Ranked overload resolution

Currently, functions are selected as overload candidates if all arguments can be implicitly converted to the expected types. If there is not exactly one candidate, resolution fails.
This makes it impossible to call different functions depending on whether one argument is a storage or a memory array, because storage arrays can be implicitly converted to both storage and memory arrays.
We should add a ranking of candidates depending on which conversions have to take place.
For this, a conversion from e.g. storage to memory should be "worse" than a widening of an integer type which in turn is worse than no conversion at all.

No files, tests, or entry points are named. Start by locating the overload-resolution implementation and its candidate-conversion tests, then trace how storage-to-memory conversion, integer widening, and exact matches are represented. Done means overloads can be selected by the stated conversion ranking, including the storage-versus-memory case, with regression coverage.

Written by the indexing model from the issue text.