Not sure what the best way to explain what I want is, but here goes...
I have a header file ('generated.h') that is generated from another header file ('interface.h') using some python script.
If I add a custom build step to generated.h, that's a circular dependency. Also, 'generated.h' doesn't even exist in a new workspace, so it gets a bit more confusing there.
Should I instead change interface.h to a custom build tool?
'generated.h' is only for use in testing (generated.h's are mock headers) and there may be several. Therefore I don't really want to add a custom build step to interface.h, since that's used in "real" code. It's not really interface.h's responsibility to generate 'generated.h' (or is it?).
I could add the script as an item next to 'generated.h', but if there are many such generated.h-like-files, I would need to modify the script to accept multiple sets of args, or otherwise find a way to add the generation script several times.
What would you recommend?