From 0e58ea12822623f12b0c3d93e7dd4876d8fa5822 Mon Sep 17 00:00:00 2001
From: Stephen Waits <steve@waits.net>
Date: Sun, 15 Feb 2015 08:07:42 -0800
Subject: [PATCH] bugfix: enable case insensitivity on REs for urls and title

---
 src/docrawler/parse.go      | 5 +++--
 src/docrawler/parse_test.go | 2 +-
 2 files changed, 4 insertions(+), 3 deletions(-)

diff --git a/src/docrawler/parse.go b/src/docrawler/parse.go
index e03a667..13a1010 100644
--- a/src/docrawler/parse.go
+++ b/src/docrawler/parse.go
@@ -7,8 +7,8 @@ import (
 )
 
 // a regexp which captures src/href/xhref="..." from text
-var reURL = regexp.MustCompile(`(src|href|xhref)\s*=\s*"([^"]+)"`) // TODO make RFC compliant
-var reTitle = regexp.MustCompile(`<\s*title\s*>([^<]*)`)
+var reURL = regexp.MustCompile(`(?i)(src|href|xhref)\s*=\s*"([^"]+)"`) // TODO make RFC compliant
+var reTitle = regexp.MustCompile(`(?i)<\s*title\s*>([^<]*)<\s*\/\s*title`)
 
 // parse takes a string and attempts to parse any html title and all links out of it,
 // and returns a sorted slice of the captures found
@@ -17,6 +17,7 @@ func parse(s string) (string, []string, error) {
 	title := ""
 	titleMatches := reTitle.FindStringSubmatch(s)
 	if titleMatches != nil {
+		// we found a title, replace our default title with the result
 		title = strings.TrimSpace(titleMatches[1])
 	}
 
diff --git a/src/docrawler/parse_test.go b/src/docrawler/parse_test.go
index eae5883..9089d55 100644
--- a/src/docrawler/parse_test.go
+++ b/src/docrawler/parse_test.go
@@ -8,7 +8,7 @@ import (
 func TestSimpleParse(t *testing.T) {
 	doc := `<html>
 	<head>
-		<title> Test Page  </title>
+		<TITle> Test Page  </title>
 	</head>
 	<body>
 		<img src="/assets/image.png"/>