Is this the expected behaviour
This may not really be a bug - but i would like to know if it is intended behaviour.
I am porting the Mu2e code to use art v2_02_02. In my first test building data products I created a test library that contains only a single data product, which is a class that contains one data member, an int.
When I ran genflex on the classes.h file I discovered that I need to link to CLHEP or else the link will fail.
The dependence on CLHEP comes from:
I have not traced how this is being included in my test. The only art/canvas headers that I include myself are the following in the classes.h file:
The genreflex command is:
genreflex Sandbox/src/classes.h \ -s Sandbox/src/classes_def.xml \ -I. \ -Iinclude \ -I/home/kutschke/products/art/v2_02_02a/include \ -I/home/kutschke/products/canvas/v1_04_05a/include \ -I/home/kutschke/products/messagefacility/v1_17_01/include \ -I/home/kutschke/products/fhiclcpp/v4_00_01/include \ -I/home/kutschke/products/sqlite/v3_12_02_00/Linux64bit+2.6-2.12/include \ -I/home/kutschke/products/cetlib/v1_19_01/include \ -I/home/kutschke/products/boost/v1_61_0/Linux64bit+2.6-2.12-e10-prof/include \ -I/home/kutschke/products/clhep/v2_2_0_8a/Linux64bit+2.6-2.12-e10-prof/include \ -I/home/kutschke/products/cppunit/v1_12_1d/include \ -I/home/kutschke/products/root/v6_06_04b/Linux64bit+2.6-2.12-e10-prof/include \ -I/home/kutschke/products/xerces_c/v3_1_3/Linux64bit+2.6-2.12-e10-prof/include \ -I/home/kutschke/products/tbb/v4_4_3a/Linux64bit+2.6-2.12-e10-prof/include \ -l lib/libmu2e_Sandbox_dict.so \ -o tmp/src/Sandbox/src/mu2e_Sandbox_dict.cpp \ --fail_on_warnings -DNDEBUG \ --rootmap-lib=lib/libmu2e_Sandbox_dict.so \ --rootmap=lib/libmu2e_Sandbox_dict.rootmap
#1 Updated by Kyle Knoepfel over 3 years ago
- Description updated (diff)
- Status changed from New to Feedback
aggregate.h dependencies have changed with art 2.03.00, which Lynn released just last night. Once a
mu build is available, can you run
genreflex again and verify if the dependency on
CLHEP is still required?