doc: fix trivial mistakes in stream document.

[fmy]

Checklist

• documentation is changed or added
• commit message follows commit guidelines

Affected core subsystem(s)

doc

Description of change

1st commit:
In stream.Writable, when decodeStrings option is false, the chunk _write() function received is not always be Buffer, it depends types of the sources.

const {Writable} = require('stream');
class MyWritable extends Writable {
  _write(chunk, enc, cb) {
    console.log(chunk);
    console.log('typeof chunk : ' + typeof chunk);
    cb();
  }
}

const w = new MyWritable({decodeStrings: false});
w.write('abc');
w.write(Buffer.from('abc'));

// => abc
// => typeof chunk : string
// => <Buffer 61 62 63>
// => typeof chunk : object

2nd commit:
This is a trivial fix, 'unexpected' fits this context.

[Fishrock123]

How can one tell what the type is? I thought these were always Buffers lol.

Do you mean it is different for Object Mode?

[fmy]

When decodeStrings is true, string is converted into Buffer here.
https://github.com/nodejs/node/blob/v6.9.3-proposal/lib/_stream_writable.js#L285-L288
Of course Buffer is still Buffer. So the chunk always be a Buffer.

When decodeStrings is false, there is no conversion before _write() received it.
So, string keeps string, not always be Buffer.
Please see upper sample codes.

[mcollina]

@fmy you are right. I would phrase that differently thought, because Writable  can be used also without pipe(). How about:

will remain the same object that is passed to .write().

[fmy]

will remain the same object that is passed to .write().

@mcollina thanks! i will change words.

[italoacasas]

@nodejs/streams

[mcollina]

@fmy you are right. I would phrase that differently thought, because Writable  can be used also without pipe(). How about:

will remain the same object that is passed to .write().

[fmy]

I changed some words.
@mcollina Thanks for your review.

[mcollina]

LGTM

[jasnell]

Landed in 37d6052