Redux and Component tests, coverage. Normalize redux actions.

[TristanWright]

• redux tests
  • for: Clusters, Projects & Simulations, and Taskflows:
  • simple actions
  • simple reducers
• component tests?

[TristanWright]

Cannot figure out why these two tests are failing:
https://travis-ci.org/Kitware/HPCCloud/jobs/135997229#L2007

[TristanWright]

Because of the odd async behavior, I'm thinking of just testing simple actions (actions which return { type: ..., someData: ... }) and the corresponding reducer. If so, redux bugs here out may be because we're calling a simple action in an async action where we don't intend to, (the effect of the action will be correct though we'll just be mistakingly calling it.)

[dmitry-zaets]

Since you are using expect - it contains spy functions.
Basic idea is that you can mock any function by creating spy on it.
This will allow you to mock client.listSimulations(id) and return fake promise with data or with error, for different test cases.The easiest way to do mocks is in beforeEach part of Jasmine tests.

Here is an example:

beforeEach(function () {
  const response = { data: 'your data'};
  expect.spyOn(client, 'listSimulations')
    .andReturn(Promise.resolve(response));
})

afterEach(function () {
  expect.restoreSpies(); // This is needed for restoring default function behaviour
})

it('works', function (done) {
  expect(fetchProjectSimulations())
    .toDispatchActions([ {type: 'test1'}, {type: 'test2'} ]);
})

[TristanWright]

@dmitry-zaets THANKS! Hadn't thought of that and I don't have to intercept xhr requests anymore. So happy with this setup now.

[TristanWright]

Cool we have codecov : https://codecov.io/gh/Kitware/HPCCloud/branch/tests-1

[cjh1]

Awesome! Well done.

[cjh1]

@TristanWright Let rename the PR to something a little more descriptive :-) The changes to the redux code, are these the results of problems uncovered by the testing? Its not clear exactly what they fix, so makes it a little hard to review.

[TristanWright]

Renamed PR,
Redux changes normalize our redux actions, for some actions what we'd return was inconsistent. Like actions that have some async component need to dispatch and sometimes didn't. (https://github.com/Kitware/HPCCloud/pull/420/files#diff-3cdd813e5fad39baac221af549782a5bL30)

[TristanWright]

Closes:

• Test something #355
• When adding clusters 'created' even doesn't always update page #421

[cjh1]

Thanks for the explanation, I will try to test this out today.

[cjh1]

When running the test ( which seem to pass ) I am seeing the following:

ERROR LOG: 'Unhandled promise rejection', TypeError{stack: 'http://localhost:9876/base/test/contexts/tests.all.js?392a3be52600af9141667a9059cfaa12e61b93f4:23375:10466
run@http://localhost:9876/base/node_modules/babel-polyfill/dist/polyfill.js?445a020deba63c9118cf9c570f97130c461f14d7:3931:29
http://localhost:9876/base/node_modules/babel-polyfill/dist/polyfill.js?445a020deba63c9118cf9c570f97130c461f14d7:3944:31
flush@http://localhost:9876/base/node_modules/babel-polyfill/dist/polyfill.js?445a020deba63c9118cf9c570f97130c461f14d7:1214:11', line: 23375, sourceURL: 'http://localhost:9876/base/test/contexts/tests.all.js?392a3be52600af9141667a9059cfaa12e61b93f4'}

Is this expected?

[cjh1]

I assume these are just formatting fixes? In future it might be go to do the formatting changes in a separate commit, to make it easier to see the logic changes.

[TristanWright]

The block is wrapped in a (dispatch) => {} now. I'll be sure to keep formatting separate.

[TristanWright]

When running the test ( which seem to pass ) I am seeing the following:

ERROR LOG: 'Unhandled promise rejection', TypeError{stack: 'http://localhost:9876/base/test/contexts/tests.all.js?392a3be52600af9141667a9059cfaa12e61b93f4:23375:10466
run@http://localhost:9876/base/node_modules/babel-polyfill/dist/polyfill.js?445a020deba63c9118cf9c570f97130c461f14d7:3931:29
http://localhost:9876/base/node_modules/babel-polyfill/dist/polyfill.js?445a020deba63c9118cf9c570f97130c461f14d7:3944:31
flush@http://localhost:9876/base/node_modules/babel-polyfill/dist/polyfill.js?445a020deba63c9118cf9c570f97130c461f14d7:1214:11', line: 23375, sourceURL: 'http://localhost:9876/base/test/contexts/tests.all.js?392a3be52600af9141667a9059cfaa12e61b93f4'}

Is this expected?

I've tried to track it down several times, by looking through the stack trace there and commenting in and out test blocks to see if it's coming from one specifically. I think it's related to some combination of how we're polyfilling Promises in PhantonJS. If we run the tests in Chrome we don't get any of the same errors. There's a headless chromium in development, when it's ready we'll switch to it. issue 546953

[cjh1]

I am trying to create a trad cluster and it never seems to move into the running state. I just see the following in the console:

cluster creating

This could be something up with environment, can you confirm you are able to create a cluster?

[TristanWright]

When I re-up my vm I usually I need to restart girder. I appended some cluster preference changes last night too if you haven't pulled.

[cjh1]

This was indeed a local configuration issue.

[TristanWright]

There are going to be some conflicts with pull #409, this should go first though.

[cjh1]

@TristanWright Your last commit does fix the new cluster creation issue. However, I still see an issue with the state transition from "created" to "running" a refresh is required.

[TristanWright]

This confusion came from about how the cluster store items list and mapById are related and used.

Added redux store documentation to issue #406 (Documentation).

[cjh1]

I am also seeing that in the "Simulation" view a newly create EC2 cluster doesn't appear without a page refresh, not sure if this is new behavior.

[TristanWright]

Simulation start or run? I can run and the log and jobs appear

On Wed, Jun 22, 2016 at 3:03 PM, Chris Harris notifications@github.com
wrote:

I am also seeing that in the "Simulation" view a newly create EC2 cluster
doesn't appear without a page refresh, not sure if this is new behavior.

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
#420 (comment), or mute
the thread
https://github.com/notifications/unsubscribe/AMs36V8u4EIcNZd-UiWuSxDFFN6soPCsks5qOaMmgaJpZM4Is4XJ
.

• Tristan Wright

R&D Engineer
Kitware Inc.

[cjh1]

This is when starting a simulation ( creating a new ec2 cluster )

[TristanWright]

I'm not sure I'm seeing it. I do recall some issue like this, for now I
wouldn't say it's new.

On Wed, Jun 22, 2016 at 3:15 PM, Chris Harris notifications@github.com
wrote:

This is when starting a simulation ( creating a new ec2 cluster )

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
#420 (comment), or mute
the thread
https://github.com/notifications/unsubscribe/AMs36RaiOQQ4pjUfNPup0E5L8jjdd1Zsks5qOaXmgaJpZM4Is4XJ
.

• Tristan Wright

R&D Engineer
Kitware Inc.

[cjh1]

I tested against master, it not new.

LGTM

[TristanWright]

excellent!