JSON formatter for cucumber-js #79

worldofchris · 2012-07-24T11:54:48Z

Hi Julien,

As discussed, here are the changes to add JSON output to cucumber.js

Let me know if you need anything else.

Cheers

Chris

…d fixed up the Jasmine tests for the latter

…h what you get from Ruby Cucumber

Have had to deprecate the uri attribute as I can't figure out where to get it from and this is holding up getting on with all the remaining good stuff.

… yet implemented

…rom progress listener

…ct. Don't know how it should be populated.

…rrect feature/scenario with its child steps. Added more Jasmine tests

Conflicts: README.md lib/cucumber/ast/feature.js lib/cucumber/listener.js lib/cucumber/parser.js spec/cucumber/ast/feature_spec.js spec/cucumber/cli_spec.js

… json_formatter

…e feature passing now. Ten to go...

…to the new Gherkin provided formatter. All features are now passing but there not all the jasmine tests.

@jbpros

…ved thanks to @jbpros.

…rent listener class. Added partial support for background. Does not deal with results of background steps yet.

Conflicts: lib/cucumber/cli.js lib/cucumber/cli/configuration.js

…erge from upstream.

jbpros · 2012-07-24T22:05:22Z

spec/cucumber/ast/feature_spec.js

@@ -56,6 +56,12 @@ describe("Cucumber.Ast.Feature", function() {
    });
  });

+  describe("getUri()", function() {


This method is already speced :) The original example description is wrong though, it is about background while we're describing feature.

Apologies. I started adding this before uris were available in cucumber.js. I then did a fetch of the upstream changes which introduced uris and clearly didn't remove this spec. Will do so and push up in the morning.

Removed duplicate:

worldofchris@80b8b56

On 24 Jul 2012, at 23:05, Julien Biezemans wrote:

@@ -56,6 +56,12 @@ describe("Cucumber.Ast.Feature", function() {
});
});

describe("getUri()", function() {

This method is already speced :) The original example description is wrong though, it is about background while we're describing feature.

Reply to this email directly or view it on GitHub:
https://github.com/cucumber/cucumber-js/pull/79/files#r1230303

jbpros · 2012-07-24T22:12:18Z

@dmcaulay Alright then, we'll close your pull request. Thank you for your work.

Ideally, I'd like all the formatters to inherit from the base formatter. However, the JSON formatter is a special beast because - as you said - it's just a wrapper around the formatter supplied by Gherkin itself. Thanks for pointing this, I'll try to find a way to use the base formatter.

jbpros · 2012-07-25T07:26:14Z

features/simple.feature

Do we need this feature?

I'm glad you asked :-) I can't see a great deal of value in it myself. Do you know how widely used it is in cucumber-ruby / jvm land?

Mmm, does it come from -ruby or -jvm?

Well I've only used it with -ruby. I tested my features against cucumber-ruby before writing the -js impl.

That's clever.

This feature is a duplicate of cli.feature#95 anyway. Let's kill it!

Sorry, I think I was talking at cross purposes there. I thought you meant the inclusion of tables in the json output.

The simple.feature is plain old cruft. Will delete it now.

Done:

worldofchris@73b52b6

… allow us to spy on this in jasmine tests

…g those strings in the actual json output that contain the paths to either the cucumber.js source or the sandbox with sanitised versions of the strings. The two elements affected are uri and error_message, the content of which are not the primary concern of the formatter nor of these tests. Replacing them with sanitised versions makes the tests more portable and resiliant to chnge.

jbpros · 2012-08-04T20:58:02Z

lib/cucumber/ast/feature.js

@@ -26,6 +26,15 @@ var Feature = function(keyword, name, description, uri, line) {
      return line;
    },

+    getUri: function getUri() {
+      //TODO: Add uri to feature


As you might have noticed, I don't like lingering TODOs in the codebase.

What can we do to remove this guy? :) Do we care about undefined values?

'undefined' is good. It does what is says on the tin.

jbpros · 2012-08-04T22:05:19Z

I still got the feature suite broken. The specs are also failing.

…t provides json formatter

worldofchris · 2012-08-06T11:43:52Z

The update to package.json, see comment above, should fix the problem with the features and specs failing. It now refers to the latest version of gherkin which includes the missing json formatter.

…no arguments. Removed 'TODO' strings from output of formatter. Where a value is undefined it is passed to the Gherkin formatter as undefined. This as the effect of supressing these values in the final JSON output. Changed underscore_separated_variable to be camelCase.

worldofchris · 2012-08-25T17:04:27Z

Any update on this pull request now the package.json has been fixed?

jbpros · 2012-08-27T07:59:59Z

Sorry for the late reply, got some pretty tough days lately.

It's all green on my side now, thank you! I'll review it properly tonight or so. I spotted some things that can be improved in the features and specs, I'll suggest some changes and hopefully we'll merge your work soon.

Thanks again!

jbpros · 2012-09-06T15:19:10Z

Hey Chris,

Sorry for all the delays.

