[cssom] LinkStyle, etc. (was: Re: [Bug 360746] The right panel has disappeared at andrewdupont.net (fwd))

On Fri, 11 May 2007 20:14:36 +0200, Boris Zbarsky <bzbarsky@MIT.EDU> wrote:
> Anne van Kesteren wrote:
>>>> 1)  Should a change of "rel" from "alternate stylesheet" to
>>>> "stylesheet" have any effect?  If so, what?
>>  You mean whether this should re-run the algorithm that decides which  
>> style sheets to apply?
> No, I mean exactly what I asked.  In Gecko, for a while, that change  
> reloaded the stylesheet (reparsed it, etc).  I've changed things so that  
> we no longer do that, but the question remains what the "right" behavior  
> is, imo.

Fair enough. I think this should be addressed by HTML5 though. The CSSOM  
doesn't define the rel= attribute.

>> I suppose it would be nice if HTMLStyleElement.disabled was changed  
>> LinkStyle.sheet.disabled would change and vice versa. Which should  
>> solve this issue, I think.
> You mean that HTMLStyleElement would have out-of-band disabled state  
> (not represented by any attribute), which would persist even if the  
> sheet linked to changes?
> What Gecko does right now is to just track the disabled state on the  
> sheet; setting HTMLStyleElement.disabled just sets  
> LinkStyle.sheet.disabled.

That should happen per HTML5.

> I could probably live with having the state in both places and keeping  
> them synced...  That way a new sheet load after disabling the previous  
> sheet will also be disabled.  Is that the behavior we want?

I'm not sure. We could either do it that updating LinkStyle.sheet.href  
will potentially give a different result from updating  
HTMLLinkElement.href or we could keep them in sync somehow.

Maybe it would be more consistent if HTML5 just defined how  
HTMLLinkElement and HTMLStyleElement work and that LinkStyle.sheet is  
"separate" from them. As in, what updating LinkStyle.sheet members would  
do would be defined in the CSSOM and would not affect HTMLLinkElement and  
such... This would mean that a style sheet could be disabled while the  
disabled attribute of HTMLLinkElement doesn't reflect that. Do we want  

Anne van Kesteren

Received on Saturday, 12 May 2007 21:03:15 UTC