I need a collection in which i can store heap-allocated objects having virtual functions.
I known about boost::shared_ptr, std::unique_ptr (C++11) and boost::ptr_(vector|list|map), but they doesn't solve duplicate pointer problem.
Just to describe a problem - i have a function which accepts heap-allocated pointer and stores it for future use:
void SomeClass::add(T* ptr)
{
_list.push_back(ptr);
}
But if i call add twice with same parameter ptr - _list will contain two pointers to same object and when _list is destructed multiple deletion of same object will occur.
If _list will count pointer which he stores and uses them at deletion time then this problem will be solved and objects will not be deleted multiple times.
So the question is:
Does somebody knows some library with collections (vector,list,map in essence) of pointer with auto-delete on destruction and support of reference counting?
Or maybe i can solve this problem using some other technique?
Update:
I need support of duplicate pointers. So i can't use std::set.
As Kerrek SB and Grizzly mentioned - it is a bad idea to use raw pointers in general and suggests to use std::make_shared and forget about instantiation via new. But this is responsibility of client-side code - not the class which i designs. Even if i change add signature (and _list container of course) to
void SomeClass::add(std::shared_ptr<T> ptr)
{
_list.push_back(ptr);
}
then somebody (who doesn't know about std::make_shared) still can write this:
SomeClass instance;
T* ptr = new T();
instance.add(ptr);
instance.add(ptr);
So this is not a full solution which i wait, but useful if you write code alone.
Update 2:
As an alternative solution i found a clonning (using generated copy constructor). I mean that i can change my add function like this:
template <typename R>
void SomeClass::add(const R& ref)
{
_list.push_back(new R(ref));
}
this will allow virtual method (R - class which extends some base class (interface)) calls and disallow duplicate pointers. But this solution has an overhead for clone.
Yes: std::list<std::shared_ptr<T>>.
The shared pointer is avaiable from <memory>, or on older platforms from <tr1/memory>, or from Boost's <boost/shared_ptr.hpp>. You won't need to delete anything manually, as the shared pointer takes care of this itself. You will however need to keep all your heap pointers inside a shared pointer right from the start:
std::shared_ptr<T> p(new T); // legacy
auto p = std::make_shared<T>(); // better
If you another shared pointer to the same object, make a copy of the shared pointer (rather than construct a new shared pointer from the underlying raw pointer): auto q = p;
The moral here is: If you're using naked pointers, something is wrong.
Realize that smart pointers are compared by comparing the underlying container. So you can just use a std::set of whatever smartpointer you prefer. Personally I use std::unique_ptr over shared_ptr, whenever I can get away with it, since it makes the ownership much clearer (whoever holds the unique_ptris the owner) and has much lower overhead too. I have found that this is enough for almost all my code. The code would look something like the following:
std::set<std::unique_ptr<T> > _list;
void SomeClass::add(T* ptr)
{
std::unique_ptr<T> p(ptr);
auto iter = _list.find(p);
if(iter == _list.end())
_list.insert(std::move(p));
else
p.release();
}
I'm not sure right now if that is overkill (have to check if insert is guaranteed not to do anything, if the insertion fails), but it should work. Doing this with shared_ptr<T> would look similar, although be a bit more complex, due to the lack of a relase member. In that case I would probably first construct a shared_ptr<T> with a do nothing deleter too pass to the call to find and then another shared_ptr<T> which is actually inserted.
Of course personally I would avoid doing this and always pass around smart pointers when the ownership of a pointer changes hands. Therefore I would rewrite SomeClass::add as void SomeClass::add(std::unique_ptr<T> ptr) or void SomeClass::add(std::shared_ptr<T> ptr) which would pretty much solve the problem of having multiple instances anyways (as long as the pointer is always wrapped).
Related
Shared pointers are good idea, no doubt. But as long as a large scale program includes raw pointers, I think there is a big risk in using shared pointers. Mainly, you will loose control of the real life-cycle of pointers to objects that hold raw pointers, and bugs will occur in locations which are more difficult to find and debug.
