A GEM software library has been written and is compatible with HCANA

Add the ability to run cppcheck on the Hall C src/ directory .... perhaps with something like 'scons cppcheck=1' (much like the debug and standalone options are already implemented).

First step to see if this is really a problem. Are there events in which the LR assignment is wrong.

A recent build of hcana, via ROOT 6.08.00 indicates a few compiler warnings that need to be addressed:

src/THcPrimaryKine.cxx: In member function ‘virtual Int_t THcPrimaryKine::ReadDatabase(const TDatime&)’:
src/THcPrimaryKine.cxx:188:28: warning: unused variable ‘here’ [-Wunused-variable]
static const char* const here = "THcPrimaryKine::ReadDatabase";

In file included from src/THcRunParameters.h:10:0,
from src/THcRunParameters.cxx:10:
podd/src/THaRunParameters.h:21:29: warning: ‘virtual THaRunParameters& THaRunParameters::operator=(const THaRunParameters&)’ was hidden [-Woverloaded-virtual]
virtual THaRunParameters& operator=( const THaRunParameters& rhs );

In file included from src/THcRunParameters.cxx:10:0:
src/THcRunParameters.h:12:7: warning: by ‘THcRunParameters& THcRunParameters::operator=(const THcRunParameters&)’ [-Woverloaded-virtual]
class THcRunParameters : public THaRunParameters {
^
src/THcSecondaryKine.cxx: In member function ‘virtual Int_t THcSecondaryKine::ReadDatabase(const TDatime&)’:
src/THcSecondaryKine.cxx:303:1: warning: no return statement in function returning non-void [-Wreturn-type]
}

In file included from podd/src/THaBeam.h:12:0,
from src/THcRasteredBeam.h:13,
from HallCDict.C:69:
podd/src/THaRunParameters.h:21:29: warning: ‘virtual THaRunParameters& THaRunParameters::operator=(const THaRunParameters&)’ was hidden [-Woverloaded-virtual]
virtual THaRunParameters& operator=( const THaRunParameters& rhs );

In file included from HallCDict.C:76:0:
src/THcRunParameters.h:12:7: warning: by ‘THcRunParameters& THcRunParameters::operator=(const THcRunParameters&)’ [-Woverloaded-virtual]
class THcRunParameters : public THaRunParameters {

Comment from Gabriel Niculescu: " Because hcana calculates fptimes separately for each plane (as it should, one would argue) the mechanism of rejecting "out-of-time" hits based on average hit times does not work as well. There is no easy fix for this but I do not anticipate this to be a problem. One can get a similar result by implementing a method of rejecting fptimes (as opposed to individual time hits) in the THcHodoscope class. This is on my "to-do" list. "

The current tracking code only handles HMS chambers. Adding compatibility for the SOS chambers will get us close to what we need for the SHMS as the SOS and SHMS use chambers with the same geometry.

In Podd, the THaSpectrometer class gets the spectrometer angle and momentum setting from a database file db_run.dat. As THcHallCSpectrometer inherits from THaSpectrometer, it also gets the angle and momentum from this file. The Hall C ENGINE got these values from a database file, which hcana also supports. So, right now, hcana needs to have these values defined in a Hall A style database file and in a Hall C style.

What do we do about this? We could override the THaSpectrometer::ReadRunDatabase. We would need to duplicate the geometry calculations done there.

There is a (semi-)new TI config related blob in Event 125 that stores some trigger/TM related settings that needs a decoder in THcConfigEvtHandler.cxx. Runs after SHMS 1295 and HMS 718 will have blobs with the information below.

The word format is defined by this code in the hcvme0[12].c ROLs:

/* Log prescales, misc TI related information from ffile */
*rol->dabufp++ = (0xd0000000);        // 'misc' header block
*rol->dabufp++ = (0xabcd0000+1);      // last +N is version ID number
*rol->dabufp++ = nped;                //
*rol->dabufp++ = scaler_period;       //
*rol->dabufp++ = syncCnt;             //
for(i=0; i<NUM_PRESCALES; i++) {      // Input Trigger prescale settings
  *rol->dabufp++ = prescales[i];
}
*rol->dabufp++ = (0xd000000f);        // 'misc' trailer block

Need to add class(es) for beam information, utilizing EPICS and raster information.

In THcTrigDet.cxx the commit d6e15d1 contains experiment specific definitions located between lines 367-394.

Hi Guys,

More than an issue it's a question. How are the values for W computed? I track it down to a TLorentzVector but couldn't find the lines. Because I'm getting a bunch of 1e+38 values from the 2397 run replay at cdaql1 hcana ROOTfiles/coin_replay_production_2397_-1.root.

My commit e9ce104, which cleaned up a lot of compiler warnings broke the code. Now THaVar spits out lots of warnings. Commit 7edb1e4 is the last working verison. Will fix soon.

Merge in Zafar's work on the Cherenkov detector.

