Fix bug in indexing for EB-inflow hyperbolics

[baperry2]

• Move definition of indices used in MOL hyp flux calculations to be defined as macros in IndexDefines.H. otherwise these must be redefined 3 times, including in a user-specified function, which likely to introduce bugs
• In pc_compute_hyp_mol_flux, instead of defining spr and spl as new arrays that must be allocated and copied into, just point to the appropriate places in qtempl and qtempr. This cuts down on code and should be a bit more efficient.
• In pc_compute_hyp_mol_flux_eb do the same thing. Note that there is currently a bug where the qtemp arrays aren't sized to store ADV/AUX/LIN quantities, but the indexing used assumes does account for these quantities. This fortunately seems to not have caused issues, but is now fixed.
• Note that in the function ProblemSpecificFunctions::problem_eb_state we pass both qtemp{l,r} and sp{l,r} as arguments. This is redundant, but shouldn't hurt anything as long as users understand this.

@mjrahimi - can you verify that this updated version of the code works for the oxy-combustor case?

[marchdf]

Good work!

[mjrahimi]

The solver cannot be compiled. I get this error:
prob.H(532): error: expected an identifier const int 0 = 0; ^