Fix support for typed arrays in remote/rpc-server

[miniak]

Adds support for ArrayBuffer and all typed array views (except for DataView) in the remote module.

Also makes the IPC payload much smaller (in the net module tests the payload goes down from approx 13 KB to 5 KB) by using base64 encoding instead of an array of values.

for example when calling electron.remote.getCurrentWindow().getNativeWindowHandle(); this is the JSON with the return value

before:
[{"type":"buffer","value":{"type":"Buffer","data":[32,72,27,0,64,96,0,0]}}]
after:
[{"type":"buffer","value":{"type":"Buffer","data":"oAzG+9B/AAA="}}]

Also calling Buffer.from(value) instead of Buffer.from(value.buffer) would create an unnecessary copy of the data instead of just creating a different view of the underlying memory.

https://nodejs.org/api/buffer.html#buffer_class_method_buffer_from_buffer

Copies the passed buffer data onto a new Buffer instance.

https://nodejs.org/api/buffer.html#buffer_class_method_buffer_from_arraybuffer_byteoffset_length

This creates a view of the ArrayBuffer without copying the underlying memory. For example, when passed a reference to the .buffer property of a TypedArray instance, the newly created Buffer will share the same allocated memory as the TypedArray.

[zcbenz]

There are tests failing related to this change.

[miniak]

@zcbenz should be fixed now

[miniak]

@zcbenz I still have issue with supports TypedArray in api-remote-spec.js. Actually it seems like Int16Array etc is not passed properly without my change, example:

var a = new Int16Array([256, 512, 768, 1024])
a -> Int16Array(4) [256, 512, 768, 1024]
Buffer.from(a) -> [0, 0, 0, 0]

so the test does not even seem to be correct. Just try changing const int16values = new Int16Array([1, 2, 3, 4]) to const int16values = new Int16Array([256, 512, 768, 1024])

you'll get this error

AssertionError [ERR_ASSERTION]: <Buffer 00 00 00 00> deepEqual Int16Array [ 256, 512, 768, 1024 ]

[miniak]

@zcbenz I would suggest adding full support for all typed array types in a separate PR, I am willing to do that.

[zcbenz]

@zcbenz I would suggest adding full support for all typed array types in a separate PR, I am willing to do that.

Sounds good to me 👍

[miniak]

@zcbenz I've changed my mind, let's do it in this PR. It's quite simple

[miniak]

@zcbenz I think I'm done with the changes. Can you please review?

[zcbenz]

👍

[poiru]

See also #8953, cc @tarruda