Fix polyline-based distance calculation if stops are present more than once

		return ( 0, 0, $distance_beeline );

	}

	my %seen;

	for my $stop ( @{$route_ref} ) {

		$seen{ $stop->[1] } //= 1;

		$stop->[2]{n} = $seen{ $stop->[1] };

		$seen{ $stop->[1] } += 1;

	}

	# Assumption: polyline entries are always [lat, lon] or [lat, lon, stop ID]

	%seen = undef;

	for my $entry ( @{$polyline_ref} ) {

		if ( $entry->[2] ) {

			$seen{ $entry->[2] } //= 1;

			$entry->[3] = $seen{ $entry->[2] };

			$seen{ $entry->[2] } += 1;

		}

	}

	$journey->{route_dep_index} = $route_start;

	$journey->{route_arr_index} = $route_end;

	my @route = @{$route_ref}[ $route_start .. $route_end ];

	if (

		@route < 2

		or ( $route[-1][0] ne $to

			and ( not $route[-1][1] or $route[-1][1] != $to_eva ) )

	  )

	{

		# I AM ERROR

		return ( 0, 0, $distance_beeline );

	# Just like the route, the polyline may contain the same stop more than

	# once. So we need to select based on the seen counter.

	my @polyline = after_incl {

		$_->[2] and $_->[2] == $from_eva and $_->[3] == $route[0][2]{n}

	}

	# TODO this always selects the first stop entry.

	# Gotta count how many duplicates we need to skip...

	# Approach: build dict route id -> polyline id of corresponding stop

	# (and bail out if route stops ≠ polyline stops),

	# then find indexes from route and select polyline range from those

	my @polyline = after_incl { $_->[2] and $_->[2] == $from_eva }

	@{ $polyline_ref // [] };

	@polyline

	  = before_incl { $_->[2] and $_->[2] == $to_eva } @polyline;

	@polyline = before_incl {

		$_->[2] and $_->[2] == $to_eva and $_->[3] == $route[-1][2]{n}

	}

	@polyline;

	# ensure that before_incl matched -- otherwise, @polyline is too long

	if ( @polyline and $polyline[-1][2] == $to_eva ) {