Merge 28e0442 into 8f9a07a

.codeclimate.yml

@@ -0,0 +1,5 @@
+version: "2"
+plugins:
+  credo:
+    enabled: true
+    channel: "beta"

.gitignore

@@ -30,3 +30,9 @@ erl_crash.dump
 # Also ignore archive artifacts (built via "mix archive.build").
 *.ez
 
+<<<<<<< 26cc7c1a7c0e718707208ab099d9d87efb8925e8
+=======
+# Ignore benchmark output
+#/bench/output
+/bench/snapshots
+>>>>>>> bench

.travis.yml

@@ -1,15 +1,26 @@
+env:
+  global:
+    - CC_TEST_REPORTER_ID=a669f9b84dcb893c8df72d85c56658c03e59c523ef619714d37b6523d35698aa
 language: elixir
 env: MIX_ENV=test
 notifications:
   recipients:
     - carlos@carloslage.net
 elixir:
   - 1.6
+before_script:
+  - curl -L https://codeclimate.com/downloads/test-reporter/test-reporter-latest-linux-amd64 > ./cc-test-reporter
+  - chmod +x ./cc-test-reporter
+  - ./cc-test-reporter before-build
 script:
+  - mix dialyzer
   - MIX_ENV=test mix test
 after_script:
   - mix deps.get --only docs
   - MIX_ENV=docs mix inch.report
+  - MIX_ENV=test mix coveralls --verbose
+  - MIX_ENV=test mix coveralls.json --verbose
+  - MIX_ENV=test mix coveralls.travis
+  - ./cc-test-reporter after-build -r "r$CC_TEST_REPORTER_ID" -t excoveralls --exit-code $TRAVIS_TEST_RESULT
 after_success:
-  - MIX_ENV=test mix coveralls
-  - MIX_ENV=test mix coveralls.post --token  ilbBz5SvaqJqUBZx9KcIcuM8C1D3BCsB8  --branch $TRAVIS_BRANCH --name "TravisCI"
+  - mix credo --strict

README.md

@@ -1,29 +1,114 @@
+# License
+The Elixir JSON library is available under the [BSD 3-Clause aka "BSD New" license](http://www.tldrlegal.com/l/BSD3)
+
 # Elixir JSON
 
-[![Coverage Status](https://coveralls.io/repos/github/cblage/elixir-json/badge.svg?branch=master)](https://coveralls.io/github/cblage/elixir-json?branch=develop)
-[![Hex.pm](https://img.shields.io/hexpm/v/json.svg?style=flat-square)](https://hex.pm/packages/json)
-[![Build Status](https://travis-ci.org/cblage/elixir-json.svg?branch=master)](https://travis-ci.org/cblage/elixir-json) [![Inline docs](http://inch-ci.org/github/cblage/elixir-json.svg)](http://inch-ci.org/github/cblage/elixir-json) 
-[![Hex.pm](https://img.shields.io/hexpm/dt/json.svg?style=flat-square)](https://hex.pm/packages/json) 
-
-
+[![Hex.pm](https://img.shields.io/hexpm/dt/json.svg?style=flat-square)](https://hex.pm/packages/json)[![Build Status](https://travis-ci.org/cblage/elixir-json.svg?branch=master)](https://travis-ci.org/cblage/elixir-json) [![Maintainability](https://api.codeclimate.com/v1/badges/43b6e8c25e036558ccb6/maintainability)](https://codeclimate.com/github/cblage/elixir-json/maintainability) [![Test Coverage](https://api.codeclimate.com/v1/badges/43b6e8c25e036558ccb6/test_coverage)](https://codeclimate.com/github/cblage/elixir-json/test_coverage) 
+
+
+[![Hex.pm](https://img.shields.io/hexpm/v/json.svg?style=flat-square)](https://hex.pm/packages/json) [![Inline docs](http://inch-ci.org/github/cblage/elixir-json.svg)](http://inch-ci.org/github/cblage/elixir-json)
+
 
 This library provides a natively implemented JSON encoder and decoder for Elixir.
 
