C++Talk.NET C++ language newsgroups   Archives   FAQ   Search   Memberlist   Usergroups   Register   Profile   Log in to check your private messages   Log in  Question on returning full copies on stack and their referen          C++Talk.NET Forum Index -> C++ Language (Moderated) View previous topic :: View next topic   Author Message AB Guest Posted: Fri Oct 03, 2003 11:42 am    Post subject: Question on returning full copies on stack and their referen Hello, I have already googled through this group but could not find any response to the following scenario. While in a method, a const reference is created for a full copy of an object that was returned on the stack. Is this a valid thing to do or not? Here is some sample code to illustrate the issue: #include <iostream> #include <string> #include <vector> class X { public: vector<int> getList() { return(vecX); } vector<int> &getList2() { return(vecX); } private: vector<int> vecX; }; void fn1(X &xinst) { xinst.getList2().push_back(1); xinst.getList2().push_back(2); const vector<int> &x2=xinst.getList(); for (int i=0; i < x2.size(); ++i) { cout << "x " << x2[i] << endl; } } int main(void) { X *xinst=new X(); fn1(*xinst); return 0; } The code seems to work fine but the question is, is the constant reference "x2" pointing to anything valid after calling xinst.getList() or is it just luck of the draw? Regards, AB. [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Ron Guest Posted: Sat Oct 04, 2003 8:36 am    Post subject: Re: Question on returning full copies on stack and their ref Quote: I have already googled through this group but could not find any response to the following scenario. While in a method, a const reference is created for a full copy of an object that was returned on the stack. Is this a valid thing to do or not? Here is some sample code to illustrate the issue: #include <iostream #include #include class X { public: vector vector<int> &getList2() { return(vecX); } private: vector<int> vecX; }; void fn1(X &xinst) { xinst.getList2().push_back(1); xinst.getList2().push_back(2); const vector<int> &x2=xinst.getList(); for (int i=0; i < x2.size(); ++i) { cout << "x " << x2[i] << endl; } } ... The code seems to work fine but the question is, is the constant reference "x2" pointing to anything valid after calling xinst.getList() or is it just luck of the draw? Your code is OK, but I wouldn't make a habit of it. Ordinarily, temporaries are destroyed at the end of the full-expression that creates them. s.12.2(3). However, when you bind a reference to a temporary, the temporary's lifetime generally is extended to match the reference's lifetime. s.12.2(5). This occurs when binding a reference to a function's return value; s.12.2(5) says, "A temporary bound to the returned value in a function return statement (6.6.3) persists until the function exits." This is a little inartfully phrased, but s.12.2(5)'s example clarifies it: ------ class C { // ... public: C(); C(int); friend C operator+(const C&, const C&); ~C(); }; C obj1; const C& cr = C(16)+C(23); C obj2; .... The temporary T3 bound to the reference cr is destroyed at the end of cr's lifetime, that is, at the end of the program. ------ -Ron [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Joshua Lehrer Guest Posted: Sat Oct 04, 2003 8:38 am    Post subject: Re: Question on returning full copies on stack and their ref [email]alb09439 (AT) hotmail (DOT) com[/email] (AB) wrote in message news:<ffcb4c91.0310021547.730dcb86 (AT) posting (DOT) google.com>... Quote: Hello, I have already googled through this group but could not find any response to the following scenario. While in a method, a const reference is created for a full copy of an object that was returned on the stack. Is this a valid thing to do or not? This initial value of a refernce to non-const must be an lvalue. T func(); const T& obj(func()); //ok const T& obj=func(); //ok, same as above const T& obj=T(1,2,3); //ok T& obj=func(); //not ok T& obj=T(1,2,3); //not ok Temporary objects returned, either from function calls or by calling constructors directly, may be bound to const references, but not to non-const references. The lifetime of the temporary object is guaranteed to be longer than the lifetime of the reference to which it is bound. joshua lehrer factset research systems NYSE:FDS [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Le Chaud Lapin Guest Posted: Sat Oct 04, 2003 8:39 am    Post subject: Re: Question on returning full copies on stack and their ref [email]alb09439 (AT) hotmail (DOT) com[/email] (AB) wrote in message news:<ffcb4c91.0310021547.730dcb86 (AT) posting (DOT) google.com>... Quote: class X { public: vector<int> getList() { return(vecX); } vector<int> &getList2() { return(vecX); } private: vector<int> vecX; }; void fn1(X &xinst) { xinst.getList2().push_back(1); xinst.getList2().push_back(2); const vector<int> &x2=xinst.getList(); for (int i=0; i < x2.size(); ++i) { cout << "x " << x2[i] << endl; } } The code seems to work fine but the question is, is the constant reference "x2" pointing to anything valid after calling xinst.getList() or is it just luck of the draw? Yes. It is best to think of xinst.getList() "yielding" an object, and that object getting slammed into another, "nameless" object of equal size [sizeof(X)] occupying some space specifically reserved for it on the stack. In this case, x2 happens to refer to that "nameless" stack object. -Chaud Lapin- [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Brian McKeever Guest Posted: Sat Oct 04, 2003 1:53 pm    Post subject: Re: Question on returning full copies on stack and their ref [email]alb09439 (AT) hotmail (DOT) com[/email] (AB) wrote in message news:<ffcb4c91.0310021547.730dcb86 (AT) posting (DOT) google.com>... [snip] Quote: const vector<int> &x2=xinst.getList(); for (int i=0; i < x2.size(); ++i) { cout << "x " << x2[i] << endl; } [snip] The code seems to work fine but the question is, is the constant reference "x2" pointing to anything valid after calling xinst.getList() or is it just luck of the draw? Yes, the code is fine. getList() returns a copy of the member variable. The temporary that is returned is bound to the reference, which extends its lifetime. If you had just called getList() and not done anything