How to write "nested if...else statement for constants" without warnings, in gnu++11 standard?

Question

When I use nested if....else statement with

if (std::is_same<T, T1>::value) 
{
  // do something
} 
else if (std::is_same<T, T2>::value)
{
  // do something else
}
.
.
.
else
{
  // print error
}

I get a QACPP static code analyzer's compiler warning qacpp-4.2.1-4090 with message "The condition in this 'if' statement is constant." How do I fix this compiler warning in gnu++11 standards ?

NOTE: I'm not an expert at C++ so please excuse me if the question sounds amateur.

I slightly modified the question, if you see it now, the above lines are called inside the constructor of a template class T and compared with other template types T1, T2 to assign a particular value for a member variable of the template class T, based on the matching if condition. — Rookie_Coder2318
– Rookie_Coder2318, Commented Apr 2, 2019 at 21:06

Bathsheba · Accepted Answer · 2019-04-02 20:51:25Z

3

For a particular instantiation of T, the if conditionals are constant. In other words std::is_same<T, int>::value &c. are constant expressions.

But that may well be a necessity if you've set the code out in this perfectly legitimate way. (You might be able to refactor to a static polymorphism technique using template specialisation if obviating the static analyser warning is important.)

Seems that the static analyser is being over-zealous to me.

edited Apr 2, 2019 at 20:51

answered Apr 2, 2019 at 20:36

Bathsheba

235k35 gold badges376 silver badges504 bronze badges

Sign up to request clarification or add additional context in comments.

4 Comments

Rookie_Coder2318 Over a year ago

Will it be any different if I tell that the actual code does not have if (std::is_same<T, int>::value) but has if (std::is_same<T1, T2>::value) ? However I the code analyzer shows warning in both cases.

Bathsheba Over a year ago

There's no difference. This is because std::is_same<T1, T2>::value is always a compile time evaluable constant expression.

Rookie_Coder2318 Over a year ago

Yea, I guess you're right... so there is nothing that can be done to fix such warnings ? NOTE: I use PRQA Source Code Analyzer v2.3.0.93377 Framework.

Bathsheba Over a year ago

@Rookie_Coder2318; Other than changing the warnings or refactoring the code (which I wouldn't do on the grounds that it would be an imperfect tail wagging a reasonably coded dog), no. Static analysers have their place but I don't think the goal is to eliminate all warnings.

Kaz · Accepted Answer · 2019-04-02 21:22:26Z

The coding tool is literally asking you to abandon good practices like:

if (static_condition) {
   // code that is effectively compiled out when static_condition is false
}

and instead use something like:

#if static_condition
   // code that is compiled out when static_condition is false
#endif

which can't even be done in the exact situation you have there, and is an inferior practice.

This is not a good, productive diagnostic, except in cases when static_condition is clearly unconditional, like if (false) ....

If the diagnostic tool is otherwise valuable, find out how to suppress that diagnostic in specific files or lines of code. Maybe it supports some directive that can be put into comments to squelch some diagnostics.

// @foo-bar tool: disable-warning(13125)
if (static_condition) ...

Collectives™ on Stack Overflow

How to write "nested if...else statement for constants" without warnings, in gnu++11 standard?

2 Answers 2

4 Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

4 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related