Bug 149516 – [Win] Support composited content in WebView render-to-context methods

• Component: Layout and Rendering
• Status: RESOLVED FIXED

[Brent Fulgham]

The current implementations of WebView::paintDocumentRectToContext and WebView::paintScrollViewRectToContextAtPoint only draw GDI-based content into the passed device context. This causes CALayer-based information to be skipped, since it typically draws on its own 60-hz timer outside of the normal paint operation.

The purpose of these two methods is to allow a WebView to be rendered into a specific device context, usually representing a bitmap or back buffer. Consequently, we should perform a draw operation from the CALayerTree into the context when this method is called.

[Brent Fulgham]

<rdar://problem/22635080>

[Brent Fulgham]

Created attachment 261862 [details]
Patch

[Simon Fraser (smfr)]

Comment on attachment 261862 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=261862&action=review

> Source/WebKit/win/WebView.cpp:6165
> +            ::ValidateRect(m_viewWindow, nullptr);

Is it correct to validate a rect in the window when you may be painting into an unrelated DC?

> Source/WebKit/win/WebView.cpp:6192
> +            ::ValidateRect(m_viewWindow, nullptr);

Ditto. Also why isn't this code shared?

[Brent Fulgham]

Comment on attachment 261862 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=261862&action=review

>> Source/WebKit/win/WebView.cpp:6165
>> +            ::ValidateRect(m_viewWindow, nullptr);
> 
> Is it correct to validate a rect in the window when you may be painting into an unrelated DC?

I think you are right. This should not be used here.

>> Source/WebKit/win/WebView.cpp:6192
>> +            ::ValidateRect(m_viewWindow, nullptr);
> 
> Ditto. Also why isn't this code shared?

It should be. I'll revise that.

[Brent Fulgham]

Committed r190235: <http://trac.webkit.org/changeset/190235>