On Thu, Mar 11, 2010 at 11:25:44PM +0200, Khaled Hosny wrote: > On Thu, Mar 11, 2010 at 09:24:06PM +0100, Heiko Oberdiek wrote: > > On Thu, Mar 11, 2010 at 10:12:25PM +0200, Khaled Hosny wrote: > > > > > What about the approach we used in etex.src; redefining \patterns and > > > \hyphenation to use corresponding lua functions, so we don't need to > > > duplicate hyphenation files and, AFAIU, solves the catcode issue. > > > > \patterns is not read by Lua but by TeX, thus the catcode > > problem remains. > > I don't claim that I understand anything, but do we have some example > for the likes of me to see the problem? * verbatim mode * babel shorthands * other active characters * ... At format generation the catcodes are known, the pattern files are written for that case. Later the catcodes can be anything. Yours sincerely Heiko <[log in to unmask]>