std/

process.rs

//! A module for working with processes.
//!
//! This module is mostly concerned with spawning and interacting with child
//! processes, but it also provides [`abort`] and [`exit`] for terminating the
//! current process.
//!
//! # Spawning a process
//!
//! The [`Command`] struct is used to configure and spawn processes:
//!
//! ```no_run
//! use std::process::Command;
//!
//! let output = Command::new("echo")
//!     .arg("Hello world")
//!     .output()
//!     .expect("Failed to execute command");
//!
//! assert_eq!(b"Hello world\n", output.stdout.as_slice());
//! ```
//!
//! Several methods on [`Command`], such as [`spawn`] or [`output`], can be used
//! to spawn a process. In particular, [`output`] spawns the child process and
//! waits until the process terminates, while [`spawn`] will return a [`Child`]
//! that represents the spawned child process.
//!
//! # Handling I/O
//!
//! The [`stdout`], [`stdin`], and [`stderr`] of a child process can be
//! configured by passing an [`Stdio`] to the corresponding method on
//! [`Command`]. Once spawned, they can be accessed from the [`Child`]. For
//! example, piping output from one command into another command can be done
//! like so:
//!
//! ```no_run
//! use std::process::{Command, Stdio};
//!
//! // stdout must be configured with `Stdio::piped` in order to use
//! // `echo_child.stdout`
//! let echo_child = Command::new("echo")
//!     .arg("Oh no, a tpyo!")
//!     .stdout(Stdio::piped())
//!     .spawn()
//!     .expect("Failed to start echo process");
//!
//! // Note that `echo_child` is moved here, but we won't be needing
//! // `echo_child` anymore
//! let echo_out = echo_child.stdout.expect("Failed to open echo stdout");
//!
//! let mut sed_child = Command::new("sed")
//!     .arg("s/tpyo/typo/")
//!     .stdin(Stdio::from(echo_out))
//!     .stdout(Stdio::piped())
//!     .spawn()
//!     .expect("Failed to start sed process");
//!
//! let output = sed_child.wait_with_output().expect("Failed to wait on sed");
//! assert_eq!(b"Oh no, a typo!\n", output.stdout.as_slice());
//! ```
//!
//! Note that [`ChildStderr`] and [`ChildStdout`] implement [`Read`] and
//! [`ChildStdin`] implements [`Write`]:
//!
//! ```no_run
//! use std::process::{Command, Stdio};
//! use std::io::Write;
//!
//! let mut child = Command::new("/bin/cat")
//!     .stdin(Stdio::piped())
//!     .stdout(Stdio::piped())
//!     .spawn()
//!     .expect("failed to execute child");
//!
//! // If the child process fills its stdout buffer, it may end up
//! // waiting until the parent reads the stdout, and not be able to
//! // read stdin in the meantime, causing a deadlock.
//! // Writing from another thread ensures that stdout is being read
//! // at the same time, avoiding the problem.
//! let mut stdin = child.stdin.take().expect("failed to get stdin");
//! std::thread::spawn(move || {
//!     stdin.write_all(b"test").expect("failed to write to stdin");
//! });
//!
//! let output = child
//!     .wait_with_output()
//!     .expect("failed to wait on child");
//!
//! assert_eq!(b"test", output.stdout.as_slice());
//! ```
//!
//! # Windows argument splitting
//!
//! On Unix systems arguments are passed to a new process as an array of strings,
//! but on Windows arguments are passed as a single commandline string and it is
//! up to the child process to parse it into an array. Therefore the parent and
//! child processes must agree on how the commandline string is encoded.
//!
//! Most programs use the standard C run-time `argv`, which in practice results
//! in consistent argument handling. However, some programs have their own way of
//! parsing the commandline string. In these cases using [`arg`] or [`args`] may
//! result in the child process seeing a different array of arguments than the
//! parent process intended.
//!
//! Two ways of mitigating this are:
//!
//! * Validate untrusted input so that only a safe subset is allowed.
//! * Use [`raw_arg`] to build a custom commandline. This bypasses the escaping
//!   rules used by [`arg`] so should be used with due caution.
//!
//! `cmd.exe` and `.bat` files use non-standard argument parsing and are especially
//! vulnerable to malicious input as they may be used to run arbitrary shell
//! commands. Untrusted arguments should be restricted as much as possible.
//! For examples on handling this see [`raw_arg`].
//!
//! ### Batch file special handling
//!
//! On Windows, `Command` uses the Windows API function [`CreateProcessW`] to
//! spawn new processes. An undocumented feature of this function is that
//! when given a `.bat` file as the application to run, it will automatically
//! convert that into running `cmd.exe /c` with the batch file as the next argument.
//!
//! For historical reasons Rust currently preserves this behavior when using
//! [`Command::new`], and escapes the arguments according to `cmd.exe` rules.
//! Due to the complexity of `cmd.exe` argument handling, it might not be
//! possible to safely escape some special characters, and using them will result
//! in an error being returned at process spawn. The set of unescapeable
//! special characters might change between releases.
//!
//! Also note that running batch scripts in this way may be removed in the
//! future and so should not be relied upon.
//!
//! [`spawn`]: Command::spawn
//! [`output`]: Command::output
//!
//! [`stdout`]: Command::stdout
//! [`stdin`]: Command::stdin
//! [`stderr`]: Command::stderr
//!
//! [`Write`]: io::Write
//! [`Read`]: io::Read
//!
//! [`arg`]: Command::arg
//! [`args`]: Command::args
//! [`raw_arg`]: crate::os::windows::process::CommandExt::raw_arg
//!
//! [`CreateProcessW`]: https://learn.microsoft.com/en-us/windows/win32/api/processthreadsapi/nf-processthreadsapi-createprocessw

#![stable(feature = "process", since = "1.0.0")]
#![deny(unsafe_op_in_unsafe_fn)]

#[cfg(all(
    test,
    not(any(
        target_os = "emscripten",
        target_os = "wasi",
        target_env = "sgx",
        target_os = "xous",
        target_os = "trusty",
        target_os = "hermit",
    ))
))]
mod tests;

use crate::convert::Infallible;
use crate::ffi::OsStr;
use crate::io::prelude::*;
use crate::io::{self, BorrowedCursor, IoSlice, IoSliceMut};
use crate::num::NonZero;
use crate::path::Path;
use crate::sys::{AsInner, AsInnerMut, FromInner, IntoInner, process as imp};
use crate::{fmt, format_args_nl, fs, str};

/// Representation of a running or exited child process.
///
/// This structure is used to represent and manage child processes. A child
/// process is created via the [`Command`] struct, which configures the
/// spawning process and can itself be constructed using a builder-style
/// interface.
///
/// There is no implementation of [`Drop`] for child processes,
/// so if you do not ensure the `Child` has exited then it will continue to
/// run, even after the `Child` handle to the child process has gone out of
/// scope.
///
/// Calling [`wait`] (or other functions that wrap around it) will make
/// the parent process wait until the child has actually exited before
/// continuing.
///
/// # Warning
///
/// On some systems, calling [`wait`] or similar is necessary for the OS to
/// release resources. A process that terminated but has not been waited on is
/// still around as a "zombie". Leaving too many zombies around may exhaust
/// global resources (for example process IDs).
///
/// The standard library does *not* automatically wait on child processes (not
/// even if the `Child` is dropped), it is up to the application developer to do
/// so. As a consequence, dropping `Child` handles without waiting on them first
/// is not recommended in long-running applications.
///
/// # Examples
///
/// ```should_panic
/// use std::process::Command;
///
/// let mut child = Command::new("/bin/cat")
///     .arg("file.txt")
///     .spawn()
///     .expect("failed to execute child");
///
/// let ecode = child.wait().expect("failed to wait on child");
///
/// assert!(ecode.success());
/// ```
///
/// [`wait`]: Child::wait
#[stable(feature = "process", since = "1.0.0")]
#[cfg_attr(not(test), rustc_diagnostic_item = "Child")]
pub struct Child {
    pub(crate) handle: imp::Process,

    /// The handle for writing to the child's standard input (stdin), if it
    /// has been captured. You might find it helpful to do
    ///
    /// ```ignore (incomplete)
    /// let stdin = child.stdin.take().expect("handle present");
    /// ```
    ///
    /// to avoid partially moving the `child` and thus blocking yourself from calling
    /// functions on `child` while using `stdin`.
    #[stable(feature = "process", since = "1.0.0")]
    pub stdin: Option<ChildStdin>,

    /// The handle for reading from the child's standard output (stdout), if it
    /// has been captured. You might find it helpful to do
    ///
    /// ```ignore (incomplete)
    /// let stdout = child.stdout.take().expect("handle present");
    /// ```
    ///
    /// to avoid partially moving the `child` and thus blocking yourself from calling
    /// functions on `child` while using `stdout`.
    #[stable(feature = "process", since = "1.0.0")]
    pub stdout: Option<ChildStdout>,

    /// The handle for reading from the child's standard error (stderr), if it
    /// has been captured. You might find it helpful to do
    ///
    /// ```ignore (incomplete)
    /// let stderr = child.stderr.take().expect("handle present");
    /// ```
    ///
    /// to avoid partially moving the `child` and thus blocking yourself from calling
    /// functions on `child` while using `stderr`.
    #[stable(feature = "process", since = "1.0.0")]
    pub stderr: Option<ChildStderr>,
}

impl AsInner<imp::Process> for Child {
    #[inline]
    fn as_inner(&self) -> &imp::Process {
        &self.handle
    }
}

impl FromInner<(imp::Process, StdioPipes)> for Child {
    fn from_inner((handle, io): (imp::Process, StdioPipes)) -> Child {
        Child {
            handle,
            stdin: io.stdin.map(ChildStdin::from_inner),
            stdout: io.stdout.map(ChildStdout::from_inner),
            stderr: io.stderr.map(ChildStderr::from_inner),
        }
    }
}

impl IntoInner<imp::Process> for Child {
    fn into_inner(self) -> imp::Process {
        self.handle
    }
}

#[stable(feature = "std_debug", since = "1.16.0")]
impl fmt::Debug for Child {
    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
        f.debug_struct("Child")
            .field("stdin", &self.stdin)
            .field("stdout", &self.stdout)
            .field("stderr", &self.stderr)
            .finish_non_exhaustive()
    }
}

/// The pipes connected to a spawned process.
///
/// Used to pass pipe handles between this module and [`imp`].
pub(crate) struct StdioPipes {
    pub stdin: Option<imp::ChildPipe>,
    pub stdout: Option<imp::ChildPipe>,
    pub stderr: Option<imp::ChildPipe>,
}

/// A handle to a child process's standard input (stdin).
///
/// This struct is used in the [`stdin`] field on [`Child`].
///
/// When an instance of `ChildStdin` is [dropped], the `ChildStdin`'s underlying
/// file handle will be closed. If the child process was blocked on input prior
/// to being dropped, it will become unblocked after dropping.
///
/// [`stdin`]: Child::stdin
/// [dropped]: Drop
#[stable(feature = "process", since = "1.0.0")]
pub struct ChildStdin {
    inner: imp::ChildPipe,
}

// In addition to the `impl`s here, `ChildStdin` also has `impl`s for
// `AsFd`/`From<OwnedFd>`/`Into<OwnedFd>` and
// `AsRawFd`/`IntoRawFd`/`FromRawFd`, on Unix and WASI, and
// `AsHandle`/`From<OwnedHandle>`/`Into<OwnedHandle>` and
// `AsRawHandle`/`IntoRawHandle`/`FromRawHandle` on Windows.