 All contributions are welcome.
-## Installing
 
-Simply add ```{:json, "~> 1.2"}``` to your project's ```mix.exs``` file, in the dependencies list and run ```mix deps.get json```.
+# Before you install
+
+When dealing with smaller `file.json ~ 14KB` payloads, `JSON v1` handles the processing consistently performant, with
+ a much smaller deviation, and absolutely no real-world absolute performance differences with `Jason`.
+
+However, with often unusually large `file.json > 5MB` payloads, and if the processing speed for those payloads is 
+paramount to you (ie, processing them in a real-time manner vs using them in migration scripts or whatever), `JSON v1` 
+currently significantly slower when compared to `Jason`.
+
+#### Small payload `bench/data/utf-8-unescaped.json < 30KB` benchmark results for `JSON v1` and `Jason`
+
+As you can see below, both libraries handle "regular" small `json` payloads beautifully.
+
+`JSON.decode` calls while usually being a bit slower than their `Jason` counterparts, they are more consistently 
+performant, with a much smaller deviation. 
+
+So I would actually advise using `JSON.decode` and `JSON.encode` for smaller payloads.
+
+For `JSON.encode` vs `Jason.encode`, the difference is so minimal, it's not worth arguing about. 
+
+| Library        	| Average  	| Deviation 	| median   	| minimum  	| maximum   	|
+|----------------	|----------	|-----------	|----------	|----------	|-----------	|
+| `JSON.decode`  	| 13.35 ms 	| ±45.95%   	| 11.47 ms 	| 8.17 ms  	| 99.49 ms  	|
+| `Jason.decode` 	| 0.26 ms  	| ±246.84%  	| 0.170 ms 	| 0.150 ms 	| 64.35 ms  	|
+| `JSON.encode`  	| 1.32 ms  	| ±225.81%  	| 0.58 ms  	| 0.28 ms  	| 103.70 ms 	|
+| `Jason.encode` 	| 0.30 ms  	| ±316.71%  	| 0.163 ms 	| 0.147 ms 	| 82.68 ms  	|
+
+#### Full `benchee` reports for `bench/data/utf-8-unescaped.json < 30KB`:
+ - `decode`: https://bit.ly/2GVV8dy
+ - `encode`: https://bit.ly/2v5W0H0
+
+
+#### Large payload `bench/data/issue-90.json ~ 8MB` benchmark results for `JSON v1` and `Jason`
+
+However, with often unusually large `file.json > 5MB` payloads, and if the processing speed for those payloads is paramount to you 
+(ie, processing them in a real-time manner vs using them in migration scripts or whatever), then 
+`JSON v1` would not be the best choice when compared to `Jason`.
+
+| Library        	| Average  	| Deviation 	| median   	| minimum  	| maximum   	|
+|----------------	|----------	|-----------	|----------	|----------	|-----------	|
+| `JSON.decode`  	| 8.93 s	| ±5.71%	 	| 8.96 s 	| 8.10 s  	| 9.57 s	  	|
+| `Jason.decode` 	| 0.182 s	| ±21.60%		| 0.171 s	| 0.139 s 	| 0.42 s	  	|
+| `JSON.encode`  	| 5.51 s	| ±18.10%	  	| 5.24 s 	| 4.32 s 	| 7.36 s 	    |
+| `Jason.encode` 	| 0.186 s 	| ±26.41%   	| 0.173 s	| 0.122 s	| 0.38 s	  	|
+
+### Full `benchee` reports for `bench/data/issue-90.json ~ 8MB`:
+ - `decode`: https://bit.ly/2HxReEP
+ - `encode`: https://bit.ly/2HuR0OM
+
+# Plan of action for `Elixir JSON v2`
+
+I am currently working on a solution for this problem in `JSON v2`.
+You can follow the process here: https://github.com/cblage/elixir-json/pull/52 
+
+# Interim Solution
+
+To processes these large payloads adding the `Jason` lib to your dependencies (without hopefully removing `JSON` 
+for the smaller payloads :sweat_smile:): 
+ - `Jason@Hex.pm`: http://hex.pm/packages/jason
+ - `Jason@GitHub`: https://github.com/michalmuskala/jason
+
+After installing `Jason`, you then use `JSON.decode` and `JSON.encode` for small your small `30KB range json` payloads due to the reasons mentioned above.
+
+While `Elixir JSON v2` is not ready to processs the bigger `>5MB json` payloads in time-sensitive operations, you go for `Jason.decode` and `Jason.encode`.
+
+Thanks for the comprehension,
+Carlos Brito Lage
+
+## Example 
 
