gecko-dev/xpcom/public/nsCOMPtr.h
1999-02-18 22:34:15 +00:00

669 lines
19 KiB
C++

/* -*- Mode: C++; tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
/*
* The contents of this file are subject to the Netscape Public License
* Version 1.0 (the "NPL"); you may not use this file except in
* compliance with the NPL. You may obtain a copy of the NPL at
* http://www.mozilla.org/NPL/
*
* Software distributed under the NPL is distributed on an "AS IS" basis,
* WITHOUT WARRANTY OF ANY KIND, either express or implied. See the NPL
* for the specific language governing rights and limitations under the
* NPL.
*
* The Initial Developer of this code under the NPL is Netscape
* Communications Corporation. Portions created by Netscape are
* Copyright (C) 1998 Netscape Communications Corporation. All Rights
* Reserved.
*/
#ifndef nsCOMPtr_h___
#define nsCOMPtr_h___
// Wrapping includes can speed up compiles (see "Large Scale C++ Software Design")
#ifndef nsDebug_h___
#include "nsDebug.h"
// for |NS_PRECONDITION|
#endif
#ifndef nsISupports_h___
#include "nsISupports.h"
// for |nsresult|, |NS_ADDREF|, et al
#endif
/*
TO DO...
+ make alternative function for |getter_AddRefs| (or something)
+ make constructor for |nsQueryInterface| explicit (suddenly construct/assign from raw pointer becomes illegal)
+ Improve internal documentation
+ mention *&
+ alternatives for comparison
+ do_QueryInterface
*/
/* USER MANUAL
See also:
<http://www.meer.net/ScottCollins/doc/nsCOMPtr.html>, or
<http://www.mozilla.org/projects/xpcom/nsCOMPtr.html>
What is |nsCOMPtr|?
|nsCOMPtr| is a `smart-pointer'. It is a template class that acts, syntactically,
just like an ordinary pointer in C or C++, i.e., you can apply |*| or |->| to it to
`get to' what it points at. |nsCOMPtr| is smart in that, unlike a raw COM
interface pointer, |nsCOMPtr| manages |AddRef|, |Release|, and |QueryInterface|
_for_ you.
For instance, here is a typical snippet of code (at its most compact) where you assign
a COM interface pointer into a member variable:
NS_IF_RELEASE(mFoop); // If I have one already, I must release it before over-writing it.
if ( mFooP = aPtr ) // Now it's safe to assign it in, and, if it's not NULL
mFooP->AddRef(); // I must |AddRef| it, since I'll be holding on to it.
If our member variable |mFooP| were a |nsCOMPtr|, however, the snippet above
would look like this:
mFoop = aPtr; // Note: automatically |Release|s the old and |AddRef|s the new
|nsCOMPtr| helps you write code that is leak-proof, exception safe, and significantly
less verbose than you would with raw COM interface pointers. With |nsCOMPtr|, you
may never have to call |AddRef|, |Release|, or |QueryInterface| by hand.
You still have to understand COM. You still have to know which functions return
interface pointers that have already been |AddRef|ed and which don't. You still
have to ensure your program logic doesn't produce circularly referencing garbage.
|nsCOMPtr| is not a panacea. It is, however, helpful, easy to use, well-tested,
and polite. It doesn't require that a function author cooperate with you, nor does
your use force others to use it.
Where should I use |nsCOMPtr|?
...
Where _shouldn't_ I use |nsCOMPtr|?
In public interfaces... [[others]]
How does a |nsCOMPtr| differ from a raw pointer?
A |nsCOMPtr| differs, syntactically, from a raw COM interface pointer in three
ways:
+ It's declared differently, e.g.,
// instead of saying // you say
IFoo* fooP; nsCOMPtr<IFoo> fooP;
+ You can't call |AddRef| or |Release| through it,
fooP->AddRef(); // OK fooP->AddRef(); // Error: no permission
fooP->Release(); // OK fooP->Release(); // Error: no permission
+ You can't just apply an |&| to it to pass it to the typical `getter' function
AcquireFoo(&fooP); AcquireFoo( getter_AddRefs(fooP) );
GetFoo(&fooP); GetFoo( getter_doesnt_AddRef(fooP) );
How do I use |nsCOMPtr|?
