Re: [dom] Need to describe the interaction of adoptNode with prototype chains

On 12/24/12 5:42 PM, David Bruant wrote:
> Chrome has a relevant market share, so testing in Chrome is never
> pointless.

I meant in terms of figuring out what the spec says t odo.

> If a web browser with relevant proves that their non-spec compliant
> behavior doesn't break the web, then the spec becomes questionable.

Sure.  I'm just saying that in my experience this particular behavior on 
Chrome's (more precisely WebKit's) part breaks things for a number of users.

Of course so does the complete lack of history traversal handling across 
document.open in WebKit.

>> It's pretty crucial last I checked.  There are various things broken
>> in Chrome because of its behavior here.
> Do you have examples of webpages in the wild?
> If they are broken, why don't they bother fixing it?

Because none of their users use Chrome, because it doesn't work on those 
pages.  Classic chicken and egg.  Plus it's actually impossible to "fix" 
short of not using document.open at all, which means a complete rewrite.

I'd have to query Bugzilla for examples, but you can do that as well as 
I can.

> Why is it web-compatible enough for Chrome?

Are you sure it is?

> Do they have a different definition?

Yes.

> if so, how does it differ?

It differs because they want to be compatible with the sites their users 
already use while we want to be compatible with the sites our users 
already use, if nothing else.

This is why we were just fine with not having a global scope polluter in 
standards mode, but apparently WebKit was not, for example.

-Boris

Received on Tuesday, 25 December 2012 03:28:47 UTC