[XeTeX] xltxtra's textsuperscript, was Re: Super/subscript

[Will Robertson]

Hi Bruno,

Many thanks for looking into this.

On 28/05/2007, at 18:29 , Bruno Voisin wrote:

> It would be better that xltxtra first
> tests whether VerticalPosition=Superior is available, and if so
> reimplement the definition of \textsuperscript to use it, and if not
> leave \textsuperscript unchanged;

This is how it's supposed to work, so the fact that it's not for you  
indicates there's been a bug introduced at some stage as either  
fontspec or XeTeX has been updated.

> or, similar to what many package do
> (think for example of all the options of the lucidabr package),
> provide options allowing the user to activate or disactivate
> selectively specific functions of the xltxtra package.

This is definitely required, and on my wish list. I hope to get  
around soon to spending some time on this package again.

> If none of the above is possible, then I think it should be
> \textsuperscript that retains the behaviour of the original
> \textsuperscript, and \textsuperscript* that offers a modified  
> version.

What do you think if (a) it fakes the superscripts if none are  
offered by the font, and (b) can be turned off separately to the  
other things offered by xltxtra -- is it then okay to override the  
default? The only time you wouldn't want the real superscripts are in  
cases like JUnicode, where only a very few superscript characters are  
offered. By default, \textsuperscript would then give bad output, and  
you'd have to activate a package option to fix things up.

By *not* going with this solution, that is, using \textsuperscript*  
for the xltxtra-extended version, you miss out on footnote symbols  
taking advantage of the real superscripts offered by the font without  
changing the definition of the underlying \@mkfootnote command (or  
whatever it's called -- I always forget!).

All the best,
Will