v4: Should send bytes instead of UTF-16 length

[MCPE-PC]

Issue

Describe your issue here

For Node Redis v4, the code seems to be using String.prototype.length to calculate the byte length of an argument to load on the payload
https://github.com/NodeRedis/node-redis/blob/c72aab2fc29a9cc994f6b810aeb10a4fe1863f29/lib/commands-queue.ts#L52-L53
It should load the byte length instead. Buffer.byteLength(String.prototype) is one of its ways.

---

Environment

• Node.js Version: 16.0.0

• Redis Version: 6.2.4

• Platform: Linux MCPEPC-DESKTOP 5.8.0-59-generic #66~20.04.1-Ubuntu SMP Thu Jun 17 11:14:10 UTC 2021 x86_64 x86_64 x86_64 GNU/Linux

[NovaAtWarren]

As per my comment here, it seems to be working correctly already. Thank you still, though, because (if the test gets merged) we'll know if it breaks.

[MCPE-PC]

@NovaAtWarren Probably it is not working correct, here. I was testing the v4 and encountered wrong bytes issue on Redis. Small patching of using Buffer.byteLength() fixed it perfectly.

[NovaAtWarren]

Turns out there were two places that calculated the length of a string
https://github.com/NodeRedis/node-redis/blob/c72aab2fc29a9cc994f6b810aeb10a4fe1863f29/lib/commands-queue.ts#L47
and
https://github.com/NodeRedis/node-redis/blob/c72aab2fc29a9cc994f6b810aeb10a4fe1863f29/lib/commands-queue.ts#L53

The fix you wrote worked perfectly fine, I just didn't encounter it because I put the special character in the wrong place to encounter the specific issue you did.

Thank you for the help, re-opening PR for review and merging.

[MCPE-PC]

@NovaAtWarren Cool. But what if the type of the original argument is ArrayBuffer view? Let's see,
https://github.com/nodejs/node/blob/eaca13ddc9092e831b9b45180c6e0ee1f3292e5e/src/node_buffer.cc#L698
Here, Buffer.byteLength() with the string typed argument is slower than the one with the ArrayBuffer view argument (it has byteLength property by design, i. e. Buffer.from(str).byteLength)

I'd like to suggest some changes to your PR(#1630).

[leibale]

@MCPE-PC nice catch, thanks