Cleanup httpmock #29
Cleanup httpmock #29
Conversation
|
|
||
| type Matcher func(req *http.Request) bool | ||
|
|
||
| func REST(method, p string) Matcher { |
There was a problem hiding this comment.
What do we think about the naming scheme here? Should we prefix/postfix Match/Matcher to this and the GQL function?
| stubs []*stub | ||
| requests []*http.Request |
There was a problem hiding this comment.
I purposefully un-exported these and added a Requests() accessor function. It felt smart to move this direction incase we want to change the underlying implementation.
| } | ||
| if stub != nil { | ||
| r.mu.Unlock() | ||
| dr, _ := httputil.DumpRequestOut(req, true) |
There was a problem hiding this comment.
I think using DumpRequestOut leads to nicer looking error messages.
|
|
||
| type Responder func(req *http.Request) (*http.Response, error) | ||
|
|
||
| func HTTPResponse(status int, header *map[string][]string, body interface{}, cb func(*http.Request)) Responder { |
There was a problem hiding this comment.
What do we think about the naming scheme for Responders? Should we prefix/postfix them with Response/Responder?
| } | ||
| } | ||
|
|
||
| func RESTResponse(body interface{}, cb func(map[string]interface{})) Responder { |
There was a problem hiding this comment.
I think having HTTPResponse and RESTResponse to be a bit confusing. Open to better naming here.
There was a problem hiding this comment.
Any less confusing if you call RESTResponse a JSONResponse? Afterall, it's just the response itself which happens to be JSON. Perhaps HTTPResponse could be RAWResponse and uses for non-JSON payloads - perhaps with a content-type? More focus on the payload than how it was retrieved.
|
Closing as we decided against exporting |
This PR gets the package formerly known as
httpmockinto a state that it can be made public and moved out of theinternalfolder.Some major changes:
transportmockI left comments for open questions inline.
cc #28