KarmaCodeSpecsMatcher expects to see a nesting for the coverage output, which is generally true. However, when there's no coverage, will be blank. As a result, the ".forEach" call invokes an error and the program outputs a warning about invalid format. It still falls back on a full mapping, which works, but I'm trying to improve run speed.
However, I surprised this hasn't come up. Wouldn't everybody always have zero base coverage? It doesn't include any specs -- there's therefore nothing to run. I'm concerned that I'm missing the point, or have something misconfigured. Heck, maybe I broke it myself with #60.
In case it's related, I'm using the karma-dojo plugin, and all of my source uses AMD-style modules.
KarmaCodeSpecsMatcher expects to see a nesting for the coverage output, which is generally true. However, when there's no coverage, will be blank. As a result, the ".forEach" call invokes an error and the program outputs a warning about invalid format. It still falls back on a full mapping, which works, but I'm trying to improve run speed.
I was able to adjust this so it would return a blank result if the "lines-covered" was zero, and that seems to do what I expect it to. I can make a pull request if desired, https://github.com/divido/grunt-mutation-testing/tree/fix-karma-code-specs
However, I surprised this hasn't come up. Wouldn't everybody always have zero base coverage? It doesn't include any specs -- there's therefore nothing to run. I'm concerned that I'm missing the point, or have something misconfigured. Heck, maybe I broke it myself with #60.
In case it's related, I'm using the karma-dojo plugin, and all of my source uses AMD-style modules.