Get calculated values from defined constants in macro

[Nickelgrass]

Hello,

I am using STM32CubeIDE (eclipse) and was wondering if it is possible to get the constant calculated value of a macro.

For example

#define A 1
#define B 2
#define R A+B   // R is now 3

#if R > 2
  #warning R is greater 
#else
  #warning R is less 2
#endif

I can only see 1+2 in the macroexpansion. But clearly the IDE has calculated the value and can evaluate the #if ... #else and gray out the unused part. So how can i access this calculated value? 

[Peter BENSCH]

A macro does not have a stored numeric value, it only performs text substitution. The preprocessor can evaluate the expression in #if, which is why the IDE can determine the active branch, but that result is not directly accessible as a macro value.

If you need a real compile-time constant, use an enum or a const instead.

Regards
/Peter

[Andrew Neil]

---

@Nickelgrass wrote:

 get the constant calculated value of a macro.? 

---

Not sure what you mean by that ?

---

@Nickelgrass wrote:

I can only see 1+2 in the macroexpansion.  

---

That's to be expected; that's what macro expansion does - it is purely text substitution.

[Nickelgrass]

Thanks for the reply. But neither an enum or a const fits my application here. I just need to check if the calculated values are what they need to be. 

[Nickelgrass]

Thanks for the reply. Yes I know it is only a text substitution. But I need to access the evaluated value of the macro to make sure it has the correct value.

[Pavel A.]

[Nickelgrass]

Thanks but that only gives me the text and not the calculated value.

[Ozone]

> A macro does not have a stored numeric value, it only performs text substitution. The preprocessor can evaluate the expression in #if, ...

And in addition to that, the preprocessor is a different "entity" than the C/C++ compiler, and uses different datatypes and ranges than the compiler. The result of preprocessor "calculations" might not be what you expect.

> If you need a real compile-time constant, use an enum or a const instead.

Thus I would definitely second that.

[Pavel A.]

Hmm maybe the upstream Eclipse CDT can calculate "constexpr" and "consteval" expressions for new C++ standard? If not yet, it should be proposed, and by the same occasion add evaluation of const expressions in C macros :)

[Andrew Neil]

That still just shows the macro expansion - it doesn't evaluate the result of the expression after expansion.

@Nickelgrass I think you're just going to have to do the evaluation manually.

Or write some test code that will do it for you ...

#define A 1
#define B 2
#define R A+B   // R is now 3 (sic)

printf( "A=%d; B=%d; R=%d", A, B, R );

Note that this comment is incorrect:

#define R A+B   // R is now 3

R is not 3.

R just expands to A+B

R does not have the numerical value 3

And, for safety, it should be

#define R (A+B)