Bug 74768 - Cache and reuse the HTMLAllCollection returned by document.all.
Summary: Cache and reuse the HTMLAllCollection returned by document.all.
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: DOM (show other bugs)
Version: 528+ (Nightly build)
Hardware: Unspecified Unspecified
: P2 Normal
Assignee: Andreas Kling
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2011-12-16 16:41 PST by Andreas Kling
Modified: 2011-12-17 19:58 PST (History)
5 users (show)

See Also:

Attachments
Patch (12.97 KB, patch)
2011-12-16 16:50 PST, Andreas Kling 		kling: review-
Details | Formatted Diff | Diff
Patch v2 (11.18 KB, patch)
2011-12-17 09:06 PST, Andreas Kling 		koivisto: review+
Details | Formatted Diff | Diff
Show Obsolete (1) View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Andreas Kling 2011-12-16 16:41:50 PST 
document.all should behave like the other collections on document (multiple calls return the same objects.)
Comment 1 Andreas Kling 2011-12-16 16:50:24 PST 
Created attachment 119699 [details]
Patch
Comment 2 Andreas Kling 2011-12-16 19:30:38 PST 
Comment on attachment 119699 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=119699&action=review

> Source/WebCore/html/HTMLCollection.cpp:70
> -    if (m_baseIsRetained)
> +    if (!m_base->isDocumentNode())

Gnarf! This isn't correct, isDocumentNode() will be false after ~Document() runs (if the HTMLCollection is being destroyed after the Document removed the last ref.)
Comment 3 Andreas Kling 2011-12-17 09:06:28 PST 
Created attachment 119726 [details]
Patch v2

Keep the m_baseIsRetained logic in HTMLCollection, since we can't depend on calling Node::isDocumentNode() after ~Document().
m_baseIsRetained can go away by making all HTMLCollections cached instead, since the collection will then never hold a ref on its base node.
Comment 4 Antti Koivisto 2011-12-17 18:45:16 PST 
Comment on attachment 119726 [details]
Patch v2

r=me

The ChangeLog could use a sentence or two about why this is a good thing. Bonus points for some performance measurements.
Comment 5 Andreas Kling 2011-12-17 19:58:05 PST 
Committed r103166: <http://trac.webkit.org/changeset/103166>