WebCore 146

[Zack Rusin]

On Tuesday 06 July 2004 18:38, Dave Hyatt wrote:
> Incremental patches for large features like HTML editing would have
> been fairly pointless.  I'd argue that even merging it now in its
> current state would be of little benefit.

That's good to know. Thanks. 
I'm forwarding your email to the Quanta guys since besides Leo they were 
the ones working on khtml editting the most on our side. 

Quanta guys, could you take a look at the html editting implementation 
in the new webcore at  
http://darwinsource.opendarwin.org/tarballs/other/WebCore-146.1.tar.gz 
and let me know what you think? 

> Breaks down primarily into:
> (1) Performance (style system improvements and atomization of
> attribute values).  We're about 30% faster.

Yeah, that's the thing we'll have to merge. I think Dirk started looking 
into it and had problems figuring out where does the speed improvement 
come from. Do you recall what exactly caused such a huge speedup?

> (2) HTML Editing (in the new subdir plus changes to line layout to
> make it smarter about only repainting changed lines)

That will wait. 

> (3) Dashboard Support (new widgets input type=search, range and
> canvas and composite attr on image)

Shame that you were not allowed to be spreading the news about this one. 
I had canvas (well, i called it painter) implementation for quite a 
while now because I was playing with client side rendering and you 
could have gotten it for free.

Those will also wait because we're already in a freeze for 3.3 and these 
are hardly essential. They will be merged though.

> (4) Bug Fixes

Obviously would love to have it, but we'll need to figure out what they 
were.

> We could not show patches for (2) or (3).  I did send out the style
> system improvements in (1) to khtml-devel.  I agree we could be
> better about (4), but often these overlapped with work in (2).

The big question is: besides the style optimization patch is there 
anything you think should be merged sooner rather than later? 
That would be really helpful to know. Even such a small thing as letting 
us know that merging html editting would be of little benefit is 
immensely helpful simply because it saves us time of having to find out 
about it the hard way :)

Zack

-- 
Failure is not an option. (It comes bundled with Windows.)