[3.11] gh-87846: test_io: Ignore OpenWrapper in test___all__#126478
Merged
pablogsal merged 1 commit intopython:3.11from Nov 6, 2024
Merged
[3.11] gh-87846: test_io: Ignore OpenWrapper in test___all__#126478pablogsal merged 1 commit intopython:3.11from
pablogsal merged 1 commit intopython:3.11from
Conversation
The deprecated `OpenWrapper` is added to the module on demand
in `__getattr__`, so it might or might not show up in `dir(io)`
depending on whether e.g. `test_openwrapper` was run.
Add it to `not_exported` so that check__all__ ignores it
when it exists.
The test consistently failed on some refleaks buildbots (but
not when the test is re-run, so it was only marked as a warning).
Locally, it can be reproduced by running `test_openwrapper` and
`test__all__`, twice:
./python -m test test_io test_io -m '*test_[o_][p_][ea][nl][wl]*' -v
pablogsal
approved these changes
Nov 6, 2024
Member
|
I'm fine to backport since this is a change in tests |
Member
Author
|
After this was merged, all the 3.11 refleaks buildbots turned red. But, that's an improvement! There's a pre-existing refleak that was hidden like this:
I'll find and write/backport a fix. (update: see #111942) |
encukou
added a commit
to encukou/cpython
that referenced
this pull request
Nov 7, 2024
…e_encoding There's an issue with the 3.11 backport commit e2421a3 The chane for the main branch was: ```diff + Py_INCREF(errors); ... Py_SETREF(self->encoding, encoding); - Py_SETREF(self->errors, Py_NewRef(errors)); + Py_SETREF(self->errors, errors); ``` but in 3.11 this became: ```diff + Py_INCREF(errors); ... Py_INCREF(errors); Py_SETREF(self->encoding, encoding); Py_SETREF(self->errors, errors); ``` i.e. there's an extra incref, but it looks -- at least to Git -- like the change that removes `Py_NewRef` was already applied. This was not caught because `test_io` refleaks tests were ineffective on 3.11 (see python#126478 (comment)). Remove the extraneous incref.
Member
Author
|
FWTW, 3.10 is unaffected: it doesn't have #111370. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
@pablogsal, are you interested in fixing 3.11 buildbots that consistently warn, rather than fail?
test_iofails iftest_openwrapperruns beforetest___all__. This consistently happens on some refleaks buildbots, for example Windows or Fedora, but it's only marked as a warning: when onlytest___all__is re-run in a fresh process, it succeeds.Locally, this can be reproduced by running
test_openwrapperandtest__all__, twice (since in a single run,test_openwrappercomes aftertest___all__). I don't know of a more elegant way than:The reason is that the deprecated
OpenWrapperis added to the module on demand in__getattr__, after which it shows up indir(io).Add it to
not_exportedso thatcheck__all__ignores it when it exists.