The Hall C ENGINE had an option to output certain information so that an external program could calculate the various hodoscope parameters such as time offsets, timewalk corrections etc. hcana needs the hooks to output the information needed by this external program. Ideally, this information should be able to be put in the ntuple/tree so that it can be analyzed by a root macro. Doing this will also allow one to change cuts without having to rerun the analyzer by doing the cuts in the root macro.

It would be good to structure the source code better.

Suggestions:

1. Put the interface headers in an include directory. This nearly all the header files I think.
2. Put (and install) the header files in their own directory. For example include/hcana/HEADER.h.
3. This leads to use cases like:

#include "hcana/THcParmList.h"

(note this doesn't have to happen in the current source code include directories can be added to compiler flags).

4. This header name space then leads to clearer class names. (Using class name prefixes is old fashioned and ignores the namespace feature of C++).

#include "hcana/ParameterList.h"

5. Clearer header file and class names leads us finally to functional separation. Perhaps by detector type: (note this is just a suggestion)

#include "hcana/cer/Cherenkov.h"
#include "hcana/cer/CherenkovHit.h"

Here is how it could look:

hcana/include
├── beam
│   ├── BCMCurrent.h
│   ├── RasteredBeam.h
│   ├── Raster.h
│   └── RasterRawHit.h
├── cer
│   ├── Aerogel.h
│   ├── AerogelHit.h
│   ├── Cherenkov.h
│   └── CherenkovHit.h
├── core
│   ├── Analyzer.h
│   ├── ConfigEvtHandler.h
│   ├── DetectorMap.h
│   ├── DummySpectrometer.h
│   ├── Formula.h
│   ├── Globals.h
│   ├── HallCSpectrometer.h
│   ├── HitList.h
│   ├── Interface.h
│   ├── ParmList.h
│   ├── PeriodicReport.h
│   ├── Run.h
│   ├── RunParameters.h
│   └── SignalHit.h
├── dc
│   ├── DC.h
│   ├── DCHit.h
│   ├── DCLookupTTDConv.h
│   ├── DCTimeToDistConv.h
│   ├── DCTrack.h
│   ├── DCWire.h
│   ├── DriftChamber.h
│   └── DriftChamberPlane.h
├── hodo
│   ├── HodoEff.h
│   ├── HodoHit.h
│   ├── Hodoscope.h
│   └── ScintillatorPlane.h
├── phys
├── raw
│   ├── RawAdcHit.h
│   ├── RawDCHit.h
│   ├── RawHit.h
│   ├── RawHodoHit.h
│   ├── RawShowerHit.h
│   └── RawTdcHit.h
├── recon
│   ├── CoinTime.h
│   ├── ExtTarCor.h
│   ├── PrimaryKine.h
│   ├── ReactionPoint.h
│   ├── SecondaryKine.h
│   └── SpacePoint.h
├── scaler
│   ├── Scaler9001.h
│   ├── Scaler9250.h
│   └── ScalerEvtHandler.h
├── shower
│   ├── ShowerArray.h
│   ├── Shower.h
│   ├── ShowerHit.h
│   └── ShowerPlane.h
└── trig
    ├── TimeSyncEvtHandler.h
    ├── TrigApp.h
    ├── TrigDet.h
    ├── TrigRawHit.h
    └── TIBlobModule.h

Again, this is just a suggestion to help others understand the code, and is common practice in C++ projects.

Eliminate or reduce the compiler warnings.

E.g.

unused variable 'x'
'x' may be used uninitialized in this function
'x' will be initialized after 'y' when initialized here
comparison between signed and unsigned integer expressions

A number arrays in THcHodosope, THcScintillatorPlane, and THcHallCSpectrometer have a fixed length of 53. (53 is the number of scintillator bars in the HMS +1). The length of these arrays should be derived from the parameters.

A number of these arrays are probably for holding processed hit information. This information would better be held in a structure or class with multiple variables. The THcDCHit class for thre wire chambers might be used as an example.

On MacOSX, I get the following warnings and errors when trying to compile hcana with the new raster classes:

g++ -o src/THcRaster.os -c -O -Wall -Woverloaded-virtual -pthread -pthread -m64 -stdlib=libc++ -fPIC -fPIC -DNDEBUG -DMACVERS -DHAS_SSTREAM -DHALLC_MODS -Isrc -Ipodd/src -Ipodd/hana_decode -Ipodd/hana_scaler -I/Users/brash/Dropbox/root/include src/THcRaster.cxx
In file included from src/THcRaster.cxx:14:
src/THcRaster.h:30:42: warning: array index 2 is past the end of the array (which contains 2 elements) [-Warray-bounds]
TVector3 GetPosition() const { return fPosition[2]; }
^ ~
src/THcRaster.h:52:3: note: array 'fPosition' declared here
TVector3 fPosition[2]; // Beam position at 1st, 2nd BPM or at the target (meters)
^
src/THcRaster.cxx:61:3: error: cannot delete expression of type 'Double_t [2]'
delete [] fPedADC;
^ ~~~~~~~
src/THcRaster.cxx:62:3: error: cannot delete expression of type 'Double_t [2]'
delete [] fAvgPedADC;
^ ~~~~~~~~~~
1 warning and 2 errors generated.
scons: *** [src/THcRaster.os] Error 1
scons: building terminated because of errors.

