but based on the graph, they look like two different heads. What am I doing wrong or misunderstanding
The graph shows quite clearly that 502 is not a head at all, right there in red. Your view filters out a set of changes, a set of changes that should show that 502 is an ancestor of 512.
Can you turn off the filter and reposition the image, or better yet provide hg glog command line hg glog ?
In the worst case scenario, you can hg update 502 make a do-nothing change whose parent is 502 and then combine it with 512, but you better get a complete picture / understanding of what is happening.
source share