-### Example for a project that already uses [Plug](https://github.com/elixir-plug/plug):
 ```elixir
-defp deps do
-  [{:cowboy, "~> 1.0.0"},
+ [
+   {:cowboy, "~> 1.0.0"},
    {:plug, "~> 1.0"},
-   {:json, "~> 1.2"}]
-end
+   {:json, "~> 1.2"},
+   {:jason, "~> 1.0"},
+ ]
+```
+
+You can find its documentation here: https://hexdocs.pm/jason/readme.html
+
+#Installing
+
+Simply add ```{:json, "~> 1.2"}``` to your project's ```mix.exs``` file, in the dependencies list and run ```mix deps.get json```.
+
+## Example for a project that already uses [Plug](https://github.com/elixir-plug/plug):
+
+```elixir
+[
+  {:cowboy, "~> 1.0.0"},
+  {:plug, "~> 1.0"},
+  {:json, "~> 1.2"},
+]
 ```
 
-## Usage
+# Usage
 
 Encoding an Elixir type
 ```elixir
@@ -54,6 +139,3 @@ Decoding a list from a string that contains JSON
   list["key"]
   # "this will be a value"
 ```
-
-## License
-The Elixir JSON library is available under the [BSD 3-Clause aka "BSD New" license](http://www.tldrlegal.com/l/BSD3)

bench/decode.exs

@@ -1,7 +1,7 @@
 decode_jobs = %{
   "Jason"  => fn {json, _} -> Jason.decode!(json) end,
   #"Poison" => fn {json, _} -> Poison.decode!(json) end,
-  "JSX"    => fn {json, _} -> JSX.decode!(json, [:strict]) end,
+  #"JSX"    => fn {json, _} -> JSX.decode!(json, [:strict]) end,
   #"Tiny"   => fn {json, _} -> Tiny.decode!(json) end,
   #"jsone"  => fn {json, _} -> :jsone.decode(json) end,
   "JSON"   => fn {json, _} -> JSON.decode!(json) end,
@@ -12,11 +12,11 @@ decode_inputs = [
   #"GitHub",
   #"Giphy",
   #"GovTrack",
-  "Blockchain",
+  #"Blockchain",
   #"Pokedex",
-  "JSON Generator",
-  "JSON Generator (Pretty)",
-  "UTF-8 escaped",
+  #"JSON Generator",
+  #"JSON Generator (Pretty)",
+  #"UTF-8 escaped",
   "UTF-8 unescaped",
   "Issue 90",
 ]

bench/encode.exs

@@ -1,21 +1,21 @@
 encode_jobs = %{
   "Jason"          => &Jason.encode_to_iodata!/1,
-  "Jason strict"   => &Jason.encode_to_iodata!(&1, maps: :strict),
-  "Poison"         => &Poison.encode_to_iodata!/1,
-  "JSX"            => &JSX.encode!/1,
+  #"Jason strict"   => &Jason.encode_to_iodata!(&1, maps: :strict),
+  #"Poison"         => &Poison.encode_to_iodata!/1,
+  #"JSX"            => &JSX.encode!/1,
   #"Tiny"           => &Tiny.encode!/1,
   #"jsone"          => &:jsone.encode/1,
   "JSON"           => &JSON.encode!/1,
-  "term_to_binary" => &:erlang.term_to_binary/1,
+  #"term_to_binary" => &:erlang.term_to_binary/1,
 }
 
 encode_inputs = [
-  "GitHub",
-  "Giphy",
-  "GovTrack",
-  "Blockchain",
-  "Pokedex",
-  "JSON Generator",
+  #"GitHub",
+  #"Giphy",
+  #"GovTrack",
+  #"Blockchain",
+  #"Pokedex",
+  #"JSON Generator",
   "UTF-8 unescaped",
   "Issue 90",
 ]