[XeTeX] Nested \textit in an italic fontinstance

Jonathan Kew jonathan_kew at sil.org
Wed Sep 13 13:34:58 CEST 2006

On 13 Sep 2006, at 12:07 pm, Will Robertson wrote:

> Bruno wrote:
>> With respect to the LaTeX way of dealing with fonts, I think then
>> that we have a problem for fontspec: \emph corresponds to the
>> behaviour of /I, but not \textit. When a face is already italic,
>> fontspec should define \textit to do nothing, and \emph to do what /I
>> does. But, of course, how could fontspec know that a face is already
>> italic?
> Well, it's possible to look at the slant of the font, but that's  
> not always reliable.
> By the way, \emph doesn't correspond to /I, because "/I" will never  
> be iteratively applied. It's only in the setup that this slightly  
> non-intuitive behaviour occurs.

Yes... and so I don't think there is really a problem here for  

The only case you'll get arguably-incorrect results from \textit is  
(I think) if you ask fontspec to set up a family, but give it the  
italic face as the starting point. This is not really analogous to  
anything a user can normally do in a LaTeX document; it's analogous  
to creating a .fd file where the default "roman" shape is actually  
the italic. Is there any convention for doing this, and for what the  
"italic" shape would be in such a case?

(Things like Zapf Chancery aren't comparable, as it has no "non- 
italic" face!)

> Jonathan wrote:
>>> Incidentally, you might expect /B to act similarly, toggling bold on
>>> or off, but it doesn't. Instead, it always tries to find a bolder
>>> version of the font that was given.
> Wah?!? I could see that being useful in the future.
> Hmm, if that's the case I've never seen it -- /B seems to me to  
> always snap to the most bold font available. But maybe my testing  
> hasn't been wide enough.

Hmm.... you're right. I thought that worked, at least with some  
families, but in a few quick tests I can't seem to find an example  
that actually does! Guess I'll take another look at it.


More information about the XeTeX mailing list