the border-radius "subtypes" border-radius-topleft, border-radius-topright, border-radius-bottomleft and border-radius-bottomright are not implemented, as shown by the forthcoming testcase.
Created attachment 8634 [details] Testcase
cc-ing Beth, as she fixed the other border-radius bugs.
This testcase is plain wrong :). Should be -webkit-border-top-left-radius etc. Only one not working seems to be bottom-right, but i'm researching that...
Created attachment 9798 [details] new test case... hangs my webkit
making P1, this hangs webkit reproducibly
Created attachment 10024 [details] sample of hang this is the end of the main thread: 696 WebCore::GraphicsContext::drawArc(WebCore::IntRect const&, float, int, int)   696 CGContextStrokePath     696 CGContextDrawPath        696 ripc_DrawPath          696 ripr_Path            696 CGPathApply             696 ripr_path_stroke                696 path_dash_iterate                  696 cube_length                   696 cube_length
whoa! what happened there?
<rdar://problem/4680169>
(In reply to comment #7) > whoa! what happened there? Looks like a regression from Bug 9630, or are you using an old WebKit build, Alice? The   entity is the same as a non-breaking space entity ( ). (See comment #6.)
*** Bug 10479 has been marked as a duplicate of this bug. ***
As far as I can tell, the hang is the only issue here (the original bug was invalid). As I mentioned in bug 10479, the problem is a consequence of division by zero in drawArc when it's asked to draw a zero-radius dashed arc. It ends up passing some huge value down to CG causing it to hand (I think someone said that it actually returns after 15 minutes).
Created attachment 12031 [details] Patch
Comment on attachment 12031 [details] Patch r=me
Landed in r18456.