So my question is, was there no attempt to add to modern c++ a "weak pointer" which does not depend on using shared pointers? I mean just a pointer which becomes NULL when deleted in any part of the program. Is there a reason not to use such a self-made wrapper?
To better explain what I mean, the following is such a "weak pointer" that I made. I named it WatchedPtr.
#include <memory>
#include <iostream>
template <typename T>
class WatchedPtr {
public:
// the only way to allocate new pointer
template <typename... ARGS>
WatchedPtr(ARGS... args) : _ptr (new T(args...)), _allocated (std::make_shared<bool>(true)) {}
WatchedPtr(const WatchedPtr<T>& other) : _ptr (other._ptr), _allocated (other._allocated) {}
// delete the pointer
void del () {delete _ptr; *_allocated = false;}
auto& operator=(const WatchedPtr<T> &other) { return *this = other; }
bool isNull() const { return *_allocated; }
T* operator->() const { return _ptr; }
T& operator*() const { return *_ptr; }
private:
T* _ptr;
std::shared_ptr <bool> _allocated;
};
struct S {
int a = 1;
};
int main () {
WatchedPtr<S> p1;
WatchedPtr<S> p2(p1);
p1->a = 8;
std::cout << p1.isNull () << std::endl;
std::cout << p2.isNull () << std::endl;
p2.del ();
std::cout << p1.isNull () << std::endl;
std::cout << p1.isNull () << std::endl;
return 0;
}
Result:
1
1
0
0
-Edited-
Thank you all. Some clarifications following the comments and answers so far:
The implementation I presented for WatchedPtr is merely to demonstrate what I mean: a pointer that does not get the copy from external allocation, cannot be deleted externally, and becomes null if it is deleted. The implementation is knowingly far from perfect and was not meant to be perfect.
Problem with mix of shared_ptr and raw pointers is very common: A* is held as raw pointer, thus created at some point of the program and explicitly deleted at some point of the program. B holds a A*, and B* is held as shared_ptr, thus B* has vague lifespan. Thus B may live long after the deletion of A* that B holds.
The main usage of "WatchedPtr" in my mind is defensive programing. i.e. check for null and do the best thing possible for continuity (and a debug error). shared_ptr can do it, but in a very dangerous way - it will hide and delay the problem.
There can also be a design usage for "WatchedPtr" (very few and explicit "owners"), but this is not the main idea. For that indeed shared pointers are doing the job.
The intention of "WatchedPtr" is not for replacing all existing raw pointers in the program at once. It is not the same effort as replacing to shared_ptr, which IMHO has be done for the whole program at once. Which is unrealistic for large scale programs.
Weak pointers rely on notifications from the smart pointer infrastructure, so you could never do this with actual raw pointers.
One could imagine an extension of, say, unique_ptr which supported weak pointers, certainly. Presumably the main reason that nobody rushed in to implement such a feature is that weak pointers are already at the "Use refcounting and everything should just work" end of the scale, while unique_ptr is at the "Manage your lifetimes through RAII or you're not a real C++ programmer" end of the scale. Weak pointers also require there to be a separate control block per allocation, meaning that the performance advantage of such a WatchedPtr would be minimal compared to shared_ptr.
I think there is a big risk in using shared pointers. Mainly, you will loose control of the real life-cycle of pointers to objects that hold raw pointers, and bugs will occur in locations which are more difficult to find and debug.
Then you say
just a pointer which becomes NULL when deleted in any part of the program.
Don't you see the contradiction?
You don't want to use shared pointer because the lifetime of objects are determined at runtime. So far so good.
However, you want a pointer that automatically becomes null when the owner deletes it. The problem is if the lifetime of your pointer is known, you should not need that at all! If you know when the lifetime of your pointer ends, then you should be able to remove all instances of that pointer, a have a mean to check if the pointer is dead.
