Core Minutes 17 May 2011

Core Minutes 5/17/2011

Present: Joanne Bogart, Anders Borgland, Stefano Ciprini, Warren Focke, Dario Gasparrini, Tom Glanzman, Elizabeth Hays, Heather Kelly, Chuck Patterson, Tom Stephens, Tracy Usher

Symposium Report (Tom G.) It was, over-all, great: great job by organizing committee, great talks. The banquet was held at the Vittorio Emmanuele monument. From the roof one gets a spectacular view of Rome, particularly striking at sunset.
(Liz) The work of the Instrument Analysis team got a very positive reception.
FSSC: (Tom S.) No particular news.
Documentation: (Chuck) sends the following:

The GBM Gamma-ray Burst Analysis tutorial has been updated, and I'm currently working on how to setup pointlike on SLAC Central Linux. For some reason, I keep bumping into CALDB problems. Other than that, it's business as usual.
Pass8: (Tracy) There will be a Pass8 meeting this week. Immediate goal is to make another tag which uses the new G4 [9.4.p01. See report from Heather below], hopefully within the next two weeks.
G4 and CLHEP Upgrades: (Heather) writes as follow) Francesco has committed his G4 upgrades to CVS. We also intend to move to CLHEP 2.1.0.1 which requires modifications to a number of packages to utilize the CLHEP namespace properly. Heather is working on the finishing touches and will commit them to CVS (hopefully today) and tag to create a new GR LATEST (and ST SCons LATEST since ST also depends on CLHEP).
Python upgrade: (Heather writes as follows) Recently there has been interest in trying out pycwsgrid2 as an additional python module. This caused a flurry of discussion concerning how we install python and suggestions were made to try out easy_install and pip. At Jim Chiang's request we'll take this opportunity to try out python 2.7.1 to inch us closer to python 3.0. Heather has installed python 2.7.1 for all the linuxes and is working on SnowLeopard and Windows versions. She has tried out easy_install and pip and both are easy to use. Some modules still require hand installation, such as numpy and scipy to ensure we indicate g77 versus gfortran for the various platforms. We must be sure that we link against the correct version of the fortran libraries, to match what was used in the system's version of BLAS and lapack. We have added the nose module to allow use of the NumPy and SciPy unit tests. We also rediscovered that the rhel4-64 version of BLAS at SLAC is missing some symbols, which causes some complaints about missing symbols in the SciPy unit test. In addition, numpy 1.5.1 has a known issue on rhel4-64 with gcc34 (the fix is to move to gcc4) For now we'll remain at python 2.6.5 on rhel4-64. (Tom G.) Is AplPy included? (Heather) Perhaps not [but when she took a closer look, she discovered it was in the new build after all].
Core talk:
- MySQL migration (Heather) We believe we're ready, but Arash is on vacation in June and the Computer Center is also waiting on some new hardware. Therefore we expect to make the move in July to centrally-maintained MySQL 5 for databases now hosted on glastlnx01.
- glast-win04 (Tom S.) was in the midst of trying out using this machine as our sole SCons Windows build machine, accessing externals and so forth locally, when a hardware problem cropped up. The machine was shut down in order to reseat its memory and this appears to have fixed the problem. Tom is able to make the builds go manually. The next step is to start them via the daemon.
- tagCollector script (Joanne) has made a new version in which tagging is done package by package, with the SConsFiles package put off until the end. This should be less subject to clashes with other operations going on in CVS; we'll know better once someone uses is. The documentation has been updated for the new version. There are a couple small changes in the call interface and a difference in the way the file packageList.txt is used. It is now a required component of any container.
- GRBAnalysis disposition (Joanne) GRBAnalysis currently is a kind of mongrel package: not quite a full-fledged container in its own right but not belonging to the ScienceTools container either. Various possibilities are under consideration. More information from the developers on exactly how GRBAnalysis is to be used is needed to choose the most suitable among them.
Announcements
- SCons meeting Tomorrow (Wednesday), 10:30 AM PDT

minutes index

J. Bogart, Last Modified: