W3C home > Mailing lists > Public > public-script-coord@w3.org > October to December 2013

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

From: <bugzilla@jessica.w3.org>
Date: Fri, 11 Oct 2013 12:55:52 +0000
To: public-script-coord@w3.org
Message-ID: <bug-20567-3890-LYtrjPXreU@http.www.w3.org/Bugs/Public/>
https://www.w3.org/Bugs/Public/show_bug.cgi?id=20567

--- Comment #34 from Bobby Holley (:bholley) <bobbyholley@gmail.com> ---
(In reply to Anne from comment #33)

> For window.open() Bobby describes in
> http://lists.w3.org/Archives/Public/www-dom/2012OctDec/0173.html what we do
> in Gecko, which seems to allow for collecting the previous tree, but not
> necessarily the entire context. Is that correct?

Mmm, not necessarily. Our setup has changed a bit since that posting, mainly
because most everything is now on the new bindings, where we don't do orphan
fixup.

For document.open, we basically just eject all the old nodes and leave them
behind (orphaned) in the old window. Inserting them into the document again
requires adopting them, just as we'd have if the nodes were non-orphan nodes
that lived in some other document.

I would guess that, in 90% of cases, the old scope can simply die, because
nobody is holding references to the old nodes. If they are holding references,
those references will keep the old window alive.

> If we still leak some of the old context it's hard to make a similar
> argument here.

Per the above, I don't think they're really analogous. Moreover, the leak issue
being discussed in this bug involves leaking the old window _even_ in the case
when the node has been properly inserted into the new document (in the new
window), just by virtue of the node's sordid past. This is very different from
the document.domain case.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
Received on Friday, 11 October 2013 12:55:58 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:14:18 UTC