-
-
Notifications
You must be signed in to change notification settings - Fork 2.7k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Pathlib compat fails under Python 3.5 #5017
Comments
Hi @khaeru, thanks for the report. We decided to use A workaround is to use |
Hi @nicoddemus — yes, that's what we're doing for now. However, it was somewhat cryptic, so still took us some sleuthing to figure out was going on. Only when the pytest-returned For less experienced pytest users, it might help to:
|
Its a unfortunate fact that pathlib pre python 3.6 is a broken mess |
I will see if we can drop |
Hi @khaeru, We have decided that reverting to standard We have updated the docs for some time now explaining Thanks again. 👍 |
Hi @nicoddemus —thanks for the valiant attempt to overcome that upstream brokenness! Hopefully anyone who encounters the same problem can stumble upon this issue and learn that they might need to explicitly cast to |
need to wrap pytest's tmp_dir in pathlib.Path for python 3.5: pytest-dev/pytest#5017
* Improve test suite handling of paths, temp files This updates most uses of `os.path` to instead use `pathlib.Path`. Relatedly, and following up from #3955 (which replaced pytest's `tmpdir` fixture with `tmp_path`), this removes most ad-hoc tempfile creation in favor of the `tmp_path` fixture. Following conversion, unnecessary `os` and `tempfile` imports were removed. Most pathlib changes involve straightforward changes from `os` functions such as `os.mkdir` or `os.path.abspath` to their equivalent methods in `pathlib.Path`. Changing ad-hoc temporary path to `tmp_path` involved removing the `tmp_dir_path` fixture and replacing its functionality with `tmp_path` in `test_save_load` and `test_guess_filename_with_tempfile`. On `test_static_route_user_home` function: * I think that the intention of this test is to ensure that aiohttp correctly expands the home path if passed in a string. I refactored it to `pathlib.Path` and cut out duplication of `relative_to()` calls. But if it's not doing anything but expanding `~`, then it's testing the functionality of `pathlib.Path`, not aiohttp. On `unix_sockname` fixture: This fixture uses `tempfile.TemporaryDirectory`. Because it's a somewhat complicated fixture used across multiple test modules, I left it as-is for now. On `str(tmp_path)` and even `pathlib.Path(str(tmp_path))`: pytest uses `pathlib2` to provide `tmp_path` for Python 3.5 (only). This is mostly fine but it fails on a couple of corner cases, such as `os.symlink()` which blocks all but `str` and `PurePath` via isinstance type checking. In several cases, this requires conversion to string or conversion to string and then into `pathlib.Path` to maintain code compatibility. See: pytest-dev/pytest/issues/5017 * Correct test_guess_filename to use file object * Update symlink in tests; more guess_filename tests
* Improve test suite handling of paths, temp files This updates most uses of `os.path` to instead use `pathlib.Path`. Relatedly, and following up from aio-libs#3955 (which replaced pytest's `tmpdir` fixture with `tmp_path`), this removes most ad-hoc tempfile creation in favor of the `tmp_path` fixture. Following conversion, unnecessary `os` and `tempfile` imports were removed. Most pathlib changes involve straightforward changes from `os` functions such as `os.mkdir` or `os.path.abspath` to their equivalent methods in `pathlib.Path`. Changing ad-hoc temporary path to `tmp_path` involved removing the `tmp_dir_path` fixture and replacing its functionality with `tmp_path` in `test_save_load` and `test_guess_filename_with_tempfile`. On `test_static_route_user_home` function: * I think that the intention of this test is to ensure that aiohttp correctly expands the home path if passed in a string. I refactored it to `pathlib.Path` and cut out duplication of `relative_to()` calls. But if it's not doing anything but expanding `~`, then it's testing the functionality of `pathlib.Path`, not aiohttp. On `unix_sockname` fixture: This fixture uses `tempfile.TemporaryDirectory`. Because it's a somewhat complicated fixture used across multiple test modules, I left it as-is for now. On `str(tmp_path)` and even `pathlib.Path(str(tmp_path))`: pytest uses `pathlib2` to provide `tmp_path` for Python 3.5 (only). This is mostly fine but it fails on a couple of corner cases, such as `os.symlink()` which blocks all but `str` and `PurePath` via isinstance type checking. In several cases, this requires conversion to string or conversion to string and then into `pathlib.Path` to maintain code compatibility. See: pytest-dev/pytest/issues/5017 * Correct test_guess_filename to use file object * Update symlink in tests; more guess_filename tests (cherry picked from commit 79fe204)
