Reading scroll listeners without drowning in the trace
A raw scroll trace is a bad place to start an argument. Every movement fires. The page looks guilty because the log is long. We have sat with teams who spent a week inside that length and still could not say which listener belonged to the header, which to a sticky table, and which to a third-party map that had been added “only on the venue page.”
The method we use is slower at the beginning. Name the surfaces. Draw the attach points on paper. Then sample fire: one short scroll on one surface, with the others held still if you can. The question is not how many lines the trace produced. The question is whether a listener attached on the venue page is still firing on the ticket page.
Pre-release passes use this method because freeze weeks do not allow a week inside a log. If a scroll listener accumulates across views, the punch-list names the view that attached it and the view that failed to detach it. That is enough for a contractor. A full essay on scroll performance is a different sitting, and we will say so rather than pad the pass.