See https://bugzilla.mozilla.org/show_bug.cgi?id=781767#c4 (In reply to Brian Hackett (:bhackett) from comment #4) > A separate thing is that even when a GC clears analysis-temporary data, > usage will not go to zero; the bucket describes an allocator that is cleared > on each GC but is then immediately refilled with certain data that has to > live across collections. Bug 778724 also improves this, by splitting the > allocator in two, and memory reporting could be improved as a followup to > distinguish the two allocators (the current patch gloms them together).

In bug 775382 I saw the analysis-temporary numbers go negative, which indicates that it's possible for something to go wrong within the size computation. In general, straight measurements of allocated memory (e.g. with malloc_usable_size) are preferable to computed sizes, because computed sizes are error-prone. (https://wiki.mozilla.org/Memory_Reporting#Two_Ways_to_Measure has more details.) So if the subtractions can be removed as part of this change, that would be great.

Nick, is this a dupe of bug 789398?

Yes! Thanks.