C++Talk.NET C++ language newsgroups   Archives   FAQ   Search   Memberlist   Usergroups   Register   Profile   Log in to check your private messages   Log in  strange behaviour of STL destructor Goto page 1, 2  Next          C++Talk.NET Forum Index -> C++ Language (Moderated) View previous topic :: View next topic   Author Message Mookie Guest Posted: Wed Sep 24, 2003 10:34 pm    Post subject: strange behaviour of STL destructor Hi group, I'm in need of some help. I don't understand why my application is crashing. I boiled down my source code and now I'm at least able to point my finger at the list destructor ~list(). The following sample code mirrors more or less my code and this is also crashing. The if-condition in my destructor is there on purpose, because in my application there is a library call to free things up and I musn't call this if the pointer is NULL. I'm thankfull if someone could explain to me why the code is crashing, and why the destructor is called so often (myList should be empty, should't it?) #include <list> class Test { public: Test(int a){pt = new int(a);} Test(const Test& rhs) { int tmp = *(rhs.pt); pt = new int(tmp); } ~Test() { cout << "Destructor was called" << endl; if (pt !=0) delete pt; pt = 0; } private: int* pt; }; int main() { list Test a(5); myList.push_back(a); /* If I comment out the following line there will be no crash! */ myList.~list(); return 0; } I'm compiling on HP-UX 10.2, with g++ (gcc version 2.95.2 19991024 (release)) Cheers, Bernd [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Erik Max Francis Guest Posted: Thu Sep 25, 2003 1:17 pm    Post subject: Re: strange behaviour of STL destructor Mookie wrote: Quote: I'm in need of some help. I don't understand why my application is crashing. I boiled down my source code and now I'm at least able to point my finger at the list destructor ~list(). The circumstances under which you want to call a destructor explicitly are extremely rare, and this is a good example why. The list is deleted automatically (since it's an auto variable), and so C++ itself will call the destructor for you. Since you're calling it unnecessarily, that means that it's getting called twice and that results in undefined behavior. Just remove the explicit call to the destructor and it will work fine. Even for objects allocated dynamically with new you shouldn't be calling the _destructor_ explicitly, you should instead be deleting it with delete. -- Erik Max Francis && [email]max (AT) alcyone (DOT) com[/email] && http://www.alcyone.com/max/ __ San Jose, CA, USA && 37 20 N 121 53 W && &tSftDotIotE / I love mankind; it's people I can't stand. __/ Charles Schultz [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Siemel Naran Guest Posted: Thu Sep 25, 2003 1:29 pm    Post subject: Re: strange behaviour of STL destructor "Mookie" <bernd_moeller (AT) hotmail (DOT) com> wrote in message Quote: class Test { public: Test(int a){pt = new int(a);} Test(const Test& rhs) ~Test() Don't forget to write operator= too . Quote: int main() { list<Test> myList; Test a(5); myList.push_back(a); /* If I comment out the following line there will be no crash! */ myList.~list(); return 0; } Your program calls the destructor of myList twice. First time is your explicit call. Second time is at the closing } where the system automatically destroys local objects. Probably the implementation of your destructor that does set m_pRoot to NULL in the destructor, but it's not required too because you're not supposed to use the object again afterwards. There is no reason to call the destructor explicitly. If you really must use placement new to construct a new list right after destroying it. But if you want to erase all elements in the list consider these two options myList.clear(); list<Test>().swap(myList); -- +++++++++++ Siemel Naran [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Heinz Ozwirk Guest Posted: Thu Sep 25, 2003 1:39 pm    Post subject: Re: strange behaviour of STL destructor "Mookie" <bernd_moeller (AT) hotmail (DOT) com> schrieb im Newsbeitrag news:47b8e6cd.0309240604.329bd680 (AT) posting (DOT) google.com... : int main() : { : list<Test> myList; : Test a(5); : myList.push_back(a); : /* If I comment out the following line there will be no crash! */ : myList.~list(); : return 0; : } Don't call destructors explicitly. The compiler will take care of that. In your code, the destructor of myList will be called twice. First you call it explicitly, and at the end of main the compiler creates a second call. And I'm almost certain that it is in the second call your program crashs. So just remove the explicit call and let the compiler do its work. Heinz [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Dhruv Guest Posted: Thu Sep 25, 2003 1:44 pm    Post subject: Re: strange behaviour of STL destructor On Wed, 24 Sep 2003 18:34:22 -0400, Mookie wrote: [snipped some code]....... Quote: int main() { list<Test> myList; Test a(5); myList.push_back(a); /* If I comment out the following line there will be no crash! */ myList.~list(); Simple, then comment it out!!! That's because local variables are destroyed automatically at the end of their scope, namely here when main() ends. What you probably ment is myList.clear(); which is again handled by the destructor. BTW the app causes a memory leak which is not very dangerous here, because the app. is going to exit anyways, and all the memory is going back to the OS. Quote: return 0; } Regards, -Dhruv. [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Gareth Stockwell Guest Posted: Thu Sep 25, 2003 1:44 pm    Post subject: Re: strange behaviour of STL destructor Hi Bernd, You seem to be a bit confused as to the nature of a destructor. You should never have to call a destructor explicitly, using code like myList.