SVG panning y-axis is flipped on Windows and WebKit2/Mac

[Tim Horton]

1. Open a large SVG document, for example, http://upload.wikimedia.org/wikipedia/commons/7/74/Timeline_of_web_browsers.svg
2. Pan up and down with shift-drag.

Expected result:

Image pans with mouse.

Actual result:

Image pans in opposite direction of mouse in y-axis.

I have a patch which I will attach shortly.

<rdar://problem/9908012> for WK2
<rdar://problem/5700027> for Windows

[Tim Horton]

Created attachment 103262 [details]
Patch

I'm going to test it in Windows first and see if it fixes the problem there before committing, and I'll add that Radar # too if it does.

[WebKit Review Bot]

Comment on attachment 103262 [details]
Patch

Attachment 103262 [details] did not pass chromium-ews (chromium-xvfb):
Output: http://queues.webkit.org/results/9328956

New failing tests:
svg/custom/pan-direction.svg

[W. James MacLean]

*** Bug 66315 has been marked as a duplicate of this bug. ***

[W. James MacLean]

Created attachment 104074 [details]
Chromium-linux pixel results for patch.

[W. James MacLean]

(In reply to comment #1)
> Created an attachment (id=103262) [details]
> Patch
> 
> I'm going to test it in Windows first and see if it fixes the problem there before committing, and I'll add that Radar # too if it does.

Any idea when this patch will land? There are outstanding bugs filed against it that would be nice to close. Let me know if I can help.

[Tim Horton]

(In reply to comment #5)
> (In reply to comment #1)
> > Created an attachment (id=103262) [details] [details]
> > Patch
> > 
> > I'm going to test it in Windows first and see if it fixes the problem there before committing, and I'll add that Radar # too if it does.
> 
> Any idea when this patch will land? There are outstanding bugs filed against it that would be nice to close. Let me know if I can help.

I'll try to get it through as soon as I can this week; I have a backlog of patches, so I should spend some time getting them landed.

[Tim Horton]

(In reply to comment #6)
> (In reply to comment #5)
> > (In reply to comment #1)
> > > Created an attachment (id=103262) [details] [details] [details]
> > > Patch
> > > 
> > > I'm going to test it in Windows first and see if it fixes the problem there before committing, and I'll add that Radar # too if it does.
> > 
> > Any idea when this patch will land? There are outstanding bugs filed against it that would be nice to close. Let me know if I can help.
> 
> I'll try to get it through as soon as I can this week; I have a backlog of patches, so I should spend some time getting them landed.

Oddly, the image you attached is pixel-identical to the one in the test; I wonder why the bot says cr-linux doesn't match my result.

[Tim Horton]

Created attachment 105543 [details]
with cr-linux results (?)

[Darin Adler]

Comment on attachment 105543 [details]
with cr-linux results (?)

View in context: https://bugs.webkit.org/attachment.cgi?id=105543&action=review

> Source/WebCore/page/EventHandler.cpp:477
> +            FrameView* view = m_frame->view();
> +            static_cast<SVGDocument*>(m_frame->document())->startPan(view->windowToContents(event.event().pos()));

Why the local variable? I think it would read better without it.

> Source/WebCore/page/EventHandler.cpp:1566
> +        FrameView* view = m_frame->view();
> +        static_cast<SVGDocument*>(m_frame->document())->updatePan(view->windowToContents(m_currentMousePosition));

Ditto.

> Source/WebCore/page/EventHandler.cpp:1688
> +        FrameView* view = m_frame->view();
> +        static_cast<SVGDocument*>(m_frame->document())->updatePan(view->windowToContents(m_currentMousePosition));

Ditto.

[Tim Horton]

Landed in 94114.

[W. James MacLean]

(In reply to comment #10)
> Landed in 94114.

Thanks, looks good!

[Simon Fraser (smfr)]

Caused bug 67318?

[Alexey Proskuryakov]

Seems unlikely.