If you have a pointer that you don't know when the owner will free it and have no way to check or no observable side effect for the point of view of the weak owner, then do you really have control over lifetime of your pointer? Not really.
In fact, your implementation rely on containing a shared pointer. This is enlightening in the sense that you need some form of shared ownership in order to implement a raw pointer that can have weak pointer to it. Then if you need shared ownership to implement a raw pointer with weak references, you are left with a shared pointer. That's why the existence of your proposed class is contradictory.
std::shared_ptr + std::weak_ptr is made to deal with the issue of "parts of your program don't know when the owner free the resouce". What you need is a single std::shared_ptr and multiple std::weak_ptr, so they know when the resource is freed. These classes have the infomation needed to check the lifetime of a variable at runtime.
Or if in the contrary you know the lifetime of your pointers, then use that knowledge and find a way to remove dangling pointers, or expose a way to check for dangling pointers.
Reading the answers and comments, along with C++ Core Guidelines by Bjarne Stroustrup & Herb Sutter, I have come to the following answer:
When following the guidelines, there is no need for a "WatchedPtr" which involves "new" and "delete". However, a way to track the validity of a raw pointer taken from a smart pointer, is still in question for me, for debug/QA purposes.
In details:
Raw pointers should continue to be used. For various reasons. However, explicit "new" and "delete" should not. The cases of calling "new" and "delete" should all be replaced by shared_ptr/unique_ptr.
At the place where a raw pointer is currently allocated, there is no point in replacing it by "WatchedPtr".
If replacing a raw pointer to something else where it is allocated, it will be in most cases to unique_ptr, and on the other cases to shared_ptr. The "WatchedPtr", if at all, will continue from that point, built from the shared/unique pointer.
Therefor I have posted a somewhat different question.
Currently I'm using some functions from the glib library. With glib also comes the gio. glib is a C library and therefore I need to delete some structures that I create.
for many of the objects I create a smartpointer eg:
std::shared_ptr<GAsyncQueue> my_queue = std::shared_ptr<GAsyncQueue>(g_async_queue_create(), g_async_queue_unref);
For this creates a shared pointer to an GAsyncQueue and this is safely destroys the queue on its end of its life.
However, I encounter a problem when I obtain a pointer from the gio library that I should not free. In the following code my_connection is a GSocketClient which implements (in glib speak) GIOStream.
std::shared_ptr<GInputStream> my_input_stream =
std::shared_ptr<GInputStream> (
g_io_stream_get_input_stream(G_IO_STREAM(my_connection.get()))
);
Because the documentation on GIOStream mentions, that the pointer obtained with g_io_stream_get_input_stream() should not be freed. That is because it is owned by the my_connection instance.
I thought about creating a lamda for the destroy object, the second parameter of a shared pointer object. eg auto deleter = [](GInputStream* ptr) {}; and then give that lambda as destroy function to the shared pointer, but that feels a kind of stupid.
Well, alternative to no-op deleter might be using aliasing shared pointer
template <class U> shared_ptr (const shared_ptr<U>& x, element_type* p) noexcept;
It shares x, but after get() you'll get back p.
Discussion: What is shared_ptr's aliasing constructor for?
You probably just don't need a std::shared_ptr. And you probably don't even need a pointer.
As I read your question and comments, I don't see any point against
auto& my_input_stream = *( g_io_stream_get_input_stream(G_IO_STREAM(my_connection.get())) )
It is true that pointers allow optional data. However, it's also true that it's mostly used the wrong way. Having
void foo( type* ptr)
{
if (!ptr)
throw exception;
}
often doesn't make sense. If the function has to to work on concrete data, allowing a NULL parameter is only useful if you then worry about providing that data. Otherwise, just require a reference (possibly const) to the object.
Smart pointers are useful; but they're still pointers. Avoiding them altogether, if possible, is even better.
From the comments:
However, a reference must always be initialized
Absolutely. Since C++11 though we've got std::reference_wrapper which can also be reassinged and stored in containers.
