fix: check container runtime is available in stop, restart and logs

run_detached and run_foreground both call check_available before they
touch the runtime, which routes a missing binary through
not_installed_error and produces an actionable message. stop, restart
and logs spawned the binary directly, so on a host without the
configured runtime they failed with a bare os error 2 that does not
name what was missing.

Author: Cintu07

crates/cli/src/local_runtime.rs

@@ -314,13 +314,15 @@ impl LocalRuntime {
     }
 
     pub fn stop(&self, instance_name: &str) -> Result<bool> {
+        Self::check_available(self.runtime)?;
         let name = self.container_name(instance_name);
         let removed_helix = self.remove_container(&name)?;
         let removed_disk_resources = self.remove_disk_resources(instance_name, false)?;
         Ok(removed_helix || removed_disk_resources)
     }
 
     pub fn restart(&self, instance_name: &str, config: &LocalInstanceConfig) -> Result<()> {
+        Self::check_available(self.runtime)?;
         if config.storage.is_disk() || config.storage.is_s3() {
             return self.run_detached(instance_name, config);
         }
@@ -341,6 +343,7 @@ impl LocalRuntime {
     }
 
     pub fn logs(&self, instance_name: &str, follow: bool) -> Result<()> {
+        Self::check_available(self.runtime)?;
         let name = self.container_name(instance_name);
         let mut command = self.runtime_command();
         command.arg("logs");