Typically, you can use a |nsCOMPtr| exactly as you would a standard COM
interface pointer:
IFoo* fooP; nsCOMPtr<IFoo> fooP;
// ... // ...
fooP->SomeFunction(x, y, z); fooP->SomeFunction(x, y, z);
AnotherFunction(fooP); AnotherFunction(fooP);
if ( fooP ) if ( fooP )
// ... // ...
if ( fooP == barP ) if ( fooP == barP )
// ... // ...
There are some differences, though. In particular, you can't call |AddRef| or |Release|
through a |nsCOMPtr| directly, nor would you need to. |AddRef| is called for you
whenever you assign a COM interface pointer _into_ a |nsCOMPtr|. |Release| is
called on the old value, and also when the |nsCOMPtr| goes out of scope. Trying
to call |AddRef| or |Release| yourself will generate a compile-time error.
fooP->AddRef(); // fooP->AddRef(); // ERROR: no permission
fooP->Release(); // fooP->Release(); // ERROR: no permission
The final difference is that a bare |nsCOMPtr| (or rather a pointer to it) can't
be supplied as an argument to a function that `fills in' a COM interface pointer.
Rather it must be wrapped with a utility call that says whether the function calls
|AddRef| before returning, e.g.,
...->QueryInterface(riid, &fooP) ...->QueryInterface(riid, getter_AddRefs(fooP))
LookupFoo(&fooP); LookupFoo( getter_doesnt_AddRef(fooP) );
Don't worry. It's a compile-time error if you forget to wrap it.
Compare the raw-pointer way...
IFoo* foo = 0;
nsresult status = CreateIFoo(&foo);
if ( NS_SUCCEEDED(status) )
{
IBar* bar = 0;
if ( NS_SUCCEEDED(status = foo->QueryInterface(riid, &bar)) )
{
IFooBar* foobar = 0;
if ( NS_SUCCEEDED(status = CreateIFooBar(foo, bar, &foobar)) )
{
foobar->DoTheReallyHardThing();
foobar->Release();
}
bar->Release();
}
foo->Release();
}
To the smart-pointer way...
nsCOMPtr<IFoo> fooP;
nsresult status = CreateIFoo( getter_AddRefs(fooP) );
if ( NS_SUCCEEDED(status) )
if ( nsCOMPtr<IBar> barP( fooP ) )
{
nsCOMPtr<IFooBar> fooBarP;
if ( NS_SUCCEEDED(status = CreateIFooBar(fooP, barP, getter_AddRefs(fooBarP))) )
fooBarP->DoTheReallyHardThing();
}
Is there an easy way to convert my current code?
...
What do I have to beware of?
VC++ < 6.0 _can't_ handle the following situation
class nsIFoo; // forward declare some class
// ...
nsCOMPtr<nsIFoo> bar; // ERROR: incomplete type nsIFoo, etc.
Instead, you must make sure that you actually defined the underlying interface class, e.g.,
#include "nsIFoo.h" // fully defines |class nsIFoo|
// ...
nsCOMPtr<nsIFoo> bar; // no problem
Why is this? It's because VC++ tries to instantiate every member of the template
as soon as it sees the template declarations. Bad compiler. No cookie!
[[Thanks to mjudge, waterson, and pinkerton on this one.]]
Why does |getter_AddRefs| have such a funny name? I.e., why doesn't it follow our
naming conventions?
|getter_AddRefs| and |getter_doesnt_AddRef| use underscores for the same
reason our special macros do, quoting from our coding conventions "...to make them
stick out like a sore thumb". Note also that since |AddRef| is one word,
|getter_AddRefs| and |getter_doesnt_AddRef| couldn't have the right spacing if only inter-
caps were used.
