cleanup patches after test in pytest plugin (#1148) #1164
Conversation
|
Apparently there is an issue in the For now pinned the offending dependency. Probably best to create an issue (label: "good first issue") so we don't forget in the future |
Codecov Report
@@ Coverage Diff @@
## master #1164 +/- ##
=========================================
Coverage 100.00% 100.00%
=========================================
Files 79 79
Lines 2415 2415
Branches 208 208
=========================================
Hits 2415 2415 Continue to review full report at Codecov.
|
sobolevn
left a comment
sobolevn
left a comment
There was a problem hiding this comment.
Great! Thank you!
A couple of questions from me 🙂
| # (1) we cannot always assume hashability and | ||
| # (2) we need to track the object identity, not its value | ||
| _ERRORS_HANDLED: Final[Dict[int, Any]] = {} # noqa: WPS407 | ||
| _ErrorsHandled = Dict[int, Any] |
There was a problem hiding this comment.
Oh, this is now a type alias. Let's leave a TODO item to use explicit TypeAlias once [email protected] is out.
There was a problem hiding this comment.
I was doubting as well whether to use it. Based on the PEP it seemed optional -- necessary in case of forward references, for example.
On the other hand, if this is looking to be the default way of declaring type aliases (explicit!) then we can already import from typing_extensions right?
There was a problem hiding this comment.
Hmmm something strange in mypy indeed, it gives:
Variable "typing_extensions.TypeAlias" is not valid as a type
mypy doesn't support them yet at all yet?
There was a problem hiding this comment.
Yes, 0.910 does not support it yet, 0.920 will (hopefully).
There was a problem hiding this comment.
Are there other features coming in 0.920 that returns finds useful?
There was a problem hiding this comment.
I am not sure about ParamSpec (it might not be released just yet), but it would remove decorator plugin! 🎉
returns/contrib/pytest/plugin.py
Outdated
|
|
||
|
|
||
| @contextmanager | ||
| def _spy_error_handling() -> Generator[_ErrorsHandled, None, None]: |
There was a problem hiding this comment.
Looks like it is just Iterator[_ErrorsHandled]
There was a problem hiding this comment.
Indeed -- I assumed contextmanager would complain -- let's test that hypothesis 🧑🔬 ...
returns/contrib/pytest/plugin.py
Outdated
| return wraps(original)(wrapper) # type: ignore | ||
|
|
||
|
|
||
| _ERROR_HANDLING_PATCHERS: Final = { # noqa: WPS407 |
There was a problem hiding this comment.
Do we need this to be mutable? Or will types.MappingProxyType work here?
There was a problem hiding this comment.
Nope. Chose the dict for simplicity. Can use a readonly proxy just fine
There was a problem hiding this comment.
Let's do this! We are kinda strict on mutability in this specific project 👍
There was a problem hiding this comment.
Naturally 😅 . It's a shameMappingProxyType isn't a full frozendict-like object, but we'll take it.
| # See: https://github.com/typlog/sphinx-typlog-theme/issues/22 | ||
| jinja2==3.0.3 | ||
|
|
||
| # TODO: Remove this lock when this dependency issue is resolved. |
| # (1) we cannot always assume hashability and | ||
| # (2) we need to track the object identity, not its value | ||
| _ERRORS_HANDLED: Final[Dict[int, Any]] = {} # noqa: WPS407 | ||
| _ErrorsHandled = Dict[int, Any] |
There was a problem hiding this comment.
I am not sure about ParamSpec (it might not be released just yet), but it would remove decorator plugin! 🎉
2 participants
I have made things!
As discussed in #1148 and #1147
excuse the large diff; I took the liberty of rearranging the functions in the module to:
Checklist
CHANGELOG.mdRelated issues
Closes #1148