Encode string once and reuse for subsequent calls

[zikaari]

For times when apps use same string multiple times with findNextMatch etc, it is very expensive to run UTF8 encoding over and over.

For such cases, it is now advised to create an instance of OnigString and keep it referenced. Then, pass that OnigString ref to findNextMatch instead of literal strings for better performance.

Following code was run against onigasm@2.0 (before) and current patch (after):

import {
    loadWASM,
    OnigScanner,
    OnigString
} from 'onigasm'


(async () => {
    await loadWASM(require(`onigasm/lib/onigasm.wasm`))

    const str = await (await fetch('node_modules/typescript/lib/typescriptServices.js')).text()

    const onigStr0 = new OnigString(str.repeat(3))
    const onigStr1 = new OnigString(str.repeat(4))

    const scanner = new OnigScanner(['searchTillTheEndButYouWontFindMe'])

    const t0 = performance.now()    
    const m0_1 = scanner.findNextMatchSync(onigStr0)
    const m1_1 = scanner.findNextMatchSync(onigStr1)
    const m0_2 = scanner.findNextMatchSync(onigStr0)
    const m1_2 = scanner.findNextMatchSync(onigStr1)
    const m0_3 = scanner.findNextMatchSync(onigStr0)
    const m1_3 = scanner.findNextMatchSync(onigStr1)

    console.log('Took ', performance.now() - t0, 'ms')
})()

Before (onigasm@2.0)

---

After (current patch)