#[stable(feature = "process", since = "1.0.0")]
impl Write for ChildStdin {
    fn write(&mut self, buf: &[u8]) -> io::Result<usize> {
        (&*self).write(buf)
    }

    fn write_vectored(&mut self, bufs: &[IoSlice<'_>]) -> io::Result<usize> {
        (&*self).write_vectored(bufs)
    }

    fn is_write_vectored(&self) -> bool {
        io::Write::is_write_vectored(&&*self)
    }

    #[inline]
    fn flush(&mut self) -> io::Result<()> {
        (&*self).flush()
    }
}

#[stable(feature = "write_mt", since = "1.48.0")]
impl Write for &ChildStdin {
    fn write(&mut self, buf: &[u8]) -> io::Result<usize> {
        self.inner.write(buf)
    }

    fn write_vectored(&mut self, bufs: &[IoSlice<'_>]) -> io::Result<usize> {
        self.inner.write_vectored(bufs)
    }

    fn is_write_vectored(&self) -> bool {
        self.inner.is_write_vectored()
    }

    #[inline]
    fn flush(&mut self) -> io::Result<()> {
        Ok(())
    }
}

impl AsInner<imp::ChildPipe> for ChildStdin {
    #[inline]
    fn as_inner(&self) -> &imp::ChildPipe {
        &self.inner
    }
}

impl IntoInner<imp::ChildPipe> for ChildStdin {
    fn into_inner(self) -> imp::ChildPipe {
        self.inner
    }
}

impl FromInner<imp::ChildPipe> for ChildStdin {
    fn from_inner(pipe: imp::ChildPipe) -> ChildStdin {
        ChildStdin { inner: pipe }
    }
}

#[stable(feature = "std_debug", since = "1.16.0")]
impl fmt::Debug for ChildStdin {
    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
        f.debug_struct("ChildStdin").finish_non_exhaustive()
    }
}

/// A handle to a child process's standard output (stdout).
///
/// This struct is used in the [`stdout`] field on [`Child`].
///
/// When an instance of `ChildStdout` is [dropped], the `ChildStdout`'s
/// underlying file handle will be closed.
///
/// [`stdout`]: Child::stdout
/// [dropped]: Drop
#[stable(feature = "process", since = "1.0.0")]
pub struct ChildStdout {
    inner: imp::ChildPipe,
}

// In addition to the `impl`s here, `ChildStdout` also has `impl`s for
// `AsFd`/`From<OwnedFd>`/`Into<OwnedFd>` and
// `AsRawFd`/`IntoRawFd`/`FromRawFd`, on Unix and WASI, and
// `AsHandle`/`From<OwnedHandle>`/`Into<OwnedHandle>` and
// `AsRawHandle`/`IntoRawHandle`/`FromRawHandle` on Windows.

#[stable(feature = "process", since = "1.0.0")]
impl Read for ChildStdout {
    fn read(&mut self, buf: &mut [u8]) -> io::Result<usize> {
        self.inner.read(buf)
    }

    fn read_buf(&mut self, buf: BorrowedCursor<'_, u8>) -> io::Result<()> {
        self.inner.read_buf(buf)
    }

    fn read_vectored(&mut self, bufs: &mut [IoSliceMut<'_>]) -> io::Result<usize> {
        self.inner.read_vectored(bufs)
    }

    #[inline]
    fn is_read_vectored(&self) -> bool {
        self.inner.is_read_vectored()
    }

    fn read_to_end(&mut self, buf: &mut Vec<u8>) -> io::Result<usize> {
        self.inner.read_to_end(buf)
    }
}

impl AsInner<imp::ChildPipe> for ChildStdout {
    #[inline]
    fn as_inner(&self) -> &imp::ChildPipe {
        &self.inner
    }
}

impl IntoInner<imp::ChildPipe> for ChildStdout {
    fn into_inner(self) -> imp::ChildPipe {
        self.inner
    }
}

impl FromInner<imp::ChildPipe> for ChildStdout {
    fn from_inner(pipe: imp::ChildPipe) -> ChildStdout {
        ChildStdout { inner: pipe }
    }
}

#[stable(feature = "std_debug", since = "1.16.0")]
impl fmt::Debug for ChildStdout {
    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
        f.debug_struct("ChildStdout").finish_non_exhaustive()
    }
}

/// A handle to a child process's stderr.
///
/// This struct is used in the [`stderr`] field on [`Child`].
///
/// When an instance of `ChildStderr` is [dropped], the `ChildStderr`'s
/// underlying file handle will be closed.
///
/// [`stderr`]: Child::stderr
/// [dropped]: Drop
#[stable(feature = "process", since = "1.0.0")]
pub struct ChildStderr {
    inner: imp::ChildPipe,
}

// In addition to the `impl`s here, `ChildStderr` also has `impl`s for
// `AsFd`/`From<OwnedFd>`/`Into<OwnedFd>` and
// `AsRawFd`/`IntoRawFd`/`FromRawFd`, on Unix and WASI, and
// `AsHandle`/`From<OwnedHandle>`/`Into<OwnedHandle>` and
// `AsRawHandle`/`IntoRawHandle`/`FromRawHandle` on Windows.

#[stable(feature = "process", since = "1.0.0")]
impl Read for ChildStderr {
    fn read(&mut self, buf: &mut [u8]) -> io::Result<usize> {
        self.inner.read(buf)
    }

    fn read_buf(&mut self, buf: BorrowedCursor<'_, u8>) -> io::Result<()> {
        self.inner.read_buf(buf)
    }

    fn read_vectored(&mut self, bufs: &mut [IoSliceMut<'_>]) -> io::Result<usize> {
        self.inner.read_vectored(bufs)
    }

    #[inline]
    fn is_read_vectored(&self) -> bool {
        self.inner.is_read_vectored()
    }

    fn read_to_end(&mut self, buf: &mut Vec<u8>) -> io::Result<usize> {
        self.inner.read_to_end(buf)
    }
}

impl AsInner<imp::ChildPipe> for ChildStderr {
    #[inline]
    fn as_inner(&self) -> &imp::ChildPipe {
        &self.inner
    }
}

impl IntoInner<imp::ChildPipe> for ChildStderr {
    fn into_inner(self) -> imp::ChildPipe {
        self.inner
    }
}

impl FromInner<imp::ChildPipe> for ChildStderr {
    fn from_inner(pipe: imp::ChildPipe) -> ChildStderr {
        ChildStderr { inner: pipe }
    }
}

#[stable(feature = "std_debug", since = "1.16.0")]
impl fmt::Debug for ChildStderr {
    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
        f.debug_struct("ChildStderr").finish_non_exhaustive()
    }
}

/// A process builder, providing fine-grained control
/// over how a new process should be spawned.
///
/// A default configuration can be
/// generated using `Command::new(program)`, where `program` gives a path to the
/// program to be executed. Additional builder methods allow the configuration
/// to be changed (for example, by adding arguments) prior to spawning:
///
/// ```
/// # if cfg!(not(all(target_vendor = "apple", not(target_os = "macos")))) {
/// use std::process::Command;
///
/// let output = if cfg!(target_os = "windows") {
///     Command::new("cmd")
///         .args(["/C", "echo hello"])
///         .output()
///         .expect("failed to execute process")
/// } else {
///     Command::new("sh")
///         .arg("-c")
///         .arg("echo hello")
///         .output()
///         .expect("failed to execute process")
/// };
///
/// let hello = output.stdout;
/// # }
/// ```
///
/// `Command` can be reused to spawn multiple processes. The builder methods
/// change the command without needing to immediately spawn the process.
///
/// ```no_run
/// use std::process::Command;
///
/// let mut echo_hello = Command::new("sh");
/// echo_hello.arg("-c").arg("echo hello");
/// let hello_1 = echo_hello.output().expect("failed to execute process");
/// let hello_2 = echo_hello.output().expect("failed to execute process");
/// ```
///
/// Similarly, you can call builder methods after spawning a process and then
/// spawn a new process with the modified settings.
///
/// ```no_run
/// use std::process::Command;
///
/// let mut list_dir = Command::new("ls");
///
/// // Execute `ls` in the current directory of the program.
/// list_dir.status().expect("process failed to execute");
///
/// println!();
///
/// // Change `ls` to execute in the root directory.
/// list_dir.current_dir("/");
///
/// // And then execute `ls` again but in the root directory.
/// list_dir.status().expect("process failed to execute");
/// ```
#[stable(feature = "process", since = "1.0.0")]
#[cfg_attr(not(test), rustc_diagnostic_item = "Command")]
pub struct Command {
    inner: imp::Command,
}

impl Command {
    /// Constructs a new `Command` for launching the program at
    /// path `program`, with the following default configuration:
    ///
    /// * No arguments to the program
    /// * Inherit the current process's environment
    /// * Inherit the current process's working directory
    /// * Inherit stdin/stdout/stderr for [`spawn`] or [`status`], but create pipes for [`output`]
    ///
    /// [`spawn`]: Self::spawn
    /// [`status`]: Self::status
    /// [`output`]: Self::output
    ///
    /// Builder methods are provided to change these defaults and
    /// otherwise configure the process.
    ///
    /// If `program` is not an absolute path, the `PATH` environment variable
    /// will be searched in an OS-defined way.
    ///
    /// # Platform-specific behavior
    ///
    /// The details below describe the current behavior, but these details
    /// may change in future versions of Rust.
    ///
    /// On Unix, the `PATH` searched comes from the child's environment:
    ///
    /// - If the environment is unmodified, the child inherits the parent's
    ///   `PATH` and that is what is searched.
    /// - If `PATH` is explicitly set via [`env`], that new value is searched.
    /// - If [`env_clear`] or [`env_remove`] removes `PATH` without a
    ///   replacement, `execvp` falls back to an OS-defined default (typically
    ///   `/bin:/usr/bin`), **not** the parent's `PATH`. This may fail to find
    ///   programs that rely on the parent's `PATH`.
    ///
    /// To avoid surprises, use an absolute path or explicitly set `PATH` on
    /// the `Command` when modifying the child's environment.
    ///
    /// On Windows, Rust resolves the executable path before spawning, rather
    /// than passing the name to `CreateProcessW` for resolution. When
    /// `program` is not an absolute path, the following locations are searched
    /// in order:
    ///
    /// 1. The child's `PATH`, if explicitly set via [`env`].
    /// 2. The directory of the current executable.
    /// 3. The system directory (`GetSystemDirectoryW`).
    /// 4. The Windows directory (`GetWindowsDirectoryW`).
    /// 5. The parent process's `PATH`.
    ///
    /// Note: when `PATH` is cleared via [`env_clear`] or [`env_remove`] on
    /// Windows, step 1 is skipped but the parent process's `PATH` is still
    /// searched at step 5, unlike on Unix.
    ///
    /// For executable files, the `.exe` extension may be omitted. Files with
    /// other extensions must include the extension, otherwise they will not be
    /// found. Note that this behavior has some known limitations
    /// (see issue #37519).
    ///
    /// [`env`]: Self::env
    /// [`env_remove`]: Self::env_remove
    /// [`env_clear`]: Self::env_clear
    ///
    /// # Examples
    ///
    /// ```no_run
    /// use std::process::Command;
    ///
    /// Command::new("sh")
    ///     .spawn()
    ///     .expect("sh command failed to start");
    /// ```
    ///
    /// # Caveats
    ///
    /// [`Command::new`] is only intended to accept the path of the program. If you pass a program
    /// path along with arguments like `Command::new("ls -l").spawn()`, it will try to search for
    /// `ls -l` literally. The arguments need to be passed separately, such as via [`arg`] or
    /// [`args`].
    ///
    /// ```no_run
    /// use std::process::Command;
    ///
    /// Command::new("ls")
    ///     .arg("-l") // arg passed separately
    ///     .spawn()
    ///     .expect("ls command failed to start");
    /// ```
    ///
    /// [`arg`]: Self::arg
    /// [`args`]: Self::args
    #[stable(feature = "process", since = "1.0.0")]
    pub fn new<S: AsRef<OsStr>>(program: S) -> Command {
        Command { inner: imp::Command::new(program.as_ref()) }
    }

