In the previous release cycle, there was an attempt to optimize TP by hiding the edit menu's items and shutting their command-update mechanism. That didn't work, but the complex-code is still in place (I tend to complexity of this code is the main reason that it didn't help perf after all) Experience shows that popups (or maybe just "heavy" poupup) may have notable affect on Txul - the startUI-poupp regressed Txul much when it landed. In that case, we fixed the regression by loading editBookmarkOverlay lazily, meaning that the overlay is applied the first time it's needed. This report is on applying this optimization on the edit menu along with the browser's context menu, and on removing the old edit-menu optimization that's mentioned above. While working this, I noticed that various attributes are set on edit/context-menu items unnecessarily. I'll fix that too

If these panels are display: none, I don't think they incur a significant startup cost, and I don't think lazy loaded XUL overlays would be considered a good idea now, given that some people would like to remove support for xul overlays completely.