Fix potential nil pointer dereference when checking container state in Remove

[Nepomuk5665]

Summary

• Add nil check for inspected.State before accessing inspected.State.Running in the Remove function
• Prevents potential panic if Docker returns a container without State information

Problem

In pkg/compose/remove.go:67, the code accesses inspected.State.Running without first checking if inspected.State is nil:

if !inspected.State.Running || (options.Stop && s.dryRun) {

While Docker containers typically always have State populated, this is inconsistent with nil checks done elsewhere in the codebase and could cause a panic in edge cases (race conditions, unusual Docker states, or API changes).

Solution

Add a nil check consistent with the pattern used elsewhere in the codebase:

• pkg/compose/convergence.go:836 - ctr.State != nil && ctr.State.Status
• pkg/compose/convergence.go:862 - ctr.State != nil && ctr.State.Status
• pkg/compose/monitor.go:150 - inspect.State != nil && (inspect.State.Restarting || ...)
• pkg/compose/ps.go:70 - if inspect.State != nil {

The fix adds a nil check that treats containers with nil State as "stopped" (which is the safe default for the Remove operation).

[ndeloof]

add nolint:gogyclo on this function, we can look into refactoring to make code simpler in a follow-up PR

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!