jeffersonlab / jana

JANA

JANA is the JLab ANAlysis framework. It is a multi-threaded C++ framework for processing event-based data from particle physics experiments. It has been primarily used by the GlueX experiment at Jefferson Lab.

The JANA project is being replaced by JANA2 which is currently under development.

jana's Issues

Compiler 'misleadingly indented' warning

There are several identical warnings appear during the compilation:

src/JANA/JApplication.cc: In member function ‘jerror_t jana::JApplication::AttachPlugins()’:
src/JANA/JApplication.cc:2394:4: warning: this ‘if’ clause does not guard... [-Wmisleading-indentation]
    if(printPaths) jout<<"Looking for \""<<fullpath<<"\" ...."; cout.flush();
    ^~
src/JANA/JApplication.cc:2394:64: note: ...this statement, but the latter is misleadingly indented as if it were guarded by the ‘if’
    if(printPaths) jout<<"Looking for \""<<fullpath<<"\" ...."; cout.flush();
                                                                ^~~~
src/JANA/JApplication.cc:2408:4: warning: this ‘if’ clause does not guard... [-Wmisleading-indentation]
    if(printPaths) jout<<"Looking for \""<<fullpath<<"\" ...."; cout.flush();
    ^~
src/JANA/JApplication.cc:2408:64: note: ...this statement, but the latter is misleadingly indented as if it were guarded by the ‘if’
    if(printPaths) jout<<"Looking for \""<<fullpath<<"\" ...."; cout.flush();
                                                                ^~~~

Reviewing them I would rather agree with the compiler. So PR is coming

Several files always rebuilding

When trying to rebuild JANA for several times I figured out that several files are being rebuilt each time. The reason is generated jana_config.h

scons: rebuilding `/home/romanov/jleic/test/jana/build/jana-drateots-master/JANA/JGeometryMYSQL.o' because:
           `src/JANA/JGeometryMYSQL.cc' is no longer a dependency
           `src/JANA/JGeometryMYSQL.cc' is a new dependency
           `/home/romanov/jleic/test/jana/jana-drateots-master/include/JANA/jana_config.h' changed

Scons uses MD5 to track file changes. And the file changes each time because it has generation time written in it:

str += '// Generation date: %s\n' % datetime.datetime.now().strftime("%I:%M%p on %B %d, %Y")

Removing the line would allow not to rebuild everything when nothing is really changed.

The question is. Where there cases, when this line really helped to debug or fix something? Otherwise I would propose to remove it.

Make installation dir changeable

When scons install is called, JANA is installed to ```#bsosname" . This line

Later build scripts move the binary directory to their preferred places and very convenient setenv.* script becomes useless

I would propose to add PREFIX scons argument. If the PREFIX is set, installdir should be set to it. Can make a PR

JANA_PLUGIN_PATH in autogenerated setenv scripts

Since both Jana and JLEIC_ANA (or whoever) tries to set JANA_PLUGIN_PATH, I would suggest changing auto-generated scripts to something like:

# Make sure JANA_PLUGIN_PATH is set
if [ -z "$JANA_PLUGIN_PATH" ]; then
   export JANA_PLUGIN_PATH=""
fi

# JANA
export JANA_PLUGIN_PATH=$JANA_PLUGIN_PATH:${JANA_HOME}/plugins