Use extensible protocol for serialisation

[garethmdavies]

We are trying to serialise dates into our dynamodb table and plan to use prismatic schema to deserialise and coerce the data out again. Using protocols allows us to achieve this by implementing a date to string conversion locally without affecting the faraday library. An example of the date serialisation that we wish to use has been included in the test.

[ptaoussanis]

Hi there, thanks for the PR!

Should be okay in principle but will need some time to take a proper look, will get back to you. Cheers :-)

[garethmdavies]

Great thanks.

[garethmdavies]

Did you get chance to look at this?

[ricardojmendez]

I just cloned it myself, but it looks like that new test is failing.

failure in (main.clj:367) : taoensso.faraday.tests.main
(expect
 (assoc t :date "Thu Jan 01 01:00:00 GMT 1970")
 (update-t {:update-map {:date [:put (java.util.Date. (long 0))]}}))

           expected: {:boolF false, :num 1, :date "Thu Jan 01 01:00:00 GMT 1970", :vec ["a" 1 false nil], :strset #{"a" "b" "c"}, :numset #{4 13 6 12}, :boolT true, :string "string", :null nil, :id 14, :map {:k1 "v1", :k2 "v2", :k3 "v3"}}
                was: {:boolF false, :num 1N, :date "Thu Jan 01 02:00:00 EET 1970", :vec ["a" 1N false nil], :strset #{"a" "b" "c"}, :numset #{4N 13N 6N 12N}, :boolT true, :string "string", :null nil, :id 14N, :map {:k1 "v1", :k2 "v2", :k3 "v3"}}

           in expected, not actual: {:date "Thu Jan 01 01:00:00 GMT 1970"}
           in actual, not expected: {:date "Thu Jan 01 02:00:00 EET 1970"}

[ricardojmendez]

There also seems to be a conflict likely because of the recent dependency updates. Could you please re-base against the latest, re-run the tests and re-submit?

Thanks!

[garethmdavies]

Rebased and fixed the test.

[ricardojmendez]

All tests pass both locally and when running them on my Gitlab CI. Merging.

[ptaoussanis]

@taffowl Sorry for the long delay in handling this, thanks again for the PR.

@ricardojmendez Great, thanks Ricardo! Please note that you'll want to rebase against master since I rewrote history on this last commit.

About that: could I possibly request that any future PRs you commit use the same repo style, e.g.: [#95] Short description (@submitter).

This style only takes a moment longer when accepting a commit, but can save significant effort when cutting new releases, and can be helpful when scanning history, etc.

Much appreciated, cheers! :-)

[ricardojmendez]

@ptaoussanis Got it, will ping you before the next one just in case.

[ptaoussanis]

Oh, no pinging necessary - please feel free to do as you've been doing. It's a huge help, thanks! :-)