Present: Vito Di Benedetto, Lynn Garren, Jim Kowalkowski, Robert Kutschke, Thomas Junk, Katherine Lato, Saba Sehrish, Luke Simons, Erica Snider
Remote: David Adams, Gleb Sinev, Others
Project Status Report [Erica Snider]
-
Mu2e uses Geant4 with Qt; LArSoft is not aware of being using it; the proposal is to adopt for LArSoft distribution the Geant4 UPS package without Qt
-
OSX 10.11 ("El Capitan") is declared as a "known to work" platform
-
needs tweaks on the user system side
-
OpenSSL should be installed as a keg package, that is not used as a system library but needs to be explicitly
-
distribution might happen via binary UPS products, not guaranteed though
-
[David Adams] should DUNE go ahead and recommend the update? [A] we would not recommend it
Actions
-
need to know from DUNE if the coming one is going to be a production release
-
will poll about distributing the Geant4 UPS with Qt and distribute the one without
-
LArSoft is not recommending the upgrade to OSX 10.11 ("El Capitan")
LArSoft/LArLite integration status [Ariana Hackenburg]
-
progress
-
asking what to do with
PhotonVisibilityService, since there are two versions (LArSoft's and LArLite's)
-
wiki page is in LArSoft wiki
-
[Lynn Garren] is it a problem to have
larsimobj data products moved into lardataobj? [A] it seems ok
Documentation report [Katherine Lato]
-
the "Using LArSoft" instructions have been revamped
-
reminder: add documentation of your algorithms/services!
Association usability project [Saba Sehrish]
-
reminder: work aimed to simplify an access pattern for associations following a specific prescription
-
this pattern preserves the order the associations were created in
-
trying to avoid overheads like temporary containers
-
presenting:
for_each_associated_group(), applying a function/callable to everyone of the "left" objects in the association, with all its associated objects
-
for example, call the function for each track, every time passing to the function all the hits associated with that track
-
the example on slide 7 feels more complicate, mostly because we are not yet very used to this programming style
-
we'll move toward it with time
-
it's very beneficial for multi-threading
-
[Thomas Junk] it's interesting that this is the style that will help with multi-threading... but the
push_back in the example seems to require some attention from that point of view [Jim Kowalkowski] this is an example which is not tuned for multi-threading in mind; this programming style is a big step toward multi-threading, but it's not enough by itself
CI integration report [Vito Di Benedetto]
-
IFDH beta version is being tested to verify the solution of the long-standing problem of lack of directory listing in the protocol
Data product review: phase II [Gianluca Petrillo]
-
four areas of intervention proposed
-
delivery time: January 2017
-
[Erica Snider] we'll need to define deliverables
-
[Lynn Garren] changes in nutools will need to be discussed with NOνA as well
Actions
-
need to define deliverables
There are minutes attached to this event.
Show them.