public void Pack(byte[] rawBytes, IPackerOutput packerOutput, Object consumerContext)
{
var hash = new PrefixHash(rawBytes, false);
int dictLen = dictionary.Length;
int previousMatchIndex = 0;
int previousMatchLength = 0;
int curr, count;
for (curr = 0, count = rawBytes.Length; curr < count; curr++)
{
int bestMatchIndex = 0;
int bestMatchLength = 0;
if (curr + PrefixHash.PrefixLength - 1 < count)
{
PrefixHash.Match match = dictHash.GetBestMatch(curr, rawBytes);
bestMatchIndex = match.BestMatchIndex;
bestMatchLength = match.BestMatchLength;
match = hash.GetBestMatch(curr, rawBytes);
// Note the >= because we prefer a match that is nearer (and a match
// in the string being compressed is always closer than one from the dict).
if (match.BestMatchLength >= bestMatchLength)
{
bestMatchIndex = match.BestMatchIndex + dictLen;
bestMatchLength = match.BestMatchLength;
}
hash.Put(curr);
}
if (bestMatchLength < MinimumMatchLength)
{
bestMatchIndex = bestMatchLength = 0;
}
if (previousMatchLength > 0 && bestMatchLength <= previousMatchLength)
{
// We didn't get a match or we got one and the previous match is better
packerOutput.EncodeSubstring(-(curr + dictLen - 1 - previousMatchIndex), previousMatchLength, consumerContext);
// Make sure locations are added for the match. This allows repetitions to always
// encode the same relative locations which is better for compressing the locations.
int endMatch = curr - 1 + previousMatchLength;
curr++;
while (curr < endMatch && curr + PrefixHash.PrefixLength < count)
{
hash.Put(curr);
curr++;
}
curr = endMatch - 1; // Make sure 'curr' is pointing to the last processed byte so it is at the right place in the next iteration
previousMatchIndex = previousMatchLength = 0;
}
else if (previousMatchLength > 0 && bestMatchLength > previousMatchLength)
{
// We have a match, and we had a previous match, and this one is better.
previousMatchIndex = bestMatchIndex;
previousMatchLength = bestMatchLength;
packerOutput.EncodeLiteral(rawBytes[curr - 1], consumerContext);
}
else if (bestMatchLength > 0)
{
// We have a match, but no previous match
previousMatchIndex = bestMatchIndex;
previousMatchLength = bestMatchLength;
}
else if (bestMatchLength == 0 && previousMatchLength == 0)
{
// No match, and no previous match.
packerOutput.EncodeLiteral(rawBytes[curr], consumerContext);
}
}
packerOutput.EndEncoding(consumerContext);
}
public void Pack(byte[] rawBytes, IPackerOutput packerOutput, Object consumerContext)
{
var hash = new PrefixHash(rawBytes, false);
int dictLen = dictionary.Length;
int previousMatchIndex = 0;
int previousMatchLength = 0;
int curr, count;
for (curr = 0, count = rawBytes.Length; curr < count; curr++)
{
int bestMatchIndex = 0;
int bestMatchLength = 0;
if (curr + PrefixHash.PrefixLength - 1 < count)
{
PrefixHash.Match match = dictHash.GetBestMatch(curr, rawBytes);
bestMatchIndex = match.BestMatchIndex;
bestMatchLength = match.BestMatchLength;
match = hash.GetBestMatch(curr, rawBytes);
// Note the >= because we prefer a match that is nearer (and a match
// in the string being compressed is always closer than one from the dict).
if (match.BestMatchLength >= bestMatchLength)
{
bestMatchIndex = match.BestMatchIndex + dictLen;
bestMatchLength = match.BestMatchLength;
}
hash.Put(curr);
}
if (bestMatchLength < MinimumMatchLength)
{
bestMatchIndex = bestMatchLength = 0;
}
if (previousMatchLength > 0 && bestMatchLength <= previousMatchLength)
{
// We didn't get a match or we got one and the previous match is better
packerOutput.EncodeSubstring(-(curr + dictLen - 1 - previousMatchIndex), previousMatchLength, consumerContext);
// Make sure locations are added for the match. This allows repetitions to always
// encode the same relative locations which is better for compressing the locations.
int endMatch = curr - 1 + previousMatchLength;
curr++;
while (curr < endMatch && curr + PrefixHash.PrefixLength < count)
{
hash.Put(curr);
curr++;
}
curr = endMatch - 1; // Make sure 'curr' is pointing to the last processed byte so it is at the right place in the next iteration
previousMatchIndex = previousMatchLength = 0;
}
else if (previousMatchLength > 0 && bestMatchLength > previousMatchLength)
{
// We have a match, and we had a previous match, and this one is better.
previousMatchIndex = bestMatchIndex;
previousMatchLength = bestMatchLength;
packerOutput.EncodeLiteral(rawBytes[curr - 1], consumerContext);
}
else if (bestMatchLength > 0)
{
// We have a match, but no previous match
previousMatchIndex = bestMatchIndex;
previousMatchLength = bestMatchLength;
}
else if (bestMatchLength == 0 && previousMatchLength == 0)
{
// No match, and no previous match.
packerOutput.EncodeLiteral(rawBytes[curr], consumerContext);
}
}
packerOutput.EndEncoding(consumerContext);
}