*/
/*
WARNING:
This file defines several macros for internal use only. These macros begin with the
prefix |NSCAP_|. Do not use these macros in your own code. They are for internal use
only for cross-platform compatibility, and are subject to change without notice.
*/
/*
Set up some |#define|s to turn off a couple of troublesome C++ features.
Interestingly, none of the compilers barf on template stuff.
Ideally, we would want declarations like these in a configuration file
that everybody would get. Deciding exactly how to do that should
be part of the process of moving from experimental to production.
Update: ramiro is working on getting these into the configuration system.
*/
#if defined(__GNUG__) && (__GNUC_MINOR__ <= 90) && !defined(SOLARIS)
#define NSCAP_NO_MEMBER_USING_DECLARATIONS
#if (defined(LINUX) || defined(__bsdi__)) && (__GNUC_MINOR__ <= 7)
#define NSCAP_NEED_UNUSED_VIRTUAL_IMPLEMENTATIONS
#endif
#endif
#if defined(SOLARIS) && !defined(__GNUG__)
#define NSCAP_NO_BOOL
#define NSCAP_NO_EXPLICIT
#define NSCAP_NO_NEW_CASTS
#define NSCAP_NO_MEMBER_USING_DECLARATIONS
#endif
#if defined(_MSC_VER) && (_MSC_VER<1100)
#define NSCAP_NO_EXPLICIT
#define NSCAP_NO_BOOL
#endif
#if defined(IRIX)
#define NSCAP_NO_MEMBER_USING_DECLARATIONS
#define NSCAP_NO_EXPLICIT
#define NSCAP_NO_NEW_CASTS
#define NSCAP_NO_BOOL
#endif
#ifdef NSCAP_NO_EXPLICIT
#define explicit
#endif
#ifndef NSCAP_NO_NEW_CASTS
#define NSCAP_REINTERPRET_CAST(T,x) reinterpret_cast<T>(x)
#else
#define NSCAP_REINTERPRET_CAST(T,x) ((T)(x))
#endif
#ifndef NSCAP_NO_BOOL
typedef bool NSCAP_BOOL;
#else
typedef PRBool NSCAP_BOOL;
#endif
#ifdef NSCAP_FEATURE_DEBUG_MACROS
#define NSCAP_ADDREF(ptr) NS_ADDREF(ptr)
#define NSCAP_RELEASE(ptr) NS_RELEASE(ptr)
#else
#define NSCAP_ADDREF(ptr) (ptr)->AddRef()
#define NSCAP_RELEASE(ptr) (ptr)->Release()
#endif
/*
WARNING:
VC++4.2 is very picky. To compile under VC++4.2, the classes must be defined
in an order that satisfies:
nsDerivedSafe < nsCOMPtr
nsDontAddRef < nsCOMPtr
nsCOMPtr < nsGetterAddRefs
The other compilers probably won't complain, so please don't reorder these
classes, on pain of breaking 4.2 compatibility.
*/
template <class T>
class nsDerivedSafe : public T
/*
No client should ever see or have to type the name of this class. It is the
artifact that makes it a compile-time error to call |AddRef| and |Release|
on a |nsCOMPtr|.
See |nsCOMPtr::operator->|, |nsCOMPtr::operator*|, et al.
*/
{
private:
#ifndef NSCAP_NO_MEMBER_USING_DECLARATIONS
using T::AddRef;
using T::Release;
#else
nsrefcnt AddRef();
nsrefcnt Release();
#endif
void operator delete( void* ); // NOT TO BE IMPLEMENTED
// declaring |operator delete| private makes calling delete on an interface pointer a compile error
nsDerivedSafe& operator=( const nsDerivedSafe& ); // NOT TO BE IMPLEMENTED
// you may not call |operator=()| through a dereferenced |nsCOMPtr|, because you'd get the wrong one
};
#if defined(NSCAP_NO_MEMBER_USING_DECLARATIONS) && defined(NSCAP_NEED_UNUSED_VIRTUAL_IMPLEMENTATIONS)
template <class T>
nsrefcnt
nsDerivedSafe<T>::AddRef()
{
return 0;
}
template <class T>
nsrefcnt
nsDerivedSafe<T>::Release()
{
return 0;
}
#endif
template <class T>
struct nsDontQueryInterface
/*
...
