Offset start time when stepping

[valentinsulzer]

Currently, when solving a simulation by stepping, the final time step of the old solution (e.g. [0,0.01 ..., 1]) is the same as the first time step of the next solution (e.g. [1,1.01,...]). We get around this by excluding the first time step of the next solution when constructing the full solution ([0,0.01,...,1,1.01,...]). However, this means that the solution naively constructed by concatenating together all the steps is different from the solution as constructed by PyBaMM. Instead, we should offset the start of each step by a relevant small number (e.g. 1e-6) and then keep the first step.

[valentinsulzer]

Is it better to offset time by a small number, save only parts [1:] of the step, or leave as is?