Improve matching algorithm

[hrsh7th]

We need tests for lua...

[kkharji]

Its a bit hard and you will need custom shell and lua scripts, if you want I can walk through running your first test busted framework

[hrsh7th]

I'm implementing it...

benchmark

v1: elapsed time: 2.401
v2: elapsed time: 0.342

local M1 = require'compe.completion.matcher'
local M2 = require'compe.completion.matcher2'

local attempts = 100000

x = os.clock()
for j = 1, attempts do
    tmp = M1.score('nvimgetbufline', 'nvimgetbufline', 'nvim_g_et_b_uf_line')
end
elapsed = os.clock() - x
print(string.format("V1: elapsed time: %.3f", elapsed))

x = os.clock()
for j = 1, attempts do
    tmp = M2.score('nvimgetbufline', 'nvim_g_et_b_uf_line')
end
elapsed = os.clock() - x
print(string.format("V2: elapsed time: %.3f", elapsed))

[kkharji]

Oh wow 😍 😍 😍 😍 😍 😍 😍 😍 😍 😍 😍 😍 😍 Nice well done @hrsh7th

[hrsh7th]

I will add the feature of remaining char match.

Currently, nvim-compe does not match fnamemodify with fmodify.
This feature makes matching it.

[hrsh7th]

I will add the feature of remaining char match.

Currently, nvim-compe does not match fnamemodify with modify.
This feature makes matching it.

Sorry. It bit hard to implement. So I will merge performance improvements before implementing them.