Home | Previous module | Next module
Get ready for a big module! Here you will learn a lot of things to speed up (and enhance) your coding techniques in Vim. Note that the content will be easy to digest if you slowly adapt the features mentioned in this module in your workflow, without trying to remember everything at once. Also, this is the module from where I will start referring you to the awesome Vim documentation, for learning the topics introduced here in detail.
Exuberant ctags and cscope are external programs that allow jumping across code-blocks, just like IDEs, in Vim. For installation, please refer to their respective websites.
To use ctags, first build tags for your project in the project's root directory using the command:
ctags -R " Build tags recursively
The above command creates a tags
file in your directory, containing a list of tags that ctags can parse very fast.
Note that you need to repeat the command regularly to update the tags after you make changes in your project.
See this to know how you can partially
automate tag-generation in git repositories.
Now, say you are at a function call foo()
in a cpp file.
To jump to the definition of foo
, just press Ctrl-]
.
Whoa! This even jumps across files!
To get back to where you were, press Ctrl-t
.
If there are multiple functions with the name foo
, you will get a list to select from.
Note that ctags maintains a stack of your jump-locations.
Thus, you can keep jumping from one function-call to another using Ctrl-]
, and pop back repeatedly using Ctrl-t
.
Check out :help tags
to learn more about using tags in Vim.
Cscope is an advanced tool that mostly supersedes what ctags does, and also lets you jump to the call-sites of a method. Post installation, similar to ctags, you first need to build a database for cscope. For example, to build a cscope-database for all the cpp files in your directory, use:
find . -name '*.cpp' > cscope.files
cscope -Rb " Build cscope-database recursively (in a file 'cscope.out')
Now let's say you want to get all the call-sites of the method foo
.
Stay over the function header and type Ctrl-\ c
to get the list of all the callers of foo
(if you are using the file containing default mappings from the cscope website).
Refer to the cscope website to learn about many other features of cscope.
Forget about code indentation worries in Vim! Vim supports indentation for programs written in almost every language (and others keep getting added). The following two options are useful in this regard:
filetype indent on " enable file-type based indentation
set autoindent " indent next line based on the previous one (e.g., whether it's a '{')
But you had not enabled these in the beginning.
Or someone gave you a very badly-indented file.
No problem!
You can indent existing code using the =
command.
Just select the text you wish to indent, and press =
.
See the magic!
You can indent code in the normal mode as well.
Say you want to indent the next 6 lines.
Just press 6==
and see the magic again!
You can also increase or decrease the indentation levels of a line using >>
and <<
, respectively, in normal mode.
Similar to the =
command, you can select text in visual mode and press >
or <
as well.
- You can indent a whole file using the shortcut
gg=G
in normal mode. Try figuring out what do the individual commands mean. - You can set the number of characters used for indentation by
>>
and<<
(say, to 4) as follows:
set shiftwidth=4
Vim can show you the matching parentheses, braces, and brackets when the cursor is on one of them. Add the following in your vimrc:
set showmatch
Now, if your cursor is on say, the opening bracket, press %
to jump to the corresponding closing bracket.
Vim has a "secret" built-in plugin (called matchit) that also allows you to jump to the corresponding keywords (for example, endif
for if
), using the %
operator.
Add the following line to your vimrc file to enable this plugin by default:
runtime! macros/matchit.vim
Once you start coding, you will find this very handy, specially when a code-block doesn't fit into your screen.
It's routine to keep editing your code, compile it, edit it again, and so on. And it's very inefficient to close Vim and run the commands, open Vim and edit again, and so on.
One simple alternative is to simply put Vim to the background using Ctrl-Z
, use the shell, and then come back using fg
.
However, you can even avoid that by running shell commands while being inside Vim.
There are two ways to do this:
-
Type
:shell
to get to a temporary shell-prompt. TypeCtrl-D
to come back to Vim. -
Run shell commands using the
!
operator in command-line mode. For example,:!javac %
will compile the file loaded in the current buffer.
When you have Makefiles, you can build your project even more efficiently using the builtin make command of Vim.
Just run :make
and see your project built!
