Table rows do not have a size in the render tree currently. This needs to be fixed, since it is causing the following issues: (1) Backgrounds don't work properly on table rows (2) offsetWidth returns 0 (3) Anchor jumping to a row doesn't work
Apple Bug: rdar://3613374/
http://bugs.kde.org/show_bug.cgi?id=70326#c17, FYI
Possibly related to this is the bug whereby tr:hover CSS styles don't apply if the mouse is within the row bounds but outside any of it's cells (i.e. there is some cellspacing, or the row doesn't have a full complement of table cells).
Created attachment 3775 [details] Test case The bug also affects TBODY. TBODY has a zero offsetWidth and correct offsetHeight, TR zero offsetWidth & offsetHeight.
Yes, this is a major architectural flaw in tables. It also affects background painting (try doing a non- repeating background on a row or body).
Created attachment 3789 [details] For the curious :) test case for non-repeating backgrounds So the TR background-image is duplicated in all TDs. There's also something fishy with a border on a TR, which gets semi-replicated (only right & bottom) on TDs, but only if table has border=1.
*** Bug 7242 has been marked as a duplicate of this bug. ***
Upon further investigation, it appears that it should not be possible to hit test a row. Other browsers do not allow you to hit test it, so tr:hover is *not* supposed to work if you are in the space between cells. Moreover background painting is also not done on the row. All browsers inherit the backgrounds into the cells. It looks like Firefox - in strict mode only - will honor the non-repeating bg, but they still don't do it correctly. In general, our agreed strategy with tables is to match WinIE, and so the background painting is not an issue any more (nor is the hover). Issues that are genuine, however, are the anchor bug (which is filed as 5393) and the offsetWidth/Height issue. These are now resolved with the patch attached to 5393.
Marking as fixed, according to the above comment.