Bug 144676 - Don't allocate a StringImpl for every Number JSValue in JSON.stringify().
Summary: Don't allocate a StringImpl for every Number JSValue in JSON.stringify().
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: JavaScriptCore (show other bugs)
Version: 528+ (Nightly build)
Hardware: Unspecified Unspecified
: P2 Normal
Assignee: Andreas Kling
URL:
Keywords: Performance
Depends on:
Blocks:
 
Reported: 2015-05-06 00:53 PDT by Andreas Kling
Modified: 2015-05-06 10:12 PDT (History)
5 users (show)

See Also:

Attachments
Patch (1.90 KB, patch)
2015-05-06 00:54 PDT, Andreas Kling 		no flags Details | Formatted Diff | Diff
Patch (1.88 KB, patch)
2015-05-06 01:39 PDT, Andreas Kling 		darin: 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 2015-05-06 00:53:46 PDT 
We can do a lot better on Kraken/json-stringify-tinderbox!

For example:
json-stringify-tinderbox        63.370+-3.319      ^      56.031+-3.150         ^ definitely 1.1310x faster
Comment 1 Andreas Kling 2015-05-06 00:54:46 PDT 
Created attachment 252461 [details]
Patch
Comment 2 WebKit Commit Bot 2015-05-06 00:56:32 PDT 
Attachment 252461 [details] did not pass style-queue:


ERROR: Source/JavaScriptCore/runtime/JSONObject.cpp:41:  Alphabetical sorting problem.  [build/include_order] [4]
Total errors found: 1 in 2 files


If any of these errors are false positives, please file a bug against check-webkit-style.
Comment 3 Andreas Kling 2015-05-06 01:39:05 PDT 
Created attachment 252462 [details]
Patch
Comment 4 Darin Adler 2015-05-06 09:23:09 PDT 
Comment on attachment 252462 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=252462&action=review

> Source/JavaScriptCore/runtime/JSONObject.cpp:399
> +            NumberToStringBuffer buffer;
> +            builder.append(numberToString(number, buffer));

StringBuilder has a function named appendECMAScriptNumber. I suggest you use that instead, which would make your patch even smaller.
Comment 5 Andreas Kling 2015-05-06 10:12:05 PDT 
Committed r183874: <http://trac.webkit.org/changeset/183874>