eliminate global variables

[mikejiang]

There are currently two global variables used by cytolib (mainly for troubleshooting purpose) https://github.com/RGLab/cytolib/blob/trunk/inst/include/cytolib/global.hpp#L30-L31

extern keyword indicates they are only the variable declaration instead of definition. Because due to the nature of header-only, cytolib can't define them in any header in order to avoid multi-definition compiling errors (when header is included multiple times).

Therefore it implicitly requires the cytolib users to define it somewhere in the user code in order for cytolib function, even if they don't necessarily need to care about this log feature.

We shouldn't force user to do this (besides it is not best practice to use global variable). Instead, we can move them into GatingSet class and initialize/change their values either through parseWorkspace call or through setter methods once GatingSet is created.

[mikejiang]

It turns out to be a lot of hassles to initialize and pass around the local version of this loglevel flag due to its universal usage among every classes defined in cytolib. We will need to figure out more elegant way to handle this (maybe the third-party logging lib).

[gfinak]

Maybe a naive question, but wouldn't a conditional definition using the preprocessor work?

[mikejiang]

I am not sure if that will allow user to switch on/off the log at runtime

[mikejiang]

For now, I defined a macro as CYTOLIB_INIT() to make it more user friendly, see https://github.com/RGLab/flowWorkspace/blob/trunk/src/R_GatingSet.cpp#L13

[mikejiang]

@gfinak , #ifdef only prevents the same source file(i.e. single translation unit) from defining it multiple times, but different source files can still include the same header independently, which will pass the compilation, and eventually causing linking error complaining the conflicting definitions among these translation units.

[gfinak]

Yes, makes sense. Thanks for the clarification.