Introduce WTF::createRef<T>(Args&&...)

[Zan Dobersek]

Introduce WTF::createRef<T>(Args&&...)

[Zan Dobersek]

Created attachment 253727 [details]
Test patch

Checking how the EWSs handle it (specifically the friend template declaration).

[Zan Dobersek]

Created attachment 253847 [details]
Test patch

Further iteration.

[Zan Dobersek]

Created attachment 253900 [details]
Test patch

Another iteration, testing out modifying AsyncRequestImpl creation.

[WebKit Commit Bot]

Attachment 253900 [details] did not pass style-queue:


ERROR: Source/WebKit2/Shared/AsyncRequest.h:77:  Extra space before ( in function call  [whitespace/parens] [4]
Total errors found: 1 in 11 files


If any of these errors are false positives, please file a bug against check-webkit-style.

[Zan Dobersek]

Created attachment 254279 [details]
Patch

Another run through the EWS.

[WebKit Commit Bot]

Attachment 254279 [details] did not pass style-queue:


ERROR: Source/WebKit2/Shared/AsyncRequest.h:77:  Extra space before ( in function call  [whitespace/parens] [4]
Total errors found: 1 in 11 files


If any of these errors are false positives, please file a bug against check-webkit-style.

[Zan Dobersek]

Created attachment 254413 [details]
Patch

More of an RFC about the idea, approach and namings.

[WebKit Commit Bot]

Attachment 254413 [details] did not pass style-queue:


ERROR: Source/WebKit2/Shared/AsyncRequest.h:77:  Extra space before ( in function call  [whitespace/parens] [4]
Total errors found: 1 in 12 files


If any of these errors are false positives, please file a bug against check-webkit-style.

[Anders Carlsson]

Comment on attachment 254413 [details]
Patch

This is an interesting idea, but we _could_ just put it as a static member function on RefCounted directly - that way the friend declaration would be less awkward.

[Zan Dobersek]

(In reply to comment #9)
> Comment on attachment 254413 [details]
> Patch
> 
> This is an interesting idea, but we _could_ just put it as a static member
> function on RefCounted directly - that way the friend declaration would be
> less awkward.

Could you please elaborate on how the friend declaration would look? I ended up with a pretty similar declaration.

I chose createRef so that the name indicates the type of the returned value, much like std::make_unique<>() indicates that std::unique_ptr<> is returned.

[Anders Carlsson]

(In reply to comment #10)
> 
> Could you please elaborate on how the friend declaration would look? I ended
> up with a pretty similar declaration.
> 

friend class RefCounted<MyClass>;

[Zan Dobersek]

Created attachment 254583 [details]
Patch

Adds RefCounted<T>::create(), friends just RefCounted<T>.