Feature request: can we have support for XMLNode_add_sibling()? about sxmlc

[matthieu-labas]

I'm still polishing v4.3.0 but here is what the final XMLNode_insert_child() will look like, if you want to use it asap:

#define CHECK_NODE(node,ret) if ((node) == NULL || (node)->init_value != XML_INIT_DONE) return (ret)

int XMLNode_insert_child(XMLNode* node, XMLNode* child, int index)
{
	int i, j;

	CHECK_NODE(node, -1);

	/* We could process cases "first" and "last" in an optimized way, but we prefer readability to (micro-)optimization */
	if (index < 0) /* Before first => first */
		index = 0;
	if (index >= node->n_children) /* After last => last */
		index = node->n_children - 1;

	for (i = 0; i < node->n_children; i++) {
		if (!node->children[i]->active || index-- > 0)
			continue;
		/* Insert it here, at 'i' */
		if (_add_node(&node->children, &node->n_children, child) >= 0) {
			node->tag_type = TAG_FATHER;
			child->father = node;
			/* Erase 'child', which is the last node ('n_children' has been incremented by '_add_node()') */
			for (j = node->n_children - 1; j >= i; j--)
				node->children[j] = node->children[j-1];
			node->children[i] = child; /* Set it */
			return true;
		} else
			return false;
	}

	return false; /* Oops! */
}

[matthieu-labas]

v4.3.0 pushed.
Some more work is coming regarding documentation (tentative switching to Doxygen) and more thorough unit testing.

[damian-m-g]

Wow, this was very fast, and I'll be using it on next days. Thanks!

[matthieu-labas]

Good idea! I will implement it in the upcoming v4.3.0.

In the meantime, if you want that type of functionality, you should directly manipulate the node->children array (with XMLNode_add_child() to allocate space, then move all nodes one by one in a loop):

XMLNode_add_child(node, child);
memmove(node->children[i+1], node->children[i], node->n_children - i - 1);
node->children[i] = child;

[damian-m-g]

This are very good news, and thanks for the explanation on the workaround.

I'll be looking forward for this new version! In the meantime, I'll make use of the workaround. Cheers!

[damian-m-g]

This is great Matthieu, I'll be using this code these days. I'll be looking forward for 4.3.0.