Description of problem: running man(1) in gnome-terminal mangles the manual
pages changing ASCII ' to a non-ASCII closing quote character.  But this is not
correct, because often that character is used in code samples and this breaks them.

Even outside code samples, the ' character is intended to be a neutral vertical
mark, leaning neither left nor right (see
<http://www.cl.cam.ac.uk/~mgk25/ucs/quotes.html>), so it is not correct to
convert it to a closing quote.

Version-Release number of selected component (if applicable):

% man --version
man, version 1.6e
% rpm -q man
man-1.6e-3.fc7

(yes, fc7 is the version included with Fedora 8, see
<http://download.fedora.redhat.com/pub/fedora/linux/releases/8/Fedora/x86_64/os/Packages/>)

How reproducible:

Install the package perl-File-Slurp to provide an example man page.  Log in to a
GNOME session and start gnome-terminal with Applications -> System Tools ->
Terminal.  Set LANG=en_GB.UTF-8 or en_US.UTF-8 and have the LC environment
variables unset.  Then say 'man File::Slurp'.
  
Actual results: the manual page appears with code example

my $text = read_file( XfilenameX ) ;

where X is a non-ASCII closing quote character (according to od -t u1, it is the
UTF-8 byte sequence 226 128 153).

Expected results: the example code should have the ASCII ' character.


Additional info: running man from the Linux console seems to work fine, even
though the value of LANG is the same.