123385 - Leaf Content Iterator

Status: RESOLVED INACTIVE

(Core :: DOM: Core & HTML, defect)

Description

[Joe Francis]

The Find code would be simplified by having a version of the content iterator
that visited only leaf nodes (nodes that have no children).   Other folks have
occassional asked for this as well so I think it is worth having.  You could
simply use the regualar iterator and ignore non-leaves, but I beleive we can
leverage the performance work in the iterator to get a speed gain in a pure leaf
iterator.

Not sure what component my content iteraotr stuff falls under.  DOM Traversal-Range?

Comment 1

[Joe Francis]

putting on 1.1 list.  We can move it up if Akkana concludes we need it for Find
performance.

Comment 2

[Akkana Peck]

Actually: just yesterday I realized that I do need to do my skipped content
check after all (which had been commented out due to lack of a bidi iterator). 
I've put in some temporary code that climbs up the parent chain for each node to
see if a parent is in the skipped list, but I don't know how this affects
performance and if we're worried about performance, the leaf iterator (or even
just the regular iterator but being able to use both pre and post order) would
help ...

If it helps, I can get some numbers on how climbing up the parent chain affects
performance.  I'm not sure how expensive GetParent is (seems like it should be
fairly cheap, but I thought that about GetNextSibling and I was dead wrong).

Comment 3

[Joe Francis]

GetParent is cheap.  And you will still have to make GetParent calls each time 
even with a leaf iterator, if you want to check for block crossings.

Comment 4

[Akkana Peck]

Two things I need the iterator to do:

1. Visit parent nodes both on the way in and on the way out.  (This would also
help in detecting crossing block borders during copy/serialization).

2. Offer a PositionAt or equivalent which takes a node, not a content.  Not all
nodes can QI to nsIContent, but the selection can be placed at such nodes so I
need to be able to iterate starting there.

I'm not sure if "leaf content iterator" is really the best name for what I want,
since it seems to contradict both of the goals. :-)

Comment 5

[Joe Francis]

The days of having a half dozen milestones out in front of us to divide bugs 
between seem to be gone, though I dont know why.  Lumping everything together as 
far out as I can.  I'll pull back things that I am working on as I go.

Comment 6

[Joe Francis]

[ushing these out as far as bugzilla will let me.  I'll pull them back as I work
on them.