Present: Heather, Joanne, Karl, Malcolm, Richard, Toby, Tracy
Core Report: (Toby) See Toby's writeup. Issues at the forefront:
Getting a unix build environment, similar to that provided by the U.W. Windows server, working smoothly at SLAC. This is a high-priority item and we're almost there.
CMT. After some experience there seems to be no reason not to move ahead to v1r9 and at least some advantage in the new, more flexible implementation of cmt broadcast. Karl will do some further testing to make sure there is no down side to our moving to the new version.
Even with this version there still is a problem with the way (recursive) cmt checkout handles version numbers in a package's requirements file: it always goes for the latest tagged version of a used package, even when some other version has been explicitly requested.
We need to look into the use of .cmtrc to handle, e.g., setting of CMTPATH. .cmtrc goes in your $HOME directory, which means that, without some extra machinery, it will not handle the case of a single user working on two projects which require different CMTPATHs.
Latest release of pdrApp is v6r5p1. This does not include the new version of TkrRecon, which doesn't yet compile on Linux. It does build and run with no problems on Windows. Since there is high-priority (Steve and Bill) short-term demand for this code on Windows only, we will go ahead and make the new release first, then finish fixing the Linux problems [ed.: and make a patch release then?].
New Recruits: (Richard) We have made a nearly-official offer to our leading candidate for the full-time regular slot and it's been nearly-officially accepted. We await the final grinding of the wheels of bureacracy.
The new person will probably take on the task of librarian (among other things), where the responsibilities of the librarian include at least
CMT guru; i.e., knowing CMT itself well, knowing GLAST-specific wrinkles in its use, and keeping abreast of other experiments' experiences with it.
Maintaining releases and documentation, both for users and for developers. (Speaking of documentation, we need to do some merging of documentation of pdrApp, userAlg, etc., into the mainstream.)
Two of our three new masters students part-time people are getting set up; the third has decided to go elsewhere, so we are again in the process of interviewing for this third slot. There is no shortage of applicants.
Upcoming virutal workshop: (Richard) It is now fixed for the week before Thanksgiving, Nov. 12-16. Focus will be on simulation/reconstruction (we're still looking for a time for a separate workshop on science tools). Topics will, at a minimum, include
Design review of the current state of things
G4 work: integration of G4 with Gaudi, hit structure, etc.
Data Manager Review: (Karl) It will be held tomorrow, Thursday, Oct. 11 at 1 PM. You might want to take a look at his spec first.
Binary Distribution Snag: (Karl) If one naively builds an executable linked to shareables, the actual path at link time to the shareable gets embedded in the executable, making it impossible to copy it all elsewhere and run successfully. This doesn't affect dynamically-loaded libraries, but does affect our external libraries. One solution is to link statically to these libraries. The size of the executable will increase noticeably.
J. Bogart Last Modified: 08/04/2004 15:41