Unique short name

The auto generation of the short names is fantastic, but it would be nice to have it automatically detect used names and choose the next best thing. So if two options are "--cheese" and "--cherries", it would be "-c" and "-C" respectively, or if that is taken, "-d", etc.

No file, test, or entry point is named in the issue. Start by locating the automatic short-name generation and inspect how names are assigned; done means conflicting options receive distinct unused short names, following the stated case-sensitive examples and continuing when candidates are taken.

Written by the indexing model from the issue text.