#89 is not fixed

[fabioel]

Hi @DavertMik!

The issue #89 is still not fixed. I tried it once again. I am currently using version 0.3.3 and the tests are not failing when they should!

I have a very simple test once again:

'use strict';
const config = require('../../config.js');
const expect = require('chai').expect;

Feature('Navigation');

Before((I) => {
  I.amOnPage(config.baseUrl);
});
Scenario.only( 'Test', (I) => {
    expect(true).to.be.eql(false);
});

Here is my terminal output:

fabio@fabio-XPS-15-9530:~/Documents/Project/git/core-feature-tests$ node ./node_modules/.bin/codeceptjs run --steps
CodeceptJS v0.3.3
Test root is assumed to be /home/fabio/Documents/Project/git/core-feature-tests

Navigation --
Test
 • I am on page "https://www.test.com"
fabio@fabio-XPS-15-9530:~/Documents/Shore/git/core-feature-tests$ 

as you can see the tests don't fail they just finish

Here is the output if I change the assertion (expect(true).to.be.eql(true)):

Navigation --
Test
• I am on page "https://www.test.com"
 ✓ OK in 1ms

[fabioel]

UPDATE:

This also happens with the CodeceptJS assertions. Like I.see('asdfasdf') when 'asdfasdf' is not on the page, the test suite does not fail properly

[trikonasana]

Hi @fabioel,

if you run node ./node_modules/.bin/codeceptjs run --steps; echo $?

i guess there is a '0' at the end of your output. So I think my bug (#98) is a duplicate of yours (or v.v.)

Is it?

Cheers
tri

[tjoskar]

I had the same problem until I manually created the output folder. I thought it was a windows-related problem (#126) since I only could reproduce the problem on a windows machine but it turned out that the output folder was missing when I tried on the windows machine and once I created the folder it starts working again.

[smozely]

As well as the output folder not existing I also encountered this being caused by helper classes not having a _failed function by default. I've added a no-op version it to the abstract Helper class so people don't need to provide it by themselves (see #140 ).

@fabioel maybe that is what your hitting?

However I've also noticed that it looks like browser.end() doesn't get called when there is a failure in a before block. Meaning you can end up with dangling browser sessions. I might raise that in a different issue.

[LaminSanneh]

Calling browser.end() in _failed in my helper fixed this for me. Hope it helps anyone else

[DavertMik]

Fixed in #162