Setting a stroke-width on a rectangle results in a stroke that is not even on all sides of the rectangle.
Created attachment 17585 [details] testcase
Bug confirmed. It's a clipping problem though. If you pan around you can clearly see the stroke is set correctly - though it's partly clipped aways.
(In reply to comment #2) > Bug confirmed. It's a clipping problem though. If you pan around you can > clearly see the stroke is set correctly - though it's partly clipped aways. > Eek I am too tired - was too quick. The rendering looks correct on Tiger! So it must be a CG/windows specific bug - as you described it on IRC. Sorry for the trouble.
After talking with Nico more, it seems that there really is a bug here (in his opinion), which is that the clipping of the top/left sides of the stroke should not happen when padding is set on the <svg> element.
(In reply to comment #4) > After talking with Nico more, Niko, even.
This behavior is 100% correct, IMO. I'm not sure if we should be respecting "padding" in an SVG stand-alone document however. <svg> has an implicit clip: http://www.w3.org/TR/SVG/masking.html#OverflowAndClipProperties When an outermost SVG 'svg' element is stand-alone or embedded inline within a parent XML grammar which does not use CSS layout [CSS2-LAYOUT] or XSL formatting [XSL], the 'overflow' property on the outermost 'svg' element is ignored for the purposes of visual rendering and the initial clipping path is set to the bounds of the initial viewport. So the stroke for the rect is getting clipped off the top and left. If the SVG were not width/height 100%, but rather only as large as the rect, the bottom and right would also have been clipped. I suggest this is closed as INVALID.
Opera agrees with our rendering. Firefox does too, but ignores padding. Closing as invalid.