--- rxvt-unicode/src/perl/selection	2006/05/20 18:17:38	1.38
+++ rxvt-unicode/src/perl/selection	2011/03/13 17:07:36	1.53
@@ -19,7 +19,6 @@
 
    for (my $idx = 0; defined (my $res = $self->x_resource ("selection.pattern-$idx")); $idx++) {
       $res = $self->locale_decode ($res);
-      utf8::encode $res;
       push @{ $self->{patterns} }, qr/$res/;
    }
 
@@ -34,20 +33,30 @@
 
 # "find interesting things"-patterns
 my @mark_patterns = (
+#   qr{ ([[:word:]]+) }x,
+   qr{ ([^[:space:]]+) }x,
+
    # common types of "parentheses"
-   qr{ (?<![^[:space:]]) ‘ ([^‘’]+) ’ (?![^[:space]]) }x,
-   qr{ (?<![^[:space:]]) ` ([^`']+) ' (?![^[:space]]) }x,
-   qr{ (?<![^[:space:]]) (" [^[:space:]] [^"]* ")                 }x,
-   qr{                   (" [^"]* [^[:space:]] ") (?![^[:space]]) }x,
-   qr{ \< ([^<>[:space:]]+) \> }x,
-   qr{ \{ ([^{}[:space:]]+) \} }x,
-   qr{ \[ ([^{}[:space:]]+) \] }x,
-   qr{ \( ([^()[:space:]]+) \) }x,
+   qr{ (?<![^[:space:]])  [`'] ([^`']+)  [`'] (?![^[:space:]]) }x,
+   qr{ (?<![^[:space:]])  ‘ ([^‘’]+)  ’ (?![^[:space:]]) }x,
+   qr{ (?<![^[:space:]])  “ ([^“”]+)  ” (?![^[:space:]]) }x,
+
+   qr{ (?<![^[:space:]]) (' [^[:space:]] [^']* ')                  }x,
+   qr{                   (' [^']* [^[:space:]] ') (?![^[:space:]]) }x,
+   qr{ (?<![^[:space:]]) (` [^[:space:]] [^']* ')                  }x,
+   qr{                   (` [^']* [^[:space:]] ') (?![^[:space:]]) }x,
+   qr{ (?<![^[:space:]]) (" [^[:space:]] [^"]* ")                  }x,
+   qr{                   (" [^"]* [^[:space:]] ") (?![^[:space:]]) }x,
+
+   qr{ \{ ([^\{\}]+) \} }x,
+   qr{ \( ([^\(\)]+) \) }x,
+   qr{ \[ ([^\[\]]+) \] }x,
+   qr{ \< ([^\<\>]+) \> }x,
 
    # urls, just a heuristic
    qr{(
-      (?:https?://|ftp://|news://|mailto:|file://)[ab-zA-Z0-9\-\@;\/?:&=%\$_.+!*\x27(),~]+
-      [ab-zA-Z0-9\-\@;\/?:&=%\$_+!*\x27()~]   # exclude some trailing characters (heuristic)
+      (?:https?://|ftp://|news://|mailto:|file://|\bwww\.)[ab-zA-Z0-9\-\@;\/?:&=%\$_.+!*\x27(),~#]+
+      [ab-zA-Z0-9\-\@;\/?:&=%\$_+*()~]   # exclude some trailing characters (heuristic)
    )}x,
 
    # shell-like argument quoting, basically always matches
@@ -83,17 +92,6 @@
    my @matches;
 
    if ($markofs < $line->l) {
-      # convert markofs form character to UTF-8 offset space
-      {
-         my $prefix = substr $text, 0, $markofs;
-         utf8::encode $prefix;
-         $markofs = length $prefix;
-      }
-
-      # not doing matches in unicode mode helps speed
-      # enourmously here. working in utf-8 should be
-      # equivalent due to the magic of utf-8 encoding.
-      utf8::encode $text;
       study $text; # _really_ helps, too :)
 
       for my $regex (@mark_patterns, @{ $self->{patterns} }) {
@@ -123,18 +121,6 @@
 
       next if $len <= $curlen;
 
-      # convert back from UTF-8 offset space to character space
-      {
-         my $length = substr "$text ", $ofs, $len;
-         utf8::decode $length;
-         $len = length $length;
-      }
-      {
-         my $prefix = substr $text, 0, $ofs;
-         utf8::decode $prefix;
-         $ofs = length $prefix;
-      }
-
       $self->selection_beg ($line->coord_of ($ofs));
       $self->selection_end ($line->coord_of ($ofs + $len));
       return 1;