Hi, we are trying to use

experimental_test_shell_command

to check to see if the developer has run our code generation step. The thing is when we run the code generation, we then run

pants fmt fix

afterwards on the generated files. This means that what was generated and then fixed is different to what we have if we just do the generation. I am wondering if there is a way we can run the

fmt

and

fix

goals either with the

adhoc_tool

or on the output of that? Hopefully this makes sense 🤣

I don't know of any mechanic to do that, except potentially by packing the tools themselves as

pex_binary

and consuming as

execution_dependencies

in an adhoc tool - but it wouldn't be a proper setup. Alternative grug solutions that I use in various similar situations: • hash the input for codegen and embed into the file and compare that instead • disable fmt/fix on the generated code • fix your codegen to output code that meets your standards

Those are some great suggestions. Thanks Tom