- 55 day run, GlastRelease: (Heather, Richard)
The Onboard Filter fix went in and background runs were made with v11r2;
Yet another fix was added in v11r3. We need to decide specifics of the
earth albedo file for the pointed run with interleave. [Update from
Richard: He and Julie believe the current scheme will not handle
pointed observations quite properly so it is not (yet) worth the effort to
refine this model. We'll go with what we have and see how it looks.]
(Toby) offered to make another release to pick up Navid's
fix to GlastPolicy,
reinstating debug symbols in optimized builds. Is there anything else that
should go in? (Heather) would like to include a rebuilt Gaudi
which requires Boost 1.31 (the one we distribute) rather than 1.32 for the
benefit of off-site Linux users. (Tracy)
There is a new tag of G4HadronSim which should take care of earlier
problems with the alternate hadronic shower model. (Heather) And Navid
has made the change to get rid of CMT-generated "config" environment variables
which we don't use.
- Windows VS 2005: (Toby, ..)
He has run into several stumbling blocks, largely involved with
external libraries. For the time being he's attempting to build only
ScienceTools, not GlastRelease, since GR requires Gaudi.
He's having trouble with Xerces 2.7 (managed
to evade a linking error, but there are run-time problems), fftw and ROOT.
Heather expects to provide him a newly-built ROOT. The fftw problem may
have to do with shared-versus-static library. Jim has offered to build the
new version of fftw as a static if it isn't already available. Currently
Toby can build ScienceTools but there are serious run-time problems, such
as memory corruption. It seems one can't load dll's built with the
old compiler.
(Toby) One way around some of the external library problems would be to
allow interface requirements files to select different native versions,
depending on
OS. [This was the case for the MySQL interface package
MYSQLEXT a while ago when we were using
3.23.49 on Linux and 3.23.52 on Windows. ed.]
(Toby) On the bright side, the Python code which constructs
VS project files
from CMT requirements files is essentially done and working.
(Heather) What are the prospects for RM builds using VS2005?
(Navid) would first like to make the switch to the new glastrm account
for Windows builds.
- RootIo: Tracy's refactoring is now complete and
committed. (David) will make another pass through the code,
reorganizing
so that non-Gaudi facilities can use the same tools.
- Sys Tests: (Heather, Leon) The changes
seen in AcdDigi
test output for GR v11r2 are consistent with updates to the
code made by Luis, hence
probably not a cause for alarm.
(Heather) An old problem with schema evolution concerning vector of
Double32 which we thought we had licked has popped up again in Sys Tests.
- Geometry (Joanne) has started to
implement a simplified side tile shingling scheme as proposed on the last page
of a
presentation
made to the Calibration & Analysis Meeting a couple weeks ago. She
made a further adjustment, extending rows 1 and 3 to avoid a substantial
gap not in the real instrument along the edges where adjacent sides
come together. See
pictures of the shingling with and without the adjustment to the width of
rows 1 and 3. A
view with some tiles filled gives the clearest idea of the new
arrangement.
(Richard) What is the prognosis for screws? There are measurements of
actual hole positions available. (Joanne) I have that information (work
done by Luis Reyes). It only concerns top tiles. It would be somewhat
more work to simulate these positions as opposed to a simplified symmetric
model, but in either case it's just a question of time and tedium. The
work to be done is straightforward.
- CCB Control (Richard, Heather) See Anders'
proposal.
Before too long we should merge EM and GR; that is, incorporate into GR those
facilities from EM it doesn't already have. In particular this would
include packages like calibGenCAL.
J. Bogart, Last Modified:
01-Jun-2010 15:47:34 -0700