`Stdio::none()` for Windows

Proposal

Problem statement

On Windows it's common to start processes without setting any stdio handles; in particular for GUI programs. This can be achieved in rust today but isn't obvious and requires unsafe.

Motivating examples or use cases

Command::new("notepad.exe")
        .arg("notes.txt")
        .stdin(unsafe { Stdio::from_raw_handle(std::ptr::null_mut()) })
        .stdout(unsafe { Stdio::from_raw_handle(std::ptr::null_mut()) })
        .stderr(unsafe { Stdio::from_raw_handle(std::ptr::null_mut()) })
        .spawn()?;

Unsafe aside, it doesn't really make sense to use from_raw_handle here because null is not a valid handle.

Solution sketch

// in std::os::windows::process

trait StdioExt {
    fn none() -> Stdio;
}
impl StdioExt for Stdio { ... }

Example

use std::os::windows::process::StdioExt;
Command::new("notepad.exe")
        .arg("notes.txt")
        .stdin(Stdio::none()),
        .stdout(Stdio::none()),
        .stderr(Stdio::none()),
        .spawn()?;

Alternatives

The status quo.

impl From<Option<OwnedHandle>> for Stdio where None means no handle.

.use_std_handles(false) to make it so we don't pass stdio handles to CreateProcess

We could make Stdio::null use null_mut() instead of opening the nul device but that would be a breaking change because the nul device can be successfully written to and read from (albeit it's a no-op in both cases) whereas a null handle cannot.

Links and related work

What happens now?

This issue contains an API change proposal (or ACP) and is part of the libs-api team feature lifecycle. Once this issue is filed, the libs-api team will review open proposals as capability becomes available. Current response times do not have a clear estimate, but may be up to several months.

Possible responses

The libs team may respond in various different ways. First, the team will consider the problem (this doesn't require any concrete solution or alternatives to have been proposed):

• We think this problem seems worth solving, and the standard library might be the right place to solve it.
• We think that this probably doesn't belong in the standard library.

Second, if there's a concrete solution:

• We think this specific solution looks roughly right, approved, you or someone else should implement this. (Further review will still happen on the subsequent implementation PR.)
• We're not sure this is the right solution, and the alternatives or other materials don't give us enough information to be sure about that. Here are some questions we have that aren't answered, or rough ideas about alternatives we'd want to see discussed.

Read the proposal's Windows process examples, solution sketch, and alternatives, then review the linked feature lifecycle for how an ACP is evaluated. No implementation file or test is identified; the work is ready only after the libs-api team reaches a decision on the API and implementation scope.