perf(core): improve vdom diffing by removing unnecessary conditional statement in getSequence function

[xxgjzftd]

Here is some simple explanation for this pr.
For the removal of conditional statement if (arrI < arr[result[u]]):
Before the while loop at line 2479, the arrI variable always less than the last item of the result array,
so after this loop, arrI always less than the result[v],
and after this loop, the v always equals to u,
so we don't need the arrI < arr[result[u]] statement.

For the result array init const result = []:
Actually, this change make the result variable different from original,but it don't effect the children patch result.
Original getSequence([0, 1, 8, 7, 9, 5]) will be[0, 1, 3, 4], now it's [1, 3, 4].

[github-actions]

Size Report

Bundles

File	Size	Gzip	Brotli
runtime-dom.global.prod.js	100 kB (-14 B)	37.7 kB (-9 B)	34 kB (-37 B)
vue.global.prod.js	159 kB (-14 B)	57.6 kB (-9 B)	51.2 kB (-5 B)

Usages

Name	Size	Gzip	Brotli
createApp	55.4 kB (-14 B)	21.3 kB (-11 B)	19.4 kB (-17 B)
createSSRApp	59.3 kB (-14 B)	23 kB (-7 B)	20.9 kB (+57 B)
defineCustomElement	60.1 kB (-14 B)	22.8 kB (-13 B)	20.7 kB (-11 B)
overall	69.1 kB (-14 B)	26.3 kB (-6 B)	24 kB (+53 B)

[pkg-pr-new]

Open in Stackblitz

@vue/compiler-core

pnpm add https://pkg.pr.new/@vue/compiler-core@3928 

@vue/compiler-dom

pnpm add https://pkg.pr.new/@vue/compiler-dom@3928 

@vue/compiler-sfc

pnpm add https://pkg.pr.new/@vue/compiler-sfc@3928 

@vue/compiler-ssr

pnpm add https://pkg.pr.new/@vue/compiler-ssr@3928 

@vue/reactivity

pnpm add https://pkg.pr.new/@vue/reactivity@3928 

@vue/runtime-core

pnpm add https://pkg.pr.new/@vue/runtime-core@3928 

@vue/runtime-dom

pnpm add https://pkg.pr.new/@vue/runtime-dom@3928 

@vue/server-renderer

pnpm add https://pkg.pr.new/@vue/server-renderer@3928 

@vue/shared

pnpm add https://pkg.pr.new/@vue/shared@3928 

@vue/compat

pnpm add https://pkg.pr.new/@vue/compat@3928 

vue

pnpm add https://pkg.pr.new/vue@3928 

commit: 46e0d4e

[skirtles-code]

I made an attempt at benchmarking this:

• Playground

where:

• getSequence is the existing implementation.
• getSequence2 is the implementation in this PR.
• getSequence3 removes the conditional but leaves the const result = [0] unchanged.

Running that benchmark in both Chrome and Firefox, it seems that getSequence2 is fairly consistently slower than getSequence.

getSequence3 is about the same as getSequence, though getSequence3 is marginally faster.

Perhaps I'm missing something, but it seems that changing const result = [0] to const result = [] has actually made things slower.