On Tue, Feb 16, 2010 at 19:13 -0600, Leutenegger, Maurice A. (GSFC-662.0)[OAK RIDGE ASSOCIATED UNIVERSITIES (ORAU)] wrote: > How about this?: > > nm /path/to/local/models/liblocalmodels.dylib > I still have the svn repository of your models from the last time we discussed this, so I did an 'svn update' and built your models on a local Mac. I then reproduced the problem you're seeing. Is the svn repository version the same as what you're using? Looking again at the list of symbols in the shared library, it seems to me that there's an inconsistency. Your lmodel.dat file associates a symbol of the form C_<model_name> with every model, but no symbols of that form are present in your shared library. Focusing on the windprof model, your shared library contains these symbols: 00010d00 t __GLOBAL__I_windprof.cpp 00010f80 T _windprof 00010e10 T _windprofisis but the lmodel.dat file says isis should call C_windprof, which is not present. Isis will look for a few variations of that name, appending or prepending an underscore, variations in case, etc. but the symbol it needs doesn't seem to appear in the shared library. For comparison, I picked a built-in xspec function that also has a C++ interface (xsbexriv.cxx). Looking at bexriv (which works in isis) I see the symbols: 0000a550 T _C_xsbexriv 0000f490 t __GLOBAL__I_xsbexriv.cxx 00009310 T __Z10f_xsbexrivPKfiS0_iPf 000b8015 d __ZZ8xsbexrivE7isFirst 0000fda0 T _xsbexriv 00009370 T _xsbexriv_ The model.dat file says isis should look for C_xsbexriv, and that name is actually present in the shared library. Could this be the problem? Thanks, -John ---- You received this message because you are subscribed to the isis-users list. To unsubscribe, send a message to isis-users-request_at_email.domain.hiddenwith the first line of the message as: unsubscribeReceived on Wed Feb 17 2010 - 08:14:46 EST
This archive was generated by hypermail 2.3.0 : Fri May 02 2014 - 08:35:46 EDT