Paul Price added a comment - 02/Jun/17 10:48 AM On Python 3.5.x, this change produces: OverflowError: timestamp too large to convert to C _PyTime_t https://bugs.python.org/issue25155

Paul Price added a comment - 02/Jun/17 10:57 AM Going to reduce the overkill factor.

Paul Price added a comment - 02/Jun/17 11:02 AM Sorry to bother you again, Russell Owen , but would you please have a look at these simple changes? price@pap-laptop:~/LSST/pipe_base (tickets/DM-10633=) $ git sub-patch commit 355fc941f07c724a4c25081f0abfa8bb661c97e6 Author: Paul Price <price@astro.princeton.edu> Date: Fri Jun 2 14:01:13 2017 -0400   TaskRunner: revise insane timeout This value is more reasonable, and doesn't trigger Y2K38 bugs (e.g., https://bugs.python.org/issue25155 ), but should be large enough to prevent annoyance.   diff --git a/python/lsst/pipe/base/cmdLineTask.py b/python/lsst/pipe/base/cmdLineTask.py index cfc4d26..34d6174 100644 --- a/python/lsst/pipe/base/cmdLineTask.py +++ b/python/lsst/pipe/base/cmdLineTask.py @@ -140,7 +140,7 @@ class TaskRunner(object): [1] http://bugs.python.org/issue8296 [2] http://stackoverflow.com/questions/1408356/keyboard-interrupts-with-pythons-multiprocessing-pool) """ - TIMEOUT = 999999999999 # Default timeout (sec) for multiprocessing + TIMEOUT = 3600*24*30 # Default timeout (sec) for multiprocessing def __init__(self, TaskClass, parsedCmd, doReturnResults=False): """!Construct a TaskRunner   commit b33ffe24b6716cee286993c3493f7da27d464296 Author: Paul Price <price@astro.princeton.edu> Date: Fri Jun 2 14:01:42 2017 -0400   TaskRunner: fix doc We don't use TIMEOUT_DEFAULT, but TIMEOUT (there is no TIMEOUT_DEFAULT).   diff --git a/python/lsst/pipe/base/cmdLineTask.py b/python/lsst/pipe/base/cmdLineTask.py index 34d6174..d5f93d1 100644 --- a/python/lsst/pipe/base/cmdLineTask.py +++ b/python/lsst/pipe/base/cmdLineTask.py @@ -135,7 +135,7 @@ class TaskRunner(object): Due to a python bug [1], handling a KeyboardInterrupt properly requires specifying a timeout [2]. This timeout (in sec) can be specified as the "timeout" element in the output from ArgumentParser (the "parsedCmd"), if - available, otherwise we use TaskRunner.TIMEOUT_DEFAULT. + available, otherwise we use TaskRunner.TIMEOUT. [1] http://bugs.python.org/issue8296 [2] http://stackoverflow.com/questions/1408356/keyboard-interrupts-with-pythons-multiprocessing-pool)

Russell Owen added a comment - 02/Jun/17 1:57 PM (The pull request looks weird because this is not rebased to master, so I used Source Tree and my comments are below.) Your changes look fine. I confess I was uneasy about the huge number you used initially, but I figured if it passed Jenkins then I had no valid grounds for complaint. So I much prefer your new "long enough but not crazy" value. Thank you! One minor suggest which you can take or leave: rename TIMEOUT to DEFAULT_TIMEOUT or DEFAULT_TIMEOUT_SEC.

Paul Price added a comment - 02/Jun/17 2:37 PM Yes, I should have listened to you the first time. But it passed Jenkins.... I blame the python bug that it triggers! Having just fixed up the docs for TIMEOUT , I think I'd prefer to leave the name (and I don't want to have to go searching for uses or RFC it). Merged to master.