> Hans Hagen <mailto:pragma@wxs.nl>
> 9. August 2015 19:02
>
>
> hm, hard to catch (esp if we want to do that everywhere it will be a 
> slow downer) ... i'll do it for cells but so far other mechanisms were 
> less sensitive for this (we'd need an extension to luatex for handling 
> such cases ... maybe i'll look into that later)
Shouldn’t be something like this enough (even though a better solution 
to prevent expension here would be nice):

\unexpanded\def\tabl_x_start_cell_delayed_one[#tag]%
% {\ifcsname\namedxtablehash{#tag}\s!parent\endcsname
- {\ifcsname\??xtablecheck#tag\endcsname
+ {\ifcsname\??xtablecheck\detokenize\expandafter{#tag}\endcsname
      \expandafter\tabl_x_start_cell_delayed_two
    \else
      \expandafter\tabl_x_start_cell_yes
    \fi[#tag]}

Wolfgang