Project

General

Profile

Support #4545

Help Martin track down undesirable behavior in NovaDDT shared memory reading.

Added by Christopher Green about 7 years ago. Updated about 7 years ago.

Status:
Closed
Priority:
Normal
Category:
Third Party
Target version:
-
Start date:
08/14/2013
Due date:
% Done:

100%

Estimated time:
1.00 h
Spent time:
Scope:
Internal
Experiment:
NOvA
SSI Package:
Duration:

Description

The NovaDDT effort has been suffering from undesired behavior by the ddt-filter program for some time now. Upon a pause in the data stream, the program hangs, and data-taking does not resume when the stream starts back up.

History

#1 Updated by Christopher Green about 7 years ago

  • Status changed from Assigned to Feedback
  • % Done changed from 0 to 90
  • Estimated time set to 1.00 h

Beyond misbehavior identified in the Nova-specific code leading to an attempted exit upon a pause in the data stream, there appears to be a real bug in the way that art::Source handles detail classes with the generator trait (Source_generator<T>::value = true) set in art 1.02.06.

There are two possible solutions: upgrade to a more recent version of art (1.07.01 or 1.08.02), or specialize the Source_wantFileServices<T> to false instead of specializing Source_generator<T> to true. Then, the number of times that T::readNext(...) will be called following a return of false will be one less than the number of dummy, "file names" in the source.fileNames FHiCL configuration variable.

Let me know what you decided to do, and whether it works.

#2 Updated by Christopher Green about 7 years ago

  • % Done changed from 90 to 100

Martin has found ways to do what he needs and will upgrade to the latest art when possible.

#3 Updated by Christopher Green about 7 years ago

  • Status changed from Feedback to Resolved

#4 Updated by Christopher Green about 7 years ago

  • Status changed from Resolved to Closed


Also available in: Atom PDF