Don't setup the Image/ContainerEngine when calling a cmd with subcmds

There is no need to setup the image and container engine when calling
a command with subcommands since we only print a usage message.
e.g `podman`,`podman container`

This also allows the remote client to show the usage message on
these commands without a running endpoint. I added a test for this.

Signed-off-by: Paul Holzinger <paul.holzinger@web.de>

1 files changed, 11 insertions, 0 deletions

diff --git a/test/system/001-basic.bats b/test/system/001-basic.bats
index a5a3324fb..1d5eb066b 100644
--- a/test/system/001-basic.bats
+++ b/test/system/001-basic.bats
@@ -69,6 +69,17 @@ function setup() {
     is "$output" "Error: unknown flag: --remote" "podman version --remote"
 }
 
+# Check that just calling "podman-remote" prints the usage message even
+# without a running endpoint. Use "podman --remote" for this as this works the same.
+@test "podman-remote: check for command usage message without a running endpoint" {
+    if is_remote; then
+        skip "only applicable on a local run since this requires no endpoint"
+    fi
+
+    run_podman 125 --remote
+    is "$output" "Error: missing command 'podman COMMAND'" "podman remote show usage message without running endpoint"
+}
+
 # This is for development only; it's intended to make sure our timeout
 # in run_podman continues to work. This test should never run in production
 # because it will, by definition, fail.