| Age | Commit message (Collapse) | Author | Lines |
|---|
|
The functionality available on Apple platforms are very similar, and
were duplicated for each platform.
Additionally, this fixes a warning when compiling the standard library
for tvOS, watchOS and visionOS by marking the corresponding code as
dead code.
|
|
Fix minor typos in std::process doc on Win argv
|
|
Rollup of 6 pull requests
Successful merges:
- #122300 (Add FileCheck annotations to mir-opt/dest-prop tests)
- #127434 (use "bootstrap" instead of "rustbuild" in comments and docs)
- #127477 (Clear `inner_attr_ranges` regularly.)
- #127558 (More attribute cleanups)
- #127659 (Use ManuallyDrop in BufWriter::into_parts)
- #127671 (rustdoc: rename `issue-\d+.rs` tests to have meaningful names (part 8))
r? `@ghost`
`@rustbot` modify labels: rollup
|
|
Use ManuallyDrop in BufWriter::into_parts
The fact that `mem::forget` takes by value means that it interacts very poorly with Stacked Borrows; generally users think of calling it as a no-op, but in Stacked Borrows, the field retagging tends to cause surprise tag invalidation.
|
|
r=Mark-Simulacrum
Remove memory leaks in doctests in `core`, `alloc`, and `std`
cc `@RalfJung` https://github.com/rust-lang/rust/issues/126067 https://github.com/rust-lang/miri/issues/3670
Should be no actual *documentation* changes[^1], all added/modified lines in the doctests are hidden with `#`,
This PR splits the existing memory leaks in doctests in `core`, `alloc`, and `std` into two general categories:
1. "Non-focused" memory leaks that are incidental to the thing being documented, and/or are easy to remove, i.e. they are only there because preventing the leak would make the doctest less clear and/or concise.
- These doctests simply have a comment like `# // Prevent leaks for Miri.` above the added line that removes the memory leak.
- [^2]Some of these would perhaps be better as part of the public documentation part of the doctest, to clarify that a memory leak can happen if it is not otherwise mentioned explicitly in the documentation (specifically the ones in `(A)Rc::increment_strong_count(_in)`).
2. "Focused" memory leaks that are intentional and documented, and/or are possibly fragile to remove.
- These doctests have a `# // FIXME` comment above the line that removes the memory leak, with a note that once `-Zmiri-disable-leak-check` can be applied at test granularity, these tests should be "un-unleakified" and have `-Zmiri-disable-leak-check` enabled.
- Some of these are possibly fragile (e.g. unleaking the result of `Vec::leak`) and thus should definitely not be made part of the documentation.
This should be all of the leaks currently in `core` and `alloc`. I only found one leak in `std`, and it was in the first category (excluding the modules `@RalfJung` mentioned in https://github.com/rust-lang/rust/issues/126067 , and reducing the number of iterations of [one test](https://github.com/rust-lang/rust/blob/master/library/std/src/sync/once_lock.rs#L49-L94) from 1000 to 10)
[^1]: assuming [^2] is not added
[^2]: backlink
|
|
Windows: Add experimental support for linking std-required system DLLs using raw-dylib
For Windows, this allows std to define system imports without needing the user to have import libraries. It's intended for this to become the default.
For now it's an experimental feature so it can be tested using build-std.
|
|
|
|
fdopendir, openat and unlinkat are available since yosemite but we
support sierra as minimum.
|
|
Rollup of 3 pull requests
Successful merges:
- #127654 (Fix incorrect NDEBUG handling in LLVM bindings)
- #127661 (Stabilize io_slice_advance)
- #127668 (Improved slice documentation)
r? `@ghost`
`@rustbot` modify labels: rollup
|
|
Stabilize io_slice_advance
Closes https://github.com/rust-lang/rust/issues/62726 (FCP completed)
Stabilized API:
```rust
impl<'a> IoSlice<'a> {
pub fn advance(&mut self, n: usize);
pub fn advance_slices(bufs: &mut &mut [IoSlice<'a>], n: usize);
}
impl<'a> IoSliceMut<'a> {
pub fn advance(&mut self, n: usize);
pub fn advance_slices(bufs: &mut &mut [IoSliceMut<'a>], n: usize);
}
```
|
|
fix interleaved output in the default panic hook when multiple threads panic simultaneously
previously, we only held a lock for printing the backtrace itself. since all threads were printing to the same file descriptor, that meant random output in the default panic hook from one thread would be interleaved with the backtrace from another. now, we hold the lock for the full duration of the hook, and the output is ordered.
---
i noticed some odd things while working on this you may or may not already be aware of.
- libbacktrace is included as a submodule instead of a normal rustc crate, and as a result uses `cfg(backtrace_in_std)` instead of a more normal `cfg(feature = "rustc-dep-of-std")`. probably this is left over from before rust used a cargo-based build system?
- the default panic handler uses `trace_unsynchronized`, etc, in `sys::backtrace::print`. as a result, the lock only applies to concurrent *panic handlers*, not concurrent *threads*. in other words, if another, non-panicking, thread tried to print a backtrace at the same time as the panic handler, we may have UB, especially on windows.
- we have the option of changing backtrace to enable locking when `backtrace_in_std` is set so we can reuse their lock instead of trying to add our own.
|
|
Guard against calling `libc::exit` multiple times on Linux.
