LHCb Core Software Meeting
Date and Location
June, 4 2008
10:30 - 11:30
CERN (2-R-030)
Attendees
Gerhard (EVO),
Hubert,
Marco Ca.,
Marco Cl. (minutes),
Patrick (EVO),
Philippe
Apologies
Thomas
Subjects
News
- LCG 54g is being built, LCG 55 is under preparation (at least the relevant CVS tags should be ready by the end of the week).
- Application Area Meeting
in the afternoon:
- ROOT data model evolution
- From Architects Forum (Marco Ca.):
- Decided to change the library suffix on MacOS 10.5 from .dylib to .so (Fons' suggestion).
- The compiler for SLC5 will be gcc4.1, if we want gcc4.3, we shall distribute it as an external.
- The MacOS 10.4 build machine will not be upgraded yet to 10.5.
Software Releases
Gaudi (Marco Cl.)
- Applied the new state machine patch after some last minute changes in the interfaces.
- A couple of bugs in gaudirun.py has been reported and will be fixed ASAP.
- Hubert changes for python binary modules and PATH cleanup work in the nightlies.
LHCb, Boole, Brunel (Marco Ca.)
- LHCb v23r7
is almost ready (Hubert will start the build tomorrow). Features:
- OT encoding/decoding
- clean up of BeginEvent and EndEvent (whenever possible, BeginEvent is used)
- Recent mail discussion triggered by a problem caused by a NaN in some part of the tracking code. Gerhard prepared an Auditor that enable floating point exceptions for specific algorithms (by name). Marco Ca. adapted the code to work on Win32 too. MacOS version is dummy (throws an exception) for the moment.
- Using the nightlies, currently only to fix compilation warnings. Unfortunately there are no builds for the last two days (disk space problems, we should not keep the builds on the build machine as soon as we have AFS copies). Once the release is done, we should move the nightlies to the new Gaudi.
HLT/DaVinci (Patrick)
- Waiting for the release of LHCb
- Translated all the option files of the tutorials to Python.
Release and Deployment (Hubert)
- Released:
- Releasing Panoptes
Round Table
- Philippe
-
- We need to understand and solve the problems occurred using the Oracle CondDB during CCRC (CORAL failing to retrieve Oracle credentials from LFC).
- During CCRC, we experienced many problems with storage systems (especially dCache). We are investigating the possibility of using "xrootd".
--
MarcoClemencic - 04 Jun 2008