579886 - Get rid of indexed DOM access in the HTML5 parser

Status: RESOLVED WORKSFORME

(Core :: DOM: HTML Parser, defect, P5)

Description

[Henri Sivonen (:hsivonen)]

Since using neighbor pointer-based access to nsINode/nsIContent is now preferred, the HTML5 parser should no longer use the old index-based access.

bz, it seems to me that the notification APIs in nsContentUtils still want to see indeces. Which indexed APIs should I avoid and which APIs should I use instead?

Comment 1

[Boris Zbarsky [:bzbarsky]]

To do this changeover incrementally, I think we need to first change all the notification API consumers to not use indices, then change the api to not need indices.

As an intermediate stage, most of that API right now takes nodes _and_ indices (though .  If you're not calling ContentRemoved, then all of it does.  So on your end you could pass both index and node for now...

Realistically, I suspect that we probably just need to fix all the mutation observer consumers before we can get rid of index use in the parser due to the notification api issue.

Comment 2

[Jens Stutte [:jstutte]]

Bulk-downgrade of unassigned, >=5 years untouched DOM/Storage bugs' priority.

If you have reason to believe this is wrong (especially for the severity), please write a comment and ni :jstutte.

Comment 3

[Henri Sivonen (:hsivonen)]

AFAICT, we've fixed this along the way.