spin does not produce correct output if tspan does not start with 0

[NBoulle]

If I run the following code

S = spinop('ks');
S.tspan = 100:1e-1:101;
[u,t] = spin(S, 256, 1e-1);

I get the following output for t:

t =

   1.0e+02 *

  Columns 1 through 7

                   0   1.001000000000000   1.002000000000000   1.003000000000000   1.004000000000000   1.005000000000000   1.006000000000000

  Columns 8 through 11

   1.007000000000000   1.008000000000000   1.009000000000000   1.010000000000000

spin replaces the first entry of tspan by 0 and u{1} = u(t=0), which sounds like a bug.

However, S.tspan = [0 100:1e-1:101]; works as expected.

There seems to be a similar issue for the other spin codes (at least spinsphere).