    /// Adds an argument to pass to the program.
    ///
    /// Only one argument can be passed per use. So instead of:
    ///
    /// ```no_run
    /// # std::process::Command::new("sh")
    /// .arg("-C /path/to/repo")
    /// # ;
    /// ```
    ///
    /// usage would be:
    ///
    /// ```no_run
    /// # std::process::Command::new("sh")
    /// .arg("-C")
    /// .arg("/path/to/repo")
    /// # ;
    /// ```
    ///
    /// To pass multiple arguments see [`args`].
    ///
    /// [`args`]: Command::args
    ///
    /// Note that the argument is not passed through a shell, but given
    /// literally to the program. This means that shell syntax like quotes,
    /// escaped characters, word splitting, glob patterns, variable substitution,
    /// etc. have no effect.
    ///
    /// <div class="warning">
    ///
    /// On Windows, use caution with untrusted inputs. Most applications use the
    /// standard convention for decoding arguments passed to them. These are safe to
    /// use with `arg`. However, some applications such as `cmd.exe` and `.bat` files
    /// use a non-standard way of decoding arguments. They are therefore vulnerable
    /// to malicious input.
    ///
    /// In the case of `cmd.exe` this is especially important because a malicious
    /// argument can potentially run arbitrary shell commands.
    ///
    /// See [Windows argument splitting][windows-args] for more details
    /// or [`raw_arg`] for manually implementing non-standard argument encoding.
    ///
    /// [`raw_arg`]: crate::os::windows::process::CommandExt::raw_arg
    /// [windows-args]: crate::process#windows-argument-splitting
    ///
    /// </div>
    ///
    /// # Examples
    ///
    /// ```no_run
    /// use std::process::Command;
    ///
    /// Command::new("ls")
    ///     .arg("-l")
    ///     .arg("-a")
    ///     .spawn()
    ///     .expect("ls command failed to start");
    /// ```
    #[stable(feature = "process", since = "1.0.0")]
    pub fn arg<S: AsRef<OsStr>>(&mut self, arg: S) -> &mut Command {
        self.inner.arg(arg.as_ref());
        self
    }

    /// Adds multiple arguments to pass to the program.
    ///
    /// To pass a single argument see [`arg`].
    ///
    /// [`arg`]: Command::arg
    ///
    /// Note that the arguments are not passed through a shell, but given
    /// literally to the program. This means that shell syntax like quotes,
    /// escaped characters, word splitting, glob patterns, variable substitution, etc.
    /// have no effect.
    ///
    /// <div class="warning">
    ///
    /// On Windows, use caution with untrusted inputs. Most applications use the
    /// standard convention for decoding arguments passed to them. These are safe to
    /// use with `arg`. However, some applications such as `cmd.exe` and `.bat` files
    /// use a non-standard way of decoding arguments. They are therefore vulnerable
    /// to malicious input.
    ///
    /// In the case of `cmd.exe` this is especially important because a malicious
    /// argument can potentially run arbitrary shell commands.
    ///
    /// See [Windows argument splitting][windows-args] for more details
    /// or [`raw_arg`] for manually implementing non-standard argument encoding.
    ///
    /// [`raw_arg`]: crate::os::windows::process::CommandExt::raw_arg
    /// [windows-args]: crate::process#windows-argument-splitting
    ///
    /// </div>
    ///
    /// # Examples
    ///
    /// ```no_run
    /// use std::process::Command;
    ///
    /// Command::new("ls")
    ///     .args(["-l", "-a"])
    ///     .spawn()
    ///     .expect("ls command failed to start");
    /// ```
    #[stable(feature = "process", since = "1.0.0")]
    pub fn args<I, S>(&mut self, args: I) -> &mut Command
    where
        I: IntoIterator<Item = S>,
        S: AsRef<OsStr>,
    {
        for arg in args {
            self.arg(arg.as_ref());
        }
        self
    }

    /// Inserts or updates an explicit environment variable mapping.
    ///
    /// This method allows you to add an environment variable mapping to the spawned process or
    /// overwrite a previously set value. You can use [`Command::envs`] to set multiple environment
    /// variables simultaneously.
    ///
    /// Child processes will inherit environment variables from their parent process by default.
    /// Environment variables explicitly set using [`Command::env`] take precedence over inherited
    /// variables. You can disable environment variable inheritance entirely using
    /// [`Command::env_clear`] or for a single key using [`Command::env_remove`].
    ///
    /// Note that environment variable names are case-insensitive (but
    /// case-preserving) on Windows and case-sensitive on all other platforms.
    ///
    /// # Examples
    ///
    /// ```no_run
    /// use std::process::Command;
    ///
    /// Command::new("ls")
    ///     .env("PATH", "/bin")
    ///     .spawn()
    ///     .expect("ls command failed to start");
    /// ```
    #[stable(feature = "process", since = "1.0.0")]
    pub fn env<K, V>(&mut self, key: K, val: V) -> &mut Command
    where
        K: AsRef<OsStr>,
        V: AsRef<OsStr>,
    {
        self.inner.env_mut().set(key.as_ref(), val.as_ref());
        self
    }

    /// Inserts or updates multiple explicit environment variable mappings.
    ///
    /// This method allows you to add multiple environment variable mappings to the spawned process
    /// or overwrite previously set values. You can use [`Command::env`] to set a single environment
    /// variable.
    ///
    /// Child processes will inherit environment variables from their parent process by default.
    /// Environment variables explicitly set using [`Command::envs`] take precedence over inherited
    /// variables. You can disable environment variable inheritance entirely using
    /// [`Command::env_clear`] or for a single key using [`Command::env_remove`].
    ///
    /// Note that environment variable names are case-insensitive (but case-preserving) on Windows
    /// and case-sensitive on all other platforms.
    ///
    /// # Examples
    ///
    /// ```no_run
    /// use std::process::{Command, Stdio};
    /// use std::env;
    /// use std::collections::HashMap;
    ///
    /// let filtered_env : HashMap<String, String> =
    ///     env::vars().filter(|&(ref k, _)|
    ///         k == "TERM" || k == "TZ" || k == "LANG" || k == "PATH"
    ///     ).collect();
    ///
    /// Command::new("printenv")
    ///     .stdin(Stdio::null())
    ///     .stdout(Stdio::inherit())
    ///     .env_clear()
    ///     .envs(&filtered_env)
    ///     .spawn()
    ///     .expect("printenv failed to start");
    /// ```
    #[stable(feature = "command_envs", since = "1.19.0")]
    pub fn envs<I, K, V>(&mut self, vars: I) -> &mut Command
    where
        I: IntoIterator<Item = (K, V)>,
        K: AsRef<OsStr>,
        V: AsRef<OsStr>,
    {
        for (ref key, ref val) in vars {
            self.inner.env_mut().set(key.as_ref(), val.as_ref());
        }
        self
    }

    /// Removes an explicitly set environment variable and prevents inheriting it from a parent
    /// process.
    ///
    /// This method will remove the explicit value of an environment variable set via
    /// [`Command::env`] or [`Command::envs`]. In addition, it will prevent the spawned child
    /// process from inheriting that environment variable from its parent process.
    ///
    /// After calling [`Command::env_remove`], the value associated with its key from
    /// [`Command::get_envs`] will be [`None`].
    ///
    /// To clear all explicitly set environment variables and disable all environment variable
    /// inheritance, you can use [`Command::env_clear`].
    ///
    /// # Examples
    ///
    /// Prevent any inherited `GIT_DIR` variable from changing the target of the `git` command,
    /// while allowing all other variables, like `GIT_AUTHOR_NAME`.
    ///
    /// ```no_run
    /// use std::process::Command;
    ///
    /// Command::new("git")
    ///     .arg("commit")
    ///     .env_remove("GIT_DIR")
    ///     .spawn()?;
    /// # std::io::Result::Ok(())
    /// ```
    #[stable(feature = "process", since = "1.0.0")]
    pub fn env_remove<K: AsRef<OsStr>>(&mut self, key: K) -> &mut Command {
        self.inner.env_mut().remove(key.as_ref());
        self
    }

    /// Clears all explicitly set environment variables and prevents inheriting any parent process
    /// environment variables.
    ///
    /// This method will remove all explicitly added environment variables set via [`Command::env`]
    /// or [`Command::envs`]. In addition, it will prevent the spawned child process from inheriting
    /// any environment variable from its parent process.
    ///
    /// After calling [`Command::env_clear`], the iterator from [`Command::get_envs`] will be
    /// empty.
    ///
    /// You can use [`Command::env_remove`] to clear a single mapping.
    ///
    /// # Examples
    ///
    /// The behavior of `sort` is affected by `LANG` and `LC_*` environment variables.
    /// Clearing the environment makes `sort`'s behavior independent of the parent processes' language.
    ///
    /// ```no_run
    /// use std::process::Command;
    ///
    /// Command::new("sort")
    ///     .arg("file.txt")
    ///     .env_clear()
    ///     .spawn()?;
    /// # std::io::Result::Ok(())
    /// ```
    #[stable(feature = "process", since = "1.0.0")]
    pub fn env_clear(&mut self) -> &mut Command {
        self.inner.env_mut().clear();
        self
    }

    /// Sets the working directory for the child process.
    ///
    /// # Platform-specific behavior
    ///
    /// If the program path is relative (e.g., `"./script.sh"`), it's ambiguous
    /// whether it should be interpreted relative to the parent's working
    /// directory or relative to `current_dir`. The behavior in this case is
    /// platform specific and unstable, and it's recommended to use
    /// [`canonicalize`] to get an absolute program path instead.
    ///
    /// # Examples
    ///
    /// ```no_run
    /// use std::process::Command;
    ///
    /// Command::new("ls")
    ///     .current_dir("/bin")
    ///     .spawn()
    ///     .expect("ls command failed to start");
    /// ```
    ///
    /// [`canonicalize`]: crate::fs::canonicalize
    #[stable(feature = "process", since = "1.0.0")]
    pub fn current_dir<P: AsRef<Path>>(&mut self, dir: P) -> &mut Command {
        self.inner.cwd(dir.as_ref().as_ref());
        self
    }

    /// Configuration for the child process's standard input (stdin) handle.
    ///
    /// Defaults to [`inherit`] when used with [`spawn`] or [`status`], and
    /// defaults to [`piped`] when used with [`output`].
    ///
    /// [`inherit`]: Stdio::inherit
    /// [`piped`]: Stdio::piped
    /// [`spawn`]: Self::spawn
    /// [`status`]: Self::status
    /// [`output`]: Self::output
    ///
    /// # Examples
    ///
    /// ```no_run
    /// use std::process::{Command, Stdio};
    ///
    /// Command::new("ls")
    ///     .stdin(Stdio::null())
    ///     .spawn()
    ///     .expect("ls command failed to start");
    /// ```
    #[stable(feature = "process", since = "1.0.0")]
    pub fn stdin<T: Into<Stdio>>(&mut self, cfg: T) -> &mut Command {
        self.inner.stdin(cfg.into().0);
        self
    }

    /// Configuration for the child process's standard output (stdout) handle.
    ///
    /// Defaults to [`inherit`] when used with [`spawn`] or [`status`], and
    /// defaults to [`piped`] when used with [`output`].
    ///
    /// [`inherit`]: Stdio::inherit
    /// [`piped`]: Stdio::piped
    /// [`spawn`]: Self::spawn
    /// [`status`]: Self::status
    /// [`output`]: Self::output
    ///
    /// # Examples
    ///
    /// ```no_run
    /// use std::process::{Command, Stdio};
    ///
    /// Command::new("ls")
    ///     .stdout(Stdio::null())
    ///     .spawn()
    ///     .expect("ls command failed to start");
    /// ```
    #[stable(feature = "process", since = "1.0.0")]
    pub fn stdout<T: Into<Stdio>>(&mut self, cfg: T) -> &mut Command {
        self.inner.stdout(cfg.into().0);
        self
    }

