C++ libxml xmlNode->children - c++

Question

Ok I have been working on a class to iterate through all nodes in an html doc and return the data that I need. This is very simple and I have achieved this in Bash but now I am trying to port the same to C++.
I started with the example on the libxml site but I have stepped through this function node by node and I can't understand how it is working.
Here is the function:
static void
print_element_names(xmlNode * a_node)
{
xmlNode *cur_node = NULL;
for (cur_node = a_node; cur_node; cur_node = cur_node->next) {
if (cur_node->type == XML_ELEMENT_NODE) {
printf("node type: Element, name: %s\n", cur_node->name);
}
print_element_names(cur_node->children);
}
}
So basically, this function takes a node assigns it to a pointer, and starts to loop through all sibling nodes, but if the current node has children, it calls the function and start over on that child node. This is all very understandable.
So it drives down the doc structure, but how does it navigate back up the structure?
Does xmlNode->children return the next parent node when it is found NULL? As far as I can tell, this is not true, but I just can't figure out how this is working.
I successfully created a class to do what I want, but it is so much more complex than this and about 10 lines longer. I actually had to check if the next node was null and either navigate down if it has children or navigate back up and to the next node if it didn't.
This example is much simpler and I would like to understand how to make my code better.

Answer 1

So it drives down the doc structure, but how does it navigate back up the structure?
What you have posted is a recursive function. From the nature of your question, I'm assuming you don't quite understand what recursion is/how it works. A quick google search should give you some good information/examples.
Does xmlNode->children return the next parent node when it is found NULL?
I am assuming that xmlNode->children returns a pointer to a linked list containing all children of the current node. If the current node has no children, it would probably return NULL.
This example is much simpler and I would like to understand how to make my code better.
I am assuming the class you wrote was purely iterative. Recursive functions can greatly simplify code, but performance wise they can cause issues on larger data sets. I would definitely recommend reading up on them; they can be quite fun.

Answer 2

Solving the problem iteratively is essentially a tree traversal. You will need a stack to accomplish this. The stack can probably be most easily implemented as a singly-linked-list.
// interface to be implemented
typedef void* Stack;
Stack stack_new(); // creates a new stack
void stack_add(Stack stack, xmlNode *element); // adds an element to the stack
int stack_size(); // returns the number of elements currently in the stack
xmlNode* stack_remove(Stack stack); // pops an element from the stack
void stack_free(Stack stack); // frees up resources used by the stack
// printing code
static void print_element_names(xmlNode *a_node)
{
Stack stack = stack_new();
stack_add(stack, a_node);
while(stack_size(stack))
{
xmlNode *cur_node = stack_remove(stack);
if(cur_node->children) stack_add(cur_node->children);
xmlNode *iter_node = NULL;
for (iter_node = cur_node; iter_node; iter_node = iter_node->next)
{
if (iter_node->type == XML_ELEMENT_NODE)
printf("node type: Element, name: %s\n", iter_node->name);
}
}
stack_free(stack);
}