LArSoft

Matt,

I notice in the dune/DataOverlay/DataOverlay/CMakeLists.txt there are a lot of library names that have ub in them, e.g.
LIBRARY_NAME ub_RawDigitAdderAna

Can you rename them to dune35_RawDigitAdderAna etc.? Also please make sure all the LIBRARY_NAMEs are different from the ones in uboonecode.

Tingjun

Hi all,

I've just pushed changes to dunetpc where the library names should all be different from the original ones in uboonecode. Could someone please test and make sure the cross-compilation issues don't still exist?

Matt

In preparation for this weeks' release, I have made the standard multiexperiment build. The build appears to be fine, but 76 of 202 tests fail!

[garren@woof geometry.d]$ lar -c ./test_geometry.fcl 
%MSG-i MF_INIT_OK:  lar 02-Nov-2016 11:48:52 CDT JobSetup
Messagelogger initialization complete.
%MSG
terminate called after throwing an instance of 'cet::coded_exception<art::errors::ErrorCodes, &art::ExceptionDetail::translate>'
  what():  ---- FatalRootError BEGIN
  Fatal Root Error: @SUB=TInterpreter::ReadRootmapFile
  class  art::Wrapper<std::vector<mix::EventMixingSummary> > found in libdune_DataOverlay_DataOverlayProducts_dict.so  is already in libuboone_DataOverlay_DataOverlayProducts_dict.so 
---- FatalRootError END

Aborted

Will a new namespace / product name solve the problem? (temporarily, of course, until someone makes a general LArSoft implementation of DataOverlay (again, not me))

All names need to be unique, so hopefully that would solve the problem.

Try pulling dunetpc and testing again.

I renamed the data product from "mix::EventMixingSummary" to "dunemix::EventMixingSummary". If that's not enough, I can rename EventMixingSummary to something else.

Thank you. That seems to have done the trick. All unit tests now pass.

I think Wes may be the original author of this code.