C++ – Why does everybody use unanchored namespace declarations (i.e. std:: not ::std::)

cnamespaces

It seems to me that using unanchored namespaces is just asking for trouble later when someone puts in a new namespace that happens to have the same name as a root level namespace and mysteriously alters the meaning of a whole lot of programs. So, why do people always say std:: instead of ::std::. Do they really mean to be saying "I want to use whatever std is handy, not the root one."?

Here is an example of what I mean:

In fred/Foo.h:

#include <string>

namespace fred {

class Foo {
 public:
   void aPublicMember(const std::string &s);
};

} // end namespace fred

In fred/Bar.h:

namespace fred {
namespace std {  // A standard fred component

class string { // Something rather unlike the ::std::string
   // ...
};

} // namespace std

class Bar {
 public:
   void aPublicMember(std::string &s);
};

} // namespace fred

In oops.cpp:

#include <string>
#include "fred/Bar.h"
#include "fred/Foo.h"  // Oops, the meaning of Foo is now different.

Is that what people want, or am I missing something?

And maybe you say that you should just never name a namespace std. And that's all well and good, but what about some other root level namespace then? Should any root level namespace anybody ever defines anywhere always be off-limits for a sub-namespace name?

To clarify, I won't consider any answer that tells me std is special because I just used it as an example. I'm talking about a general issue, and I'm using std as a prop to illustrate it, though I do admit it's a rather startling prop.

Best Answer

The practical reason for unanchored namespaces is that one level of namespaces usually is enough. When it isn't, a second level is usually going to be used for implementation details. And finally, even when using multiple levels, they are still usually specified implicitly from root level. ie. even inside namespace ns1, you'd typically refer to ns1::ns2::foo instead of ns2::foo or ::ns1::ns2::foo.

So, for these three reasons the ::ns1 form is redundant in normal cases. The only case where I'd consider it would be in submissions to Boost, because as a Boost author I won't know where my software will be used.

Related Solutions

C++ – Why use static_cast(x) instead of (int)x

The main reason is that classic C casts make no distinction between what we call static_cast<>(), reinterpret_cast<>(), const_cast<>(), and dynamic_cast<>(). These four things are completely different.

A static_cast<>() is usually safe. There is a valid conversion in the language, or an appropriate constructor that makes it possible. The only time it's a bit risky is when you cast down to an inherited class; you must make sure that the object is actually the descendant that you claim it is, by means external to the language (like a flag in the object). A dynamic_cast<>() is safe as long as the result is checked (pointer) or a possible exception is taken into account (reference).

A reinterpret_cast<>() (or a const_cast<>()) on the other hand is always dangerous. You tell the compiler: "trust me: I know this doesn't look like a foo (this looks as if it isn't mutable), but it is".

The first problem is that it's almost impossible to tell which one will occur in a C-style cast without looking at large and disperse pieces of code and knowing all the rules.

Let's assume these:

class CDerivedClass : public CMyBase {...};
class CMyOtherStuff {...} ;

CMyBase  *pSomething; // filled somewhere

Now, these two are compiled the same way:

CDerivedClass *pMyObject;
pMyObject = static_cast<CDerivedClass*>(pSomething); // Safe; as long as we checked

pMyObject = (CDerivedClass*)(pSomething); // Same as static_cast<>
                                     // Safe; as long as we checked
                                     // but harder to read

However, let's see this almost identical code:

CMyOtherStuff *pOther;
pOther = static_cast<CMyOtherStuff*>(pSomething); // Compiler error: Can't convert

pOther = (CMyOtherStuff*)(pSomething);            // No compiler error.
                                                  // Same as reinterpret_cast<>
                                                  // and it's wrong!!!

As you can see, there is no easy way to distinguish between the two situations without knowing a lot about all the classes involved.

The second problem is that the C-style casts are too hard to locate. In complex expressions it can be very hard to see C-style casts. It is virtually impossible to write an automated tool that needs to locate C-style casts (for example a search tool) without a full blown C++ compiler front-end. On the other hand, it's easy to search for "static_cast<" or "reinterpret_cast<".

pOther = reinterpret_cast<CMyOtherStuff*>(pSomething);
      // No compiler error.
      // but the presence of a reinterpret_cast<> is 
      // like a Siren with Red Flashing Lights in your code.
      // The mere typing of it should cause you to feel VERY uncomfortable.

That means that, not only are C-style casts more dangerous, but it's a lot harder to find them all to make sure that they are correct.

C++ – Why use apparently meaningless do-while and if-else statements in macros

The do ... while and if ... else are there to make it so that a semicolon after your macro always means the same thing. Let's say you had something like your second macro.

#define BAR(X) f(x); g(x)

Now if you were to use BAR(X); in an if ... else statement, where the bodies of the if statement were not wrapped in curly brackets, you'd get a bad surprise.

if (corge)
  BAR(corge);
else
  gralt();

The above code would expand into

if (corge)
  f(corge); g(corge);
else
  gralt();

which is syntactically incorrect, as the else is no longer associated with the if. It doesn't help to wrap things in curly braces within the macro, because a semicolon after the braces is syntactically incorrect.

if (corge)
  {f(corge); g(corge);};
else
  gralt();

There are two ways of fixing the problem. The first is to use a comma to sequence statements within the macro without robbing it of its ability to act like an expression.

#define BAR(X) f(X), g(X)

The above version of bar BAR expands the above code into what follows, which is syntactically correct.

if (corge)
  f(corge), g(corge);
else
  gralt();

This doesn't work if instead of f(X) you have a more complicated body of code that needs to go in its own block, say for example to declare local variables. In the most general case the solution is to use something like do ... while to cause the macro to be a single statement that takes a semicolon without confusion.

#define BAR(X) do { \
  int i = f(X); \
  if (i > 4) g(i); \
} while (0)

You don't have to use do ... while, you could cook up something with if ... else as well, although when if ... else expands inside of an if ... else it leads to a "dangling else", which could make an existing dangling else problem even harder to find, as in the following code.

if (corge)
  if (1) { f(corge); g(corge); } else;
else
  gralt();

The point is to use up the semicolon in contexts where a dangling semicolon is erroneous. Of course, it could (and probably should) be argued at this point that it would be better to declare BAR as an actual function, not a macro.

In summary, the do ... while is there to work around the shortcomings of the C preprocessor. When those C style guides tell you to lay off the C preprocessor, this is the kind of thing they're worried about.

Best Answer

Related Solutions

C++ – Why use static_cast(x) instead of (int)x

C++ – Why use apparently meaningless do-while and if-else statements in macros

Related Topic