In some cases, when your code uses timers (
clearInterval), your tests may become unpredictable, slow and
To solve these problems, or if you need to rely on specific timestamps in your code, most testing frameworks offer the option to replace the real timers in your tests with fake ones. This should be used sporadically and not on a regular basis since using it contains some overhead.
When using fake timers in your tests, all of the code inside your test uses fake timers.
The common pattern to setup fake timers is usually within the
When using fake timers, you need to remember to restore the timers after your test runs.
The main reason to do that is to prevent 3rd party libraries running after your test finishes (e.g cleanup functions), from being coupled to your fake timers and use real timers instead.
For that you usually call
It's important to also call
runOnlyPendingTimers before switching to real
timers. This will ensure you flush all the pending timers before you switch to
real timers. If you don't progress the timers and just switch to real timers,
the scheduled tasks won't get executed and you'll get an unexpected behavior.
This is mostly important for 3rd parties that schedule tasks without you being
aware of it.
Here's an example of doing that using jest: