Strip the result body when creating default HEAD responses

[aturon]

If you don't specify a HEAD behavior but you do provide GET, Tide will automatically use the GET behavior. But it retains the body, which should be stripped.

[tzilist]

@aturon did a little diggin into this, not sure if I am reading this right either, but it seems like hyper handles this for us. In the current code on master here are some example responses from curl
I'm running the named_path example.

The get request:

url --get --url http://127.0.0.1:8000/add_two/2 --header 'content-type: application/json' -v

Returns

*   Trying 127.0.0.1...
* TCP_NODELAY set
* Connected to 127.0.0.1 (127.0.0.1) port 8000 (#0)
> GET /add_two/2 HTTP/1.1
> Host: 127.0.0.1:8000
> User-Agent: Mozilla/5.0 (compatible; MSIE 9.0; Windows NT 6.1; Trident/5.0)
> Accept: */*
> Referer:
> content-type: application/json
>
< HTTP/1.1 200 OK
< content-type: text/plain
< content-length: 15
< date: Mon, 19 Nov 2018 04:53:55 GMT
<
* Connection #0 to host 127.0.0.1 left intact
2 plus two is 4

and the head request:

curl --head --url http://127.0.0.1:8000/add_two/2 --header 'content-type: application/json' -v

returns

*   Trying 127.0.0.1...
* TCP_NODELAY set
* Connected to 127.0.0.1 (127.0.0.1) port 8000 (#0)
> HEAD /add_two/2 HTTP/1.1
> Host: 127.0.0.1:8000
> User-Agent: Mozilla/5.0 (compatible; MSIE 9.0; Windows NT 6.1; Trident/5.0)
> Accept: */*
> Referer:
> content-type: application/json
>
< HTTP/1.1 200 OK
HTTP/1.1 200 OK
< content-type: text/plain
content-type: text/plain
< content-length: 15
content-length: 15
< date: Mon, 19 Nov 2018 04:53:33 GMT
date: Mon, 19 Nov 2018 04:53:33 GMT

<
* Connection #0 to host 127.0.0.1 left intact

Perhaps I am missing something though. Thoughts?

[DeltaManiac]

@tzilist

I had observed the same behavior while testing #31.
Guess i missed something too :)

[aturon]

That's very interesting. I haven't been able to track down where this is happening in Hyper...

[aturon]

@seanmonstar maybe you can provide a quick answer: does Hyper have built-in treatment for HEAD requests and if so, what is it?

[seanmonstar]

hyper will try to prevent illegal HTTP semantics, like sending a body in response to HEAD requests, or in 204/304 status codes.

It may still be useful to recognize a HEAD request is different from a GET in cases where preparing the body may have been expensive.

[tzilist]

@aturon if this is the case, should we just write a test case to ensure that HEAD requests always strip the body? We can just rely on hyper to handle this and the test case to make sure it doesn't break.

[aturon]

@tzilist Yep, that seems reasonable for now! I'd also suggest adding a comment to make clear that's what we're doing.

[fairingrey]

I attempted to write a test for this given the mock service provided through http_service_mock but I'm running into something strange... There's more details in my PR: #179

[WSeegers]

Has this issue been fully resolved?

[yoshuawuyts]

@WSeegers I think it may have been in #179. Going to go ahead and close this!