* Improve test suite handling of paths, temp files This updates most uses of `os.path` to instead use `pathlib.Path`. Relatedly, and following up from aio-libs#3955 (which replaced pytest's `tmpdir` fixture with `tmp_path`), this removes most ad-hoc tempfile creation in favor of the `tmp_path` fixture. Following conversion, unnecessary `os` and `tempfile` imports were removed. Most pathlib changes involve straightforward changes from `os` functions such as `os.mkdir` or `os.path.abspath` to their equivalent methods in `pathlib.Path`. Changing ad-hoc temporary path to `tmp_path` involved removing the `tmp_dir_path` fixture and replacing its functionality with `tmp_path` in `test_save_load` and `test_guess_filename_with_tempfile`. On `test_static_route_user_home` function: * I think that the intention of this test is to ensure that aiohttp correctly expands the home path if passed in a string. I refactored it to `pathlib.Path` and cut out duplication of `relative_to()` calls. But if it's not doing anything but expanding `~`, then it's testing the functionality of `pathlib.Path`, not aiohttp. On `unix_sockname` fixture: This fixture uses `tempfile.TemporaryDirectory`. Because it's a somewhat complicated fixture used across multiple test modules, I left it as-is for now. On `str(tmp_path)` and even `pathlib.Path(str(tmp_path))`: pytest uses `pathlib2` to provide `tmp_path` for Python 3.5 (only). This is mostly fine but it fails on a couple of corner cases, such as `os.symlink()` which blocks all but `str` and `PurePath` via isinstance type checking. In several cases, this requires conversion to string or conversion to string and then into `pathlib.Path` to maintain code compatibility. See: pytest-dev/pytest/issues/5017 * Correct test_guess_filename to use file object * Update symlink in tests; more guess_filename tests (cherry picked from commit 79fe204)
* Improve test suite handling of paths, temp files This updates most uses of `os.path` to instead use `pathlib.Path`. Relatedly, and following up from aio-libs#3955 (which replaced pytest's `tmpdir` fixture with `tmp_path`), this removes most ad-hoc tempfile creation in favor of the `tmp_path` fixture. Following conversion, unnecessary `os` and `tempfile` imports were removed. Most pathlib changes involve straightforward changes from `os` functions such as `os.mkdir` or `os.path.abspath` to their equivalent methods in `pathlib.Path`. Changing ad-hoc temporary path to `tmp_path` involved removing the `tmp_dir_path` fixture and replacing its functionality with `tmp_path` in `test_save_load` and `test_guess_filename_with_tempfile`. On `test_static_route_user_home` function: * I think that the intention of this test is to ensure that aiohttp correctly expands the home path if passed in a string. I refactored it to `pathlib.Path` and cut out duplication of `relative_to()` calls. But if it's not doing anything but expanding `~`, then it's testing the functionality of `pathlib.Path`, not aiohttp. On `unix_sockname` fixture: This fixture uses `tempfile.TemporaryDirectory`. Because it's a somewhat complicated fixture used across multiple test modules, I left it as-is for now. On `str(tmp_path)` and even `pathlib.Path(str(tmp_path))`: pytest uses `pathlib2` to provide `tmp_path` for Python 3.5 (only). This is mostly fine but it fails on a couple of corner cases, such as `os.symlink()` which blocks all but `str` and `PurePath` via isinstance type checking. In several cases, this requires conversion to string or conversion to string and then into `pathlib.Path` to maintain code compatibility. See: pytest-dev/pytest/issues/5017 * Correct test_guess_filename to use file object * Update symlink in tests; more guess_filename tests (cherry picked from commit 79fe204)
…hs, temp files (#8083) **This is a backport of PR #3957 as merged into master (79fe204).** * Improve test suite handling of paths, temp files This updates most uses of `os.path` to instead use `pathlib.Path`. Relatedly, and following up from #3955 (which replaced pytest's `tmpdir` fixture with `tmp_path`), this removes most ad-hoc tempfile creation in favor of the `tmp_path` fixture. Following conversion, unnecessary `os` and `tempfile` imports were removed. Most pathlib changes involve straightforward changes from `os` functions such as `os.mkdir` or `os.path.abspath` to their equivalent methods in `pathlib.Path`. Changing ad-hoc temporary path to `tmp_path` involved removing the `tmp_dir_path` fixture and replacing its functionality with `tmp_path` in `test_save_load` and `test_guess_filename_with_tempfile`. On `test_static_route_user_home` function: * I think that the intention of this test is to ensure that aiohttp correctly expands the home path if passed in a string. I refactored it to `pathlib.Path` and cut out duplication of `relative_to()` calls. But if it's not doing anything but expanding `~`, then it's testing the functionality of `pathlib.Path`, not aiohttp. On `unix_sockname` fixture: This fixture uses `tempfile.TemporaryDirectory`. Because it's a somewhat complicated fixture used across multiple test modules, I left it as-is for now. On `str(tmp_path)` and even `pathlib.Path(str(tmp_path))`: pytest uses `pathlib2` to provide `tmp_path` for Python 3.5 (only). This is mostly fine but it fails on a couple of corner cases, such as `os.symlink()` which blocks all but `str` and `PurePath` via isinstance type checking. In several cases, this requires conversion to string or conversion to string and then into `pathlib.Path` to maintain code compatibility. See: pytest-dev/pytest/issues/5017 * Correct test_guess_filename to use file object * Update symlink in tests; more guess_filename tests (cherry picked from commit 79fe204) <!-- Thank you for your contribution! --> ## What do these changes do? This updates most uses of `os.path` to instead use `pathlib.Path`. Relatedly, and following up from #3955 (which replaced pytest's `tmpdir` fixture with `tmp_path`), this removes most ad-hoc tempfile creation in favor of the `tmp_path` fixture. Following conversion, unnecessary `os` and `tempfile` imports were removed. Most pathlib changes involve straightforward changes from `os` functions such as `os.mkdir` or `os.path.abspath` to their equivalent methods in `pathlib.Path`. Changing ad-hoc temporary path to `tmp_path` involved removing the `tmp_dir_path` fixture and replacing its functionality with `tmp_path` in `test_save_load` and `test_guess_filename_with_tempfile`. On `test_static_route_user_home` function: * I think that the intention of this test is to ensure that aiohttp correctly expands the home path if passed in a string. I refactored it to `pathlib.Path` and cut out duplication of `relative_to()` calls. But if it's not doing anything but expanding `~`, then it's testing the functionality of `pathlib.Path`, not aiohttp. On `unix_sockname` fixture: This fixture uses `tempfile.TemporaryDirectory`. Because it's a somewhat complicated fixture used across multiple test modules, I left it as-is for now. On `str(tmp_path)` and even `pathlib.Path(str(tmp_path))`: pytest uses `pathlib2` to provide `tmp_path` for Python 3.5 (only). This is mostly fine but it fails on a couple of corner cases, such as `os.symlink()` which blocks all but `str` and `PurePath` via isinstance type checking. In several cases, this requires conversion to string or conversion to string and then into `pathlib.Path` to maintain code compatibility. See: pytest-dev/pytest/issues/5017 ## Are there changes in behavior for the user? These changes only affect the test suite and have no impact on the end user. ## Related issue number This is intended to address discussion following the simplistic changes from tmpdir to tmp_path of #3955. ## Checklist - [X] I think the code is well written - [X] Unit tests for the changes exist - [X] Documentation reflects the changes - [X] If you provide code modification, please add yourself to `CONTRIBUTORS.txt` * The format is <Name> <Surname>. * Please keep alphabetical order, the file is sorted by names. - [X] Add a new news fragment into the `CHANGES` folder * name it `<issue_id>.<type>` for example (588.bugfix) * if you don't have an `issue_id` change it to the pr id after creating the pr * ensure type is one of the following: * `.feature`: Signifying a new feature. * `.bugfix`: Signifying a bug fix. * `.doc`: Signifying a documentation improvement. * `.removal`: Signifying a deprecation or removal of public API. * `.misc`: A ticket has been closed, but it is not of interest to users. * Make sure to use full sentences with correct case and punctuation, for example: "Fix issue with non-ascii contents in doctest text files." Co-authored-by: Matt VanEseltine <matvan@umich.edu>
…s, temp files (#8084) **This is a backport of PR #3957 as merged into master (79fe204).