===================================================================
RCS file: /cvs/mandoc/mandoc_char.7,v
retrieving revision 1.40
retrieving revision 1.47
diff -u -p -r1.40 -r1.47
--- mandoc/mandoc_char.7	2010/10/29 00:05:53	1.40
+++ mandoc/mandoc_char.7	2011/08/18 08:58:44	1.47
@@ -1,4 +1,4 @@
-.\"	$Id: mandoc_char.7,v 1.40 2010/10/29 00:05:53 schwarze Exp $
+.\"	$Id: mandoc_char.7,v 1.47 2011/08/18 08:58:44 kristaps Exp $
 .\"
 .\" Copyright (c) 2009 Kristaps Dzonsons <kristaps@bsd.lv>
 .\"
@@ -14,7 +14,7 @@
 .\" ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
 .\" OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
 .\"
-.Dd $Mdocdate: October 29 2010 $
+.Dd $Mdocdate: August 18 2011 $
 .Dt MANDOC_CHAR 7
 .Os
 .Sh NAME
@@ -288,6 +288,9 @@ Mathematical:
 .It \e(Re    Ta \(Re        Ta real
 .It \e(pd    Ta \(pd        Ta partial differential
 .It \e(-h    Ta \(-h        Ta Planck constant over 2\(*p
+.It \e[12]   Ta \[12]       Ta one-half
+.It \e[14]   Ta \[14]       Ta one-fourth
+.It \e[34]   Ta \[34]       Ta three-fourths
 .El
 .Pp
 Ligatures:
@@ -481,8 +484,13 @@ Greek letters:
 .It \e(ts    Ta \(ts        Ta sigma terminal
 .El
 .Sh PREDEFINED STRINGS
-These are not recommended for use, as they differ across
-implementations:
+Predefined strings are inherited from the macro packages of historical
+troff implementations.
+They are
+.Em not recommended
+for use, as they differ across implementations.
+Manuals using these predefined strings are almost certainly not
+portable.
 .Pp
 .Bl -column -compact -offset indent "Input" "Rendered" "Description"
 .It Em Input Ta Em Rendered Ta Em Description
@@ -512,39 +520,82 @@ implementations:
 .It \e*(>=   Ta \*(>=       Ta greater-than-equal
 .It \e*(aa   Ta \*(aa       Ta acute
 .It \e*(ga   Ta \*(ga       Ta grave
+.It \e*(Px   Ta \*(Px       Ta POSIX standard name
+.It \e*(Ai   Ta \*(Ai       Ta ANSI standard name
 .El
-.Sh COMPATIBILITY
-This section documents compatibility of
-.Nm
-with older or existing versions of groff.
+.Sh UNICODE CHARACTERS
+The escape sequence
 .Pp
-The following render differently in
-.Fl T Ns Ar ascii
-output mode:
-.Bd -ragged -offset indent
-\e(ss, \e(nm, \e(nb, \e(nc, \e(ib, \e(ip, \e(pp, \e[sum], \e[product],
-\e[coproduct], \e(gr, \e(-h, \e(a.
-.Ed
+.Dl \e[uXXXX]
 .Pp
-The following render differently in
-.Fl T Ns Ar html
-output mode:
-.Bd -ragged -offset indent
-\e(~=, \e(nb, \e(nc
-.Ed
+is interpreted as a Unicode codepoint.
+The codepoint must be in the range above U+0080 and less than U+10FFFF.
+For compatibility, points must be zero-padded to four characters; if
+greater than four characters, no zero padding is allowed.
+Unicode surrogates are not allowed.
+.\" .Pp
+.\" Unicode glyphs attenuate to the
+.\" .Sq \&?
+.\" character if invalid or not rendered by current output media.
+.Sh NUMBERED CHARACTERS
+For backward compatibility with existing manuals,
+.Xr mandoc 1
+also supports the
 .Pp
-Finally, the following have been omitted by being poorly documented or
-having no known representation:
-.Bd -ragged -offset indent
-\e[radicalex], \e[sqrtex], \e(ru
-.Ed
+.Dl \eN\(aq Ns Ar number Ns \(aq
+.Pp
+escape sequence, inserting the character
+.Ar number
+from the current character set into the output.
+Of course, this is inherently non-portable and is already marked
+as deprecated in the Heirloom roff manual.
+For example, do not use \eN'34', use \e(dq, or even the plain
+.Sq \(dq
+character where possible.
+.Sh COMPATIBILITY
+This section documents compatibility between mandoc and other other
+troff implementations, at this time limited to GNU troff
+.Pq Qq groff .
+.Pp
+.Bl -dash -compact
+.It
+The \eN\(aq\(aq escape sequence is limited to printable characters; in
+groff, it accepts arbitrary character numbers.
+.It
+In
+.Fl T Ns Cm ascii ,
+the
+\e(ss, \e(nm, \e(nb, \e(nc, \e(ib, \e(ip, \e(pp, \e[sum], \e[product],
+\e[coproduct], \e(gr, \e(\-h, and \e(a. special characters render
+differently between mandoc and groff.
+.It
+In
+.Fl T Ns Cm html
+and
+.Fl T Ns Cm xhtml ,
+the \e(~=, \e(nb, and \e(nc special characters render differently
+between mandoc and groff.
+.It
+The
+.Fl T Ns Cm ps
+and
+.Fl T Ns Cm pdf
+modes format like
+.Fl T Ns Cm ascii
+instead of rendering glyphs as in groff.
+.It
+The \e[radicalex], \e[sqrtex], and \e(ru special characters have been omitted
+from mandoc either because they are poorly documented or they have no
+known representation.
+.El
 .Sh SEE ALSO
 .Xr mandoc 1
 .Sh AUTHORS
 The
 .Nm
 manual page was written by
-.An Kristaps Dzonsons Aq kristaps@bsd.lv .
+.An Kristaps Dzonsons ,
+.Mt kristaps@bsd.lv .
 .Sh CAVEATS
 The
 .Sq \e*(Ba