Translate UTF-8 files into ISO Latin 1 (ISO-8859-1)

This is very easy as ISO-8859-1 is the first codepage of UTF-8.

This warns about ignored characters which are not in the Latin1
codepage.  Ignore stderr if you don't want the warnings.

Note that this was only tested within the Latin1 codepage, so if you
find some values outside Latin1 the UTF-8 reading code might be
broken.

If you want to have non-Latin1 characters in your database, perhaps
better store UTF-8 in it to avoid confusion.

Usage:
	utf8-latin1 [error-string] 2>/dev/null

If [error-string] is given (first argument) this string is printed
instead of the non-Latin1-characters (else the character is skipped).

SECURITY WARNING:

The [error-string] is a format string to printf, where the one and
only printf-arg is a (signed) long which holds the UTF-8 character which
is not in range of Latin1 (thus <0 or >0xff, where <0 denotes an
illegal UTF-8 encoding was found, see source code).  So it is
definitively a bad idea to install this program SUID or give users the
option to pass any first argument.

Example error-string which prints the hex value of the non-latin
character:

./utf8-latin "[[non-latin code %04lx]]"

Example error-string which makes the program react unpredictable
(perhaps even segv, buffer overrun or delete the whole universe):

./utf8-latin "%04lx %s"		YOU SHOULD NEVER DO THIS!!!

Example error-string which should put out a percent sign:

./utf8-latin "%%"

Example error-string which should put out a percent sign and then the
hex value of the non-latin character:

./utf8-latin "%%%lx"

(You've got it now, hopefully)