Linear Collider Forum: Tracking & Vertexing

Re: Pulse (and Hit) quailty [message #1967 is a reply to message #1966]

Thu, 06 May 2010 09:22

killenberg
Messages: 125
Registered: July 2005
Location: CERN

Hello Christoph,

thanks for your reply. I completely agree that the definition should be done externally. But we should also have it implemented as constants because the source code is much clearer if it reads

if (quality & OVERFLOW) {...}

instead of

if (quality & (1<<4) ) {...}

Quote:

An older, first idea (I'm behind the updating schedule) is at http://www-flc.desy.de/flc/flcwiki/MarlinTPCStructure

Yes, that's the way I would have implemented it. I just proposed the "eye candy" thing because the 1, 10, 100 looked like it.

Quote:

Please don't do the definition of numbers in header files; rather use the initialisation list.

Yes, you are right. However, for static const I sometimes make an exception because

They are not real variables. I think the compiler "hardcodes" them during the optimisation anyway and does not even allocate memory for them.
You have the variable names and the values in the same place.
Doxygen includes the values.
You don't need a .cpp and object file.

I usually interpret them as the "C++-version" of #define

Cheers

Martin

Martin Killenberg

CERN
martin.killenberg@cern.ch

Report message to a moderator