Bug 165086 – Make CanvasRenderingContext2D use WebIDL unions / Variants for createPattern and drawImage

• Status: RESOLVED
• Resolution: FIXED
• Component: New Bugs

[Sam Weinig]

Make CanvasRenderingContext2D use WebIDL unions / Variants for createPattern and drawImage

[Sam Weinig]

Created attachment 295475 [details]
Patch

[Chris Dumez]

Comment on attachment 295475 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=295475&action=review

r=me but please check GTK/EFL build.

> Source/WebCore/html/canvas/CanvasRenderingContext2D.idl:137
> +    // FIXME: All the unrestricted float arguments below should be unrestricted doubles.

I think we can get away with doing this change in tis patch (only in the IDL).

[Darin Adler]

Comment on attachment 295475 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=295475&action=review

Some compilation problems with gcc on the GTK and EFL bots. Not sure what’s going on there. Maybe [&] isn’t supposed to capture "this"?

> Source/WebCore/html/canvas/CanvasRenderingContext2D.h:58
> +using CanvasImageSource = Variant<RefPtr<HTMLImageElement>, RefPtr<HTMLVideoElement>, RefPtr<HTMLCanvasElement>>;

So sad that these all have to be RefPtr, when none are allowed to be null.

[Sam Weinig]

Created attachment 295501 [details]
Patch

[Sam Weinig]

Created attachment 295539 [details]
Patch

[Sam Weinig]

(In reply to comment #3)
> Comment on attachment 295475 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=295475&action=review
> 
> Some compilation problems with gcc on the GTK and EFL bots. Not sure what’s
> going on there. Maybe [&] isn’t supposed to capture "this"?

It's a GCC bug (https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61636) I can work around it using an explicit this->.

> 
> > Source/WebCore/html/canvas/CanvasRenderingContext2D.h:58
> > +using CanvasImageSource = Variant<RefPtr<HTMLImageElement>, RefPtr<HTMLVideoElement>, RefPtr<HTMLCanvasElement>>;
> 
> So sad that these all have to be RefPtr, when none are allowed to be null.

Yeah, it's something I would like to improve at some point, but I am not quite sure how yet.

[Sam Weinig]

Committed r209048: <http://trac.webkit.org/changeset/209048>