Once we have built our node tree we need some way to parse it in order to produce a pretty html formatted page.  When I was developing this it took way to much thinking time for the number of lines of code, but there are literally 2 functions that we use, and out of them, only one is called externally (the process_nodes function).

The first function is used to look for the first open sibling.  If it finds one then it returns a pointer to that node, if not then it returns nullptr

//**************************************************************//
// ARE ALL DIRECT DECENDENTS OF THE CURRENT NODE CLOSED?        //
//**************************************************************//
html_tree *pretty_html::find_first_open_sibling(html_tree *parent) {

    html_tree *workNode = nullptr;

    for(html_tree *it : parent->childNodes) {
        if(it->closed() != html_tree::status::CLOSED) {
            workNode = it;
            break;
        }
    }

    return(workNode);
}

 

And last but not least, the process node function, this is the function that actually creates the html page.

//**************************************************************//
// ARE ALL DIRECT DECENDENTS OF THE CURRENT NODE CLOSED?	//
//**************************************************************//
void pretty_html::process_nodes(html_tree *primaryNode) {

    html_tree *curNode = primaryNode;
    unsigned tabs =0;
    while(primaryNode->closed() != html_tree::status::CLOSED) {

        if(curNode->closed() == html_tree::status::UNPROCESSED) {
            curNode->open_node(tabs);
            ++tabs;
        }

        //if the current node is a leaf node and it is open then close it
        if((curNode->is_leaf_node()) && (curNode->closed() == html_tree::status::OPEN)) {

            --tabs;
            curNode->close_node(tabs);

        }

        //NEXT NODE
        //find the next node to work on
        //first of all we need to see if the node has any open children
        html_tree *sibling = find_first_open_sibling(curNode);

        //if the node does have children then we need to drill down to the first unclosed child
        if(sibling){
            curNode = sibling;
        }

        //if it has no children then it can be closed regardless of whether it is a leaf node
        //or not.
        else{

            //if the current node has no open children then close it
            if(curNode->closed() == html_tree::status::OPEN) {
                --tabs;
                curNode->close_node(tabs);
            }

            //we can then make our way back up to the parent node
            curNode = curNode->previousBranch;
        }

    }

}

How this function works is explained in the comments, once called the string passed to the primary iteration of the html_tree class will contain a valid html page (with any luck).