Waste a few minutes occasionally on cleanup, and maybe we can make this
thing look ok in a decade or two..
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This code still assumes a latin1 kind of "one byte, one character" setup.
UTF-8 input/output (even if the data is encoded in latin-1) is a separate
issue.
make sure to include <errno.h>, and allow for the fact that newer
gcc's don't allow function declarations in function scope (don't
ask me why, but there you have it..)
This is a slightly updated version of uemacs-PK (PK is Pekka
Kutvonen) which was used at Helsinki University a long time
ago. My fingers cannot be retrained.
