casm enum in 2D case

[sjtuzhanglei]

When I do casm enum on only a-b plane, it could generates the weird transformation matrix. Is it a bug of the program? Should I avoid transformation matrix like this:

This puts my c lattice vector upside down:
"SCEL3_3_1_1_0_0_0" : [
[ 0, 3, 0 ],
[ 1, 0, 0 ],
[ 0, 0, -1 ]

This swaps my c lattice vector to the new b axis:
"SCEL5_1_5_1_0_0_0" : [
[ 1, 0, 0 ],
[ 0, 0, 5 ],
[ 0, -1, 0 ]

Also, when I do casm init, it suggests me to use a prim.niggli.json file, but I couldn't see a difference between this and my original prim.json file.

[bpuchala]

casm enum always saves supercells in a CASM canonical form because it is primarily intended to only keep symmetrically unique configurations, although non-primitive configurations may also be kept as an option.

I could inspect the prim init suggestion if you're willing to share by email. CASM will suggest a Niggli cell and a particular orientation based on the same lattice canonical form method linked above so that the volume 1 supercell is simply the prim lattice x identity matrix.