[XeTeX] Another tipa/xunicode puzzle: interaction with array.sty?

[Alan Munn]

On Aug 30, 2010, at 12:12 PM, Alan Munn wrote:

> Hi again, I've encountered another puzzle.
>
> % !TEX TS-program = XeLaTeX
> \listfiles
> \documentclass{article}
> \usepackage{xltxtra}
> \usepackage{array}
> \newfontfamily{\ipafont}{Doulos SIL}
> \def\useTIPAfont{\ipafont}
> \newenvironment{ipa}{%
> \let\stone\TIPAstonebar
> \let\tone\TIPAtonebar
> \setTIPAcatcodes\activatetipa
> \csname useTIPAfont\endcsname
> }{}
>
> \newcolumntype{I}{>{\begin{ipa}[}l<{]\end{ipa}}}
>
> \begin{document}
> \begin{ipa}[PoP]\end{ipa}
>
> \begin{tabular}{I}
> PoP
> \end{tabular}
>
> \end{document}
>
> For some reason, the first character of the the IPA environment  
> (whether with a newcolumntype or explicitly encoded into the column)  
> specification fails to do the correct TIPA mapping.  Enclosing the  
> character in braces solves the problem, but this doesn't seem ideal.
>
> Any suggestions?

And I'll add another problem while I'm at it:

I have lots of tables of the sort described above.  I've defined \NL  
to be equivalent to \tabularnewline, but this no longer works with  
these tables: (copy into the document above).

\newcommand\NL{\tabularnewline}
\begin{tabular}{lI}
foo & {B}AR \tabularnewline % fails if \NL instead of \tabularnewline
foo & {B}AR
\end{tabular}



-- 
Alan Munn
amunn at gmx.com