The error traces to the THcRaster destructor ... commenting out the two delete statements allows hcana to compile, but may cause runtime errors I suppose. I wonder why fPedADC and fAvgPedADC are defined at Double_t types in the first place ...

I also note that the warning appears to be a serious one ... fPosition is indeed defined to be an array of size 2, and so in principle only fPosition[0] and fPosition[1] are meaningful.

As I don't really understand at this point any of the raster code, I won't attempt a fix ... just raising this issue for Buddhini's benefit.

High priority

Valgrind reports and "invalid read of size 4" at this location (see below). Debugging reveals that the array index nscal is zero, causing the code to read scal_prev_read[-1]. In fact, in the test replay I am doing, run 4914 with the UTIL_KAONLT/scripts_Replay/replay_production_coin.C script, nscal is always zero at this point in the code.

Looks like a logic error.

As a temporary fix, I have changed that line to

if(nscal > 0 && scaldata < scal_prev_read[nscal-1]) {

but I doubt that that does what is intended here.

Partial output of

valgrind --suppressions=/apps/root/6.08.00/root/etc/valgrind-root.supp --leak-check=yes --track-origins=yes ./hcana -q -b .x UTIL_KAONLT/scripts_Replay/replay_production_coin.C(4914,10000)

544 errors in context 3 of 44:
==38116== Invalid read of size 4
==38116==    at 0x4FA1C8D: THcScalerEvtHandler::AnalyzeBuffer(unsigned int*, bool) (THcScalerEvtHandler.cxx:538)
==38116==    by 0x4F9F168: THcScalerEvtHandler::Analyze(THaEvData*) (THcScalerEvtHandler.cxx:234)
==38116==    by 0x53AE125: THaAnalyzer::MainAnalysis() (THaAnalyzer.cxx:1274)
==38116==    by 0x53AE984: THaAnalyzer::Process(THaRunBase*) (THaAnalyzer.cxx:1425)
==38116==  Address 0x2966cd9c is 4 bytes before a block of size 2,048 alloc'd
==38116==    at 0x4C2A1E3: operator new(unsigned long) (vg_replace_malloc.c:334)
==38116==    by 0x4FA84B1: __gnu_cxx::new_allocator<unsigned int>::allocate(unsigned long, void const*) (new_allocator.h:104)
==38116==    by 0x4FA7BCB: std::allocator_traits<std::allocator<unsigned int> >::allocate(std::allocator<unsigned int>&, unsigned long) (alloc_traits.h:360)
==38116==    by 0x4FA6A8F: std::_Vector_base<unsigned int, std::allocator<unsigned int> >::_M_allocate(unsigned long) (stl_vector.h:170)
==38116==    by 0x4FA6B8B: void std::vector<unsigned int, std::allocator<unsigned int> >::_M_emplace_back_aux<unsigned int>(unsigned int&&) (vector.tcc:412)
==38116==    by 0x4FA5C82: void std::vector<unsigned int, std::allocator<unsigned int> >::emplace_back<unsigned int>(unsigned int&&) (vector.tcc:101)
==38116==    by 0x4FA4EC3: std::vector<unsigned int, std::allocator<unsigned int> >::push_back(unsigned int&&) (stl_vector.h:932)
==38116==    by 0x4F9FF23: THcScalerEvtHandler::AnalyzeBuffer(unsigned int*, bool) (THcScalerEvtHandler.cxx:370)
==38116==    by 0x4F9F168: THcScalerEvtHandler::Analyze(THaEvData*) (THcScalerEvtHandler.cxx:234)
==38116==    by 0x53AE125: THaAnalyzer::MainAnalysis() (THaAnalyzer.cxx:1274)
==38116==    by 0x53AE984: THaAnalyzer::Process(THaRunBase*) (THaAnalyzer.cxx:1425)

Trying to build hcana according to instructions on the wiki. When doing

./podd/scons/scons.py debug=1 -j4

the build eventually errors out with

./libHallC.so: undefined reference to THcFormula::THcFormula(char const*, char const*, THaVarList const*, THaCutList const*)' ./libHallC.so: undefined reference totypeinfo for THcFormula'

(The same happens without any debug or -j parameters.)

Building with "make -j4" works.

Not sure if this is related to #23, which seems to be about shared library version numbers.

