X-Git-Url: http://git.shiar.net/perl/list-index.git/blobdiff_plain/e269a677a2adb49fc79e6c2267551a716a55c559..1c9de66914ea7bf8525df6f5285b649c006477b9:/lib/List/Index.pm

diff --git a/lib/List/Index.pm b/lib/List/Index.pm
index 89fdc75..dca0d1f 100644
--- a/lib/List/Index.pm
+++ b/lib/List/Index.pm
@@ -6,7 +6,7 @@ use warnings;
 
 use Exporter 'import';
 
-our $VERSION = '1.00';
+our $VERSION = '1.01';
 our @EXPORT_OK = qw(rangematch);
 
 sub new {
@@ -23,31 +23,53 @@ sub ranges {
 	my $pages    = $options->{pages   } || 1 + int $#$self / $pagesize;
 
 	$pagesize = @$self / $pages;
-	my $offset = 0;
-	my @links;
+	my $offset = $pagesize;
+	my @links = ('');
 	while ($offset < @$self) {
 		my $link = substr $self->[$offset], 0, $length;
 		if ($context) {
-			my $trim = 1;
-			my $before = $offset > $context ? $self->[$offset - $context] : '';
-			for my $match (split //, $before) {
-				scalar $link =~ /\G\Q$match/g or last;
-				$trim++;
+			{
+				# take a value slightly before the current offset
+				my $before = $offset > $context ? $self->[$offset - $context] : '.';
+				# see how much of it matches the current link
+				my $trim = 1;
+				for my $match (split //, $before) {
+					scalar $link =~ /\G\Q$match/g or last;
+					$trim++;
+				}
+				# truncate link upto where the earlier value starts to differ
+				substr($link, $trim) = '' unless $trim > length $link;
+			}
+
+			if ($offset + $context < $#$self) {
+				# take a value after the current offset
+				my $after = $self->[$offset + $context];
+				# see how much of it matches the current link
+				my $trim = 1;
+				for my $match (split //, $after) {
+					scalar $link =~ /\G\Q$match/g or last;
+					$trim++;
+				}
+				# use this link if it's shorter
+				if ($trim < length $link) {
+					$link = substr $after, 0, $trim;
+				}
 			}
-			substr($link, $trim) = '' unless $trim > length $link;
 		}
 
-		push @links, [$link];
+		push @links, $link;
 		$offset += $pagesize;
 	}
 
+	use List::MoreUtils 'uniq';
+	@links = uniq @links;
 	for my $i (0 .. $#links - 1) {
-		my ($link, $lastchar) = $links[$i + 1]->[0] =~ /(.*)(.)/;
-		$link .= $lastchar eq '.' ? 'z' : chr( ord($lastchar) - 1 )
-			unless $lastchar eq 'a';
-		$links[$i]->[1] = $link;
+		my ($link, $lastchar) = $links[$i + 1] =~ /(.*)(.)/;
+		$link .= $lastchar le 'a' ? '.' : chr( ord($lastchar) - 1 );
+		next if $link eq $links[$i] and $i;
+		$links[$i] .= '-'.$link;
 	}
-	$links[-1]->[1] = '';
+	$links[-1] .= '-';
 
 	return \@links;
 }
@@ -56,36 +78,80 @@ sub rangematch {
 	my ($link) = @_;
 	my ($s1, $s2) = $link =~ /([^-]*) - ([^-]*)/x
 		or return qr/^\Q$link/i;
+	$s1 =~ s/\.$//;
 	my @allow;
 
 	if (length $s1) {
+		if (length $s2) {
+			$s1 le $s2 or $s1 =~ /^\Q$s2/ or return undef;
+		}
+
 		my $prefix = '';
-		my $c1;
+		my $char;
 		for my $i (0 .. length($s1) - 1) {
-			$c1 = substr $s1, $i, 1;
-			my $c2 = length $s2 <= $i ? undef : substr $s2, $i, 1;
-			my $next = $i + 1 >= length($s1) ? $c1 : chr( ord($c1) + 1 );
-			my $last = defined $c2 && $i == 0 ? chr( ord($c2) - (length $s2 > 1) ) : 'z';
-			$next le $last or next if defined $c2;
+			my $lasti = $i == length($s1) - 1;
+			$char = substr $s1, $i, 1;
+			my $next = $char;
+			# do not include prefix character in final range
+			$next = chr( ord($char) + 1 ) unless $lasti;
+
+			my $last = 'z';
+			next if $next gt $last;
+			if (length $s2 > $i) {
+				if ($s2 =~ /^\Q$prefix/) {
+					$last = substr $s2, $i, 1;
+					next if $char eq $last;
+					$last = chr( ord($last) - (length $s2 > 1) );
+					next if $next gt $last;
+				}
+			}
+
+			if ($char eq '.') {
+				if ($last eq 'z') {
+#					push @allow, $prefix if $i and $lasti;
+#					next;
+				}
+#				if ($last eq 'z') {
+#					push @allow, $prefix if $i and $lasti;
+#					next;
+#				}
+				$next = 'a';
+			}
+
 			push @allow, $prefix."[$next-$last]";
 		}
 		continue {
-			$prefix .= $c1;
+			$prefix .= $char eq '.' ? '[^a-z]' : $char;
 		}
 	}
 
 	if (length $s2) {
 		my $prefix = '';
+		my $char;
 		for my $i (0 .. length($s2) - 1) {
-			my $c1 = length $s1 <= $i ? undef : substr $s1, $i, 1;
-			my $c2 = substr $s2, $i, 1;
+			$char = substr $s2, $i, 1;
 			my $last = 'z';
-			push @allow, "$prefix(?![$c2-$last])"
+			if (length $s1 > $i) {
+				my $c1 = substr $s1, $i, 1;
+				if ($s1 =~ /^\Q$prefix/) {
+					next if $c1 le $char;
+				}
+			}
+
+			if ($char eq '.') {
+				next if $i < length($s2) - 1;
+			}
+
+			push @allow, $prefix.'(?!['.($char eq '.' ? 'a' : $char)."-$last])"
 				if $i or $s1 eq '';
-			$prefix .= $c2;
 		}
+		continue {
+			$prefix .= $char eq '.' ? '[^a-z]' : $char;
+		}
+
 		push @allow, $prefix
-			unless length $s1 > length $s2 or length $s1 != 0 && length $s2 == 1; #TODO
+			if $s2 =~ /^\Q$prefix/ and $s1 le $s2
+			and not (length $s2 == 1 && length $s1 >= length $s2 && $s1 ne $s2);
 	}
 
 	my $match = sprintf @allow <= 1 ? '%s' : '(?:%s)', join('|', @allow);
@@ -105,7 +171,11 @@ List::Index - Paginate alphabetic entries by finding minimal prefixes
 	use List::Index;
 	my $index = List::Index->new(\@values);
 	my @pages = $index->ranges({pagesize => 50});
-	printf '<a href="?start=%s&amp;end=%s">%1$s</a> ', @$_ for @pages;
+	printf '<a href="?q=%s-%s">%1$s</a> ', @$_ for @pages;
+
+	use List::Index 'rangematch';
+	my $limit = rangematch('b-bmq');  # matches prefix like 'baa'..'bmq'
+	@results = grep { $limit } @results;
 
 =head1 DESCRIPTION