    /// Configuration for the child process's standard error (stderr) handle.
    ///
    /// Defaults to [`inherit`] when used with [`spawn`] or [`status`], and
    /// defaults to [`piped`] when used with [`output`].
    ///
    /// [`inherit`]: Stdio::inherit
    /// [`piped`]: Stdio::piped
    /// [`spawn`]: Self::spawn
    /// [`status`]: Self::status
    /// [`output`]: Self::output
    ///
    /// # Examples
    ///
    /// ```no_run
    /// use std::process::{Command, Stdio};
    ///
    /// Command::new("ls")
    ///     .stderr(Stdio::null())
    ///     .spawn()
    ///     .expect("ls command failed to start");
    /// ```
    #[stable(feature = "process", since = "1.0.0")]
    pub fn stderr<T: Into<Stdio>>(&mut self, cfg: T) -> &mut Command {
        self.inner.stderr(cfg.into().0);
        self
    }

    /// Executes the command as a child process, returning a handle to it.
    ///
    /// By default, stdin, stdout and stderr are inherited from the parent.
    ///
    /// # Errors
    ///
    /// This method returns an [`io::Error`] if the child process could not be
    /// spawned. Common reasons include:
    ///
    /// * the program could not be found (for example, it does not exist, or,
    ///   when given a bare name, it is not present in the `PATH`);
    /// * the current process does not have permission to execute the program
    ///   (for example, the file is not marked executable, or execution is
    ///   denied by a security policy such as `seccomp`);
    /// * the operating system could not create the new process because of
    ///   resource exhaustion (for example, a limit on the number of processes
    ///   was reached).
    ///
    /// An error is only returned for failures that occur while the child is
    /// being spawned. Once the child has started successfully, anything that
    /// happens to it afterwards — including being terminated by a signal — is
    /// reported through its [`ExitStatus`] rather than as an error from the
    /// spawning method.
    ///
    /// # Examples
    ///
    /// ```no_run
    /// use std::process::Command;
    ///
    /// Command::new("ls")
    ///     .spawn()
    ///     .expect("ls command failed to start");
    /// ```
    #[stable(feature = "process", since = "1.0.0")]
    pub fn spawn(&mut self) -> io::Result<Child> {
        self.inner.spawn(imp::Stdio::Inherit, true).map(Child::from_inner)
    }

    /// Executes the command as a child process, waiting for it to finish and
    /// collecting all of its output.
    ///
    /// By default, stdout and stderr are captured (and used to provide the
    /// resulting output). Stdin is not inherited from the parent and any
    /// attempt by the child process to read from the stdin stream will result
    /// in the stream immediately closing.
    ///
    /// # Errors
    ///
    /// Like [`spawn`], this method returns an [`io::Error`] if the child
    /// process could not be spawned; see [`spawn`] for the common reasons. It
    /// may also return an error if reading the child's output or waiting on the
    /// child fails.
    ///
    /// Note that this method does **not** return an error if the child runs and
    /// then exits unsuccessfully, or is terminated by a signal. In those cases
    /// it still returns [`Ok`], and the outcome is reflected in the
    /// [`ExitStatus`] stored in the returned [`Output`].
    ///
    /// [`spawn`]: Command::spawn
    ///
    /// # Examples
    ///
    /// ```should_panic
    /// use std::process::Command;
    /// use std::io::{self, Write};
    /// let output = Command::new("/bin/cat")
    ///     .arg("file.txt")
    ///     .output()?;
    ///
    /// println!("status: {}", output.status);
    /// io::stdout().write_all(&output.stdout)?;
    /// io::stderr().write_all(&output.stderr)?;
    ///
    /// assert!(output.status.success());
    /// # io::Result::Ok(())
    /// ```
    #[stable(feature = "process", since = "1.0.0")]
    pub fn output(&mut self) -> io::Result<Output> {
        let (status, stdout, stderr) = imp::output(&mut self.inner)?;
        Ok(Output { status: ExitStatus(status), stdout, stderr })
    }

    /// Executes a command as a child process, waiting for it to finish and
    /// collecting its status.
    ///
    /// By default, stdin, stdout and stderr are inherited from the parent.
    ///
    /// # Errors
    ///
    /// Like [`spawn`], this method returns an [`io::Error`] if the child
    /// process could not be spawned; see [`spawn`] for the common reasons. It
    /// may also return an error if waiting on the child fails.
    ///
    /// Note that this method does **not** return an error if the child runs and
    /// then exits unsuccessfully, or is terminated by a signal. In those cases
    /// it still returns [`Ok`], and the outcome is reflected in the returned
    /// [`ExitStatus`].
    ///
    /// [`spawn`]: Command::spawn
    ///
    /// # Examples
    ///
    /// ```should_panic
    /// use std::process::Command;
    ///
    /// let status = Command::new("/bin/cat")
    ///     .arg("file.txt")
    ///     .status()
    ///     .expect("failed to execute process");
    ///
    /// println!("process finished with: {status}");
    ///
    /// assert!(status.success());
    /// ```
    #[stable(feature = "process", since = "1.0.0")]
    pub fn status(&mut self) -> io::Result<ExitStatus> {
        self.inner
            .spawn(imp::Stdio::Inherit, true)
            .map(Child::from_inner)
            .and_then(|mut p| p.wait())
    }

    /// Returns the path to the program that was given to [`Command::new`].
    ///
    /// # Examples
    ///
    /// ```
    /// use std::process::Command;
    ///
    /// let cmd = Command::new("echo");
    /// assert_eq!(cmd.get_program(), "echo");
    /// ```
    #[must_use]
    #[stable(feature = "command_access", since = "1.57.0")]
    pub fn get_program(&self) -> &OsStr {
        self.inner.get_program()
    }

    /// Returns an iterator of the arguments that will be passed to the program.
    ///
    /// This does not include the path to the program as the first argument;
    /// it only includes the arguments specified with [`Command::arg`] and
    /// [`Command::args`].
    ///
    /// # Examples
    ///
    /// ```
    /// use std::ffi::OsStr;
    /// use std::process::Command;
    ///
    /// let mut cmd = Command::new("echo");
    /// cmd.arg("first").arg("second");
    /// let args: Vec<&OsStr> = cmd.get_args().collect();
    /// assert_eq!(args, &["first", "second"]);
    /// ```
    #[stable(feature = "command_access", since = "1.57.0")]
    pub fn get_args(&self) -> CommandArgs<'_> {
        CommandArgs { inner: self.inner.get_args() }
    }

    /// Returns an iterator of the environment variables explicitly set for the child process.
    ///
    /// Environment variables explicitly set using [`Command::env`], [`Command::envs`], and
    /// [`Command::env_remove`] can be retrieved with this method.
    ///
    /// Note that this output does not include environment variables inherited from the parent
    /// process. To see the full list of environment variables, including those inherited from the
    /// parent process, use [`Command::get_resolved_envs`].
    ///
    /// Each element is a tuple key/value pair `(&OsStr, Option<&OsStr>)`. A [`None`] value
    /// indicates its key was explicitly removed via [`Command::env_remove`]. The associated key for
    /// the [`None`] value will no longer inherit from its parent process.
    ///
    /// An empty iterator can indicate that no explicit mappings were added or that
    /// [`Command::env_clear`] was called. After calling [`Command::env_clear`], the child process
    /// will not inherit any environment variables from its parent process.
    ///
    /// # Examples
    ///
    /// ```
    /// use std::ffi::OsStr;
    /// use std::process::Command;
    ///
    /// let mut cmd = Command::new("ls");
    /// cmd.env("TERM", "dumb").env_remove("TZ");
    /// let envs: Vec<(&OsStr, Option<&OsStr>)> = cmd.get_envs().collect();
    /// assert_eq!(envs, &[
    ///     (OsStr::new("TERM"), Some(OsStr::new("dumb"))),
    ///     (OsStr::new("TZ"), None)
    /// ]);
    /// ```
    #[stable(feature = "command_access", since = "1.57.0")]
    pub fn get_envs(&self) -> CommandEnvs<'_> {
        CommandEnvs { iter: self.inner.get_envs() }
    }

    /// Returns an iterator of the environment variables that will be set when the process is spawned.
    ///
    /// This returns the environment as it would be if the command were executed at the time of calling
    /// this method. The returned environment includes:
    /// - All inherited environment variables from the parent process (unless [`Command::env_clear`] was called)
    /// - All environment variables explicitly set via [`Command::env`] or [`Command::envs`]
    /// - Excluding any environment variables removed via [`Command::env_remove`]
    ///
    /// Note that the returned environment is a snapshot at the time this method is called and will not
    /// reflect any subsequent changes to the `Command` or the parent process's environment. Additionally,
    /// it will not reflect changes made in a `pre_exec` hook (on Unix platforms).
    ///
    /// Each element is a tuple `(OsString, OsString)` representing an environment variable key and value.
    ///
    /// # Examples
    ///
    /// ```
    /// #![feature(command_resolved_envs)]
    /// use std::process::Command;
    /// use std::ffi::{OsString, OsStr};
    /// use std::env;
    /// use std::collections::HashMap;
    ///
    /// let mut cmd = Command::new("ls");
    /// cmd.env("TZ", "UTC");
    /// unsafe { env::set_var("EDITOR", "vim"); }
    ///
    /// let resolved: HashMap<OsString, OsString> = cmd.get_resolved_envs().collect();
    /// assert_eq!(resolved.get(OsStr::new("TZ")), Some(&OsString::from("UTC")));
    /// assert_eq!(resolved.get(OsStr::new("EDITOR")), Some(&OsString::from("vim")));
    /// ```
    #[unstable(feature = "command_resolved_envs", issue = "149070")]
    pub fn get_resolved_envs(&self) -> CommandResolvedEnvs {
        self.inner.get_resolved_envs()
    }

    /// Returns the working directory for the child process.
    ///
    /// This returns [`None`] if the working directory will not be changed.
    ///
    /// # Examples
    ///
    /// ```
    /// use std::path::Path;
    /// use std::process::Command;
    ///
    /// let mut cmd = Command::new("ls");
    /// assert_eq!(cmd.get_current_dir(), None);
    /// cmd.current_dir("/bin");
    /// assert_eq!(cmd.get_current_dir(), Some(Path::new("/bin")));
    /// ```
    #[must_use]
    #[stable(feature = "command_access", since = "1.57.0")]
    pub fn get_current_dir(&self) -> Option<&Path> {
        self.inner.get_current_dir()
    }

    /// Returns whether the environment will be cleared for the child process.
    ///
    /// This returns `true` if [`Command::env_clear`] was called, and `false` otherwise.
    /// When `true`, the child process will not inherit any environment variables from
    /// its parent process.
    ///
    /// # Examples
    ///
    /// ```
    /// #![feature(command_resolved_envs)]
    /// use std::process::Command;
    ///
    /// let mut cmd = Command::new("ls");
    /// assert_eq!(cmd.get_env_clear(), false);
    ///
    /// cmd.env_clear();
    /// assert_eq!(cmd.get_env_clear(), true);
    /// ```
    #[must_use]
    #[unstable(feature = "command_resolved_envs", issue = "149070")]
    pub fn get_env_clear(&self) -> bool {
        self.inner.get_env_clear()
    }
}

#[stable(feature = "rust1", since = "1.0.0")]
impl fmt::Debug for Command {
    /// Format the program and arguments of a Command for display. Any
    /// non-utf8 data is lossily converted using the utf8 replacement
    /// character.
    ///
    /// The default format approximates a shell invocation of the program along with its
    /// arguments. It does not include most of the other command properties. The output is not guaranteed to work
    /// (e.g. due to lack of shell-escaping or differences in path resolution).
    /// On some platforms you can use [the alternate syntax] to show more fields.
    ///
    /// Note that the debug implementation is platform-specific.
    ///
    /// [the alternate syntax]: fmt#sign0
    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
        self.inner.fmt(f)
    }
}

impl AsInner<imp::Command> for Command {
    #[inline]
    fn as_inner(&self) -> &imp::Command {
        &self.inner
    }
}

impl AsInnerMut<imp::Command> for Command {
    #[inline]
    fn as_inner_mut(&mut self) -> &mut imp::Command {
        &mut self.inner
    }
}

/// An iterator over the command arguments.
///
/// This struct is created by [`Command::get_args`]. See its documentation for
/// more.
#[must_use = "iterators are lazy and do nothing unless consumed"]
#[stable(feature = "command_access", since = "1.57.0")]
#[derive(Debug)]
pub struct CommandArgs<'a> {
    inner: imp::CommandArgs<'a>,
}

