Measure performance impact of the membrane

[Gr8Gatsby]

SOMA#431

Migrated issue

Notes:

Today querySelector and querySelectorAll and the respective APIs for the shadow root returns a Membrane wrapping HTMLElements. It is used today to enforce the ShadowDOM semantic at the framework level.

Need to measure:

• Performance cost (speed): handler and piercingHook
• Memory cost: Proxy creation for the replica

To be defined:

• We need to define what would be a good benchmark to measure the membrane cost? A javascript library requiring a lot of interaction with the DOM, like a javascript text editor (CKEditor, Code Mirror)?

[caridy]

I don't think this is relevant anymore since our main goal is to eliminate the traverse membrane as soon as possible. @davidturissini what do you think?

[davidturissini]

I don't think we need to measure the membrane per say, but having perf tests for shadow dom methods like querySelector and querySelector all would be useful so we can measure perf as we migrate away from the membrane.

[caridy]

since we don't use membranes for the DOM anymore, closing this one.