[whatwg] <table> Relation between "bordercolor" and "border" attributes

Kang-Hao (Kenny) Lu kennyluck at csail.mit.edu
Wed Mar 7 19:08:54 PST 2012

(12/03/06 17:58), Kishore Bolisetty wrote:
> But it doesn't talks about the behaviour - what if border is not specified
> but bordercolor is specified? Looks like browsers have taken their own
> implementations on this, Opera and Mozilla displays bordercolor only if
> border is specified, where as Safari displays bordercolor assuming a
> default border.

I'll add one more testing result. IE9 and IE quirks mode match Opera and
Mozilla's behavior.

> Which one is correct behaviour? Is it not necessary to explicitly state
> this in the spec?

I don't know what is correct but Safari is certainly less interoperable
here, and the spec matches the interoperable behavior. That is, I think
the spec is pretty clear: bordercolor only changes the border color, not
whether the border shows up or not ('border-style') or border width

(12/03/07 0:35), Tab Atkins Jr. wrote:
> This behavior depends on the browser's default stylesheet, and whether
> they default borders to "border-style: solid; border-color:
> currentColor;" or "border-style: solid; border-width: 0px;".  Browsers
> are allowed to do either.

I don't understand your examples. I think you are probably suggesting
that browsers are allowed to put "table[bordercolor] { border-style:
solid}" in the default stylesheet, but I don't think that's the case
since it makes the browser that does this not "supporting the suggested
default rendering".


More information about the whatwg mailing list