Previous status report; Joanne's meeting minutes
As promised, we have changed to CMT v1r10p20011126, at SLAC linux and UW TS. All collaborators are urged to download and build it. (Appreciation to Joanne for pointing out that the web instructions were not appropriate: since corrected.)
A little caveat for Windows users: it requires GlastPolicy v4 to build Glast code. (Perhaps Gaudi too, need to check.)
Glitch at SLAC: the installation was destroyed somehow, had to be rebuild. Some vulnerability apparently not understood.
Problem in general: since private set statements are lost, this change breaks the setup for some of the programs in pdrApp v7r2, including pdrApp itself. Not backward compatible as hoped.
Installed at SLAC, validated at CERN for Red Hat. Attempting to make it our standard. (Heather and Alex.) Note that we must change from the Red Hat default in order to compile Gaudi! (Burden on people maintaining own boxes, sorry! Are there RPM's for them?)
Note that we must assure that code developed under windows compiles with this, before expecting our European colleagues to collaborate. (But a big question is at what point: check in? tag? or only release?)
Just in: Heather has had to backtrack to 2.95.
Under study. Visual studio files are now in XML format, Toby has preliminary fragments that appear to work ok. Some simple packages compiled. But CMT has to be modified, or need a wrapper (like vcmt) to rename output files.
Glitch: the current (version 6) Visual studio files that we create cannot be converted.
Imported to our cvs, basic stuff compiles and runs test programs.
Change in interface breaks GlastSvc, Heather looking into what we have to do.
Upgraded to allow for transients. Package flux has been merged into FluxSvc. (Sean)
About to institute a new scheme that deals with the predicted time interval to the next particle, rather than inferring from the flux.
Johann in the process of including a GRB model.
I've updated a page that I was using to keep track of the resources and comments about the situation. We have been promising Windows developers a semi-automated tool to easily check compilation of a packaged under development on linux. A subset of the RM will satisfy this need, which is important as long as the Linux compiler is more strict.
No change with G4. (Still works!) Complaints from RD not addressed yet :-(.
GismoGenerator checked in and tagged (v6r0).
- Breaks current head of TkrRecon! But Tracy has tried out THB suggested mods, says OK. Stay tuned.
- Also, it turns out that the order of the Si diodes associated with each log is different from the assumption of the current digitizer code. Need to have a CAL guy fix and test.
- Detector display is not good: no colors, default is off, many phantom enclosing volumes showing.
Riccardo's new package that will introduce HepRep to our environment, allowing much nicer graphics. Here is a block diagram of the proposed design.
How does one control G4's use of random numbers? Anyone??? Please????
T. Burnett Last update:
Wednesday, 04. August 2004 15:42:10 -0700
