Text Selection Speed

Dirk Mueller mueller at kde.org
Fri Jun 14 10:03:55 BST 2002


On Don, 13 Jun 2002, Matt Newell wrote:

> understanding to improve the horribly slow text selection algorithm.  I am 
> starting to get a pretty decent understanding of how it works, but I still am 
> not sure where all of the cpu cycles are going.  I just thought that you guys 
> could give me a few pointers on what functions need optimizing, or how to 
> profile khtml to find out for myself.

There is no index about onscreen position -> rendertree object. 

Partially the text selection code simply scans through the tree for finding 
the starting and end position, which is really slow. That part needs 
rewriting, but up to now nobody volunteered. 


Dirk




More information about the kfm-devel mailing list