WebMock is on globally now; no need to disable it

[pd]

Introduced by conflicts in how #175 and #196 juggled webmock usage.

Closes #207.

[iainbeeston]

Ah I see - that's because of me.

👍

[pd]

Eek, still some errors for the other JSON backends. Will figure it out soon.

[iainbeeston]

Each json backend has it’s own parsing error classes - they all inherit from standard error though (I had the same problem when writing tests to check initialize_data, across different backends)=

[iainbeeston]

Incidentally, I put in place a “refute_block” helper method in the initialize_data test - that might help you (and I meant to move it into test_helper.rb, so it could be reused, but never did)=

[pd]

I'm not thrilled with how I just dealt with it in f263094, but I really did want the test to be fairly explicit about "the JSON parse error should be allowed to bubble up", not just "something could raise here". Checking for the currently defined modules is the best I could come up with ... =\

[RST-J]

If someone has a better idea some time we can change that then.
👍

[iainbeeston]

Actually, that’s not bad - can you put that in test_helper? Perhaps rename it as “assert_raises_parse_error”? I’d like to reuse it in a test I wrote for initialize_data=

[iainbeeston]

Tell you what, I'll refactor both.

Merging