@sawjlab Please just delete the cmake build. It is not worth it. The analyzer should also just remove it but Ole blocked me from filing issues on the analyzer and has totally thrown away all the effort I (and others) put into the initial PR by rewriting a cmake build in the worst way. Please just forget I even brought it up because it only breaks things for me now.

Need to update podd to get updated code for 2D histograms DEFs.

Medium priority

CMake Error at CMakeLists.txt:11 (include):
  include could not find load file:

    PoddCMakeEnv


CMake Error at CMakeLists.txt:20 (find_package):
  No "FindROOT.cmake" found in CMAKE_MODULE_PATH.

FindROOT.cmake should not be installed by analyzer. ROOT comes fully equipped. This is literally undoing the point of using cmake.

With the changes to podd/site_scons/construct.py to use curl to download the evio library, scons
fails on cdaql1. When running curl , it can not connect to github.com .

Currently the only x and y positions for the drift chambers are global for full chamber. Each plane in a drift chamber should be able to shifted in x and y position for optimizing the alignment.

Various files for building podd clutter the top hcana directory when running scons from the top hcana directory. These files include:

haDict.*
NormAnaDict.*
THaDecDict.*
THaScalDict.*

Figure how to handle EPICS events. The Hall A analyzer does have some code for EPICS. Need to determine if this can be reused for the way Hall C puts EPICS events into the data stream and how the analysis uses the EPICS events.

For now it would be in the Monte Carlo

Valgrind reports that some operations in the above line refer to uninitialized memory and hence will likely produce garbage results. It might the matrix AA or the vector TT, which may not even be uninitialized per se, but be initialized with values which come from function calls that retrieve uninitialized data. It's not obvious, and I haven't been able to track this down yet. It is unlikely to be false positive though, based on my experience with valgrind.

Partial output of replaying 10000 events (note: 2.8 such errors per event):

valgrind --suppressions=/apps/root/6.08.00/root/etc/valgrind-root.supp --leak-check=yes --track-origins=yes ./hcana -q -b .x UTIL_KAONLT/scripts_Replay/replay_production_coin.C(4914,10000)

28240 errors in context 4 of 44:
==38116== Conditional jump or move depends on uninitialised value(s)
==38116==    at 0x6186055: TObject::operator=(TObject const&) (TObject.cxx:107)
==38116==    by 0x83E38BD: TVectorT<double>::operator=(TVectorT<double> const&) (TVectorT.cxx:686)
==38116==    by 0x83E398B: TVectorT<double>::TVectorT(TVectorT<double> const&) (TVectorT.cxx:217)
==38116==    by 0x83E3B2F: TVectorT<double> operator*<double>(TMatrixT<double> const&, TVectorT<double> const&) (TVectorT.cxx:1436)
==38116==    by 0x4F4E082: THcDC::TrackFit() (THcDC.cxx:1011)
==38116==    by 0x4F4BBDC: THcDC::CoarseTrack(TClonesArray&) (THcDC.cxx:604)
==38116==    by 0x540281D: THaSpectrometer::CoarseTrack() (THaSpectrometer.cxx:290)
==38116==    by 0x53AD225: THaAnalyzer::PhysicsAnalysis(int) (THaAnalyzer.cxx:1098)
==38116==    by 0x53AE190: THaAnalyzer::MainAnalysis() (THaAnalyzer.cxx:1281)
==38116==    by 0x53AE984: THaAnalyzer::Process(THaRunBase*) (THaAnalyzer.cxx:1425)
==38116==  Uninitialised value was created by a stack allocation
==38116==    at 0x4F4D853: THcDC::TrackFit() (THcDC.cxx:926)

When compiling on Mac OS X Yosemite, I get the following error:

g++ -o src/THcHodoHit.os -c -O -Wall -Woverloaded-virtual -pthread -pthread -m64 -stdlib=libc++ -fPIC -fPIC -DNDEBUG -DMACVERS -DHAS_SSTREAM -DHALLC_MODS -I/Users/brash/Dropbox/Research/analyzer-evio/evio-4.4.5/Darwin-x86_64/include -Isrc -Ipodd/src -Ipodd/hana_decode -Ipodd/hana_scaler -I/Users/brash/Dropbox/root/include src/THcHodoHit.cxx
src/THcHodoHit.cxx:18:40: error: addition of default argument on redeclaration makes this constructor a default constructor
THcHodoHit::THcHodoHit( THcRawHodoHit hit=NULL, Double_t posPed=0.0,
^ ~~~~
src/THcHodoHit.h:19:3: note: previous declaration is here
THcHodoHit( THcRawHodoHit hit, Double_t posPed,
^
1 error generated.
scons: *** [src/THcHodoHit.os] Error 1
scons: building terminated because of errors.

If I simply remove the "=NULL" definition on *hit, it compiles without issue. I am just not sure how important this might be ... comments?

Best,
E.

High priority.
Needed for computer deadtime.

Modify HCANA so that it uses the same drift chamber debug parameters as ENGINE and that the printout is the same.

