Neither of encodings are correct

This is related to, but separate from #113 as it also affects non-browsers and is about correctness, not performance

From the Readme:

https://github.com/mafintosh/streamx/blob/d04ef953539f46b665d11c2be298790f636725ad/README.md?plain=1#L49

Node.js setEncoding doc: https://nodejs.org/api/stream.html#readablesetencodingencoding
Node.js Bufffer encodings doc: https://nodejs.org/api/buffer.html#buffers-and-character-encodings

But neither of encodings work:

• hex is unsupported in browsers
• utf16le, base64 and base64url are unsupported at all
• latin1 / ascii behavior on browsers differs from Node.js
• utf-8 is incorrect in general until text-encoding is bumped to ^1.2.6
  See https://github.com/holepunchto/text-decoder/issues/10, https://github.com/holepunchto/text-decoder/issues/11
• utf-8 is still incorrect even on 1.2.6 when the input is truncated mid-character
  e.g. if this line in tests is replaced with b4a.from('hællå wø').subarray(0, -1), the test fails (misses the last codepoint)

meta: I opened a single issue for all of those inconsistencies, that should be cleaner than filing that as separate ones.
This can be split into sub-issues.

Start by comparing the README encoding claims with the linked Node.js encoding documentation and inspect the encoding tests in test/readable.js. Check each listed encoding across browser and non-browser behavior, including truncated UTF-8 input. Done means the supported encodings match Node.js behavior and the cited test cases pass consistently.