with the returned vector, the temporary would have been destroyed immediately. It's a special case to make it work the way you think it should. Brian [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Francis Glassborow Guest Posted: Sun Oct 05, 2003 12:40 am    Post subject: Re: Question on returning full copies on stack and their ref In article <c668c1f8.0310030940.12390bb4 (AT) posting (DOT) google.com>, Ron <rcrane (AT) ictv (DOT) com> writes Quote: Your code is OK, but I wouldn't make a habit of it. Ordinarily, temporaries are destroyed at the end of the full-expression that creates them. s.12.2(3). However, when you bind a reference to a temporary, the temporary's lifetime generally is extended to match the reference's lifetime. s.12.2(5). Note that you cannot bind a temporary to a reference but only to a const reference. -- Francis Glassborow ACCU If you are not using up-to-date virus protection you should not be reading this. Viruses do not just hurt the infected but the whole community. [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Maciej Sobczak Guest Posted: Mon Oct 06, 2003 12:50 pm    Post subject: Re: Question on returning full copies on stack and their ref Hi, Joshua Lehrer wrote: Quote: The lifetime of the temporary object is guaranteed to be longer than the lifetime of the reference to which it is bound. Of course, we have to be aware of the following gotcha: struct A {}; struct B { B(const A &a) : ref(a) {} const A &ref; }; A fun() { return A(); } int main() { B b(fun()); // b.ref is dangling here } Here, the temporary object A() is bound to the const reference in the B's constructor, *but no longer*. In other words, after the B's constructor finishes, the temporary can be destroyed and from that time the b's ref member is a dangling reference. So - carefully with refs to temps. -- Maciej Sobczak [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top AB Guest Posted: Tue Oct 07, 2003 6:13 pm    Post subject: Re: Question on returning full copies on stack and their ref Maciej Sobczak <maciej (AT) maciejsobczak (DOT) com> wrote Quote: Hi, Joshua Lehrer wrote: The lifetime of the temporary object is guaranteed to be longer than the lifetime of the reference to which it is bound. Of course, we have to be aware of the following gotcha: struct A {}; struct B { B(const A &a) : ref(a) {} const A &ref; }; A fun() { return A(); } int main() { B b(fun()); // b.ref is dangling here } Here, the temporary object A() is bound to the const reference in the B's constructor, *but no longer*. In other words, after the B's constructor finishes, the temporary can be destroyed and from that time the b's ref member is a dangling reference. So - carefully with refs to temps. Hello again, How about the following example based on your example above(which is where the original problem occurred): class C { public: C() { ax=new(fun()); } A getA() { return(*ax); } ~C() { delete ax; } private: A *ax; }; /* class C */ int main() { C *cx=new C(); B bx(cx->getA()); // do some stuff with bx here using its (const A& ref) delete cx; } Does this also cause the dangling reference issue in B that you mentioned in your previous example? Regards, AB. [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Ron Natalie Guest Posted: Tue Oct 07, 2003 7:14 pm    Post subject: Re: Question on returning full copies on stack and their ref "AB" <alb09439 (AT) hotmail (DOT) com> wrote Quote: Does this also cause the dangling reference issue in B that you mentioned in your previous example? No, as long as the A object is still existant that's fine. Of course your code is fraught with peril because bx lives longer than *cx. If you declared the C variable locally it would make it safer: int main() { C cx; B bx(cx->getA()); } bx will be destructed before cx. By the way, if you use the example code, make sure you implement the copy constructor and copy-assignment operator for your C class, because if they ever get invoked, your class will get trashed. [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Glen Low Guest Posted: Sat Nov 29, 2003 4:50 pm    Post subject: Re: Question on returning full copies on stack and their ref Quote: Of course, we have to be aware of the following gotcha: struct A {}; struct B { B(const A &a) : ref(a) {} const A &ref; }; A fun() { return A(); } int main() { B b(fun()); // b.ref is dangling here } Here, the temporary object A() is bound to the const reference in the B's constructor, *but no longer*. In other words, after the B's constructor finishes, the temporary can be destroyed and from that time the b's ref member is a dangling reference. Minor pedantry: it should be "after the statement B b(fun()); completes, the temporary can be destroyed." In your case it is the same as "after the constructor finishes", but you can also do: B (fun()).something (); and the temporary should last until the something completes. According to 12.2/5, there are two possibilities: "a temporary bound to a reference member in a constructor's ctor-initializer (class.base.init) persists until the constructor exits" OR "a temporary bound to a reference parameter in a function call (expr.call) persists until the completion of the full expression containing the call" However in a previous thread discussing this, the consensus was the first applies to temporaries created within a ctor-initializer, so the first statement does not apply. Therefore the second must apply, since a constructor is also a function call. Some empirical tests on gcc seem to confirm this, but I have yet to experiment in detail with other compilers. Cheers, Glen Low, Pixelglow Software www.pixelglow.com [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Display posts from previous: All Posts1 Day7 Days2 Weeks1 Month3 Months6 Months1 Year Oldest FirstNewest First         C++Talk.NET Forum Index -> C++ Language (Moderated) All times are GMT Page 1 of 1   Jump to: Select a forum C/C++----------------C LanguageC Language (Moderated)C++ Language (Moderated)C++ language (comp.lang.c++)C++ language, library and standards Non-English C/C++ Newsgroups----------------C++ (German)C++ (French)  You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum Powered by phpBB © 2001, 2006 phpBB Group SEO toolkit © 2004-2006 webmedic.