From 7b394d2acece3204c956a5c36f1047cbd5da9e6f Mon Sep 17 00:00:00 2001
From: Arne Georg Gleditsch <argggh@lxr.linpro.no>
Date: Thu, 21 Aug 2008 21:37:54 +0200
Subject: Searching tests.

---
 lxr-genxref | 7 +++++++
 1 file changed, 7 insertions(+)

(limited to 'lxr-genxref')

diff --git a/lxr-genxref b/lxr-genxref
index 3173bdd..1642c2d 100755
--- a/lxr-genxref
+++ b/lxr-genxref
@@ -240,6 +240,13 @@ sub hash_file($$$) {
 	    my $pos = 0;
 	    # Latin-1 word characters.
 	    foreach my $term (/([0-9a-zA-Z\300-\326\330-\366\370-\377]+)/g) {
+		# TODO: For foo_bar_zoo_ack, index
+		# - foo_bar_zoo_ack
+		# - foo_bar bar_zoo zoo_ack
+		# - foo bar zoo ack
+		# This enables subcomponent searches without running
+		# into the stopword problem that earlier
+		# reduce-to-phrase approaches suffered from.
 		$term = lc($term);
 		next if length($term) > 128;
 		$doc->add_posting($term, $.*100 + $pos++);
-- 
cgit v1.2.3