xAH with CMake

Setup

To set up, create a directory to work in:

mkdir rel21testing && cd $_

then clone me with the submodules fetched as well:

git clone --recursive <github url>

Next, create a build directory to build from CMake, set up the release, then try making:

mkdir build && cd $_
setupATLAS
asetup AnalysisBase,2.6.1
cmake ../xAODAnaHelpersCMake
make install DESTDIR=../install

After installing, my top-level looks like this:

$ ls
total 28K
drwxr-xr-x  5 kratsg  117 Apr 19 17:24 ./
drwxr--r-- 53 kratsg 4.0K Apr 19 17:17 ../
-rw-r--r--  1 kratsg  13K Apr 19 10:56 .asetup.save
drwxr-xr-x  6 kratsg 4.0K Apr 19 16:18 build/
drwxr-xr-x  3 kratsg   32 Apr 19 16:12 install/
drwxr-xr-x  4 kratsg  144 Apr 19 17:24 xAODAnaHelpersCMake/

where everything was installed to install. Now, to run code, I updated xAH_run.py to check if we are on 2.5/2.6 release series (the ones that use CMake) and requires an extra option: -L/--library-linking which needs a path to the installed shared library file you just made! So creating an example config.py for testing purposes, I run the code like this:

./install/InstallArea/x86_64-slc6-gcc49-opt/bin/xAH_run.py --config config.py \
--files input.root \
-L install/InstallArea/x86_64-slc6-gcc49-opt/lib/libxAODAnaHelpers.so \
direct

Notice that the path to the xAH_run.py needs to be specified. This isn't automatically in your $PATH yet, but I'm sure given time, a solution to make this happen in a cleaner way will come up. For now, just know that install/InstallArea/x86_64-slc6-gcc49-opt contains everything you've made, including bin/, lib/, data/, and docs/. The last thing is that I gave xAH_run.py the full path to where it can find the shared library, then inside xAH_run.py, it executes the following calls:

import ROOT
ROOT.gSystem.Load(args.ext_library)
xAH_logger.info("loading packages")
ROOT.gROOT.Macro("$ROOTCOREDIR/scripts/load_packages.C")

where the shared library that was made needs to be loaded in first before you can load in the RootCore macro here.