Coveralls.js vs. Istanbul

Question

I'm confused, I see people using both. They're both code coverage reporting tools. So is it just that people are using the Istanbul functionality and want to use coveralls UI instead of the istanbul html output files as just a nicer coverage runner, is that it? is that the reason to use both??

Answer 1

Istanbul generates coverage information and coveralls provides historical coverage reporting. Istanbul provides a snapshot of where you are; coveralls tells you where you have been.

Typically, you use coveralls as part of a CI/CD pipeline: local build, push to Git, Travis build, push results to coveralls, ...

When you build your project, you will look at your lcov html report to review coverage. How do you know if your coverage has increased or reduced? Look at coveralls for the history.

Shields.io provides badges for Coveralls coverage that you can wear on your GitHub README.md which also shows on npmjs.com if you publish there. It is a nice quality indicator for people using your product and equally nice as a note-to-self that your coverage is slipping (badges are colored and show a % coverage).