[XeTeX] Font protrusion --- new or old ?

[Jonathan Kew]

On 9 Feb 2009, at 04:57, Will Robertson wrote:

> On 2009-02-09 01:06:16 +1030, Jonathan Kew
> <jonathan at jfkew.plus.com> said:
>
>> Basically, borrow the ideas from Appendix D but use the transition
>> between class 255 (non-character) and whatever classes you assign to
>> the punctuation chars to insert the kerns.
>
> Well! That's quite amazing.
> I never even knew that class 255 was encountered at line start and  
> end.

It's "encountered" much more frequently than that; there's nothing  
special about line start/end. The interchar-toks mechanism considers  
there to be a class 255 pseudo-item wherever there is a boundary  
between a "run" of characters and something else -- glue, kern, math,  
box, etc., etc. So in practice that means at the start and end of each  
space-separated "word", as well as other places where markup and other  
non-character content is found.

>
> (Well, actually I never knew where it was found at all!)
>
> Will class 255 turn up in places where hanging punctuation won't be
> appropriate?

The idea of the kerns from Appendix D is that if a line-break isn't  
chosen here, they'll just cancel each other out; the hanging behavior  
arises from the way discardable items are removed when a break is  
chosen. So it shouldn't matter that this happens throughout the line,  
not just at the edges. After all, that's what Knuth's active  
characters did, too.

JK