Mitigates (but does not fix) #126600 by ensuring only one thread which calls Rust `exit` actually calls `libc::exit`, and all other callers of Rust `exit` block.
|
|
|
|
|
|
previously, we only held a lock for printing the backtrace itself. since all threads were printing to the same file descriptor, that meant random output in the default panic hook would be interleaved with the backtrace. now, we hold the lock for the full duration of the hook, and the output is ordered.
|
|
Use pidfd_spawn for faster process spawning when a PidFd is requested
glibc 2.39 added `pidfd_spawnp` and `pidfd_getpid` which makes it possible to get pidfds while staying on the CLONE_VFORK path.
verified that vfork gets used with strace:
```
$ strace -ff -e pidfd_open,clone3,openat,execve,waitid,close ./x test std --no-doc -- pidfd
[...]
[pid 2820532] clone3({flags=CLONE_VM|CLONE_PIDFD|CLONE_VFORK|CLONE_CLEAR_SIGHAND, pidfd=0x7b7f885fec6c, exit_signal=SIGCHLD, stack=0x7b7f88aff000, stack_size=0x9000}strace: Process 2820533 attached
<unfinished ...>
[pid 2820533] execve("/home/the8472/bin/sleep", ["sleep", "1000"], 0x7ffdd0e268d8 /* 107 vars */) = -1 ENOENT (No such file or directory)
[pid 2820533] execve("/home/the8472/.cargo/bin/sleep", ["sleep", "1000"], 0x7ffdd0e268d8 /* 107 vars */) = -1 ENOENT (No such file or directory)
[pid 2820533] execve("/usr/local/bin/sleep", ["sleep", "1000"], 0x7ffdd0e268d8 /* 107 vars */) = -1 ENOENT (No such file or directory)
[pid 2820533] execve("/usr/bin/sleep", ["sleep", "1000"], 0x7ffdd0e268d8 /* 107 vars */ <unfinished ...>
[pid 2820532] <... clone3 resumed> => {pidfd=[3]}, 88) = 2820533
[pid 2820533] <... execve resumed>) = 0
[pid 2820532] openat(AT_FDCWD, "/proc/self/fdinfo/3", O_RDONLY|O_CLOEXEC) = 4
[pid 2820532] close(4) = 0
```
Tracking issue: #82971
|
|
|
|
Rename `lazy_cell_consume` to `lazy_cell_into_inner`
Name this something that is less confusable with an atomic consume API for `{Lazy,Once}Lock`.
|
|
Name this something that is less confusable with an atomic consume API for
`{Lazy,Once}Lock`.
|
|
|
|
|
|
That would give it alignment of 1 which is ABI-incompatible with its C
definition.
|
|
|
|
|
|
control how a new process should display its window (normal, minimized, maximized, etc)
|
|
clarify `sys::unix::fd::FileDesc::drop` comment
closes #66876
simply clarifies some resource-relevant things regarding the `close` syscall to reduce the amount of search needed in other parts of the web.
|
|
|
|
|
|
|
|
|
|
once_lock: make test not take as long in Miri
Allocating 1000 list elements takes a while (`@zachs18` reported >5min), so let's reduce the iteration count when running in Miri. Unfortunately due to this clever `while let i @ 0..LEN =` thing, the count needs to be a constants, and constants cannot be shadowed, so we need to use another trick to hide the `cfg!(miri)` from the docs. (I think this loop condition may be a bit too clever, it took me a bit to decipher. Ideally this would be `while let i = ... && i < LEN`, but that is not stable yet.)
|
|
Improve std::Path's Hash quality by avoiding prefix collisions
This adds a bit rotation to the already existing state so that the same sequence of characters chunked at different offsets into separate path components results in different hashes.
The tests are from #127255
Closes #127254
|
|
|
|
|
|
impl PathBuf::add_extension and Path::with_added_extension
See the ACP for motivation and discussions - https://github.com/rust-lang/libs-team/issues/368
|
|
|
|
|
|
Co-authored-by: Ralf Jung <post@ralfj.de>
|
|
Signed-off-by: tison <wander4096@gmail.com>
|
|
Signed-off-by: tison <wander4096@gmail.com>
|
|
For Windows, this allows defining imports without needing the user to have import libraries. It's intended for this to become the default.
|
|
|
|
Update windows-bindgen to 0.58.0
This also switches from the bespoke `std` generated bindings to the normal `sys` ones everyone else uses.
This has almost no difference except that the `sys` bindings use the `windows_targets::links!` macro for FFI imports, which we implement manually. This does cause the diff to look much larger than it really is but the bulk of the changes are mostly contained to the generated code.
|
|
|
|
|
|
Remove unqualified form import of io::Error in process_vxworks.rs and fallback on remove_dir_impl for vxworks
Hi all,
This is to address issue #127084. On inspections it was found that io::Error refrences were all of qualified form and there was no need to add a unqualified form import. Also to successfully build rust for vxworks, we need to fallback on the remove_impl_dir implementations.
Thank you.
|
|
wasm64 build with target-feature=+simd128,+atomics
Fixes https://github.com/rust-lang/rust/issues/126778
|
|
Signed-off-by: tison <wander4096@gmail.com>
|
|
|
|
|
