Much of the information in this chapter involves integration between Emacs and the operating system. Emacs is most commonly a Unix editor, so forgive us for a bias in that direction. But we are happy to report that for users of GNU Emacs on other platforms, integration with the operating system is still available; you can use shell mode to run commands and can edit directories with Dired. There's no reason to leave Emacs no matter what your platform is.
5.1 Executing Commands in Shell Buffers
One of the most important features of Emacs is its ability to run a command shell in a buffer. Once you have started a shell buffer, you can do all of your normal command-line work within Emacs. What does this buy you?
• You don't have to leave Emacs to get a command prompt. If you want to print or compile a file that you're editing, you can do it immediately.
• You can use Emacs editing features to write your commands.
• You can use Emacs editing features to "back up" through your command list, copy an old command, modify it, and execute it again.
• You can save your shell buffer, keeping a transcript of your editing session—which automatically includes the output from every command that you ran. For debugging or remembering commands you run infrequently, this can be invaluable.
• You can copy output from commands into a file or into another command.
• You can save complex commands in a file and insert the file at the prompt, rather than retyping the command.
As you get used to working within Emacs, you will undoubtedly discover more and more ways to put shell mode to use.
In this section, we discuss shell mode. Later in this chapter, we discuss directory editing, printing, and calendar and diary features for doing simple time management in Emacs. Right now, we'll start with a simple variation on shell mode, a feature that lets you execute commands one at a time.
5.1.1 Running One Command at a Time
To run a command while you're in an Emacs session, type M-!. Emacs asks for the command you want to run. Type the command and press Enter. Emacs then opens a window called *Shell Command Output* where it displays the results of your command.
Type: M-!
Emacs prompts you for a command to execute.
Type: diff joyce joyce2
Emacs executes the diff command and puts the output into a *Shell Command Output* buffer.
Because the output from the diff command is in a buffer, you can edit it, save it, or do anything else you would like with it. Of course, if the operating system has no diff command or cannot access it for some reason, this command fails.
An interesting twist to the shell command facility is that you can use a region of a buffer rather than a traditional file as input to the command. For example, let's say we want to sort a phone list. First, we put the cursor somewhere in the list (say, on the first character of Liam), then we give the mark-paragraph command (M-h). This command defines the phone list as a region, with the cursor at the beginning of the paragraph and the mark at the end.
In the following example, the shaded area shows the extent of the region we want to sort. After selecting a region, we press M-| (for shell-command-on-region); Emacs prompts for the shell command to run.
Type: M-h M-|
Emacs prompts you for a command to execute (Windows).
Now we give the command sort without specifying any input file. Emacs is taking care of the input for us.
Type: sort Enter
Emacs runs a sort on the region (Windows).
Emacs has sorted the phone list (i.e., everything within the region).
A useful variation for M-! puts the output directly into the current buffer, rather than into a *Shell Command Output* buffer. To do so, precede the command with C-u: for example, C-u M-! runs a shell command and puts the output in the current buffer.
Type: C-u M-! ls -la Enter
Emacs runs ls and inserts the result at your current location (Mac OS X).
5.1.2 Using Shell Mode
Now we're ready to discuss shell mode, the interactive facility for running commands. To start a shell buffer, type M-x shell Enter. This creates a buffer named *shell*. You see the prompt for your shell within this buffer. (This defaults to your usual shell; you can substitute another shell to use in Emacs. See "Which shell?" later in this chapter.)
Figure 5-1. Shell buffers for Linux, Mac OS X, and Windows
For the most part, shell mode is exactly like the normal command interface, except that you can use Emacs to edit the commands as you type them. You can copy commands from one place to another, copy the results into a file, save the whole shell buffer to a file, and so on. Note in Figure 5-1 that Emacs has added a few items to the menu bar (Complete, In/Out, and Signals).
A few tricks are worth knowing, though. For example, you normally interrupt a command by typing C-c. If you type C-c in shell mode, Emacs thinks that the C-c is part of a command meant for it, because many Emacs commands start with C-c. Therefore, you have to type C-c C-c to terminate the current job. Likewise, under Unix, you type C-c C-z to stop a job, instead of C-z, and C-c C-d instead of C-d, and so on. (C-c C-d is not strictly necessary because Emacs understands C-d in context. If you're at the end of the buffer, C-d means "end of file"; if you're anywhere else, it deletes a character.) Alternatively, you can select options from the Signals menu rather than using control characters, if desired (for example, selecting EOF instead of typing C-d).
Shell mode also provides a few convenient shortcuts. The command M-p retrieves the last shell command you typed, no matter how far back in the buffer it is. Typing successive M-p's brings back earlier commands.
Type: M-p
M-p retrieves the last command, even if it isn't on the screen (Mac OS X).
In this example, the previous command was more dickensxmas.tex. It's no longer on the screen; its output has pushed it off the top. M-p (for comint-previous-input) retrieves the command, but doesn't execute it; you can edit the command before pressing Enter. To find subsequent commands, type M-n.