Fixed segfault when monster tries to grab-drag another monster out of reality bubble

[HadeanLake]

Summary

Bugfixes "Fixed segfault when monster tries to grab-drag another monster out of reality bubble"

Purpose of change

fix #75793

I believe, someone was writing the code for selecting a random square, then discovered that function random_entry already exists and decided to use it instead, but forgot to remove already written and no longer needed code

There are 2 lines of code at mattack_actors.cpp:617

            std::set<tripoint>::iterator intersect_iter = intersect.begin();
            std::advance( intersect_iter, rng( 0, intersect.size() - 1 ) );

Those lines are optimized away by some configs since variable intersect_iter is never used after this

set intersect is not checked for being empty.
intersect is empty when monster that does dragging is right at the edge of reality bubble and the next tile in the direction of dragging is outside of it.
when intersect is empty second line causes segfault (if built for linux using gcc or built for windows using mxe) or debug assetion failure "cannot decrement begin map/set iterator" (if built using msvc for debugging)

this crash does not occur in android and msvc-release builds

function random_entry called on the next line after those 2 has container.empty() check, 2 same lines and return of it is assigned to target_square and actually used down the code

Describe the solution

remove redundant code

Describe alternatives you've considered

Testing

Compiled, verified that monsters can grab and drag and save from linked issue does not trigger segfault

Additional context

Should be backported to 0.H-branch since those lines are present there

[inogenous]

Wow, that's such a nice find. Fixing a bug by only removing (unused) code, thereby also making it easier to maintain. Hats off to you @HadeanLake ! 🥇