#[stable(feature = "command_access", since = "1.57.0")]
impl<'a> Iterator for CommandArgs<'a> {
    type Item = &'a OsStr;
    fn next(&mut self) -> Option<&'a OsStr> {
        self.inner.next()
    }
    fn size_hint(&self) -> (usize, Option<usize>) {
        self.inner.size_hint()
    }
}

#[stable(feature = "command_access", since = "1.57.0")]
impl<'a> ExactSizeIterator for CommandArgs<'a> {
    fn len(&self) -> usize {
        self.inner.len()
    }
    fn is_empty(&self) -> bool {
        self.inner.is_empty()
    }
}

const fn assert_send<T: core::marker::Send>() {}
const fn assert_sync<T: core::marker::Sync>() {}

const _: () = assert_send::<CommandArgs<'static>>();
const _: () = assert_sync::<CommandArgs<'static>>();

/// An iterator over the command environment variables.
///
/// This struct is created by
/// [`Command::get_envs`][crate::process::Command::get_envs]. See its
/// documentation for more.
#[must_use = "iterators are lazy and do nothing unless consumed"]
#[stable(feature = "command_access", since = "1.57.0")]
pub struct CommandEnvs<'a> {
    iter: imp::CommandEnvs<'a>,
}

#[stable(feature = "command_access", since = "1.57.0")]
impl<'a> Iterator for CommandEnvs<'a> {
    type Item = (&'a OsStr, Option<&'a OsStr>);

    fn next(&mut self) -> Option<Self::Item> {
        self.iter.next()
    }

    fn size_hint(&self) -> (usize, Option<usize>) {
        self.iter.size_hint()
    }
}

#[stable(feature = "command_access", since = "1.57.0")]
impl<'a> ExactSizeIterator for CommandEnvs<'a> {
    fn len(&self) -> usize {
        self.iter.len()
    }

    fn is_empty(&self) -> bool {
        self.iter.is_empty()
    }
}

#[stable(feature = "command_access", since = "1.57.0")]
impl<'a> fmt::Debug for CommandEnvs<'a> {
    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
        self.iter.fmt(f)
    }
}

#[unstable(feature = "command_resolved_envs", issue = "149070")]
pub use imp::CommandResolvedEnvs;

/// The output of a finished process.
///
/// This is returned in a Result by either the [`output`] method of a
/// [`Command`], or the [`wait_with_output`] method of a [`Child`]
/// process.
///
/// [`output`]: Command::output
/// [`wait_with_output`]: Child::wait_with_output
#[derive(PartialEq, Eq, Clone)]
#[stable(feature = "process", since = "1.0.0")]
pub struct Output {
    /// The status (exit code) of the process.
    #[stable(feature = "process", since = "1.0.0")]
    pub status: ExitStatus,
    /// The data that the process wrote to stdout.
    #[stable(feature = "process", since = "1.0.0")]
    pub stdout: Vec<u8>,
    /// The data that the process wrote to stderr.
    #[stable(feature = "process", since = "1.0.0")]
    pub stderr: Vec<u8>,
}

impl Output {
    /// Returns an error if a nonzero exit status was received.
    ///
    /// If the [`Command`] exited successfully,
    /// `self` is returned.
    ///
    /// This is equivalent to calling [`exit_ok`](ExitStatus::exit_ok)
    /// on [`Output.status`](Output::status).
    ///
    /// Note that this will throw away the [`Output::stderr`] field in the error case.
    /// If the child process outputs useful informantion to stderr, you can:
    /// * Use `cmd.stderr(Stdio::inherit())` to forward the
    ///   stderr child process to the parent's stderr,
    ///   usually printing it to console where the user can see it.
    ///   This is usually correct for command-line applications.
    /// * Capture `stderr` using a custom error type.
    ///   This is usually correct for libraries.
    ///
    /// # Examples
    ///
    // Ferrocene annotation: QNX does not have the binaries
    /// ```ignore-qnx
    /// # #![allow(unused_features)]
    /// #![feature(exit_status_error)]
    /// # #[cfg(all(unix, not(target_os = "android"), not(all(target_vendor = "apple", not(target_os = "macos")))))] {
    /// use std::process::Command;
    /// assert!(Command::new("false").output().unwrap().exit_ok().is_err());
    /// # }
    /// ```
    #[unstable(feature = "exit_status_error", issue = "84908")]
    pub fn exit_ok(self) -> Result<Self, ExitStatusError> {
        self.status.exit_ok()?;
        Ok(self)
    }
}

// If either stderr or stdout are valid utf8 strings it prints the valid
// strings, otherwise it prints the byte sequence instead
#[stable(feature = "process_output_debug", since = "1.7.0")]
impl fmt::Debug for Output {
    fn fmt(&self, fmt: &mut fmt::Formatter<'_>) -> fmt::Result {
        let stdout_utf8 = str::from_utf8(&self.stdout);
        let stdout_debug: &dyn fmt::Debug = match stdout_utf8 {
            Ok(ref s) => s,
            Err(_) => &self.stdout,
        };

        let stderr_utf8 = str::from_utf8(&self.stderr);
        let stderr_debug: &dyn fmt::Debug = match stderr_utf8 {
            Ok(ref s) => s,
            Err(_) => &self.stderr,
        };

        fmt.debug_struct("Output")
            .field("status", &self.status)
            .field("stdout", stdout_debug)
            .field("stderr", stderr_debug)
            .finish()
    }
}

/// Describes what to do with a standard I/O stream for a child process when
/// passed to the [`stdin`], [`stdout`], and [`stderr`] methods of [`Command`].
///
/// [`stdin`]: Command::stdin
/// [`stdout`]: Command::stdout
/// [`stderr`]: Command::stderr
#[stable(feature = "process", since = "1.0.0")]
pub struct Stdio(imp::Stdio);

impl Stdio {
    /// A new pipe should be arranged to connect the parent and child processes.
    ///
    /// # Examples
    ///
    /// With stdout:
    ///
    /// ```no_run
    /// use std::process::{Command, Stdio};
    ///
    /// let output = Command::new("echo")
    ///     .arg("Hello, world!")
    ///     .stdout(Stdio::piped())
    ///     .output()
    ///     .expect("Failed to execute command");
    ///
    /// assert_eq!(String::from_utf8_lossy(&output.stdout), "Hello, world!\n");
    /// // Nothing echoed to console
    /// ```
    ///
    /// With stdin:
    ///
    /// ```no_run
    /// use std::io::Write;
    /// use std::process::{Command, Stdio};
    ///
    /// let mut child = Command::new("rev")
    ///     .stdin(Stdio::piped())
    ///     .stdout(Stdio::piped())
    ///     .spawn()
    ///     .expect("Failed to spawn child process");
    ///
    /// let mut stdin = child.stdin.take().expect("Failed to open stdin");
    /// std::thread::spawn(move || {
    ///     stdin.write_all("Hello, world!".as_bytes()).expect("Failed to write to stdin");
    /// });
    ///
    /// let output = child.wait_with_output().expect("Failed to read stdout");
    /// assert_eq!(String::from_utf8_lossy(&output.stdout), "!dlrow ,olleH");
    /// ```
    ///
    /// Writing more than a pipe buffer's worth of input to stdin without also reading
    /// stdout and stderr at the same time may cause a deadlock.
    /// This is an issue when running any program that doesn't guarantee that it reads
    /// its entire stdin before writing more than a pipe buffer's worth of output.
    /// The size of a pipe buffer varies on different targets.
    ///
    #[must_use]
    #[stable(feature = "process", since = "1.0.0")]
    pub fn piped() -> Stdio {
        Stdio(imp::Stdio::MakePipe)
    }

    /// The child inherits from the corresponding parent descriptor.
    ///
    /// # Examples
    ///
    /// With stdout:
    ///
    /// ```no_run
    /// use std::process::{Command, Stdio};
    ///
    /// let output = Command::new("echo")
    ///     .arg("Hello, world!")
    ///     .stdout(Stdio::inherit())
    ///     .output()
    ///     .expect("Failed to execute command");
    ///
    /// assert_eq!(String::from_utf8_lossy(&output.stdout), "");
    /// // "Hello, world!" echoed to console
    /// ```
    ///
    /// With stdin:
    ///
    /// ```no_run
    /// use std::process::{Command, Stdio};
    /// use std::io::{self, Write};
    ///
    /// let output = Command::new("rev")
    ///     .stdin(Stdio::inherit())
    ///     .stdout(Stdio::piped())
    ///     .output()?;
    ///
    /// print!("You piped in the reverse of: ");
    /// io::stdout().write_all(&output.stdout)?;
    /// # io::Result::Ok(())
    /// ```
    #[must_use]
    #[stable(feature = "process", since = "1.0.0")]
    pub fn inherit() -> Stdio {
        Stdio(imp::Stdio::Inherit)
    }

    /// This stream will be ignored. This is the equivalent of attaching the
    /// stream to `/dev/null`.
    ///
    /// # Examples
    ///
    /// With stdout:
    ///
    /// ```no_run
    /// use std::process::{Command, Stdio};
    ///
    /// let output = Command::new("echo")
    ///     .arg("Hello, world!")
    ///     .stdout(Stdio::null())
    ///     .output()
    ///     .expect("Failed to execute command");
    ///
    /// assert_eq!(String::from_utf8_lossy(&output.stdout), "");
    /// // Nothing echoed to console
    /// ```
    ///
    /// With stdin:
    ///
    /// ```no_run
    /// use std::process::{Command, Stdio};
    ///
    /// let output = Command::new("rev")
    ///     .stdin(Stdio::null())
    ///     .stdout(Stdio::piped())
    ///     .output()
    ///     .expect("Failed to execute command");
    ///
    /// assert_eq!(String::from_utf8_lossy(&output.stdout), "");
    /// // Ignores any piped-in input
    /// ```
    #[must_use]
    #[stable(feature = "process", since = "1.0.0")]
    pub fn null() -> Stdio {
        Stdio(imp::Stdio::Null)
    }

    /// Returns `true` if this requires [`Command`] to create a new pipe.
    ///
    /// # Example
    ///
    /// ```
    /// #![feature(stdio_makes_pipe)]
    /// use std::process::Stdio;
    ///
    /// let io = Stdio::piped();
    /// assert_eq!(io.makes_pipe(), true);
    /// ```
    #[unstable(feature = "stdio_makes_pipe", issue = "98288")]
    pub fn makes_pipe(&self) -> bool {
        matches!(self.0, imp::Stdio::MakePipe)
    }
}

impl FromInner<imp::Stdio> for Stdio {
    fn from_inner(inner: imp::Stdio) -> Stdio {
        Stdio(inner)
    }
}

#[stable(feature = "std_debug", since = "1.16.0")]
impl fmt::Debug for Stdio {
    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
        f.debug_struct("Stdio").finish_non_exhaustive()
    }
}

#[stable(feature = "stdio_from", since = "1.20.0")]
impl From<ChildStdin> for Stdio {
    /// Converts a [`ChildStdin`] into a [`Stdio`].
    ///
    /// # Examples
    ///
    /// `ChildStdin` will be converted to `Stdio` using `Stdio::from` under the hood.
    ///
    /// ```rust,no_run
    /// use std::process::{Command, Stdio};
    ///
    /// let reverse = Command::new("rev")
    ///     .stdin(Stdio::piped())
    ///     .spawn()
    ///     .expect("failed reverse command");
    ///
    /// let _echo = Command::new("echo")
    ///     .arg("Hello, world!")
    ///     .stdout(reverse.stdin.unwrap()) // Converted into a Stdio here
    ///     .output()
    ///     .expect("failed echo command");
    ///
    /// // "!dlrow ,olleH" echoed to console
    /// ```
    fn from(child: ChildStdin) -> Stdio {
        Stdio::from_inner(child.into_inner().into())
    }
}

