You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
To summarise previous discussions as far as I can tell:
We want to test type promotion anyway in our primary tests.
There's still value of type promotion for developers of compliant libraries, given it pinpoints the issues.
The fact is though, I don't care about maintaining it. We already have a hacky conftest.py just to skip some flaky cases we identified, and I don't see it worth my time supporting complex cases. Testing type promotion isn't simple for a lot of functions, given other rules for inputs (namely broadcasting).
I think leaning on the primary tests solely is just a more sane approach anyway, as whilst bugs can sometimes be lost in the noise, they're usually getting reported and easier to identify once you fix other bugs you'll have to fix anyway.
The text was updated successfully, but these errors were encountered:
To summarise previous discussions as far as I can tell:
The fact is though, I don't care about maintaining it. We already have a hacky
conftest.py
just to skip some flaky cases we identified, and I don't see it worth my time supporting complex cases. Testing type promotion isn't simple for a lot of functions, given other rules for inputs (namely broadcasting).I think leaning on the primary tests solely is just a more sane approach anyway, as whilst bugs can sometimes be lost in the noise, they're usually getting reported and easier to identify once you fix other bugs you'll have to fix anyway.
The text was updated successfully, but these errors were encountered: