Closed tdyas closed 2 weeks ago
Out of curiosity, could this be codified in a test?
I suppose given we have examples of integration tests already which invoke Pants and check the resulting output.
What would we test? And what would be the benefit vs the cost?
What would we test? And what would be the benefit vs the cost?
I agree with Benjy. Nothing really to test here.
Sorry maybe I'm being naive, but isn't this that we think an option shouldn't be marked advanced AND not have a default? Seems like you could do that either in a test (by instrumenting the constructor) or just at type check time?
All options have a registration-level default I think. We manually check that interpreter_constraints are explicitly provided, as a special case.
Ah I see! Ok ignore me 🙃
Pants currently requires users to set
--python-interpreter-constraints
and raises an exception if they do not. The option, however, is marked "advanced" which really should only apply to optional configuration with defaults most users will not touch, not to mandatory options with no default.Remove the "advanced" marking so that this option is more discoverable by being output for
./pants help python
.