It would be preferable if SLSQP Fortran code didn't exceed the bounds at all.

@lukasheinrich, @kratsg can you try out this proposed fix for #11403 and see how it goes on your issues?

One possibility is that we issue no warning and just clip by default.

Last commit also applies bounds to the numerical differentiation of the jacobian, as mentioned in @davidpasquale comments in #12949.

@larsoner I made the suggested changes and rebased. Everything you reviewed is now in the first commit, the changes I made are in the second.

Travis is just unhappy about some linting errors that I guess we can safely ignore. Thanks @andyfaff !

Any new on this issue?
I am having the follow warming when I use the SLSQP minimization

RuntimeWarning: Values in x were outside bounds during a minimize step, clipping to bounds
  warnings.warn("Values in x were outside bounds during a "

all issues I read in the matter are from older versions of scipy, I am using scipy=1.7.3 and numpy=1.21.2. How can I maintain the minimization steps within the bounds? And any idea in why is this happening?