Thanks - I will try to get at it tonight!
The workaround to the "reading huge files" problem seems like a good
one. Now that you mention it, I did notice that before as well and
the description also explains a number of peculiar behaviors that I
have been seeing with the input framework.
cheers,
scott