Add mandoc-manpage for expr(1)

and mark it as finished in README.
Upon further checking, expr(1) turns out to be implicitly UTF-8 compliant.

README

@@ -29,7 +29,7 @@ The following tools are implemented ('*' == finished, '#' == UTF-8 support,
 =* echo            yes                             none
 =* env             yes                             none
 #* expand          yes                             none
-   expr            yes                             none
+=* expr            yes                             none
 =* false           yes                             none
    fold            yes                             none
    grep            yes                             none

expr.1

@@ -1,12 +1,5 @@
-.\"	$OpenBSD: src/bin/expr/expr.1,v 1.22 2014/02/23 18:13:27 schwarze Exp $
+.Dd January 25, 2015
-.\"	$NetBSD: expr.1,v 1.9 1995/04/28 23:27:13 jtc Exp $
+.Dt EXPR 1 sbase\-VERSION
-.\"
-.\" Written by J.T. Conklin <jtc@netbsd.org>.
-.\" Public domain.
-.\"
-.Dd $Mdocdate: September 3 2010 $
-.Dt EXPR 1
-.Os
 .Sh NAME
 .Nm expr
 .Nd evaluate expression
@@ -14,143 +7,95 @@
 .Nm expr
 .Ar expression
 .Sh DESCRIPTION
-The
 .Nm
-utility evaluates
+evaluates
 .Ar expression
-and writes the result on standard output.
+and writes the result to stdout.
-All operators are separate arguments to the
-.Nm
-utility.
-Characters special to the command interpreter must be escaped.
 .Pp
-Operators are listed below in order of increasing precedence.
+There are two elemental expressions,
-Operators with equal precedence are grouped within { } symbols.
+.Sy integer
-.Bl -tag -width indent
+and
-.It Ar expr1 | expr2
+.Sy string.
-Returns the evaluation of
+Let
-.Ar expr1
+.Sy expr
-if it is neither an empty string nor zero;
+be a non-elemental expression and
-otherwise, returns the evaluation of
+.Sy expr1 ,
-.Ar expr2 .
+.Sy expr2
-.It Ar expr1 Li & Ar expr2
+arbitrary expressions. Then
-Returns the evaluation of
+.Sy expr
-.Ar expr1
+has the recursive form
-if neither expression evaluates to an empty string or zero;
+.Sy expr = [(] expr1 operand expr2 [)].
-otherwise, returns zero.
+.Pp
-.It Ar expr1 Li "{=, >, >=, <, <=, !=}" Ar expr2
+With
-Returns the results of integer comparison if both arguments are integers;
+.Sy operand
-otherwise, returns the results of string comparison using the locale-specific
+being in oder of increasing precedence:
-collation sequence.
+.Bl -tag -width Ds
-The result of each comparison is 1 if the specified relation is true,
+.It |
-or 0 if the relation is false.
+Evaluate to
-.It Ar expr1 Li "{+, -}" Ar expr2
+.Sy expr1
-Returns the results of addition or subtraction of integer-valued arguments.
+if it is neither an empty string nor 0; otherwise evaluate to
-.It Ar expr1 Li "{*, /, %}" Ar expr2
+.Sy expr2 .
-Returns the results of multiplication, integer division, or remainder of
+.It &
-integer-valued arguments.
+Evaluate to
-.It Ar expr1 Li : Ar expr2
+.Sy expr1
-The
+if
-.Ql \&:
+.Sy expr1
-operator matches
+and
-.Ar expr1
+.Sy expr2
+are neither empty strings nor 0; otherwise evaluate to 0.
+.It = > >= < <= !=
+If
+.Sy expr1
+and
+.Sy expr2
+are integers, evaluate to 1 if the relation is true and 0 if it is false.
+If
+.Sy expr1
+and
+.Sy expr2
+are strings, apply the relation to the return value of
