23987 2013-12-03 18:09:44 +0000 Empty path and drawSystemFocusRing 2014-01-20 23:46:32 +0000 1 1 1 Unclassified HTML WG CR HTML Canvas 2D Context unspecified PC Windows NT CLOSED FIXED a11y_canvas P2 minor --- 1 plh jaymunro jmann mark.sadecki+w3c mike public-html-admin public-html-wg-issue-tracking schwer public-html-bugzilla oldest_to_newest 97067 0 plh 2013-12-03 18:09:44 +0000 What is expected from drawSystemFocusRing if the path is empty, e.g. [[ var ctx = canvas.getContext("2D"); ctx.beginPath(); myInput.focus(); ctx.drawSystemFocusRing(myInput); ]] 97078 1 jaymunro 2013-12-03 21:29:37 +0000 Reassigning to Canvas 2D Context Level 2 since drawSystemFocusRing and drawCustomFocusRing are at risk for CR. 97130 2 schwer 2013-12-04 15:14:55 +0000 There is always a default path: The context always has a current default path. There is only one current default path, it is not part of the drawing state. The current default path is a path, as described above. I am trying to understand the error condition you are getting at. 97132 3 plh 2013-12-04 16:24:47 +0000 My example has an empty path. Unless I'm misreading the specification, the default path doesn't enter into the equation here. beginPath() was invoked and nothing has been added into the path. You may be suggesting that, if the path is empty, then the previous intended path in the context is used. 97157 4 jaymunro 2013-12-04 19:49:37 +0000 I jumped the gun on moving to L2, there are still ongoing discussions on the focus ring methods. Moving back to CR until discussions are complete. 98097 5 jaymunro 2014-01-07 17:39:33 +0000 If the path was empty, there isn't a path to act on at this point, though the intent is there with beginPath(). The issue appears to be whether to use the last path created before beginPath() was called? I would expect it to just do nothing, and treat it as a bug in code if an empty path was created. 98140 6 plh 2014-01-08 18:31:51 +0000 > I would expect it to just do nothing, and treat it as a bug in code if an > empty path was created. That would be my expectation as well and a clarification in the specification would be appropriate imho. 98634 7 jaymunro 2014-01-16 17:34:28 +0000 Updated the spec to include a new first step stating that if the current path had zero subpaths, then abort the steps. 98800 8 jmann 2014-01-20 23:46:32 +0000 Based on Canvas SubGroup discussion, http://www.w3.org/2014/01/20-html-a11y-minutes.html, we have agreed to close this bug as the processing model step 1 resolves this issue.