Podd version should be updated.

This would allow me to change some code to avoid core dumps when trying to read non-existing parameter using the THcParmList::LoadParmValues.

When using any type of command-line flag for scons to build hcana, these flags are apparently not used for building the podd submodul: Example:

scons -j8 debug=1

compiles podd sources like

c++ -o src/THaADCHelicity.os -c -O -Wall -Woverloaded-virtual -Wextra -Wno-missing-field-initializers -Wno-unused-parameter -std=c++11 -m64 -pthread -rdynamic -fPIC -DNDEBUG -DLINUXVERS -DHAS_SSTREAM -I/home/ole/evio-4.4.6/Linux-x86_64/include -Isrc -Ihana_decode -I/opt/ROOT/6.10-04/include src/THaADCHelicity.C

but the hcana sources later like

g++ -o src/THcAerogel.os -c -g -O0 -Wall -Woverloaded-virtual -std=c++11 -m64 -pthread -rdynamic -fPIC -fPIC -DLINUXVERS -DHAS_SSTREAM -DHALLC_MODS -I/home/ole/evio-4.4.6/Linux-x86_64/include -Isrc -Ipodd/src -Ipodd/hana_decode -I/opt/ROOT/6.10-04/include src/THcAerogel.cxx

and the hcana build, but not the podd build, is very obviously parallelized.

The top SConscript.py file has a list, "hcheaders", for all the .h files in src. src/SConscript.py has a list, "list" of all the .cxx files in src. When a new class is added, both of these files must be updated. One of these lists should be automatically generated from the other.

Furthermore, it should be possible to avoid hand writing either of these lists. They could be generating from the file names of all the .cxx files in src.

Even more ideally, the file src/HallC_LinkDef.h which has a line for each class in it, could be generated automatically. If this is done, then Makefile should also be updated to autogenerate HallC_LinkDef.h and it's source code list "SRC".

The Cebaf Test Package, used in the ENGINE, as a report system. The report system allows one to define a template for an end of run analysis Printout/Report. Everything in the template file is simply output to an output file, except that what is inside braces ({}) is evaluated as an expression. The variables in the expression could be parameters, values from the hardware scalers, or "test scalers". (A test scaler is the number of times a given test was true/passed.)

A similar report system, ideally one that can use existing report templates, should be added to hcana. It should take advantage of existing infrastructure in podd as much as possible.

Below is my replay script. There is a segfault (see stacktrace below). I think this is likely a design flaw associated with the podd's use of static for the list of "post processes".
If std::cout isn't used, then there is no problem.

In addition to the class SimplePostProcess, the only other change from the source adopted from is adding the post process via

  auto* pp = new SimplePostProcess([](){ std::cout << "qwerty\n"; return 0;});
  analyzer->AddPostProcess(pp);

#include "THttpServer.h"
#include <functional>
#include <iostream>
#include <string>
 
class SimplePostProcess : public THaPostProcess {
public:
  using Function_t = std::function<int()>;
  Function_t flambda;
  SimplePostProcess(Function_t&& f) : flambda(std::forward<Function_t>(f)) { }
  //SimplePostProcess()  { }
  virtual ~SimplePostProcess(){ }

  virtual Int_t Init(const TDatime& ) {
    //std::cout << derp;
    return 0;
  }
  virtual Int_t Process( const THaEvData*, const THaRunBase*, Int_t code ){ 
    printf("asdf\n");
    flambda();
    return 0;
  }
  virtual Int_t Close(){ return 0; }
};

#if defined(__ROOTCLING__)
#pragma link C++ class SimplePostProcess+;
#pragma link C++ class std::vector<SimplePostProcess*>+;
#endif

void replay_production_all_shms (Int_t RunNumber = 0, Int_t MaxEvent = 0) {

  // Get RunNumber and MaxEvent if not provided.
  if(RunNumber == 0) {
    cout << "Enter a Run Number (-1 to exit): ";
    cin >> RunNumber;
    if( RunNumber<=0 ) return;
  }
  if(MaxEvent == 0) {
    cout << "\nNumber of Events to analyze: ";
    cin >> MaxEvent;
    if(MaxEvent == 0) {
      cerr << "...Invalid entry\n";
      exit;
    }
  }

  // Create file name patterns.
  const char* RunFileNamePattern = "shms_all_%05d.dat";
  vector<TString> pathList;
  pathList.push_back(".");
  pathList.push_back("./raw");
  pathList.push_back("./raw/../raw.copiedtotape");
  pathList.push_back("./cache");

  const char* ROOTFileNamePattern = "ROOTfiles/shms_replay_production_all_%d_%d.root";
  
