Bug 155002 - [[GetPrototypeOf]] should be a fully virtual method in the method table
Summary: [[GetPrototypeOf]] should be a fully virtual method in the method table
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: JavaScriptCore (show other bugs)
Version: WebKit Nightly Build
Hardware: Unspecified Unspecified
: P2 Normal
Assignee: Saam Barati
URL:
Keywords:
Depends on: 161534 155097
Blocks:
  Show dependency tree / graph
 
Reported: 2016-03-03 19:20 PST by Saam Barati
Modified: 2016-09-02 09:46 PDT (History)
10 users (show)

See Also:

Attachments
patch (59.30 KB, patch)
2016-03-05 15:31 PST, Saam Barati 		fpizlo: review+
Details | Formatted Diff | Diff
perf numbers (66.95 KB, text/plain)
2016-03-05 15:35 PST, Saam Barati 		no flags Details
patch for landing (58.05 KB, patch)
2016-03-06 14:35 PST, Saam Barati 		no flags Details | Formatted Diff | Diff
patch for landing v2 (119.99 KB, patch)
2016-03-06 16:07 PST, Saam Barati 		no flags Details | Formatted Diff | Diff
v3 (121.00 KB, patch)
2016-03-06 16:12 PST, Saam Barati 		no flags Details | Formatted Diff | Diff
v4 (128.08 KB, patch)
2016-03-06 16:25 PST, Saam Barati 		no flags Details | Formatted Diff | Diff
Show Obsolete (4) 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 Saam Barati 2016-03-03 19:20:54 PST 
This is needed for Proxy.
To make things fast, we will probably have to change
some JIT code to respect this and to introduce a getPrototypeInline
which will try to perform the default getPrototype which will probably
be named getPrototypeDirect.
Comment 1 Saam Barati 2016-03-05 15:31:01 PST 
Created attachment 273102 [details]
patch
Comment 2 Filip Pizlo 2016-03-05 15:32:32 PST 
Comment on attachment 273102 [details]
patch

View in context: https://bugs.webkit.org/attachment.cgi?id=273102&action=review

> Source/JavaScriptCore/dfg/DFGOperations.cpp:1392
>  
> +#if USE(JSVALUE64)
> +EncodedJSValue 
> +#else
> +int32_t
> +#endif

Yuck!  Why not just return size_t like other functions that are in this situation?
Comment 3 Saam Barati 2016-03-05 15:35:19 PST 
Created attachment 273104 [details]
perf numbers
Comment 4 Saam Barati 2016-03-06 14:35:54 PST 
Created attachment 273147 [details]
patch for landing
Comment 5 Saam Barati 2016-03-06 15:14:25 PST 
The compiler is confused about the ClassName::getPrototype when
building the method table :(
Comment 6 Saam Barati 2016-03-06 16:07:33 PST 
Created attachment 273151 [details]
patch for landing v2

I want to see EWS results
Comment 7 Saam Barati 2016-03-06 16:12:08 PST 
Created attachment 273152 [details]
v3

more method renaming. Hopefully no more is needed.
Comment 8 Saam Barati 2016-03-06 16:25:33 PST 
Created attachment 273153 [details]
v4
Comment 9 Saam Barati 2016-03-06 17:00:56 PST 
landed in:
http://trac.webkit.org/changeset/197648