In _cmd_deploy.py, before passing the OCI image to _docker_run, add a fallback for localhost:32000/... references. If an image is available from a registry on port 32000, use it as-is. Otherwise, strip the localhost:32000/ prefix and resolve the image from the local Docker image cache (e.g. localhost:32000/api-demo-server:integration → api-demo-server:integration). This lets charm integration tests written for real Juju (which pulls OCI resources from a local registry) run unchanged under jjx, where the local Docker image cache acts as the registry. Scope the translation to port 32000 only.
In _cmd_deploy.py, before passing the OCI image to
_docker_run, add a fallback forlocalhost:32000/...references. If an image is available from a registry on port 32000, use it as-is. Otherwise, strip thelocalhost:32000/prefix and resolve the image from the local Docker image cache (e.g.localhost:32000/api-demo-server:integration→api-demo-server:integration). This lets charm integration tests written for real Juju (which pulls OCI resources from a local registry) run unchanged under jjx, where the local Docker image cache acts as the registry. Scope the translation to port 32000 only.