René's URL Explorer Experiment

{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"Potential memory leak in SimpleHeap's removed_node_tuples","articleBody":"Hey there, I've been investigating a memory issue in the SimpleHeap implementation and found that `removed_node_tuples` can grow indefinitely during pathfinding, especially on larger grids.\n\nThe problem is that when nodes are \"removed\" via `remove_node()`, they get added to `removed_node_tuples` but are never cleaned up. Over time, this set just keeps growing.\n\nAfter digging into the code, I'm wondering if `removed_node_tuples` is actually necessary at all? From what I can see, the A* algorithm already marks processed nodes with the closed flag, so we could potentially simplify things by:\n\n1. Getting rid of `removed_node_tuples` completely\n2. Having `pop_node()` just skip any nodes that are already closed\n3. Making `remove_node()` do nothing (or removing it entirely)\n\nSomething like:\n```\ndef pop_node(self):\n    while self.open_list:\n        node_tuple = heapq.heappop(self.open_list)\n        # ... get node from tuple ...\n        if not node.closed:\n            return node\n    return None\n```\n\nI tested this approach and it seems to work correctly while eliminating the memory leak. The pathfinding results are identical, but without the growing memory usage.\n\nIs there a specific reason removed_node_tuples was implemented this way, or would you be interested in a PR that simplifies this? I might be missing some edge case that requires the current approach, so wanted to check first.\n\nThanks!","author":{"url":"https://github.com/codelion","@type":"Person","name":"codelion"},"datePublished":"2025-08-15T01:27:23.000Z","interactionStatistic":{"@type":"InteractionCounter","interactionType":"https://schema.org/CommentAction","userInteractionCount":3},"url":"https://github.com/71/python-pathfinding/issues/71"}