[rspec-users] [Cucumber] any examples of navigating cucumber parse tree from external script

[Korny Sietsma]

Hi folks,
I'm writing some scripts to integrate our cucumber features with stories stored in a wiki, and I'm hoping to use cucumber's parser to parse the features rather than doing it manually.  (I don't really care about the feature contents much, just scenario titles and tags)

I've worked out how to parse the features:
      Cucumber.load_language('en')
      features = Cucumber::Ast::Features.new
      parser = Cucumber::Parser::FeatureParser.new

      feature_files = Dir["#{FEATURE_DIR}/**/*.feature"]

      feature_files.each do |f|
        puts "parsing feature file #{f}"
        features.add_feature(parser.parse_file(f))
      end

But now I'm digging in to the whole ast visitor thing, and it's getting quite complex to *do* stuff with the features once I've parsed them.

I'm sure I can work this out myself, given time, but I was wondering if there are any code examples out there to save me some of the time/effort?  Anyone else tried parsing features like this from outside Cucumber itself?

Thanks!

- Korny

--
Kornelis Sietsma  korny at my surname dot com
"Every jumbled pile of person has a thinking part
that wonders what the part that isn't thinking
isn't thinking of"

[Korny Sietsma]

Hmm - on digging further, I might be better off writing a custom formatter as described at http://wiki.github.com/aslakhellesoy/cucumber/custom-formatters, and just invoking cucumber with --dry-run and my formatter...  Though as I want to use the html formatter to format steps for insertion into the wiki, I've still got some work to do :)

- korny

[Korny Sietsma]

In case anyone followed this: I got everything working pretty nicely - now I have a cucumber formatter that automatically updates a confluence wiki.

I still have to cover some bits like table outputs and the like, but the basics are pretty nice - I use cucumber to parse the features and create a wiki page per feature file with the (complete) feature, and I also check for scenarios with tags like "@story-blah", and update the corresponding story page in the wiki.

I'm not sure this stuff is much use generally, it's pretty tightly coupled to how we have our wiki set up - but I'm happy to share the (messy) code if anyone is interested.

- Korny

[aslak hellesoy]

On Tue, Apr 14, 2009 at 6:06 AM, Korny Sietsma <ko...@sietsma.com> wrote:

In case anyone followed this: I got everything working pretty nicely - now I have a cucumber formatter that automatically updates a confluence wiki.

I still have to cover some bits like table outputs and the like, but the basics are pretty nice - I use cucumber to parse the features and create a wiki page per feature file with the (complete) feature, and I also check for scenarios with tags like "@story-blah", and update the corresponding story page in the wiki.

I'm not sure this stuff is much use generally, it's pretty tightly coupled to how we have our wiki set up - but I'm happy to share the (messy) code if anyone is interested.

Why not share it as a http://gist.github.com/ ?

Aslak

_______________________________________________
rspec-users mailing list
rspec...@rubyforge.org
http://rubyforge.org/mailman/listinfo/rspec-users

[Korny Sietsma]

Good idea.  It's at http://gist.github.com/95033
And as I said before, it's pretty hacky :)

- Korny

[Korny Sietsma]

I should clarify what might not be obvious from the code:
This builds two sets of pages in the wiki:

- Features, which are basically just a dump of the features files into wiki format, added as children of the main "Features" page (confluence lets you have parent/child relationships, which is great for this sort of thing - no need to create an index page)

- Stories, which are our user stories; these are written by hand, and the script looks for story tags on each scenario, and if it finds a matching story in the wiki, it inserts the scenario into that wiki page (and links it back to the matching feature page).

So, for example, I might have features:
----
Feature: top level tabs (in tabs.feature)
  @story-f-001 @story-f-002 @qa
  Scenario: visit the foo tab
    Given I am viewing the main page
    When I navigate to the "Foo" tab
    Then I should be on the "Foo" panel
----
Feature: foo wrangling (in foo.feature)
  @story-f-001 @qa
  Scenario: display all foos
     Given I am viewing the "Foo" panel
     When I select "view all"
     Then I should see all foos in the system

Running "cucumber --dry-run --format WikiFormatter" will build new wiki pages "tabs" and "foo", and if there is a story page called "foo-1" it will add the tagged scenarios to it as well. (in a special "qa" section in this case)

- Korny

[aslak hellesoy]

On Tue, Apr 14, 2009 at 11:34 AM, Korny Sietsma <ko...@sietsma.com> wrote:

I should clarify what might not be obvious from the code:
This builds two sets of pages in the wiki:

- Features, which are basically just a dump of the features files into wiki format, added as children of the main "Features" page (confluence lets you have parent/child relationships, which is great for this sort of thing - no need to create an index page)

- Stories, which are our user stories; these are written by hand, and the script looks for story tags on each scenario, and if it finds a matching story in the wiki, it inserts the scenario into that wiki page (and links it back to the matching feature page).

Sounds like an excellent example of adaptation for a team that has already settled on an external tool to manage the development process and communication in general. Cool stuff!

Aslak