Shrinking

Shrinking is a crucial process in property-based testing that finds the smallest value that causes an error. This process simplifies test cases, enabling efficient debugging and error analysis. However, the minimum value found by shrinking is only an approximation, and it's being the absolute minimum is not guaranteed. Particularly with continuous values, it's relatively easier to find the nearest minimum approximation, but this is not necessarily the case with discrete values.

The shrinking process involves reducing values towards a baseline, which varies depending on the arbitrary used. Generally, values are reduced towards zero or empty. The shrunk values generated are not random but are produced according to specific rules, ensuring the same shrunk value is generated for the same error value each time.

Mostly, there is no need to deeply consider the details of shrinking. Each arbitrary is designed to perform shrinking as efficiently as possible. Although there is a limit to the number of shrinking attempts by default, if you wish to perform more shrinking, you can specify the number in settings or set it to unlimited under the shrinking policy. Nonetheless, for most cases, the default setting is adequate. Rather than increasing the number of shrinking attempts, it is more important to specify an appropriate range for generating data.