From 035d116c8bd48483681251aebb3277ebb3ecf698 Mon Sep 17 00:00:00 2001
From: Doug Felt <dougfelt@gmail.com>
Date: Mon, 10 Apr 2006 21:54:40 +0000
Subject: [PATCH] ICU-5156 perf enhancements

X-SVN-Rev: 19526
---
 .../src/com/ibm/icu/util/StringTokenizer.java | 79 +++++++++++++++----
 1 file changed, 62 insertions(+), 17 deletions(-)

diff --git a/icu4j/src/com/ibm/icu/util/StringTokenizer.java b/icu4j/src/com/ibm/icu/util/StringTokenizer.java
index 00fcfc584bd..5418443af5d 100644
--- a/icu4j/src/com/ibm/icu/util/StringTokenizer.java
+++ b/icu4j/src/com/ibm/icu/util/StringTokenizer.java
@@ -233,6 +233,7 @@ public final class StringTokenizer implements Enumeration
         if (delim != null && delim.length() > 0) {
             m_delimiters_ = new UnicodeSet();
             m_delimiters_.addAll(delim);
+            checkDelimiters();
         }
         m_coalesceDelimiters_ = coalescedelims;
         m_source_ = str;
@@ -318,7 +319,10 @@ public final class StringTokenizer implements Enumeration
             if (m_returnDelimiters_) {
                 int tokenlimit = 0;
                 int c = UTF16.charAt(m_source_, m_nextOffset_);
-                if (m_delimiters_.contains(c)) {
+                boolean contains = delims == null 
+                    ? m_delimiters_.contains(c) 
+                    : c < delims.length && delims[c];
+                if (contains) {
                      if (m_coalesceDelimiters_) {
                         tokenlimit = getNextNonDelimiter(m_nextOffset_);
                      } else {
@@ -417,6 +421,7 @@ public final class StringTokenizer implements Enumeration
     public String nextToken(UnicodeSet delim) 
     {
         m_delimiters_ = delim;
+        checkDelimiters();
         m_tokenOffset_ = -1;
         m_tokenSize_ = -1;
         if (!m_returnDelimiters_) {
@@ -490,8 +495,11 @@ public final class StringTokenizer implements Enumeration
                 }
                 m_tokenStart_[result] = m_nextOffset_;
                 if (m_returnDelimiters_) {
-                    if (m_delimiters_.contains(UTF16.charAt(m_source_, 
-                                                            m_nextOffset_))) {
+                    int c = UTF16.charAt(m_source_, m_nextOffset_);
+                    boolean contains = delims == null 
+                        ? m_delimiters_.contains(c) 
+                        : c < delims.length && delims[c];
+                    if (contains) {
                         m_tokenLimit_[result] = getNextNonDelimiter(
                                                                 m_nextOffset_);
                     }
@@ -592,13 +600,23 @@ public final class StringTokenizer implements Enumeration
         if (offset >= 0) {
             int result = offset; 
             int c = 0;
-            do {
-                c = UTF16.charAt(m_source_, result);
-                if (m_delimiters_.contains(c)) {
-                    break;
-                }
-                result ++;
-            } while (result < m_length_);
+            if (delims == null) {
+                do {
+                    c = UTF16.charAt(m_source_, result);
+                    if (m_delimiters_.contains(c)) {
+                        break;
+                    }
+                    result ++;
+                } while (result < m_length_);
+            } else {
+                do {
+                    c = UTF16.charAt(m_source_, result);
+                    if (c < delims.length && delims[c]) {
+                        break;
+                    }
+                    result ++;
+                } while (result < m_length_);
+            }                
             if (result < m_length_) {
                 return result;
             }
@@ -618,17 +636,44 @@ public final class StringTokenizer implements Enumeration
         if (offset >= 0) {
             int result = offset; 
             int c = 0;
-            do {
-                c = UTF16.charAt(m_source_, result);
-                if (!m_delimiters_.contains(c)) {
-                    break;
-                }
-                result ++;
-            } while (result < m_length_);
+            if (delims == null) {
+                do {
+                    c = UTF16.charAt(m_source_, result);
+                    if (!m_delimiters_.contains(c)) {
+                        break;
+                    }
+                    result ++;
+                } while (result < m_length_);
+            } else {
+                do {
+                    c = UTF16.charAt(m_source_, result);
+                    if (!(c < delims.length && delims[c])) {
+                        break;
+                    }
+                    result ++;
+                } while (result < m_length_);
+            }
             if (result < m_length_) {
                 return result;
             }
         }
         return -1 - m_length_;
     }
+
+    void checkDelimiters() {
+        if (m_delimiters_ == null || m_delimiters_.size() == 0) {
+            delims = new boolean[0];
+        } else {
+            int maxChar = m_delimiters_.getRangeEnd(m_delimiters_.getRangeCount()-1);
+            if (maxChar < 0x7f) {
+                delims = new boolean[maxChar+1];
+                for (int i = 0, ch; -1 != (ch = m_delimiters_.charAt(i)); ++i) {
+                    delims[ch] = true;
+                }
+            } else {
+                delims = null;
+            }
+        }
+    }
+    private boolean[] delims;
 }