🚀 go-pugleaf

Hello,

      GDB seems to report erroneous type information when my c++ code is being
debugged.  I do NOT believe this a name mangling problem.  I confess however
that I am working on the untested assumption that name mangling is merely a
superficial problem.  When I get to the arrow below

                       DivyNextLowerE(double ll, double ul)         
		          {                                         
		          double TotalXsecAtE=0;                    
		          double GivenRange;                        
		          double dtmp[nStates];                     
		                                                    
		    =>    GivenRange=ul-ll;                         
		          for (int ii=0; ii<nStates; ii++)          
		             {                                      
		             *GasSpecRB[ii] >> dtmp[ii];     
		             TotalXsecAtE+= dtmp[ii];               
		                                                    
		    // and so on                                    

and try to get information on dtmp I am amazed to see ...

                   (gdb) print dtmp
                   $1 = (double (*)[]) 0xeffec280

Could this truly mean that the debugger sees dtmp as the base address of an
array of pointers to type double.

Info:
   -The name dtmp is not declared anywhere else.
   -I am using g++(gcc vs. 2.6.0),  Sun ld release 4.1, and GDB 4.13
    (sparc-sun-sunos4.1.3_U1).
   -Yes, I get name mangling.
   -Possibly unrelated (still testing): when crash finally comes the following
    message is received this is the message,
		   
		   Program received signal SIGBUS, Bus error.
		   0x5a74 in GasSpecies::DivyNextLowerE (this=0xefffa430, ll=0,
		       ul=0.99826065384036533) at gas.h:259
		   Source file is more recent than executable.
		   (gdb)               


Wild guesses:
   -Could my mix of compiler, linker and debugger software be the cause of my
    problem?
   -Could programming errors elsewhere cause GDB type information to be
    changed?
   -I am still too much a newbie and missing the obvious?

                                               
                                                  Thanks,

                                                  Phillip MacLean
	                                          <pgmst@pitt.edu>