Bug 77925
| Summary: | Introduces Element::youngestShadowRoot() and Element::oldestShadowRoot() | ||
|---|---|---|---|
| Product: | WebKit | Reporter: | Shinya Kawanaka <shinyak> |
| Component: | DOM | Assignee: | Nobody <webkit-unassigned> |
| Status: | RESOLVED INVALID | ||
| Severity: | Normal | CC: | dglazkov, dominicc, morrita |
| Priority: | P2 | ||
| Version: | 528+ (Nightly build) | ||
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
Shinya Kawanaka
Instead of shadowRoot(), we should implement youngestShadowRoot() and oldestShadowRoot().
youngestShadowRoot() will be used in renderer, and oldestShadowRoot() will be used in elements.
| Attachments | ||
|---|---|---|
| Add attachment proposed patch, testcase, etc. |
Dominic Cooney
I think shadow roots should be a doubly linked list. new ShadowRoot() pushes on the front of the list. Tree flattening iterator uses the back pointer to get from the last child of a SR to the next sibling of the <shadow> element that output it.
Do we need oldestShadowRoot? Could callers just walk the list of roots themselves. I feel like this is builtInShadowRoot with a different name.
Shinya Kawanaka
(In reply to comment #1)
> I think shadow roots should be a doubly linked list. new ShadowRoot() pushes on the front of the list. Tree flattening iterator uses the back pointer to get from the last child of a SR to the next sibling of the <shadow> element that output it.
>
> Do we need oldestShadowRoot? Could callers just walk the list of roots themselves. I feel like this is builtInShadowRoot with a different name.
Yeah, it's almost the same as builtinShadowRoot().
I and morrita discussed it. We concluded that it would be better to have it instead of having a reference for now, because it's too error-prone. Of course we can remove oldestShadowRoot later...
Shinya Kawanaka
Since we will introduce ShadowRootList in Bug 78069, I make this INVALID.