On Sat, Mar 04, 2006 at 11:26:28PM +0100, Heiko Oberdiek wrote:

> That means, the command tokens in LICR are limited to
> commands defined by the nfss2 \Declare... commands?

What is "\ "?
In the old 8bit encoding files, this is often used for
lonely accents, examples (latin2.def):
  \DeclareInputText{178}{\k\ }
  \DeclareInputText{184}{\c\ }
But the use is not constent, sometimes {} is used instead
(again latin2.def):

One disadvantage of the "\ " method over "{}" can be seen
    [\k\ ] = [\k{}]

In the error recovery:
  ! LaTeX Error: Command \k unavailable in encoding OT1.
the argument of \k is not eaten, thus there is a space
between the square brackets in the first place.

utf8enc.dfu handles this much better by using \textascii...:
  % vs. \"{}
  % vs. \@tabacckludge={}
  % vs. \@tabacckludge'{}
  % vs. \^{}
  % vs. \v{}
  % vs. \~{}
  % vs. \u{}
  % vs. \H{}

  \DeclareUnicodeCharacter{00B8}{\c\ }
Why is not \textasciicedilla used?

Yours sincerely
  Heiko <[log in to unmask]>