Re: Foster parenting characters (detailed review of parsing algorithm)

On Jul 10, 2007, at 21:35, Henri Sivonen wrote:

> Since space characters are unconditionally not foster parented

A possible solution:

Perhaps text node coalescing needs to be handled by making the  
buffering of character tokens and explicit part of the tree  
construction algorithm. This isn't a big deal in practice for  
implementations that promise coalesced text nodes anyway.

Whether the buffer should be flushed upon a non-character token or  
upon tree mutation (i.e. the token was not ignored) would need  
research. Either way, the buffer could be checked for non-space  
character content upon flush.

-- 
Henri Sivonen
hsivonen@iki.fi
http://hsivonen.iki.fi/

Received on Tuesday, 10 July 2007 18:48:01 UTC