Shell is not executed for some containers

**Describe the bug**
Shell is not executed for some containers

**To Reproduce**
Steps to reproduce the behavior:
1. Pull `bitnami/openldap` image with `docker pull bitnami/openldap`
2. Launch it (`docker run --rm bitnami/openldap`)
3. Launch lazydocker
4. Select running openldap container and press `E` to get into container shell

**Expected behaviour**
You are dropped into container shell.

**Actual behaviour**
I see `id: cannot find name for user ID 1001` message.

This happens because image maintainer used `USER 1001` command [in the Dockerfile](https://github.com/bitnami/bitnami-docker-openldap/blob/master/2.6/debian-10/Dockerfile#L26) to set custom uid for the image. Obviously this user id is not present in `/etc/passwd` file - hence lazydocker command fails to detect shell.

I think lazydocker should try to detect preferred shell and if it fails to find one or execute one `/bin/sh` should be launched instead.

This approach will solve another shell-related problem (steps to reproduce below)
1. Pull `quay.io/keycloak/keycloak:18.0` image with `docker pull quay.io/keycloak/keycloak:18.0`
2. Launch it (`docker run --rm quay.io/keycloak/keycloak:18.0 start-dev`)
3. Launch lazydocker
4. Try to drop into shell by pressing `E` against keycloak container

You will see an error because Red Hat developers put `/sbin/nologin` (`/bin/false` seems to be another popular option in such cases though) as the shell for image user to tighten security. However, you still can directly exec `/bin/sh` and get access to the shell.

Also I should mention that I have an image without grep in it which also makes `exec shell` command to fail. Yes, it is a very specific image and I have custom command which simply execs `/bin/sh` against container but I would really like to see some bulletproof solution implemented in lazydocker.