Bug 45107 – SVG atlas map slider doesn't work properly

• Status: RESOLVED
• Resolution: FIXED
• Component: SVG

[Nikolas Zimmermann]

Created attachment 66360 [details]
Partly reduced testcase

The SVG Atlas Map from IE9 Test Center (vis.uell.net) contains a slider, which doesn't respond to move events as expected.
A mousemove listener is attached to the <svg> root element, and events are only fired, when moving the mouse over areas that contain graphical elements.

SVG 1.1 2nd edition also allows to fire events when hovering over container elements (like <svg>). The old behaviour implemented in WebKit now, also doesn't make any sense.

This was brought to attention @ SVG Open 2010, Michael Neutze the author of atlas map was so kind to reduce the problem, attaching his testcase (not really reduced, but I'm containing two reduced testcases in my patch)

[Nikolas Zimmermann]

*** Bug 45170 has been marked as a duplicate of this bug. ***

[Nikolas Zimmermann]

Created attachment 66490 [details]
Patch

Fixes all problems with the SVG "Bevoelkerungspyramide".

[Dirk Schulze]

Comment on attachment 66490 [details]
Patch

I'd write the comment before the if, but that's  matter of opinion. r=me.

[Nikolas Zimmermann]

Thanks Dirk, landed in r66731.

Michael, please retry with a nightly >= 66731, once it's available.

[Michael N.]

ok, this looks very good and is a huge improvement.
tested in r66819 on OSX 10.6.4

I am still not sure if text that's set to "pointer-events: none" should be selectable but I now know how to cover that with prventDefault.

[Nikolas Zimmermann]

(In reply to comment #5)
> ok, this looks very good and is a huge improvement.
> tested in r66819 on OSX 10.6.4
Good to hear, glad it works fine.

> I am still not sure if text that's set to "pointer-events: none" should be selectable but I now know how to cover that with prventDefault.
Hm, can you try to look up a relevant section in the SVG spec? My feeling is that you're right, if you can find any part of the spec that supports this view, please file another bug.
Happy to fix, it you find evidence that we're wrong :-) (despite what other browsers do)

[Andy Estes]

<rdar://problem/8432245>