Control, Oh!

So I’m brushing up on my shell-fu. In other words, working my way through the excellent and highly recommended From Bash to Z Shell, skimming the bits I know and experimenting with the bits I didn’t know. Here’s something I didn’t know; you may not either.

By default in bash and zsh the control-o key has some very useful behaviour. Let’s say you’re in a compile/edit/test cycle and you’re basically in a tight loop of shell commands. In this case you’re probably doing what I would have in the past, namely using up-arrow or control-r to manually locate the next command to execute from the history list.

% make
...
% ./test
...

Try this at home (use “echo make” or similar). You quickly get bored of the double up-arrow to alternate between the two commands.

Control-o to the rescue!

Hit up-arrow twice to get back the to the make command, but don’t hit return. Instead, hit control-o. This will execute the make command and leave you at the next in the history list — the ./test command — ready to hit return again (or another control-o).

You can keep iterating through these two (or more!) commands with a single keystroke, each time previewing what you’re going to execute. I like it.

An interesting tip. Thanks, Alastair.

I have to say, though, that I find it easiest to use the “bang, first letter” sequence when I’m developing; e.g.,

!m <RET>

!t <RET>

to run make and then the program itself. (I allow . to be in my path when logged in as a regular user.)

When things get more complicated, I find it easier to write a short script that does the sequence of commands. I name this script “x” and put it in the directory that I’m working in.

But, always good to know another way.

Shortcuts? History expansions? Scripts!? What sort of universe do you people come from? When I get tired multiple up-arrow taps, I use the power of Ctrl-U/Ctrl-Y to compose the commands into a single &&-chain such as “make && ./test” so I can thenceforth repeat a single command.