~list(). In the case of on-stack objects (of which your myList is an example), the destructor is called automatically when the object goes out of scope - in this case, at the end of the main() function. If, on the other hand, the list had been created on the heap, using new, then calling delete on a pointer to the list would invoke the destructor. Since destructors are designed to be called at the end of an object's life, most are not written in such a way that permits multiple calls, so some unpleasant behaviour can result if this occurs. So, leave that line commented out and your code is fine. I don't think I have ever written code using the idiom 'myObj.~Obj()' - but I would be interested to hear if any others on this list could cite a case in which it would be valid! Gareth [email]bernd_moeller (AT) hotmail (DOT) com[/email] (Mookie) wrote in message news:<47b8e6cd.0309240604.329bd680 (AT) posting (DOT) google.com>... Quote: Hi group, I'm in need of some help. I don't understand why my application is crashing. I boiled down my source code and now I'm at least able to point my finger at the list destructor ~list(). The following sample code mirrors more or less my code and this is also crashing. The if-condition in my destructor is there on purpose, because in my application there is a library call to free things up and I musn't call this if the pointer is NULL. I'm thankfull if someone could explain to me why the code is crashing, and why the destructor is called so often (myList should be empty, should't it?) #include <list class Test { public: Test(int a){pt = new int(a);} Test(const Test& rhs) { int tmp = *(rhs.pt); pt = new int(tmp); } ~Test() { cout << "Destructor was called" << endl; if (pt !=0) delete pt; pt = 0; } private: int* pt; }; int main() { list Test a(5); myList.push_back(a); /* If I comment out the following line there will be no crash! */ myList.~list(); return 0; } [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top John Potter Guest Posted: Thu Sep 25, 2003 2:26 pm    Post subject: Re: strange behaviour of STL destructor On 24 Sep 2003 18:34:22 -0400, [email]bernd_moeller (AT) hotmail (DOT) com[/email] (Mookie) wrote: Quote: I'm thankfull if someone could explain to me why the code is crashing, and why the destructor is called so often (myList should be empty, should't it?) It does not exist. There is a difference. Quote: int main() { list<Test> myList; This is an automatic variable and will be destructed automatically. Quote: Test a(5); myList.push_back(a); /* If I comment out the following line there will be no crash! */ myList.~list(); You have destroyed myList, not cleared it. You now have space for an automatic list which does not contain a list object. Quote: return 0; Destructor called for non-object. KaBOOM! Either replace the destruction with myList.clear(); // It exists and is empty or create a new list there to destroy. new ((void*)&myList) list<Test>(); // There is a new empty list there John [ 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: Thu Sep 25, 2003 3:39 pm    Post subject: Re: strange behaviour of STL destructor In article <47b8e6cd.0309240604.329bd680 (AT) posting (DOT) google.com>, Mookie <bernd_moeller (AT) hotmail (DOT) com> writes Quote: int main() { list<Test> myList; Test a(5); myList.push_back(a); /* If I comment out the following line there will be no crash! */ myList.~list(); return 0; } The problem has nothing to do with your implementation of Test, or the creation of myList. You have explicitly destroyed an auto variable which is a 'No, No, Never' action. You would get the same error if you wrote: int main(){ Test t; t.~Test(); return 0; } You simply must not call destructors on automatic (local, on stack or whatever term you use) objects. -- Francis Glassborow ACCU 64 Southfield Rd Oxford OX4 1PA +44(0)1865 246490 All opinions are mine and do not represent those of any organisation [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Gerhard Menzl Guest Posted: Thu Sep 25, 2003 7:06 pm    Post subject: Re: strange behaviour of STL destructor Mookie wrote: Quote: I'm in need of some help. I don't understand why my application is crashing. I boiled down my source code and now I'm at least able to point my finger at the list destructor ~list(). The following sample code mirrors more or less my code and this is also crashing. The if-condition in my destructor is there on purpose, because in my application there is a library call to free things up and I musn't call this if the pointer is NULL. I don't know what you mean exactly by "free things up", but deleting a null pointer is safe. There is no need to check yourself, as the check is already built-in. Quote: #include <list class Test { public: Test(int a){pt = new int(a);} Test(const Test& rhs) { int tmp = *(rhs.pt); pt = new int(tmp); } ~Test() { cout << "Destructor was called" << endl; if (pt !=0) delete pt; pt = 0; } private: int* pt; }; It is equally pointless to set pt to 0 in the destructor, as it cannot be legally accessed after that anyway. Quote: int main() { list Test a(5); myList.push_back(a); /* If I comment out the following line there will be no crash! */ myList.