Struct ManagedProcess 

pub struct ManagedProcess { /* private fields */ }

Full process lifecycle driven by a caller-owned reactor.

ManagedProcess preserves the blocking spawn semantics while allowing an application reactor to stay responsive: Core owns timeout/cancellation escalation, process-group signaling, pipe draining, overflow reporting, and waitpid reaping; the caller only routes readiness events and polls on Self::next_deadline.

Implementations

impl ManagedProcess

pub fn pid(&self) -> pid_t

Return the child PID.

The PID is captured at spawn time, so this remains available after the process has completed (unlike the running handle, which is consumed).

pub fn register_with_reactor( &mut self, reactor: &mut Reactor, ) -> Result<(), CoreError>

Register active child I/O descriptors with the caller’s reactor.

pub fn handle_reactor_event( &mut self, reactor: &mut Reactor, event: &Event, ) -> Result<(), CoreError>

Route one reactor event to the child’s I/O drain state.

pub fn stdout_paused(&self) -> bool

Return whether the stdout stream is paused on a full sink queue.

pub fn stderr_paused(&self) -> bool

Return whether the stderr stream is paused on a full sink queue.

pub fn resume_stdout( &mut self, reactor: &mut Reactor, ) -> Result<bool, CoreError>

Re-deliver the held stdout chunk (if any) and re-register the fd when the sink has room again. Returns true when the stream is resumed.

pub fn resume_stderr( &mut self, reactor: &mut Reactor, ) -> Result<bool, CoreError>

Re-deliver the held stderr chunk (if any) and re-register the fd when the sink has room again. Returns true when the stream is resumed.

pub fn request_cancel(&mut self)

Request cancellation using the daemon-owned policy from SpawnOptionsBuilder::cancel. Repeated requests are idempotent.

pub fn next_deadline(&self) -> Option<Instant>

Earliest time at which Self::poll_completion should run again.

A bounded reap tick is returned while the child is live, and exact timeout / TERM-to-KILL deadlines take precedence. None means the completion was already consumed.

pub fn poll_completion( &mut self, reactor: &mut Reactor, ) -> Result<Option<Output>, CoreError>

Advance timeout/cancellation, reap state, and completion.

Returns Ok(None) while work remains, the normal Output once the child is reaped and its pipes are drained, or EOVERFLOW when the configured combined output limit was exceeded on the fully-drained path. A forced-close (timeout/cancel with a wedged pipe) returns the partial output and the timed_out flag instead, matching blocking spawn.

pub fn resize_pty(&self, rows: u16, cols: u16) -> Result<(), CoreError>

Apply a new terminal window size to a pty-spawned child.

Only valid for a SpawnOptionsBuilder::pty spawn with an active stdout stream; callers typically follow this with a SIGWINCH to the child (or its foreground group) so the program can re-read the size.

Errors

• EINVAL: The spawn was not a pty spawn, the stream is already closed, or rows/cols is zero.
• ENOTTY: The pty master is unexpectedly not a terminal.

pub fn write_input(&self, bytes: &[u8]) -> Result<usize, CoreError>

Write bytes to a pty-spawned child’s stdin (the master end).

Only valid for a SpawnOptionsBuilder::pty spawn with an active stdout stream. See RunningProcess::write_input.

pub fn write_input_nonblock( &self, bytes: &[u8], ) -> Result<Option<usize>, CoreError>

Write bytes to a pty-spawned child’s stdin without blocking.

Only valid for a SpawnOptionsBuilder::pty spawn with an active stdout stream. See RunningProcess::write_input_nonblock.

Trait Implementations

impl Drop for ManagedProcess

fn drop(&mut self)

Executes the destructor for this type.

fn pin_drop(self: Pin<&mut Self>)

🔬This is a nightly-only experimental API. (pin_ergonomics)

Execute the destructor for this type, but different to Drop::drop, it requires self to be pinned.