fix(check): don't raise error on pypi reference #9475
Conversation
Secrus
force-pushed
the
pypi-ref
branch
3 times, most recently
from
e0a4ce0
to
a320010
Compare
| @@ -111,19 +127,55 @@ def test_check_valid_legacy( | |||
| assert tester.io.fetch_error() == expected | |||
|
|
|||
|
|
|||
| def test_check_invalid( | |||
| def test_check_invalid_dep_name_same_as_project_name( | |||
There was a problem hiding this comment.
What is the reason for splitting (almost duplicating) the test and how is it related to the fix?
There was a problem hiding this comment.
The old one was checking the dependency name vs project name. This is not possible with a fixture being a Poetry instance, so I moved this to another test to keep this case tested.
There was a problem hiding this comment.
So if we do not introduce the fixture, we do not have to split the test? Then, why introduce a fixture in the first place? (Further, it still seems unrelated to the test case for the fix.)
There was a problem hiding this comment.
We need to introduce the fixture because otherwise (with patching pyproject file) we get the wrong list of sources for the project.
There was a problem hiding this comment.
I still do not get why this is an issue. We now have two almost similar tests (test_check_invalid_dep_name_same_as_project_name and test_check_invalid). Both provide the same errors except for that the first one provides two more errors:
Error: Project name (invalid) is same as one of its dependenciesError: Invalid source "exists" referenced in dependencies.
And both tests are unrelated to the issue that is resolved by this PR. This is tested in test_check_does_not_error_on_pypi_reference, which uses another fixture.
Pull Request Check List
Resolves: #9436