Delete Leaves With a Target Value

Given a binary tree and an integer target, repeatedly delete every leaf whose value equals target. After deletions, a parent may become a new target-valued leaf and must also be deleted.
Return the final root. Modify and reuse the surviving input nodes.

Function
removeLeafNodes(root: TreeNode, target: int) → TreeNode

Examples
Example 1
root = [1,2,3,2,null,2,4]
target = 2
return = [1,null,3,null,4]
All target-valued leaves are removed, including the left child after its child disappears.
Example 2
root = [1,3,3,3,2]
target = 3
return = [1,3,null,null,2]
The left 3 remains because it still has a non-target child.
Example 3
root = [1,2,null,2,null,2]
target = 2
return = [1]
Deletion cascades upward through the chain of 2-valued nodes.
Example 4
root = []
target = 1
return = []
An empty tree remains empty.

Constraints
The tree contains between 0 and 100000 nodes.
-10^9 <= node.val, target <= 10^9.

The trick is processing children before the parent. Recurse into left and right first, and assign the results back to node.left and node.right. Then check the current node: if both children are now null and node.val equals target, return null. Otherwise return the node. That single postorder pass handles the cascade, because by the time you judge a parent, its children have already been pruned. The common pitfall is checking the node before recursing, which misses the chains in Example 3 and forces a repeat loop. Another pitfall is forgetting to reassign the child pointers, so the deletion never sticks. With up to 100000 nodes, a skewed tree can make recursion deep, so mention an iterative option or note the stack depth. Time is O(n), space is O(h). If the postorder ordering slips away mid-assessment, StealthCoder is the hedge that hands you the working solution while you stay in the editor.

If this hits your live OA and you blank, StealthCoder solves it in seconds, invisible to the proctor.