In discussions with Sam Weinig, Dave Hyatt, and John Sullivan, I've come up with a new SPI for dealing with user scripts and stylesheets and isolated worlds. This bug covers implementing this SPI.
Created attachment 43066 [details] Small clean-up in WebView's user content functions
Comment on attachment 43066 [details] Small clean-up in WebView's user content functions r=me
Created attachment 43084 [details] Remove worldIDs from WebKit's SPI
Comment on attachment 43084 [details] Remove worldIDs from WebKit's SPI I don't have time to do a full review. But WebJSWorld is hard to read. This would have been better named WebScriptWorld. (To match WebScriptObject.)
Comment on attachment 43084 [details] Remove worldIDs from WebKit's SPI collectionWithWorld should be collectionWithScriptWorld if you take my WebScriptWorld suggestion. Same with initWithWorld.
- (JSGlobalContextRef)globalContextInWorld:(WebJSWorld *)world; reads better as: - (JSGlobalContextRef)globalContextForScriptWorld:(WebScriptWorld *)world;
Thanks for the reviews, Sam and Hyatt! I'm going to make Tim's suggested renames and then land these two patches.
Comment on attachment 43066 [details] Small clean-up in WebView's user content functions Landed in r50906 <http://trac.webkit.org/changeset/50906>
Comment on attachment 43084 [details] Remove worldIDs from WebKit's SPI Landed in r50907 <http://trac.webkit.org/changeset/50907>
This doesn't block bug 31124 anymore, now that the patch in attachment 43084 [details] has landed.
These patches claim to have been landed.
Yes, but there are more coming (eventually).
Why not use one patch per bug? It's hard to follow these bugs once they go epic.
Comment on attachment 43084 [details] Remove worldIDs from WebKit's SPI Remove from pending-commit.
(In reply to comment #13) > Why not use one patch per bug? It's hard to follow these bugs once they go > epic. OK, I'll do that.
OK, I have one more teeny patch that really does belong in this bug. But then I'll move to a new one, I promise!
Created attachment 43162 [details] Finish replacing worldIDs with world objects
Comment on attachment 43162 [details] Finish replacing worldIDs with world objects r=me
Committed r50943: <http://trac.webkit.org/changeset/50943>