Fetching tasks give an invalid endpoint error

[matthijsgroen]

When fetching a story, and calling story.tasks the code gives an error:

{:status=>404, :headers=>{"server"=>"nginx/1.6.0", "date"=>"Fri, 26 Sep 2014 06:22:49 GMT", "content-type"=>"application/octet-stream", "content-length"=>"135", "connection"=>"close", "etag"=>"\"5314f7e6-87\""}, :body=>{"code"=>"route_not_found", "error"=>"The path you requested has no valid endpoint.", "kind"=>"error", "http_status"=>"404"}}
tracker_api-0.2.5/lib/tracker_api/client.rb:162:in `rescue in request'
tracker_api-0.2.5/lib/tracker_api/client.rb:149:in `request'
tracker_api-0.2.5/lib/tracker_api/client.rb:70:in `paginate'
tracker_api-0.2.5/lib/tracker_api/endpoints/tasks.rb:11:in `get'
tracker_api-0.2.5/lib/tracker_api/resources/story.rb:43:in `tasks'

It does not matter if you include tasks in the list of fields when fetching a story.

[matthijsgroen]

https://www.pivotaltracker.com/help/api/rest/v5#projects_project_id_stories_story_id_tasks_get the example in curl works fine

[forest]

I re-ran the specs for getting tasks and it is working. If you delete the vcr/cassette for getting tasks the spec will make real requests to the test project.

https://github.com/dashofcode/tracker_api/blob/master/test/story_test.rb#L12

Are you sure you have the project_id and story_id correct?

[matthijsgroen]

Yeah these should by fine, I used the following code:

  stories = project.stories(with_state: 'unstarted', fields: 'name,story_type,owners')
  stories.each do |story|
    story.tasks
  end

But I will try by using project.story('id').tasks instead

[forest]

Hmmm. That should work. I'll try to dig in a bit more this week.

[matthijsgroen]

project.story('id').tasks this syntax works, only downside is there seems to be no story.project, so I have to pass around to project to get the tasks. But I'm no longer impeded. Thanks!

(not having story.tasks work is still an issue offcourse)

[matthijsgroen]

I've found what is causing this, I specify the fields I want from a userstory. When I do not ask for the project_id it is not assigned to the story when asking project.stories. This way the request to ask for the tasks fails because it does not know the project id (which could be assigned beforehand since the stories are asked from the project in the first place)

[matthijsgroen]

See https://github.com/dashofcode/tracker_api/blob/master/lib/tracker_api/endpoints/stories.rb#L14

Changing it into: data.map { |story| Resources::Story.new({ client: client, project_id: project_id }.merge(story)) } could solve it

[forest]

Good find. I'll add some specs to make sure I can break it by excluding the project_id and then make the fix.