Chris Backhouse, Martin Frank, Brendan Kiburg, Rob Kutschke, Adam Lyon, Chris Jones, Brian Rebel, Kanika Sachdev
artists: Lynn Garren, Chris Green, Jim Kowalkowski, Kyle Knoepfel, Marc Paterno
art release 1.13.00
Chris reports that art release 1.13.00 is very close to being ready. Lynn should be able to start building art on Monday 2/16.
 Yosemite build:
Lynn reports that she has built everything through ROOT on Yosemite. Lynn is working on Geant4.9.6.p04. v4.10.1 is not yet ready for building.
 Added enhancements to buildtool (package cetbuildtools):
Chris reports issue of build failing when trying to remake the dictionaries, whose include references have changed. This problem stems from the dictionaries being generated files. The newest version of cetbuildtools, shipped with art v1.13.00 will include a – -deleted-header option that will clean up (generated) files with that header before building.
 Additional build options
Chris has added commands to tell cmake to operate in trace mode, or debug output mode, in case users are having trouble with composing and parsing cmake files.
 Improvements to FHiCL usability
Chris has been able to improve FHiCL location information so that when there’s a complaint of a parse error, it tells use the line and character, and the place from which that file was included (in base file), it shows you the line at issue, along with a caret underneath.
Rob asked if FHiCL is robust against a missing newline at end-of-file. Chris says he thinks it is not a problem. Marc says that if it is, it should be reported.
Mu2e requests that ProductNotFound rethrow by default, which currently if thrown, skips to next event. Chris Backhouse agrees that this should be the desired behavior. The behavior change to art that all exceptions would rethrow by default. This could be overridden if desired.
Current art exec options:
– -rethrow-default (makes all exceptions rethrow by default, overridable by FHiCL),
– -rethrow-all (not overridable by FHiCL).
Chris Backhouse also voiced a suggestion that the default behavior should be for modules that call produces<> should also call put, otherwise throw exception. This behavior could be overridden in FHiCL. This suggestion was received favorably by those in attendance, and Marc suggested that Chris file a new issue in the tracker.
Discussion about making quicker decisions:
Marc suggested a better technique should be developed to resolve issues quickly. The stakeholders meeting is not an ideal forum since official computing representatives from experiments are often not present. Rob suggested that the appropriate discussions can be had in the issues tracker, but notices of possible/probable changes to art behavior should go to a broader distribution. Possibility of creating a limited email list of individuals that represent each experiment for issues like this.