~list(); return 0; } That's no surprise. Except for the relatively exotic case of in-place construction, you are *never* supposed to call a destructor explicitly. myList is an automatic object, hence its destructor will be called automatically when it goes out of scope at teh end of main(). Calling the destructor explicitly before that invokes undefined behaviour, which can mean anything and, in your case, luckily results in a crash. Quote: I'm thankfull if someone could explain to me why the code is crashing, and why the destructor is called so often (myList should be empty, should't it?) Which destructor do you mean? The list<Test> destructor? This should be called once, if it were not for your erroneous extra invocation. Or the destructor of Test? You should normally see (at least) two invocations here: one for the instance named a, and one for the copy in the list that gets destroyed with the container. Gerhard Menzl [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Tonci Tomic Guest Posted: Thu Sep 25, 2003 8:40 pm    Post subject: Re: strange behaviour of STL destructor "Mookie" <bernd_moeller (AT) hotmail (DOT) com> wrote Quote: int main() { list<Test> myList; Test a(5); myList.push_back(a); /* If I comment out the following line there will be no crash! */ myList.~list(); return 0; } You answered your own question. If you don't invoke destructor twice on myList, everything will be OK. In you case, destructor is invoked twice. Once explicitly, once at the end of main body. Mayby I don't see something obvious but I wonder: Why did you invoke list destructor explicitly?? [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Robert Klemme Guest Posted: Thu Sep 25, 2003 8:42 pm    Post subject: Re: strange behaviour of STL destructor "Mookie" <bernd_moeller (AT) hotmail (DOT) com> schrieb im Newsbeitrag news:47b8e6cd.0309240604.329bd680 (AT) posting (DOT) google.com... Quote: int main() { list<Test> myList; Test a(5); myList.push_back(a); /* If I comment out the following line there will be no crash! */ myList.~list(); You practically never invoke a destrutor manually. myList is destroyed automatically at the end of the scope, i.e. after return. Quote: return 0; } Regards robert [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Ulrich Eckhardt Guest Posted: Thu Sep 25, 2003 8:49 pm    Post subject: Re: strange behaviour of STL destructor Mookie wrote: Quote: Hi group, I'm in need of some help. I don't understand why my application is crashing. I boiled down my source code and now I'm at least able to point my finger at the list destructor ~list(). The following sample code mirrors more or less my code and this is also crashing. Your example violates the requirement of std::list and good programming-practice to supply an assignment-operator. The new'ed int was already deleted by a copy but the original's pointer still left pointing nowhere. Quote: The if-condition in my destructor is there on purpose, because in my application there is a library call to free things up and I musn't call this if the pointer is NULL. FYI: You can always call delete on a null-pointer, at least with a non-broken implementation. Ulrich Eckhardt -- Questions ? see C++-FAQ Lite: http://parashift.com/c++-faq-lite/ first ! [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Radu Grigore Guest Posted: Thu Sep 25, 2003 11:29 pm    Post subject: Re: strange behaviour of STL destructor Quote: int main() { list<Test> myList; Test a(5); myList.push_back(a); /* If I comment out the following line there will be no crash! */ myList.~list(); return 0; } AFAIK the standard says that calling a destructor twice is undefined. And here you call the destructor once explicitely and once implicitely (because the list is on the stack - auto variable). Why would you EVER need to call the destructor explicitely? For a STL container clear() is probably what you want. regards, radu http://www.geocities.com/_rgrig [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Siemel Naran Guest Posted: Sat Sep 27, 2003 12:40 am    Post subject: Re: strange behaviour of STL destructor "Erik Max Francis" <max (AT) alcyone (DOT) com> wrote in message Quote: The circumstances under which you want to call a destructor explicitly are extremely rare, and this is a good example why. The list is deleted So when is an explicit call to the destructor warranted? -- +++++++++++ Siemel Naran [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Siemel Naran Guest Posted: Sat Sep 27, 2003 12:42 am    Post subject: Re: strange behaviour of STL destructor "Ulrich Eckhardt" <doomster (AT) knuut (DOT) de> wrote Quote: Your example violates the requirement of std::list and good programming-practice to supply an assignment-operator. The new'ed int was already deleted by a copy but the original's pointer still left pointing nowhere. Not so. The std::list will use the copy constructor to copy elements. It cannot construct a default Test object and then set it using Test::operator= because a default constructor Test::Test() does not exist. Test(const Test& rhs) { int tmp = *(rhs.pt); pt = new int(tmp); } The copy constructor makes a deep copy of the integer, and the destructor releases it. So there's no memory leak here. However, it's a good practice to have an assignment operator that does the right thing, which here is like the copy constructor to make a deep copy. Never know who might want to use it, as in Test test2 = test1. -- +++++++++++ Siemel Naran [ 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 Goto page 1, 2  Next Page 1 of 2   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.