Tests and partial fix for suppressing cancellations. #230
Conversation
Dreamsorcerer
changed the title
| @@ -198,6 +198,8 @@ def _do_exit(self, exc_type: Type[BaseException]) -> None: | |||
| return None | |||
|
|
|||
| def _on_timeout(self, task: "asyncio.Task[None]") -> None: | |||
| if task._fut_waiter and task._fut_waiter.cancelled(): # type: ignore[attr-defined] # noqa: E501 | |||
There was a problem hiding this comment.
This hack looks dangerous and potentially not future compatible; while I don't see a better solution.
For Python 3.9+ task.cancel(SENTINEL) variant can be used though to filter out all cancellation events not sent by async_timeout library itself.
There was a problem hiding this comment.
I think putting a 3.9+ fix in would be acceptable, as I don't see a good way to do this otherwise.
Codecov Report
@@ Coverage Diff @@
## master #230 +/- ##
===========================================
- Coverage 100.00% 98.13% -1.87%
===========================================
Files 1 1
Lines 105 107 +2
Branches 15 16 +1
===========================================
Hits 105 105
- Misses 0 1 +1
- Partials 0 1 +1
Flags with carried forward coverage won't be shown. Click here to find out more.
Continue to review full report at Codecov.
|
Dreamsorcerer
force-pushed
the
fix-cancel-overrule
branch
from
7e136a2 to
e7d4620
Compare
|
Ok, let me merge and make a new release |
Partially fixes #229.
I don't really see any way to get enough information to fully fix this, not really sure how best to proceed.
For an overview of why this is a serious problem, see: #229 (comment)