FPC and bitwise negation

[Mike James]

I'm interested in language design and I'm writing a compiler for a Oberon-like language using Coco-R. The bitwise negation is define as :

Code: [Select]

Factor =  /* others */ | "NOT" Factor.
This means that the following code fragment is legit:

Code: [Select]

mybool := NOT NOT NOT NOT mybool;
I checked the same code fragment under FreePascal and this allows it also.

It got me thinking: Should it be flagged as a warning (or even a mild rebuke  ) and when would it ever be useful to allow such a construction?

-=mike=-

[marcov]

Repeated unary NOTs are fairly easily optimizable, and logically sound (though redundant) so why would one unnecesarily burden the compiler (and increase warning/hint noise)?

Both the compiler developer and user have better things to do.

Only exception I can imagine is compilers and, more importantly, source analysers for training/educational purposes.