You can use a deleter type that does nothing, but it will need to be passed as an argument to the shared_ptr's constructor
struct DoNothing {
template <typename T>
void operator()(T*) const noexcept { }
};
When creating a shared_ptr you will need to create one of these deleters and pass it in the constructor (as you're doing with the lambda). You can make this easier on yourself with an intermediate function
template <typename T>
std::shared_ptr<T> non_deleting_shared_ptr(T* ptr) {
return {ptr, DoNothing};
}
auto my_input_stream =
non_deleting_shared_ptr(
g_io_stream_get_input_stream(G_IO_STREAM(my_connection.get()));
However the bigger question is why you're using smart pointers when you don't want ownership to be a part of it. You'd almost certainly be better off with just a GAsyncQueue*, unless of course you're in a situation where you have a shared_ptr that needs to free sometimes. Like a data member maybe?
I have some code that currently uses raw pointers, and I want to change to smart pointers. This helps cleanup the code in various ways. Anyway, I have factory methods that return objects and its the caller's responsibility to manager them. Ownership isn't shared and so I figure unique_ptr would be suitable. The objects I return generally all derive from a single base class, Object.
For example,
class Object { ... };
class Number : public Object { ... };
class String : public Object { ... };
std::unique_ptr<Number> State::NewNumber(double value)
{
return std::unique_ptr<Number>(new Number(this, value));
}
std::unique_ptr<String> State::NewString(const char* value)
{
return std::unique_ptr<String>(new String(this, value));
}
The objects returned quite often need to be passed to another function, which operates on objects of type Object (the base class). Without any smart pointers the code is like this.
void Push(const Object* object) { ... } // push simply pushes the value contained by object onto a stack, which makes a copy of the value
Number* number = NewNumber(5);
Push(number);
When converting this code to use unique_ptrs I've run into issues with polymorphism. Initially I decided to simply change the definition of Push to use unique_ptrs too, but this generates compile errors when trying to use derived types. I could allocate objects as the base type, like
std::unique_ptr<Object> number = NewNumber(5);
and pass those to Push - which of course works. However I often need to call methods on the derived type. In the end I decided to make Push operate on a pointer to the object stored by the unique_ptr.
void Push(const Object* object) { ... }
std::unique_ptr<Object> number = NewNumber(5);
Push(number.get());
Now, to the reason for posting. I'm wanting to know if this is the normal way to solve the problem I had? Is it better to have Push operate on the unique_ptr vs the object itself? If so how does one solve the polymorphism issues? I would assume that simply casting the ptrs wouldn't work. Is it common to need to get the underlying pointer from a smart pointer?
Thanks, sorry if the question isn't clear (just let me know).
edit: I think my Push function was a bit ambiguous. It makes a copy of the underlying value and doesn't actually modify, nor store, the input object.
Initially I decided to simply change the definition of Push to use
unique_ptrs too, but this generates compile errors when trying to use
derived types.
You likely did not correctly deal with uniqueness.
void push(std::unique_ptr<int>);
int main() {
std::unique_ptr<int> i;
push(i); // Illegal: tries to copy i.
}
If this compiled, it would trivially break the invariant of unique_ptr, that only one unique_ptr owns an object, because both i and the local argument in push would own that int, so it is illegal. unique_ptr is move only, it's not copyable. It has nothing to do with derived to base conversion, which unique_ptr handles completely correctly.
If push owns the object, then use std::move to move it there. If it doesn't, then use a raw pointer or reference, because that's what you use for a non-owning alias.
Well, if your functions operate on the (pointed to) object itself and don't need its address, neither take any ownership, and, as I guess, always need a valid object (fail when passed a nullptr), why do they take pointers at all?
Do it properly and make them take references:
void Push(const Object& object) { ... }
Then the calling code looks exactly the same for raw and smart pointers:
auto number = NewNumber(5);
Push(*number);
EDIT: But of course no matter if using references or pointers, don't make Push take a std::unique_ptr if it doesn't take ownership of the passed object (which would make it steal the ownership from the passed pointer). Or in general don't use owning pointers when the pointed to object is not to be owned, std::shared_ptr isn't anything different in this regard and is as worse a choice as a std::unique_ptr for Push's parameter if there is no ownership to be taken by Push.
