Jan 152019
 

Now that the click-bait is out of the way, vi movement keys are perfectly reasonable particularly to those who have been using them for decades (which includes me). But for ages, vi itself has supported the arrow keys for movement as well as the tradition cursor movement keys.

For the benefit of those who have not used vi and are wondering what those traditional cursor movement keys are, they are H (left), J (down), K (up), L (right). A bit like the gamer’s set of movement keys – W, A, S, and D, except that the vi movement keys pre-date arrow keys.

There are those who will claim that the traditional movement keys are more efficient as they require less hand movement. And they are. So it is perfectly understandable that many tiling window managers and other keyboard-centric software uses these movement keys.

But someone who hasn’t spend decades hard-wiring the vi movement keys into their brain, will find vi-style key bindings inscrutable. And the fix? Just use the arrow keys as well.

There is no harm in having two key sequences do the same thing; no harm in emphasising that the arrow keys work too. And indeed no harm encouraging the use of vi-style movement keys by emphasising their efficiency.

Don’t forget that someone learning a new tiling window manager (or most other things) can be put off by the silliest of things – such as inscrutable control keys.

Rusty Handrail
Dec 112016
 

Vi (or vim) is one of those editors that at first appearance appears to be insanely user-hostile, and some will say it looks the same way on a second, and third glance. Yet it remains one of the most popular editors under Linux, and even if you choose another editor as your mainstay, you are likely to encounter it in use a fair bit.

(Apologies for the little line at the top of that screenshot; lazy editing)

The strange thing about vi is that nobody uses the real thing any more (well almost nobody), but instead clones of one kind or another. That is mostly because vi originated in the commercial world of Unix, and clones were written to be open-source. It is perhaps worth remembering that vi has remained the mainstay of editing under Unix-based operating systems for three decades despite there being many alternatives.

It may look plain, but vi has almost every editor feature you can think of and almost certainly a few you never thought of.

The Modes

Almost every every other editor currently in use is a modeless editor, or at least mostly modeless. Vi is different in that it has three different modes that operate differently – insert mode, command mode, and ex-mode (essentially for extended commands). Of the three modes, the insert mode is the most like other editors, although are relatively few commands to use. As you can see in the screenshot above, the words “- INSERT -” appears at the bottom of the screen whenever you are insert mode (and the same for replace mode which is effectively the same).

Most commands are performed in the command mode, which can be thought of as the default mode – there is nothing saying that you are in command mode. If in doubt, you can press Esc to get from insert mode into command mode. There are some who will argue that in fact vi is modeless and that the “insert mode” is in fact a parameter to the insert command. This has a certain ring of validity to it – if you enter 32iHelloEsc in command mode, you will end up with 32 copies of the word “Hello” inserted.

But conventionally vi is written of as a mode-based editor, so it is best to think of it as such until you have learned enough to throw off conventional wisdom and go your own way.

The last mode is ex-mode, which at a basic level is covered only in enough detail to tell you how to get out of it! It is entered from command mode with the “:” command, at which point the cursor moves to the end of the screen leaving you free to type in lengthier commands. To exit simply hit the backspace key until the cursor returns to its normal location.

The Insert Mode

(and the replace mode)

The insert mode is started with a variety of commands, but the simplest is i(nsert).

Once in insert mode, you can start typing normal text without worrying about what commands it will run. There are a fair few things you can do with the control keys, but we’ll skip over those for the “basics”. To correct a few historical limitations :-

  1. You can move the cursor around with the arrow keys. It might seem a bit strange to say so, but the original vi didn’t allow you to move in insert mode partially because it pre-dates arrow keys (yes, really!) and had to use commands to move the cursor around the screen.
  2. You can move anywhere within the file and make changes anywhere; not just where the original change was intended. This may seem like an unnecessary feature to explain, but when you are changing a single word, it can seem wrong to also go somewhere else in the file and make changes elsewhere.

Without going into the more esoteric features, there is not a great deal more to say about the insert mode except it is exited with the Esc key.

