Warm start does no longer work in 1.5.0

[EmJay276]

I am solving my problen in a loop.
In a previous version (I don't know exactly which one it was June 2023) I was using the last result as the initial guess:

    if previous_res is not None:
        x1, info = nlp.solve(x0,
                             lagrange=previous_res['mult_g'],
                             zl=previous_res['mult_x_L'],
                             zu=previous_res['mult_x_U'])
    else:
        x1, info = nlp.solve(x0)

previous_res is identical to info, but this does no longer work and raises the exeption

    x1, info = nlp.solve(x=x0,
               ^^^^^^^^^^^^^^^
  File "cyipopt\\cython\\ipopt_wrapper.pyx", line 628, in ipopt_wrapper.Problem.solve
ValueError: operands could not be broadcast together with shapes (208,) (0,) 

Is warm start no longer supported?

[moorepants]

I just tried the tests for warm start from #271 in version 1.3.0, 1.4.1, 1.5.0, and the tip of master and it ran. I don't think anything has changed about warm starting other that adding the documentation.

You'll have to provide a full reproducer for more help.

[EmJay276]

Ok that is really weird, I will investigate this and report again.
But it will take at least some weeks until I am able to test it again, sorry.