--- rxvt-unicode/src/perl/selection	2006/01/12 03:32:56	1.20
+++ rxvt-unicode/src/perl/selection	2007/11/19 09:27:08	1.44
@@ -1,6 +1,6 @@
 #! perl
 
-sub on_keyboard_command {
+sub on_user_command {
    my ($self, $cmd) = @_;
 
    $cmd eq "selection:rot13"
@@ -9,24 +9,60 @@
    ()
 }
 
-# "find interetsing things"-patterns
+sub on_init {
+   my ($self) = @_;
+
+   if (defined (my $res = $self->resource ("cutchars"))) {
+      $res = $self->locale_decode ($res);
+      push @{ $self->{patterns} }, qr{\G [\Q$res\E[:space:]]* ([^\Q$res\E[:space:]]+) }x;
+   }
+
+   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/;
+   }
+
+   $self->{enabled} = 1;
+
+   push @{ $self->{term}{option_popup_hook} }, sub {
+      ("new selection" => $self->{enabled}, sub { $self->{enabled} = shift })
+   };
+
+   ()
+}
+
+# "find interesting things"-patterns
 my @mark_patterns = (
-   qr{([[:alnum:]]+)},
+   qr{ ([[:word:]]+) }x,
+   qr{ ([^[:space:]]+) }x,
 
-   # common "parentheses"
-   qr{(?:^|\s) ‘  ([^‘’]+?) ’ (?:\s|\)|$)}x,
-   qr{(?:^|\s) `  ([^`']+?) ' (?:\s|\)|$)}x,
-   qr{         \{ ([^{}]+?) \} }x,
-   qr{         \[ ([^{}]+?) \] }x,
+   # common types of "parentheses"
+   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
-   qr{\G [\ \t|&;<>()] *(
+   qr{\G [\ \t|&;<>()]* (
       (?:
          [^\\"'\ \t|&;<>()]+
          | \\.
@@ -43,7 +79,10 @@
 );
 
 sub on_sel_extend {
-   my ($self) = @_;
+   my ($self, $time) = @_;
+
+   $self->{enabled}
+      or return;
 
    my ($row, $col) = $self->selection_mark;
    my $line = $self->line ($row);
@@ -54,39 +93,63 @@
 
    my @matches;
 
-   for my $regex (@mark_patterns) {
-      while ($text =~ /$regex/g) {
-         if ($-[1] <= $markofs and $markofs <= $+[1]) {
-            my $ofs = $-[1];
-            my $match = $1;
-
-            push @matches, [$ofs, length $match];
-
-            for my $regex (@simplify_patterns) {
-               if ($match =~ $regex) {
-                  $match = $1;
-                  $ofs += $-[1];
+   if ($markofs < $line->l) {
+      # convert markofs from 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, @{ $self->{patterns} }) {
+         while ($text =~ /$regex/g) {
+            if ($-[1] <= $markofs and $markofs <= $+[1]) {
+               my $ofs = $-[1];
+               my $match = $1;
+
+               for my $regex (@simplify_patterns) {
+                  if ($match =~ $regex) {
+                     $match = $1;
+                     $ofs += $-[1];
+                  }
                }
-            }
 
-            push @matches, [$ofs, length $match];
+               push @matches, [$ofs, length $match];
+            }
          }
       }
    }
 
+   # whole line
+   push @matches, [0, ($line->end - $line->beg + 1) * $self->ncol];
+
    for (sort { $a->[1] <=> $b->[1] or $b->[0] <=> $a->[0] } @matches) {
       my ($ofs, $len) = @$_;
 
       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;
    }
 
-   # whole line
-   $self->selection_beg ($line->beg, 0);
-   $self->selection_end ($line->end, $self->ncol);
-
-   return 1;
+   ()
 }