Bug 87872

Summary: [Qt] Implement a shim layer to wrap v8 behind the JavaScriptCore API
Product: WebKit Reporter: Balazs Kelemen <kbalazs>
Component: PlatformAssignee: Balazs Kelemen <kbalazs>
Status: RESOLVED WONTFIX    
Severity: Normal CC: abarth, ap, gaborb, hausmann, levin+threading, sam, vestbo, webkit.review.bot
Priority: P2    
Version: 528+ (Nightly build)   
Hardware: Unspecified   
OS: Unspecified   
Bug Depends on:    
Bug Blocks: 76778, 84457    
Attachments:
Description Flags
Patch none

Balazs Kelemen
Reported 2012-05-30 10:22:41 PDT
The goal is to be able to use v8 in WebKit2 without the need to fork the InjectedBundle (of WTR).
Attachments
Patch (87.26 KB, patch)
2012-05-31 07:52 PDT, Balazs Kelemen 		no flags
DetailsFormatted DiffDiff
View All Add attachment proposed patch, testcase, etc.
Balazs Kelemen
Comment 1 2012-05-31 07:52:59 PDT
Created attachment 145093 [details] Patch
Alexey Proskuryakov
Comment 2 2012-05-31 08:48:57 PDT
This is part of "make WebKit2 work with v8" project that has not been accepted on webkit-dev. Why is it ok to post this for review?
Balazs Kelemen
Comment 3 2012-05-31 09:00:27 PDT
(In reply to comment #2) > This is part of "make WebKit2 work with v8" project that has not been accepted on webkit-dev. Why is it ok to post this for review? It has not been rejected unambiguously as well. I finished implementing this so the community can at least see how it looks like. It's better to decide on something that actually exist.
Alexey Proskuryakov
Comment 4 2012-05-31 09:17:42 PDT
That's not how I read the discussion. Anyway, if this is your goal, I suggest clearing the review flag and posting to webkit-dev again with a link. This patch is not something that should be r+'ed and landed without community buy-in.
Balazs Kelemen
Comment 5 2012-05-31 09:39:54 PDT
(In reply to comment #4) > That's not how I read the discussion. > > Anyway, if this is your goal, I suggest clearing the review flag and posting to webkit-dev again with a link. This patch is not something that should be r+'ed and landed without community buy-in. Mail sent. I would better like to keep the r? because otherwise nobody will notice the patch (according to my experiences).
Adam Barth
Comment 6 2012-05-31 10:51:33 PDT
This wrapper looks like it's pretty slow. Have you benchmarked it?
Balazs Kelemen
Comment 7 2012-06-01 11:32:53 PDT
(In reply to comment #6) > This wrapper looks like it's pretty slow. Have you benchmarked it? Not yet. Considering that the wrapper is only taking part in binding layouttestcontroller and friends to the js environment I don't really care about speed. Note that the actual script execution is happening in WebCore, not WebKit2. Most of the js related thinks in WebKit2 are just for serving the needs of the injected bundle. As long as a test session does not takes considerably more time than before, I don't think we should worry about speed.
Alexey Proskuryakov
Comment 8 2012-06-05 00:00:58 PDT
WONTFIX per complete lack of buy-in on webkit-dev.
Note You need to log in before you can comment on or make changes to this bug.