Make "Scale" command optionally multi-threaded

With large datasets nowadays, often an ImageJ stack is virtual, and lacking a cache, often takes too much time to scroll back and forth across Z. A simple solution is to duplicate a range of sections. But since these are virtual, fetching the files, scaling, interpolating, etc. takes time, so the duplication takes long. Many of these operations could run entirely in parallel, hence, a simple solution is to add a checkbox "multithreaded" to the `Scaler.java` plugin `GenericDialog`, with the corresponding flag for the macro recorder, to duplicate the stack using one `Thread` per slice to duplicate, for as many threads as specified in the ImageJ options.

Start with the Scaler.java plugin and its GenericDialog to trace how stack duplication and macro recording are handled. Check the ImageJ options for the configured thread count and determine how the multithreaded choice should be represented in recorded macros. Done means the optional mode duplicates slices concurrently while preserving the existing behavior when unchecked.

Written by the indexing model from the issue text.