Bug 90050 – Rename Element::rareData() to Element::elementRareData(), and Element::ensureRareData() to Element::ensureElementRareData()

• Component: DOM
• Status: RESOLVED FIXED

[Kentaro Hara]

Element::rareData()/Element::ensureRareData() and Node::rareData()/Node::ensureRareData() are confusing. They are not virtual methods. For clarification, we can rename Element::rareData() to Element::elementRareData(), and Element::ensureRareData() to Element::ensureElementRareData().

c.f. SVGRareData uses SVGElement::rareSVGData() and SVGElement::ensureRareSVGData(). (We might want to rename them to SVGElement::svgRareData() and SVGElement::ensureSVGRareData() though.)

[Kentaro Hara]

Created attachment 149690 [details]
Patch

[Hajime Morrita]

Is that confusing? How could you misuse it?
Another random idea is to unify then to rareData(bool ensure);

[Kentaro Hara]

(In reply to comment #2)
> Is that confusing? How could you misuse it?

Misuse would lead to a compile error (because Node::rareData() returns NodeRareData and Element::rareData() returns ElementRareData), and thus it would not cause any serious issue in consequence. That being said, distinguishing their names would be less confusing. For example, SVGElement uses rareSVGData() to retrieve SVGRareData (I am trying to rename it to svgRareData() though).

Specific context: I've tried to virtualize Node::rareData() to cache NodeRareData* on Document to optimize Dromaeo. Then Element::rareData() prevented me from doing the virtualization. (Whether the optimization is acceptable or not is another topic.)

[Ryosuke Niwa]

(In reply to comment #3)
> Specific context: I've tried to virtualize Node::rareData() to cache NodeRareData* on Document to optimize Dromaeo. Then Element::rareData() prevented me from doing the virtualization. (Whether the optimization is acceptable or not is another topic.)

I don't think we want to virtualize rareData(). That function is already slow. We shouldn't be making it even slower by virtualizing it.

[Ryosuke Niwa]

Comment on attachment 149690 [details]
Patch

Regardless, this sounds like a good idea.

[Kentaro Hara]

morrita-san: If you think the change is OK, I'm happy to commit it.

[WebKit Review Bot]

Comment on attachment 149690 [details]
Patch

Clearing flags on attachment: 149690

Committed r121335: <http://trac.webkit.org/changeset/121335>

[WebKit Review Bot]

All reviewed patches have been landed.  Closing bug.