[XeTeX] xunicode.sty -- pinyin and TIPA shortcuts

[Robert Spence]

Something that had been literally staring me in the face all the time...

> I'm pretty sure the Chinese government prefers the upright italic  
> shape for lowercase a and g in pinyin, but I'm not sure how strict  
> this is.

... I've only just now noticed that when typing inside a \textipa  
environment as defined by xunicode.sty and using the Gentium font the  
lowercase g I'm getting is the _wrong_ one, i.e. the one that looks  
like a pair of spectacles rotated ninety degrees, not the one the  
teacher writes on the blackboard in primary school.  For phonetics it  
really has to be the blackboard-type g, because the amount of  
topological variation involved in not only changing the shape of a  
curve but actually closing it on itself is systematically meaningful  
within IPA, i.e. whenever you put a little extra hook or turn  
somewhere on a glyph, it turns it into a different glyph standing for  
a different sound.

Tricky...  In this case XeTeX needs to know what font you're using  
before it can decide which unicode point the g in your input text  
should be addressing: plain old 0067 if you're using a sans serif  
font or an italic roman font, otherwise whatever unicode point (1D4D  
maybe?) corresponds to the special lowercase g phonetic letter  
(provided the font supports it); otherwise XeTeX has to fall back on  
using the wrong glyph, but in that case should perhaps issue a  
console warning and record it in the log file.

Hmmm.. from memory Fukui Rei had a command \textg and it would give  
you _either_ the one shape _or_ the other depending on which input  
mode you were in at the time: in what could be called the "longhand"  
input mode you typed \textg for the special phonetic glyph, but in  
the "express" input mode typing just g would give you the special  
phonetic glyph and if you needed the normal (serif-) roman font glyph  
you typed \textg .

At line 752 of xunicode.sty I see that \textg is assigned to unicode  
0067, and that's the only occurrence of \textg in the file.  So that  
would mean... perhaps... having to write some kind of patch that  
would do, the XeTeX unicode way, what Rei's \textg does?

Does anyone have an out-of-the-box solution for this?
:-)

And does this (admittedly almost maximally minor) problem relate to  
the issue of fallback fonts mentioned in Malte Rosenau's XeTeX list  
posting of 06/04/2006 ?

Curious to find out more about all this...

-- Rob Spence

P.S.  I just saw Will Robertson's latest posting in which he suggests  
a "xetipa" package.. ("Curiouser and curiouser!")