This is the 3rd post of my post series on nifty Nvim/Vim techniques that will make my editing experience easier.
As is often the case, when we want to replace some patterns, we may not be able
to write the correct regular expression in one pass and have to refine it many
times. I often search first to make sure that the search pattern is correct.
Then I manually type the search pattern again in substitute command to
replace it with the text I want. This is acceptable when the pattern is simple
to type. It becomes unacceptable once the search patterns are complicated.
To use the last search pattern, you can keep the search part empty in
substitute command, e.g., %s//REPLACE/g, which will replace all occurrences
of last search pattern with REPLACE. Alternatively, you could use <C-R>/ to
input your last search pattern, that is, first press Ctrl+
R and then press /1.
A least three ways are possible (take function
strftime() and command
:scriptnames (:h :scriptnames) for an example).
Use <C-R>= in insert mode. In insert mode, press
Ctrl+R followed by =, then type the
function (e.g., strftime('%c')) or use execute() to run a command (e.g.,
execute('scriptnames')). The output will inserted right after the current
cursor.
Use :put= in command mode. In command mode, first input put=, then input
the function or execute() command. The output will be put below the
current cursor line.
Use redir. We first use redir to redirect the output to a Vim register
and then paste the content of the register to current buffer. For functions,
use the following
:redir @a | echo strftime('%c') | redir END<CR>For commands, use:
:redir @a | silent scriptnames |redir END<CR>This will redirect the output of function strftime() or command
scriptnames to register a. To paste the content of register a, use
"ap in normal mode.
You can also directly redirect the output to current file:
:redir >> % | silent scriptnames|redir END|edit, which may be more convenient sometimes.
Use let with a register. This is similar to using redir, but more
concise. For function outputs, use:
:let @a = strftime("%c")<CR>For command outpus, use:
:let @a = execute("silent scriptnames")<CR>Then you can paste the register content to current buffer with "ap
Using <C-R>= in insert mode is not good. It will write the output line by
line, thus are very slow if your output has hundreds or thousands of the lines.
The line indentation may also break if you use this method.
For example, given the following text:
免费 免费 🆓️️if you place the cursor on the emoji and use ga, the output is:
<🆓> 127379, Hex 0001f193, Octal 370623 < ️> 65039, Hex fe0f, Octal 177017 < ️> 65039, Hex fe0f, Octal 177017It shows that this character is actually made up of three Unicode characters!
This is called variant form and the Unicode
character U+FE0F is called a variant selector . There are other forms of
multi-Unicode-char characters, for example, combining characters,
where the character is made up of a base character and one or more decoration
characters. In Vim, the main character is called base character and the other
character that follows is called composing character.
If you use search the base character or the composing character using their Unicode code points, you will find disappointedly that no pattern are found.
It turns out the way we search the characters are wrong. To search the base
character, you should follow it by \%C (in no-magic mode) or %C (in magic
mode). For example, to search the above base character in the emoji, use
\v%U1f193%C\%U1f193\%CTo search the composing character, it is more complicated. You should type
/<Ctrl-V>ufe0f:
<Ctrl-v>u: start Unicode inputfe0f: the code point of the composing characterThen press Enter to search.
If you set the option laststatus to 2,
after searching a word using *, you will find that the text is not cleared
automatically. To clear the text , you can press Ctrl+L .
You can use :execute, which
will evaluate the expression followed before executing it. Suppose that you
have defined a download URL variable URL, in order to download it using
wget inside Neovim, you can use the following command:
:execute '!wget ' . URLNote the space inside '!wget ' because we use . to concatenate the
arguments behind execute to form an expression. Without the space, the
command you are going to run is !wget{the_content_of_URL}, which is
apparently wrong.
system() function?#When I wanted to assign the output of an external shell command to a Vim
variable using system() function, I found that a trailing newline character
was present in the returned string, which breaks a lot of things. For example,
let python_path = system('which python')the variable python_path will contain an annoying newline. To remove this
newline character, you can use various methods:
substitute(): you can replace the trailing newline with an empty string to remove it:
substitute(python_path, '\n$', '', 'g')string indexing: you can also utilize string indexing (see :h expr-[:]) if
you are sure that there is only one newline at the end of the string:
" remove last byte from the string
let new_path=python_path[:-2]system()?see :h c_CTRL-R for more information. ↩︎
This is the 3rd post of my post series on nifty Nvim/Vim techniques that will make my editing experience easier.