There are even better ways to compile your code using plugins. We will learn more about them in later modules.
In module 3, we had learnt that we can search in the current buffer using the /
and ?
commands, and replace text in the current buffer using the %s/oldword/newword/g
command.
To search across multiple files, we can use grep (yes, the one you use in a terminal) along with all its options, in the command-line mode of Vim.
For example, to recursively search for the word sophisticated
in the current directory, use:
:grep -R 'sophisticated' . " search the word 'sophisticated' recursively (-R) in the current directory (.)
The advantage of using grep within Vim is that it loads all the search-results in the quickfix window.
To understand this, once you have searched for a word using grep
, type :copen
in the command-line mode.
This creates a new split in your Vim with all the search occurrences.
You can either use the arrow keys with <CR>
to jump to any results (use <C-w>w
to switch windows),
or use the following commands to jump to the next and the previous search-result, respectively:
:cnext " jump to the next entry in the quickfix window
:cprevious " jump to the previous entry in the quickfix window
Cool, right!
But this is not all.
You can apply a command to all the entries in the quickfix window using :cdo
.
This means you can replace all the occurrences of the searched word as follows:
:grep -R 'oldword' . " search 'oldword' recursively
:cdo %s/oldword/newword/c " replace occurrences one-by-one
OR
:cdo %s/oldword/newword/g | update " replace all occurrences and save
Note that in the last command, as we have replaced all the occurrences in one shot, we have used the bar operator (|
) to also update (i.e., save) all the files.
In the former case, you need to save the buffers explicitly -- either one-by-one, or using :wall
.
Combine the above commands with regular expressions, and see Vim do wonders with searching and replacement. Check out this article to learn how to simplify the above process further.
Similar to the diff
command in your terminal, Vim provides a vimdiff
program to compare and merge two files.
To see the differences between file1
and file2
, type the following at your command prompt:
vimdiff file1 file2
This will open Vim with the two files open side-by-side, in a vertical split.
The changes will be highlighted using different colors.
If you have used diff
before, you will find vimdiff
a boon.
Vim provides the following shortcuts to navigate among the changes, and push or pull changes from one file to another:
]c
: Jump to next change.[c
: Jump to previous change.:diffput
ordp
: push the current change to the other split.:diffget
ordo
: pull the corresponding change from the other split.
You can also diff multiple files using vimdiff.
Use :help vimdiff
to learn the details.
Now we are moving on to some sophisticated, not-known-to-basic-users, features of Vim. Let's start with insert-mode completions.
Vim provides superb completion support. You can avoid typing literally anything you have typed before, and also complete things that you have not even typed yet! In insert-mode, just type a few characters and press the following to complete items (you will get a navigable list if multiple choices are available):
Ctrl-p
: Search what you have already typed, in reverse order (backwards from the current location).Ctrl-n
: Search what you have already typed, forwards from the current location.Ctrl-x Ctrl-f
: Complete file pathsCtrl-x Ctrl-]
: Complete tagsCtrl-x Ctrl-l
: Complete whole linesCtrl-x Ctrl-k
: Complete from Vim's built-in dictionaryCtrl-x Ctrl-o
: Omni-completion
The last entry (omni-completion) needs special mention.
Omni-completion lets you complete keywords based on special characters.
For example, in c/cpp programs, omni-completion will let you complete after typing .
or ->
operators.
To see a list of languages for which Vim provides omni-completion, use :help omni<Tab>
.
These are not the only completion-types available.
Use :help ins-completion
to get the comprehensive list.
In the next module, we will learn about some plugins that simplify as well as extend the completion types available, even further.
Vim allows you to create abbreviations as shortcuts to enter text.
Say you want to expand syso
in Java to System.out.println
.
Create an abbreviation as follows:
:iabbrev syso System.out.println
You would find that once you type syso
and then anything else, it gets expanded to System.out.println
.
The i
in the above command says that the abbreviation works only in insert mode.
Another common use-case of abbreviations is to correct spellings.
Say you find that you repeatedly type teh
instead of the
.