If Push does not take owenrship, it should probably take reference instead of pointer. And most probably a const one. So you'll have
Push(*number);
Now that's obviously only valid if Push isn't going to keep the pointer anywhere past it's return. If it does I suspect you should try to rethink the ownership first.
Here's a polymorphism example using unique pointer:
vector<unique_ptr<ICreature>> creatures;
creatures.emplace_back(new Human);
creatures.emplace_back(new Fish);
unique_ptr<vector<string>> pLog(new vector<string>());
for each (auto& creature in creatures)
{
auto state = creature->Move(*pLog);
}
I have a set of shared pointers:
std::set<boost::shared_ptr<T>> set;
And a pointer:
T* p;
I would like to efficiently remove the element of set equal to p, but I can't do this with any of the members of set, or any of the standard algorithms, since T* is a completely different type to boost::shared_ptr<T>.
A few approaches I can think of are:
somehow constructing a new shared_ptr from the pointer that won't take ownership of the pointed to memory (ideal solution, but I can't see how to do this)
wrapping / re-implementing shared_ptr so that I can do the above
just doing my own binary search over the set
Construct a shared_ptr<T> from T with a null_deleter (see boost:::shared_ptr FAQ).
struct null_deleter {
void operator()(void const *) const { }
};
size_t remove_ptr_from_set(std::set<boost::shared_ptr<T>> &set, X* x)
{
shared_ptr<X> px(x, null_deleter());
return set.erase(px);
}
That way the types are compatible and you don't have to worry about your temporary shared_ptr deleting any object.
Or, as one of the comments say, if you can change T to inherit from enable_shared_from_this you could get a correct shared ptr from your object.
If the reason for using the set is that you need to efficiently find pointers of type T, then the obvious answer is not to make it a set of shared pointers! Instead, wrap the set in a class which manages the lifetimes of the pointers that the set contains.
You can use boost::ptr_set if you want the set to have ownership of the objects, or boost::reference_wrapper if you just want the set to store references to them. If you use shared_ptr in one place in your code, you will have to use it in all places, or risk terrible crashes (dangling pointers, already deleted objects etc.). The exception is weak_ptr, a pointer that points to an object held by a shared_ptr but does not share ownership.
What are some ways you can shoot yourself in the foot when using boost::shared_ptr? In other words, what pitfalls do I have to avoid when I use boost::shared_ptr?
Cyclic references: a shared_ptr<> to something that has a shared_ptr<> to the original object. You can use weak_ptr<> to break this cycle, of course.
I add the following as an example of what I am talking about in the comments.
class node : public enable_shared_from_this<node> {
public :
void set_parent(shared_ptr<node> parent) { parent_ = parent; }
void add_child(shared_ptr<node> child) {
children_.push_back(child);
child->set_parent(shared_from_this());
}
void frob() {
do_frob();
if (parent_) parent_->frob();
}
private :
void do_frob();
shared_ptr<node> parent_;
vector< shared_ptr<node> > children_;
};
In this example, you have a tree of nodes, each of which holds a pointer to its parent. The frob() member function, for whatever reason, ripples upwards through the tree. (This is not entirely outlandish; some GUI frameworks work this way).