  // Load global parameters
  gHcParms->Define("gen_run_number", "Run Number", RunNumber);
  gHcParms->AddString("g_ctp_database_filename", "DBASE/SHMS/standard.database");
  gHcParms->Load(gHcParms->GetString("g_ctp_database_filename"), RunNumber);
  gHcParms->Load(gHcParms->GetString("g_ctp_parm_filename"));
  gHcParms->Load(gHcParms->GetString("g_ctp_kinematics_filename"), RunNumber);
  // Load parameters for SHMS trigger configuration
  gHcParms->Load("PARAM/TRIG/tshms.param");
  // Load fadc debug parameters
  gHcParms->Load("PARAM/SHMS/GEN/p_fadc_debug.param");

  // Load the Hall C detector map
  gHcDetectorMap = new THcDetectorMap();
  gHcDetectorMap->Load("MAPS/SHMS/DETEC/STACK/shms_stack.map");

  // Add trigger apparatus
  THaApparatus* TRG = new THcTrigApp("T", "TRG");
  gHaApps->Add(TRG);
  // Add trigger detector to trigger apparatus
  THcTrigDet* shms = new THcTrigDet("shms", "SHMS Trigger Information");
  TRG->AddDetector(shms);

  // Set up the equipment to be analyzed
  THcHallCSpectrometer* SHMS = new THcHallCSpectrometer("P", "SHMS");
  gHaApps->Add(SHMS);
  // Add Noble Gas Cherenkov to SHMS apparatus
  THcCherenkov* ngcer = new THcCherenkov("ngcer", "Noble Gas Cherenkov");
  SHMS->AddDetector(ngcer);
  // Add drift chambers to SHMS apparatus
  THcDC* dc = new THcDC("dc", "Drift Chambers");
  SHMS->AddDetector(dc);
  // Add hodoscope to SHMS apparatus
  THcHodoscope* hod = new THcHodoscope("hod", "Hodoscope");
  SHMS->AddDetector(hod);
  // Add Heavy Gas Cherenkov to SHMS apparatus
  THcCherenkov* hgcer = new THcCherenkov("hgcer", "Heavy Gas Cherenkov");
  SHMS->AddDetector(hgcer);
  // Add Aerogel Cherenkov to SHMS apparatus
  THcAerogel* aero = new THcAerogel("aero", "Aerogel");
  SHMS->AddDetector(aero);
  // Add calorimeter to SHMS apparatus
  THcShower* cal = new THcShower("cal", "Calorimeter");
  SHMS->AddDetector(cal);

  // Add rastered beam apparatus
  THaApparatus* beam = new THcRasteredBeam("P.rb", "Rastered Beamline");
  gHaApps->Add(beam);
  // Add physics modules
  // Calculate reaction point
  THaReactionPoint* prp = new THaReactionPoint("P.react", "SHMS reaction point", "P", "P.rb");
  gHaPhysics->Add(prp);
  // Calculate extended target corrections
  THcExtTarCor* pext = new THcExtTarCor("P.extcor", "HMS extended target corrections", "P", "P.react");
  gHaPhysics->Add(pext);
  // Calculate golden track quantites
  THaGoldenTrack* gtr = new THaGoldenTrack("P.gtr", "SHMS Golden Track", "P");
  gHaPhysics->Add(gtr);
  // Calculate primary (scattered beam - usually electrons) kinematics
  THcPrimaryKine* kin = new THcPrimaryKine("P.kin", "SHMS Single Arm Kinematics", "P", "P.rb");
  gHaPhysics->Add(kin);
  // Calculate the hodoscope efficiencies
  THcHodoEff* peff = new THcHodoEff("phodeff", "SHMS hodo efficiency", "P.hod");
  gHaPhysics->Add(peff);

  // Add event handler for prestart event 125.
  THcConfigEvtHandler* ev125 = new THcConfigEvtHandler("HC", "Config Event type 125");
  gHaEvtHandlers->Add(ev125);
  // Add event handler for EPICS events
  THaEpicsEvtHandler* hcepics = new THaEpicsEvtHandler("epics", "HC EPICS event type 180");
  gHaEvtHandlers->Add(hcepics);
  // Add event handler for scaler events
  THcScalerEvtHandler* pscaler = new THcScalerEvtHandler("P", "Hall C scaler event type 1");
  pscaler->AddEvtType(1);
  pscaler->AddEvtType(129);
  pscaler->SetDelayedType(129);
  pscaler->SetUseFirstEvent(kTRUE);
  gHaEvtHandlers->Add(pscaler);
  // Add event handler for DAQ configuration event
  THcConfigEvtHandler *pconfig = new THcConfigEvtHandler("pconfig", "Hall C configuration event handler");
  gHaEvtHandlers->Add(pconfig);

  // Set up the analyzer - we use the standard one,
  // but this could be an experiment-specific one as well.
  // The Analyzer controls the reading of the data, executes
  // tests/cuts, loops over Acpparatus's and PhysicsModules,
  // and executes the output routines.
  THcAnalyzer* analyzer = new THcAnalyzer;

  auto* pp = new SimplePostProcess([](){ std::cout << "qwerty\n"; return 0;});
  analyzer->AddPostProcess(pp);

