Details
-
Type:
Story
-
Status: Done
-
Resolution: Done
-
Fix Version/s: None
-
Labels:
-
Story Points:0.5
-
Epic Link:
-
Sprint:AP S19-5
-
Team:Alert Production
Description
In running some tests involving pipe_base I got this warning:
/Volumes/ExternalSSD/Users/timj/work/lsstsw37/stack/DarwinX86/pipe_base/16.0-10-g0b41441+6/python/lsst/pipe/base/timer.py:91: DeprecationWarning: time.clock has been deprecated in Python 3.3 and will be removed from Python 3.8: use time.perf_counter or time.process_time instead
|
Not urgent yet but the Python developers have now given an explicit timeline for when it will become a problem.
Attachments
Activity
I didn't fix it because I had no idea what the right answer was. Russell Owen wrote the initial code so maybe he has some thoughts. time.process_time seems to be closest.
Good, that makes three of us: Russell Owen also said he thought process_time was most appropriate.
Tim Jenness: can you do this small review, please? It got rid of the warnings I was seeing in jointcal.
Jenkins run: https://ci.lsst.codes/blue/organizations/jenkins/stack-os-matrix/detail/stack-os-matrix/29612/pipeline
Looks okay to me given what we know about the problem.
Thanks for the quick review.
Merged and done.
Having just seen a mess of these warnings running jointcal tests, I'd like to clean them up. However, it's not clear to me whether our intended usage here is of the perf_counter or process_time variety. In fact, given that we're logging that value and not doing differences in this code, it seems to me that neither of those options is correct for us. Is the value of cpuTime logged by pipe.base.timer.logInfo() used anywhere?