I've cleaned up the code a bit and re-styled it to follow the project conventions.

In addition to making the JSON feature pass with Cucumber-ruby, I changed the way paths are handled within JSON payloads: placeholders are substituted at runtime with the real paths. I think it's more robust this way.

I'm doing a final review of this; I will probably push a few more small changes before merging this too-long-awaited thing :)

jbpros · 2012-09-06T15:21:28Z

For your convenience, here is the diff between your branch and mine.

worldofchris · 2012-09-06T15:52:43Z

Nice, I like the use of the as you say, more robust.

I've been using 'Should' like this for some time now. It goes back to this from the RSpec book which made a big impression on me when I first read it back in 2010:

"In executable code examples, we are setting an expectation of what should happen rather than what will happen. In fact we've embedded the word should right into RSpec's expectation framework..." p157 - The RSpec Book, Chelimsky et al. 2010

I use 'should' in Specification Workshops to make it clear we are talking about an expectation rather than an assertion.

That's just me giving you the back story for my choice of language, which believe me is something I agonise over.

Makes sense not to use 'should' here it it makes for consistency with the existing code / features.

jbpros · 2012-09-06T19:27:24Z

I used to adhere to these theoretical ideas about should and I still agree with them.

However, Then already expresses the expectation aspect of the steps it prefixes. Also, lots of (not-so-useful) repetitions in scenarios lead to boring scenarios. With time and experience I came to appreciate shorter step sentences in which every single word carries some piece of information or clarification. Should is just a redundancy of Then.

The only solid argument in favor of should in gherkin scenarios is to differenciate a step from a synonymous one used as a Given or When:

Scenario: ambiguous steps
  The first and last steps of this scenario would resolve to the same stepdef while
  we'd probably use a factory in the Given and an assertion in the Then.

  Given I have 3 cucumbers
  When I eat a cucumber
  Then I have 2 cucumbers

Scenario: ambiguity solved thanks to "should"
  Given I have 3 cucumbers
  When I eat a cucumber
  Then I should have 2 cucumbers

However, after 2 years of not using should, I came up with this solution: using clarification words:

Scenario:
  Given I already have 3 cucumbers
  When I eat a cucumber
  Then I have 2 cucumbers

The word already clearly differentiates the context (Given) step from the outcome (When) step.

In the end, it really is a subjective appreciation and I'm definitely not against should.

As you wrote: let's be consistent with the rest of the feature suite.

jbpros · 2012-09-06T20:03:01Z

About time.

jbpros · 2012-09-06T20:05:04Z

Oh and thank you Chris, this is a great contribution.

worldofchris · 2012-09-06T20:53:58Z

Julien,

Thank you so much for your help and support with this contribution, it's been a real pleasure to work on cucumber.js.

Having been a technology manager for the last ten or so years I made a conscious decision this year to return to writing code.

About twenty years ago I cut my teeth writing 6502 assembler on a Commodore 64. Contributing to cucumber.js is the most fun I've had programming since then.

This is down to a number of things, first off the sheer quality of the code base. It is clear, well factored and easy to navigate. Second the quality of the tests, both at the Feature level and the Specs.

Finally, and most importantly, the quality of your reviewing and mentoring / directing of my code once the pull request had gone in. Peer review and mentoring through collaboration is at the heart of building and maintaining great software.

Cheers

Chris

jbpros · 2012-09-06T21:52:37Z

I'm glad you enjoyed contributing to this project. This kind of feedback and cheerful words is what makes it exhilarating and making me think we're doing something useful here.

I know new features are sometimes slow to get into the codebase, compared to some other OSS projects. However, since day one I wanted Cucumber.js to be reliable, efficient, easy to use, close to its brothers (-rb, -jvm, ...) and easy for contributors to pick up. I think all of these criteria are fulfilled at the moment and I mean to keep it that way. I plan on publishing precise rules as to how pull requests should be articulated, what kinds of tests are mandatory, etc.

The build is rarely broken (well.. not the 3 last ones, there were a few issues with the JSON formatter specs on Travis ;)) and very few defects are being reported, even though the userbase is growing. I guess we're in the right direction.

So, thanks again for helping and complying to the rather strong constraints of the project. :)

lock · 2018-10-25T13:16:11Z

This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.

worldofchris added 30 commits May 21, 2012 22:46

Skeletal JSON formatter

a23d0b1

Getting my head around Jasmine Spys...

3f23299

Tests now ignore whitespace / indentation and just check JSON is valid

b4f1b0c

Updated README to show status of my hacks

953f824

Removed redundant ProgressFormatter methods from the JsonFormatter an…

8eb1074

…d fixed up the Jasmine tests for the latter

Updated the output of the tests in the json_formatter feature to matc…

fd940fa

…h what you get from Ruby Cucumber

Command line says there are no commits GUI says there are…

58cd346

Now have two passing scenarios.

e613101

Have had to deprecate the uri attribute as I can't figure out where to get it from and this is holding up getting on with all the remaining good stuff.

