Octet-stream response bodies should be decoded to a hex Buffer

[kevinburke]

I have a server that responds with binary data and an octet-stream header:

$ http POST localhost:4249/v1/sign 'data=aGVsbG8gd29ybGQ='
HTTP/1.1 200 OK
Content-Length: 64
Content-Type: application/octet-stream
Date: Thu, 16 Nov 2017 05:04:11 GMT
Server: signing-server/0.1
X-Request-Duration: 141 µs

�̰3郞�W�5-��)�Z��`�׋�l����Ʃ2 ���n�|qu��<�4>

The response is a Buffer, so nock converts it to hex before storing in a file. I've truncated the response here and above, but it should be enough to get the idea.

    {
        "scope": "http://localhost:4249",
        "method": "POST",
        "path": "/v1/sign",
        "body": {
            "data": "aGVsbG8gd29ybGQ=",
            "name": "WebApp"
        },
        "status": 200,
        "response": "f0ccb033e9839eb657ae352d9bd329cf5a0193e660ded78bcf6cacbf4950acd3c6a9322098f9dc6e9a107c7175c3e83c19c7343e39d91d",
        "rawHeaders": [
            "Content-Type",
            "application/octet-stream",
            "Server",
            "signing-server/0.1",
            "X-Request-Duration",
            "178 µs",
            "Date",
            "Thu, 16 Nov 2017 05:00:17 GMT",
            "Content-Length",
            "64",
            "Connection",
            "close"
        ]
    }

When reading back from the file, I expect it to be converted back into a Buffer via Buffer.from(response.body, 'hex'). Instead it gets converted into a normal Buffer, e.g. the first byte is 'f', the second byte is '0', instead of one byte 0xf0.

I can work around this for the time being by setting Content-Encoding: identity, which triggers the right nock logic. But it would seem preferable to e.g. read the application/octet-stream response header and try to convert from hex based on that.

[gr2m]

Hey @kevinburke can you check if #1067 fixes your problem?

[kevinburke]

Apologies but I've since moved on from that engagement and probably won't have time to test :( best of luck!

A way to test would probably be to start with the JSON I put up above, read it back into memory, and verify the length of the buffer in memory matches the Content-Length header.

[lock]

This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue and add a reference to this one if it’s related. Thank you!