[iOS] Programmatic paste access should be granted when copying and pasting within the same origin

[Wenson Hsieh]

<rdar://problem/48134710>

[Wenson Hsieh]

Created attachment 363009 [details]
First pass

[EWS Watchlist]

Thanks for the patch. If this patch contains new public API please make sure it follows the guidelines for new WebKit2 GTK+ API. See http://trac.webkit.org/wiki/WebKitGTK/AddingNewWebKit2API

[Wenson Hsieh]

(In reply to Build Bot from comment #2)
> Thanks for the patch. If this patch contains new public API please make sure
> it follows the guidelines for new WebKit2 GTK+ API. See
> http://trac.webkit.org/wiki/WebKitGTK/AddingNewWebKit2API

Doesn't seem relevant here, since the method is private.

[Ryosuke Niwa]

Comment on attachment 363009 [details]
First pass

View in context: https://bugs.webkit.org/attachment.cgi?id=363009&action=review

r=me. Please be sure to fix WPE build failure before landing.

> Source/WebCore/dom/DOMPasteAccess.h:28
> +#import <wtf/Forward.h>

We should be using #include instead. That's why WPE isn't happy.

> Source/WebCore/page/Frame.cpp:686
> +        return response != DOMPasteAccessResponse::DeniedForGesture;

I think we should use a switch and explicitly list all values.
Otherwise, we could miss updating this location when we add a new kind of denial.

> Source/WebKit/UIProcess/ios/WKContentViewInteraction.mm:5032
> +    NSIndexSet *indices = [NSIndexSet indexSetWithIndexesInRange:NSMakeRange(0, [pasteboard numberOfItems])];

Use auto?

> Source/WebKit/UIProcess/ios/WKContentViewInteraction.mm:5033
> +    NSArray<NSData *> *allCustomData = [pasteboard dataForPasteboardType:@(WebCore::PasteboardCustomData::cocoaType()) inItemSet:indices];

Ditto.

> Source/WebKit/UIProcess/ios/WKContentViewInteraction.mm:5037
> +    BOOL foundAtLeastOneMatchingIdentifier;

Huh, does objective-C always initialize BOOL to NO?

[Wenson Hsieh]

Comment on attachment 363009 [details]
First pass

View in context: https://bugs.webkit.org/attachment.cgi?id=363009&action=review

>> Source/WebCore/dom/DOMPasteAccess.h:28
>> +#import <wtf/Forward.h>
> 
> We should be using #include instead. That's why WPE isn't happy.

Indeed — fixed.

>> Source/WebCore/page/Frame.cpp:686
>> +        return response != DOMPasteAccessResponse::DeniedForGesture;
> 
> I think we should use a switch and explicitly list all values.
> Otherwise, we could miss updating this location when we add a new kind of denial.

Good idea — changed to use switch case.

>> Source/WebKit/UIProcess/ios/WKContentViewInteraction.mm:5032
>> +    NSIndexSet *indices = [NSIndexSet indexSetWithIndexesInRange:NSMakeRange(0, [pasteboard numberOfItems])];
> 
> Use auto?

👍

>> Source/WebKit/UIProcess/ios/WKContentViewInteraction.mm:5033
>> +    NSArray<NSData *> *allCustomData = [pasteboard dataForPasteboardType:@(WebCore::PasteboardCustomData::cocoaType()) inItemSet:indices];
> 
> Ditto.

👍

>> Source/WebKit/UIProcess/ios/WKContentViewInteraction.mm:5037
>> +    BOOL foundAtLeastOneMatchingIdentifier;
> 
> Huh, does objective-C always initialize BOOL to NO?

Good catch — BOOL is only initialized to NO for instance variables! Changed this to be NO.

[Wenson Hsieh]

Created attachment 363387 [details]
Patch for EWS

[Wenson Hsieh]

Created attachment 363396 [details]
Fix WPE build

[WebKit Commit Bot]

Comment on attachment 363396 [details]
Fix WPE build

Clearing flags on attachment: 363396

Committed r242317: <https://trac.webkit.org/changeset/242317>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.