#[stable(feature = "stdio_from", since = "1.20.0")]
impl From<ChildStdout> for Stdio {
    /// Converts a [`ChildStdout`] into a [`Stdio`].
    ///
    /// # Examples
    ///
    /// `ChildStdout` will be converted to `Stdio` using `Stdio::from` under the hood.
    ///
    /// ```rust,no_run
    /// use std::process::{Command, Stdio};
    ///
    /// let hello = Command::new("echo")
    ///     .arg("Hello, world!")
    ///     .stdout(Stdio::piped())
    ///     .spawn()
    ///     .expect("failed echo command");
    ///
    /// let reverse = Command::new("rev")
    ///     .stdin(hello.stdout.unwrap())  // Converted into a Stdio here
    ///     .output()
    ///     .expect("failed reverse command");
    ///
    /// assert_eq!(reverse.stdout, b"!dlrow ,olleH\n");
    /// ```
    fn from(child: ChildStdout) -> Stdio {
        Stdio::from_inner(child.into_inner().into())
    }
}

#[stable(feature = "stdio_from", since = "1.20.0")]
impl From<ChildStderr> for Stdio {
    /// Converts a [`ChildStderr`] into a [`Stdio`].
    ///
    /// # Examples
    ///
    /// ```rust,no_run
    /// use std::process::{Command, Stdio};
    ///
    /// let reverse = Command::new("rev")
    ///     .arg("non_existing_file.txt")
    ///     .stderr(Stdio::piped())
    ///     .spawn()
    ///     .expect("failed reverse command");
    ///
    /// let cat = Command::new("cat")
    ///     .arg("-")
    ///     .stdin(reverse.stderr.unwrap()) // Converted into a Stdio here
    ///     .output()
    ///     .expect("failed echo command");
    ///
    /// assert_eq!(
    ///     String::from_utf8_lossy(&cat.stdout),
    ///     "rev: cannot open non_existing_file.txt: No such file or directory\n"
    /// );
    /// ```
    fn from(child: ChildStderr) -> Stdio {
        Stdio::from_inner(child.into_inner().into())
    }
}

#[stable(feature = "stdio_from", since = "1.20.0")]
impl From<fs::File> for Stdio {
    /// Converts a [`File`](fs::File) into a [`Stdio`].
    ///
    /// # Examples
    ///
    /// `File` will be converted to `Stdio` using `Stdio::from` under the hood.
    ///
    /// ```rust,no_run
    /// use std::fs::File;
    /// use std::process::Command;
    ///
    /// // With the `foo.txt` file containing "Hello, world!"
    /// let file = File::open("foo.txt")?;
    ///
    /// let reverse = Command::new("rev")
    ///     .stdin(file)  // Implicit File conversion into a Stdio
    ///     .output()?;
    ///
    /// assert_eq!(reverse.stdout, b"!dlrow ,olleH");
    /// # std::io::Result::Ok(())
    /// ```
    fn from(file: fs::File) -> Stdio {
        Stdio::from_inner(file.into_inner().into())
    }
}

#[stable(feature = "stdio_from_stdio", since = "1.74.0")]
impl From<io::Stdout> for Stdio {
    /// Redirect command stdout/stderr to our stdout
    ///
    /// # Examples
    ///
    // Ferrocene annotation: QNX does not have a `whoami` binary
    /// ```rust,ignore-qnx
    /// #![feature(exit_status_error)]
    /// use std::io;
    /// use std::process::Command;
    ///
    /// # fn test() -> Result<(), Box<dyn std::error::Error>> {
    /// let output = Command::new("whoami")
    // "whoami" is a command which exists on both Unix and Windows,
    // and which succeeds, producing some stdout output but no stderr.
    ///     .stdout(io::stdout())
    ///     .output()?;
    /// output.status.exit_ok()?;
    /// assert!(output.stdout.is_empty());
    /// # Ok(())
    /// # }
    /// #
    /// # if cfg!(all(unix, not(target_os = "android"), not(all(target_vendor = "apple", not(target_os = "macos"))))) {
    /// #     test().unwrap();
    /// # }
    /// ```
    fn from(inherit: io::Stdout) -> Stdio {
        Stdio::from_inner(inherit.into())
    }
}

#[stable(feature = "stdio_from_stdio", since = "1.74.0")]
impl From<io::Stderr> for Stdio {
    /// Redirect command stdout/stderr to our stderr
    ///
    /// # Examples
    ///
    // Ferrocene annotation: QNX does not have a `whoami` binary
    /// ```rust,ignore-qnx
    /// #![feature(exit_status_error)]
    /// use std::io;
    /// use std::process::Command;
    ///
    /// # fn test() -> Result<(), Box<dyn std::error::Error>> {
    /// let output = Command::new("whoami")
    ///     .stdout(io::stderr())
    ///     .output()?;
    /// output.status.exit_ok()?;
    /// assert!(output.stdout.is_empty());
    /// # Ok(())
    /// # }
    /// #
    /// # if cfg!(all(unix, not(target_os = "android"), not(all(target_vendor = "apple", not(target_os = "macos"))))) {
    /// #     test().unwrap();
    /// # }
    /// ```
    fn from(inherit: io::Stderr) -> Stdio {
        Stdio::from_inner(inherit.into())
    }
}

#[stable(feature = "anonymous_pipe", since = "1.87.0")]
impl From<io::PipeWriter> for Stdio {
    fn from(pipe: io::PipeWriter) -> Self {
        Stdio::from_inner(pipe.into_inner().into())
    }
}

#[stable(feature = "anonymous_pipe", since = "1.87.0")]
impl From<io::PipeReader> for Stdio {
    fn from(pipe: io::PipeReader) -> Self {
        Stdio::from_inner(pipe.into_inner().into())
    }
}

/// Describes the result of a process after it has terminated.
///
/// This `struct` is used to represent the exit status or other termination of a child process.
/// Child processes are created via the [`Command`] struct and their exit
/// status is exposed through the [`status`] method, or the [`wait`] method
/// of a [`Child`] process.
///
/// An `ExitStatus` represents every possible disposition of a process.  On Unix this
/// is the **wait status**.  It is *not* simply an *exit status* (a value passed to `exit`).
///
/// For proper error reporting of failed processes, print the value of `ExitStatus` or
/// `ExitStatusError` using their implementations of [`Display`](crate::fmt::Display).
///
/// # Differences from `ExitCode`
///
/// [`ExitCode`] is intended for terminating the currently running process, via
/// the `Termination` trait, in contrast to `ExitStatus`, which represents the
/// termination of a child process. These APIs are separate due to platform
/// compatibility differences and their expected usage; it is not generally
/// possible to exactly reproduce an `ExitStatus` from a child for the current
/// process after the fact.
///
/// [`status`]: Command::status
/// [`wait`]: Child::wait
//
// We speak slightly loosely (here and in various other places in the stdlib docs) about `exit`
// vs `_exit`.  Naming of Unix system calls is not standardised across Unices, so terminology is a
// matter of convention and tradition.  For clarity we usually speak of `exit`, even when we might
// mean an underlying system call such as `_exit`.
#[derive(PartialEq, Eq, Clone, Copy, Debug)]
#[stable(feature = "process", since = "1.0.0")]
pub struct ExitStatus(imp::ExitStatus);

/// The default value is one which indicates successful completion.
#[stable(feature = "process_exitstatus_default", since = "1.73.0")]
impl Default for ExitStatus {
    fn default() -> Self {
        // Ideally this would be done by ExitCode::default().into() but that is complicated.
        ExitStatus::from_inner(imp::ExitStatus::default())
    }
}

impl ExitStatus {
    /// Was termination successful?  Returns a `Result`.
    ///
    /// # Examples
    ///
    /// ```
    /// #![feature(exit_status_error)]
    /// # if cfg!(all(unix, not(all(target_vendor = "apple", not(target_os = "macos"))))) {
    /// use std::process::Command;
    ///
    /// let status = Command::new("ls")
    ///     .arg("/dev/nonexistent")
    ///     .status()
    ///     .expect("ls could not be executed");
    ///
    /// println!("ls: {status}");
    /// status.exit_ok().expect_err("/dev/nonexistent could be listed!");
    /// # } // cfg!(unix)
    /// ```
    #[unstable(feature = "exit_status_error", issue = "84908")]
    pub fn exit_ok(&self) -> Result<(), ExitStatusError> {
        self.0.exit_ok().map_err(ExitStatusError)
    }

    /// Was termination successful? Signal termination is not considered a
    /// success, and success is defined as a zero exit status.
    ///
    /// # Examples
    ///
    /// ```rust,no_run
    /// use std::process::Command;
    ///
    /// let status = Command::new("mkdir")
    ///     .arg("projects")
    ///     .status()
    ///     .expect("failed to execute mkdir");
    ///
    /// if status.success() {
    ///     println!("'projects/' directory created");
    /// } else {
    ///     println!("failed to create 'projects/' directory: {status}");
    /// }
    /// ```
    #[must_use]
    #[stable(feature = "process", since = "1.0.0")]
    pub fn success(&self) -> bool {
        self.0.exit_ok().is_ok()
    }

    /// Returns the exit code of the process, if any.
    ///
    /// In Unix terms the return value is the **exit status**: the value passed to `exit`, if the
    /// process finished by calling `exit`.  Note that on Unix the exit status is truncated to 8
    /// bits, and that values that didn't come from a program's call to `exit` may be invented by the
    /// runtime system (often, for example, 255, 254, 127 or 126).
    ///
    /// On Unix, this will return `None` if the process was terminated by a signal.
    /// [`ExitStatusExt`](crate::os::unix::process::ExitStatusExt) is an
    /// extension trait for extracting any such signal, and other details, from the `ExitStatus`.
    ///
    /// # Examples
    ///
    /// ```no_run
    /// use std::process::Command;
    ///
    /// let status = Command::new("mkdir")
    ///     .arg("projects")
    ///     .status()
    ///     .expect("failed to execute mkdir");
    ///
    /// match status.code() {
    ///     Some(code) => println!("Exited with status code: {code}"),
    ///     None => println!("Process terminated by signal")
    /// }
    /// ```
    #[must_use]
    #[stable(feature = "process", since = "1.0.0")]
    pub fn code(&self) -> Option<i32> {
        self.0.code()
    }
}

impl AsInner<imp::ExitStatus> for ExitStatus {
    #[inline]
    fn as_inner(&self) -> &imp::ExitStatus {
        &self.0
    }
}

impl FromInner<imp::ExitStatus> for ExitStatus {
    fn from_inner(s: imp::ExitStatus) -> ExitStatus {
        ExitStatus(s)
    }
}

#[stable(feature = "process", since = "1.0.0")]
impl fmt::Display for ExitStatus {
    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
        self.0.fmt(f)
    }
}

/// Describes the result of a process after it has failed
///
/// Produced by the [`.exit_ok`](ExitStatus::exit_ok) method on [`ExitStatus`].
///
/// # Examples
///
// Ferrocene annotation: QNX does not have the binaries
/// ```ignore-qnx
/// #![feature(exit_status_error)]
/// # if cfg!(all(unix, not(target_os = "android"), not(all(target_vendor = "apple", not(target_os = "macos"))))) {
/// use std::process::{Command, ExitStatusError};
///
/// fn run(cmd: &str) -> Result<(), ExitStatusError> {
///     Command::new(cmd).status().unwrap().exit_ok()?;
///     Ok(())
/// }
///
/// run("true").unwrap();
/// run("false").unwrap_err();
/// # } // cfg!(unix)
/// ```
#[derive(PartialEq, Eq, Clone, Copy, Debug)]
#[unstable(feature = "exit_status_error", issue = "84908")]
// The definition of imp::ExitStatusError should ideally be such that
// Result<(), imp::ExitStatusError> has an identical representation to imp::ExitStatus.
pub struct ExitStatusError(imp::ExitStatusError);

