[texhax] Clash between \emph and \xspace [corrected, improved]

[Uwe Lueck]

Uwe Lueck <uwe.lueck at web.de> wrote 15.11.2006 14:47:46:
> What clash? (Try additional tests attached below.) 

Sorry, I am unable to reproduce my former (\emph{\lin}) miracle. 
Probably it resulted from code of which I was not aware of, 
different from the code code I posted. 

To make up for this, I add to Phil Ratcliff's explanation and solution 
(to do without {} at each occurrence): 

> \documentclass{article} 
> \usepackage{xspace} 
> \newcommand{\lin}{Lindel\"of\xspace} 
> \begin{document}
> 
> (property we call \emph{amply \lin})
> 
> (property we call {\it amply \lin})
> 
> (property we call {\it\lin\/})
> 
> \end{document}

The last case is OK since \/ is in the exeptions list. 

At 23:27 04.11.06, Philip G. Ratcliffe wrote:
> Well this seems ok:
> 
> (property we call \emph{amply \lin{}})
> 
> The problem with \xspace is that it looks at the following character, which
> is a problem if it appears at the end of an argument to some macro.  Sorry,
> I've nothing more to offer.

Indeed, the LaTeX font commands like \emph{...} and \textit{...} 
put the token \check at icr and some other things behind their argument, 
so \xspace cannot see the closing bracket. 

With the recent version of xspace, you can do 

    \makeatletter
    \xspaceaddexeptions{\check at icr}
    \makeatother

so the final \xpace won't add a \space at the end of the text font 
argument. 

I wondered why current xspace package doesn't give \check at icr 
as an exeption already. 

At 00:06 05.11.06, Michael Barr wrote:
> The thing is that \emph also looks ahead 
> and this is what I meant by saying it seems to be a clash of uses of 
> \futurelet.

This is not the case. The spurious space before the bracket 
comes from the \xspace at the end of the argument, due to \check at icr. 
\emph doesn't look ahead. 

Hope this helps, 

  Uwe Lueck.