Hi Derek, Thanks for taking a look.
> If we really want to keep the current behavior, I think we would > need to update the doc string for > `org-columns-inhibit-recalculation' to match, since right now if I > read it I would assume that nothing gets recomputed. Agreed. I have pushed the change to main as 03bb8b2bf, including a docstring update: "Inhibit recomputing of columns when building column view. When non-nil, neither column summaries nor clock summaries are recomputed." > While I think your proposal is the right approach, we would need to > add something in NEWS since I think this could be considered a > breaking change. I don't think a NEWS entry is necessary here. The variable is a plain defvar, not a defcustom, and it is not documented in the manual. More importantly, there should be no user-visible change: all the commands that bind it to t (deleting, resizing or reordering columns, moving rows) cannot change clock sums, and the `:org-clock-minutes' text properties set by the initial computation survive those operations, so the skipped recomputation was redundant work. And, as you note, the new behavior is what a reader of the docstring would expect in the first place. That said, if you feel strongly about it, I'm happy to add a short entry. Best, -- Slawomir Grochowski