#[unstable(feature = "exit_status_error", issue = "84908")]
#[doc(test(attr(allow(unused_features))))]
impl ExitStatusError {
    /// Reports the exit code, if applicable, from an `ExitStatusError`.
    ///
    /// In Unix terms the return value is the **exit status**: the value passed to `exit`, if the
    /// process finished by calling `exit`.  Note that on Unix the exit status is truncated to 8
    /// bits, and that values that didn't come from a program's call to `exit` may be invented by the
    /// runtime system (often, for example, 255, 254, 127 or 126).
    ///
    /// On Unix, this will return `None` if the process was terminated by a signal.  If you want to
    /// handle such situations specially, consider using methods from
    /// [`ExitStatusExt`](crate::os::unix::process::ExitStatusExt).
    ///
    /// If the process finished by calling `exit` with a nonzero value, this will return
    /// that exit status.
    ///
    /// If the error was something else, it will return `None`.
    ///
    /// If the process exited successfully (ie, by calling `exit(0)`), there is no
    /// `ExitStatusError`.  So the return value from `ExitStatusError::code()` is always nonzero.
    ///
    /// # Examples
    ///
    // Ferrocene annotation: QNX does not have the binaries
    /// ```ignore-qnx
    /// #![feature(exit_status_error)]
    /// # #[cfg(all(unix, not(target_os = "android"), not(all(target_vendor = "apple", not(target_os = "macos")))))] {
    /// use std::process::Command;
    ///
    /// let bad = Command::new("false").status().unwrap().exit_ok().unwrap_err();
    /// assert_eq!(bad.code(), Some(1));
    /// # } // #[cfg(unix)]
    /// ```
    #[must_use]
    pub fn code(&self) -> Option<i32> {
        self.code_nonzero().map(Into::into)
    }

    /// Reports the exit code, if applicable, from an `ExitStatusError`, as a [`NonZero`].
    ///
    /// This is exactly like [`code()`](Self::code), except that it returns a <code>[NonZero]<[i32]></code>.
    ///
    /// Plain `code`, returning a plain integer, is provided because it is often more convenient.
    /// The returned value from `code()` is indeed also nonzero; use `code_nonzero()` when you want
    /// a type-level guarantee of nonzeroness.
    ///
    /// # Examples
    ///
    // Ferrocene annotation: QNX does not have the binaries
    /// ```ignore-qnx
    /// #![feature(exit_status_error)]
    ///
    /// # if cfg!(all(unix, not(target_os = "android"), not(all(target_vendor = "apple", not(target_os = "macos"))))) {
    /// use std::num::NonZero;
    /// use std::process::Command;
    ///
    /// let bad = Command::new("false").status().unwrap().exit_ok().unwrap_err();
    /// assert_eq!(bad.code_nonzero().unwrap(), NonZero::new(1).unwrap());
    /// # } // cfg!(unix)
    /// ```
    #[must_use]
    pub fn code_nonzero(&self) -> Option<NonZero<i32>> {
        self.0.code()
    }

    /// Converts an `ExitStatusError` (back) to an `ExitStatus`.
    #[must_use]
    pub fn into_status(&self) -> ExitStatus {
        ExitStatus(self.0.into())
    }
}

#[unstable(feature = "exit_status_error", issue = "84908")]
impl From<ExitStatusError> for ExitStatus {
    fn from(error: ExitStatusError) -> Self {
        Self(error.0.into())
    }
}

#[unstable(feature = "exit_status_error", issue = "84908")]
impl fmt::Display for ExitStatusError {
    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
        write!(f, "process exited unsuccessfully: {}", self.into_status())
    }
}

#[unstable(feature = "exit_status_error", issue = "84908")]
impl crate::error::Error for ExitStatusError {}

/// This type represents the status code the current process can return
/// to its parent under normal termination.
///
/// `ExitCode` is intended to be consumed only by the standard library (via
/// [`Termination::report()`]). For forwards compatibility with potentially
/// unusual targets, this type currently does not provide `Eq`, `Hash`, or
/// access to the raw value. This type does provide `PartialEq` for
/// comparison, but note that there may potentially be multiple failure
/// codes, some of which will _not_ compare equal to `ExitCode::FAILURE`.
/// The standard library provides the canonical `SUCCESS` and `FAILURE`
/// exit codes as well as `From<u8> for ExitCode` for constructing other
/// arbitrary exit codes.
///
/// # Portability
///
/// Numeric values used in this type don't have portable meanings, and
/// different platforms may mask different amounts of them.
///
/// For the platform's canonical successful and unsuccessful codes, see
/// the [`SUCCESS`] and [`FAILURE`] associated items.
///
/// [`SUCCESS`]: ExitCode::SUCCESS
/// [`FAILURE`]: ExitCode::FAILURE
///
/// # Differences from `ExitStatus`
///
/// `ExitCode` is intended for terminating the currently running process, via
/// the `Termination` trait, in contrast to [`ExitStatus`], which represents the
/// termination of a child process. These APIs are separate due to platform
/// compatibility differences and their expected usage; it is not generally
/// possible to exactly reproduce an `ExitStatus` from a child for the current
/// process after the fact.
///
/// # Examples
///
/// `ExitCode` can be returned from the `main` function of a crate, as it implements
/// [`Termination`]:
///
/// ```
/// use std::process::ExitCode;
/// # fn check_foo() -> bool { true }
///
/// fn main() -> ExitCode {
///     if !check_foo() {
///         return ExitCode::from(42);
///     }
///
///     ExitCode::SUCCESS
/// }
/// ```
#[derive(Clone, Copy, Debug, PartialEq)]
#[stable(feature = "process_exitcode", since = "1.61.0")]
pub struct ExitCode(imp::ExitCode);

#[stable(feature = "process_exitcode", since = "1.61.0")]
impl ExitCode {
    /// The canonical `ExitCode` for successful termination on this platform.
    ///
    /// Note that a `()`-returning `main` implicitly results in a successful
    /// termination, so there's no need to return this from `main` unless
    /// you're also returning other possible codes.
    #[stable(feature = "process_exitcode", since = "1.61.0")]
    pub const SUCCESS: ExitCode = ExitCode(imp::ExitCode::SUCCESS);

    /// The canonical `ExitCode` for unsuccessful termination on this platform.
    ///
    /// If you're only returning this and `SUCCESS` from `main`, consider
    /// instead returning `Err(_)` and `Ok(())` respectively, which will
    /// return the same codes (but will also `eprintln!` the error).
    #[stable(feature = "process_exitcode", since = "1.61.0")]
    pub const FAILURE: ExitCode = ExitCode(imp::ExitCode::FAILURE);

    /// Exit the current process with the given `ExitCode`.
    ///
    /// Note that this has the same caveats as [`process::exit()`][exit], namely that this function
    /// terminates the process immediately, so no destructors on the current stack or any other
    /// thread's stack will be run. Also see those docs for some important notes on interop with C
    /// code. If a clean shutdown is needed, it is recommended to simply return this ExitCode from
    /// the `main` function, as demonstrated in the [type documentation](#examples).
    ///
    /// # Differences from `process::exit()`
    ///
    /// `process::exit()` accepts any `i32` value as the exit code for the process; however, there
    /// are platforms that only use a subset of that value (see [`process::exit` platform-specific
    /// behavior][exit#platform-specific-behavior]). `ExitCode` exists because of this; only
    /// `ExitCode`s that are supported by a majority of our platforms can be created, so those
    /// problems don't exist (as much) with this method.
    ///
    /// # Examples
    ///
    /// ```
    /// #![feature(exitcode_exit_method)]
    /// # use std::process::ExitCode;
    /// # use std::fmt;
    /// # enum UhOhError { GenericProblem, Specific, WithCode { exit_code: ExitCode, _x: () } }
    /// # impl fmt::Display for UhOhError {
    /// #     fn fmt(&self, _: &mut fmt::Formatter<'_>) -> fmt::Result { unimplemented!() }
    /// # }
    /// // there's no way to gracefully recover from an UhOhError, so we just
    /// // print a message and exit
    /// fn handle_unrecoverable_error(err: UhOhError) -> ! {
    ///     eprintln!("UH OH! {err}");
    ///     let code = match err {
    ///         UhOhError::GenericProblem => ExitCode::FAILURE,
    ///         UhOhError::Specific => ExitCode::from(3),
    ///         UhOhError::WithCode { exit_code, .. } => exit_code,
    ///     };
    ///     code.exit_process()
    /// }
    /// ```
    #[unstable(feature = "exitcode_exit_method", issue = "97100")]
    pub fn exit_process(self) -> ! {
        exit(self.to_i32())
    }
}

impl ExitCode {
    // This is private/perma-unstable because ExitCode is opaque; we don't know that i32 will serve
    // all usecases, for example windows seems to use u32, unix uses the 8-15th bits of an i32, we
    // likely want to isolate users anything that could restrict the platform specific
    // representation of an ExitCode
    //
    // More info: https://internals.rust-lang.org/t/mini-pre-rfc-redesigning-process-exitstatus/5426
    /// Converts an `ExitCode` into an i32
    #[unstable(
        feature = "process_exitcode_internals",
        reason = "exposed only for libstd",
        issue = "none"
    )]
    #[inline]
    #[doc(hidden)]
    pub fn to_i32(self) -> i32 {
        self.0.as_i32()
    }
}

/// The default value is [`ExitCode::SUCCESS`]
#[stable(feature = "process_exitcode_default", since = "1.75.0")]
impl Default for ExitCode {
    fn default() -> Self {
        ExitCode::SUCCESS
    }
}

#[stable(feature = "process_exitcode", since = "1.61.0")]
impl From<u8> for ExitCode {
    /// Constructs an `ExitCode` from an arbitrary u8 value.
    fn from(code: u8) -> Self {
        ExitCode(imp::ExitCode::from(code))
    }
}

impl AsInner<imp::ExitCode> for ExitCode {
    #[inline]
    fn as_inner(&self) -> &imp::ExitCode {
        &self.0
    }
}

impl FromInner<imp::ExitCode> for ExitCode {
    fn from_inner(s: imp::ExitCode) -> ExitCode {
        ExitCode(s)
    }
}

impl Child {
    /// Forces the child process to exit. If the child has already exited, `Ok(())`
    /// is returned.
    ///
    /// The mapping to [`ErrorKind`]s is not part of the compatibility contract of the function.
    ///
    /// This is equivalent to sending a SIGKILL on Unix platforms.
    ///
    /// # Examples
    ///
    /// ```no_run
    /// use std::process::Command;
    ///
    /// let mut command = Command::new("yes");
    /// if let Ok(mut child) = command.spawn() {
    ///     child.kill().expect("command couldn't be killed");
    /// } else {
    ///     println!("yes command didn't start");
    /// }
    /// ```
    ///
    /// [`ErrorKind`]: io::ErrorKind
    /// [`InvalidInput`]: io::ErrorKind::InvalidInput
    #[stable(feature = "process", since = "1.0.0")]
    #[cfg_attr(not(test), rustc_diagnostic_item = "child_kill")]
    pub fn kill(&mut self) -> io::Result<()> {
        self.handle.kill()
    }

    /// Returns the OS-assigned process identifier associated with this child.
    ///
    /// # Examples
    ///
    /// ```no_run
    /// use std::process::Command;
    ///
    /// let mut command = Command::new("ls");
    /// if let Ok(child) = command.spawn() {
    ///     println!("Child's ID is {}", child.id());
    /// } else {
    ///     println!("ls command didn't start");
    /// }
    /// ```
    #[must_use]
    #[stable(feature = "process_id", since = "1.3.0")]
    #[cfg_attr(not(test), rustc_diagnostic_item = "child_id")]
    pub fn id(&self) -> u32 {
        self.handle.id()
    }

