible when the call unwinds being
removed, possibly leading to calls to drop() functions with b0rked
memory contents.
As there is no fix for this in LLVM yet and we want to keep
compatibility to current LLVM versions anyways, we have to workaround
this bug by omitting the noalias attribute on &mut function arguments.
Benchmarks suggest that the performance loss by this change is very
small.
Thanks to @RalfJung for pushing me towards not removing too many
noalias annotations and @alexcrichton for helping out with the test for
this bug.
Fixes#29485
Previously when breaking tokens into smaller pieces, the replace_token
function have been used. It replaced current token and updated span
information, but it did not clear the list of expected tokens, neither
did it update remaining info about last token. This could lead to
incorrect error message, like one described in the issue #24780:
expected one of ... `>` ... found `>`
This commit implements the `exec` function proposed in [RFC 1359][rfc] which is
a function on the `CommandExt` trait to execute all parts of a `Command::spawn`
without the `fork` on Unix. More details on the function itself can be found in
the comments in the commit.
[rfc]: https://github.com/rust-lang/rfcs/pull/1359
cc #31398
