Jean-Marc Lasgouttes wrote:
Richard heck <rgh...@bobjweil.com> writes:
I could be wrong, but what you report certainly seems like a bug. I
think maybe around this part of the function:
// The last added char caused a mismatch, because
// we didn't exhaust the chars in cmd and didn't
// exceed the maximum size of the current unicmd
if (k < cmdend && cur_size > tmp.size())
tmp.resize(tmp.size() - 1);
we ought to check whether that last added char was alpha and, if so,
figure it's a command of some other sort.
Yes, I think it is a bug. But actually I do not understand why this
macro is written as it is here. The code should read a token and look
whether it is part of a unicode symbol, or something like that.
The logic of the thing suggests that the author didn't think about the
case where it was a longer command, just about the case of, say: \O23,
rather than: \OtherCmd. Anyway, it does seem to me that this thing could
be simplified, but I guess Enrico will know.
rh
