Core Minutes 2/11/2003
Present: Joanne Bogart, Toby Burnett, Richard Dubois, Heather Kelly, Leon Rochester,
Alex Schlessinger, Tracy Usher, Karl Young
-
GlastRelease/Nightly builds:
(Alex, Toby, et. al.)
- Somewhat mysteriously, last night's build using latest tags (in
particular, using Geant 5.0) succeeded. Except for FluxSvc,
G4Propagator and ntupleWriterSvc, test programs ran successfully.
Perhaps this signals an end to our difficulties with Geant
5.0.
- For Geant, we only need the "compound libraries".
This accounts for the shorter list we're now using for Geant 5.0.
- Now that we appear to have a workable IExternal,
it's time to put
together a complete set of external library binaries in our ftp
area.
Also need to make new binary for RootHistCnv.
- We should try to get Tracy's CMTification (TM) of the Geant build
in use on Linux as well as Windows.
- Compiler oddities: the default SLAC-supplied gcc 2.95.3
binary was built in a Redhat 6.2 environment, which means it looks
for different versions of certain things than those normally
present on our officially-recommended Linux OS, Redhat 7.2.
Alex has made his own build of the compiler which is OK for
now.
We need to make a request to SCS for better compiler
support; if we do, it should be forthcoming.
- (Toby) Output of test programs sometimes just stops without any
diagnostic information. (Alex) This happens when there is an
untrapped core dump.
- (Richard) Are we ready to
re-enable email notification of nightly
build problems? (Alex) Yes, just about. Some file (e.g.,
release.notes or a new file contacts.txt in the doc directory) could
be used to find email addresses of people who should be notified
- (Toby) Current format of nightly build output makes it awkward to
see results for packages contained in IExternal. (Alex) Should be
possible to
fix this.
- It would help in reproducing and debugging nightly build problems
if
a new option was added to glastpack to check out latest tags of
everything.
-
G4Propagator needs to be added to Gleam requirements and a proper
scheme for job options has to be designed and implemented.
System tests: (Karl) There is still
a little work to do on scripts to get this going with Root 3.04.02. Should
have some results later today.
Calibration infrastructure:
(Joanne)
-
Calibration-related packages have been integrated with the standard
stuff: MYSQLEXT is now in IExternal (binaries for Windows and Linux are
available from the
ftp site, but not yet incorporated in the Grand Unified
Tarball). calibUtil, CalibData and CalibSvc are in
GlastRelease.
-
There was some discussion of policies,
but we're not done yet. Especially could use input on 4 and 5.
Keep those cards and letters coming! Richard pointed out that 3
(how data is organized) should not really be a concern to Calibration
Infrastructure. Note that although flavors may be used for a variety of
functions, they will not be used to specify instrument (LAT, EM, etc.)
That is done through a job options parameter. Flavor and instrument are
two separate columns in the MySQL table used for metadata.
-
Joanne has written a ROOTless skeleton for the ROOT Conversion
Service; Heather is just about ready to dive in and make it realistic.
-
News and notes: (Richard)
-
We have an estimated 10 weeks of Heather-availability before she is
otherwise occupied.
-
Most of our CHEP03 abstracts have been accepted; still waiting to
hear about a couple.
-
Efforts are in progress to get access to Insightful Miner at an
affordable price.
- Form habit of reviewing this list at each meeting. (2/4/03) We forgot all
about this, or maybe we just ran out of time.
is used above to mark possible action items.(2/11/03)
- Respond to Traudl's release
management proposal.(11/20/02)
- Come up with procedures for development within release branch (11/20/02)
- Resolve calibration/timestamp issue. (11/20/02)
- Update Gleam User's Guide. (9/25) (10/30/02) Mostly
done.
- Assemble minutes, etc. from Core Week reviews of AcdRecon and GLAST Gaudi
use. (9/18/02)
- Alex will look into technical solutions (front-end script?) for
controlling CVS commit access. (6/5/02) Doable; also circumventable. A
back-burner project. (7/10/02)
- CAL geometry changes (5/8/02) Partly done.
(7/3/02) Awaiting further information from CAL (7/24/02)
J. Bogart Last Modified:
04-Aug-2004 15:40:08 -0700