From bc9543fe181147a2b12f0267a7a3aded3a97e29c Mon Sep 17 00:00:00 2001
From: Sebastian Nagel <snagel@apache.org>
Date: Tue, 5 Jul 2016 15:09:59 +0200
Subject: [PATCH] Fix StringIndexOutOfBoundsException in patternCSSExtract  -
 correct check for min. required URL lenght when stripping 4 characters (2 at
 each end)  - simplified code, use non-capturing groups in regular expression

---
 .../html/ExtractingParseObserver.java         | 31 ++++++-------------
 .../html/ExtractingParseObserverTest.java     |  3 +-
 2 files changed, 12 insertions(+), 22 deletions(-)

diff --git a/src/main/java/org/archive/resource/html/ExtractingParseObserver.java b/src/main/java/org/archive/resource/html/ExtractingParseObserver.java
index e1f57b55..a29744c8 100644
--- a/src/main/java/org/archive/resource/html/ExtractingParseObserver.java
+++ b/src/main/java/org/archive/resource/html/ExtractingParseObserver.java
@@ -23,7 +23,7 @@ public class ExtractingParseObserver implements ParseObserver {
 	protected static String cssUrlPatString = 
 		"url\\s*\\(\\s*([\\\\\"']*.+?[\\\\\"']*)\\s*\\)";
 	protected static String cssImportNoUrlPatString = 
-		"@import\\s+(('[^']+')|(\"[^\"]+\")|(\\('[^']+'\\))|(\\(\"[^\"]+\"\\))|(\\([^)]+\\))|([a-z0-9_.:/\\\\-]+))\\s*;";
+		"@import\\s+((?:'[^']+')|(?:\"[^\"]+\")|(?:\\('[^']+'\\))|(?:\\(\"[^\"]+\"\\))|(?:\\([^)]+\\))|(?:[a-z0-9_.:/\\\\-]+))\\s*;";
 
 	protected static Pattern cssImportNoUrlPattern = Pattern
 			.compile(cssImportNoUrlPatString);
@@ -372,36 +372,25 @@ private void patternCSSExtract(HTMLMetaData data, Pattern pattern, String conten
 		Matcher m = pattern.matcher(content);
 		int idx = 0;
 		int contentLen = content.length();
-		while((idx < contentLen) && m.find(idx)) {
+		while((idx < contentLen) && m.find()) {
+			idx = m.end();
 			String url = m.group(1);
-			int origUrlLength = url.length();
-			int urlStart = m.start(1);
-			int urlEnd = m.end(1);
-			idx = urlEnd;
 			if(url.length() < 2) {
 				continue;
 			}
 			if ((url.charAt(0) == '(') 
-					&& (url.charAt(origUrlLength-1) == ')')) {
-				url = url.substring(1, origUrlLength - 1);
-				urlStart += 1;
-				origUrlLength -= 2;
+					&& (url.charAt(url.length()-1) == ')')) {
+				url = url.substring(1, url.length() - 1);
 			}
-			if (url.charAt(0) == '"') {
-				url = url.substring(1, origUrlLength - 1);
-				urlStart += 1;
-			} else if (url.charAt(0) == '\'') {
-				url = url.substring(1, origUrlLength - 1);
-				urlStart += 1;
+			if (url.charAt(0) == '"' || url.charAt(0) == '\'') {
+				url = url.substring(1, url.length() - 1);
 			} else if (url.charAt(0) == '\\') {
-				if(url.length() == 2)
+				if(url.length() <= 4) {
 					continue;
-				url = url.substring(2, origUrlLength - 2);
-				urlStart += 2;
+				}
+				url = url.substring(2, url.length() - 2);
 			}
-			int urlLength = url.length();
 			data.addHref("path","STYLE/#text","href",url);
-			idx += urlLength;
 		}
 	}
 }
diff --git a/src/test/java/org/archive/resource/html/ExtractingParseObserverTest.java b/src/test/java/org/archive/resource/html/ExtractingParseObserverTest.java
index 24b6c18a..7d8c7ea8 100644
--- a/src/test/java/org/archive/resource/html/ExtractingParseObserverTest.java
+++ b/src/test/java/org/archive/resource/html/ExtractingParseObserverTest.java
@@ -19,7 +19,8 @@ public void testHandleStyleNodeExceptions() throws Exception {
 				"url (' ')",
 				"url('\")",
 				"url(')",
-				"url('\"')"
+				"url('\"')",
+				"url('\\\"\"')"
 		};
 		boolean except = false;
 		HTMLMetaData md = new HTMLMetaData(new MetaData());