The shell

It is possible to run a UNIX shell under Emacs, and this allows any number of useful effects. The command M-x shell moves to a buffer named *shell* running a UNIX shell (creating it if necessary). Anything typed into this buffer is sent to the shell, just it would be outside of Emacs. Any output produced as a result of the commands sent to the shell is placed at the end of the shell buffer. Because the shell is running in an Emacs window, the contents of the shell can be edited and navigated freely, and the entire record of the input and output to the shell is available at all times. A few keys have slightly different-from-usual meanings in the shell buffer.

RET

sends whatever line the cursor is on to the shell and moves to the end of the shell buffer. Hence, one can repeat a command by placing the cursor anywhere in it and typing RET.

TAB

attempts to complete the immediately preceding file name.

C-c C-c

is the same as a single C-c outside Emacs.

C-c C-d

is the same as C-d (end-of-file) outside Emacs.

C-c C-z

is the same as C-z outside Emacs.

C-c C-u

kills the current line of input to the shell.

It is sometimes useful to run a single shell command over a region of text in a buffer.

M-|

prompts for a shell command and executes it, giving the current region as the standard input. If the M-| is preceded by C-u, the output of the command replaces the region. Otherwise, the output goes to a separate buffer. For example, to sort the lines in the current region, enter the command C-u M-| sort.