Changes in this section took place 13-Aug-98:
A list of all modified files:
General notes:
mvd98run.kumac and mvd98ini.kumac were changed to make them compatible
with the new module calling sequences. These could be used as examples
of what needs to be done in luxor.
The new cvs version was tested on the LANL hp's and on rlnx02.
I had a minor problem on rlnx02, which I hope is fixed. staf could
not find the definition of new table dMvdVertexOut when I used it.
It was there, but not "seen" somehow. I put in a comment line
in dMvdVertexOut.idl and remade my shared libraries. Everything
was suddenly OK. I put this new and improved dMvdVertexOut back into
cvs and had no problems with a freshly checkout version on the hp. I
assume this problem is gone. I suspect some confusion about file
creation dates when I put everything into cvs and took it back out.
This code does not currently work under staf97. I will see if it
is easy to fix (I did not even try so far). It could not find
an include file for the staf histogram factory. This is easy
to fix, but it may be the first symptom of a series of related
problems. I suggest using staf98 if you can.
The staf histogram factory worked
as advertised.
http://www.phenix.bnl.gov/~markp/staf/histos/histogram.htm
I had to use some root stuff because this generic interface
does not have profile histograms or ntuples.
Instructions for building the MVD code in staf can be found at
this link.
The following idl files were changed. The changes will require
changes in kumac files which run the MVD software.
src/mMvbdNdEta.c and .idl
src/mMvcdNdEta.cc and .idl (new file replaces mMvcdNdEta.c)
src/mMvdHst.cc and .idl (new file replaces mMvdHst.c)
src/mMvdPseudoTrk.cc and .idl (new file replaces mMvdPseudoTrk.c)
src/mMvdSetHst.cc and .idl (new file replaces mMvdSetHst.c)
src/mMvdZcor.cc and .idl (new file replaces mMvdZcor.c)
wrk/mvd98run.kumac
wrk/mvd98ini.kumac
inc/dMvdMultOut.idl (new file)
inc/dMvdVertexOut.idl (new file, replaces dMvdPseudoTrk.idl and
dMvdZcor.idl)
module | tables removed from calling sequence | tables added to calling sequence |
mMvcdNdEta | dMVdHst dMvdPseudoTrk dMvdZcor | dMvdVertexOut |
mMvdHst | dMvdHst dMvdPseudoTrk dMvdZcor | dMvdVertexOut |
mMvdPseudoTrk | dMvdHst dMvdPseudoTrk | dMvdVertexOut |
mMvdSetHst | dMvdPar dMvbPar | -- |
mMvdZcor | dMvdZcor dMvdHst | dMvdVertexOut |
mMvbdNdEta | dMvdZcor dMvdPseudoTrk | dMvdVertexOut |
Other code changes:
mMvcdNdEta -- Add code to use staf table dMvdVertexOut. This is the "official" MVD vertex output table, which was not really used until now.
mMvdHst -- Add code to fill staf table dMvdVertexOut. This is the "official" MVD vertex output table, which was not really used until now. Moved code which "booked" the histograms from (staf module) mMvdSetHst to this module. Add code to use root for profile histograms and ntuple. Added code to fill a few histograms which were booked but not filled before.
mMvdPseudoTrk -- Add code to fill staf table dMvdVertexOut. Use a local array to find the vertex. Previously an hbook histogram was used. Remove hbook. Add diagnostic histograms (not used to find vertex) using the staf histogram factory. Fixed a 1 bin (300 micron) bug in vertex location.
mMvdSetHst -- Removed almost all code from this routine (it is now in mMvdHst). Now, all it does is create a directory called MVD for MVD histograms (with staf histogram factory calls).
mMvdZcor -- Add code to fill staf table dMvdVertexOut. Use a local array to find the vertex. Previously an hbook histogram was used. Remove hbook. Add diagnostic histograms (not used to find vertex) using the staf histogram factory.
The previous version of this page (dated 13-Jul-98) is here.