Awesome
Screencast
If you'd like a visual demonstration of the plugin, you can find a screencast here.
Usage
This plugin remaps the standard p
and P
mappings to enhance their functionality. When pasting, it compresses all blank lines that result from the paste to a single one (or none, at the top and bottom of the file). That way, even if you copy any leftover whitespace, it'll be neatly trimmed to just one line. This takes effect only for linewise pasting, since it's not entirely clear what the behaviour should be for characterwise and blockwise pasting.
Mappings
If you don't want to clobber your default p
and P
mappings, you can make whitespaste use different ones by setting two predefined variables:
let g:whitespaste_before_mapping = ',P'
let g:whitespaste_after_mapping = ',p'
If you need more fine-grained control, you can disable mappings altogether by setting both of these variables to empty strings. You can then use the three provided <Plug>
mappings for your purposes. For example:
let g:whitespaste_before_mapping = ''
let g:whitespaste_after_mapping = ''
nmap ,P <Plug>WhitespasteBefore
nmap ,p <Plug>WhitespasteAfter
xmap ,P <Plug>WhitespasteVisual
xmap ,p <Plug>WhitespasteVisual
Special cases
The plugin also takes care of special cases like pasting functions/methods, if-clauses and so on. Currently, these special cases work only with ruby and vimscript, but see below in "Extending" to find out how you can extend the plugin for a different language or change it to fit your own coding style. If you're wondering how this could be useful, consider a ruby example:
class Test
def one
end
def two
end
end
There's no simple way to swap these two methods' positions without having to adjust blank lines around them. For instance, if you delete the two
method definition completely (with its surrounding whitespace) and try to paste it above, you'd end up with:
class Test
def two
end
def one
end
end
Whitespaste takes care of that by detecting that a pasted block stats with a "def" line and tweaks the resulting whitespace accordingly, so the above example is automatically corrected to:
class Test
def two
end
def one
end
end
Compatibility
Whitespaste automatically integrates with the vim-pasta plugin. If you want to disable this, you can do it like so:
let g:whitespaste_pasta_enabled = 0
Extending
The global variable g:whitespaste_linewise_definitions
controls the behaviour of the plugin. It's a hash with two keys: top
and bottom
. Each of these keys points to a list of definitions that are attempted to decide how much space to leave at the top of the pasted text and at the bottom, respectively. This looks a bit like this:
let g:whitespaste_linewise_definitions = {
\ 'top': [
\ { ... }, { ... }
\ ],
\ 'bottom': [
\ { ... }, { ... }
\ ]
\ }
Each of the definitions in the list is a dictionary that can hold several different keys.
target_line
: If this key is set, the definition matches only for a specific line number of the area that the text is pasted in (the "target"). The special value -1 denotes the last line + 1.target_text
: A pattern to match the target line's contents with.pasted_line
: Only matches when the first/last nonblank line of the pasted text is positioned on this line.pasted_text
: Only matches when the first/last nonblank line of the pasted text matches this pattern.blank_lines
: the exact amount of blank lines to set at the top or bottom of the pasted text.compress_blank_lines
: same asblank_lines
, except only enforced if the current amount of blank lines is larger than the given number.
The target_line
and pasted_line
keys would probably not be very useful,
but they help with the edge case definitions:
let g:whitespaste_linewise_definitions = {
\ 'top': [
\ { 'target_line': 0, 'blank_lines': 0 },
\ ],
\ 'bottom': [
\ { 'target_line': -1, 'blank_lines': 0 },
\ ]
\ }
This set of definitions ensures that, when pasting at the top and bottom of the buffer, whitespace is reduced to 0. Usually, though, you'll want to use target_text
and pasted_text
.
The definitions are attempted in order, which means that you should put stricter definitions at the top and fallbacks at the bottom.
For an example, illustrating the "target" and "pasted" lines, let's assume that we've yanked the following text:
puts "one"
puts "two"
puts "three"
Now, we'd like to paste it into the following area:
something {
}
Pasting the text on any line between the curly braces sets the target lines to something {
and }
respectively -- the target lines are always the first non-blank lines upwards and downwards of the pasted position. Similarly, regardless of the whitespace we've pasted along with the given text, the "pasted" lines' texts will be puts "one"
(for the top) and puts "three"
(for the bottom). For this example, if we wanted the paste to always result in this, regardless of leftover blank lines:
something {
puts "one"
puts "two"
puts "three"
}
Then we could define a whitespaste definition like so:
let g:whitespaste_linewise_definitions = {
\ 'top': [
\ { 'target_text': '{$', 'blank_lines': 0 }
\ ],
\ 'bottom': [
\ { 'target_text': '^}$', 'blank_lines': 0 }
\ ]
\ }
Note that, for a catchall definition, you could simply make a definition without any conditions -- only with a blank_lines
key or compress_blank_lines
key. Such a definition exists by default, and is set to
{ 'compress_blank_lines': 1 }
That way, no more than 1 blank line is allowed on any paste, as long as some other definition doesn't override it earlier on.
To add definitions for only a specific filetype, assign the buffer-local b:whitespaste_linewise_definitions
variable instead. Buffer-local definitions are assumed to be of higher priority, so be careful when putting catchall definitions in there.
The default definitions can be seen in plugin/whitespaste.vim
. At this time, the global definitions handle curly braces, as demonstrated in the above example, remove whitespace at the top and bottom of the buffer, and compress all other pasting operations to only one blank line. There are also specific definitions for ruby and vimscript.
Contributing
Pull requests are welcome, but take a look at CONTRIBUTING.md first for some guidelines.