Bug 140198 – WKWebViewConfiguration should provide public API to enable private browsing

• Component: WebKit2
• Status: RESOLVED FIXED

[Eugene But]

WKWebView does not support custom protocol handlers, so there is no way to implement Private browsing for a Web Browser app. Private Browsing is very important privacy feature and it is supported by Safari.

This change shows an example of Private Browsing implementation, but it uses private API:
https://github.com/WebKit/webkit/commit/512bc88e867086713820d4eb86d736586c20586f

[Eugene But]

Radar ID: 17238307

[Eugene But]

Created attachment 244524 [details]
Patch

[Beth Dakin]

Comment on attachment 244524 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=244524&action=review

Sorry, meant to r-! The name _privateBrowsingEnabled is not good.

> Tools/MiniBrowser/mac/AppDelegate.m:-106
> -    privateConfiguraton._websiteDataStore = [_WKWebsiteDataStore nonPersistentDataStore];

Why is the removed?

[Eugene But]

Comment on attachment 244524 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=244524&action=review

>> Tools/MiniBrowser/mac/AppDelegate.m:-106
>> -    privateConfiguraton._websiteDataStore = [_WKWebsiteDataStore nonPersistentDataStore];
> 
> Why is the removed?

privateConfiguraton._privateBrowsingEnabled = YES; code will substitute internal data store with nonPersistentDataStore. So this change should have no functional effect.

[Eugene But]

>>> The name _privateBrowsingEnabled is not good.
The name follows the same patten as [WKWebViewConfiguration _featureCounterEnabled]. Please let me know if you don't like presence of leading underscore or using "privateBrowsing" term? Would appreciate if you can suggest a name which is better from your perspective.

[Sam Weinig]

Instead of using a property, I think the direction we will be going is exposing WKWebsiteDataStore as API and allowing setting a non persistent data store on the configuration.

[Eugene But]

Thank you for reply. In case if there already exists a private API which solves the problem, would it be useful to submit a patch which makes that API public?

[Sam Weinig]

(In reply to comment #7)
> Thank you for reply. In case if there already exists a private API which
> solves the problem, would it be useful to submit a patch which makes that
> API public?

No, we will make it public when we feel it is ready from prime time. No need for a patch.

[Anders Carlsson]

This has been done now.