``meteor test`` NOT running some tests and NOT throwing errors


Hi guys,

has anyone else experienced that using meteor test * tests simetimes simply don’t show up in the testrunner, when hidden erros occur?

I am running my tests from within imports/modules/* and right now a lot of tests do not show up.
I am sure that I have an error somewhere in my code BUT I would still expect the testrunner to run all tests.

Anyone else having that issue?

Testing guide article

Update. I have found the reason for the tests NOT showing up.

The Scenario:

  • Out of 65 client tests ONLY 21 where showing up.
  • The browser-console showed a couldn't import module xxxx-error within an anonynous function call, so I could NOT see which line of code this was related to
  • There was NO OTHER error message shown in the browser, indicating that I was having a problem. It seems like the testrunner just stopped.

In my case I just had 21 passing tests and was NOT really awaire (for quite some time) that the other 44 were NOT running. This is really bad - the testrunner should ALWAYS show that something went wrong with the other tests.

The reason for the testrunner failing me trying to import a server file from the client.

The remaining question is: will there be a fix for the testrunner ALWAYS showing ALL tests?


I want to second this, this is a very challenging bug as it has a cascading effect in alphabetical order on your test cases. If you have a test that starts M (monkies.test.js) that throws a “cannot find module” error. tests starting with A-L will execute but any tests starting with M-Z will not run, without any sort of error.


Right now you really have to pay close attention to this as you might just loose your tests when commiting to version-control and later wonder what has happened.

A quick fix for the testrunner might be to simply count the number of “describe > it” in all test.js-files and the diff the number to the actual number of recieved test results. if it diffs then it should give an warning like “warning: we expected to recieve 22 test-results but just got 10”.


I’m making the assumption here that all of you are using pm mocha as your testrunner. I had very similar issues though mine were limited to server-side testing.

I developed a simple test case using the todos app and reported my issue on PM Mocha’s gethub site. I was pleasantly surprised at the quickness with which @jsep verified the issue and created a fix.

I would suggest you try the version of PM Mocha with that fix,

    meteor add practicalmeteor:mocha@2.4.6-rc.2

and if that does not resolve your issue, figure out a simple change that can be made to the latest version of todos to recreate the problem and report it to the PM Mocha site as I did. My guess is you’ll be pleased with the result.


If you are using practicalmeteor:mocha update to latest version. I just release a new version of the package the version is 2.4.5_2.

Thanks @rlivingston for the support :smiley:


I’m using 2.4.5_2 on my boilerplate code it works well until you add the dispatch:mocha-phantomjs package and the server side tests stop working. I opened up a bug on dispatch:mocha-phantomjs and it is closed in favor of https://github.com/practicalmeteor/meteor-mocha/issues/23

My workaround right now for CI is to add dispatch:mocha-phantomjs as part of the build https://github.com/trajano/meteor-boilerplate/blob/meteor-skeleton/.travis.yml#L20


@trajano now there are no conflicts with dispatch:mocha-phantomjs see HERE for more details


has it been tested on windows 10?


@trajano Yes is an issue that its going to be solve with the next release.

Please try with this release candidate and tell me if its working:

Run this:

meteor add practicalmeteor:mocha@=2.4.5-rc5.5

For more details go here-


practicalmeteor:mocha@2.4.5-rc5.5 seems to load now.

Loads well so far (at least on Windows 7, will test Windows 10 later)


Yes muy bad @trajano the right version is practicalmeteor:mocha@2.4.5-rc5.5


Works on Windows 10 Chrome, but not Firefox.