[webkit-dev] HTML5 tokenizer landing soon

[David Hyatt]

On Jun 14, 2010, at 2:44 PM, Geoffrey Garen wrote:

> 
>> The old parser was un-touchable for fear of breaking the world.  
> 
> Once again, I believe in hackability, but we don't need to rewrite history in order to justify it. In the past, we have indeed made changes to the old parser to improve its performance, to good effect.

Yeah and most times we touched it to improve performance, we introduced subtle regressions or tripped over nasty edge cases caused by the horrible state of the code and the poor documentation of that code.  I think Eric's statement is pretty justified actually.  Either you considered the parser un-touchable and just stayed out of it, or you hacked on it, ultimately with good effect, but often with great pain and suffering involved in the regression department just to achieve those improvements.

I really do consider the current code to be "barely hackable," so any new code that follows the HTML5 spec (especially one that has a document.write / pending script model that is easier to understand) is a huge win in my book.

dave
(hyatt at aple.com)

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-dev/attachments/20100614/edc58423/attachment.html>