Bug 43749 - Eliminate extra traversal of the tree when determining ability to lazy-attach a Node.
Summary: Eliminate extra traversal of the tree when determining ability to lazy-attach...
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: New Bugs (show other bugs)
Version: 528+ (Nightly build)
Hardware: Other OS X 10.5
: P2 Normal
Assignee: Dimitri Glazkov (Google)
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2010-08-09 15:10 PDT by Dimitri Glazkov (Google)
Modified: 2010-08-17 10:58 PDT (History)
5 users (show)

See Also:


Attachments
Patch (1.61 KB, patch)
2010-08-09 15:12 PDT, Dimitri Glazkov (Google)
eric: review+
eric: commit-queue-
Details | Formatted Diff | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Dimitri Glazkov (Google) 2010-08-09 15:10:13 PDT
Eliminate extra traversal of the tree when determining ability to lazy-attach a Node.
Comment 1 Dimitri Glazkov (Google) 2010-08-09 15:12:32 PDT
Created attachment 63933 [details]
Patch
Comment 2 Dimitri Glazkov (Google) 2010-08-09 15:16:50 PDT
Since neither SVG nor Input shadow DOM aren't using this path (they use home-baked attachement), the check for whether we're inside of the shadow tree isn't needed.

This gives a small, but consistent win in Dromaeo's DOM Modification.

If you're really mean, you can make me write layout tests to prove that this check isn't meaningful inside of the shadow DOM.
Comment 3 Dimitri Glazkov (Google) 2010-08-17 09:03:26 PDT
Adding more peeps.
Comment 4 Eric Seidel (no email) 2010-08-17 09:06:01 PDT
Comment on attachment 63933 [details]
Patch

Lipstick on a pig.  Eager attach needs to die.

Can you explain better why this check is not needed in your ChangeLog?
Comment 5 Eric Seidel (no email) 2010-08-17 09:07:23 PDT
Sorry, meant to CC adele, not beth.  Adele is the queen of the world of shadow DOMs, iirc.
Comment 6 Dimitri Glazkov (Google) 2010-08-17 09:08:01 PDT
(In reply to comment #4)
> (From update of attachment 63933 [details])
> Lipstick on a pig.  Eager attach needs to die.
> 
> Can you explain better why this check is not needed in your ChangeLog?

Will do.
Comment 7 Dimitri Glazkov (Google) 2010-08-17 09:15:27 PDT
Committed r65508: <http://trac.webkit.org/changeset/65508>
Comment 8 Dimitri Glazkov (Google) 2010-08-17 10:58:53 PDT
I bungled landing this and had to land the actual change in http://trac.webkit.org/changeset/65517. Sorry :-\