LHRS replay is tested to be crashing even loading only LHRS; seems very sick. Checking with OLE, he found cint crashed at a "cout<<" line for no reason. There was an analyzer update on Monday. However the update is checked to be very reasonable.
There seems to be a problem with root C interpreter (cint). Therefore, core part of replay script is compiled to avoid going through cint. Tested with do_all() for both LHRS and bigbite, the problem seems to be fixed.
The advantage of compiling the replay script is that it's mroe stable and more efficient, while disadvantage is that it's more complicated to maintenance. Pi0 replay scripts fully compiled; Gep and Transversity ones are non-compiled. Now d2n scripts are compiled for only the core parts: The user part of replay is still easy to modify without bothering any compiling.
Special Notice for d2n: During next analyzer/root update, or moving replay to farm, ReplayCore.C should be recompiled by calling ".L ReplayCore.C++" in analyzer.
A copy of this log entry has been emailed to: ole, vasulk, moffit, gjin