Comment 0 for bug 1054

I had a very simple spreadsheet: two columns of data (x-axis and y-axis), a third column modeling the y-axis data, a fourth column computing the squared error between the model and the data, and two cells with parameters to the model. I wanted the Solver to change those parameters until the sum of the squared error was minimized. So I set the cells in the Solver like you'd expect, then pressed Solve. The Solver told me I needed to set constraints (see the other usability bug I just filed), so I tried to put in the totally perfunctory constraints that the two parameters needed to be nonnegative. Gnumeric crashed. I restarted gnumeric and tried again. It crashed again. I'm willing to say that this is a highly repeatable bug.