encoder.encode returns a Map instead of a Buffer when data is large

[CherryDT]

In case CBOR fails, a Map is unexpectedly returned from encoder.encode. This is because errors in _cborEncode are not handled, but the original object is returned.

Further debugging showed that the swallowed error from CBOR was:

Error: Push Error
    at Encoder._transform (/home/ubuntu/app/project/node_modules/cbor/lib/encoder.js:72:33)
    at Encoder.Transform._read (_stream_transform.js:186:10)
    at Encoder.Transform._write (_stream_transform.js:174:12)
    at doWrite (_stream_writable.js:387:12)
    at writeOrBuffer (_stream_writable.js:373:5)
    at Encoder.Writable.write (_stream_writable.js:290:11)
    at JSONStreamCompressEncoder._cborEncode (/home/ubuntu/app/project/node_modules/json-stream-compression/lib/encoder.js:21:24)
    at JSONStreamCompressEncoder.encode (/home/ubuntu/app/project/node_modules/json-stream-compression/lib/encoder.js:84:42)
    at saveHistory (/home/ubuntu/app/project/lib/history.js:64:27)
    at /home/ubuntu/app/project/lib/history.js:131:4

Example:

const { Encoder } = require('json-stream-compression')
const arr = []
for(let i = 0; i < 20000; i++) arr.push(1)
const result = new Encoder().encode(arr) // throws error

This actually comes from the fact that you are exploiting the stream buffer here instead of processing incoming data in the stream.