Lib/test/lock_tests.py should not use time.time(), but time.monotonic()

[vstinner]

BPO	35513
Nosy	@vstinner, @pablogsal, @miss-islington
PRs	bpo-35513, unittest: TextTestRunner uses time.perf_counter() #11180 bpo-35513: Replace time.time() with time.monotonic() in tests #11182 [3.7] bpo-35513: Replace time.time() with time.monotonic() in tests (GH-11182) #11185 [3.7] bpo-35513, unittest: TextTestRunner uses time.perf_counter() (GH-11180) #11188

^{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 = <Date 2018-12-17.10:54:06.897>
created_at = <Date 2018-12-16.22:05:49.526>
labels = ['3.8', 'tests']
title = 'Lib/test/lock_tests.py should not use time.time(), but time.monotonic()'
updated_at = <Date 2018-12-17.10:54:06.896>
user = 'https://github.com/vstinner'

bugs.python.org fields:

activity = <Date 2018-12-17.10:54:06.896>
actor = 'vstinner'
assignee = 'none'
closed = True
closed_date = <Date 2018-12-17.10:54:06.897>
closer = 'vstinner'
components = ['Tests']
creation = <Date 2018-12-16.22:05:49.526>
creator = 'vstinner'
dependencies = []
files = []
hgrepos = []
issue_num = 35513
keywords = ['patch']
message_count = 6.0
messages = ['331939', '331951', '331954', '331968', '331975', '331976']
nosy_count = 3.0
nosy_names = ['vstinner', 'pablogsal', 'miss-islington']
pr_nums = ['11180', '11182', '11185', '11188']
priority = 'normal'
resolution = 'fixed'
stage = 'resolved'
status = 'closed'
superseder = None
type = None
url = 'https://bugs.python.org/issue35513'
versions = ['Python 3.8']

[vstinner]

https://buildbot.python.org/all/#/builders/145/builds/956

Unhandled exception in thread started by <function Bunch.__init__.<locals>.task at 0x111775cb0>
Traceback (most recent call last):
  File "/Users/buildbot/buildarea/3.x.billenstein-sierra/build/Lib/test/lock_tests.py", line 41, in task
    f()
  File "/Users/buildbot/buildarea/3.x.billenstein-sierra/build/Lib/test/lock_tests.py", line 591, in f
    self.assertTimeout(dt, 0.1)
  File "/Users/buildbot/buildarea/3.x.billenstein-sierra/build/Lib/test/lock_tests.py", line 80, in assertTimeout
    self.assertGreaterEqual(actual, expected * 0.6)
  File "/Users/buildbot/buildarea/3.x.billenstein-sierra/build/Lib/unittest/case.py", line 1283, in assertGreaterEqual
    self.fail(self._formatMessage(msg, standardMsg))
  File "/Users/buildbot/buildarea/3.x.billenstein-sierra/build/Lib/unittest/case.py", line 719, in fail
    raise self.failureException(msg)
AssertionError: -0.24049997329711914 not greater than or equal to 0.06
test_waitfor_timeout (test.test_threading.ConditionTests) ... FAIL

test_waitfor_timeout():

            ...
            dt = time.time()
            result = cond.wait_for(lambda : state==4, timeout=0.1)
            dt = time.time() - dt
            self.assertFalse(result)
            self.assertTimeout(dt, 0.1)
            ...

with:

    def assertTimeout(self, actual, expected):
        ...
        self.assertGreaterEqual(actual, expected * 0.6)
        ...

It seems like time.time() gone backward on the buildbot. The test must use time.monotonic() to measure time difference.

Attached PR fix the issue.

[vstinner]

New changeset 2cf4c20 by Victor Stinner in branch 'master':
bpo-35513: Replace time.time() with time.monotonic() in tests (GH-11182)
2cf4c20

[miss-islington]

New changeset be69ff2 by Miss Islington (bot) in branch '3.7':
bpo-35513: Replace time.time() with time.monotonic() in tests (GH-11182)
be69ff2

[vstinner]

New changeset 8db5b54 by Victor Stinner in branch 'master':
bpo-35513, unittest: TextTestRunner uses time.perf_counter() (GH-11180)
8db5b54

[miss-islington]

New changeset 9ade4cb by Miss Islington (bot) in branch '3.7':
bpo-35513, unittest: TextTestRunner uses time.perf_counter() (GH-11180)
9ade4cb

[vstinner]

I searched from "time.time" and "from time import(...)time" in the Python standard library (in the master branch) and it seems like all usage of time.time() are now appropriate. I close the issue.

Sometimes, I'm not sure that if time.monotonic() or time.perf_counter() should be used, but at least both functions are monotonic and so are not affected by this issue (clock going backwards).