I'm writing a program as an assignment for school and I though I had worked out all the bugs until I decided to call my copy constructor. The program is interactive (CLI) which basically has two moduals: a .h and .cpp file for the LList class and a .h and .cpp file for the structure of the program and also a 3rd cpp file just for main(). It is suppose to be a program for a hydropower engineering company (fake company) in which the LList nodes hold data for their annual water flow in a river(year and flow). Here is some insight on the class:
//list.h
struct ListItem {
int year;
double flow;
};
struct Node {
ListItem item;
Node *next;
};
class FlowList {
public:
FlowList();
// PROMISES: Creates empty list
FlowList(const FlowList& source);
// REQUIRES: source refers to an existing List object
// PROMISES: create the copy of source
~FlowList();
// PROMISES: Destroys an existing list.
FlowList& operator =(const FlowList& rhs);
// REQUIRES: rhs refers to an existing FlowList object
// PROMISES: the left hand side object becomes the copy of rhs
//....Other member functions
private:
// always points to the first node in the list.
Node *headM;
// Initially is set to NULL, but it may point to any node.
Node *cursorM;
//For node manipulation within interactive CLI
void copy(const FlowList& source);
void destroy();
I belive the memory leak or collision is taking place somewhere within the copy function but cant pin point where.
//list.cpp
FlowList::FlowList() : headM(0), cursorM(0) {}
FlowList::FlowList(const FlowList& source)
{
copy(source);
}
FlowList::~FlowList()
{
destroy();
}
FlowList& FlowList::operator =(const FlowList& rhs)
{
if (this != &rhs)
{
destroy();
copy(rhs);
}
return (*this);
}
//...more function definitions
void FlowList::copy(const FlowList& source)
{
if (source.headM == NULL)
{
headM = NULL;
cursorM = NULL;
return;
}
Node* new_node = new Node;
assert(new_node);
new_node->item.year = source.headM->item.year;
new_node->item.flow = source.headM->item.flow;
new_node->next = NULL;
headM = new_node;
Node* thisptr = new_node;
for(Node* srcptr = source.headM->next; srcptr != NULL; srcptr = srcptr->next)
{
new_node = new Node;
assert(new_node);
new_node->item.year = srcptr->item.year;
new_node->item.flow = srcptr->item.flow;
new_node->next = NULL;
thisptr->next = new_node;
thisptr = thisptr->next;
}
}
void FlowList::destroy()
{
Node* ptr = headM;
Node* post = headM->next;
while (ptr != NULL)
{
delete ptr;
ptr = post;
if (post)
post = ptr->next;
}
headM = NULL;
}
The program works fine if I create a FlowList object, fill it with data from a .dat file; i can then manipulate the data within the program (display, perform calculations, add to the list, remove from list and save data to file) but program crashes (segmentation fault) if I create another FlowList object (within main.cpp). Any help would be really appreciated.
1 Answer 1
The initial thing I spot is that it looks like your destroy() function will always segmentation fault if the list is empty:
void FlowList::destroy()
{
Node* ptr = headM;
Node* post = headM->next;
//...
}
If the list is empty, headM is NULL and then you're trying to do headM->next which will always produce a segmentation fault in that case.
I think you might also have a memory leak in your copy constructor if you pass in an empty list. If you look at this code:
void FlowList::copy(const FlowList& source)
{
if (source.headM == NULL)
{
headM = NULL;
cursorM = NULL;
return;
}
//...
}
What if the current list contained 20 items and source is an empty list? You set the current list's headM and cursorM to NULL, but you never call delete on any of the nodes in the current list that you originally used new to create. You probably want to work your destroy() function somewhere into your copy constructor too (you did it for the operator= function).
The last thing I noticed is that you don't initialize cursorM for a non-empty list in your copy() function (@Akusete pointed that out as well). I think I'd recommend that at the beginning of your copy constructor, just initialize cursorM and headM to NULL just to cover your bases.
It looks like you're really close, I think you just really need to think through the boundary case of dealing with empty lists (both on the LHS and RHS) and you'll probably find most of these bugs. Good luck!
copyand adestroyin multiple places in the code. It's pretty common to split them out.