Currently, the fuzzer creates statements by evaluating random percentages. This works, but if we want the fuzzer to follow a specific path to generate an expression, function_expression, or anything else, we have to modify the code in many places to make sure the control flow ends up where we want it. This works, but requires lots of recompilation and code editing that can get tiring.
If we move the hardcoded percentages out and into a file, we can then modify the percentages to nudge the fuzzer to generate statements that are more to our liking. Since these percentages can be read from a file, we also don't have to recompile, and we can just revert any changes to our percentages file.
Currently, the fuzzer creates statements by evaluating random percentages. This works, but if we want the fuzzer to follow a specific path to generate an expression, function_expression, or anything else, we have to modify the code in many places to make sure the control flow ends up where we want it. This works, but requires lots of recompilation and code editing that can get tiring.
If we move the hardcoded percentages out and into a file, we can then modify the percentages to nudge the fuzzer to generate statements that are more to our liking. Since these percentages can be read from a file, we also don't have to recompile, and we can just revert any changes to our percentages file.