[Webkit-unassigned] [Bug 32499] Add client based Geolocation provider

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=32499





--- Comment #10 from Sam Weinig <sam at webkit.org>  2009-12-15 10:54:49 PST ---
(In reply to comment #9)
> Eric, thanks for adding me to the bug.
> 
> Sam, I have a few questions about the patch ...
> - Can you explain the motivation behind adding an alternative to the current
> Geolocation service? What are the advantages of the client-based provider?

The intent of the client based provider is to allow an embedding layer more
control of the geolocation implementation without egregious layering
violations.

> Do you intend that we support both types in the future, or will the existing
> Geolocation service be deprecated?

It will only be deprecated if the new way works better for others.

> - Why do you introduce new position and error types? Why can't the existing
> Geoposition and PositionError types be used with the client provider? I see
> you've excluded the TIMEOUT error code from the new error type as it's not
> needed by the provider. Why have you left in the PERMISSION_DENIED code? This
> too is handled separately by the Geolocation class.

I did this because it is an existing layering violation for the platform layer
to know about these DOM types.  In general, I don't think it is a good idea to
conflate data coming from the platform and DOM types due to the caching
behavior of JS objects.

> - What's the reason for making Geolocation::suspend/resume() private? These
> methods are used as public methods on Android, where we suspend the Geolocation
> service when the browser tab is in the background. Also, it looks like the
> implementation of these methods is missing.

I made them private to find out if anyone was using it, and since it compiled I
assumed no one was.  I meant to remove them.  I will add them back if you are
using them.

> Finally, could you CC me on all future Geolocation changes please? I'm working
> on Geolocation in Android and have sent a number of patches recently.
> 

Sure.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.