fixtures,runner: fix tracebacks getting longer and longer #12264
Conversation
| @@ -1126,7 +1127,7 @@ def pytest_fixture_setup( | |||
| # Don't show the fixture as the skip location, as then the user | |||
| # wouldn't know which test skipped. | |||
| e._use_item_location = True | |||
| fixturedef.cached_result = (None, my_cache_key, e) | |||
| fixturedef.cached_result = (None, my_cache_key, (e, e.__traceback__)) | |||
There was a problem hiding this comment.
Would it be possible to just pass the dunder traceback into the method to avoid tracking the concrete one for reraise?
There was a problem hiding this comment.
I don't think so, the exception is saved on the FixtureDef and can be needed again at an arbitrary point in the future without a caller-callee relation to the time it is saved.
|
This wasn't backported to 8.2.x, right? Because I see a similar issue on the recently released 8.2.1 |
|
@bluetech was this not backported on purpose, or we just forgot to do it? |
|
I didn't backport because of the changes that may affect plugins. |
| - ``FixtureDef.cached_result[2]`` is now a tuple ``(exc, tb)`` instead of ``exc``. | ||
| - ``SetupState.stack`` failures are now a tuple ``(exc, tb)`` instead of ``exc``. |
There was a problem hiding this comment.
@bluetech FYI this list seems to need indentation or something like that. See the preview of how it's rendered: https://docs.pytest.org/en/latest/changelog.html#bug-fixes.
P.S. I integrated my changelog draft preview Sphinx extension during the sprint. So now, you're going to be able to preview the rendering in new pull requests, even before merging.
There was a problem hiding this comment.
I submitted a fix via #12563. It's better with that patch: https://pytest--12563.org.readthedocs.build/en/12563/changelog.html#bug-fixes.
It was being rendered inline and not it's not.
It was being rendered inline and now it's not.
It was being rendered inline and now it's not.
It was being rendered inline and now it's not.
It was being rendered inline and now it's not.
It was being rendered inline and now it's not.
…64--reraise-with-original-tb 📝🚑 Polish the PR #12264 changelog entry
It was being rendered inline and now it's not.
Fix #12204 - please see the explanation in #12204 (comment)