From 77be8178483003a4caa6079a10f2aba2d38eef09 Mon Sep 17 00:00:00 2001
From: Ben East <benjamin-east@uiowa.edu>
Date: Mon, 3 Jul 2017 15:57:17 -0500
Subject: [PATCH 1/3] [Issue #16] Updated process colon to solve
 StringIndexOutOfBoundsException when input string contains the sub-string
 "//:@."

---
 .../com/linkedin/urls/detection/UrlDetector.java | 16 ++++++++++------
 1 file changed, 10 insertions(+), 6 deletions(-)

diff --git a/url-detector/src/main/java/com/linkedin/urls/detection/UrlDetector.java b/url-detector/src/main/java/com/linkedin/urls/detection/UrlDetector.java
index 43815f6..1098ba9 100644
--- a/url-detector/src/main/java/com/linkedin/urls/detection/UrlDetector.java
+++ b/url-detector/src/main/java/com/linkedin/urls/detection/UrlDetector.java
@@ -154,7 +154,6 @@ private void readDefault() {
     while (!_reader.eof()) {
       //read the next char to process.
       char curr = _reader.read();
-
       switch (curr) {
         case ' ':
           //space was found, check if it's a valid single level domain.
@@ -188,7 +187,8 @@ private void readDefault() {
         case '\uFF61':
         case '.': //"." was found, read the domain name using the start from length.
           _buffer.append(curr);
-          readDomainName(_buffer.substring(length));
+          String substring = _buffer.substring(length);
+          readDomainName(substring);
           length = 0;
           break;
         case '@': //Check the domain name after a username
@@ -277,10 +277,15 @@ private void readDefault() {
   private int processColon(int length) {
     if (_hasScheme) {
       //read it as username/password if it has scheme
-      if (!readUserPass(length) && _buffer.length() > 0) {
+      if (!readUserPass(length)) {
         //unread the ":" so that the domain reader can process it
         _reader.goBack();
-        _buffer.delete(_buffer.length() - 1, _buffer.length());
+        
+        if (_buffer.length() > 0) {
+        	_buffer.delete(_buffer.length() - 1, _buffer.length());
+        } else {
+        	length = 0;
+        }
 
         int backtrackOnFail = _reader.getPosition() - _buffer.length() + length;
         if (!readDomainName(_buffer.substring(length))) {
@@ -470,10 +475,9 @@ private boolean readScheme() {
    * @return True if a valid username and password was found.
    */
   private boolean readUserPass(int beginningOfUsername) {
-
     //The start of where we are.
     int start = _buffer.length();
-
+    
     //keep looping until "done"
     boolean done = false;
 

From 9f267214885c6f82fad0915ddb42db33fbddccd2 Mon Sep 17 00:00:00 2001
From: Ben East <benjamin-east@uiowa.edu>
Date: Mon, 3 Jul 2017 15:59:19 -0500
Subject: [PATCH 2/3] [Issue #16] Updated process colon to solve
 StringIndexOutOfBoundsException when input string contains the sub-string
 "//:@."

---
 .../src/main/java/com/linkedin/urls/detection/UrlDetector.java | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/url-detector/src/main/java/com/linkedin/urls/detection/UrlDetector.java b/url-detector/src/main/java/com/linkedin/urls/detection/UrlDetector.java
index 1098ba9..f85873c 100644
--- a/url-detector/src/main/java/com/linkedin/urls/detection/UrlDetector.java
+++ b/url-detector/src/main/java/com/linkedin/urls/detection/UrlDetector.java
@@ -187,8 +187,7 @@ private void readDefault() {
         case '\uFF61':
         case '.': //"." was found, read the domain name using the start from length.
           _buffer.append(curr);
-          String substring = _buffer.substring(length);
-          readDomainName(substring);
+          readDomainName(_buffer.substring(length));
           length = 0;
           break;
         case '@': //Check the domain name after a username

From d405160c26b3e041fc3478ff09b6784dba3ef7bc Mon Sep 17 00:00:00 2001
From: Ben East <benjamin-east@uiowa.edu>
Date: Mon, 3 Jul 2017 16:42:06 -0500
Subject: [PATCH 3/3] [Issue #12] Fixed StringIndexOutOfBoundsException when
 given 'http://user:pass@host.com host.com' as input string

Updated url-detector.processColon to set length to 0 when readUserPass(length) is true.
All unit tests passed after this update.
---
 .../main/java/com/linkedin/urls/detection/UrlDetector.java | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/url-detector/src/main/java/com/linkedin/urls/detection/UrlDetector.java b/url-detector/src/main/java/com/linkedin/urls/detection/UrlDetector.java
index f85873c..1b6bc10 100644
--- a/url-detector/src/main/java/com/linkedin/urls/detection/UrlDetector.java
+++ b/url-detector/src/main/java/com/linkedin/urls/detection/UrlDetector.java
@@ -280,10 +280,11 @@ private int processColon(int length) {
         //unread the ":" so that the domain reader can process it
         _reader.goBack();
         
+        // Check buffer length before clearing it; set length to 0 if buffer is empty
         if (_buffer.length() > 0) {
-        	_buffer.delete(_buffer.length() - 1, _buffer.length());
+          _buffer.delete(_buffer.length() - 1, _buffer.length());
         } else {
-        	length = 0;
+          length = 0;
         }
 
         int backtrackOnFail = _reader.getPosition() - _buffer.length() + length;
@@ -293,6 +294,8 @@ private int processColon(int length) {
           readEnd(ReadEndState.InvalidUrl);
         }
         length = 0;
+      } else {
+    	length = 0;
       }
     } else if (readScheme() && _buffer.length() > 0) {
       _hasScheme = true;