Table_Value fails when X is in decreasing order

[nicriz]

Hi @willend,

I was trying to debug a weird behavior of Source_gen4 component with a flux_file, when I found out that the root of the problem was in Table_value returning the wrong value for the flux from the table. The file Lambda/Brilliance I was using had a constant step of 0.1 AA and the values of Lambda in decreasing order from 15AA to 0.2 AA. Now this is what was happening:

1. Source_gen4 extracts a lambda between Lmin and Lmax, let's say 1.9AA
2. When Source_gen4 calls Table_value to fix the weight, the constant-time lookup is satisfied

   McCode/common/lib/share/read_table-lib.c

   Lines 837 to 842 in 7f77bfb

   	if(Table.constantstep) {
   	Index = (long)floor(
   	(X - Table.min_x) / (Table.max_x - Table.min_x) * (Table.rows-1));
   	X1 = Table_Index(Table,Index-1,0);
   	X2 = Table_Index(Table,Index ,0);
   	}

   so Index = 17. But this value of Index is correct only for monotonically increasing bins!
3. This ultimately breaks when the code checks for the linear search

   McCode/common/lib/share/read_table-lib.c

   Lines 862 to 870 in 7f77bfb

   	// Fall back to linear search, if no-one else has set X1, X2 correctly
   	if (!((X1 <= X) && (X < X2))) {
   	/* look for index surrounding X in the table -> Index */
   	for (Index=1; Index <= Table.rows-1; Index++) {
   	X1 = Table_Index(Table, Index-1,0);
   	X2 = Table_Index(Table, Index ,0);
   	if ((X1 <= X) && (X < X2)) break;
   	} /* end for Index */
   	}

   which returns always the last bin because the break condition is never fulfilled (X2 is always smaller than X1)

So it seems to me that the function requires monotonically increasing X bins to work, but this is neither enforced anywhere nor documented (as far as I could search). So maybe a flag could be added in 'Table_value' or 'Table_Stat' ?

I could help with that in case,
Best

Nico