Now have three scenarios passing albeit with steps:match:location not…

a4a6b75

… yet implemented

Four passing scenrios. Need to add error_message to EVENT though

deb9e6b

Added Scenarios for failing and passing steps

f91bea1

Added more scenarios to the tests. Removed redundant methods copied f…

51c52af

…rom progress listener

Removing redundant dir

b3dc4ad

Added more tests. Pushed the placeholder URI down into the event obje…

e042dae

…ct. Don't know how it should be populated.

Added tests for multiple features. Fixed functionality to populate co…

7035ea0

…rrect feature/scenario with its child steps. Added more Jasmine tests

Added note to readme to clarify current status of JSON formatter

3a03a91

Merge remote-tracking branch 'upstream/master'

7dfe6e3

Conflicts: README.md lib/cucumber/ast/feature.js lib/cucumber/listener.js lib/cucumber/parser.js spec/cucumber/ast/feature_spec.js spec/cucumber/cli_spec.js

Removed redunant require of Cucumber-js

516c5b9

Updated README to show that we are now using the new Gherkin provided…

109d47d

… json_formatter

Merge remote-tracking branch 'upstream/master'

ebafde0

Refactored to use the newly available Gherking json_formatter. Got on…

a7d9747

…e feature passing now. Ten to go...

Adding renamed classes which now wrap the Gherkin JSON formatter

e0cf8a6

Removed my now redundant json_formatter code and replaced with calls …

95b7aba

…to the new Gherkin provided formatter. All features are now passing but there not all the jasmine tests.

Jasmine tests all passing. Ugly relative path to require Gherkin remo…

31f260b

…ved thanks to @jbpros.

Refactored json_formatter_wrapper to use functionality provided by pa…

0e5d28b

…rent listener class. Added partial support for background. Does not deal with results of background steps yet.

Merge remote-tracking branch 'upstream/master'

6ef524d

Conflicts: lib/cucumber/cli.js lib/cucumber/cli/configuration.js

Added Work In Progress DocString formatting. Fixed fail in previous m…

f5c2112

…erge from upstream.

Added DocString support

8272e2f

Added support for scenario level tags

dd5eb0a

Added support for Feature level tags

0b8d93e

jbpros reviewed Jul 24, 2012
View reviewed changes

jbpros reviewed Jul 25, 2012
View reviewed changes

worldofchris added 6 commits July 25, 2012 10:36

Moved Gherkin JSON formatter back into Wrapper class. Added getter to…

243bce8

… allow us to spy on this in jasmine tests

Removed redundant spec for uri method

80b8b56

Removed spurious spec which has nothing to do with JSON formatter

b827efb

Renamed JSON Formatter to make it consistent with existing formatters

5753058

Removed spurious feature file

73b52b6

jbpros reviewed Aug 4, 2012
View reviewed changes

worldofchris added 2 commits August 6, 2012 10:14

Merge remote-tracking branch 'upstream/master'

8cab836

Updated package.json to indicate dependency on version of gherkin tha…

64d5438

…t provides json formatter

jbpros closed this in 7fd3bdd Sep 6, 2012

lock bot locked as resolved and limited conversation to collaborators Oct 25, 2018

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

JSON formatter for cucumber-js #79

JSON formatter for cucumber-js #79

worldofchris commented Jul 24, 2012

jbpros Jul 24, 2012

worldofchris Jul 24, 2012

worldofchris Jul 25, 2012

jbpros commented Jul 24, 2012

jbpros Jul 25, 2012

worldofchris Jul 25, 2012

jbpros Jul 25, 2012

worldofchris Jul 25, 2012

jbpros Jul 25, 2012

worldofchris Jul 25, 2012

worldofchris Jul 25, 2012

jbpros Aug 4, 2012

worldofchris Aug 6, 2012

jbpros commented Aug 4, 2012

worldofchris commented Aug 6, 2012

worldofchris commented Aug 25, 2012

jbpros commented Aug 27, 2012

jbpros commented Sep 6, 2012

jbpros commented Sep 6, 2012

worldofchris commented Sep 6, 2012

jbpros commented Sep 6, 2012

jbpros commented Sep 6, 2012

jbpros commented Sep 6, 2012

worldofchris commented Sep 6, 2012

jbpros commented Sep 6, 2012

lock bot commented Oct 25, 2018

JSON formatter for cucumber-js #79

JSON formatter for cucumber-js #79

Conversation

worldofchris commented Jul 24, 2012

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

jbpros commented Jul 24, 2012

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

jbpros commented Aug 4, 2012

worldofchris commented Aug 6, 2012

worldofchris commented Aug 25, 2012

jbpros commented Aug 27, 2012

jbpros commented Sep 6, 2012

jbpros commented Sep 6, 2012

worldofchris commented Sep 6, 2012

jbpros commented Sep 6, 2012

jbpros commented Sep 6, 2012

jbpros commented Sep 6, 2012

worldofchris commented Sep 6, 2012

jbpros commented Sep 6, 2012

lock bot commented Oct 25, 2018