*/
{
explicit
nsDontQueryInterface( T* aRawPtr )
: mRawPtr(aRawPtr)
{
// nothing else to do here
}
T* mRawPtr;
};
template <class T>
inline
nsDontQueryInterface<T>
dont_QueryInterface( T* aRawPtr )
{
return nsDontQueryInterface<T>(aRawPtr);
}
struct nsQueryInterface
{
// explicit
nsQueryInterface( nsISupports* aRawPtr, nsresult* error = 0 )
: mRawPtr(aRawPtr),
mErrorPtr(error)
{
// nothing else to do here
}
nsISupports* mRawPtr;
nsresult* mErrorPtr;
};
inline
nsQueryInterface
do_QueryInterface( nsISupports* aRawPtr, nsresult* error = 0 )
{
return nsQueryInterface(aRawPtr, error);
}
template <class T>
struct nsDontAddRef
/*
...cooperates with |nsCOMPtr| to allow you to assign in a pointer _without_
|AddRef|ing it. You would rarely use this directly, but rather through the
machinery of |getter_AddRefs| in the argument list to functions that |AddRef|
their results before returning them to the caller.
See also |getter_AddRefs()| and |class nsGetterAddRefs|.
*/
{
explicit
nsDontAddRef( T* aRawPtr )
: mRawPtr(aRawPtr)
{
// nothing else to do here
}
T* mRawPtr;
};
template <class T>
inline
nsDontAddRef<T>
dont_AddRef( T* aRawPtr )
/*
...makes typing easier, because it deduces the template type, e.g.,
you write |dont_AddRef(fooP)| instead of |nsDontAddRef<IFoo>(fooP)|.
Like the class it is shorthand for, you would rarely use this directly,
but rather through |getter_AddRefs|.
*/
{
return nsDontAddRef<T>(aRawPtr);
}
class nsCOMPtr_base
{
public:
nsCOMPtr_base( nsISupports* rawPtr = 0 )
: mRawPtr(rawPtr)
{
// nothing else to do here
}
~nsCOMPtr_base()
{
if ( mRawPtr )
NSCAP_RELEASE(mRawPtr);
}
NS_EXPORT void assign_with_AddRef( nsISupports* );
NS_EXPORT void assign_with_QueryInterface( nsISupports*, const nsIID&, nsresult* );
NS_EXPORT void** begin_assignment();
protected:
nsISupports* mRawPtr;
};
template <class T>
class nsCOMPtr : private nsCOMPtr_base
/*
...
*/
{
public:
typedef T element_type;
nsCOMPtr()
// : nsCOMPtr_base(0)
{
// nothing else to do here
}
nsCOMPtr( const nsQueryInterface& aSmartPtr )
// : nsCOMPtr_base(0)
{
assign_with_QueryInterface(aSmartPtr.mRawPtr, T::IID(), aSmartPtr.mErrorPtr);
}
nsCOMPtr( const nsDontAddRef<T>& aSmartPtr )
: nsCOMPtr_base(aSmartPtr.mRawPtr)
{
// nothing else to do here
}
nsCOMPtr( const nsDontQueryInterface<T>& aSmartPtr )
: nsCOMPtr_base(aSmartPtr.mRawPtr)
{
if ( mRawPtr )
NSCAP_ADDREF(mRawPtr);
}
nsCOMPtr( const nsCOMPtr<T>& aSmartPtr )
: nsCOMPtr_base(aSmartPtr.mRawPtr)
{
if ( mRawPtr )
NSCAP_ADDREF(mRawPtr);
}
nsCOMPtr<T>&
operator=( const nsQueryInterface& rhs )
{
assign_with_QueryInterface(rhs.mRawPtr, T::IID(), rhs.mErrorPtr);
return *this;
}
nsCOMPtr<T>&
operator=( const nsDontAddRef<T>& rhs )
{
if ( mRawPtr )
NSCAP_RELEASE(mRawPtr);
mRawPtr = rhs.mRawPtr;
return *this;
}
nsCOMPtr<T>&
operator=( const nsDontQueryInterface<T>& rhs )
{
assign_with_AddRef(rhs.mRawPtr);
return *this;
}
nsCOMPtr<T>&
operator=( const nsCOMPtr& rhs )
