- From: Philip Walton <philip@philipwalton.com>
- Date: Mon, 1 Dec 2014 15:35:02 -0800
- To: "Tab Atkins Jr." <jackalmage@gmail.com>
- Cc: www-style list <www-style@w3.org>
- Message-ID: <CAGRhNhUJqciVkv4SqPHJ8fkT8N7xW2Rp=cJs9YRPjk+T-orMaQ@mail.gmail.com>
> Per <http://dev.w3.org/csswg/css-scoping/#cascading>, rules from outside a > shadow tree always win against rules within a shadow tree, when they > attempt to style the same element. Hmmm, then it seems like there may be a bug in blink. In my example, the selector `:host(element-name)` wins over the universal selector. Should that not be the case? On Mon, Dec 1, 2014 at 3:06 PM, Tab Atkins Jr. <jackalmage@gmail.com> wrote: > On Fri, Nov 28, 2014 at 5:35 PM, Philip Walton <philip@philipwalton.com> > wrote: > > I've put together a demo (working in Chrome), and in each of the three > > examples, the text is red. It seems to me like it shouldn't be red, but > > perhaps I'm not clear on the specificity of the host and host-context > > selectors. > > > > http://jsbin.com/tacidonere/1/edit?html,output > > > > The gist of the demo is that given the following rule in the <head> > > > > * { > > color: red; > > } > > > > And the following rule in a shadow root: > > > > :host { > > color: blue; > > } > > > > It's surprising to me that the universal selector "wins". I was under the > > impression that pseudo-classes carried the same specificity as regular > > classes. Is there an exception inside shadow DOM? > > > > FWIW, I can see this being a pretty big problem since a lot of CSS resets > > use the universal selector, and in such cases the various :host selectors > > will be essentially useless or have to resort to !important. > > The specificity is standard for pseudo-classes, but this case is > actually covered by something higher in the cascade. Per > <http://dev.w3.org/csswg/css-scoping/#cascading>, rules from outside a > shadow tree always win against rules within a shadow tree, when they > attempt to style the same element. > > The fact that styling the host element doesn't mix well with reset > stylesheets is just yet another entry in the long list of reasons why > reset stylesheets are the devil. ^_^ > > ~TJ >
Received on Monday, 1 December 2014 23:35:29 UTC