--- rxvt-unicode/src/perl/selection 2006/01/12 23:11:23 1.24 +++ rxvt-unicode/src/perl/selection 2006/01/17 16:53:47 1.30 @@ -14,6 +14,7 @@ for (my $idx = 0; defined (my $res = $self->x_resource ("selection.pattern-$idx")); $idx++) { no re 'eval'; # just to be sure + $res = utf8::encode $self->locale_decode ($res); push @{ $self->{patterns} }, qr/$res/; } @@ -23,11 +24,14 @@ # "find interetsing things"-patterns my @mark_patterns = ( # common types of "parentheses" - qr{(?:^|\s) ‘ ([^‘’]+?) ’ (?:\s|\)|$)}x, - qr{(?:^|\s) ` ([^`']+?) ' (?:\s|\)|$)}x, - qr{ \{ ([^{}]+?) \} }x, - qr{ \[ ([^{}]+?) \] }x, - qr{ \( ([^()]+?) \) }x, + qr{ (?[:space:]]+) \> }x, + qr{ \{ ([^{}[:space:]]+) \} }x, + qr{ \[ ([^{}[:space:]]+) \] }x, + qr{ \( ([^()[:space:]]+) \) }x, # urls, just a heuristic qr{( @@ -64,14 +68,18 @@ my @matches; + # 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 = $1; - push @matches, [$ofs, length $match]; - for my $regex (@simplify_patterns) { if ($match =~ $regex) { $match = $1;