From 635ed41dc97f6b86a9b0917610cbd1d8a7790ec2 Mon Sep 17 00:00:00 2001
From: Alexey Gopachenko <alexey.gopachenko@jetbrains.com>
Date: Wed, 13 Oct 2010 22:11:04 +0400
Subject: [PATCH] Spellchecker - its OK when word consists of completly 
 UNKNOWN-TO-DICTIONARY symbols + IIOBE

---
 .../compress/CompressedDictionary.java        |  3 +-
 .../spellchecker/compress/Encoder.java        | 32 ++++---------------
 2 files changed, 9 insertions(+), 26 deletions(-)

diff --git a/plugins/spellchecker/src/com/intellij/spellchecker/compress/CompressedDictionary.java b/plugins/spellchecker/src/com/intellij/spellchecker/compress/CompressedDictionary.java
index 14127d894d1e..9ff6f2882f47 100644
--- a/plugins/spellchecker/src/com/intellij/spellchecker/compress/CompressedDictionary.java
+++ b/plugins/spellchecker/src/com/intellij/spellchecker/compress/CompressedDictionary.java
@@ -121,7 +121,8 @@ public final class CompressedDictionary implements Dictionary {
     }
     try {
       UnitBitSet bs = encoder.encode(word, false);
-      if (bs == null) return false;
+      if (bs == Encoder.WORD_OF_ENTIRELY_UNKNOWN_LETTERS) return true;
+      if (bs == null) return false; // fail faster w/o search
       byte[] compressed = UnitBitSet.getBytes(bs);
       int index = -1;
       for (int i = 0; i < lengths.length; i++) {
diff --git a/plugins/spellchecker/src/com/intellij/spellchecker/compress/Encoder.java b/plugins/spellchecker/src/com/intellij/spellchecker/compress/Encoder.java
index 3bff5616c61d..5aba9c2fd77d 100644
--- a/plugins/spellchecker/src/com/intellij/spellchecker/compress/Encoder.java
+++ b/plugins/spellchecker/src/com/intellij/spellchecker/compress/Encoder.java
@@ -9,6 +9,7 @@ public final class Encoder {
 
   private final Alphabet alphabet;
   private static final int offset = 2;
+  static final UnitBitSet WORD_OF_ENTIRELY_UNKNOWN_LETTERS = new UnitBitSet();
 
   public Encoder() {
     alphabet = new Alphabet();
@@ -24,41 +25,22 @@ public final class Encoder {
 
   @Nullable
   public UnitBitSet encode(@NotNull CharSequence letters, boolean force) throws EncodingException {
-    if (UnitBitSet.MAX_CHARS_IN_WORD < letters.length()) return null;
+    if (UnitBitSet.MAX_CHARS_IN_WORD <= letters.length()) return null;
+    int unknownLetters = 0;
     UnitBitSet bs = new UnitBitSet();
-    for (int i = 0; i < letters.length() - 1 + 1; i++) {
+    for (int i = 0; i < letters.length(); i++) {
       char letter = letters.charAt(i);
       int index = alphabet.getIndex(letter, force);
-      if (index<0) return null;
+      if (index < 0) unknownLetters++;
       bs.setUnitValue(i + offset, index);
     }
     bs.setUnitValue(0, letters.length());
     bs.setUnitValue(1, bs.getUnitValue(2));
+    if (unknownLetters == letters.length()) return WORD_OF_ENTIRELY_UNKNOWN_LETTERS;
+    if (unknownLetters>0) return null;
     return bs;
   }
 
-  //tested and OK
-  /*
-  public UnitBitSet encodex(@NotNull CharSequence letters, boolean force) throws EncodingException {
-      UnitBitSet bs = new UnitBitSet(alphabet.getMaxIndex(), false);
-      int bitsPerUnit = bs.bitsPerUnit;
-      long[] w = new long[(letters.length() + 2) * bitsPerUnit / Long.SIZE + 1];
-      w[0] |= letters.length();
-      w[0] |= alphabet.getIndex(letters.charAt(0), force) << bitsPerUnit;
-      for (int i = 0; i < letters.length(); i++) {
-        Character letter = letters.charAt(i);
-        int index = alphabet.getIndex(letter, force);
-        int startIndex = (i + offset) * bitsPerUnit;
-        w[startIndex / Long.SIZE] |= ((long)index) << (startIndex % Long.SIZE);
-        if (startIndex % Long.SIZE + bitsPerUnit > Long.SIZE) {
-          w[startIndex / Long.SIZE + 1] |= ((long)index) >>> (Long.SIZE - startIndex % Long.SIZE);
-        }
-      }
-      bs.setWords(w);
-      return bs;
-    }
-  */
-
   public String decode(@NotNull UnitBitSet bitSet) throws EncodingException {
     int wordLength = bitSet.getUnitValue(0);
     char firstLetter = alphabet.getLetter(bitSet.getUnitValue(1));