What if instead of random, we used a fixed buffer which is mutated at the start of each mutation and used like FuzzDataProvider

status-im / nim-drchaos

A powerful and easy-to-use fuzzing framework in Nim for C/C++/Obj-C targets

Other

68 stars 3 forks source link

What if instead of random, we used a fixed buffer which is mutated at the start of each mutation and used like FuzzDataProvider #17

Closed planetis-m closed 2 years ago

planetis-m commented 2 years ago

First we need to ensure that mutations are unique, each seed creates the same mutation. Currently we don't even test that. Adapt the benchmark to do it.

planetis-m commented 2 years ago

Could try to remove it from parts of the process like the sampler or mutators. If it works maybe we can get rid of repeatmutate.

planetis-m commented 2 years ago

I think the recent benchmark showed that everything not picked by a random uniform distribution performs badly.