Imported nxcomp-3.1.0-4.tar.gznxcomp/3.1.0-4

Summary: Imported nxcomp-3.1.0-4.tar.gz
Keywords:

Imported nxcomp-3.1.0-4.tar.gz
into Git repository

1 files changed, 77 insertions, 0 deletions

diff --git a/nxcomp/TextCompressor.cpp b/nxcomp/TextCompressor.cpp
new file mode 100755
index 000000000..71eb23a9f
--- /dev/null
+++ b/nxcomp/TextCompressor.cpp
@@ -0,0 +1,77 @@
+/**************************************************************************/
+/*                                                                        */
+/* Copyright (c) 2001, 2007 NoMachine, http://www.nomachine.com/.         */
+/*                                                                        */
+/* NXCOMP, NX protocol compression and NX extensions to this software     */
+/* are copyright of NoMachine. Redistribution and use of the present      */
+/* software is allowed according to terms specified in the file LICENSE   */
+/* which comes in the source distribution.                                */
+/*                                                                        */
+/* Check http://www.nomachine.com/licensing.html for applicability.       */
+/*                                                                        */
+/* NX and NoMachine are trademarks of NoMachine S.r.l.                    */
+/*                                                                        */
+/* All rights reserved.                                                   */
+/*                                                                        */
+/**************************************************************************/
+
+#include "TextCompressor.h"
+
+#include "EncodeBuffer.h"
+#include "DecodeBuffer.h"
+
+//
+// The compression obtained by this class is
+// very poor. In newer versions the text is
+// simply appended to the encode buffer and
+// compressed by leveraging the final stream
+// compression.
+//
+
+void
+TextCompressor::encodeChar(unsigned char ch, EncodeBuffer& encodeBuffer)
+{
+  // encode each successive character of text using
+  // a predictive model where most of the last 3 characters
+  // (low order 7 bits of the previous character, plus the
+  // low order 5 bits of the character before that, plus
+  // the low order 3 bits of the character before that)
+  // are used to find the right cache...
+
+  CharCache& cache =  cache_[key_ % cacheSize_];
+  if ((key_ >= 128) && (cache.getSize() == 0))
+  {
+    // 3rd-order model doesn't have any statistics yet,
+    // so use the 1st-order one instead
+    CharCache& cache2 =  cache_[(key_ & 0x7f) % cacheSize_];
+    encodeBuffer.encodeCachedValue((unsigned int) ch, 8, cache2);
+    cache.insert(ch);
+  }
+  else
+  {
+    encodeBuffer.encodeCachedValue((unsigned int) ch, 8, cache);
+  }
+
+  key_ = (((key_ & 0x1f) << 7) | ((key_ & 0x380) << 5) | (ch & 0x7f));
+}
+
+
+unsigned char
+TextCompressor::decodeChar(DecodeBuffer& decodeBuffer)
+{
+  unsigned char nextChar;
+  CharCache& cache =  cache_[key_ % cacheSize_];
+  if ((key_ >= 128) && (cache.getSize() == 0))
+  {
+    CharCache& cache2 =  cache_[(key_ & 0x7f) % cacheSize_];
+    decodeBuffer.decodeCachedValue(nextChar, 8, cache2);
+    cache.insert(nextChar);
+  }
+  else
+  {
+    decodeBuffer.decodeCachedValue(nextChar, 8, cache);
+  }
+
+  key_ = (((key_ & 0x1f) << 7) | ((key_ & 0x380) << 5) | (nextChar & 0x7f));
+  return nextChar;
+}