+.Xr strcmp 3 .
+.It + -
+If
+.Sy expr1
+and
+.Sy expr2
+are integers, evaluate to their sum or subtraction.
+.It * / %
+If
+.Sy expr1
+and
+.Sy expr2
+are integers, evaluate to their mulitplication, division or remainder.
+.It :
+Evaluate to the number of characters matched in
+.Sy expr1
 against
-.Ar expr2 ,
+.Sy expr2 . expr2
-which must be a basic regular expression.
+is anchored with an implicit '^'.
-The regular expression is anchored
-to the beginning of the string with an implicit
-.Ql ^ .
 .Pp
-If the match succeeds and the pattern contains at least one regular
+You can't directly match the empty string, since zero matched characters
-expression subexpression
+resolve equally to a failed match. To work around this limitation, use "expr X'' : 'X$' instead of "expr '' : '$'"
-.Dq "\e(...\e)" ,
-the string corresponding to
-.Dq "\e1"
-is returned;
-otherwise, the matching operator returns the number of characters matched.
-If the match fails and the pattern contains a regular expression subexpression
-the null string is returned;
-otherwise, returns 0.
-.Pp
-Note: the empty string cannot be matched using
-.Bd -literal -offset indent
-expr '' : '$'
-.Ed
-.Pp
-This is because the returned number of matched characters
-.Pq zero
-is indistinguishable from a failed match, so
-.Nm
-returns failure
-.Pq 0 .
-To match the empty string, use a structure such as:
-.Bd -literal -offset indent
-expr X'' : 'X$'
-.Ed
 .El
-.Pp
-Parentheses are used for grouping in the usual manner.
 .Sh EXIT STATUS
-The
+.Bl -tag -width Ds
-.Nm
-utility exits with one of the following values:
-.Pp
-.Bl -tag -width Ds -offset indent -compact
 .It 0
-The expression is neither an empty string nor 0.
+.Ar expression
+is neither an empty string nor 0
 .It 1
-The expression is an empty string or 0.
+.Ar expression
+is an empty string or 0
 .It 2
-The expression is invalid.
+.Ar expression
-.It \*(Gt2
+is invalid
-An error occurred (such as memory allocation failure).
+.It > 2
+An error occured
 .El
-.Sh EXAMPLES
-Add 1 to the variable
-.Va a :
-.Bd -literal -offset indent
-$ a=`expr $a + 1`
-.Ed
-.Pp
-Return the filename portion of a pathname stored
-in variable
-.Va a .
-The
-.Ql //
-characters act to eliminate ambiguity with the division operator:
-.Bd -literal -offset indent
-$ expr "//$a" \&: '.*/\e(.*\e)'
-.Ed
-.Pp
-Return the number of characters in variable
-.Va a :
-.Bd -literal -offset indent
-$ expr $a \&: '.*'
-.Ed
 .Sh SEE ALSO
-.Xr test 1 ,
+.Xr test 1
-.Xr re_format 7
 .Sh STANDARDS
 The
 .Nm
 utility is compliant with the
 .St -p1003.1-2008
 specification.
-.Sh HISTORY
-The
-.Nm
-utility first appeared in the Programmer's Workbench (PWB/UNIX)
-and has supported regular expressions since
-.At v7 .
-It was rewritten from scratch for
-.Bx 386 0.1
-and again for
-.Nx 1.1 .
-.Sh AUTHORS
-.An -nosplit
-The first free version was written by
-.An Pace Willisson
-in 1992.
-This version was written by
-.An John T. Conklin
-in 1994.

expr.c

@@ -182,11 +182,11 @@ lex(char *p)
 	}
 
 	/* one-char operand */
-	if (*p && !*(p+1) && strchr(ops, *p))
+	if (*p && !*(p + 1) && strchr(ops, *p))
 		return *p;
 
 	/* two-char operand */
-	if (*p && *(p+1) == '=' && !*(p+2)) {
+	if (*p && *(p + 1) == '=' && !*(p + 2)) {
 		switch (*p) {
 		case '>':
 			return GE;