    /// Waits for the child to exit completely, returning the status that it
    /// exited with. This function will continue to have the same return value
    /// after it has been called at least once.
    ///
    /// The stdin handle to the child process, if any, will be closed
    /// before waiting. This helps avoid deadlock: it ensures that the
    /// child does not block waiting for input from the parent, while
    /// the parent waits for the child to exit.
    ///
    /// # Examples
    ///
    /// ```no_run
    /// use std::process::Command;
    ///
    /// let mut command = Command::new("ls");
    /// if let Ok(mut child) = command.spawn() {
    ///     child.wait().expect("command wasn't running");
    ///     println!("Child has finished its execution!");
    /// } else {
    ///     println!("ls command didn't start");
    /// }
    /// ```
    #[stable(feature = "process", since = "1.0.0")]
    pub fn wait(&mut self) -> io::Result<ExitStatus> {
        drop(self.stdin.take());
        self.handle.wait().map(ExitStatus)
    }

    /// Attempts to collect the exit status of the child if it has already
    /// exited.
    ///
    /// This function will not block the calling thread and will only
    /// check to see if the child process has exited or not. If the child has
    /// exited then on Unix the process ID is reaped. This function is
    /// guaranteed to repeatedly return a successful exit status so long as the
    /// child has already exited.
    ///
    /// If the child has exited, then `Ok(Some(status))` is returned. If the
    /// exit status is not available at this time then `Ok(None)` is returned.
    /// If an error occurs, then that error is returned.
    ///
    /// Note that unlike `wait`, this function will not attempt to drop stdin.
    ///
    /// # Examples
    ///
    /// ```no_run
    /// use std::process::Command;
    ///
    /// let mut child = Command::new("ls").spawn()?;
    ///
    /// match child.try_wait() {
    ///     Ok(Some(status)) => println!("exited with: {status}"),
    ///     Ok(None) => {
    ///         println!("status not ready yet, let's really wait");
    ///         let res = child.wait();
    ///         println!("result: {res:?}");
    ///     }
    ///     Err(e) => println!("error attempting to wait: {e}"),
    /// }
    /// # std::io::Result::Ok(())
    /// ```
    #[stable(feature = "process_try_wait", since = "1.18.0")]
    pub fn try_wait(&mut self) -> io::Result<Option<ExitStatus>> {
        Ok(self.handle.try_wait()?.map(ExitStatus))
    }

    /// Simultaneously waits for the child to exit and collect all remaining
    /// output on the stdout/stderr handles, returning an `Output`
    /// instance.
    ///
    /// The stdin handle to the child process, if any, will be closed
    /// before waiting. This helps avoid deadlock: it ensures that the
    /// child does not block waiting for input from the parent, while
    /// the parent waits for the child to exit.
    ///
    /// By default, stdin, stdout and stderr are inherited from the parent.
    /// In order to capture the output into this `Result<Output>` it is
    /// necessary to create new pipes between parent and child. Use
    /// `stdout(Stdio::piped())` or `stderr(Stdio::piped())`, respectively.
    ///
    /// # Examples
    ///
    /// ```should_panic
    /// use std::process::{Command, Stdio};
    ///
    /// let child = Command::new("/bin/cat")
    ///     .arg("file.txt")
    ///     .stdout(Stdio::piped())
    ///     .spawn()
    ///     .expect("failed to execute child");
    ///
    /// let output = child
    ///     .wait_with_output()
    ///     .expect("failed to wait on child");
    ///
    /// assert!(output.status.success());
    /// ```
    ///
    #[stable(feature = "process", since = "1.0.0")]
    pub fn wait_with_output(mut self) -> io::Result<Output> {
        drop(self.stdin.take());

        let (mut stdout, mut stderr) = (Vec::new(), Vec::new());
        match (self.stdout.take(), self.stderr.take()) {
            (None, None) => {}
            (Some(mut out), None) => {
                let res = out.read_to_end(&mut stdout);
                res.unwrap();
            }
            (None, Some(mut err)) => {
                let res = err.read_to_end(&mut stderr);
                res.unwrap();
            }
            (Some(out), Some(err)) => {
                let res = imp::read_output(out.inner, &mut stdout, err.inner, &mut stderr);
                res.unwrap();
            }
        }

        let status = self.wait()?;
        Ok(Output { status, stdout, stderr })
    }
}

/// Terminates the current process with the specified exit code.
///
/// This function will never return and will immediately terminate the current
/// process. The exit code is passed through to the underlying OS and will be
/// available for consumption by another process.
///
/// Note that because this function never returns, and that it terminates the
/// process, no destructors on the current stack or any other thread's stack
/// will be run. If a clean shutdown is needed it is recommended to only call
/// this function at a known point where there are no more destructors left
/// to run; or, preferably, simply return a type implementing [`Termination`]
/// (such as [`ExitCode`] or `Result`) from the `main` function and avoid this
/// function altogether:
///
/// ```
/// # use std::io::Error as MyError;
/// fn main() -> Result<(), MyError> {
///     // ...
///     Ok(())
/// }
/// ```
///
/// In its current implementation, this function will execute exit handlers registered with `atexit`
/// as well as other platform-specific exit handlers (e.g. `fini` sections of ELF shared objects).
/// This means that Rust requires that all exit handlers are safe to execute at any time. In
/// particular, if an exit handler cleans up some state that might be concurrently accessed by other
/// threads, it is required that the exit handler performs suitable synchronization with those
/// threads. (The alternative to this requirement would be to not run exit handlers at all, which is
/// considered undesirable. Note that returning from `main` also calls `exit`, so making `exit` an
/// unsafe operation is not an option.)
///
/// ## Platform-specific behavior
///
/// **Unix**: On Unix-like platforms, it is unlikely that all 32 bits of `exit`
/// will be visible to a parent process inspecting the exit code. On most
/// Unix-like platforms, only the eight least-significant bits are considered.
///
/// For example, the exit code for this example will be `0` on Linux, but `256`
/// on Windows:
///
/// ```no_run
/// use std::process;
///
/// process::exit(0x0100);
/// ```
///
/// ### Safe interop with C code
///
/// On Unix, this function is currently implemented using the `exit` C function [`exit`][C-exit]. As
/// of C23, the C standard does not permit multiple threads to call `exit` concurrently. Rust
/// mitigates this with a lock, but if C code calls `exit`, that can still cause undefined behavior.
/// Note that returning from `main` is equivalent to calling `exit`.
///
/// Therefore, it is undefined behavior to have two concurrent threads perform the following
/// without synchronization:
/// - One thread calls Rust's `exit` function or returns from Rust's `main` function
/// - Another thread calls the C function `exit` or `quick_exit`, or returns from C's `main` function
///
/// Note that if a binary contains multiple copies of the Rust runtime (e.g., when combining
/// multiple `cdylib` or `staticlib`), they each have their own separate lock, so from the
/// perspective of code running in one of the Rust runtimes, the "outside" Rust code is basically C
/// code, and concurrent `exit` again causes undefined behavior.
///
/// Individual C implementations might provide more guarantees than the standard and permit concurrent
/// calls to `exit`; consult the documentation of your C implementation for details.
///
/// For some of the on-going discussion to make `exit` thread-safe in C, see:
/// - [Rust issue #126600](https://github.com/rust-lang/rust/issues/126600)
/// - [Austin Group Bugzilla (for POSIX)](https://austingroupbugs.net/view.php?id=1845)
/// - [GNU C library Bugzilla](https://sourceware.org/bugzilla/show_bug.cgi?id=31997)
///
/// [C-exit]: https://en.cppreference.com/w/c/program/exit
#[stable(feature = "rust1", since = "1.0.0")]
#[cfg_attr(not(test), rustc_diagnostic_item = "process_exit")]
pub fn exit(code: i32) -> ! {
    crate::rt::cleanup();
    crate::sys::exit::exit(code)
}

/// Terminates the process in an abnormal fashion.
///
/// The function will never return and will immediately terminate the current
/// process in a platform specific "abnormal" manner. As a consequence,
/// no destructors on the current stack or any other thread's stack
/// will be run, Rust IO buffers (eg, from `BufWriter`) will not be flushed,
/// and C stdio buffers will (on most platforms) not be flushed.
///
/// This is in contrast to the default behavior of [`panic!`] which unwinds
/// the current thread's stack and calls all destructors.
/// When `panic="abort"` is set, either as an argument to `rustc` or in a
/// crate's Cargo.toml, [`panic!`] and `abort` are similar. However,
/// [`panic!`] will still call the [panic hook] while `abort` will not.
///
/// If a clean shutdown is needed it is recommended to only call
/// this function at a known point where there are no more destructors left
/// to run.
///
/// The process's termination will be similar to that from the C `abort()`
/// function.  On Unix, the process will terminate with signal `SIGABRT`, which
/// typically means that the shell prints "Aborted".
///
/// # Examples
///
/// ```no_run
/// use std::process;
///
/// fn main() {
///     println!("aborting");
///
///     process::abort();
///
///     // execution never gets here
/// }
/// ```
///
/// The `abort` function terminates the process, so the destructor will not
/// get run on the example below:
///
/// ```no_run
/// use std::process;
///
/// struct HasDrop;
///
/// impl Drop for HasDrop {
///     fn drop(&mut self) {
///         println!("This will never be printed!");
///     }
/// }
///
/// fn main() {
///     let _x = HasDrop;
///     process::abort();
///     // the destructor implemented for HasDrop will never get run
/// }
/// ```
///
/// [panic hook]: crate::panic::set_hook
#[stable(feature = "process_abort", since = "1.17.0")]
#[cold]
#[cfg_attr(not(test), rustc_diagnostic_item = "process_abort")]
#[cfg_attr(miri, track_caller)] // even without panics, this helps for Miri backtraces
pub fn abort() -> ! {
    crate::sys::abort_internal();
}

#[doc(inline)]
#[unstable(feature = "abort_immediate", issue = "154601")]
pub use core::process::abort_immediate;

/// Returns the OS-assigned process identifier associated with this process.
///
/// # Examples
///
/// ```no_run
/// use std::process;
///
/// println!("My pid is {}", process::id());
/// ```
#[must_use]
#[stable(feature = "getpid", since = "1.26.0")]
pub fn id() -> u32 {
    imp::getpid()
}

/// A trait for implementing arbitrary return types in the `main` function.
///
/// The C-main function only supports returning integers.
/// So, every type implementing the `Termination` trait has to be converted
/// to an integer.
///
/// The default implementations are returning `libc::EXIT_SUCCESS` to indicate
/// a successful execution. In case of a failure, `libc::EXIT_FAILURE` is returned.
///
/// Because different runtimes have different specifications on the return value
/// of the `main` function, this trait is likely to be available only on
/// standard library's runtime for convenience. Other runtimes are not required
/// to provide similar functionality.
#[cfg_attr(not(any(test, doctest)), lang = "termination")]
#[stable(feature = "termination_trait_lib", since = "1.61.0")]
#[rustc_on_unimplemented(on(
    cause = "MainFunctionType",
    message = "`main` has invalid return type `{Self}`",
    label = "`main` can only return types that implement `{This}`"
))]
pub trait Termination {
    /// Is called to get the representation of the value as status code.
    /// This status code is returned to the operating system.
    #[stable(feature = "termination_trait_lib", since = "1.61.0")]
    fn report(self) -> ExitCode;
}

#[stable(feature = "termination_trait_lib", since = "1.61.0")]
impl Termination for () {
    #[inline]
    fn report(self) -> ExitCode {
        ExitCode::SUCCESS
    }
}

#[stable(feature = "termination_trait_lib", since = "1.61.0")]
impl Termination for ! {
    fn report(self) -> ExitCode {
        self
    }
}

#[stable(feature = "termination_trait_lib", since = "1.61.0")]
impl Termination for Infallible {
    fn report(self) -> ExitCode {
        match self {}
    }
}

#[stable(feature = "termination_trait_lib", since = "1.61.0")]
impl Termination for ExitCode {
    #[inline]
    fn report(self) -> ExitCode {
        self
    }
}

#[stable(feature = "termination_trait_lib", since = "1.61.0")]
impl<T: Termination, E: fmt::Debug> Termination for Result<T, E> {
    fn report(self) -> ExitCode {
        match self {
            Ok(val) => val.report(),
            Err(err) => {
                io::attempt_print_to_stderr(format_args_nl!("Error: {err:?}"));
                ExitCode::FAILURE
            }
        }
    }
}