  // A simple event class to be output to the resulting tree.
  // Creating your own descendant of THaEvent is one way of
  // defining and controlling the output.
  THaEvent* event = new THaEvent;

  // Define the run(s) that we want to analyze.
  // We just set up one, but this could be many.
  THcRun* run = new THcRun( pathList, Form(RunFileNamePattern, RunNumber) );

  // Set to read in Hall C run database parameters
  run->SetRunParamClass("THcRunParameters");
  
  // Eventually need to learn to skip over, or properly analyze the pedestal events
  run->SetEventRange(1, MaxEvent); // Physics Event number, does not include scaler or control events.
  run->SetNscan(1);
  run->SetDataRequired(0x7);
  run->Print();

  // Define the analysis parameters
  TString ROOTFileName = Form(ROOTFileNamePattern, RunNumber, MaxEvent);
  analyzer->SetCountMode(2);  // 0 = counter is # of physics triggers
                              // 1 = counter is # of all decode reads
                              // 2 = counter is event number
  analyzer->SetEvent(event);
  // Set EPICS event type
  analyzer->SetEpicsEvtType(180);
  // Define crate map
  analyzer->SetCrateMapFileName("MAPS/db_cratemap.dat");
  // Define output ROOT file
  analyzer->SetOutFile(ROOTFileName.Data());
  // Define DEF-file
  analyzer->SetOdefFile("DEF-files/SHMS/PRODUCTION/pstackana_production_all.def");
  // Define cuts file
  analyzer->SetCutFile("DEF-files/SHMS/PRODUCTION/CUTS/pstackana_production_cuts.def");  // optional
  // File to record accounting information for cuts
  analyzer->SetSummaryFile(Form("REPORT_OUTPUT/SHMS/PRODUCTION/summary_all_production_%d_%d.report", RunNumber, MaxEvent));  // optional
  // Start the actual analysis.
  analyzer->Process(run);
  // Create report file from template
  analyzer->PrintReport("TEMPLATES/SHMS/PRODUCTION/pstackana_production.template",
  			Form("REPORT_OUTPUT/SHMS/PRODUCTION/replay_shms_all_production_%d_%d.report", RunNumber, MaxEvent));  // optional

}

Starting analysis
asdf

 *** Break *** segmentation violation



===========================================================
There was a crash.
This is the entire stack trace of all threads:
===========================================================
#0  0x00007ff9142b68fa in __GI___waitpid (pid=1286, stat_loc=stat_loc
entry=0x7ffcd4073358, options=options
entry=0) at ../sysdeps/unix/sysv/linux/waitpid.c:29
#1  0x00007ff9142259cb in do_system (line=<optimized out>) at ../sysdeps/posix/system.c:148
#2  0x00007ff919552173 in TUnixSystem::Exec (shellcmd=<optimized out>, this=0x1a297c0) at /opt/software/sources/root/core/unix/src/TUnixSystem.cxx:2119
#3  TUnixSystem::StackTrace() () at /opt/software/sources/root/core/unix/src/TUnixSystem.cxx:2413
#4  0x00007ff919554a44 in TUnixSystem::DispatchSignals (this=0x1a297c0, sig=kSigSegmentationViolation) at /opt/software/sources/root/core/unix/src/TUnixSystem.cxx:3644
#5  <signal handler called>
#6  0x00007ff914c38616 in std::ostream::sentry::sentry (this=0x7ffcd4075ad0, __os=...) at /opt/software/builds/gcc_8.1.0_build/x86_64-pc-linux-gnu/libstdc++-v3/include/bits/ostream.tcc:46
#7  0x00007ff914c38c28 in std::__ostream_insert<char, std::char_traits<char> > (__out=..., __s=__s
entry=0x7ff919eb2d7e "qwerty\n", __n=7) at /opt/software/builds/gcc_8.1.0_build/x86_64-pc-linux-gnu/libstdc++-v3/include/bits/ostream_insert.h:76
#8  0x00007ff914c39037 in std::operator<< <std::char_traits<char> > (__out=..., __s=0x7ff919eb2d7e "qwerty\n") at /opt/software/builds/gcc_8.1.0_build/x86_64-pc-linux-gnu/libstdc++-v3/include/bits/char_traits.h:320
#9  0x00007ff919eb7a7c in ?? ()
#10 0x0000000006d8e238 in ?? ()
#11 0x0000000006d8e238 in ?? ()
#12 0x00007ffcd4075b70 in ?? ()
#13 0x00007ff919eb78e3 in ?? ()
#14 0x00007ff919eb7a50 in ?? ()
#15 0x0000000006d8e238 in ?? ()
#16 0x00007ffcd4075b90 in ?? ()
#17 0x00007ff919eb7f15 in ?? ()
#18 0x0000000006d8e238 in ?? ()
#19 0x0000000006d8e238 in ?? ()
#20 0x00007ffcd4075bd0 in ?? ()
#21 0x00007ff919eb7d33 in ?? ()
#22 0x0000000500000000 in ?? ()
#23 0x0000000006d8e220 in ?? ()
#24 0x00000000d4075bd0 in ?? ()
#25 0x000000000728bb00 in ?? ()
#26 0x0000000006dfea10 in ?? ()
#27 0x0000000006d8e220 in ?? ()
#28 0x00007ffcd4075c20 in ?? ()
#29 0x00007ff915b680b0 in THaAnalyzer::PostProcess(int) () from /home/whit/lib/libHallA.so
===========================================================

