Skip to main content
  1. Posts/

Nifty Nvim/Vim Techniques That Make My Life Easier -- Series 9

··684 words·4 mins·
Table of Contents

This is the 9th post of my post series on nifty Nvim/Vim techniques that will make my editing experience easier.

Click here to check other posts in this series.

How do I return a key press from a function and use it in a mapping

I want to write a function to return <Tab> key or <Ctrl-N> based on whether completion menu is available, and use the return value in an insert mode mapping. The initial code is:

inoremap <expr> <Tab> MyTabFun()
function! MyTabFun()
  if pumvisible()
    return "<C-N>"
    return "<Tab>"

However, the function returns those characters literally instead of as key press. This is because Vim thinks that you want to insert those keys literally. To signal a key press, we need to escaped it. Like the following:

return "\<C-N>"
" or
" return "\<Tab>"

The relevant vim doc on this topic is :h expr-quote.


Get character at a specific index in a multi-byte aware fashion

Unlike Python, in Vim script, string indexing uses byte index by default, not character indexing. Byte indexing works well for ASCII characters. Once your string contains multi-byte characters, things no long works as expected. For example, if we run the following code

let a = '你好吗'
echo a[0]

Vim prints <e4>, the first byte of in UTF-8 encoding (the binary representation for 你 using UTF-8 encoding is \xe4\xbd\xa0).

How to we get the character at a specific index? Using strcharpart() instead. For example,

let my_str = '你好吗'
"result will be '你', the first char in my_str
echo strcharpart(my_str, 0, 1)

" result will be '好', the second char in my_str
echo strcharpart(my_str, 1, 1)

We can also use the following convenience function:

function!  CharAtIdx(str,  idx)  abort
    " Get char at idx from str. Note that this is based on character indexing
    " instead of the byte index.
    return  strcharpart(a:str,  a:idx,  1)

Then, to get first char of a string, use CharAtIdx(my_str, 0).

Get string length regardless of ASCII or not

This is related to the previous tip. There is function strlen() or len() in Vim, but they only calculates byte length of a string, instead of character length, like what len() in Python does. We can instead use the strchars() to get string length. I consider this one of the many hidden quirks of Vim. We just need to get used to it.

Use neovim as git diff and merge tool

Here is how to set up neovim as a git diff and git merge tool. Add the following config to the file $HOME/.gitconfig:

    tool = nvimdiff
    prompt = false
[difftool "nvimdiff"]
    cmd = "nvim -d \"$LOCAL\" \"$REMOTE\""
    tool = nvimdiff
    prompt = true
[mergetool "nvimdiff"]
    cmd = "nvim -d \"$LOCAL\" \"$REMOTE\" \"$MERGED\" -c 'wincmd w' -c 'wincmd J'"


Move the view horizontally

If we do not wrap the text and the line text length exceed the window size, some text will be hidden beyond the current view port. To move the view port horizontally, Vim has the following normal mode command:

  • {count}zl: move the current view port {count} characters to the right, default is 1 if no {count} provided.
  • {count}zh: move the current view port {count} characters to the left, default is 1 if no {count} provided.
  • zL: move the current view port half screen width to the right.
  • zH: move the current view port half screen width to the left.

The default behavior for zL and zH is to move the view port half screen width, which may be too much. We can map these shortcuts to use smaller steps:

nnoremap zL 10zl
nnoremap zH 10zh



Using Diffs in Vim
··235 words·2 mins
Nifty Nvim/Vim Techniques That Make My Life Easier -- Series 8
··566 words·3 mins
Migrating from Packer.nvim to Lazy.nvim
··651 words·4 mins