The Command Mode

The movement commands :-

Arrow Keys Moves the cursor
h & l Moves the cursor left and right.
j & k Moves the cursor down and up.

The use of the h,j,k, and l keys to move the cursor around the screen seems rather bizarre except when you realise that some early terminals connected to Unix systems lacked cursor keys. They remain for compatibility reasons and because some people feel that they can be quicker to use as they require less hand movement than the cursor keys, or like me that those keys are burnt into muscle memory and so they are used almost without thought.

b(ack) Move backwards one word.
e(nd) Move forward to the end of the word.
f(orward){char} Move forwards on the same line to the next occurrence of {char}.
n(ext) Move to the next occurrence of the last search.
 / Search for something.

The most basic command for deleting text is “x” which deletes the character under the cursor, but a hint of what can be done with vi comes with the d(elete) command. The d(elete) command takes a movement as a parameter, and deletes from the current cursor position until where there movement takes you :-

dd Deletes line.
d$ Deletes to the end of the line.
d0 Deletes until the beginning of the line.
dw Delete until the end of the word.
diw Delete “in” word – deletes the current word.
df{char} Deletes until the next occurrence of {char}

But we now move finally to adding text :-

i(nsert) Insert at the current cursor.
o(pen) Open a new line below the current line.
O(pen) Open a new line above the current line.
a(ppend) Append text after the cursor.
A(ppend) Append new text at the end of the line.

Lastly, we can save and exit vi with “ZZ”.

Ex-Mode

This is going to be even more truncated than the last section (I know the last section doesn’t seem truncated, but trust me – it is!). There is a great deal more to this mode than just the three commands below :-

:write (or :w)

Writes the file being edited. Two options I am going to mention here.

Firstly you can add a filename to the command to write to an alternate file (:write new-filename) – very handy if you find you’re making changes to a file that you do not have permission to overwrite.

If you need to override a warning vi has about overwriting the current file, you can do so by appending an exclamation (!). Just don’t do it automatically (I’ve a sorry song to sing about doing that!).

:quit

And to quit vi, simply use the :quit command. If there are unsaved changes in the file you are editing, it will stop you, in which case if you really want to lose your changes add an exclamation (!).

Oct 272016
 

I have recently been ‘entertaining’ myself with watching some videos on the vim editor which to the uninitiated is an extremely powerful if somewhat ‘unusual’ editor that is popular amongst Linux power users. One of the surprising things that came up was that apparently there are experienced vim users who are not aware of why the ex mode exists.

Or probably why the ex command exists.

In the dim and distant past (and in fact even longer than I’ve used Unix!), one of the possible ways of interacting with computers was with a printing terminal :-

On such a terminal, using a visual editor like vim (or it’s predecessor vi) would have been painful. Redrawing the screen would take a couple of minutes or more; imaging moving the cursor across from the beginning of the line to the end!

So it was common to use an alternative kind of editor – the line editor. The process of creating a file is somewhat clumsy :-

$ ex ~/Foo
"~/Foo" 1L, 4C
Entering Ex mode.  Type "visual" to go to Normal mode.
:p
Foo
:a
Bar
.
:p
Bar
:1
Foo
:write
:quit

Now for a quick explanation (although this is no tutorial on line editors!): The ex ~/Foo is the command given to start editing a pre-existing file called Foo in the ex editor. After the editor starts up, I enter the “p” command to print the current line. I then use the “a” command to append text after the first line, and enter a “.” on it’s own to finish adding lines. Again I use “p” to print the current line, and then “1” to print the first line.

Which is more than you’ll ever want to know about how to use ex, so why does it still exist?

The first reason is simply because it’s possible. It’s almost certainly fairly easy to support the ex mode with vim; after all the ex-mode is effectively the commands you get when you enter “:” within vim.

The next reason is that line editors were sometimes used within shell scripts to batch edit files, and somewhere out there is a shell script from hell that relies on ex to keep running.