Make sure formatter errors are emitted by the default Write::write_fmt
Previously, if an error was returned from the formatter that did not
originate in an underlying writer error, Write::write_fmt would return
successfully even if the formatting did not complete (was interrupted by
an `fmt::Error` return).
Now we choose to emit an io::Error with kind Other for formatter errors.
Since this may reveal error returns from `write!()` and similar that
previously passed silently, it's a kind of a [breaking-change].
Fixes#31879
Previously, if an error was returned from the formatter that did not
originate in an underlying writer error, Write::write_fmt would return
successfully even if the formatting did not complete (was interrupted by
an `fmt::Error` return).
Now we choose to emit an io::Error with kind Other for formatter errors.
Since this may reveal error returns from `write!()` and similar that
previously passed silently, it's a kind of a [breaking-change].
You can now group tests into directories like `run-pass/borrowck` or `compile-fail/borrowck`. By default, all `.rs` files within any directory are considered tests: to ignore some directory, create a placeholder file called `compiletest-ignore-dir` (I had to do this for several existing directories).
r? @alexcrichton
cc @brson
This will correctly add the thread_local attribute to the external static variable ```errno```:
```rust
extern {
#[thread_local]
static errno: c_int;
}
```
Before this commit, the thread_local attribute is ignored. Fixes#30795.
Thanks @alexcrichton for pointing out the solution.
This PR changes the visibility of extern crate declarations to match that of items (fixes#26775).
To avoid breakage, the PR makes it a `public_in_private` lint to reexport a private extern crate, and it adds the lint `inaccessible_extern_crate` for uses of an inaccessible extern crate.
The lints can be avoided by making the appropriate `extern crate` declaration public.
Windows is not #[cfg(target_thread_local)] and as such should link
to the external symbol. But it fails with:
thread '<main>' panicked at 'assertion failed: `(left == right)` (left: `272246271`, right: `3`)', C:/bot/slave/auto-win-msvc-64-opt/build/src/test/run-pass/thread-local-extern-static.rs:24
This warning was introduced on Nov 28, 2015 and got into 1.6 stable, it was later requalified from a hardwired warning to a warn-by-default lint.
If this patch is landed soon enough, then `match_of_unit_variant_via_paren_dotdot` will get into 1.8 stable as a deny-by-default lint.
My intention is to turn it into a hard error after March 3, 2016, then it will hit stable at 1.9.
r? @nikomatsakis
cc @pnkfelix
In #31717 we rebased our LLVM fork over the 3.8 release branch, and it was
thought that this fixed#31702. The testing, however, must have been erroneous,
as it unfortunately didn't fix the issue! Our MUSL nightly builders are failing
from the same assertion reported in the issue, so we at least know the test case
is a reproduction!
I believe the failure is only happening on the MUSL nightly builders because
none of the auto builders have LLVM assertions enabled, and the Linux nightly
builder *does* have assertions enabled for the binaries we generate but the
distcheck run doesn't test a compiler with LLVM assertions enabled.
In #31717 we rebased our LLVM fork over the 3.8 release branch, and it was
thought that this fixed#31702. The testing, however, must have been erroneous,
as it unfortunately didn't fix the issue! Our MUSL nightly builders are failing
from the same assertion reported in the issue, so we at least know the test case
is a reproduction!
I believe the failure is only happening on the MUSL nightly builders because
none of the auto builders have LLVM assertions enabled, and the Linux nightly
builder *does* have assertions enabled for the binaries we generate but the
distcheck run doesn't test a compiler with LLVM assertions enabled.
Remove alternate stack with sigaltstack before unmaping it.
Also reuse existing signal stack if already set, this is especially
useful when working with sanitizers that configure alternate stack
themselves.
This change depends on SS_DISABLE recently introduced in libc crate and updates
this git submodule accordingly.
This commit rebases our LLVM submodule on the most recent tip of the
`release_38` branch of LLVM. There's been a few fixes and this notably fixes the
assertion error in #31702.
Closes#31702
This commit rebases our LLVM submodule on the most recent tip of the
`release_38` branch of LLVM. There's been a few fixes and this notably fixes the
assertion error in #31702.
this improves typeck performance by 5% (LLVM times are still huge).
Basically fixes#25916 (still O(n^2), but the example takes <1s to
compile).
r? @nikomatsakis
The issue was that the const evaluator was returning an error because
the feature flag const_indexing wasn't turned on. The error was then
reported as a bug.
Fixes#29914
The issue was that the const evaluator was returning an error because
the feature flag const_indexing wasn't turned on. The error was then
reported as a bug.
Fixes#29914
Ignores 82 rpass tests that use threads.
I took care to only ignore tests that call `thread::spawn`. Some tests, for example `issue-16597`, also do fail because of lack of threads support, but for other reasons.
With this PR, we're down to 49 failures.
r? @brson
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
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
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
This is a Unix-specific function which adds the ability to register a closure to
run pre-exec to configure the child process as required (note that these
closures are run post-fork).
cc #31398
