Sorry, but I do not agree:
I try to use the floating point libraries provided by MicroChip. In the header file they assign - let's say - AARGB7 to 0x20 as well as ACCB7. There's also an application note available to embedd the librariers into C. (see AN669; well, I know they don't use C2C)
So when this is not a bug and the statement makes no sense, why there's no error/warning generated?
Defining the variable/address overlap with #define - statements makes the code hard to debug in case of a problem. Changing the library code is time consuming ...
Maybe you can handle this as a feature request?