{
assign_with_AddRef(rhs.mRawPtr);
return *this;
}
nsDerivedSafe<T>*
get() const
// returns a |nsDerivedSafe<T>*| to deny clients the use of |AddRef| and |Release|
{
return NSCAP_REINTERPRET_CAST(nsDerivedSafe<T>*, mRawPtr);
}
nsDerivedSafe<T>*
operator->() const
// returns a |nsDerivedSafe<T>*| to deny clients the use of |AddRef| and |Release|
{
NS_PRECONDITION(mRawPtr != 0, "You can't dereference a NULL nsCOMPtr with operator->().");
return get();
}
nsDerivedSafe<T>&
operator*() const
// returns a |nsDerivedSafe<T>*| to deny clients the use of |AddRef| and |Release|
{
NS_PRECONDITION(mRawPtr != 0, "You can't dereference a NULL nsCOMPtr with operator*().");
return *get();
}
operator nsDerivedSafe<T>*() const
{
return get();
}
#if 0
private:
friend class nsGetterAddRefs<T>;
/*
In a perfect world, the following member function, |StartAssignment|, would be private.
It is and should be only accessed by the closely related class |nsGetterAddRefs<T>|.
Unfortunately, some compilers---most notably VC++5.0---fail to grok the
friend declaration above or in any alternate acceptable form. So, physically
it will be public (until our compilers get smarter); but it is not to be
considered part of the logical public interface.
*/
#endif
T**
StartAssignment()
{
return NSCAP_REINTERPRET_CAST(T**, begin_assignment());
}
};
template <class T>
class nsGetterAddRefs
/*
...
This class is designed to be used for anonymous temporary objects in the
argument list of calls that return COM interface pointers, e.g.,
nsCOMPtr<IFoo> fooP;
...->QueryInterface(iid, nsGetterAddRefs<IFoo>(fooP))
...->QueryInterface(iid, getter_AddRefs(fooP))
When initialized with a |nsCOMPtr|, as in the example above, it returns
a |void**| (or |T**| if needed) that the outer call (|QueryInterface| in this
case) can fill in. When this temporary object goes out of scope, just after
the call returns, its destructor assigned the resulting interface pointer, i.e.,
|QueryInterface|s result, into the |nsCOMPtr| it was initialized with.
See also |nsGetterDoesntAddRef|.
*/
{
public:
explicit
nsGetterAddRefs( nsCOMPtr<T>& aSmartPtr )
: mTargetSmartPtr(aSmartPtr)
{
// nothing else to do
}
operator void**()
{
// NS_PRECONDITION(mTargetSmartPtr != 0, "getter_AddRefs into no destination");
return NSCAP_REINTERPRET_CAST(void**, mTargetSmartPtr.StartAssignment());
}
T*&
operator*()
{
// NS_PRECONDITION(mTargetSmartPtr != 0, "getter_AddRefs into no destination");
return *(mTargetSmartPtr.StartAssignment());
}
operator T**()
{
// NS_PRECONDITION(mTargetSmartPtr != 0, "getter_AddRefs into no destination");
return mTargetSmartPtr.StartAssignment();
}
private:
nsCOMPtr<T>& mTargetSmartPtr;
};
template <class T>
inline
nsGetterAddRefs<T>
getter_AddRefs( nsCOMPtr<T>& aSmartPtr )
/*
Used around a |nsCOMPtr| when
...makes the class |nsGetterAddRefs<T>| invisible.
*/
{
return nsGetterAddRefs<T>(aSmartPtr);
}
#endif // !defined(nsCOMPtr_h___)