When lots of functions are dropped, they're not easy to manage in the UI about profiler

[mstange]

Also, having the ability "un-drop" a specific function would be very useful.

For example, you may want to drop samples in GC, find the heaviest JS function (by JS self-time), focus that function, and un-drop GC.

[mstange]

Proposed UI:

---

[mstange]

I've noticed some additional cases where reordering the drop transform with the rest of the transform stack doesn't preserve the result.

Case 1: Focus + Drop with recursion

If you have two samples:
A -> B -> C
A -> B -> A
Focus on B, Drop A. Now only the A->B->C sample remains.
Unfocus B. Now all samples are gone! This is somewhat unexpected.

Case 2: Drop + Collapse

If you have two samples
A -> B -> C
A -> B -> D
Drop C. Then collapse B. There should be only one sample remaining, but with the proposed implementation (dropping at the end, after all transforms are applied), both samples would remain!

I'm not yet sure what to do about this.

[gregtatum]

I feel like creating a drop Set and merge Set preserves the non-commutative nature of transforms. This way if you do 10 drops in a row, it would all be in a single set on the transform stack, e.g.

{ type: "drop", funcIndex: Set<IndexIntoFuncTable> }

[gregtatum]

The UI you propose could be done inline in the existing bar. I'm wondering if you could also make it overflow: auto so that even if you have 100 entries, you could still navigate it. Or maybe it's worth not putting it in a bar and just have some UI similar to your 51 dropped functions that opens a modal or dropdown.