late eval of the skip condition

[kristjanvalur]

at module import time, the REDIS_INFO dict hasn't been initialized.

Pull Request check-list

Please make sure to review and check all of these items:

• Does $ tox pass with this change (including linting)?
• Do the CI tests pass with this change (enable it first in your forked repo and wait for the github action build to finish)?
• Is the new or changed code fully tested?
• Is a documentation update included (if this change modifies existing APIs, or introduces new ones)?
• Is there an example added to the examples folder (if applicable)?
• Was the change added to CHANGES file?

NOTE: these things are not required to open a PR and can be done
afterwards / while the PR is open.

Description of change

when running the unittests manually, e.g. with pytest tests from the command line, test discovery would fail because the
REDIS_INFO dict, although existing, hadn't been initialized at import time. The skip tests need to evaluate the REDIS_INFO as initialized during session setup, hence the need for a late init (done via skip string)

Additionally, changed two instances of

assert round(time1) == round(time2)

to

assert abs(time1 - time2) < 1.0

The former will fail, if for example, time1==1000.4999 and time2==1000.5001 which is clearly not what is intended.

[codecov-commenter]

Codecov Report

Merging #2248 (4e7a3a9) into master (6da8086) will increase coverage by 0.00%.
The diff coverage is 100.00%.

@@           Coverage Diff           @@
##           master    #2248   +/-   ##
=======================================
  Coverage   91.85%   91.86%           
=======================================
  Files         108      108           
  Lines       27776    27787   +11     
=======================================
+ Hits        25515    25527   +12     
+ Misses       2261     2260    -1     

Impacted Files	Coverage Δ
tests/test_asyncio/conftest.py	93.38% <ø> (ø)
tests/conftest.py	86.50% <100.00%> (+0.05%)	⬆️
tests/test_asyncio/test_timeseries.py	100.00% <100.00%> (ø)
tests/test_timeseries.py	100.00% <100.00%> (ø)
redis/commands/core.py	82.11% <0.00%> (ø)
tests/test_cluster.py	97.08% <0.00%> (+0.01%)	⬆️
redis/cluster.py	90.33% <0.00%> (+0.02%)	⬆️
redis/asyncio/connection.py	84.07% <0.00%> (+0.11%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 6da8086...4e7a3a9. Read the comment docs.

[kristjanvalur]

Unit-tests sporatically staling or failing is a big problem in this repo. This PR aims to fix that....

We have a stalled unittest somewhere, unfortunately, the logs are not quite up to identifying it...

platform linux -- Python 3.9.13, pytest-6.2.5, py-1.11.0, pluggy-1.0.0
rootdir: /home/runner/work/redis-py/redis-py, configfile: tox.ini
plugins: asyncio-0.18.3, cov-3.0.0, timeout-2.0.1
asyncio: mode=legacy
collected 3961 items / 779 deselected / 3182 selected
tests/test_bloom.py .................
tests/test_command_parser.py ...s.
Error: The operation was canceled.```

This appears to be all of the tests in `test_command_parser.py`, so the hanging test must have something to do with the next file, test_commands.py.  

I'll try pushing this again and see if it goes through now.

[kristjanvalur]

so, yeah, spurious unit tests fail :)

[dvora-h]

@kristjanvalur Thank you for the fix!

[kristjanvalur]

Uh-oh! I just noticed a problem in the pr. Please see #2253 for a follow-up fix