Core Minutes 2/07/2006
Present: Joanne Bogart, Philippe Bruel, Toby Burnett,
David Chamont, Eric Charles,
Jim Chiang, Johann Cohen-Tanugi, Jan Conrad,
Dan Flath, Warren Focke, Berrie Giebels, Navid Golpayegani,
Heather Kelly, David Landriu, Bryson Lee, Julie McEnery,
Igor Pavlin, Chuck Patterson, James Peachey, Leon Rochester, Tracy Usher
- DC2:
- Pulsar problem Jim has suggested and Nicola has implemented
a fix concerning use of random numbers.
- GRB problem Nicola found and fixed a memory leak. We'll find
out soon whether this also cures the crashes which occur with GRBs and
interleaving.
- External libraries: Leon requested that a
newer version of FRED be incorporated into GlastRelease. This caused
no problems for RM builds of GlastRelease, but it did break the Installer.
Not only could the (GUI) Installer not find the newer FRED binary, but
it would hang in the attempt. The reason is that Navid has to manually
copy the external library binaries to the proper place for the Installer
to find them. In fact, normally he also has some work to do even to
get RM to work with a new library, but in this case that had already
been done.
Navid must be informed when
any change to external libraries is contemplated.
The copy has now been done so everything is in order again.
- Build order on Windows: (Heather, Navid,..)
AcdUtil has two libraries: AcdUtilCommon and AcdUtil.
The first is [was]
a non-component shareable and the second is a Gaudi component.
They must be built in the order listed, but Visual Studio (used by MRvcmt)
builds them in an apparently random order.
RM builds them in the order they appear
in the requirements file, hence correctly. One way to force the build
order for Visual Studio is with an explicit dependence in the
requirements file.
- GlastRelease HEAD: Toby asks that everyone
take a look at it and let him know if there are other tags which ought
to go in. (Leon) Would be nice to get in a fix for a memory leak in
Beamtest, but it isn't quite ready yet.
- FSW data readiness: (Heather) Eric
has dealt with the TDS-related matters, ROOT classes, etc.
She has the other end, involving use of Bryson's merging library.
Run id will come from LICOS. We need to change the way it's
stored (can no longer use float) which will affect at least tuples
and FITS. Ultimately she would like to represent
it as unsigned int and,
if it doesn't cause too much trouble in the short
run, she would prefer to make that switch now; otherwise we could
go to just int right away and make
another change later. In either case, Leon will have to engage in
some fancy footwork involving branches.
J. Bogart, Last Modified:
01-Jun-2010 15:48:15 -0700