Bug 173150 – Add API::GeolocationProvider

• Component: WebKit2
• Status: RESOLVED FIXED

[Carlos Garcia Campos]

We will use this instead of the C API in GTK+.

[Carlos Garcia Campos]

Created attachment 312423 [details]
Patch

[Alex Christensen]

Comment on attachment 312423 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=312423&action=review

> Source/WebKit2/UIProcess/WebGeolocationManagerProxy.h:84
> +    std::unique_ptr<API::GeolocationProvider> m_provider;

I think this is wrong.  Usually we have the API object own the WebKit namespace object.

[Carlos Garcia Campos]

Created attachment 312677 [details]
Rebased patch

It should apply now

[Alex Christensen]

Comment on attachment 312677 [details]
Rebased patch

View in context: https://bugs.webkit.org/attachment.cgi?id=312677&action=review

> Source/WebKit2/UIProcess/WebGeolocationManagerProxy.h:50
> +    void setProvider(std::unique_ptr<API::GeolocationProvider>);

&&

> Source/WebKit2/UIProcess/WebGeolocationProvider.h:44
> +class WebGeolocationProvider : public API::GeolocationProvider,  API::Client<WKGeolocationProviderBase> {

extra space

[Carlos Garcia Campos]

(In reply to Alex Christensen from comment #4)
> Comment on attachment 312677 [details]
> Rebased patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=312677&action=review
> 
> > Source/WebKit2/UIProcess/WebGeolocationManagerProxy.h:50
> > +    void setProvider(std::unique_ptr<API::GeolocationProvider>);
> 
> &&

That was also my idea at first, but then again I followed what we do for all other clients. In this particular case, I think that what we have is equivalent to && because we always pass a std::make_unique<> (or nullptr) to the function, so there aren't copies anyway.

> > Source/WebKit2/UIProcess/WebGeolocationProvider.h:44
> > +class WebGeolocationProvider : public API::GeolocationProvider,  API::Client<WKGeolocationProviderBase> {
> 
> extra space

Good catch!

[Darin Adler]

(In reply to Carlos Garcia Campos from comment #5)
> (In reply to Alex Christensen from comment #4)
> > > Source/WebKit2/UIProcess/WebGeolocationManagerProxy.h:50
> > > +    void setProvider(std::unique_ptr<API::GeolocationProvider>);
> > 
> > &&
> 
> That was also my idea at first, but then again I followed what we do for all
> other clients. In this particular case, I think that what we have is
> equivalent to && because we always pass a std::make_unique<> (or nullptr) to
> the function, so there aren't copies anyway.

I don’t know if we ever resolved this. Here is what Scott Meyers has to say:

http://scottmeyers.blogspot.com/2014/07/should-move-only-types-ever-be-passed.html

[Carlos Garcia Campos]

Created attachment 312751 [details]
Patch for landing

[Carlos Garcia Campos]

Committed r218165: <http://trac.webkit.org/changeset/218165>