Compiler Robustness ... Naivete - c++

I use Visual Studio Professional 2012. I pre-compiled a class (header and source) successfully. Days later, when compiling another class (for the moment header only) that is utilizing the previous one, the compiler caught a missing reference if(this != &rhs) and semicolon rhs.root = nullptr;.
Perhaps it is my naivete and lack of knowledge about how compilers work but I thought a compiler was robust to catch errors such as these. It appeared to me that only when a specific block of code was required did the compiler feel the need to check it.
I have read about just-in-time compilation and learned how assembly compilers perform a two-pass compilation with symbols first and then syntax. I have not taken a course in compiler construction at my university and I know such courses give great insight into parsers, etc.
The code section where it failed to catch the error is this move assignment operator:
Tree &operator=(Tree &&rhs)
{
if(this != rhs) <--------- no reference to the rhs
{
root = std::move(rhs.root);
rhs.root = nullptr <----------- no semicoln
}
return *this;
}
The errors were generated while compiling boost variant, as well as my visitor class member:
bool operator() (Tree<std::string>& tree) const {
return tree.load(tree);
}
as well as a host of other errors related to boost serialization. The fix was to, of course, correct the missing reference and semicolon but I want to understand why this was caught apparently only when the compiler needed to touch this code?

Is it a template class?
Because semantic analysis of templates makes only sense when they are instantiated. That is, if it is a template, the compiler should generate an error at the missing semicolon (syntactic error), but not at the == operator.
The following code compiles with g++:
template<typename T>
struct A {
void q(A &a) {
if (this == a) {}
}
};
int main(int argc, char **argv) {
A<int> x;
//x.q(x);
}
But doesn't compile when
x.q(x);
is uncommented.

Related

VS2017 version 15.8.3 successfully compiles inline method without returning a required value

I compiled the following code with VS2017 version 15.8.3. Its Warning Level is set to /W4. The code contains two simple getters, one of them GetM() is inline.
The GetM() inline getter does not have a return statement. However, VS2017 happily compiled the code without any warning or error.
The GetN() method will result in error C4716: 'Simple::GetN': must return a value if its return n; statement is commented out.
class Simple
{
int m = 0;
int n = 0;
public:
int GetM() const { /* No return here. */ }
int GetN() const;
};
int Simple::GetN() const
{
return n;
// No return here results in compiler error below.
// error C4716: 'Simple::GetN': must return a value
}
int main()
{
Simple obj;
}
Question: Should the compiler also generate error C4716 for the inline method GetM()?
A method fully defined inside the class definition is sort-of inlined. If it isn't expanded directly inline where it's used, it is compiled outside of the class body. This is the magic sauce that allows a method to see any members that were defined after it in the class.
If it is not used, maybe the compiler doesn't look at it deeply enough to spot the mistake. Maybe it doesn't look at it at all. Maybe it generates a warning and maybe not. That's up to the compiler. Visual Studio seems to have elected report a missing return statement as an error, but to not inspect an unused inline (or sort-of inlined) function.
By changing main to
int main()
{
Simple obj;
obj.GetM();
}
I can make Visual Studio produce error C4716 for gGetM as the function now must be compiled, inlined or not.
I can also
inline int Simple::GetN() const
{
}
to explicitly make GetN inline and "eliminate" the error.
This is all highly compiler, and possibly even compiler option, specific.

TBB compiler error - 'my_task': references must be initialized

I use TBB at multiple places in my project. But it seems that since I update Visual Studio from 15.6.X (X beeing the latest version) to 15.7.1 I get a compiler error at several places, telling me
[...]tbb\task_group.h(94): error C2530: 'my_task': references must be initialized
Looking at the referenced code (tbb/task_group.h):
//! Base class for types that should not be assigned.
class no_assign {
// Deny assignment
void operator=( const no_assign& );
public:
#if __GNUC__
//! Explicitly define default construction, because otherwise gcc issues gratuitous warning.
no_assign() {}
#endif /* __GNUC__ */
};
//! Base class for types that should not be copied or assigned.
class no_copy: no_assign {
//! Deny copy construction
no_copy( const no_copy& );
public:
//! Allow default construction
no_copy() {}
};
// ...
class ref_count_guard : internal::no_copy {
task& my_task; // compiler error occurs here
public:
ref_count_guard( task& t ) : my_task(t) {
my_task.increment_ref_count();
}
~ref_count_guard() {
my_task.decrement_ref_count();
}
};
I don't see why the compiler is complaining there, as the reference is initialized by the constructor. Finding the problem in my code is also not that easy, because the compiler error occurs in every single source file that uses TBB and I don't think I changed anything since my last successful compilation (besides updating VS).
One possibility that comes to my mind is related to this question. If msvc somehow inherits the base class constructors by default, a default constructor would be inherited explaining the error. But testing this scenario seems to disprove it (as the code compiles).
Why is msvc complaining here?
Update
This minimal example reproduces the error on my system:
#include <vector>
#include <tbb/tbb.h>
#include <tbb/flow_graph.h>
void main()
{
std::vector<int> src{1, 2, 3, 4, 5};
tbb::parallel_for_each(src.begin(), src.end(), [](int) { });
}
Update 2
Looks like just including tbb/tbb.h causes the error to occur. I don't even need to call anything. Rebuilding tbb with the new compiler version didn't help either.
Edit
Cross issue on github.
Simply removing /permissive- (e.g. set Comformance Mode to No in the C/C++ options) gets past this issue. I presume Intel will fix this soon.
This is a compiler bug when /permissive- option is used. It can be reproduced with the following code:
struct U {
template<typename T>
void foo() {
class A {
int& iref;
public:
A(int& ir) : iref(ir) { }
};
int the_answer = 42;
A a(the_answer);
}
};
int main() {
U u;
u.foo<int>();
return 0;
}
The code is perfectly valid, compliant C++. As you can see, the reference is explicitly initialized in the member initializer list of the constructor. Also this is seemingly a regression in VS, because at least the initial release of VS 2017 (15.0.something) compiles this code with /permissive-.

