[Bug]: Orphaned processes are never reaped (no --init support)

[Ninja3047]

I have done the following

• I have searched the existing issues
• If possible, I've reproduced the issue using the 'main' branch of this project

Steps to reproduce

container run --name reapro -d alpine:latest sleep 3600

# Create 5 orphans: parent forks child, parent exits, child sleeps 1s then exits
container exec reapro sh -c '
for i in 1 2 3 4 5; do
  sh -c "sh -c \"sleep 1\" &" &
  wait
done'

sleep 3

# All 5 are now permanent zombies
container exec reapro sh -c '
for p in /proc/[0-9]*/status; do
  grep -q "^State:.*Z" "$p" 2>/dev/null && echo "zombie: $(echo $p | cut -d/ -f3)"
done'

container rm -f reapro

Current behavior

All 5 orphaned processes become permanent zombies parented to PID 1. They are never reaped and accumulate over the container's lifetime.

zombie: 8
zombie: 10
zombie: 12
zombie: 14
zombie: 16

The user's entrypoint runs directly as PID 1 inside the container's PID namespace. It never calls wait(), so orphaned children that exit become permanent zombies.

Expected behavior

Orphaned processes should be reaped. Docker solves this with --init, which injects tini as PID 1 to reap children. container has no equivalent flag.

Environment

- OS: macOS 26.2 (25C56)
- Xcode: 26.2 (17C52)
- Container: container CLI version 0.9.0 (build: release)

Relevant log output

$ container exec reapro sh -c '
for p in /proc/[0-9]*/status; do
  grep -q "^State:.*Z" "$p" 2>/dev/null && echo "zombie: $(echo $p | cut -d/ -f3)"
done'
zombie: 8
zombie: 10
zombie: 12
zombie: 14
zombie: 16

Code of Conduct

• I agree to follow this project's Code of Conduct

[Ninja3047]

Probably related to #403

[Ronitsabhaya75]

@Ninja3047 I've investigated this issue. The zombie processes accumulate because the command running as PID 1 often doesn't have child reaping capabilities (e.g. a simple shell script or a binary not designed as init).

To fix this, I propose adding an --init flag to container run. When enabled, this flag will wrap the user's command with a lightweight init process.

lemme know if this make sense to you

[Ninja3047]

Yep, that makes sense and I think that's how docker handles it

[Ronitsabhaya75]

@Ninja3047 I'll pick up on this later this week after my tests are completed :)

[dcantah]

apple/containerization#537 we can support this here after this is in.

[Ronitsabhaya75]

apple/containerization#537 we can support this here after this is in.

Thank you so much danny for letting me know : )