ok    1  product mix: maximize 3x + 5y -> (2, 6), value 36, exact
ok    2  minimize with a >= row (phase 1): (3, 1), cost 9, exact
ok    3  x + y <= 1 and x + y >= 3: INFEASIBLE
ok    4  max x + y with only x - y <= 1: UNBOUNDED, the detail says arbitrarily large
ok    5  a fractional vertex: (5/4, 5/4), value 5/2, exact
ok    6  a degenerate vertex: (2, 0), value 4
ok    7  the bakery (two-decimal data): 100 croissants, 120 muffins, 580.00 exact
ok    8  three variables: (2, 0, 1), value 13
ok    9  minimize -x under x + y <= 3, x - y <= 1: x = 2
ok    10  exact: false on the product mix: (2, 6) within 1e-12
ok    11  exact: false on the bakery: (100, 120) within 1e-12
ok    12  the option value may be a variable (exact: true)
ok    13  stats$sum(maximize(...)) = 2 + 6
ok    14  12 x 20 random: feasible, x >= 0, exact and double optima agree to 1e-9 (127.9397367813438005)
ok    15  80 variables + 60 constraints (140), no exact: -> EXACTCHOICE naming both options
ok    16  ...and exact: false solves it (feasible, x >= 0)
ok    18  64 variables + 60 constraints (124): still the exact simplex by default, feasible to 1e-12
ok    17  a 3-column a with a 2-value c: ARRAYSHAPE naming the mismatch
passed =  18   failed =  0 
