I think Alv\'s suggestion of implementing the BrDetectorList class is the thing to do. I had discussed with Christian when the BrDetectorList was coming into being the possibility of doing just that, but as with many things it got away. I think if someone is going to go to the trouble of making a change, it ought to be done the right way, ie with BrDetectorList. That will put BrRawDataInput on a consistent footing with other objects that use it. Of course if it breaks applications, they need to be fixed before this is reintroduced into the repository. Kris Quoting Alv Kjetil Holme <a.k.holme@fys.uio.no>: > On Tue, 25 Jul 2000, Djamel Ouerdane wrote: > > > Hi Alv, > > > > I\'m a bit suprised by what you described but I understand your point: it > > is due to the fact that the SetxxxOff methods are based on the SetxxxOn > > methods. So I can see 2 ways to remedy to that: > > > > 1- one way is to decouple these methods and have the set of SetxxxOff > > completly independant from the SetxxxOn > > > > 2- the second is to keep fAllOn, fFSOn or fMRSOn kTRUE (instead of the > > passed value) when you set off a > > particular detector, that\'s to say: > > > > for instance MTP1: > > void SetMTP1On(const Bool_t val = kTRUE) { fMTP1On = val; fMRSOn = > > kTRUE; fAllOn = kTRUE;} instead of val all the time. > > > > Does it sound reasonable? > > Djamel > > There is one more possibility, to rewrite the code and take advantage of > the BrDetectorList class. For fun, I implemented the following changes: > > In BrRawDataInput.h: > > a) Add include statement > #ifndef BRAT_BrDetectorList_h > #include \"BrDetectorList.h\" > #endif > > b) Replace all the members > Bool_t fXXXOn > by one member > BrDetectorList fDetectors > > c) Introduce methods to set, clear and test the state > > void SetOn(EBrDetectorBit bit) { fDetectors.SetDetectorOn(bit); } > void SetOff(EBrDetectorBit bit) { fDetectors.SetDetectorOff(bit); } > > void SetOn(EBrSectorBit mask=kBrAll) > { fDetectors.AddToDetectorMask(mask); } > void SetOff(EBrSectorBit mask=kBrAll) > { fDetectors.StripDetectorMask(mask); } > > Bool_t IsOn(EBrDetectorBit bit) const > { return fDetectors.IsOn(bit); } > Bool_t IsOn(EBrSectorBit mask) const > { return (fDetectors.GetDetectorMask() & mask) != 0; } > > Since fDetector is a private member, there is no need for these to > be virtual methods. > > d) Remove the old SetXXXOn() and SetXXXOff() methods. This is OK with > BRAT. However, it could break applications. In that case one could > reimplement the old SetXXXOn() using the new methods, but I think it > would be better to change the applications to use the new methods. > > void SetAllOn(Bool_t val=kTRUE) > { if (val) SetOn(kBrAll); else SetOff(kBrAll); } > > void SetC1On(Bool_t val=kTRUE) > { if (val) SetOn(kBrC1); else SetOff(kBrC1); } > > void SetH1On(Bool_t val=kTRUE) > { if (val) SetOn(kBrTOF1); else SetOff(kBrTOF1); } > > In BrRawDataInput.cxx: > > a) Remove the three SetXXXOn() in Init() > > b) Replace the tests > if (fXXXOn) > by > if (IsOn(kBrXXX)) > > e.g. > fAllOn -> IsOn(kBrAll) > : > fC1On -> IsOn(kBrC1) > fH1On -> IsOn(kBrTOF1) > : > fMultOn -> IsOn(kBrTile) OR IsOn(kBrSi) > > > Note that BrDetectorList used different names for hodoscopes and MRS TPCs > Old name Name in BrDetectorList > H1 TOF1 > H2 TOF2 > MTP1 TPM1 > MTP2 TPM2 > > I have used these modifications in a test version of tpc3d and it works > well. > > -Alv Kjetil >
This archive was generated by hypermail 2b29 : Tue Jul 25 2000 - 17:09:24 EDT