[Bug 20567] Change [[Prototype]] for concept-node-adopt?

https://www.w3.org/Bugs/Public/show_bug.cgi?id=20567

--- Comment #58 from Bobby Holley (:bholley) <bobbyholley@gmail.com> ---
(In reply to Dominic Cooney from comment #57)
> What's the units of this measurement? 2.5 globals per... ?

Per running instance of Firefox. We maintain a count of live Windows whose
Documents have adopted at least one node from another Window. During every GC,
telemetry-enabled Firefox builds record this value as a data point in the
telemetry report, which gets sent to Mozilla's servers every 24 hours or so. So
it's an average over time, users, workloads, browsing patterns, etc.

> I will try to take another stab at this. Let me talk to my web developer
> brain trust. Apart from "no leaks", are there any other constraints a
> solution needs to incorporate?

Well, I'm pretty sure to get "no leaks" we need to spec (2) for the general DOM
case (unless there's another solution that has yet to be raised). So the main
question is how Web Components wants to treat itself in this whole charade.

I think the ideal thing would be to treat it as a first-class citizen and find
a way to make it work right cross-global. This involves structuring the
callbacks to take this into a account, and figuring out a way to either
re-create the binding in the new scope, or remove it gracefully.

The easier option is just to spec that these things can't be adopted, and
decide where/how to fail when people try. But it's not a very forward-thinking
strategy if we really believe in Web Components as a foundational technology
for the future web.

I'm certainly no expert in Web Components or webdev ergonomics, but I'd
nevertheless enjoy being CCed for these discussions to gain a fuller
understanding of the issues at play here. :-)

-- 
You are receiving this mail because:
You are on the CC list for the bug.

Received on Friday, 1 November 2013 12:28:46 UTC