[Bug 23887] [Shadow] Put only the final destination insertion point to the event path

https://www.w3.org/Bugs/Public/show_bug.cgi?id=23887

--- Comment #167 from Koji Ishii <kojiishi@gmail.com> ---
(In reply to Olli Pettay from comment #165)
> (In reply to Koji Ishii from comment #164)
> > Hayato and I discussed today and came up with a new proposal, in the hope of
> > achieving both-win.
> Don't understand what "both-win" refers to.

I meant to make all impls simple and fast regardless of underlying
architecture, and make all insertion points appear in the event path.

> What is wrong with the proposed change + the fix similar to comment 163?
>
> I very much thought we had agreement here and just had to tweak the proposed
> algorithm a bit.
> 
> > On the other hand, we implemented the proposed new algorithm, but we figured
> > out that it makes our code more complex, and we are still unable to fix 3
> > regressions.
> > 
> What regressions?

The Comment 160 reported failures of 4 test case at <http://jsbin.com/casoyi>.
I saw your comment to work on not-distributed case, and I suppose comment 163
is the one, but I don't think it fixes other cases.

I can implement it to re-test if you believe it fixes all 4 issues, but can't
figure out what the proposal in comment 163 is. Could you clarify:

> ...and CURRENT is distributed to a shadow insertion point:

Did you mean "...and at least one shadow insertion point is the destination
insertion points" or "...and the final destination point is a shadow insertion
point"?

> > Our conclusion is to allow implementation dependency in the manner of not to
> > trouble web developers is the only way to make all of us happy.
>
> I hope you don't mean we would have implementation specific event dispatch?
> That is something Gecko will not do.
> Shadow DOM will not be enabled by default before we have a good enough spec
> for it and the implementation follows it.

Hm, looks like we have some misunderstanding. Can you clarify what I miss?

1. You'd like to change the algorithm in the current spec.
2. The proposed algorithm showed 4 regressions.
3. You kindly tweaked the proposed algorithm for one of them (not really sure
"one" until I understand comment 163 and re-test), but nobody has raised hands
to solve other issues.

>From these understandings, options I can see are:
A. Someone fixes all other issues in the proposed algorithm
B. No insertion points
C. Go with the current spec
D. Stuck

If A is not happening, the new proposal looks still better than B-D, no?

-- 
You are receiving this mail because:
You are the QA Contact for the bug.

Received on Friday, 27 February 2015 03:08:14 UTC