As is often the case, when we want to replace some patterns, we may not be able
to write the correct regular expression in one pass and have to refine it many
times. I often search first to make sure that the search pattern is correct.
Then I manually type the search pattern again in substitute command to
replace it with the text I want. This is acceptable when the pattern is simple
to type. It becomes unacceptable once the search patterns are complicated.
To use the last search pattern, you can keep the search part empty in
substitute command, e.g., %s//REPLACE/g, which will replace all occurrences
of last search pattern with REPLACE. Alternatively, you could use <C-R>/ to
input your last search pattern, that is, first press Ctrl+
R and then press /1.
A least three ways are possible (take function
strftime() and command
:scriptnames (:h :scriptnames) for an example).
Use <C-R>= in insert mode. In insert mode, press
Ctrl+R followed by =, then type the
function (e.g., strftime('%c')) or use execute() to run a command (e.g.,
execute('scriptnames')). The output will inserted right after the current
cursor.
Use :put= in command mode. In command mode, first input put=, then input
the function or execute() command. The output will be put below the
current cursor line.
Use redir. We first use redir to redirect the output to a Vim register
and then paste the content of the register to current buffer. For functions,
use the following
:redir @a | echo strftime('%c') | redir END<CR>For commands, use:
:redir @a | silent scriptnames |redir END<CR>This will redirect the output of function strftime() or command
scriptnames to register a. To paste the content of register a, use
"ap in normal mode.
You can also directly redirect the output to current file:
:redir >> % | silent scriptnames|redir END|edit, which may be more convenient sometimes.
Use let with a register. This is similar to using redir, but more
concise. For function outputs, use:
:let @a = strftime("%c")<CR>For command outpus, use:
:let @a = execute("silent scriptnames")<CR>Then you can paste the register content to current buffer with "ap
Using <C-R>= in insert mode is not good. It will write the output line by
line, thus are very slow if your output has hundreds or thousands of the lines.
The line indentation may also break if you use this method.
For example, given the following text:
免费 免费 🆓️️if you place the cursor on the emoji and use ga, the output is:
<🆓> 127379, Hex 0001f193, Octal 370623 < ️> 65039, Hex fe0f, Octal 177017 < ️> 65039, Hex fe0f, Octal 177017It shows that this character is actually made up of three Unicode characters!
This is called variant form and the Unicode
character U+FE0F is called a variant selector . There are other forms of
multi-Unicode-char characters, for example, combining characters,
where the character is made up of a base character and one or more decoration
characters. In Vim, the main character is called base character and the other
character that follows is called composing character.
If you use search the base character or the composing character using their Unicode code points, you will find disappointedly that no pattern are found.
It turns out the way we search the characters are wrong. To search the base
character, you should follow it by \%C (in no-magic mode) or %C (in magic
mode). For example, to search the above base character in the emoji, use
\v%U1f193%C\%U1f193\%CTo search the composing character, it is more complicated. You should type
/<Ctrl-V>ufe0f:
<Ctrl-v>u: start Unicode inputfe0f: the code point of the composing characterThen press Enter to search.
If you set the option laststatus to 2,
after searching a word using *, you will find that the text is not cleared
automatically. To clear the text , you can press Ctrl+L .
You can use :execute, which
will evaluate the expression followed before executing it. Suppose that you
have defined a download URL variable URL, in order to download it using
wget inside Neovim, you can use the following command:
:execute '!wget ' . URLNote the space inside '!wget ' because we use . to concatenate the
arguments behind execute to form an expression. Without the space, the
command you are going to run is !wget{the_content_of_URL}, which is
apparently wrong.
system() function?#When I wanted to assign the output of an external shell command to a Vim
variable using system() function, I found that a trailing newline character
was present in the returned string, which breaks a lot of things. For example,
let python_path = system('which python')the variable python_path will contain an annoying newline. To remove this
newline character, you can use various methods:
substitute(): you can replace the trailing newline with an empty string to remove it:
substitute(python_path, '\n$', '', 'g')string indexing: you can also utilize string indexing (see :h expr-[:]) if
you are sure that there is only one newline at the end of the string:
" remove last byte from the string
let new_path=python_path[:-2]system()?see :h c_CTRL-R for more information. ↩︎