Accessing Muted Files for Research: Dumping In-Memory Mutants (if Possible)?

[Ali94k]

I'm conducting research that requires access to muted files and analyze them. As far as I know mutants are kept in-memory.

Here's my question:
Is there a way to dump these in-memory mutants into a file format suitable for analysis?

I'm open to exploring alternative approaches as well. If modifying the source code to facilitate mutant export is an option, I'd be happy to contribute with some guidance.

I appreciate any insights or suggestions you might have.

[rouke-broersma]

Mutants can be exported as json with the json reporter. Doesn't that fit your needs?

[Ali94k]

Thanks, dev-mode and Json exporter are what I needed

[dupdob]

What kind of format are you expecting ?
source files?

[Ali94k]

ideally something like the file below.

https://github.com/Flash0ver/F0-Talks-MutationTesting/blob/main/resources/FizzBuzzBuilder.Stryker.cs

[dupdob]

Thanks
In the curren Stryker version, you can use `--dev-mode' CLI option: the mutated source code will be part of the logs:

Mutated MyFile.cs:
.... mutated code

It still requires to extract it from the log files.

Adding an option to dump/export the mutated code can be discussed. As it is trickier than it looks for the general case, the first thing to understand is what is the intended use of it, and what are the general requirements: should it output net projects (that compile) or only source code? if it is only source code, what is the expected (folder) structure ? Which code variant should be dumped : the raw mutated one or the one Stryker uses for testing after removing 'compilation error' mutants ?

Note that the actual source code will be a bit more messy than your example as Stryker generates rather long and random namespace to avoid any naming conflicts.

[Ali94k]

Thanks! For now, dev-mode and the JSON exporter meet my requirements. I'm still in the early stages of my research, but I may share any new requirements that arise as I progress.