full function for various boundary conditions

[shivin9]

The full(A) functionality to retrieve the matrix of transpositions needs to be updated to correctly accommodate the various boundary conditions eg. Dirichlet and Neumann.
For Dirichlet with a specific boundary condition it is working as follows:-

julia> A = LinearOperator{Float64}(2,2,1.0,10,:D1,:D1;bndry_fn=(u[1],u[end]));

julia> full(A)
10×10 Array{Float64,2}:
  1.84135   2.84135     1.84135    …   1.84135     1.84135   1.84135
 -1.84135  -3.84135    -0.841347      -1.84135    -1.84135  -1.84135
  0.0       1.0        -2.0            0.0         0.0       0.0    
  0.0       0.0         1.0            0.0         0.0       0.0    
  0.0       0.0         0.0            0.0         0.0       0.0    
  0.0       0.0         0.0        …   0.0         0.0       0.0    
  0.0       0.0         0.0            1.0         0.0       0.0    
  0.0       0.0         0.0           -2.0         1.0       0.0    
  0.0      -0.0136593  -0.0136593      0.986341   -2.01366   1.0    
  0.0       0.0136593   0.0136593      0.0136593   1.01366  -1.0    

with the matrix multiplication not getting the right values at the boundaries.

julia> res = A*u
401-element Array{Float64,1}:
  0.0392531  
 -0.00115466 
 -0.00111008 
 -0.00106763 
 -0.0010272  
 -0.000988661
 -0.000951915
 -0.000916862
 -0.00088341 
 -0.00085147 
  ⋮          
 -1.54722e-6 
 -1.53094e-6 
 -1.51485e-6 
 -1.49893e-6 
 -1.48318e-6 
 -1.46761e-6 
 -1.45221e-6 
 -1.43698e-6 
 -0.00013916 

julia> full(A)*u
401-element Array{Float64,1}:
 -215.342      
  215.38       
   -0.00111008 
   -0.00106763 
   -0.0010272  
   -0.000988661
   -0.000951915
   -0.000916862
   -0.00088341 
   -0.00085147 
    ⋮          
   -1.54722e-6 
   -1.53094e-6 
   -1.51485e-6 
   -1.49893e-6 
   -1.48318e-6 
   -1.46761e-6 
   -1.45221e-6 
    1.55873    
   -1.55887    

julia> res[3:398]≈(full(A)*u)[3:398]
true

julia> res≈(full(A)*u)
false

[ChrisRackauckas]

sparse and BandedMatrix will also have problems. It should just error when trying to create a banded matrix from periodic.