anybody mind if I commit my changes?

[David Hyatt]

On Friday, August 8, 2003, at 12:34PM, Dirk Mueller wrote:

> c) TextSlave -> TextRun rename. While I again miss the point of 
> renaming
>    it to something that is remarkable similiar to BidiRun (which is 
> however
>    a different thing), I'm wondering why it has to inherit InlineBox 
> (which
>    is quite a huge class). There are tons of TextSlaves created, and 
> this
>    is a nontrivial overhead IMHO. Also, for consistency I would suggest
>    a name line InlineTextBox instead (similiar to InlineFlowBox).
>

You're basically seeing two independent Safari patches at work here.  
TextSlaves were renamed to TextRuns before the full inline box model 
landed.  I like the name InlineTextBox as well.

> d) Why is it named InlineBox anyway? CSS specs talk about Line boxes, 
> not
>    Inline Boxes..
>

Section 9.2.2.  The term "inline box" is used repeatedly to refer to 
the boxes for inline elements like <em>s, and also to refer to 
"anonymous inline boxes" created for text runs.

The spec basically uses both, so I don't feel strongly either way.  Do 
you like LineBox better?

> l) It amazes me that RenderFlow::constructLine takes a start and end
>    BidiIterator, but never actually makes use of it. Instead it does
>    a full iteration over the complete BidiRun.
>
>

The function does actually use end, but not start.  I had it on my list 
of things to clean up to remove the start argument.  The function only 
iterates over the runs constructed for a given line.

dave