Latex accents like \'e are translated to InsetLatexAccent by tex2lyx. This is unfortunate, since some converters produce such latex accents (from swriter, I believe).
Actually, tex2lyx has a built-in hack to handle german umlauts (surprise!), but this hacks assumes that the encoding is latin1 or compatible, which is not very good. Georg, how difficult would it be to transform all such accents to native characters when supported by the current encoding? Could we use the tables in bdb/*.cdef for that? JMarc
