Test failures when CPython is built without docstrings

[ZackerySpytz]

BPO	37501
Nosy	@brettcannon, @terryjreedy, @ericvsmith, @ZackerySpytz
PRs	bpo-37501: Fix test failures when CPython is built without docstrings #14592
Files	test_failures_without_docstrings.txt

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = None
closed_at = None
created_at = <Date 2019-07-05.03:30:04.113>
labels = ['3.7', '3.8', 'type-bug', 'tests', '3.9']
title = 'Test failures when CPython is built without docstrings'
updated_at = <Date 2019-07-09.07:46:22.932>
user = 'https://github.com/ZackerySpytz'

bugs.python.org fields:

activity = <Date 2019-07-09.07:46:22.932>
actor = 'terry.reedy'
assignee = 'none'
closed = False
closed_date = None
closer = None
components = ['Tests']
creation = <Date 2019-07-05.03:30:04.113>
creator = 'ZackerySpytz'
dependencies = []
files = ['48456']
hgrepos = []
issue_num = 37501
keywords = ['patch']
message_count = 4.0
messages = ['347298', '347361', '347363', '347527']
nosy_count = 4.0
nosy_names = ['brett.cannon', 'terry.reedy', 'eric.smith', 'ZackerySpytz']
pr_nums = ['14592']
priority = 'normal'
resolution = None
stage = 'patch review'
status = 'open'
superseder = None
type = 'behavior'
url = 'https://bugs.python.org/issue37501'
versions = ['Python 3.7', 'Python 3.8', 'Python 3.9']

Linked PRs

• gh-81682: Fix test failures when CPython is built without docstrings #113410
• [3.12] gh-81682: Fix test failures when CPython is built without docstrings (GH-113410) #113429
• [3.11] gh-81682: Fix test failures when CPython is built without docstrings (GH-113410) #113430

[ZackerySpytz]

test_coroutines, test_dataclasses, test_idle, test_importlib, test_module, and test_pydoc fail when CPython is built without docstrings.

It seems that most of the failures occur simply because the test.support.requires_docstrings decorator is missing for some tests.

See also bpo-17041.

[brettcannon]

Bit surprised by the test_importlib failure. What has crept into there that requires docstrings?

[ericvsmith]

Also surprised by the marked dataclasses tests. I don't see anything there that requires docstrings. See my review on the PR.

[terryjreedy]

Compiling without docstrings only affects compiled functions. (Is this option unix only, for servers?) This caused failures in 6 /test test functions. Running with -00 only affects python functions. It causes failure in 19 test function on Windows. For IDLE, it causes many more failures than the configuration change.

__doc__ == '' might have been less of a nuisance than == None, but too late to change that glocally now, although it can be a local solution.