thanks for the PR. You need to add some tests to explicitly test for static using.

May I also suggest you should check first if the method name is of interest before resolving its symbol ?
As it is written, this mutator will ask to resolve the full symbol for every method invocation which may add to the total execution time.

Alternatively, having a benchmark (on a significant/large project) that demonstrates the cost being negligible is ok too .

thanks for the PR. You need to add some tests to explicitly test for static using.

May I also suggest you should check first if the method name is of interest before resolving its symbol ? As it is written, this mutator will ask to resolve the full symbol for every method invocation which may add to the total execution time.

Alternatively, having a benchmark (on a significant/large project) that demonstrates the cost being negligible is ok too .

Hi @dupdob ,

Thanks for your feedback. I've added explicit tests for static using calls and updated MathMutator to check the method name before resolving the symbol. Unfortunately, I haven't been able to run benchmarks on a large project yet, but based on theoretical analysis, the filtering should mitigate performance concerns.

Let me know if you have any further suggestions.

thanks @jacinyan. This is exactly what I had in mind; filtering on the name and then checking for actual type looks like the optimized approach.
I added some remarks, in line, but I will reproduce them here as it looks you did not see them:

Thanks for the optimization. There is still one small issue: this mutator contains two versions of the list of mutateable mathematical functions: one as strings (KnownMathMethodNames) and one as an enumeration (MathExpression). Which means that if someone changes one of the list, she will have to do it on both list, otherwise strange issues may appear.
I think you can remove the strings version using Enum.TryParse(methodNameText, out _) instead of KnownMathMethodNames.Contains(methodNameText)).
This way, we just need one list which means easier and safer maintenance.
That being done, please fix failing unit tests and this PR should be good to merge.

thanks @jacinyan. This is exactly what I had in mind; filtering on the name and then checking for actual type looks like the optimized approach. I added some remarks, in line, but I will reproduce them here as it looks you did not see them:

Thanks for the optimization. There is still one small issue: this mutator contains two versions of the list of mutateable mathematical functions: one as strings (KnownMathMethodNames) and one as an enumeration (MathExpression). Which means that if someone changes one of the list, she will have to do it on both list, otherwise strange issues may appear.
I think you can remove the strings version using Enum.TryParse(methodNameText, out _) instead of KnownMathMethodNames.Contains(methodNameText)).
This way, we just need one list which means easier and safer maintenance.
That being done, please fix failing unit tests and this PR should be good to merge.

Thanks @dupdob . I've made some changes accordingly. All tests passed. Please review.