Present: Joanne Bogart, Toby Burnett, Xin Chen, Richard Dubois, Heather Kelly, Sean Robinson, Leon Rochester, Alex Schlessinger, Karl Young
Random report: (Karl) Has a program with the proposed scheme, including initialization, using 4 separate streams. Still to do:
try different engines to see which performs best for our situation
see about setting up the facility as a (Gaudi) tool.
Nightly builds/Release manager: (Alex) Output is generated from results of the build and test process which have been stored in a MySQL database.
Gleam status: (Toby)
Udine folk had to change the main program in order to get FRED to behave reasonably. Toby had a similar problem with GuiSvc which he fixed by implementing the (Gaudi) IRunable interface there. The same technique should work for FRED.
The build failures to be seen in Alex's automated build output above are due to mundane matters like a misspelling, now fixed, and the clean up of the Event package. Because of the latter, a file which was #included unnecessarily from TkrRecon no longer exists. The reference has been deleted, but TkrRecon needs a new tag. Leon will see to it, after checking that everything else in TkrRecon HEAD is ready to go.
(Heather) One other dangling item is the need for access methods for the (Event) McIntegratingHit class so that it can be written and read in the minimal MC tree case.
Incomplete identifiers: (Joanne) Changes as proposed in support of using a top volume other than LAT have been implemented in the detModel and GlastSvc packages. They will be tagged as soon as package owners Riccardo and Toby (in that order) have a chance to look them over. The last step is for Riccardo to make some mods to G4Generator to use the new service getIDPrefix() to generate full volume identifiers in all cases.
ROOT scheme evolution: (Heather) Ostensibly ROOT handles schema evolution automatically, however we had a counter-example when a TRef was removed from a class in version v2r4 of the mcRootData package. This was actually a bug in ROOT which Rene has now fixed. Until we move to a newer version of ROOT, one can avoid the problem by using mcRootData v2r5, which has reinstated the TRef.
Rootcint regularization: (Toby) About ready for release, pending OK from Heather.
Schedule: (Richard) Our software schedule has been recently updated. Mark Strickman will be incorporating current CAL information shortly.
Relational tables: (Richard for Marco) Marco has added a usage example to CalDigi. In the process he noticed a distinct lack of digis, perhaps due to incorrect thresholds [Ed. more problems with the test Root file; energies and moments were trashed].
Unhelpful test running: (Leon, Toby) What about instituting some way to avoid the building and running of test programs which currently occurs with VCMT? We need a way to just build dll's of the used packages. Selecting Gleam-app as the project rather than all (the default) avoids the Gleam test program, but not test programs of used packages.
Orbit simulation: (Toby) See this status report, including pretty picture. Toby had fruitful discussions with Gino Tosti and other Perugia collaborators on this topic. Sean will be incorporating some of their code.
previous | minutes index | next |
J. Bogart Last Modified: 04-Aug-2004 15:39:57 -0700