Remote control was what expect-lite was born to do. In the early days of expect-lite, there was a need to log into an embedded linux machine and run some tests. There was also a need to log into different remote hosts with same script, so switching which remote host to test had to be easy, as easy as putting it on the command line.
Of course there are several different methods to access a remote host, and to that end, expect-lite supports telnet, ssh, both will username and password options, and a third method, ssh with keys, which requires no password, but the keys have to be set up ahead of time. These methods are selected by setting an environment variable EL_CONNECT_METHOD.
Over the years of using expect-lite I have found while handy to connect to the remote embedded system, it isn't always the best method. Often embedded systems don't have a full linux (or other OS) on them, after all it is by its nature a small system. So I have found that logging into my desktop linux system (aka localhost) provides a rich full set of utilities, and the script can then log into the embedded system (using a variable $IP to control which embedded system).
My .expect-literc which creates my environment looks like this:
Of course, I have set up ssh keys to my localhost, allowing expect-lite to login without using a password. This is easily done by using the configuration-only option of the installation script, install.sh, in /usr/share/doc/expect-lite (if you installed using the deb or rpm file), or in expect-lite.proj directory where ever you untarred the tarball.
$ cd <install.sh directory>
$ ./install.sh -c
Installing expect-lite version "4.6.0"
Configuration Only selected, skipping install steps 1,2,3
Fine print: the install script really only sets up the environment for bash shell (or sh, or ksh). It does not set up csh (or similar shells). I use bash, and that is the default shell for most distros today. expect-lite can and does work with csh, but you will need to adjust the environment using csh syntax
Before using the environment created by install.sh it is necessary to either a) relogin, or b) source your ~/.bashrc from the shell. Only then will the new environment take effect.
With this environment, expect-lite will log into the localhost using ssh keys (yellow arrow in diagram above), and then the script will log into the remote host, such as an embedded system (the blue arrow). For additional flexibility I use the *FORK command to create another session to the localhost linux system.
And switching back to the remote host with:
Now the script has access to the rich set of linux utilities, and the embedded system under test.
Of course, you don't have to use this method. expect-lite supports connecting directly to the remote host as I stated earlier. Either way works well, and for the script, it is just like being there.