Improve bitmap fuzzer

Help
,
1

Currently I’m just using Smith.valueRangeLessThan. I haven’t started using the weighted variant yet. Should I be trying to use them everywhere?

The goal is code coverage. So I guess I should be trying to produce data for each container variant: array, bitset and run? I guess you use weighted values to pick between sparse:array, dense:run and large:bitmap?

I hope that makes sense. Any advice on this stuff or any other fuzzer critiques?

https://codeberg.org/archaistvolts/zroaring/src/branch/main/src/fuzz.zig

Another goal for the fuzzer is to get rid most of this code and HashMapOracle in this file and figure out how to do AFL fuzzing with c sources

@gooncreeper, @vulpesx maybe you guys have some ideas here (sorry for ping)?

2

I have not used fuzzing, like at all. All I can tell you is the API and patterns I know of which you already seem to be using from a quick glance.