The problem is that, if you lose reference to the topmost node, then the topmost node still holds strong references to its children, and all its children also hold a strong reference to their parents. This means that there are circular references keeping all the instances from cleaning themselves up, while there is no way of actually reaching the tree from the code, this memory leaks.
class node : public enable_shared_from_this<node> {
public :
void set_parent(shared_ptr<node> parent) { parent_ = parent; }
void add_child(shared_ptr<node> child) {
children_.push_back(child);
child->set_parent(shared_from_this());
}
void frob() {
do_frob();
shared_ptr<node> parent = parent_.lock(); // Note: parent_.lock()
if (parent) parent->frob();
}
private :
void do_frob();
weak_ptr<node> parent_; // Note: now a weak_ptr<>
vector< shared_ptr<node> > children_;
};
Here, the parent node has been replaced by a weak pointer. It no longer has a say in the lifetime of the node to which it refers. Thus, if the topmost node goes out of scope as in the previous example, then while it holds strong references to its children, its children don't hold strong references to their parents. Thus there are no strong references to the object, and it cleans itself up. In turn, this causes the children to lose their one strong reference, which causes them to clean up, and so on. In short, this wont leak. And just by strategically replacing a shared_ptr<> with a weak_ptr<>.
Note: The above applies equally to std::shared_ptr<> and std::weak_ptr<> as it does to boost::shared_ptr<> and boost::weak_ptr<>.
Creating multiple unrelated shared_ptr's to the same object:
#include <stdio.h>
#include "boost/shared_ptr.hpp"
class foo
{
public:
foo() { printf( "foo()\n"); }
~foo() { printf( "~foo()\n"); }
};
typedef boost::shared_ptr<foo> pFoo_t;
void doSomething( pFoo_t p)
{
printf( "doing something...\n");
}
void doSomethingElse( pFoo_t p)
{
printf( "doing something else...\n");
}
int main() {
foo* pFoo = new foo;
doSomething( pFoo_t( pFoo));
doSomethingElse( pFoo_t( pFoo));
return 0;
}
Constructing an anonymous temporary shared pointer, for instance inside the arguments to a function call:
f(shared_ptr<Foo>(new Foo()), g());
This is because it is permissible for the new Foo() to be executed, then g() called, and g() to throw an exception, without the shared_ptr ever being set up, so the shared_ptr does not have a chance to clean up the Foo object.
Be careful making two pointers to the same object.
boost::shared_ptr<Base> b( new Derived() );
{
boost::shared_ptr<Derived> d( b.get() );
} // d goes out of scope here, deletes pointer
b->doSomething(); // crashes
instead use this
boost::shared_ptr<Base> b( new Derived() );
{
boost::shared_ptr<Derived> d =
boost::dynamic_pointer_cast<Derived,Base>( b );
} // d goes out of scope here, refcount--
b->doSomething(); // no crash
Also, any classes holding shared_ptrs should define copy constructors and assignment operators.
Don't try to use shared_from_this() in the constructor--it won't work. Instead create a static method to create the class and have it return a shared_ptr.
I've passed references to shared_ptrs without trouble. Just make sure it's copied before it's saved (i.e., no references as class members).
Here are two things to avoid:
Calling the get() function to get the raw pointer and use it after the pointed-to object goes out of scope.
Passing a reference of or a raw pointer to a shared_ptr should be dangerous too, since it won't increment the internal count which helps keep the object alive.
We debug several weeks strange behavior.
The reason was:
we passed 'this' to some thread workers instead of 'shared_from_this'.
Not precisely a footgun, but certainly a source of frustration until you wrap your head around how to do it the C++0x way: most of the predicates you know and love from <functional> don't play nicely with shared_ptr. Happily, std::tr1::mem_fn works with objects, pointers and shared_ptrs, replacing std::mem_fun, but if you want to use std::negate, std::not1, std::plus or any of those old friends with shared_ptr, be prepared to get cozy with std::tr1::bind and probably argument placeholders as well. In practice this is actually a lot more generic, since now you basically end up using bind for every function object adaptor, but it does take some getting used to if you're already familiar with the STL's convenience functions.
This DDJ article touches on the subject, with lots of example code. I also blogged about it a few years ago when I first had to figure out how to do it.