Now you know how to correct yourself automatically:
:iabbrev teh the
Mappings are one of the most useful features of Vim. You can map almost any sequence of keys to a shorter one.
Say, you want to map Ctrl-l
to clear highlights (created say because of some search).
You know from the previous module that the command to clear highlights is :nohlsearch
.
The mapping goes as follows:
nmap <C-l> :nohlsearch<CR> " clear search-highlights using `<C-l>`
Let's break down the mapping:
- The first
n
means that the mapping should work only in normal mode. If the mapping should work in all the modes, drop then
. Similarly,imap
is for the insert mode. - The first string till a space (
C
in<C-l>
stands for theCtrl
key) is the map that we are creating. - The second string is the command that we wish to map the first string to.
Here we have mapped
<C-l>
to first go to the command-line mode, typenohlsearch
, and then press enter (<CR>
).
It is advisable to always use noremap
instead of map
to avoid recursive mappings (map within a map).
Thus, the above command becomes:
nnoremap <C-l> :nohlsearch<CR> " clear search-highlights using `<C-l>`
Following is a small list of the keystrokes used to represent the special-keys on our keyboards:
C
stands forCtrl
M
stands forAlt/Option
CR
stands forEnter
Tab
is for theTAB
keySpace
is for theSPACE
bar
Mappings are mostly created by appending them to a leader key (in order to avoid conflicts with the existing commands).
The default leader is the \
(backslash) key.
Thus, the following map would clear the search highlights using \l
:
nnoremap <leader>l :nohlsearch<CR> " clear search-highlights using `\l`
You can even change the leader key, globally as well as locally. See :help mapleader
for details.
You can add multiple commands in a map using the <bar>
(or |
) operator.
Thus, to save your code and then build it using \m
, you can have the following mapping:
nnoremap <leader>m :w<CR> <bar> :make<CR>
Mappings are very useful. With time, you will keep finding keystrokes or commands that you often use in sequence. Those will be good times to add shortcuts using maps in your vimrc.
We will wrap up this module by getting a glimpse of the many ways code (or any text) can be folded for better visibility and movement in large files, in Vim.
In programs, folding is useful to wrap, say functions, so that you can easily glance over and move among the functions in your file. In text, latex, or markdown files, folding is useful to wrap, say sections or figures, and look over the way your file is organized.
There are many ways to fold text. First, we need to enable folding, and then select one of the following modes:
- manual: Create folds manually, as and when needed (default).
- marker: Create folds using fold-markers (
{{{
and}}}
). - indent: Create folds based on indentation levels.
- syntax: Create folds based on syntax-highlighting items.
- expr: Create folds based on custom expressions.
We will learn about manual folding and markers-based folding in this tutorial.
As manual mode is the default, just enable folding as follows:
set foldenable " Enable folding
To fold a block of text, select it (using visual-mode), and then press zf
.
This creates a fold over the selected region.
The following is a subset of commands to work with folds (in normal mode):
zo
: Open a foldzO
: Open all folds under the cursor recursivelyzc
: Close a foldzC
: Close all folds under the cursor recursivelyza
: Toggle a foldzA
: Toggle all folds under the cursor recursively
The disadvantage of manual folds is that they get lost once you close a file. Let's learn about one of the folding methods that creates persistent folds -- marker. Enable it as follows:
set foldmethod=marker " Enable marker-based folding
Now you can define a fold by enclosing the intended region of text between {{{
and }}}
(in comments around the text-block).
You can create nested marker-based folds by defining a level next to the fold-markers; e.g., {{{1
and }}}1
, {{{2
and }}}2
, and so on.
Remember to close the folds properly with the correct level-numbers.
Folding is a very broad subject.
Learn more about the other fold-methods using :help folding
.
In this module, we learnt a lot of options and facilities provided by Vim for coders. The aim was to introduce you to the kind of things that can be done; mastering all of them will take months of practise. I suggest you keep coming back to this module whenever you need a specific feature, instead of trying to remember everything and complicating your workflow. In the next module, we will see how some of the existing functionalities can be extended using the plugins created by the Vim community.
Home | Previous module | Next module
Star this repository on GitHub if you like the tutorial.