$MYDIR is not being defined if "source setup.csh" is called from a script (in interactive mode there is no problem).

You can reproduce the issue by running a shell script such a:

#!/bin/csh -vx
# _FILE_ SetHcana.csh
set HCANADIR="/path/to/hcana/"
source "$HCANADIR/setup.csh"
hcana -l -b -q SomeScript.C

Then,

./SetHcana.csh

Which will display:

/path/to/hcana/: error while loading shared
libraries: libHallC.so: cannot open shared object file: No such file or
directory.

Analyze scalers.
Put scaler totals somewhere that reports (THcAnalyzer::PrintReport) can use them to calculate rates, etc.

I am getting a bunch of warnings of the type:

HitList(event=350): refindex 4 (3, 2, 127) missing for (3, 2, 12)
HitList(event=350): refindex 4 (3, 2, 127) missing for (3, 2, 13)
HitList(event=350): refindex 4 (3, 2, 127) missing for (3, 2, 17)
HitList(event=350): refindex 4 (3, 2, 127) missing for (3, 2, 24)
HitList(event=350): refindex 4 (3, 2, 127) missing for (3, 2, 23)
HitList(event=350): refindex 4 (3, 2, 127) missing for (3, 2, 26)
HitList(event=350): refindex 4 (3, 2, 127) missing for (3, 2, 16)
HitList(event=350): refindex 4 (3, 2, 127) missing for (3, 2, 27)
HitList(event=350): refindex 4 (3, 2, 127) missing for (3, 2, 18)
HitList(event=350): refindex 4 (3, 2, 127) missing for (3, 2, 19)
HitList(event=350): refindex 4 (3, 2, 127) missing for (3, 2, 70)
HitList(event=350): refindex 4 (3, 2, 127) missing for (3, 2, 73)

I chased this down as being part of the "present" mechanism for some detector variables. Is anybody else seeing this?

I started looking into using cppcheck to analyze source code. I started with the hcana/src directory, but of course this should be applied to the entire Hall A analyzer as well, in principle.

Cppcheck produced a lot of suggestions for improvement, which break down into several categories:

a) warnings about unused variables, etc.
b) potential performance hits
c) style issues

I took some time to fix a few of the things that showed up from categories a) and b) ... that was pretty simple. However, there remain to broad types of things that cppcheck complains about:

1. C-style pointer casting
2. Failure to initialize member variables in the constructor of a class.

Below is the output of cppcheck from just one of the pieces of source code in hcana/src:

[src/THcAerogel.cxx:263]: (style) C-style pointer casting
[src/THcAerogel.cxx:267]: (style) C-style pointer casting
[src/THcAerogel.cxx:273]: (style) C-style pointer casting
[src/THcAerogel.cxx:279]: (style) C-style pointer casting
[src/THcAerogel.cxx:285]: (style) C-style pointer casting
[src/THcAerogel.cxx:305]: (style) C-style pointer casting
[src/THcAerogel.cxx:454]: (style) C-style pointer casting
[src/THcAerogel.cxx:52]: (warning) Member variable 'THcAerogel::fPosGain' is not initialized in the constructor.
[src/THcAerogel.cxx:52]: (warning) Member variable 'THcAerogel::fNegGain' is not initialized in the constructor.
[src/THcAerogel.cxx:52]: (warning) Member variable 'THcAerogel::fA_Pos' is not initialized in the constructor.
[src/THcAerogel.cxx:52]: (warning) Member variable 'THcAerogel::fA_Neg' is not initialized in the constructor.
... and so on.

Looking at the first few entries, the code looks like this:

.
.
while(ihit < fNhits) {
THcAerogelHit* hit = (THcAerogelHit *) fRawHitList->At(ihit);
.
.

So, you can see the type of thing that it is complaining about. The question is: what do we do about this? One option, as I understand it, is to replace these casts with one of the C++-style casts (static_cast<...>, const_cast<...>, etc.). A second (better?) option is to use STL vectors. A third option is to just ignore this and proceed ahead.

I have to admit I don't understand at this point if the uninitialized member variable warnings are so-called "false positives" of cppcheck, or if there is some non-standard procedure that is being followed in our code.