Using shared_ptr for really small objects (like char short) could be an overhead if you have a lot of small objects on heap but they are not really "shared". boost::shared_ptr allocates 16 bytes for every new reference count it creates on g++ 4.4.3 and VS2008 with Boost 1.42. std::tr1::shared_ptr allocates 20 bytes. Now if you have a million distinct shared_ptr<char> that means 20 million bytes of your memory are gone in holding just count=1. Not to mention the indirection costs and memory fragmentation. Try with the following on your favorite platform.
void * operator new (size_t size) {
std::cout << "size = " << size << std::endl;
void *ptr = malloc(size);
if(!ptr) throw std::bad_alloc();
return ptr;
}
void operator delete (void *p) {
free(p);
}
Giving out a shared_ptr< T > to this inside a class definition is also dangerous.
Use enabled_shared_from_this instead.
See the following post here
You need to be careful when you use shared_ptr in multithread code. It's then relatively easy to become into a case when couple of shared_ptrs, pointing to the same memory, is used by different threads.
The popular widespread use of shared_ptr will almost inevitably cause unwanted and unseen memory occupation.
Cyclic references are a well known cause and some of them can be indirect and difficult to spot especially in complex code that is worked on by more than one programmer; a programmer may decide than one object needs a reference to another as a quick fix and doesn't have time to examine all the code to see if he is closing a cycle. This hazard is hugely underestimated.
Less well understood is the problem of unreleased references. If an object is shared out to many shared_ptrs then it will not be destroyed until every one of them is zeroed or goes out of scope. It is very easy to overlook one of these references and end up with objects lurking unseen in memory that you thought you had finished with.
Although strictly speaking these are not memory leaks (it will all be released before the program exits) they are just as harmful and harder to detect.
These problems are the consequences of expedient false declarations: 1. Declaring what you really want to be single ownership as shared_ptr. scoped_ptr would be correct but then any other reference to that object will have to be a raw pointer, which could be left dangling. 2. Declaring what you really want to be a passive observing reference as shared_ptr. weak_ptr would be correct but then you have the hassle of converting it to share_ptr every time you want to use it.
I suspect that your project is a fine example of the kind of trouble that this practice can get you into.
If you have a memory intensive application you really need single ownership so that your design can explicitly control object lifetimes.
With single ownership opObject=NULL; will definitely delete the object and it will do it now.
With shared ownership spObject=NULL; ........who knows?......
If you have a registry of the shared objects (a list of all active instances, for example), the objects will never be freed. Solution: as in the case of circular dependency structures (see Kaz Dragon's answer), use weak_ptr as appropriate.
Smart pointers are not for everything, and raw pointers cannot be eliminated
Probably the worst danger is that since shared_ptr is a useful tool, people will start to put it every where. Since plain pointers can be misused, the same people will hunt raw pointers and try to replace them with strings, containers or smart pointers even when it makes no sense. Legitimate uses of raw pointers will become suspect. There will be a pointer police.
This is not only probably the worst danger, it may be the only serious danger. All the worst abuses of shared_ptr will be the direct consequence of the idea that smart pointers are superior to raw pointer (whatever that means), and that putting smart pointers everywhere will make C++ programming "safer".
Of course the mere fact that a smart pointer needs to be converted to a raw pointer to be used refutes this claim of the smart pointer cult, but the fact that the raw pointer access is "implicit" in operator*, operator-> (or explicit in get()), but not implicit in an implicit conversion, is enough to give the impression that this is not really a conversion, and that the raw pointer produced by this non-conversion is an harmless temporary.
C++ cannot be made a "safe language", and no useful subset of C++ is "safe"
Of course the pursuit of a safe subset ("safe" in the strict sense of "memory safe", as LISP, Haskell, Java...) of C++ is doomed to be endless and unsatisfying, as the safe subset of C++ is tiny and almost useless, as unsafe primitives are the rule rather than the exception. Strict memory safety in C++ would mean no pointers and only references with automatic storage class. But in a language where the programmer is trusted by definition, some people will insist on using some (in principle) idiot-proof "smart pointer", even where there is no other advantage over raw pointers that one specific way to screw the program state is avoided.