Compile error 'nullptr' undeclared identifier

I'm trying to compile a source with Visual Studio 2008 Express, but I'm getting this error:
Error C2065: 'nullptr' undeclared identifier.
My code:
if (Data == nullptr) {
show("Data is null");
return 0;
}
I read on Google that I should upgrade to Visual Studio 2010, but I don't want to do this because of the IntelliSense in Visual Studio 2008. Can this be repaired or replaced?
The error you are getting is because the compiler doesn't recognize the nullptr keyword. This is because nullptr was introduced in a later version of visual studio than the one you are using.
There's 2 ways you might go about getting this to work in an older version. One idea comes from Scott Meyers c++ book where he suggests creating a header with a class that emulates nullptr like this:
const // It is a const object...
class nullptr_t
{
public:
template<class T>
inline operator T*() const // convertible to any type of null non-member pointer...
{ return 0; }
template<class C, class T>
inline operator T C::*() const // or any type of null member pointer...
{ return 0; }
private:
void operator&() const; // Can't take address of nullptr
} nullptr = {};
This way you just need to conditionally include the file based on the version of msvc
#if _MSC_VER < 1600 //MSVC version <8
#include "nullptr_emulation.h"
#endif
This has the advantage of using the same keyword and makes upgrading to a new compiler a fair bit easier (and please do upgrade if you can). If you now compile with a newer compiler then your custom code doesn't get used at all and you are only using the c++ language, I feel as though this is important going forward.
If you don't want to take that approach you could go with something that emulates the old C style approach (#define NULL ((void *)0)) where you make a macro for NULL like this:
#define NULL 0
if(data == NULL){
}
Note that this isn't quite the same as NULL as found in C, for more discussion on that see this question: Why are NULL pointers defined differently in C and C++?
The downsides to this is that you have to change the source code and it is not typesafe like nullptr. So use this with caution, it can introduce some subtle bugs if you aren't careful and it was these subtle bugs that motivated the development of nullptr in the first place.
nullptr is part of C++11, in C++03 you simply use 0:
if (!Data)
When it's not recommended to edit sources, you can just give your c++ compiler appropriate define. For example, in CMakeLists.txt I've added line:
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Dnullptr=0")
And everything worked well

Why does scope of using directive make a difference?

Below are two simplified code examples: the first one compiles just fine, the second one emits a compile error (no operator << found which takes a left hand operator ByteVector...).
The only difference in between the two examples is the placement of the using directive.
I don't want to know why it fails (you got not enough information to answer this), I am only interested in why it does make any difference where I place the using.
I would have expected the exact same behaviour in both examples.
Compiles without error
ByteVector Test()
{
using Base::operator <<;
ByteVector foo;
int bar = 1;
foo << bar;
return foo;
}
Compiles with error
using Base::operator <<;
ByteVector Test()
{
...same as above, without using
}
Extra information:
The operator<< used is defined as follows
template<typename T>
ByteVector& operator<<(ByteVector &, const T&){...};
The only thing that comes to my mind is Visual Studio. If that was the case, you may want to put the using directive right after the corresponding #include. As the remark on the page says:
Putt your using directive at the beginning of the source code file to
reduce the potential for unexpected behavior with IntelliSense.
Otherwise, it shouldn't make any difference.

Error C2275 caused by template member function. Is this code wrong?

I think I've run into a (possible) VC6 (I know. It's what we use.) compiler error, but am open to the fact that I've just missed something dumb. Given the following code (It's just an example!):
#include <iostream>
// Class with template member function:
class SomeClass
{
public:
SomeClass() {};
template<class T>
T getItem()
{
return T();
};
};
// Dummy just used to recreate compiler error
class OtherClass
{
public:
OtherClass() {};
};
std::ostream& operator<<( std::ostream& oStr, const OtherClass& obj )
{
return oStr << "OtherClass!";
};
// Main illustrates the error:
int main(int argc, char* argv[])
{
SomeClass a;
OtherClass inst2 = a.getItem<OtherClass>(); // Error C2275 happens here!
std::cout << inst2 << std::endl;
return 0;
}
If I try to compile this code VC6, dies on a.getItem<OtherClass>() yielding:
Error C2275: 'OtherClass' : illegal use of this type as an expression.
Have I overlooked some trivial syntax issue? Am I breaking a rule?
This code compiles just fine under gcc 4.3.4. Is it yet another compliance issue with VC6?
Thanks!
Among many other things with the word template in it, VC6 couldn't deal with function templates where the template parameters aren't also function parameters. The common workaround was to add a dummy function parameter:
template<class T>
T getItem(T* /*dummy*/ = NULL)
{
return T();
} // note: no ; after function definitions
However, in general, VC6 is pretty lame and often chokes as soon as a TU contains the template keyword. I had to beat my head against it for several years (big code base compiled with several compilers/compiler versions; VC6 giving us an endless amount of trouble) and was very glad when I got rid of it in 2003.
This is likely to be a VC6 issue. Although VC6 compiles most basic templates correctly it is known to have many issues when you start to move towards the more advanced template uses. Member templates are an area where VC6 is known to be weak on conformance.
I believe that's another bug in VC6, you should really switch to a more up-to-date compiler.