** * Improve test suite handling of paths, temp files This updates most uses of `os.path` to instead use `pathlib.Path`. Relatedly, and following up from #3955 (which replaced pytest's `tmpdir` fixture with `tmp_path`), this removes most ad-hoc tempfile creation in favor of the `tmp_path` fixture. Following conversion, unnecessary `os` and `tempfile` imports were removed. Most pathlib changes involve straightforward changes from `os` functions such as `os.mkdir` or `os.path.abspath` to their equivalent methods in `pathlib.Path`. Changing ad-hoc temporary path to `tmp_path` involved removing the `tmp_dir_path` fixture and replacing its functionality with `tmp_path` in `test_save_load` and `test_guess_filename_with_tempfile`. On `test_static_route_user_home` function: * I think that the intention of this test is to ensure that aiohttp correctly expands the home path if passed in a string. I refactored it to `pathlib.Path` and cut out duplication of `relative_to()` calls. But if it's not doing anything but expanding `~`, then it's testing the functionality of `pathlib.Path`, not aiohttp. On `unix_sockname` fixture: This fixture uses `tempfile.TemporaryDirectory`. Because it's a somewhat complicated fixture used across multiple test modules, I left it as-is for now. On `str(tmp_path)` and even `pathlib.Path(str(tmp_path))`: pytest uses `pathlib2` to provide `tmp_path` for Python 3.5 (only). This is mostly fine but it fails on a couple of corner cases, such as `os.symlink()` which blocks all but `str` and `PurePath` via isinstance type checking. In several cases, this requires conversion to string or conversion to string and then into `pathlib.Path` to maintain code compatibility. See: pytest-dev/pytest/issues/5017 * Correct test_guess_filename to use file object * Update symlink in tests; more guess_filename tests (cherry picked from commit 79fe204) <!-- Thank you for your contribution! --> ## What do these changes do? This updates most uses of `os.path` to instead use `pathlib.Path`. Relatedly, and following up from #3955 (which replaced pytest's `tmpdir` fixture with `tmp_path`), this removes most ad-hoc tempfile creation in favor of the `tmp_path` fixture. Following conversion, unnecessary `os` and `tempfile` imports were removed. Most pathlib changes involve straightforward changes from `os` functions such as `os.mkdir` or `os.path.abspath` to their equivalent methods in `pathlib.Path`. Changing ad-hoc temporary path to `tmp_path` involved removing the `tmp_dir_path` fixture and replacing its functionality with `tmp_path` in `test_save_load` and `test_guess_filename_with_tempfile`. On `test_static_route_user_home` function: * I think that the intention of this test is to ensure that aiohttp correctly expands the home path if passed in a string. I refactored it to `pathlib.Path` and cut out duplication of `relative_to()` calls. But if it's not doing anything but expanding `~`, then it's testing the functionality of `pathlib.Path`, not aiohttp. On `unix_sockname` fixture: This fixture uses `tempfile.TemporaryDirectory`. Because it's a somewhat complicated fixture used across multiple test modules, I left it as-is for now. On `str(tmp_path)` and even `pathlib.Path(str(tmp_path))`: pytest uses `pathlib2` to provide `tmp_path` for Python 3.5 (only). This is mostly fine but it fails on a couple of corner cases, such as `os.symlink()` which blocks all but `str` and `PurePath` via isinstance type checking. In several cases, this requires conversion to string or conversion to string and then into `pathlib.Path` to maintain code compatibility. See: pytest-dev/pytest/issues/5017 ## Are there changes in behavior for the user? These changes only affect the test suite and have no impact on the end user. ## Related issue number This is intended to address discussion following the simplistic changes from tmpdir to tmp_path of #3955. ## Checklist - [X] I think the code is well written - [X] Unit tests for the changes exist - [X] Documentation reflects the changes - [X] If you provide code modification, please add yourself to `CONTRIBUTORS.txt` * The format is <Name> <Surname>. * Please keep alphabetical order, the file is sorted by names. - [X] Add a new news fragment into the `CHANGES` folder * name it `<issue_id>.<type>` for example (588.bugfix) * if you don't have an `issue_id` change it to the pr id after creating the pr * ensure type is one of the following: * `.feature`: Signifying a new feature. * `.bugfix`: Signifying a bug fix. * `.doc`: Signifying a documentation improvement. * `.removal`: Signifying a deprecation or removal of public API. * `.misc`: A ticket has been closed, but it is not of interest to users. * Make sure to use full sentences with correct case and punctuation, for example: "Fix issue with non-ascii contents in doctest text files." --------- Co-authored-by: Matt VanEseltine <matvan@umich.edu> Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
pytest's compatibility code chooses to use
pathlib2
except for Python 3.6. This means thatpathlib2.Path
objects are returned for fixtures/fixture methods liketmp_path_factory.mktemp()
. When the user tests code using Python 3.5's built-inpathlib
, the built-in package chokes on the objects frompathlib2
.Put the following in
test_foo.py
and execute using pytest under Python 3.5:…this occurs:
The same command succeeds under Python 3.6.
Discovered by @gidden
The text was updated successfully, but these errors were encountered: