CommandThis connector executes a command and communicates the I/O.
Note that the command will be executed as the same O/S user account as the application. The executed command will not inherit any shell environment, so if that is necessary, you should wrap your target command in a shell script. Call that shell script from this plug-in.
In other words,
shell environmental variables (e.g. ${HOME}), and built-in commands (e.g. 'ls')
are not directly available. Instead, they may need to be wrapped in a shell
script that is executed within a shell. For example, ls -al will not work,
but /bin/sh -c 'ls -al' will launch a Bourne shell to execute the ls -al
command and produce a long-form file listing. Notice the extra level of
quoting around the ls -al which needs to be considered a single argument to
the -c flag. The spaces embedded in this sub-command are later interpreted
by the shell.
For Unix-type systems, "/bin/sh -c" is the recommended shell execution. For Windows 95, use "command.exe /C", and for Windows NT, XP, or newer, use "cmd.exe /C"
The present working directory (PWD) of the command is set to the application's home directory.
If a command terminates (normally or because of an error), the "source.retry" property is used to determine how often to re-run the command. This can be used to repeatedly schedule a command that polls for some device status and terminates.