michaelbull / karma-fail-fast-reporter Goto Github PK

Hello, when I use fail-fast-reporter by adding the reporter to karma.conf.js I have this console terminal result:

$ node_modules/karma/bin/karma start
29 04 2017 22:31:00.979:INFO [karma]: Karma v1.5.0 server started at http://0.0.0.0:9876/
29 04 2017 22:31:00.982:INFO [launcher]: Launching browser Firefox with unlimited concurrency
29 04 2017 22:31:00.993:INFO [launcher]: Starting browser Firefox
29 04 2017 22:31:03.452:INFO [Firefox 53.0.0 (Ubuntu 0.0.0)]: Connected on socket zRI61PCPJnJEOLfZAAAA with id 66900193
Firefox 53.0.0 (Ubuntu 0.0.0): Executed 9 of 9 SUCCESS (0.195 secs / 0.17 secs)

Attention to the last line informing that 9 tests passed with success

But when I try to use fail-fast-reporter through the command line the result is this:

$ node_modules/karma/bin/karma start --reporters fail-fast
29 04 2017 22:31:42.407:INFO [karma]: Karma v1.5.0 server started at http://0.0.0.0:9876/
29 04 2017 22:31:42.410:INFO [launcher]: Launching browser Firefox with unlimited concurrency
29 04 2017 22:31:42.422:INFO [launcher]: Starting browser Firefox
29 04 2017 22:31:44.935:INFO [Firefox 53.0.0 (Ubuntu 0.0.0)]: Connected on socket BFVnKzhMVGjJSAQwAAAA with id 14405303

Notice that the last line has disappeared. This is worse when I try to use along with reporters like the karma-spec-reporter who gives detailed information.
The information on how many tests run and failed always disappears when trying to use fast-fail from the command line.

Karma version: 1.5.0

Thanks for your time.

There is a conflict between fail-fast reporter and singleRun property.
If my configuration looks like this

module.exports = function (config) {
  config.set({
    // some stuff
    reporters: ['fail-fast'],
    singleRun: false,
    // some stuff

the reporter ends running after the first failed spec anyway and test must be run again, which means new compilation etc., which is not friendly. Optimal behavior would be to look at singleRun parameter and exit the test according to the parameter's value.

This is great but enabling it causes testing to end even if auto-watch is enabled. It would be good to have an option to continue watching after the failure.