#if defined() Usage

[dcm684]

I have the following block of code in my program

[Ttelmah]

What compiler version?.

It'll have to evaluate the statement, but the ANIMAL==COW, should still be accepted.

It works OK, on current compilers, and even on back to 4.118.

Are you sure you are typing both '&' symbols?. Remember a single & will involve a bitwise combination, while the double gives a logical and. The former will require ANIMAL and COW to be defined.

Best Wishes

[dcm684]

Actually, "ANIMAL" is not defined. Its a value that I periodically #define for certain production environments. I'm sorry that I did not state that earlier.

I copied and pasted my code straight from MPLAB X.

My compiler version is 5.020 PCH.

What I expect would happen is that the compiler would see "defined(ANIMAL)" realize that "ANIMAL" is not defined and as a result realize the statement will never be true so it would skip the "ANIMAL == COW" comparison.

I would not expect any warning to be thrown since there is a check to see if "ANIMAL" was defined before trying to use the not defined "ANIMAL".

[Ttelmah]

Seriously, the statement, works as posted in every compiler I have tried. With both values undefined.

Are you sure there isn't a fault in an earlier line?.

There is a classical problem with CCS, where it'll accept things that are wrong, and then error a lot of lines later, when something reaches a 'limit' internally, so makes it realise something is wrong.

[Ttelmah]

Also worth pointing out, that warnings, are only warnings. They don't stop the compiler compiling. If you don't want the warning to display, Unidentified identifier is 215, so
#IGNORE_WARNINGS 215

would stop it displaying.
However I have to repeat that this code compiles correctly, without this warning. If you define 'ANIMAL, then it'll give:

">>> Warning 215 location: Undefined identifier COW"

However without this defined, it compiles without warnings, so something earlier is wrong....