zooniverse / Panoptes-Front-End / 8060 / 1

57%

master: 57%

push

travis-ci

Classification queue: save each queued classification once only (#4923)

* Add a test for slow network connections
Stub resource save to return an unresolved Promise, then check that save is only called once for each classification added to the queue.

* Add a failed classification queue
Create an empty array of failed classifications when flushing the queue. Add to that array each time a classification fails, unless the error status is 422.

* Fix async test for updating the queue
Return a promise when flushing the classification queue.
Wait for that promise to resolve, or fail, before testing the queue.

* Remove unused code
Successful saves don't need to update the failed classification queue, so remove that code completely.

* Fix the sent classification test
Wait for add() to resolve before testing the results of a successful save.

* Add ClassificationQueue.store(classification)
Add a store method, which takes a classification and adds it to the saved queue.

* Handle 422 errors
Discard malformed classifications, as they will never be saved by Panoptes.

* Remove FakeLocalStorage
window.localStorage is supported by JSDOM now, so use that in the tests.

* Clear localStorage between tests
Otherwise, each test loads the stored classifications from the previous test.

* Fix FakeResourceType.save
Add a missing return to the save method.

* Catch assertion errors
Catch errors thrown by failing tests.

* Set a default for onClassificationSaved
Make sure the classification save callback is defined before invoking it.

* Reorganise tests
Add some describe and beforeEach blocks.
Break up expectations into separate it blocks, to make them easier to debug when they fail.

* Remove redundant if block
Promise.all will resolve with an empty array if the pending classifications map is empty, so remove the if block that wraps it.

* Fix tests
The second set of tests were failing because apiClient wasn't properly defin... (continued)

1642 of 5194 branches covered (31.61%)

Branch coverage included in aggregate %.

5689 of 11555 relevant lines covered (49.23%)

5.79 hits per line