Poor performance when calling toShort repeatedly (uncached repeated regexp processing)

[JoshuaKGoldberg]

👋 I learned about this project recently - it's very useful, thanks for publishing it!

emoji-toolkit is used in the emoji plugin for Draft.js. That plugin calls to emoji-toolkit's toShort repeatedly, on each emoji in emojibase-data/en/compact.json [shortnameToUnicode.ts link].

Doing so results in ~2 seconds processing time (!) (Mac Studio M1 Max), largely spent in escapeRegexp. You can see a full breakdown of the investigation here: https://github.com/JoshuaKGoldberg/repros/tree/draft-js-emoji-plugin-emoji-toolkit-performance#description-of-issue

]

@marvinhagemeister and I took a look at it and found that replaceAll was calling ns.escapeRegExp repeatedly on the same regular expression cached by unicodeCharRegex. Changing toShort to not need that call reduced that processing time to <200ms.

I took the liberty of sending #57 preemptively - hope that's ok with you! 🙂

Since GitHub doesn't support uploading .json files directly, here's a .zip with before & after profiles: Performance Profiles.zip