FIX: Minimizer: CompositionSet objects set for removal are not removed

[bocklund]

The new_free_stable_compset_indices are created after phase removal here, but these new stable indices with the phases removed are essentially reset here (by the code removed in this change).

This is just one way to address this issue, maybe the most straightforward? An alternative could be set the phase amounts to the phases scheduled for removal to zero and to leave in the code that reconstructs the new_stable_compset_indices (which won't include the phases with new zero phase amounts).

The removed code could also fix a bug where a fixed phase is included in the free stable phase indices if it has non-zero phase amount.

[codecov]

Codecov Report

Merging #343 (3232734) into develop (657f5ba) will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff            @@
##           develop     #343   +/-   ##
========================================
  Coverage    89.70%   89.70%           
========================================
  Files           44       44           
  Lines         4255     4255           
========================================
  Hits          3817     3817           
  Misses         438      438           

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 657f5ba...3232734. Read the comment docs.

[richardotis]

I checked this with the six-component steel calculation and the binary phase diagram examples, and it looks like a good change.