Closed sed-i closed 1 year ago
WDYT, @addyess, overall? Does this stand a chance to land in main?
@sed-i seems likely we could get something like this in
Question -- why stop at only .model
and model_name
?
Others which could be addressed are:
- model_full_name
- model_config
If we're gonna make a breaking change why not go whole-hog?
@sed-i I think I've pounded on this to a point where I think it's not really benefiting anyone anymore.
This package now uses type checking with mypy for internal consistency type checking during the lint stage I've made a number of the previous Optional[] properties return type instead.
Gimme you're thoughts
Lets consider merging this into 1.x branch so we can version it with 1.0.0.b1
such that it's not picked up by normal installs until we've done some testing with it as a beta
Hey @addyess, could you please re-trigger the workflow?
@sed-i anything else needed here for a 1.0 beta release?
@addyess looks good to me :)
@sed-i I released https://pypi.org/project/pytest-operator/1.0.0b1/ from the 1.x branch
You'll have to explicitly request the package because pip won't install pre-release packages by default.
Let's try it out on some projects and see it work before going 1.0.0 final release
Statically checking integration tests before attempting to run them can save some time.
85 added a py.typed marker, but this results in a lot of the following:
The changes proposed in this PR are problematic because they breaks backwards compatibility. That being said, I assume that very few (if any) use cases exist relying on getting a None for the model, etc.