e2e tests: additional checks for application/json response headers

[piotr-iohk]

Issue Number

ADP-853

Overview

• I have added additional checks in e2e tests checking that responses return 'content-type' => ['application/json;charset=utf-8'] header

Comments

Only 204 reponses do not return 'content-type' => ['application/json;charset=utf-8'] header. I assume that it is OK, since 204 returns no content?

I've temporarily cherry-picked c832eb5 from #2608, to be able to run all tests.

[rvl]

Could we refactor the response assertions to keep it maintainable?
Almost every request will have application/json content-type, except for a handful.

[piotr-iohk]

Could we refactor the response assertions to keep it maintainable?

@rvl you mean something like this -> 9e65a6c

[rvl]

Yes - a little. But it would be better if the response content type were checked automatically, so there doesn't need to be this assertion in every single test case. It could be tested when parsing the response (i.e. only decode as json if that's the content-type), or as part of the http status code check.

[piotr-iohk]

Ok, gotcha. Currently it is like this:

expect(response).to have_http 201
expect(response).to have_expected_headers

which I could change to something like:

expect(response).to be_correct 201

where I could check both assertions, with additional logic that when response code is 204 the application/json header is not expected.

I guess the pros is that it is more concise, cons that two assertions combined in one... which is sort of anti-pattern I suppose. 🤔

[piotr-iohk]

Yes - a little. But it would be better if the response content type were checked automatically, so there doesn't need to be this assertion in every single test case.

Done in 27df4ce.

[rvl]

👍

[rvl]

I don't think it's an anti-pattern. I think it's better to explicitly test more high level things.

[piotr-iohk]

bors r+

[iohk-bors]

Build succeeded:

• buildkite/cardano-wallet
• ci/hydra-build:required