[Webkit-unassigned] [Bug 53055] REGRESSION: Rendering artifacts on a rotated, pattern filled shape

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=53055


Nikolas Zimmermann <zimmermann at kde.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
 Attachment #104612|                            |review-
               Flag|                            |




--- Comment #20 from Nikolas Zimmermann <zimmermann at kde.org>  2011-08-20 00:13:54 PST ---
(From update of attachment 104612)
View in context: https://bugs.webkit.org/attachment.cgi?id=104612&action=review

I'm not sure about the SystemInterface stuff, as I have no idea how it works, but I trust you on that. Still some questions:

> Source/WebCore/ChangeLog:11
> +        Introduce wkCGPatternCreateWithImageAndTransform, and make use of it
> +        when tiling patterns in both directions. This helps to prevent
> +        pixel-cracking along pattern tiling boundaries.

It's so wonderful to have someone from Apple working on SVG, with the abilities to do such things as extend the WebKitSystemInterface :-)

> Source/WebCore/ChangeLog:15
> +        Fix a rounding error in the size of the temporary image created to
> +        render SVG pattern tiles into, which was introduced in r65665.
> +        We now round SVG tile image size up in all cases.

I think I introduced the ceilf logic on purpose. The ceilf alone makes no sense, but it's used in conjunction with following code: in PassOwnPtr<ImageBuffer> RenderSVGResourcePattern::createTileImage:

...
    if (!SVGImageBufferTools::createImageBuffer(absoluteTileBoundaries, clampedAbsoluteTileBoundaries, tileImage, ColorSpaceDeviceRGB))
        return nullptr;

    GraphicsContext* tileImageContext = tileImage->context();
    ASSERT(tileImageContext);

    // The image buffer represents the final rendered size, so the content has to be scaled (to avoid pixelation).
    tileImageContext->scale(FloatSize(absoluteTileBoundaries.width() / tileBoundaries.width(),
                                      absoluteTileBoundaries.height() / tileBoundaries.height()));

Is this obsolete now? How are other platforms affected by your change to avoid the ceil?
Maskers contain similar code to avoid fix-up rounding differences by the ceil.

> Source/WebCore/ChangeLog:17
> +        No new tests. (OOPS!)

Can you add a test?

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