Bug 79903 – WTF: Micro-optimize cleanup of empty vectors and hash tables

• Status: RESOLVED
• Resolution: FIXED
• Component: JavaScriptCore

[Andy Wingo]

The patch to be attached has the following changelog:

        This patch speeds up cleanup of vectors and hash tables whose
        backing store was never allocated.  This is the case by default
        for most vectors / hash tables that never had any entries added.

        The result for me is that calling checkSyntax 1000 times on
        concat-jquery-mootools-prototype.js goes from 6.234s to 6.068s, a
        2.4% speedup.

        * wtf/HashTable.h:
        (WTF::HashTable::~HashTable):
        (WTF::::clear): Don't deallocate the storage or frob member
        variables if there is no backing storage.
        * wtf/Vector.h:
        (WTF::VectorBufferBase::deallocateBuffer): Likewise.

[Andy Wingo]

Created attachment 129436 [details]
Patch

[Andy Wingo]

Comment on attachment 129436 [details]
Patch

Thanks for the review!

[Geoffrey Garen]

Comment on attachment 129436 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=129436&action=review

> Source/JavaScriptCore/wtf/HashTable.h:983
> +        if (m_table) {

WebKit style is to use early return here:

if (!m_table)
    return;

This reduces overall indentation.

> Source/JavaScriptCore/wtf/Vector.h:284
> +            if (bufferToDeallocate) {

WebKit style is to use early return here:

if (!bufferToDeallocate)
    return;

This reduces overall indentation.

[Andy Wingo]

Comment on attachment 129436 [details]
Patch

clearing cq+ in anticipation of landing a patch with ggaren's feedback incorporated

[Andy Wingo]

Created attachment 129484 [details]
Patch for landing

[WebKit Review Bot]

Comment on attachment 129484 [details]
Patch for landing

Rejecting attachment 129484 [details] from commit-queue.

New failing tests:
fast/js/kde/RegExp.html
Full output: http://queues.webkit.org/results/11770092

[Andy Wingo]

Comment on attachment 129484 [details]
Patch for landing

Flipping cq+ again; I think that the previous run was when the queues were flaky.  Will look into this if it does a cq- again.

[WebKit Review Bot]

Comment on attachment 129484 [details]
Patch for landing

Clearing flags on attachment: 129484

Committed r109731: <http://trac.webkit.org/changeset/109731>

[WebKit Review Bot]

All reviewed patches have been landed.  Closing bug.

[Darin Adler]

Moving all JavaScriptGlue bugs to JavaScriptCore. The JavaScriptGlue framework itself is long gone. And most of the more recent bugs put in this component were put there by people who thought this was for some other aspect of “JavaScript glue” and have nothing to do with the actual original reason for the existence of this component, which was an OS-X-only framework named JavaScriptGlue.